S53B-0194 1340h
Crustal Structure in Southern Korea From Joint Analysis of Teleseismic Receiver Functions and Surface Wave Dispersion
We estimated crustal structures at eighteen broadband stations in southern Korea by combining receiver functions and surface wave dispersion with the genetic algorithm (GA). The trend of the Moho depths estimated from the GA inversion generally coincides with the topography of the surface, ranging from 26 km to 36 km in inland. However, the Moho depth distribution does not agree with the topography in the region around the Chugaryeong Fault located approximately in NNE-SSW direction in the central area of the Korean peninsula. The shallow Moho depth under this region may be related to magmatic rift and consequential crustal thinning processes along the fault caused by an extensional tectonic movement. Another discrepancy is found in the Gyeongsang Basin, which was formed by the sedimentary deposits accumulated during Cretaceous extension caused by retroarc environment. The thick crust may be caused by both the maturity of basin and underplating of magma materials after the closure of the basin. The average crustal velocity is variable from 6.02 km/sec to 6.51 km/sec over southern Korea. This indicates that crustal structures in southern Korea involve diverse velocity profiles, changing rapidly with distance. It is also a remarkable phenomenon that an obvious discontinuity of the velocity is observed at 8-10 km depths under several stations.
S53B-0195 1340h
Surface-Wave Focal Mechanism, Velocity and Q Models of the 29 May 2004 South Korea Earthquake
The South Korea earthquake event of 29 May 2004 was reported by the USGS as follows: mb = 5.3, Origin Time = 10:14:28.44, Latitude = 36.63N, Longitude = 129.93E. In this study, I will present the focal mechanism of this event, as well as the average velocity and Q models from inversions. Corrected Love-wave and Rayleigh-wave amplitude data from 15 GSN stations (MDJ, YSS, MAJO, TATO, QIZ, SSE, KMI, ENH, XAN, LSA, INCN, BJT, WMQ, ULN, and HIA) were used in the search for focal mechanism employing the technique of Nguyen and Herrmann (1992, SRL). Eigenfunctions were computed from the average crustal model that was obtained by inversion from surface-wave group velocities of these stations. Surface-wave attenuation coefficients were obtained using the technique of Tsai and Aki (1969). The focal mechanism of this event has a nodal plane with dip = 55 deg, slip = 65 deg, and strike = 160 deg. P-wave first motions were used to constrain the nodal planes. The seismic moment obtained is 3.79E+23 dyne-cm. The source depth is 11 km. The average velocity and Qb models obtained by inversions are tabulated as follows: Layer Number, Layer Thickness (km), P (km/sec), S (km/sec), Density (g/cm3), Poisson's Ratio, Qa/Qb: L1, 3.0, 4.4826, 2.5880, 2.3876, 0.25, 354/157, L2, 9.0, 6.1420, 3.5461, 2.7426, 0.25, 364/162, L3, 8.0, 6.2784, 3.6249, 2.7835, 0.25, 377/168, L4, 25.0, 7.2880, 4.2078, 3.0722, 0.25, 434/193, L5, 0.0, 8.1984, 4.6019, 3.3814, 0.27, 508/226.
http://home.cfl.rr.com/bao/AGU
S53B-0196 1340h
Study Of Attenuation Structure Based On The Dense Networks Around South Korea
There are few seismic studies of South Korea because this region has relatively quiet seismicity as a typical intraplate environment and lack of seismic stations (e.g. Lee et al., 2003). Recently, however, the densification of regional seismic network has been proceeded in S. Korea since 1999. Based on the dense network data, we first reported not only focal mechanisms of nationwide microseismicity (Chung et al., 2004) but also body wave attenuation structure for several regions (Chung and Sato, 2001; Kim et al., 2004) in S. Korea. These attenuation studies using the coda normalization method (CNM) showed very low {\it Q$_{P}$$^{-1}$} and {\it Q$_{S}$$^{-1}$} values comparable to those of seismically stable regions. By using the CNM, we also studied {\it Q$_{Lg}$$^{-1}$} value (Chung and Lee, 2003), which is similarly low value with {\it Q$_{S}$$^{-1}$}. To compensate the lack of data, the source pair/receiver pair method (SPRP) was applied to 71 earthquakes for further {\it Q$_{Lg}$$^{-1}$} attenuation study (Chung et al., 2004). The SPRP provided nearly 1500 source-receiver pairs and reproduced {\it Q$_{Lg}$$^{-1}$} for all frequencies less than 0.002, which is a reasonable value for a seismically inactive region. Lately, we have continued to analyze {\it Lg} attenuation structure by adding dense Japanese network located in Kyushu, near southeast of S. Korea. In this analysis, we used the reversed two-station method for collinearly aligned event and receiver pairs to correct azimuthal effect of {\it Lg} in the previous SPRP study. {\bf Refernces} Chung, T.W. and, H. Sato, 2001, BSSA, v.91, 1867-1874. Chung, T.W. and, K. Lee, 2003, BSSA, v.93, 1401-1406. Lee, K., N.S. Chung, and, T.W. Chung, 2003, BSSA, v.93, 2131-2145. Kim, K.D., T.W.Chung, J.B. Kyung, 2004, BSSA, v.94, 1070-1078. Chung, T.W., J.C. Park, and W. Kim, 2004, PEPI, in submitted. Chung, T.W., Y.K. Park, I.B.Kang, and K. Lee, 2004, BSSA, in submitted.
