S11A-0263
Seismic Hazard Characteristics Derived from Ground Motion Simulation
Discrete numerical simulation for wave equation has been widely employed in study on the seismic wave propagation and strong ground motion in heterogeneous media by using the pseudospectral method with staggered grid RFFT differentiation (SGRFFTD) with different source models. Some characteristics of seismic hazard derived from ground motion simulations may benefit the disaster prevention and mitigating hazard. The simulation results show that the geophysical structure strongly affects the distribution of maximum amplitudes of seismic waveforms and the seismic damage. The amplification effects of shallow sedimentary layers vary with the sites at the surface. The influences on ground motions from seismological dynamic characteristics of the medium structure were analyzed. The influences upon the amplitudes of seismic waveforms by sedimentary layer were analyzed employing the soil core samples. Amplification effects of soil core samples of ten drill holes with 40 m deep are quite different among the sites. The maximum rate is 3.123, displaying soil amplified effect. The minimum is 0.864, displaying reduced effect. Such ratios display the nonlinear property of amplification effect of sediments in the heavily damaged zone too. The simulation results of seismic wave propagation in a basin structure using SGRFFTD show that the seismic energy transmitted into a sediment-filled basin from the bedrock may not refract back into the rock region. The seismic energy converted into the energy of the oscillations of the medium within the basin. The seismic wave oscillating up and down within the basin may result in large damage to the buildings. It is a notable characteristic in aseismic research. Ground motion simulation shows that interference between the body wave and the secondary surface wave occurring nearly the basin result in peak ground motions and high collapse ratios of buildings away from the earthquake fault. The interference strongly affects the distribution of maximum amplitudes of seismic waveforms and the seismic damage in the surface and around the basin and dam structure.
S11A-0264
Initiation of a Database of CEUS Ground Motions for NGA East
The Nuclear Regulatory Commission has funded the first stage of development of a database of central and eastern US (CEUS) broadband and accelerograph records, along the lines of the existing Next Generation Attenuation (NGA) database for active tectonic areas. This database will form the foundation of an NGA East project for the development of CEUS ground-motion prediction equations that include the effects of soils. This initial effort covers the development of a database design and the beginning of data collection to populate the database. It also includes some processing for important source parameters (Brune corner frequency and stress drop) and site parameters (kappa, Vs30). Besides collecting appropriate earthquake recordings and information, existing information about site conditions at recording sites will also be gathered, including geology and geotechnical information. The long-range goal of the database development is to complete the database and make it available in 2010. The database design is centered on CEUS ground motion information needs but is built on the Pacific Earthquake Engineering Research Center's (PEER) NGA experience. Documentation from the PEER NGA website was reviewed and relevant fields incorporated into the CEUS database design. CEUS database tables include ones for earthquake, station, component, record, and references. As was done for NGA, a CEUS ground- motion flat file of key information will be extracted from the CEUS database for use in attenuation relation development. A short report on the CEUS database and several initial design-definition files are available at https://umdrive.memphis.edu:443/xythoswfs/webui/_xy-7843974_docstore1. Comments and suggestions on the database design can be sent to the author. More details will be presented in a poster at the meeting.
S11A-0265
Enhanced Ground Motions and Seismic Hazards due to Crustal Structures in Korea
Seismic waves, in general, attenuate with distances they travel, so does the potential seismic hazard. Although damages due to large earthquakes are limited by distances within a few tens of kilometers from the source region, some areas far from the epicenter may experience strong shakings and severe damages. A moderate size earthquake with ML 4.8 occurred in the central eastern Korean Peninsula in the evening of January 20, 2007. The earthquake was the largest inland earthquake ever recorded since the modern Seismic Network has been fully operational in the peninsula and provided unprecedented opportunity to observe strong ground shakings throughout the country. At the closest station from the epicenter, 6.5%g, 13.2%g, 15.6%g were registered on the vertical, east-west, and north-south components, respectively. While the observed PGAs during the earthquake follow the general trend of exponential decay with the distance, some strong motion stations experienced several times larger ground shakings than those expected from the empirically determined attenuation relation. The observed strong ground shakings may be caused by site responses which amplify the wave in the thick basin, directivity due to the rupture processes, dipping crustal structures, a simple crustal structure lacking mid- continental discontinuities, or the combination of above mentioned. After the examination of several potential factors, we propose Moho-reflected S wave, SmS, to be responsible for the enhanced amplitude observed in the strong motion stations at the distance range from 80 to 180 km. Empirical PGA predictions may have underestimated strong ground motions in this distance range. Thus, Moho-reflections should be taken into consideration in the prediction of the strong ground motion and the evaluation of seismic hazard mitigation efforts.
