S41A-0228
Mapping Large Structures Using the Shallow Reflection Seismics
The shallow reflection seismics is designed to map shallow structures, say, above a depth of 1000m. The method has the merits of efficiency, high-resolution, and most important, cost-effective. It is a small system, as comparing with the oil-exploration seismics, however, which is affordable by a regular institute and can be used to study relatively large structures to a certain extent. This report describes a case to detect an active fault in Central Taiwan. This fault is unclear in the surface geology. Eight profiles across the major fault are used to map this 20 km long fault. Two concepts from the shallow reflection seismics are applied: 1) Several short lines, each 400m long and 1km apart, are aligned to compose a profile with several km long. 2) High resolution factors are kept, such as small geophone intervals (2m or 4m), high-frequency and high-power sources. A large channel number (e.g., 192) is needed to extend the receiver spray range, which helps to cover larger depths. This results in "1 sec" records which can trace detailed shallow structures as well as the parts as deep as 2000m. The "1 sec" reflection seismics is demonstrated to be a powerful tool to extend the surface geologic observation to substantial depths. The method is cheap and efficient, thus, can be applied extensively as a useful tool to aid the regional geologic mapping.
S41A-0229
Tomography of the Alps using seismic ambient noise
We use seismic ambient noise correlation to study the lithosphere in western Europe. Cross correlation of one year of noise recorded at 150 3-components broadband stations yields more than 3000 Rayleigh and Love wave group velocity measurements. These measurements are used to construct Rayleigh and Love group velocity models of the Alpine region and surrounding area in the 5-50s period band. The seismic noise recorded in Europe is dominated by noise originating from the Northern Atlantic ocean. This anisotropy of the noise and the uneven station distribution affect the azimuthal distribution of the paths where we obtain reliable group velocity measurements. As a consequence our group velocity models have better resolution in the South-East direction than in the North-West direction. Finally we invert the resulting Rayleigh wave group velocity maps to determine the Moho depth. Our results are in good agreement with the result of the numerous active experiments in the Alps.
S41A-0230
A new catalog of eigenfrequencies from the 26th December 2004 Sumatra-Andaman mega- event and first perspectives
The high quality of the records obtained after the Sumatra-Andaman earthquake of December 2004 from the FDSN broad-band seismological stations offers a good opportunity for studying the free oscillations of the Earth and the behaviour of some particular modes. The observation and clear identification of singlets of the gravest seismic normal modes (frequency lower than 1 mHz) and of some particular anomalously split modes should offer strong constraints on Earth structure, and their analysis is particularly promising. The curiosity of scientists for these modes is due to the fact that they are rarely observed. Their interest is the key role they play for constraining Earth's models. Since the 1960 Chile event these modes have been identified and studied after large events recorded from broad-band seismometers or supraconducting gravimeters, but their excitation never reached the signal to noise ratio of the 2004 event and the amount of data never reached the number of present available recordings. The 2004 Sumatra-Andaman event provides us individual spectra exhibiting a clear splitting of the gravest spheroidal modes and a clear identification of each of the individual singlets, with an unprecedented resolution. We carefully analyzed 21 different modes and report the detailed results of some of them, showing an enhancement on the eigenfrequencies determination associated with the high quality of the measurements. Our results correspond mainly to analysis of vertical component recordings but also to additional horizontal components. We present modal frequencies measurements of some spheroidal fundamental gravest modes and modal frequencies measurements of some particular higher modes, inner-core modes and inner-core anomalously split modes. The results are compared to the theoretical frequencies computed for the PREM model, taking into account both rotation and ellipticity effects of the Earth. Clear observations of the individual singlets of rarely observed modes such as Slichter mode's overtones or anomalously split inner core modes allow to construct a new useful eigenfrequencies dataset. At low frequency, the misfit between our measurements and the theoretical corresponding values (computed for the PREM model in a rotating Earth with ellipticity) point out that some parameters (density or attenuation values) used for constructing the PREM theoretical model are not constrained enough to explain our observations. A small change in the density profile model (or the attenuation profile model) may provide new eigenfrequencies values and a better fit with our observations. At larger frequency the lateral heterogeneities of the Earth can also affect the results and we have to discriminate among the different effects. That new catalog of modal eigenfrequencies offers a plethora of perspectives, it is an important step towards the resolution of a density (or attenuation) profile model.
S41A-0231
Experimental Study of the Convergence of Two-Point Cross-Correlation Toward the Green's Function
It has been shown theoretically by several authors that cross-correlation of the seismic motion recorded at two points could yield the Green's Function (GF) between these points. Convergence of cross-correlations toward the GF depends on sources positions and/or the nature of the wavefield. Direct waves from an even distribution of sources can be used to retrieve the GF. On the other hand, in an inhomogeneous medium, recording the diffuse field (coda) is theoretically sufficient to retrieve the GF whatever the sources distribution is. Since none of these two conditions (even distribution of sources or a perfectly diffuse field) is satisfied in practice, the question of convergence toward the GF has to be investigated with real data. A 3D exploration survey with sources and receivers on a dense grid offers such an opportunity. We used a high- resolution survey recorded by Petroleum Development Oman in North Oman. The data have been obtained in a 1x1~km area covered with 1600 geophones located on a 25x25~m-cell grid. Records are 4-seconds long. A unique feature of this survey is that vibrators (working in the [8-120~Hz] frequency band), were located on a similar grid shifted with respect to the receiver grid by half a cell (12.5~m) in both directions. This allows us to compare estimated GF's with measured direct waves (GF's) between the geophones. The shallow subsurface is highly heterogeneous and records include seismic coda. From this dataset, we selected two receiver locations (Ra and Rb) distant from d=158~m. We used both different sets of source locations and time windows to compute the cross-correlation between these two receivers. Then we compared the derivatives of correlation functions with the actual GF measured in Rb (resp.~Ra) for a source close to Ra (resp.~Rb). By doing so, we show the actual influence of source locations and scattering (governed by the records' selected time window) on the Signal-to-Noise Ratio (SNR) of the reconstructed GF. When using direct waves, the result illustrates the dominant contribution of the sources located in the endfire lobes of the receiver pair (as defined from the stationary phase theorem). If sources are located all around the receivers, both direct and coda waves yield very good recovery of the GF, with a comparable SNR. The agreement between actual GF and cross- correlation derivative is then very good for direct surface waves, and extends to later arrivals, which are likely scattered waves. If all sources are located outside the endfire lobes, the GF can nevertheless be retrieved when using the scattered waves contained in the coda. Acknowledgment: The authors would like to thank the Ministry of Oil and Gas of the Sultanate of Oman and Shell Research for their permission to publish these results.