S53B-0197 1340h
SH Velocity Structures in the Transition Zone Beneath South America and Northeast China
The stable phase assemblages in the transition zone region are sensitive to mantle composition, temperature and chemical interactions between the olivine- and pyroxene-normative components. With the accumulation of in-situ measurements of elastic properties and accurate determination of phase equilibria data, we can now explore various chemical interactions and quantitatively calculate seismic velocity profiles for various mineralogical models. Mantle composition and thermal models, and their regional variations, can thus be quantitatively constrained by jointly modeling mineral physics data and seismic observations sampling different regions. In this study, we constrain fine seismic structures in the transition zone beneath South America and northeast China, and explore compositional and thermal models in these two regions based on mineral physics modeling. We use the triplicated phases recorded in the epicentral distance range of 10$^{\circ}$~- 30$^{\circ}$~ for a deep event occurring in South America subduction zone and two deep events occurring in northeast Asia. The triplication phases near a discontinuity place tight constraints on the velocity gradients above and below the discontinuity and the velocity jump across the discontinuity. The termination distance of the AB phases (the branch turning above the discontinuity) is sensitive to the velocity gradient above the discontinuity and the move-out of the CD phases (the branch traveling below the discontinuity) is controlled by the velocity gradient below the discontinuity. Seismic observations indicate that the SH velocity gradients above the 660 km discontinuity are larger than the Preliminary Reference Earth Model (PREM) beneath both South America and northeast China, But model differences exist for the two regions. The gradient above the 660 km discontinuity is larger for South America than for northeast China; while the model for South American has a shear velocity jump across the discontinuity and a velocity gradient below this discontinuity similar to PREM, the model for northeast China has a velocity jump larger than PREM and a velocity gradient smaller than PREM below the discontinuity. The observed triplications near the 410-km discontinuity beneath northeast China also indicate that the region has, compared to PREM, a larger velocity gradient above the 410 km discontinuity, a larger velocity jump across the 410 km discontinuity and a smaller velocity gradient below the discontinuity. In joint modeling of the mineral physics and seismic data, we calculate seismic velocity profiles for a variety of compositional and thermal models using the Tr660 program developed by Weidner and Wang (1998). We then obtain the most appropriate compositional and thermal models based on the criterion that their predicted seismic profiles best match those inferred from the seismic data and the synthetics of those predicted seismic profiles best fit the seismic data.
S53B-0198 1340h
Crustal Structure of China and Surrounding Regions from P and S Wave Travel-time Tomography
We have constructed a 3-D P and S velocity structure of the crust and uppermost mantle of China and surrounding areas using high quality first arrivals extracted from the Annual Bulletin of Chinese Earthquakes (ABCE). We used 345,000 P wave, and 230,000 shear wave arrivals. A preliminary 3-D model has been generated by combining 2400 1-D models (Sun et al., 2004). By applying the tomography method (Zhao et al., 1992), we obtain a detailed 3-D P and S wave velocity model of the crust and uppermost mantle in China with a spatial resolution of 1$\deg$ $\times$ 1$\deg$ in horizontal direction and 10 km in depth. This method is adaptable to a velocity structure which includes several complex-shaped velocity discontinuities and allows 3-D velocity variations everywhere in the model. Velocity discontinuities represent known geological boundaries such as the Moho and/or a subducting slab boundary, etc. A 3D grid net is set up in the model to express the 3D structure. Velocity perturbations at the grid nodes are taken as unknown parameters. The velocity perturbation at any point in the model is calculated by linearly interpolating the velocity perturbations at the eight grid nodes surrounding that point. To calculate travel times and ray paths accurately and rapidly, an efficient 3D ray-tracing technique is employed to iteratively use the pseudo-bending technique (Um and Thurder, 1987) and Snell's law. Station elevations and the sediment layers are taken into account in the ray tracing. The LSQR algorithm (Paige and Saunders, 1982) with a damping regularization is used to solve the large and sparse system of observation equations. Our tomographic model provides new insights into the geological structure and tectonics of the region, such as the lithological variations and large fault zones across the major geological terrenes. The velocity images of the upper crust represent the surface geological and topographic features accurately. Basins and the Tibetan Plateau are clearly depicted in our model. The crust beneath the Precambrian regions (Tarim Basin, Ordos Basin, Sichuan Basin and western half of Songliao Basin) shows high velocity anomalies. The highest velocity anomaly is beneath the Sichuan Basin. The Bohai Gulf shows both high and low velocity anomalies due to a major Cenozoic rift system through it. P velocities are lower in the lower crust beneath the northern part of the South China Block than that beneath the southern part of the South China Block. The Indochina Block demonstrates low velocities both in the crust and uppermost mantle due to volcanism. The Pn velocities in the Tibet area are higher than those in other areas largely due to thicker crust thickness. Both 3-D travel time and broadband seismogram fits show that our model is well determined and can be applied to determine the source parameters of earthquakes and to generate synthetic travel times.