S11A-0266
Loss estimation in southeast Korea from a scenario earthquake using the deterministic method in HAZUS
Strong ground motion attenuation relationship represents a comprehensive trend of ground shakings at sites with distances from the source, geology, local soil conditions, and others. It is necessary to develop an attenuation relationship with careful considerations of characteristics of the target area for reliable seismic hazard/risk assessments. In the study, observed ground motions from the January 2007 magnitude 4.9 Odaesan earthquake and the events occurring in the Gyeongsang provinces are compared with the previously proposed ground attenuation relationships in the Korean Peninsula to select most appropriate one. In the meantime, a few strong ground motion attenuation relationships are proposed and introduced in HAZUS, which have been designed for the Western United States and the Central and Eastern United States. The selected relationship from the ones for the Korean Peninsula has been compared with attenuation relationships available in HAZUS. Then, the attenuation relation for the Western United States proposed by Sadigh et al. (1997) for the Site Class B has been selected for this study. Reliability of the assessment will be improved by using an appropriate attenuation relation. It has been used for the earthquake loss estimation of the Gyeongju area located in southeast Korea using the deterministic method in HAZUS with a scenario earthquake (M=6.7). Our preliminary estimates show 15.6% damage of houses, shelter needs for about three thousands residents, and 75 life losses in the study area for the scenario events occurring at 2 A.M. Approximately 96% of hospitals will be in normal operation in 24 hours from the proposed event. Losses related to houses will be more than 114 million US dollars. Application of the improved methodology for loss estimation in Korea will help decision makers for planning disaster responses and hazard mitigation.
S11A-0267
Natural Hazards As An Indicator Of Post Seismic Activites In The Armenian Upland
In seismically active areas the risk is high due to possible structural destruction and population loss as a result of a strong earthquake impact. Moreover, there are secondary earthquake related hazards such as landslides, avalanches, probable fires, explosions, chemical and radioactive contamination that make the situation worse. For seismically active mountainous countries like Armenia the induced post-seismic activity of landslides is still very important. Post-seismic activation of landslides and extensive deforestation are widely developed in the zone of catastrophic earthquake M=7.0 in Spitak in 1988 which occurred on the north-east of Armenia and destroyed a big number cities, settlements and special and significant facilities. After Spitak 07.12.1988; M=7.0 earthquake, the observers' interest was concentrated on gradually activating landslides that lie within the territory of Pambak-Sevan and Spitak active faults adjoining the Spitak earthquake epicentral zone. The Spitak fault was activated during the earthquake and formed a 37 km long surface rupture. From the tectonic point of view that phenomenon is known for long time. Under the impact of strong earthquake the slops that have "favorable" slipping angle loose their gravitational balance and slide-down destroying everything on their way. The lithological and structural criteria of geodynamic activity of these regions on the boundary of Late Mesozoic – Early Cenozoic are considered. In particular, data are presented on geological structure of the sites of "wild" coarse flysch, olistostrome horizons and intensive folding development. For the reconstruction of geodynamic situation in the region on the Meso-Cenozoic boundary the olistostrome horizons' formation is considered. The spatial relation of contemporary active structures, including the seismogenic ones, (scarps, ledges, ruptures, landslides with the areas of Meso-Cenozoic activity manifestations) is shown and their inherited character is proved. For reliable seismic hazard assessment in young seismo-active folded regions, including the zones of Alpine tectogenesis, along with recognized seismological and seismotectonic criteria it is proposed to use the considered lithological and structural preconditions. On the other hand the task becomes more complicated if we consider it in environmental aspect. During the observations that we have done in previous years we occasionally fixed that in the mentioned zone nearly all landslide bodies had forest cover. It forces us to consider the landslide activity also as a result of extensive deforestation during the post-communistic and transitional period in our country. The comprehensive interpretation of this problem we propose and new results that can provide important ways for landslide disaster risk reduction strategy.