S41A-0232
Super-wide Angle One-way Method and its Applications in Imaging Steep Salt Boundaries
Although a variety of methods have been proposed to improve the wide-angle accuracy of one-way wave propagators, the accuracy of wide-angle waves for strong contrast media is still a serious problem and put a practical limitation on applying these methods to steep reflector imaging. One-way wave equation has a preferred direction for wave propagation which is either the z-axis or the x-axis in 2D Cartesian coordinate system. For a certain preferred direction, the large-angle waves carry some errors both in phase and amplitude. However, large-angle waves with respect to z-axis become small-angle waves to x-axis and vice versa. Based on that, the super-wide angle one-way method has been developed to extend the capability of one-way propagators by a wavefront reconstruction method which combines and interpolates the two orthogonally propagated one-way wavefields to rebuild the distorted wavefront to good accuracy. In this paper, a more complex reconstruction scheme is developed for wave propagation to improve the accuracy and reduce the artifacts. The new scheme employs the wavefield gradients to determine the weighting function for the two orthogonally propagated wavefields. We first propagate the wavefield downward completely to get both the downward wavefield and its gradients. For the horizontal propagator, the wavefield will be obtained by combining the downward propagated wavefield at each step with the weights calculated from both wavefields. This scheme can be considered as an intermediate step of the complete wavefront reconstruction. Numerical examples of impulse response demonstrate the good accuracy of the weighting scheme and the application in modeling turning waves, compared with the regular one-way method. The large-angle waves can be modeled correctly. We also show the good performance of this method on imaging the salt dome with overhanging flanks in the BP 2D benchmark model. The energy has been rebuilt on almost all boundaries of the salt by this method.
S41A-0233
A Study of Earthquake Location by Artificial Explosive Data
A good quality of earthquake location is very important factor to all of the earthquake studies. On October 18, 2006, an explosion was taken by the Taiwan Integrated Geodynamics Research (TAIGER) project at the Dili Village in central Taiwan to test seismic signal receiving condition for the future project. The explosion signals were recorded by the Central Weather Bureau Seismic Network (CWBSN). It offers a good opportunity to study the location accuracy of the CWBSN. We relocated this event using the layered model that routinely used in the CWBSN and a recently 3-D model by Wu et al. (2007). Our results show that the event locations determined by layered and 3-D model were caused 5.0 and 1.5 km from the true location, respectively. Base on this result, we suggest that the CWBSN location may have about 5 km error in routinely operation. The location error may shorten to 1.5 km by using 3-D model. Those are good reference values for the Taiwan earthquake studies using the CWBSN catalog.
S41A-0234
Spatial Variation of Surface Wave Q and Body Wave t* in North America
We estimate the spatial variation of the seismic parameter t* using teleseismic (30°--90°) P wave recordings of about 300 deep (> 200 km) earthquakes at broadband stations in North America. We determine the P wave spectral ratio Rij for about 600,000 station pairs i-j with high signal-to-noise ratio P wave signals. The linear fit to lnRij between f= 0.1--1.0 Hz is measured to estimate differential Δt* assuming that lnRij is proportional to π fΔt* (e.g., Aki and Richards, 1980). The measurements are inverted for t* at each station by least-squares inversion. Preliminary inversions indicate that the variation of t* correlate with the tectonic terrains of North America. Predominantly low values of t* are obtained for stations in the Canadian Shield and high t* values in the North American Cordillera. This variation is similar to Q variations inferred from global surface wave amplitude data (e.g., Dalton and Ekström, 2006), suggesting that intrinsic attenuation is the common cause. We will discuss the robustness of our t* estimates (including the effects of scattering on P wave ratios) and make a detailed comparison with surface wave Q maps.
S41A-0235
Understanding Extension Within a Convergent Orogen: Initial Results From the Carpathian Basins Seismic Project
The Carpathian Basins Project (CBP) aims to understand the origin of Miocene-age extensional basins, of which the Pannonian Basin is the largest, within the arc of the Alpine-Carpathian Mountain Ranges – a compressional structure. Analysis of the subsidence history of the Pannonian Basin shows that its mantle lithosphere has undergone a much greater degree of extension than the overlying crust. We describe the results of a temporary seismic deployment to test competing theories of how the continental lithosphere evolved in the region. We deployed a 46-element seismic network, 450 km x 80 km, oriented in a NW-SE direction, crossing the Vienna and western Pannonian Basins in Austria, Hungary and Serbia. The network ran for 14 months from early May 2006. The stations were broadband to 30s and spaced at ~30 km along 3 parallel lines, which are 40 km apart. The principal object of this network is to use P and S-wave teleseismic tomography to image the upper mantle. P- wave residuals from sources perpendicular to the tectonic grain show a ~1s variation across the Mid-Hungarian High in to the Pannonian Basin. This delay cannot be explained by sedimentary or crustal thickness variations, which are well-controlled by boreholes, deep seismic soundings and our own receiver function analyses. We must infer significant lithospheric thinning and anomalously low asthenospheric velocities underlying the Pannonian Basin to explain our observations. These travel time delays are accompanied by a dramatic change in the orientation of SKS splitting measurements from E-W to NW-SE across the Mid-Hungarian High. We have also installed a more broadly distributed regional broadband array of 10 instruments (broadband to 120 sec) for 2 years from September 2005, spaced at ~100km within Hungary, Croatia and Serbia to augment the data available from permanent broadband networks in central Europe. Preliminary interstation surface wave dispersion results from across the Pannonian Basin imply lithospheric thicknesses of the order of 60km.