S53B-0199 1340h
Three-component receiver function study of seismic array data from the Tibetan plateau
Abstract. Dense seismic arrays recently deployed on the Tibetan plateau have been widely used to study crust and mantle structure through traditional receiver function analysis. In this study we stack over all vertical waveforms associated with one earthquake to form the best estimate of the effective source time function. Stacking significantly cancels out the scattered wavefield after the direct P wave. The deconvolution of the stacked waveform from all three components then provides information both for P wave velocity structure from analysis of the vertical component and shear wave velocity structure from the horizontal components. Preliminary results show that three component receiver functions beneath INDEPTH-III stations are quite complicated. Strong reverberations within the upper crust are observed for most of the stations implying complicated 3D crust structure beneath the array. Vertical component receiver functions are often quite complex showing secondary arrival comparable in amplitude to the first P wave arrival suggesting large P wave reverberations in the shallow crust. Corresponding horizontal receiver functions are also very complex. It is likely that these upper crustal wave propagation effects much converted arrivals from deeper levels in the crust and mantle.
S53B-0200 1340h
High-Resolution Subducting Slab Structure Beneath Northern Chile Using the Double-Difference Tomography Method
The body-wave velocity structure beneath northern Chile was obtained using the double-difference tomography method with the relative arrival times derived from absolute catalogue data. We used 3992 events distributed within the coverage area of the RESISTE local permanent network (18.1$\deg$-20.2$\deg$S and 68.8$\deg$-70.5$\deg$W) between 1997 and 2000. During this period about 75.000 phase arrivals ($\sim$50$%$ each P and S waves) were observed at 14 stations. From these absolute picks, about 415.000 catalogue differential times $\sim$50$%$ each P and S waves) were constructed from event pairs with interevent distance $<$10 km observed at common stations. Both the mantle wedge and subducting slab are well sampled by the absolute and differential ray paths. In this study we used 1728 nodes distributed between the surface and 140 km depth in 16 layers, with an average separation of 10 km. The inversion strategy started from with 1D velocity model, and the velocities obtained in the 1D inversion were used as the initial model for the 3D inversion. The inversion was done for P and S wave velocities, with the initial 3D model starting with a Vp/Vs ratio of 1.73, for 12 iterations. The hypocenters and Vp and Vp/Vs models exhibit more detailed features of the subducting slab and the continental lithosphere than those observed before with the same dataset, but with different relocation and tomography inversion techniques. The intermediate depth double seismic zone dipping at $\sim$30$\deg$E, reported previously, appears with a clear separation of 20 km, and the thickness of each plane is less than 5 km. Vp between the two layers of the double seismic zone is higher than that obtained for each layer of seismicity, similar to that observed in northern Honshu, Japan. The double seismic zone is better observed close to the Arica Bend, and its expression is diminished southward. The shallower structures obtained are in good agreement with active fault systems present in the forearc of the region.
S53B-0201 1340h
Implications of Seismic Waveforms: Complex Physical Properties Associated with Stagnant Slab
Regional broadband waveform modeling can elaborate the structure in the areas where seismic tomography studies captured the bulk of high velocity anomaly (HVA) in the upper mantle transition zone. Our recent study shows that P-wave velocity model M3.11 or M2.0, if not iasp91, generally explains regional waves that strongly sampled the transition zone in the northwestern Pacific subduction zones (Tajima and Grand, 1995, 1998). Here M3.11 is consistent with the thermo-chemical condition of a deflected cold slab in the transition zone, and explains the structure in the southern Kuriles to northeastern Japan subduction zones. M2.0 has HVA in the deeper part of the transition zone without broad depression of the 660 km discontinuity, and explains the structure beneath the northern most Philippine Sea Plate. It is noted, however, some waves observed for deep focus events whose source processes are basically simple, show substantial broadening after the direct P arrivals and were not modeled by either of these models. The sampling rays of these data are not very different from those whose waveforms were used to derive M3.11 or M2.0. The waveform broadening probably indicates excitations of reflected, refracted or converted waves at the velocity heterogeneities associated with the slab. Whether the heterogeneities are located around the slab boundaries or involve in the chemical composition within the slab such as metastable phases has not been determined yet as the present data sampling is not dense enough to draw definite conclusions. However, we show how sensitive regional seismic body waves are to the structure associated with stagnant slab, and present some models that were tested for the causes of waveform broadening.
S53B-0202 1340h
3D Body Wave Velocity Tomography in Southern Peru: Seismotectonic Implications
The studied region corresponds to the southern segment of the 1868 rupture area, that did not break with the last Mw=8.4 Arequipa earthquake in southern Peru. A temporary network of 19 short period, continuous recording seismic stations was deployed (16.5$\deg$-18.5$\deg$S; 69.5$\deg$-72$\deg$W) between December 2002-March 2003 in this region. The 1093 select events were used for a joint hypocentral and velocity structure inversion. The first obtained 1D velocity model was used as the initial model for the 3D inversion, consisting in 231 blocks distributed along layers separated by 10 km for depths lower than 80 km, and by 20 km for depths between 80 and 160 km. North of the Arica Bend, between the Coast and Andean Range the high Vp velocity and Vp/Vs ratio observed at depths lower than 10 km could represent the Precambrian basement uplifted by the Incapuquio sinistral fault system, which develops a positive flower structure typical for transpressional zones, that raised the basement. This behavior is in good agreement with the uplift of the Cambrian metamorphic complex, in the Precordillera south of the Arica Bend (northern Chile), by the west-vergent thrust system. Beneath the anomalous high velocities (between 20 and 30 km depth), a low velocity zone (Vp$<$4.5 km/s, and Vp/Vs$<$1.5) can be observed; this LVZ can be associated with a zone of continuous metamorphism and/or partial melting that the decouples the upper-crustal imbrication from a thickening lower-crust. The Wadati-Benioff zone exhibits a dip angle of about 20$\deg$ between 10 to 60 km depth, and about 45° for depths higher than 80 km. The double seismic zone known for northern Chile is not present in southern Peru. Cross sections perpendicular to the trench, define a very active seismic zone which dips about 45$\deg$ trenchward, and is almost perpendicular to the subducting slab between 0 and 60 km depths, similar to the one existing in northern Chile, but with a lower rate of seismicity; the higher activity of this zone in southern Peru can be associated with the 2001 earthquake post-seismic period.