S11A-0268
Reconciling Neotectonic and Seismic Recurrence Rates in SW Australia
The southwest corner of Western Australia is an area of stable continental crust (SCC) with the geology being Protozoic and the area undergoing no tectonic activity in the last 40Ma. However in the last 50 years it has had the highest level of seismicity of any area of worldwide SCC. It not been glaciated in the last 20Ma and has had a cool dry climate for at least 200ka, thus providing an ideal environment to preserve fault scarps. The availability of high resolution digital elevation model data (DEM) in this area has lead to the identification of over 50 new features that are thought to be scarps of surface rupturing earthquakes. Half of these features have been the subject of field work and all are thought to be fault scarps. One feature has been trenched and the presence of a fault scarp verified. Using recently developed fault scaling relations the fault length and displacement are used to estimate magnitude and in some cases identifying multiple events, these scarps have been used to generate a neotectonic earthquake catalogue. Between M6.5 and M7.2 the Log N vs Mag plot of the data has a slope of 1, indicating that the neotectonic derived catalogue is complete above M6.4. Below M6.4 the number of earthquakes rapidly decreases, with 20% of the expected number of M6.2 earthquakes and 3% of M6.0 earthquakes being identified via neotectonic methods. This is likely a combination of the lower likelihood of the earthquakes causing a surface rupture and the smaller rupture more rapidly reducing in size to below the detection threshold. Scarps formed by earthquakes of ≥M7.3 appear over represented with M7.4-7.5 earthquakes being over represented by a factor of perhaps 3 to 6. This could be explained by the longer preservation age of these large scarps. Alternatively, as the majority of these large scarps are in a a small section of the study area which could be considered extended continental crust, they might be representing a separete tectonic regime. The three key results of this study are: 1. Under the right geological and climatological conditions fault scarps from ≥M6.5 earthquakes can be preserved for 100ka or more and scarps from ≥M7.3 earthquakes for perhaps >150ka. With approximately 50% of M6.3 and 100% ≥M6.4 earthquakes in SCC are expected to form scarps. 2. Mmax in stable continental crust is more likely to be M7.3-7.5 than the M7.0-7.2 generally used in hazard studies. If the extended continental crust hypotesis is applied then the Mmax will be more like M7.4-7.5 and M7.3- 7.4 for extended and non-extended SCC respectively. 3. The recurrence rate for this neotectonic catalogue and historical earthquakes in the whole SCC of Australia are similar. However the contemporary level of seismicty in this area is an order of magnitude higher than that needed to generate the scarps. This supports the hypothesis that seismicity in SCC is episodic and migrates on time frames of 1-10ka.
S11A-0269
Site Response and Liquefaction Risk Analysis for Bucharest, Romania
Bucharest, the capital of Romania, with more than 2 million inhabitants, is considered, after Istanbul, the second- most earthquake-endangered metropolis in Europe. Four major earthquakes with moment-magnitudes between 6.9 and 7.7 hit Bucharest in the last 65 years. All disastrous earthquakes are generated within a small epicentral area - the Vrancea region - about 150 km northeast of Bucharest. Thick unconsolidated sedimentary layers in the area of Bucharest amplify the arriving seismic shear waves causing severe destruction. Thus, pertinent site response analysis combined with liquefaction risk analysis for the city area are of highest priority for the disaster prevention and mitigation of earthquake effects. Within the frame of the Collaborative Research Center (CRC) 461: "Strong Earthquakes: A Challenge for Geosciences and Civil Engineering, at the University of Karlsruhe, Germany, recently detailed field investigations for the near-surface soil layers in Bucharest were performed. These include Seismic Cone Penetration Tests (SCPTU) and seismic refraction measurements for shallow depths. SCPTU is used to obtain a detailed distribution of the shear wave velocities and in situ state parameters of soils. The results are used for site response analysis with linear and non-linear wave-propagation models as well as for a liquefaction risk analysis. Commonly linear-equivalent models are used to simulate the response of the soil during earthquakes. Nevertheless, the linear-equivalent model cannot reproduce non-linear effects like liquefaction, layer isolation, and consolidation during shaking. Thus, a 1-D wave propagation model was used to study the influence of non- linear effects in the site response analysis. Cone resistance, sleeve friction and pore water pressure registered continuous by the SCPTU method until depths of 35m permit to determine by simplified empirical methods the factor of safety and the probability of liquefaction for different soil layers at representative sites in Bucharest. To take the depth of the liquefied layers and the thickness of the covering cohesive layer in account, the liquefaction potential index and the liquefaction severity index were computed for different sites. Contour maps for the whole city area were outlined, showing the risk of ground failure at the ground surface during a specific earthquake.