S41A-0236
Converted phases migration using active and passive seismic data at the Campi Flegrei caldera
We applied a type of depth migration for converted seismic phases to active and passive seismic data sets from the Campi Flegrei volcanic region in southern Italy. The converted phases analysis is based on the diffraction summation migration technique. Travel times to grid points in a volume are calculated in smooth P and S-velocity models from tomography and trace energy near the calculated converted phase time is stacked over multiple sources at one receiver. Weighting factors based on Snell's Law at an arbitrarily oriented local interface are applied to aid in focusing of trace energy. PP and PS reflection images from the active data set show features near 2-3 km in depth, which may be associated with an over-pressured gas volume suggested by tomographic investigations. Possible deeper features near 4 km depth may be related to the presence of the carbonate basement. A portion of the caldera rim near 1-2 km depth is imaged considering PP reflections. We found that the passive earthquake data set from the 1984 bradiseismic crisis was not well suited to the converted phases analysis due to the acquisition geometry, however, we can confirm some features found in our investigation of the active data using the earthquake traces and images from two selected stations.
S41A-0237
Free Oscillations of the Earth Observed by Closed Borehole Wells
We have made observations of pore pressure under undrained condition by an airtight borehole penetrating an artesian, or a confined aquifer in the Atotsu tunnel excavated in the Kamioka Mine, central Japan. We confirmed that the relation between pore pressure change and stress change is a zero-order system for a wide range of frequency and that stress change, strictly speaking strain change, induced within the rock mass shared by the skeletal framework of rock and pore fluid. Examining the pore pressure measured using closed borehole wells, we detected free oscillations of the Earth excited by earthquakes such as the 26 December 2004 Mw = 9.1 Sumatra-Andaman Islands earthquake (epicentral distance Δ= 51.1°) and other M7 to 8 events. We made a Fourier analysis of the pore pressure record produced by the earthquakes. We examined (1) whether the closed borehole has sufficient sensitivity to identify free oscillations, and (2) how the closed borehole responds to spheroidal modes and troidal modes. The poroelastic theory predicts that pore pressure should respond only to spheroidal modes since pore pressure change is proportional to volumetric strain change. No pore pressure response is expected from shear strain that is produced by troidal modes. However, it is controversial whether pore pressure responds to shear strain, since phases corresponding S- and Love waves have been usually detected on hydroseismograms. We calculated the spectrum of the 24 hours time windows (86400 points) with shifting the time window by 1 hour from 24 hours before the origin time of the event to 24 hours after that. The spectrum peaks correspond to entire fundamental spheroidal modes were clearly observed. The Q of each mode is calculated by fitting the decay of the amplitude of each peak. The peaks whose eigenfrequencies are less than 1 mHz (0S0, 0S2, 0S3, 0S4, and 0S5) clearly appear 5 hours after the event. On the other hand, no spectrum peak corresponding troidal modes was observed. These results confirm that the poroelastic theory correctly predicts the pore pressure response.
S41A-0238
Full Waveform 3D Synthetic Seismic Algorithm for 1D Layered Anelastic Models
Numerical calculation of synthetic seismograms for 1D layered earth models remains a significant aspect of amplitude-offset investigations, surface wave studies, microseismic event location approaches, and reflection interpretation or inversion processes. Compared to 3D finite-difference algorithms, memory demand and execution time are greatly reduced, enabling rapid generation of seismic data within workstation or laptop computational environments. We have developed a frequency-wavenumber forward modeling algorithm adapted to realistic 1D geologic media, for the purpose of calculating seismograms accurately and efficiently. The earth model consists of N layers bounded by two halfspaces. Each layer/halfspace is a homogeneous and isotropic anelastic (attenuative and dispersive) solid, characterized by a rectangular relaxation spectrum of absorption mechanisms. Compressional and shear phase speeds and quality factors are specified at a particular reference frequency. Solution methodology involves 3D Fourier transforming the three coupled, second- order, integro-differential equations for particle displacements to the frequency-horizontal wavenumber domain. An analytic solution of the resulting ordinary differential system is obtained. Imposition of welded interface conditions (continuity of displacement and stress) at all interfaces, as well as radiation conditions in the two halfspaces, yields a system of 6(N+1) linear algebraic equations for the coefficients in the ODE solution. An optimized inverse 2D Fourier transform to the space domain gives the seismic wavefield on a horizontal plane. Finally, three-component seismograms are obtained by accumulating frequency spectra at designated receiver positions on this plane, followed by a 1D inverse FFT from angular frequency ω to time. Stress-free conditions may be applied at the top or bottom interfaces, and seismic waves are initiated by force or moment density sources. Examples reveal that including attenuation stabilizes the numerical calculations, and reduces wraparound artifacts associated with the spatially-periodic 2D Fourier transform. Sandia National Laboratories is a multiprogram science and engineering facility operated by Sandia Corporation, a Lockheed-Martin company, for the US DOE under contract DE-AC04-94AL85000.
S41A-0239
High Resolution Receiver Functions From the Southern Great Basin, Nevada
The crustal structure based on teleseismic receiver functions from the 30 station short period Southern Great Basin Digital Seismic Network is our focus. The Southern Great Basin Seismic Network is centered on the proposed high level nuclear waste facility at Yucca Mountain, Nevada, on the western edge of the Nevada Test Site. There are approximately thirty 3-component seismometers in the Yucca Mountain region, some of which are located in the Nevada Test Site. Yucca Mountain is located in the Basin and Range physiographic province, and geology in the immediate vicinity consists mostly of Miocene ash flow tuffs. A large north to south variation in Bouguer gravity is centered at latitude 37N, just north of Yucca Mountain, with a 50 mGal variation in Bouguer gravity over 50 km. We seek to determine whether there are crustal thickness variations that correspond to the dramatic gravity variations over these short spatial scales. Radial and transverse component receiver functions for P arrivals are calculated from approximately 400 teleseismic events coving a wide backazimuthal range using a time-domain iterative deconvolution method. A conversion consistent with a sediment layer is seen at a few stations, creating a more complex velocity model for the area. The sediment layer also produces multiple reverberations, interfering with the useful signal. From moveout plots the Moho is visible around four seconds at most stations along with a midcrustal boundary near two seconds at a number of stations. We see large variation with backazimuth of the converted arrivals with polarity changes on both the radial and transverse components, suggesting scattering and complex crustal structure. Robust features are made visible by interstation stacking. Throughout the area we find little variation in crustal thickness, suggesting an origin of the gravity anomaly other than simple crustal thickness. Velocity models from past refraction experiments were used to migrate the observed interfaces to depth. Some of the more interesting azimuthal signals are also discussed.