S53B-0203 1340h
Upper-mantle attenuation structure as revealed by surface-wave amplitudes
Three-dimensional models of attenuation in the Earth's interior can provide important constraints on the thermal and chemical state of the planet, particularly when interpreted together with velocity models. Large lateral variations in attenuation (1/Q) will affect velocity dispersion in a way that also varies laterally and must be considered when constructing and comparing velocity models derived from seismic observations from different portions of the seismic frequency band. However, attenuation in the mantle is a difficult quantity to study due to the various factors, in addition to anelasticity, that can affect wave amplitude, including focusing and defocusing by elastic structure, uncertainty in the source scalar moment and radiation pattern, and inaccuracies in the instrument response. These extraneous effects must be accounted for before the amplitude data can be confidently interpreted in terms of anelastic structure. We present a 3-D model of shear attenuation in upper mantle derived from more than 250,000 measurements of fundamental-mode surface-wave amplitude that were recorded at 170 stations across the globe. The model shows a strong correlation with surface tectonics in the uppermost mantle that becomes weaker in the transition zone. The amplitude data are corrected for source, instrument, and focusing effects, a critical consideration that isolates the signal of attenuation in the data and enables the development of this improved model. We also invert measurements of phase delay together with the amplitudes to obtain 2-D maps of surface-wave attenuation and phase velocity. We show that as a result of focusing, the amplitude data contain significant and retrievable information about elastic velocity that should be included as an additional constraint in the development of phase-velocity maps.
S53B-0204 1340h
Mantle Discontinuities Beneath the United States and Gulf of Mexico From ScS Reverberations
We modeled mantle discontinuities beneath the United States and Gulf of Mexico using multiple ScS reverberations from Central and South American earthquakes captured by 65 stations of the IRIS Global Seismograph Network, GEOSCOPE, and the Terrascope Network across the United States. Data were separated into 8 geographic paths, 6 from Central American events and 2 from South American events, creating a fan pattern across the United States and Gulf of Mexico. Discontinuity depths and impedance contrasts across the discontinuities were calculated using the hierarchical waveform inversion method described by Revenaugh and Jordan (1989, 1991). The path-averaged depth of the 410-km discontinuity varies between the paths and is particularly shallow ($\sim$395 km) beneath the eastern United States. The path-averaged depth of the 660-km discontinuity also varies, but only by about $\pm$ 5 km. The 520-km discontinuity is seen in all of the paths, though the depth of the discontinuity and the impedance contrast across the boundary vary significantly across the study area. The easternmost paths have strong 520-km discontinuities, with impedance contrasts greater than 2.5%. The mid-continent and western paths have impedance contrasts of less than 2.0%. Further modeling of the ScS reverberations includes mantle discontinuities below the transition zone as well. In the mid-continent region, a discontinuity near 1375-km is present in 2 paths, and a 1525-km discontinuity appears in an adjacent path to the west. Further east, two paths show a discontinuity between 920-960 km. The easternmost path also contains a discontinuity below the transition zone, with an approximate depth of 1125 km. Mantle discontinuities below the transition zone are not seen beneath the western United States. Envelope stacks, with the modeled discontinuities and noise subtracted, were created for each path, which describe the degree of residual scattering in the mantle with depth. All of the profiles have a paired-caternary curve appearance, with a lower amount of scattering in the transition zone, a rise in scattering in the mid-mantle, and a high degree of scattering at the crust and core-mantle boundaries.
S53B-0205 1340h
Seismic Attenuation Anisotropy in the Crust Beneath ANZA (Southern California)
We present a method to study crustal anisotropy using polarization-dependence attenuation of {\it S}-coda waves. We have tested our algorithm on band-passed ( 2 - 4 Hz ) waveforms of a {\it M$_{L}$} = 5.1 local crustal earthquake ( {\it h} = 18.7 km ) recorded by the ANZA broadband network in southern California. We analyzed the records of 10 stations, of which 7 are $<$ 20 km away. For all stations {\it S}-coda waves with a polarization direction of $184\deg$ ( from North ) have the maximum value of the quality factor {\it Q$_{c}$} ( or the smallest attenuation ). This orientation is in agreement with the axes of the maximum compression obtained from fault mechanism analysis ( http://www.trinet.org/shake/9718013/zhu.html ) and results of seismogenic deformation field ( Unruh {\it et al}., 1996, {\it JGR} ) in southern California. The polarization directions with the biggest values of {\it Q$_{c}$} orient in a plane with the strike $159\deg$ and dip $74\deg$. This result suggests that the cracks are oriented $\sim$ NWN - SES and they are responsible for the induced attenuation anisotropy in the crust beneath ANZA. The cracks are parallel the maximum compressional stress. We found the coefficient of attenuation anisotropy at 3 Hz equals 9%.