S11A-0270
Study of the Intensity and Acceleration Attenuations in Spain and Their Impact in the Seismic Hazard Determination
In the seismic hazard determination, one of the most crucial steps is the selection of an estimation of the ground motion according to the size and distance of the selected event. Up to now, we in Spain used a set of different attenuation laws on intensity values, depending of the tectonic area in which the event is included. Those attenuation relationships have been obtained using different isoseismal maps for the Iberian Peninsula. However if we perform an analysis of the isoseismal maps used in the different studies following the Bakun & Wenthworth (1999) technique we found a similar coefficient and no regionalization can be established as was done systematically by different authors. Moreover, the range of Q values estimates for the different areas also confirms this conclusion. As in the last years a considerable number of strong ground motion data were recorded in Spain, we obtained a new PGA attenuation relationship for the territory. The relationship is obtained with 273 acceleration values of events of moment magnitude M greater than 3 in the epicentral distance range between 5 and 100 kilometers and is compared with other relations for similar tectonic environments. Finally, we have calculated the seismic hazard of Spain using both types of data, intensity values and intensity attenuation and magnitude values and the acceleration attenuation obtained. A description of the impact of using those data is performed, showing an appreciable increase near 10 % of the hazard values in the southwestern part of the Iberian Peninsula using the new acceleration attenuation obtained.
S11A-0271
Multiresolution Structured and Unstructured Finite Element Method for Three-Dimensional Attenuated Earthquake Ground Motion Modeling in Basins including Topography
We present an efficient finite element method (FEM) for large-scale attenuated earthquake ground motion simulation, which reduces modeling and computational costs by using a hybrid structured/unstructured mesh capable of treating underground and above surface topography. The meshing process starts with a uniform high- resolution background cell discretization, after which, using an octree structure, the grid is thinned to tailor the local mesh size to a prescribed fraction of the local shear wavelength. A small number of unstructured finite elements are then introduced to model the irregular free-surface topography. We have also implemented an arbitrary intrinsic attenuation model through Standard Linear Solid (SLSs) models, and verified its accuracy, with a constant Q, for a wave propagation problem in a layered system against the corresponding Greenfs function solution. Finally, we demonstrate the usefulness of the new procedure with a simulation of earthquake ground motion in a realistic 3D model of the Los Angeles basin.
S11A-0272
Structural Seismic Risk at David City, Panama
At the southwest margin of the Isthmus of Panama, the Cocos, Nazca and Panama Microplate join in a triple junction. In this tectonic setting, the Panama Fracture Zone (PFZ) which bounds the Nazca and Cocos plate, subducts in an oblique and shallow manner. This zone is one of the most active seismic areas of Central America. On July 18, 1934, the largest earthquake in Panama in historical times (Ms= 7.7) occurred in its northern region. This event caused extensive damage to towns located in the border of Panama and Costa Rica and during the two subsequent days, six aftershocks with magnitude greater than 6.0 were recorded. David City, with 130,000 inhabitants is the most important urban center in southwestern Panama and lies at the northern end of the PFZ. This city was impacted by the strong destructive earthquakes (Ms>7.0) that took place in 1879 and 1934, both with epicenters located on the northern terminus of the PFZ. In this work, we collected and digitized historical seismograms to measure the centroid moment tensor (CMT) from the main 1934 earthquake. Additionally, we gathered new macroseismic information to create improved and more complete isoseismals maps of the 1879 and 1934 events. We determined the probabilistic seismic hazard for David City using records of historical and recent seismicity and the characteristics of local faults. The hazard computation results are presented as peak iso-acceleration curves for rock/hard soil for a recurrence time of 500 years. An elastic response spectrum was obtained with a uniform exceedance probability of 10% in 50 years with one degree of freedom and 5% of damping. Our results indicate maximum peak ground acceleration (PGA) in downtown David of 3.8 and 4.5 m/s2 with a probability annual exceedance of 0.002 and 0.001, respectively. Structural vulnerability was determined analyzing the quality and construction design of housing, buildings, and critical facilities as well as the type of soil where these structures were located. Soil was classified as soft, intermediate, and hard from in-situ dynamic standard penetration test (SPT) measurements. For proper interpretation, all our observations were incorporated in a Geographical Information System (GIS). From a combined analysis of seismic hazard and vulnerability results, we identified six areas with different levels of structural seismic risk increasing from north to south along the David River. Central David has intermediate level risk. The highest risk is measured southwest and south of the city. The southern areas increase their risk due to the thickness of their sedimentary soils, low quality of housing construction and high population density.