S41A-0240
Proposed, Standardized Multimode Representation of 3-D Varying Structures Having Variable Surface Curvatures
Our problem is the computational treatment, with multimode seismic procedures, of the true 3-D heterogeneous structure having variable surface curvature. Since these multimode seismic procedures are based on a laterally- homogeneous structure with fixed surface curvature, our problem becomes that of devising, implementing, and testing a satisfactory means for assigning locally, a laterally-homogeneous structure with fixed surface curvature, to the true structure. The availability of modern computational hardware, and Internet communications for network computations, have led us to develop multimode procedures for application to large-scale 3-D mapping of the upper mantle by an international group. There are four main parts of this multimode mapping: (1) an initial 3-D structural specification (the most difficult part of the mapping), (2) static computations (whose algorithms, computations, and network are treated here), (3) wavefront-propagation computations (whose input is the output of (2), and whose output is theoretical seismograms to compare with those recorded experimentally), and (4) inversion computations (to obtain an improved structure from the above comparison). To apply these mapping procedures, a standardized multimode seismic representation of a geographical region, with surface dimensions of a few thousand km, is critical. This requirement enters in the static computations which assign a full, propagating-mode (spheroidal and torsional) specification to each latitude-longitude location of the geographical region. Explicit, fundamental assumption involved in treating a 3-D varying structure, having variable surface curvature, with modal procedures: it requires that for a given latitude-longitude location, each multimode triplet (frequency, mode number, surface azimuthal direction of propagation) be assigned its own specific laterally-homogeneous structure and its own specific radius of surface curvature. This assumption reduces to the specification of the extent of the true structure that is used to determine the triplet-specific, laterally- homogeneous structure and constant surface cuvature. The vertical extent is obvious: the triplet's depth of penetration; for the triplet's lateral extent, we use a surface diameter equal to the depth of penetration (key detail in the fundamental assumption). Within the assumptions, large-scale network computations in the extended Himalaya have been used to test the feasibility of the procedures: graphics are employed to illustrate and quantify the effects of (1) lateral heterogeneity of the structure, and of (2) variable surface curvature. The second feasibility tests concern computation time and storage requirements. Seeking 10 km lateral resolution in our mapping, static computations are carried out on a 10 km x 10 km grid of surface locations. This and the surface dimensions of the region being mapped (say, 3000 km x 4000 km) govern computation-time and storage requirements. For 2000 technology having network nodes with 64-bit-arithmetic chips and 0.6 GHz speed, the static computations that treat the 120,000 locations in our target time of 2 months require 211 nodes; for 2007 technology (nodes with 2.4 GHz speed), 53 nodes are needed. Storage requirements: the 80 or 120 GB of external disk storage per node, is not a problem. Our standardized representations: these proposed standards are illustrated as (1) multimode graphical representations at a set of fixed azimuths, and as (2) multi-azimuth graphical representations.
S41A-0241
Evaluation of Installation Methods for STS-2 Seismometers
In the course of the upgrade of the Gräfenberg Array (GRF) from Streckeisen STS-1 seismometers with 20 sec free period to STS-2 seismometers with 120 sec free period the question of how best to install the sensors was reconsidered. It was understood early on that the Streckeisen STS-1 seismometers need elaborate shielding in order to reach their full potential. Because some of the experience gained with the shielding of the STS-1 entered the design of the casing of the STS-2, it was not clear what kind of additional shielding was needed for getting best results with the STS-2. Since the first deployments of STS-2s starting with the German Regional Seismic Network (GRSN) in 1991, different types of insulations have been tried and it became apparent that the data quality at low frequencies can be markedly improved by extensive shielding of the sensors. In contrast to the STS-1 the STS-2 has a sealed casing such that variable air pressure should not lead to any buoyancy forces on the sensor masses. However, since the warp free design of the sensor does not completely remove pressure induced tilt, Wielandt proposed to install the STS-2 in a sealed container consisting of a thick gabbro base plate and an upside down stainless steel pot as cover (Stuttgart shielding). The German Regional Seismic Network (GRSN) and the Gräfenberg Array (GRF) are so far the only networks which have adopted this shielding. Its principal benefits are three fold: it reduces pressure fluctuations by approximately a factor of 30, reduces temperature fluctuations and keeps the sensor dry. We inspect two types of signals to make inferences about the STS-2 shielding used by different networks: Horizontal free mode spectra from the 2004 Sumatra earthquake and vertical component noise spectra tuned for the detection of the permanently excited background free oscillations (hum). The networks considered here are the GRSN, the GEOFON network of the GFZ Potsdam, the Swiss SDSnet and the Japanese F-net. Only few of the STS-2s operated by the GEOFON network, the F-net and the SDSnet stations can detect the hum (28, 26 and 19%, respectively), whereas 78% of the GRSN and GRF stations equipped with the Stuttgart shielding give positive results. We primarily attribute this difference to the different kinds of sensor shieldings in these networks. In our comparison of horizontal component spectra of the 2004 Sumatra event we find that the F-net ranks in quality second behind the GRSN but before SDS-net and GEOFON. The elevated noise level in the GEOFON spectra is probably due to the fact that the GEOFON shield is not stiff enough so that variations in ambient air pressure induce large tilt noise. In spite of the lack of side-by-side comparisons of differently shielded seismometers we have found clear indications that the shielding proposed by Wielandt yields by far the least noisy low frequency vertical and horizontal component data. While our study concentrated on the free mode band our findings are certainly also relevant at the frequencies of surface waves and long-period body waves as noise levels in these frequency bands are highly correlated. Based on our study we recommend the use of the Stuttgart shielding for STS-2 based networks. http://www.geophys.uni- stuttgart.de/~widmer/wws.pdf
S41A-0242
Introduction about Seismic array Laboratory, IGGCAS
Seismic Array Laboratory (SAL, IGGCAS) was built up in 2000, whose initiative was to study the earth structures through portable seismic observation and it has been operating for about 7 years till now. Totally, the lab have 272 sets of seismographs, including 152 CMG-3ESP seismometers, 18 CMG-3T seismometers, 2 STS-2 seismometers, 100 domestic seismometers, 172 Reftek recording system and 100 domestic recording system, respectively. The seismographs in IGGCAS are only used for temporary seismic observations and are open to users who aim at seismic observation based research and, in principle, are able to afford instrument rental. For the past seven years, more than 10 scientific research groups have used these seismographs and set up more than 470 portable stations in Chinese mainland. Up to now, IGGCAS has carried out 15 temporary seismic station arrays which are mainly distributed in North and Southwest China. The data collected at these seismic stations are currently managed by the SAL of IGGCAS and some regulations have been set down for data usage. Users who implement seismic observations with the seismographs in IGGCAS have three years priority to use the corresponding seismic data after field operation. Then, those data should be released by the SAL of IGGCAS. Now, parts of the data have been open to public through the website http://www.seislab.cn/data, and currently, only some event data in SAC format are released due to the lack of high-level technicians dealing with data pre- processing and quality control. In the near future, the lab will release more data in multiple formats to facilitate researchers in relative area via internet connection.