S53B-0206 1340h
The Onshore-Offshore LARSE I Transect: San Clemente Island to the Mojave Desert--Crustal blocks and the Moho
The 1994 seismic transect of the Los Angeles Region Seismic Experiment (LARSE I) crosses the Inner California Continental Borderland, Los Angeles basin, San Gabriel Mountains, and Mojave Desert. In addition to LARSE I seismic refraction and reflection data, we have used gravity, geologic, petrophysical, oil-test well, earthquake, and receiver-function data and models to constrain our geologic/tectonic model of this transect. Several blocks can be defined along the transect and are described from south to north: (1) The Borderland block, seaward of the Palos Verdes fault, is characterized by generally thin Cenozoic sedimentary rocks ($<$ 1 km), and a basement (2.8-5.9 km/s in the upper few km, increasing to 6.4 km/s at depth) that is interpreted as chiefly Catalina Schist. A lower-crustal layer (6.6-6.8 km/s), 4-5 km thick, is 18 km deep, and Moho appears relatively flat at ~22-km depth. (2) A transitional block, between the Palos Verdes and Newport-Inglewood faults, is characterized by moderately thick Cenozoic sedimentary rocks (3 to 4 km), and a basement of chiefly Catalina Schist (velocity poorly constrained). A lower-crustal layer thickens landward from ~5 to ~10 km thickness above a Moho that deepens in a sharp ramp from ~23 km, slightly west of the Palos Verdes fault, at the shelf edge to ~28 km at the Newport-Inglewood fault. (3) The Los Angeles basin block, between the Newport-Inglewood and Sierra Madre faults, is characterized by thick Cenozoic sedimentary rocks (10 km maximum in Los Angeles basin). Basement is Peninsular Ranges igneous and metamorphic rocks with velocities ranging from 6.1 to 6.8 km/s at or near the sedimentary contact to 6.8 to 6.9 km/s in the lower crust. Moho is flat at 28-km depth. (4) The San Gabriel block, extending from the Sierra Madre fault to the San Andreas fault (SAF), consists of 3 distinct rock units. Mylonitic lower-crustal rocks, ranging in thickness along the transect from 0 to 5 km, overlie Pelona Schist along the inactive Vincent thrust fault. The Pelona Schist (similar to the Catalina Schist) is characterized by a relatively low velocity (~6.0 km/s) to ~20-km depth. A bright reflective zone, interpreted as a fluid-lubricated, active decollement, is observed at the interpreted base of the Pelona Schist. The lower crust (6.7 km/s ?) has a maximum thickness of 14 km. Moho deepens in a sharp ramp in the southern block from 28-km depth to 36-km depth. Moho depth remains at ~36 km depth from 20-km south of the SAF to 10-km north of the SAF. (5) The Mojave block is characterized by very thin Cenozoic sediments (0 to ~1 km, within a few km of the transect) overlying a chiefly granitic basement (2.5 to 5 km/s near the surface). This block has no high-velocity lower crust ($>$ 6.5 km/s), and Moho is flat at ~32 km, except in the southern Mojave Desert, where it deepens in a sharp ramp to ~36 km at the SAF. The shape of the Moho as given above has been established by raytracing of PmP; further constraints and/or alternate models are being developed from a receiver-function study.
S53B-0207 1340h
Evidence for Localized Down Welling of Lithospheric Material Through the Upper Mantle Transition Zone Beneath Southern California
Common conversion point stacking of receiver functions has been applied to data from 21 seismic stations located throughout southern California to produce an image of the upper mantle transition zone. This method has gained wide acceptance in the investigation of the upper mantle transition zone (TZ) which is generally believed to be the depth range in the mantle where olivine undergoes a series of phase changes. The olivine to spinel phase change is elieved to be responsible for the 410 km discontinuity and the perovskite and magnesiowüstite is generally accepted to be associated with the 660 km discontinuity. An investigation of depth variations in the 410 and 660 km discontinuities is generally considered to be a means to identify variations in mantle temperatures associated with the respective phase change. That is that the 410 will become shallow in cooler regions of the mantle and deeper in warm regions. The 660 will behave oppositely in response to thermal variation in the mantle. As a result we would find a thick transition zone in cooler regions and a thin transition zone in warm regions. The transverse ranges in southern California have been subject to much investigation because they do not appear to have an adequate crustal root to account for observed topography. Humphreys and Hager (1990) have found evidence of very localized subduction or a "drip" of lithospheric material beneath this region, as a result of convergence due to a kink in the San Andreas fault system, which can account for the observed topography. Using seismic tomography they imaged a high velocity anomaly (generally considered to be evidence for cooler mantle temperatures) directly beneath the transverse ranges to a depth of almost 300 km. We have found in receiver function images that there is an anomalously thick TZ directly beneath this "drip" beneath the transverse range which would indicate a small localized low temperature anomaly no more than 100 km across. We interpret this feature as evidence that the "drip" feature penetrate the TZ or that the drip feature became detached and is stagnant in the TZ. Our TZ image also suggests other localized down-welling features beneath the Southern Sierra Nevada Range.