S11A-0273
Contribution of Seismology to the Protection of the Cultural Heritage of Greece
Greece is an earthquake prone country, as evidenced by the fact that almost half the seismic energy in Europe is released in its territory both on land and offshore. Numerous large earthquakes have occurred since the antiquity causing considerable damages. Furthermore, as it is well known, Greece is a country of rich cultural heritage, that incorporates a large number of archeological (more than 50%), Byzantine and recent sites and monuments, as well as a continuously growing number of museums. The Greek Seismic Design Code (NEAK, 2004) which governs building, repairing and retrofitting practices divides the country into three zones corresponding to acceleration values of 0.16g, 0.24g and 0.36g. According to the results obtained from a detailed GIS implementation the majority of Greek monuments lie within the boundaries of the second zone (more than 60%). In addition, their geographical distribution in association with the available geological data reveals that more than 50% are sited on soft soils, in contrast to a 30% on rock formations. The effects of these factors are reflected to earthquake related damages observed (evidenced) or reported at various cultural sites. Therefore, the detailed estimation of the seismic hazard for each monument, microzonation studies in the vicinity of each site and estimation of hazard due to tsunamis, liquefaction and landslides related to earthquakes, should be considered as a necessary step towards retrofitting. Applications that reveal the contribution of seismology to the preservation of Greek monuments will be presented and discussed, such as studies performed in Knossos (Crete) leading to representative design values, the church of Kapnikarea (Athens), where the influence of man-made seismic energy sources was examined and the Cathedral of Athens where it was found that local geological conditions amplify the ground velocity for frequencies close to the dominant frequency, a fact that was taken into account during the reconstruction works. ACKNOWLEDGMENTS The present study was co-funded by the European Social Fund and National Resources - (EPEAEK II) PYTHAGORAS, contract No. 70/3/7306.
S11A-0274
Long-Period simulations of Osaka basin wave propagation due to subduction earthquakes
Long-period ground motion characteristics of the Osaka basin are important for strong ground motion prediction due to the large subduction earthquakes of the Nankai trough. Among the previous studies for the Osaka basin ground motion characteristics, Hatayama et al. (1995) pointed out distinctive later phases observed at the OSA station approximately 30s after the S-wave arrival due to the deep events. They interpreted this later arrival is the basin-induced Love waves generated by the basin edge of the Eastern direction by the 2D anti-plane modeling. We tried to model the Osaka-basin induced waves of predominant frequency 0.3 Hz by the 3d FD hybrid method using the 3D basin velocity structure by Kagawa et al.(2004). The two of the three earthquakes investigated in Hatayama et al. (1995) are studied. The later arrivals of S waves (namely SL1 phase) in 3D model at the OSA station are in a good agreement with observation. The time difference between SL1(observed) and SL1"(synthesized) is smaller in full 3D case synthetics, than in 2D case modeling (Hatayama et al., 1995) for EW profile. We investigate a possibility of the later SL1 phase arrival due to it's possible arrival from other than Eastern directions.
S11A-0275
Strong motion: an extended NCQ model for non linear wave propagation simulations
In this work, we consider an extended viscoelastic NCQ model to simulate seismic wave propagation in alluvial basins in the case of strong motions. This constitutive model involves both non linear elasticity and non linear viscous behaviour. The main objective of this model is to reproduce the dependence of the shear modulus and damping on the motion amplitude. To do so, the non linear elastic part of the model is described by a hyperbolic law. The non linear viscous part combines a Nearly Constant-Q model for linear damping (generalized Maxwell body, Moczo et Kristek, 2005) and a non linear viscous contribution described by a hyperbolic variation with the strain level. Furthermore, this model complies with the thermodynamic principles of continuum mechanics (e.g. derivation from potentials and dissipation function). Starting from this extended NCQ model, the analysis of strong motion amplification in alluvial deposits is then performed thanks to a finite element formulation. Various numerical issues have been carefully considered to describe the whole algorithmic procedure (for both elastic and viscous non linear components of the constitutive law). Detailed validations of the model have shown its ability to recover low amplitude ground motion response. For larger excitation levels, the model includes the main features of the non linear behaviour of alluvial deposits. Realistic simulations are performed for Kushiro-oki earthquake. The analysis of seismic wave propagation in surface layers leads to interesting results: at the free-surface the amplification peaks are shifted to lower frequency values (when compared to the input motion); higher frequency components are not overdamped as with classical linear models and, finally, the global amplification level is generally lower than for weak motions.