S41A-0243
Local Effects Observed on the Earth's Gravest Free Oscillations
At tidal frequencies, recordings of closely-spaced horizontal seismometers, even if separated only by a few meters, can present large discrepancies. These perturbations are due to local effects such as topography, local geology or cavity effects (when instruments are installed in tunnels or cavities), which imply strain-tilt coupling. These local effects are well observed in the Earth tide recordings and have been well understood since the 70s. At very low frequencies (tide), the horizontal seismometers are mainly sensitive to tilt. As the frequency increases, the inertial sensitivity becomes larger and competes with tilt, until at high frequency the local effects mentioned above should eventually disapear. The question arises then about the importance of such local effects in the seismic band (T < 1h). This is the problem we address here by studying horizontal recordings of the Earth's normal modes from co-located instruments. The gravest free oscillations of the Earth are excited by large earthquakes but are often masked by the high noise level induced by atmospheric activity. The long period recordings of the giant 2004 Sumatra earthquake (Mw ~ 9.3) offer an excellent opportunity to study these local effects on the Earth's gravest free oscillations (f < 1.2 mHz). In order to better understand the problem, we also compare the observations with synthetic spectra that include the modeling of Earth's global 3D structure. Using records from several seismic stations with high signal/noise ratio, we demonstrate for the first time the presence of these local effects for the Earth's gravest free oscillations. Our observations indicate that the fundamental modes are the most affected. These local effects can be characterized by 3 coefficients which express the coupling between the tilt and the strain field. For each station and each instrument, we estimate a set of coefficients by comparing the numerical predictions with the observations. Once estimated, these coefficients can be directly used to reduce the local effects in the normal modes data. The obtained coefficients for the free oscillations are comparable to those found in previous analyses at tidal frequencies.
S41A-0244
Seismic Observations in Extreme Cold Environments: IRIS Instrumentation Takes to the Cold
In 2006, the National Science Foundation (NSF) awarded a Major Research Initiative (MRI) grant to UNAVCO and the Incorporated Research Institutions for Seismology (IRIS) to develop a power and communications system that will improve remote autonomous geophysical observations in the polar environments. Currently in the second year of a three year program, field developments and designs have proven that a high-quality seismic station can be operated and maintained in an extremely cold environment utilizing recent manufacturing breakthroughs in light weight battery designs and insulating materials. With modern, state-of-the-art seismic equipment now being designed to operate at very low power in more extreme temperature ranges, we have the opportunity to exploit new opportunities in polar environments with increased reliability and reduced logistics requirements on the polar field support agencies (primarily, the NSF's Office of Polar Programs). As a result of intermediate results and successes with the autonomous station design, NSF has awarded another grant to IRIS to begin to establish a pool of seismic instrumentation and station infrastructure packages designed to operate PASSCAL experiments in the polar-regions. Procurement has begun on this new pool and support of field operations have already begun on projects in Greenland (Helheim Glacier) and Antarctica (Gamburtsev Mountains Project – AGAP; Polenet; and a tomographic study of Mt Erebus). Along with the equipment, PASSCAL has now established a dedicated staff to polar projects to further enhance the support and quality of the data return for these challenging projects.
S41A-0245
Low Intrinsic Attenuation and High Scattering Attenuation Near 1Hz in the Crust of the Tanzanian Craton
A 2-dimensional Energy Flux model method of Frankel and Wennerberg (1987) was used to obtain intrinsic and scattering attenuation models from regional and local earthquakes recorded by the 1994-1995 Tanzanian passive seismic experiment. Data for three earthquakes of magnitudes 5.9, 5.0 and 5.2, that occurred on 08/18/1994, 08/31/1994 and 11/12/1994, respectively, were obtained from IRIS. The E-W and N-S component data recorded by the array of stations were rotated to the great circle path to obtain transverse and radial components. The transverse component is chosen for two reasons (1) it is rich in coda, (2) and, most earthquake damage is from the SH component. Our results indicate low scattering Q (high scattering attenuation) and high intrinsic Q (low intrinsic attenuation) of the crust between an earthquake source and a station in the array. The high crustal intrinsic Q (QI) in the Crust may be explained by the absence of intrinsic attenuation mechanisms such as passage of fluids in interconnected cracks with in the crust. The absence of shallow very high temperature in the Tanzanian Craton may also contribute to the efficient passage of seismic wave energy in the crust. The low scattering Q (QS) may be explained by the presence of lateral heterogeneities, such as fractures, that redistribute energy from the main arrival phase into the coda by scattering mechanism.