S53B-0208 1340h
CANOE: A Broadband Array in Northwestern Canada
The Canadian Northwest Experiment (or CANOE) is a nearly sixty broadband-instrument array extending from the Slave Craton in the Canadian NWT, across the Canadian Rockies in northern British Columbia and Yukon and south to Edmonton Alberta where the FLED (Florida to Edmonton) array terminated. The array crosses 4 Ga of geologic time and a series of compressive orogens undisrupted by later periods of extension or extensive hotspot volcanism. Coupled with excellent shallow structural control from Lithoprobe active-source transects, CANOE offers an unparalleled window into deep continental lithosphere structural expression and history. The array also offers excellent deep-mantle sampling of the central Pacific and Hawaii. A subset of the array was installed in May, 2003. The remaining two-thirds of the array were deployed in May and June of 2004 and will remain until October, 2005. Array endpoints are anchored by permanent stations of the CNSN; typical station spacing within the array is less than 50 km. Data are recorded continuously at 20 samples per second on a mixture of Guralp 3T, 3ESP and 40T instruments. We will present specifics of the deployment, examples records of the array and some preliminary applications of the data. Instruments for CANOE were provided by PASSCAL/IRIS. The members of CANOE wish to thank the PASSCAL Team for training, extensive field assistance and critical logistical support.
http://canoe.asu.edu
S53B-0209 1340h
Large Proterozoic Upper Crustal Intrusions Within the Eastern Athabasca Basin of Northern Saskatchewan, Canada
Regional and high-resolution reflection surveys, in three separate areas of the eastern part of the Paleoproterozoic Athabasca Basin, have identified zones of anomalously high reflectivity. These bright reflections extend through all the seismic profiles, some with lengths of over 160 km. The depths of these reflective bodies range form ~ 6 to 13.5 km (2.0 - 4.0 sec TWT). In several locations, the reflector is offset by shear/fault zones. In some localities, these shear/fault zones ascend to or approach the surface. The highly comparable seismic signatures of the isolated surveys, suggest that these are tabular bodies that extend over at least 25,000 km2. The anomalous large amplitude of the seismic signal indicates significant acoustic contrast in a vertical zone of 50 to 150 m. The origin of these enigmatic sheet-like complex zones is not resolved. One interpretation is that they are sill-like bodies, subsurface expressions of the circa 1.265 Ga (post-Hudsonian) Mackenzie diabase suite. However, instantaneous frequency, amplitude strength, and other seismic signal attribute analyses, reveal that a variety of geologically acceptable acoustic contrasts, including remnant brine fluids, cannot be ruled out. The region of the anomalous subsurface reflectivity property is under extensive exploration for unconformity-type uranium deposits, thus resolution of the true nature of this complex interval of the upper crust is of major importance.
S53B-0210 1340h
USArray Array Network Facility (ANF): Metadata, Network and Data Monitoring, and Quality Assurance During the First Year of Operations
The deployment of Transportable Array stations for the Earthscope USArray project has begun and will ramp up to 400 real-time telemetered stations over the next four years. The role of the Array Network Facility (ANF) in the USArray project is to guarantee delivery of all Transportable Array stations (400) and telemetered Flexible Array stations (200) to the IRIS Data Management Center, ensure proper calibration and metadata are always up to date, and provide quality control for all data. In support of these goals, we use the Antelope software package to facilitate data collection and transfer, generation and merging of the metadata, monitoring of dataloggers, generation of noise spectra, and analyst review of individual events. The SRB, and a newly developed Antelope/SRB interface, is used for off-site backup of the data. A freeware package, Nagios, adapted to include Antelope plug-ins, as well as newly developed in-house tools allow network and data flow monitoring. Quality control checks include: daily review by an analyst to look for obvious discrepancies in channel polarization or change in noise characteristics; associations made against regional network and global bulletins to help spot timing errors; spectral noise plots; and daily reports of outages and data gaps.
S53B-0211 1340h
Bridging Long-period and Hand-picked Arrival Times of Teleseismic Body Waves
It is well known that the S travel times of long-period Earth models delay several seconds to those of standard travel time tables. This discrepancy can be largely removed by taking the physical dispersion into consideration, where the reference frequency of standard travel time tables or short-period Earth models has been conventionally taken to be 1 Hz. Oki et. al. [2004] measured S-P travel times on the unfiltered broadband seismograms by cross-correlating the observed S waveform with those synthesized from the observed P waveform. The measured S-P times were added to the hand-picked P times to obtain the broadband S times, some of which were able to compare to the hand-picked S times. A reference frequency of 2 Hz was required to make the broadband and hand-picked S times consistent with each other. On the other hand, Bolton and Masters [2001] measured P and S travel times by cross-correlating the observed low-pass filtered P and S waveforms with those synthesized for a reference Earth model. The reference frequency was set at 0.5 Hz. We compared these two datasets in different frequency ranges to see their mutual consistency. In the two datasets there existed discrepancies of about 2, 4 and 2 sec for P, S and S-P times, respectively. After a correction for finite source duration for the long-period data, the P arrival times became consistent between the two datasets. There still remained 2 sec of discrepancy for S and S-P times. Assuming that this discrepancy represents the effect of physical dispersion, we searched for the reference frequency that can remove the 2-sec offset. This search yielded a value of 2 Hz in agreement with the result of Oki et. al [2004]. The reference frequency of 2 Hz can bridge consistently long-period, broadband and hand-picked S arrival times, and hence can be regarded as the characteristic frequency defining the onset of teleseismic body waves.