S11A-0276
Diffraction of seismic waves from 3-D canyons and alluvial basins modeled using the Fast Multipole-accelerated BEM
Seismic wave propagation and amplification in complex media is a major issue in the field of seismology. To compute seismic wave propagation in complex geological structures such as in alluvial basins, various numerical methods have been proposed. The main advantage of the Boundary Element Method (BEM) is that only the domain boundaries (and possibly interfaces) are discretized, leading to a reduction of the number of degrees of freedom. The main drawback of the standard BEM is that the governing matrix is full and non- symmetric, which gives rise to high computational and memory costs. In other areas where the BEM is used (electromagnetism, acoustics,…), considerable speedup of solution time and decrease of memory requirements have been achieved through the development, over the last decade, of the Fast Multipole Method (FMM). The goal of the FMM is to speed up the matrix-vector product computation needed at each iteration of the GMRES iterative solver. Moreover, the governing matrix is never explicitly formed, which leads to a storage requirement well below the memory necessary for holding the complete matrix. The FMM-accelerated BEM therefore achieves substantial savings in both CPU time and memory. In this work, the FMM is extended to the 3-D frequency-domain elastodynamics and applied to the computation of seismic wave propagation in 3-D. The efficiency of the present FMM-BEM is demonstrated on seismology- oriented examples. First, the diffraction of a plane wave or a point source by a 3-D canyon is studied. The influence of the size of the meshed part of the free surface is studied, and computations are performed for non- dimensional frequencies higher than those considered in other studies (thanks to the use of the FM-BEM), with which comparisons are made whenever possible. The method is also applied to analyze the diffraction of a plane wave or a point source by a 3-D alluvial basin. A parametrical study is performed on the effect of the shape of the basin and the interaction of the wavefield with the basin edges is analyzed.
S11A-0277
Velocity Structure Near the Site of the Northern 2002 Embayment Seismic Excitation Experiment Explosion From Reflection, Refraction, and Strong Motion Measurements
P wave reflection and refraction data and SH wave refraction data were taken in November 2006 near the site of the northern 5000 lb explosion of the 2002 Embayment Seismic Excitation Experiment to provide independent constraints on detailed velocity structure between the explosion source and an array of strong motion seismographs that recorded the explosion. Reflection images display prominent reflectors throughout the 650m thick section of unconsolidated Mississippi embayment sediments in addition to the major Paleozoic/Upper Cretaceous unconformity at the base. Reflection statics can be constrained by the unusual and fortuitous recording of relatively deep microearthquakes on the P wave reflection spread and show that static effects in the P wave data are very small. P and S wave velocity models for the sediments are constructed using refraction travel times, Love and Rayleigh wave dispersion measurements, and vertical velocity variations inferred from the P wave reflection images. These models are used to check and refine a velocity model inferred from well-log data elsewhere in the embayment and from modeling waveform data from the nearby strong motion array. Synthetic models for the strong motion waveforms are robust and clearly define the major body and surface wave propagation effects due to the unconsolidated sediments. These results reinforce the conclusion that the unconsolidated sediments of the embayment will form a significant waveguide for high-frequency seismic waves induced by near-surface faulting.
S11A-0278
Ground Motions Simulations of the 2004 Les Saintes Earthquake (Mw 6.4) and its Largest Aftershock (Mw 5.9)
The Guadeloupe Island situated in the Caribbean arc has been struck in November 2004 by a Mw 6.4 earthquake that occurred 20km offshore on a shallow crustal fault. This event caused the death of one person and serious damages to several buildings. The mainshock has been followed by a large number of aftershocks including about 20 events of magnitude larger than 4. The largest aftershocks occurred three month after the mainshock and reached a magnitude Mw 5.9. Most of the earthquake larger than 4 were well recorded by an accelerometric network situated on the nearby islands (the RAP network locally managed by IPGP and BRGM). Data from these stations are available freely on the web site of the RAP network (www-rap.obs.ujf-grenoble.fr). We used the recordings of ten earthquakes of magnitude 4 to 5.2 successively as empirical Green's functions to generate accelerograms for the mainshock and the largest aftershock at 13 stations. For each small event chosen as empirical Green's function, we simulated 500 accelerometers at each station using a two step stochastic approach (Kohrs-Sansorny et al, 2005, BSSA). We compared the results obtained with the signals recorded. We show that, if the parameter C (static stress drop ratio between the mainshock and the aftershock) is properly chosen, this method enables us to generate simulations that fit well the observed signals. We obtained particularly a good reproduction of site effects. This is true for almost all the small event taken as empirical Green's function, as if they are relatively far away from the mainshock. The choice of the parameter C is discussed in this paper by a comparison of the results obtained on rock stations and the values predicted by ground motion prediction equations.