S41A-0246
The Virtual Seismic Atlas Project: sharing the interpretation of seismic data
Through the activities of academic research programs, national institutions and corporations, especially oil and gas companies, there is a substantial volume of seismic reflection data. Although the majority is proprietary and confidential, there are significant volumes of data that are potentially within the public domain and available for research. Yet the community is poorly connected to these data and consequently geological and other research using seismic reflection data is limited to very few groups of researchers. This is about to change. The Virtual Seismic Atlas (VSA) is generating an independent, free-to-use, community based internet resource that captures and shares the geological interpretation of seismic data globally. Images and associated documents are explicitly indexed using not only existing survey and geographical data but also on the geology they portray. By using "Guided Navigation" to search, discover and retrieve images, users are exposed to arrays of geological analogues that provide novel insights and opportunities for research and education. The VSA goes live, with evolving content and functionality, through 2008. There are opportunities for designed integration with other global data programs in the earth sciences. http://www.seismicatlas.org
S41A-0247
UNAVCO-PBO Strainmeters on Mt. Saint Helens: The What, Where, When, and Why of a Strainmeter Array.
In the summer of 2007 UNAVCO installed four Gladwin bore hole strainmeters and seismometers for the purpose of observing Mt. Saint Helens geophysical processes. The strainmeters are co-located with tiltmeters and near continuously operating PBO GPS sites. When combined the data sets will record the time-dependent three-dimensional deformation field at the volcano and has the capability of capturing strain transients on the periods of hours to months. This array will be used as a case study of how UNAVCO strainmeters are located for volcanic monitoring and observation as well to give a general idea of how a strainmeter array is selected. What are criteria used to select an array? Where are the ideal locations for a site within array? Why is the array important? The model for the array emplaced on Mt St Helens can be transferred to other volcanic setting, such as Yellowstone. This array will provide long-term (20+ year) data with little maintenance and access to the data through UNAVCO the IRIS DMC and the NCEDC. This array is an excellent example of the power of borehole strainmeters and the suite of instruments that accompany it in understanding geophysical processes.
S41A-0248
New STS-1 Electronics: Development and Test Results
The STS-1 seismometer is currently the principal very broad-band (VBB) seismometer used in global or regional seismic networks operated by members of the Federation of Digital Broad-Band Seismograph Networks (FDSN). It is widely viewed as the finest VBB sensor in the world, Unfortunately, many of the STS-1's, which were manufactured and installed 10-20 years ago, are encountering both operational failures and age-related degradation. This problem is exacerbated by the fact that sensors are no longer being produced or supported by the original manufacturer, G. Streckeisen AG. In a first step towards assuring continued high quality of VBB data for decades to come, we have developed and tested new electronics and methods for mechanical repair for the STS-1 very broadband seismometer. This is a collaborative project with Tom VanZandt of Metrozet, LLC (Redondo Beach, CA) and Erhard Wielandt (original designer of the STS-1), and has been funded by a grant from NSF through the IRIS/GSN program. A primary goal of this effort was to develop a fully-tested, modern electronics module that will be a drop-in replacement for the original electronics. This new electronics design addresses environmental packaging problems that have led to operational degradation and failures in the existing instruments. This effort also provided the opportunity to implement a set of electronic improvements that will make the installation and operation of the sensors more efficient. Metrozet developed the first prototype new electronics for the STS-1, while the BSL engineering staff constructed a test-bed at the Byerly Vault (BKS) and developed the capability to simultaneously test 6-8 STS-1 components. BSL staff then tested successive versions of the electronics. The first generation prototype electronics did not include centering or calibration functionality. The second generation prototype included remote centering functionality as well as calibration functions. After some observations and refinements, this generation of electronics was operated on two seismometers concurrently and successfully run through swept sine and step calibration functions on four seismometers. During this final phase, the Metrozet electronics included the ability to initiate and operate the calibrations via a network (Ethernet) connection. Most of the calibration testing was performed remotely from Metrozet's Southern California office over the BSL network. Metrozet was able to remotely log into the Berkeley network, establish a connection to the test bed in the Byerly seismic vault and initiate control of the seismometer including remote centering and calibration functions. Finally, after BSL tests were completed and the development appeared complete and satisfactory, the new electronics were tested at the Albuquerque Seismological Laboratory's seismic vault, which is located in a quieter environment than BKS. The new electronics package was also field tested at the BDSN broadband station HOPS. We present detailed results of the calibrations.
S41A-0249
P-wave Tomographic Image Across the Central Transverse Ranges Region From Inversion of Local Earthquake and Active Source Data, Southern California
The central Transverse Ranges region is considered as a place that has documented significant evidence of tectonic history of southern California. To investigate such a region associated with extensive Cenozoic tectonic activities, we apply a new deformable-layer tomography approach to determine layers of varying thickness as a direct product of tomographic imaging. 1,546 P-wave first arrival data from local earthquakes and 10,869 from active shots of Los Angles Regional Seismic Experiment (LARSE) I have been used together to invert the velocity structure along a northeastern trending 2-D profile across the Central Transverse Ranges region. The combined data result in highly dense ray path distribution, especially within the upper crust. In addition, Moho depths from previous study have been incorporated into the inversion to improve solving the model in the deep and sparsely sampled areas. The tomographic inversion procedure implemented in this study is as follows: first, a long wavelength model is constructed by inverting local earthquake arrivals only with sparse Moho constraints; second, this tomographic model is refined by adding surface seismic arrivals and introducing lateral velocity variations within layers. Along the profile, our tomographic image shows high resolution, detailed near-surface geological features besides those long-wavelength features resolved in the previous studies. The existence of the San Andreas Fault is clearly evidenced, which is characterized by a vertical low-velocity zone to at least 20-km depth. The sedimentary basin boundaries are observed unconformably lying above the high velocity basement. Near the northern edge of the Los Angeles Basin and southern foothills of the San Gabriel Mountains in the upper crust, the high velocity gradient units on two sides of this high velocity feature represent that the transition from sedimentary rocks to crystalline rocks is abrupt, which could be indicative of the major basin-bounding faults, such as the Whittier fault and the Sierra Madre fault zones. Our new tomographic image represented by both layers and cells has provided new insights across the Central Transverse Ranges region, and the resolution of the tomographic solution has been greatly improved by using local earthquake data and surface seismic data jointly.