S53B-0212 1340h
Automated Analysis of Seismic Data Quality at the IRIS DMC
We present an automatic quality control system (QUACK) that monitors seismic data flowing into the IRIS DMC BUD (Buffer of Uniform Data). The BUD currently receives 4,000+ seismic channels and records 12+ GB of data per day. With this automated system, problems within this large data set can be more easily identified. QUACK is written in Java and features a modular design that allows third parties to write quality control software "plug-ins" that measure custom quality parameters. QUACK stores the plug-in measurements in a relational database from which they can later be retrieved. Currently, we are running software plug-ins that measure or detect: data availability, daily signal mean and variance, timing problems, PSD noise levels, and dead and missing channels. QUACK consists of three primary software components: (1) a framework which orchestrates the operation of the software plug-ins, (2) a reporting mechanism which sends daily e-mails to system operators reporting data issues discovered by the plug-ins and (3) a web-based query tool which allows users to explore different aspects of the recorded quality control measurements.
http://www.iris.washington.edu/
S53B-0213 1340h
Distributed Computing for Seismic Data Retrieval using FISSURES DHI Client FMI Implimented from MatSeis 1.9
Three software development efforts have recently combined to allow a seamless method for obtaining data from the IRIS DMC over the Internet and analyzing it in MATLAB. During the past few years IRIS has constructed the initial framework and definitions for FISSURES servers and Data Handling Interface (DHI) clients. FISSURES uses the distributed computing language Common Object Request Broker Architecture (CORBA), which allows software systems to work across the Internet in a platform independent and computer-language neutral manner. Starting from the FISSURES object model three seismic services have been defined: event, network and seismogram. From these three object types a DHI client is used to access information through FISSURES services. Using the University of Washington developed FISSURES-Matlab-Interface (FMI) DHI client, all three FISSURES objects can be accessed via the Internet and downloaded into local MATLAB objects, allowing access to earthquake event catalogs, seismograms, network and instrument information. The DHI client FMI can be implemented from MatSeis. MatSeis, developed at Sandia National Laboratory, is a MATLAB based, Graphical User Interface (GUI) controlled software package with seismic data visualization, signal processing, and database importing and exporting capabilities, to list just a few. In this poster, we will discuss the seismic data retrieval currently prototyped for the IRIS DMC, as well as the planned connections to the Northern California Earthquake Data Center (NCEDC), Southern California Earthquake Data Center (SCEDC), and South Carolina Earth Physics Project (SCEPP) data center.
S53B-0214 1340h
The GEOSCOPE Program : state of the art in 2004
The GEOSCOPE program was launched in 1982 by the National Institute of Sciences of Universe (INSU), a department of the French National Center of Scientific Research (CNRS), at the instigation of the Institute of Physics of the Earth of Paris (IPGP). The purpose of the GEOSCOPE program was the installation of about 25 stations well distributed worldwide (in particular in the southern hemisphere), in the standard configuration defined by the FDSN (very broad-band 24 bit, continuous recording at 20sps). The GEOSCOPE program is operating 28 digital very-broadband stations. Data from large events are teletransmitted for some stations (by phone line or through internet) and made available within one day. A satellite transmission system is now working, in cooperation with the french military agency CEA/DASE, in cooperation with CTBTO (Dzumac in New Caledonia). An agreement between GEOSCOPE and CTBTO allows us to get data continuously and with a low gain. The next CTBTO stations to be installed are ATD (Djibouti) and MBO (Senegal) in 2005. In terms of siting locations, the aim of the GEOSCOPE program is almost fulfilled; we plan to install a new station in MARQ (Marquesas Islands), one in Russia at high latitude at VOR (Vorkouta), one in Patagonia (COY in Chile), in order to fill some geographical gaps in the southern hemisphere. We installed in 2004 a joint station with IRIS and the CTBTO at TRIS (Tristan Da Cunha), in the southern Atlantic ocean. DCC in Antarctica is a joint EOST-Strasbourg/Concordia-Italy station. Our goal is now to replace our old digitizers by Quanterra ones, and to transmit all data in near real-time. At present 4 stations are sending their data in quasi-real time (DZM, ECH, SSB, FDF and RER). An inversion method for the fundamental mode Rayleigh wave spectra has made possible the rapid determination of the mechanism and the seismic moments. This determination is done routinely for all events with Ms > 6.5 from the teletransmitted stations data. The estimate Power Spectral Density plots have been computed for each station and are available on the Web site. Some small to medium earthquakes are not detected and thus are not referenced in the earthquake catalogues. Most of these events are in the southern hemisphere where the lack of seismic stations creates a detection sensitivity gap. We estimate that more than a hundred southern hemisphere events per year with magnitude between 4.5 and 5.5 go undetected by the worldwide networks. We use a surface wave analysis method to effectively detect and locate these earthquakes, particularly near-ridge events. Most GEOSCOPE stations are equipped with STS1 seismometers, only a few ones with STS2 seismometers. We are planning to move some stations from the northern hemisphere to the southern one, inorder to fill a geographical instrumental gap. We are equipping all stations with seismometers but also with microthermometers, microbarometers, in order to clean the seismic signal and to study potential correlations between the seismic signal and these environmental parameters. Some of our stations have long seismic time series (SSB in France and TAM in Algeria). Removing the atmospheric pressure effect is absolutely necessary for scientists using low frequency free oscillation modes, helping studies on the 'hum'.