S11A-0279
Study of the Probability Law Governing Ground Motion Metrics Recorded During the 2004 Parkfield Earthquake
Based on the superposition of seismic waves and the Central Limit Theorem, we developed the basis for a unified picture of earthquake variability from its recording in the ground motions to its inference in source models. According to this theory, the random properties of the ground motions and the source for a single earthquake should be both (approximately) distributed according to the Levy law. Computation of the probability density function (PDF) of the peak ground acceleration (PGA) of the 1999 Chi-Chi, the PDF of the PGA and the PDF of the peak ground velocity (PGV) of the 2004 Parkfield earthquakes confirms this theory. As predicted by the theory, we found that the tails of the PDF, characterizing the slip and the PGA, are attenuated according to power laws with exponents (denoted Levy indexes) that take almost the same values close to 1. Computations of the PDF of the PGA recorded at the surface and the PDF of the PGA recorded in borehole during the 2003 Tokachi-oki earthquake lead to a similar conclusion. The PDF tail measures the frequency at which large events occurred and thus quantifies the probability to observe large acceleration values and large velocity values during an earthquake. We extend our analysis of the random properties to other ground motion metrics. To lessen the dependency due to the source-to-site distance, we consider the ratio of the PGV to the PGA, the ratio of the two horizontal components of the PGA to the vertical component of the PGA, and the ratio of the horizontal components of the PGV to the vertical component of the PGV. In this analysis, we use the ground motions recorded during the 2004 Parkfield earthquake, arguably the best-recorded earthquake in history for the density of near-source data. We select stations located within a closest distance to the rupture surface that varies from 0 to 180 km. To test the effect of the distance on the computed random properties, these stations are divided into several subsets or windows. For each subset of stations, we compute the PDF and characteristic function (CF) of the ground motion metrics and compile the parameters of the Levy law that best fit the PDF and CF. We find that the tails of PDF of ground motions metrics decrease with power law behaviors controlled by Levy indexes with values close to 1.
S11A-0280
Testing Community Velocity Models of Southern California Using Ambient Seismic Noise
Correlation of ambient seismic noise has been shown both experimentally and theoretically to reveal robust and accurate Green's function between station pairs. Such data have led to the emerging field of ambient noise tomography. In this study, we correlate monthly data recorded on about 100 broadband seismic stations in Southern California with dense coverage in the Los Angeles area, to determine a large number of station-to-station Green's functions that are dominated by surface waves in the frequency band we are using. These data provide strong constraints on the shallow velocity structure of Southern California in areas of particular interest for seismic hazard analysis. We compare Green's functions from noise correlation (up to 1 Hz) with Green's functions calculated with a versatile finite-element method in the current community velocity models for Southern California (SCEC CVM 4.0 and SCEC CVM-H 4.0). The comparisons provide a comprehensive test of community velocity models for Southern California. The results show the potential for using these methods to target areas of poorly known velocity structure or for ray paths not constrained by regional seismicity. This is particularly relevant for extended faults, like the San Andreas, which are of high concern, but exhibit little to no background seismicity. Deploying seismic stations along the fault would allow the Green's function to the LA Basin to be recovered, and be used for accurate modeling of wave propagation from dynamic rupture models.
S11A-0281
Effects of realistic surface topography on seismic wave propagation: An example from northern Taiwan
Surface topography can influence the ground motion when seismic waves propagate along an irregular free surface. In general, topography increases the amplitude of shaking at mountaintops and ridges, whereas valleys usually have reduced ground motions, as is evident from both numerical simulations and from data recorded after real earthquakes. However, recent attention is mainly focused on mountain regions, and the surrounding areas have received less attention. Here we develop a new spectral-element mesh implementation to accommodate realistic topography and the complex Taipei basin, which is in close proximity to the Central Mountain Range in northern Taiwan. Spectral-element simulations demonstrate that high-resolution topography can change PGA values in mountainous areas by 50% compared to a half-space response. We demonstrate that large-scale surface topography affects the propagation of seismic waves. If a shallow earthquake occurs in the I-lan region, the Central Mountain Range will significantly scatter the surface waves, and in turn reduce the amplitude of the ground motion in the Taipei basin. However, as the hypocenter becomes deeper, topography will scatter the body waves, which subsequently propagate as surface waves into the basin. These waves further interact with the basin and the surrounding mountains, which finally results in complex amplification patterns (PGA increase > 50%) in Taipei city. The complex interactions between mountains and nearby areas, especially sedimentary basins, cannot be ignored. Our simulations demonstrate that surface topography should be taken into account for seismic hazard assessment.