S41A-0250
The ZH Ratio Analysis of Rayleigh Waves Using Broadband Seismograms in Southern California
With the advent of dense, three-component seismic networks, there are many new approaches that can be devised to analyze surface waves. We have been developing a technique to use the ratio of vertical to horizontal component of Rayleigh waves, termed here the ZH ratio, as an alternative approach to phase and group velocity analysis (e.g., Tanimoto and Alvizuri, 2006; Tanimoto and Rivera, 2007). In this approach, the ZH ratio is measured as a function of frequency. This method is more sensitive to shallow structure than the traditional dispersion data and therefore has unique features in the results. In this paper, we report our results for Southern California, using the microseismic signals found for frequencies 0.1-0.4 Hz. The ZH ratio is controlled only by the local structure beneath a seismic station, but the station density in southern California seems to be high enough to let us construct a cohesive model, particularly in the L.A. Basin. Our analysis shows that, even though there are some variations from station to station, the ZH ratios can be measured at most stations for frequencies 0.13-0.37 Hz. This is much wider than the frequency band used in Tanimoto and Alvizuri (2006), which restricted the range to 0.1-0.2 Hz. This narrow range was used in the study because phase shift of 90E#8249; between the horizontal and vertical components was positively identified only for this range. We have extended this range because the ZH ratio above 0.2 Hz is often observed to be fairly constant throughout the year even though amplitudes of microseisms vary from season to season by a factor of 10. Constancy seems to indicate that the signals consist of Rayleigh waves even though phase shift of 90E#8249; between the horizontal and vertical components is hard to confirm above 0.2 Hz (in Southern California) due to low signal levels. There seem to be a few different and distinct patterns in frequency dependence of the ZH ratio in the area. For example, stations in Los Angeles basin typically show a minimum below 0.20 Hz. In comparison, some stations outside the basin show a minimum at much higher frequencies while others show fairly constant ratios throughout the frequency band. These features seem to be related to the thickness of sediments under the seismic stations and may be critical for deriving the underlying shallow structure. It seems also true that the theoretical ZH ratios of the standard velocity model in the area (SCEC Community Velocity Models) are often systematically off at many stations. This means that these stations are underlain by sediments whose thickness is not correctly represented by the models. Depth sensitivity kernels of the ZH ratio for 0.13-0.37 Hz are sensitive to S-wave velocity for depths 0-10km. Inversion of this data for depths seems to provide valuable information for shallow structures. It is well known that the thickness of sediments leads to variations in ground motion by some factor and is thus critical to ground motion estimates by future earthquakes in Southern California. We will report on the shallow S-wave velocity structure, derived by this approach.
S41A-0251
The ZH ratio Analysis of Global Seismic Data
The ZH ratio, the ratio of vertical to horizontal component of the fundamental Rayleigh wave as a function of frequency, is an alternative approach to phase/group velocity analysis for constructing the S-wave velocity structure. In this study, teleseismic Rayleigh wave data for the frequency range between 0.004Hz to 0.04Hz is used to investigate the interior structure. We have analyzed most of the GEOSCOPE network data and some IRIS GSN stations using a technique developed by Tanimoto and Rivera (2007). Stable estimates of the ZH ratios were obtained for the frequency range for most stations. We have performed the inversion of the measured ZH ratios for the structure in the crust and mantle by using nonlinear iterative scheme. The depth sensitivity kernels for inversion are numerically calculated. Depth sensitivity of the lowest frequency extends to depths beyond 500 km but the sensitivity of the overall data for the frequency band extends down to about 300km. We found that an appropriate selection of an initial model, particularly the depth of Mohorovicic discontinuity, is important for this inversion. The inversion result depends on the initial model and turned out to be non-unique. We have constructed the initial model from the CRUST 2.0. Inversion with equal weighting to each data point tends to reduce variance of certain frequency range only. Therefore, we have developed a scheme to increase weighting to data points that do not fit well after the fifth iteration. This occurs more often for low frequency range, 0.004-0.007Hz. After fitting the lower frequency region, the low velocity zone around a depth of 100km is observed under some stations such as KIP (Kipapa, Hawaii) and ATD (Arta Cave, Djibouti). We have also carried out an analysis on the resolving power of data by examining the eigenvalues-eigenvectors of the least-squares problem. Unfortunately, the normal matrix usually has 1-2 very large eigenvalues, followed by much smaller eigenvalues. The third one is often an order of magnitude smaller. The largest eigenvalue is always dominated by an eigenfunction that has the peak at the surface. It indicates that the ZH ratio is sensitive to shallow structure but it has limited form in resolving power for underlying structure. We will report on the details on the resolving capabilities of the ZH ratios.
S41A-0252
Viscoelastic Surface Waves
General theoretical solutions for Rayleigh- and Love-Type surface waves in viscoelastic media describe physical characteristics of the surface waves in elastic as well as anelastic media with arbitrary amounts of intrinsic absorption. In contrast to corresponding physical characteristics for Rayleigh waves in elastic media, Rayleigh- Type surface waves in anelastic media demonstrate; 1) tilt of the particle motion orbit that varies with depth, and 2) amplitude and volumetric strain distributions with superimposed sinusoidal variations that decay exponentially with depth. Each characteristic is dependent on the amount of intrinsic absorption and the chosen model of viscoelasticity. Distinguishing characteristics of anelastic Love-Type surface waves include: 1) dependencies of the wave speed and absorption coefficient on the chosen model and amount of intrinsic absorption and frequency, and 2) superimposed sinusoidal amplitude variations with an exponential decay with depth. Numerical results valid for a variety of viscoelastic models provide quantitative estimates of the physical characteristics of both types of viscoelastic surface waves appropriate for interpretations pertinent to models of earth materials ranging from low-loss in the crust to moderate- and high-loss in water-saturated soils.
S41A-0253
Lg and Pg Attenuation in the Middle East
In order to construct reliable frequency-dependent Q models for both Lg and Pg phases, we have analyzed approximately 8000 waveforms from approximately 500 events recorded by 11 permanent and temporary networks throughout the Middle East. We have developed a frequency dependent-Q tomographic model for both direct Lg and Pg. In general, we found low Q within the plateau and high Q within the stable Arabian plate. We found that the Pg Q does not vary as much as Lg Q, however, the general trend is the same. The frequency dependent exponent for Lg is less than that of Pg in the Middle East. Resolution tests of 2x2 cell size for Q tomography in our work indicate that we have very good resolution throughout much of the Anatolian Plateau resolution of anomalies larger than 200km in the northern Arabian plate. We have begun the work on validating our Lg and Pg attenuation models. We are developing a catalog of ¡°amplitude ground truth" seismic events. We are applying these events to test whether path corrections from our attenuation models can successfully predict the source spectra (i.e., moment and corner frequency). In order to develop a catalog of events with reliable event spectra, a search was performed on 373 events at 89 stations in the Middle East to look for doublets and repeating earthquakes. Only one pair met the same criteria for repeating events as were used in China. This is 1% of the seismicity, as compared to 9% from all the currently available waveforms in China (1,301 events out of ~14,000). This difference may be due to the absence of magnitude 3.0s from the Middle East data set; these events would probably have a greater chance of having similar waveforms due to simpler rupture areas and closer proximity to one another. This catalog of similar waveforms will also allow us to calibrate and adjust our model to better predict reliable source spectra.
S41A-0254
Modeling Nonlinear Site Response Uncertainty in Broadband Ground Motion Simulations for the Los Angeles Basin
The assessment of strong motion site response is of great significance, both for mitigating seismic hazard and for performing detailed analyses of earthquake source characteristics. There currently exists, however, large degree of uncertainty concerning the mathematical model to be employed for the computationally efficient evaluation of local site effects, and the site investigation program necessary to evaluate the nonlinear input model parameters and ensure cost-effective predictions; and while site response observations may provide critical constraints on interpretation methods, the lack of a statistically significant number of in-situ strong motion records prohibits statistical analyses to be conducted and uncertainties to be quantified based entirely on field data. In this paper, we combine downhole observations and broadband ground motion synthetics for characteristic site conditions the Los Angeles Basin, and investigate the variability in ground motion estimation introduced by the site response assessment methodology. In particular, site-specific regional velocity and attenuation structures are initially compiled using near-surface geotechnical data collected at downhole geotechnical arrays, inverse low-strain velocity and attenuation profiles at these sites obtained by inversion of weak motion records and the crustal velocity structure at the corresponding locations obtained from the Southern California Earthquake Centre Community Velocity Model. Successively, broadband ground motions are simulated by means of a hybrid low/high-frequency finite source model with correlated random parameters for rupture scenaria of weak, medium and large magnitude events (M =3.5-7.5). Observed estimates of site response at the stations of interest are first compared to the ensemble of approximate and incremental nonlinear site response models. Parametric studies are next conducted for each fixed magnitude (fault geometry) scenario by varying the source-to-site distance and source parameters for the ensemble of site conditions. Elastic, equivalent linear and nonlinear simulations are implemented for the deterministic description of the base-model velocity and attenuation structures and nonlinear soil properties, to examine the variability in ground motion predictions as a function of ground motion amplitude and frequency content, and nonlinear site response methodology. The modeling site response uncertainty introduced in the broadband ground motion predictions is reported by means of the COV of site amplification, defined as the ratio of the predicted peak ground acceleration (PGA) and spectral acceleration (SA) at short and long periods to the corresponding intensity measure on the ground surface of a typical NEHRP BC boundary profile (Vs30=760m/s), for the ensemble of approximate and incremental nonlinear models implemented. A frequency index is developed to describe the frequency content of incident ground motion. In conjunction with the rock-outcrop acceleration level, this index is used to identify the site and ground motion conditions where incremental nonlinear analyses should be employed in lieu of approximate methodologies. Finally, the effects of modeling uncertainty in ground response analysis is evaluated in the estimation of site amplification factors, which are successively compared to recently published factors of the New Generation Attenuation Relations (NGA) and the currently employed Seismic Code Provisions (NEHRP).
S41A-0255
Modeling the Dominant Noise Sources in Southern Italy
We present the result of noise correlation and source modeling using recordings from the CATSCAN (Calabrian- Apennine-Tyrrhenian/Subduction-Collision-Accretion Network) broadband deployment in southern Italy. Using more than two months of data, we were able to extract potentially dominant noise sources and validate them using various processing algorithms. Our mid-frequency data (0.06-0.1 Hz) indicate at least two dominant noise sources, likely originating from the Adriatic and Tyrrhenian coasts of Italy. The findings indicate that locating ambient noise source is possible even in complex regions, such as southern Italy, that are surrounded by bodies of water. Further analysis shows that asymmetric correlation stacks could be produced even under the influence of reasonably uniform source distributions. We believe that changes in the position of stations relative to sources may produce notable differences in the symmetry of the correlated data. Results from further numerical modeling of sources for different locations, and for different arrangements (both for location of sources, and number of sources), were compared against the actual data to validate the proposed locations of dominant noise source. While the source modeling results may not be unique, there is still supporting evidence that the two proposed dominant sources, the Gargano promontory and the eastern Tyrrhenian sea, are likely responsible for the observed Rayleigh waves. Further investigations of secondary sources using a larger data set are currently underway.
S41A-0256
Surface-Wave Amplitude Tomography from Worldwide Magnitude Measurements
Surface-wave amplitudes are routinely collected by the International Seismological Centre to determine earthquake magnitudes. These form a large database for amplitude tomography studies. I selected data from vertical instruments with data averaging around 20 seconds. These waves have their primary sensitivity in the crust and cover most of the globe. Average amplitude decay rates indicate an average global surface wave Q near 380 with cylindrical spreading on a spherical earth. The tomographic inversion estimates station corrections, event size corrections, and regional variations in surface-wave Q. Both site corrections and event size corrections have rms values of 0.2 magnitude units; although some events have corrections exceeding a full magnitude unit. Q values range from less than 100 to over 1000 with resolution as good as several degrees in densely covered areas. Regional variations show low Q values in tectonic and sedimentary regions, such as the western United States, the Middle East, and basins of the North China Block. Cratons show high Q in Canada, north Asia, and Europe.