S11A-0282
Strong Ground Motion Simulation of December 26th, 2006 Ping-Tung Earthquake Based on Empirical Green's Function
Ping-Tung earthquake sequence (ML=7.0) occurred at the southwestern offshore of Taiwan on December 26, 2006. This strong earthquake sequence caused many devastating effects at the Heng-Chun area. In the study, we examine the source models of this event using the observed seismograms by CWBSN and Microbs arrays at some stations surrounding the source area. An objective estimation method was used to obtain the parameters N and C which are needed for the empirical Green's function method by Irikura (1986). This method was called ¡§source spectral ratio fitting method¡¨ which gives estimate of seismic moment ratio between a large and a small event and their corner frequencies by fitting the observed source spectral ratio with the estimated one which obeys the ω-2 model (Miyake et al., 1999). It has an advantage of removing site effects in evaluating those parameters. The best source model of the Ping-Tung mainshock was estimated by comparing the observed waveforms with the synthetics using the empirical Green's function method.
S11A-0283
Temporal changes in S-wave velocity structure at a borehole site after strong ground motion shock
We analyzed a pair of surface and downhole accelerographs at station SMNH01 of KiK-net, which experienced
strong ground motion up to 844 gal by the 2000 Western Tottori Earthquake (06/10/2000, MW6.7), Japan, to
examine how the shallow subsurface structure changes with time. Station SMNH01 has a borehole of 100 m
depth, where sandy gravel distributes from 0 to 11 m depths, and solid basalt from 11 m to the bottom. We
calculated the average spectral ratio and the average cross-correlation function of coda waves of more than 100
local earthquakes which occurred in six periods: before the mainshock, 50 to 280s, 0 to 10 days, 10 to 100 days,
100 to 1000 days, and 1000 to 2000 days after the mainshock. Before the mainshock, the lowest peak frequency
of the average spectral ratio was 4.5 Hz. In the period from 50 to 280 s after the strong ground motion shock, the
lowest peak frequency decreased to 4.0 Hz, and the other peaks also decreased their frequencies. After that, the
peak frequencies have continued to recover to their original values for over 1 year. The peak level of the average
coda spectral ratio dropped just after the direct S-wave arrival of the mainshock to the station, but it quickly
recovered within a few minutes. The average cross-correlation function of coda waves shows a clear peak at a
positive lag time in each of the six periods. Before the mainshock, the peak lag time was 0.09s. The peak lag time
increased to 0.105s in the period from 50 to 280s after the strong motion shock, however, it has continued to
recover to the original peak lag time for over 1 year. We estimate the S-wave velocities (VS [m/s]) of shallow
structure beneath the station for each period by fitting the theoretical coda spectral ratio and peak lag time to the
observed ones. The theoretical coda spectral ratio is calculated by using the modified propagator matrix method,
where coda waves are composed of SH and SV waves isotropically incident on the borehole sensor with random
phases. We divide the shallow structure into six layers and fix P-wave velocity (VP [m/s]) and thickness of them
based on the well-log data. We further suppose that attenuation parameters are expressed as
QP=QS=0.008·VS·f, where f is frequency in Hz and that density is d=310·VP0.25 [kg/m
S11A-0284
Triggering Factors Susceptibility Of Earthquake-induced Landslides in 1976 Longling Earthquake
Analysis on earthquake-induced landslide triggering factors susceptibility is an important step to every slope safety estimation method. Based on GIS, the susceptibility of earthquake-induced landslide triggering factors of Longling Earthquake in1976 is analyzed by means of the slope certainty factors (CF). The factors contributing the landslide occurrence include lithology, faults£¬earthquake intensity, epicenter distance, slope angle£¬slope aspect, elevation£¬river etc£®The results show that different factor contribution to landslide is due to its value, and the most numerical intervals of each factor which are inclined to landslides are determined. This work is the basement for further study of earthquake-induced landslide susceptibility analysis.
S11A-0285
Dyke Intrusion and the effect on fault in Northern Hainan Island, China
The A.D.1605 earthquake near Qiongshan caused 70 villages collapsed and were drown by seawater. Many parallel faults and high frequency Cenozoic volcanic activity is vivid feature in Northern Hainan Island. The world volcano map shows the area is in a NS-trending intraplate volcanic belt. In order to understand the form mechanism of the volcanic belt zone and investigate the fault hazard in relation with the magma activity beneath Hainan island, finite element method has been applied to simulate the double subduction (indian plate oblique subduction and philippine plate subduction). The results show double subduction induce double convection and between the two double subduction form an upwelling from the upper mantle, which is the reason of the NS- trending intraplate volcanic belt. The accurate location earthquake from 2000 to 2006 show a vertical earthquake plane, we assume it is due to the magma intrusion in a vertical dyke which has existed, in order to analyse the fault hazard, we calculate the coulomb stress on the faults due to the dyke intrusion. The modeling results show Qionghai-Wenchang fault and Puqian-Qinglan fault have the most potential hazard.