S53B-0215 1340h
Providing Seismic Data to the Public: Evaluation and Impact of IRIS/USGS Museum Displays
IRIS data has had an important impact on the public understanding of geophysics as well as on research advances. One example of this is the IRIS/USGS museum display program, where 16 million museum visitors per year can view near-real-time earthquake locations and ground motions. An evaluation of displays at the American Museum of Natural History in New York City (AMNH) and the Smithsonian Institution National Museum of Natural History in Washington, DC (NMNH) was conducted in the summer of 2004 to assess the display's ability to increase the public's understanding of seismology and to determine how the displays might be improved. The evaluation involved tracking and timing museum visitors to see what attracted them and what held their attention. The tracking and timing was conducted within a single gallery in both museums. Visitors were also interviewed in order to learn what they liked and disliked about the display, and to assess what they learned about seismology. The results show that the IRIS/USGS display was the top attraction in both the AMNH and NMNH galleries (in terms of the percentage of visitors that stopped at the exhibit). It was also first at both galleries for cumulative visitor stop time (the sum of the time spent by all visitors) and in the total number of visitors counted in front of the display during random sweeps of the gallery. In both galleries, visitors were attracted to the display most often by the map on a large plasma monitor that shows the last 2 weeks of seismicity with alternating views of the of the world and the continental US. Smaller numbers of visitors were attracted by the triple-drum recorder. When asked what they liked about the display, the greatest number of visitors (31% at each museum) replied that they liked that it is real-time/up to date. Visitors also liked the map of recent earthquakes, the list of earthquakes on the small monitor and the triple drum. A large majority of the visitors were interested to know that similar information is available on-line. When asked what they found most interesting or surprising, the frequency of earthquakes was listed most often at both the AMNH (49%) and NMNH (54%). While the majority of visitors understood that the triple-drums display real-time information, they were less clear about what was being presented, with only 38% (NMNH) and 16% (AMNH) understanding that each drum can record earthquakes from all over the world. Small additions of contextual information could greatly increase visitor understanding of this part of the exhibit. The evaluation shows that the general public is interested in presentations of real-time earthquake activity that combine state-of-the-art plasma monitors and traditional mechanical displays. Interview responses show that the displays are successfully conveying the message of an active planet that is continually in motion.
S53B-0216 1340h
The Role of 3D Crust and Mantle Structure in Predicting North-American Seismograms
We compare the predictive capabilities for seismic waveforms of several 3D tomographic models for the North American upper mantle in an attempt to more narrowly identify the data and theory necessary to model upper mantle structure in efficient ways. To this end, we compare recent (1995-2004) North American S and Rayleigh wave trains to synthetic wave trains computed with different velocity models and different modeling strategies. Synthetics computed, using great-circle averaging, for 3D model NA00 match the observed recent waveforms better than those for standard 1D continental and oceanic models. Despite NA00 being significantly less smooth than global models, for most waveforms NA00 does not account for all of the observed positive and negative delays. The improvements in fundamental mode matching provided by NA00 diminish with increasing frequencies. We identify two possible reasons: 1) High-frequency energy cannot be modeled by great-circle averaging, or 2) the model parameterization is not suitable for shallow, crustal depths. We will investigate the role of scattering and of the crustal model in this waveform matching. Data from Earthscope's USArray will be sufficiently dense to model lateral variations in the crust and depth to the Moho discontinuity. We prepare for these data by targeting regional crustal structures, such as the Illinois Basin. To allow for the resolution of crustal structures, we adopt a new grid capable of supporting fine crustal structure, and analyze high-frequency Rayleigh waveforms from regional earthquakes down to magnitudes of 4. Combining regional surface waves with the data that produced model NA00, which include the constraints on crustal structure by Chulick and Mooney (2002), we aim at 1) learning more about the crust in targeted regions (e.g. the Illinois Basin), 2) a better separation of crust and upper-mantle structures, and 3) the importance of the 3D crustal structure in predicting North-American seismograms.
S53B-0217 1340h
How Thick is the Earth's Crust?
In order to understand continental evolution and other geological processes, it is first necessary to develop a thorough knowledge of the Earth's crustal structure. We present a recently-updated contour map of the thickness of the Earth's crust using a 10-km contour interval, with the 45-km contour also included. This contour map was created from about 8000 individual crustal data points that have been acquired during the past 65 years. The contour map honors all available seismic refraction measurements for features with a dimension greater than 2 degrees. Crustal thicknesses in Eurasia, North America, and Australia are well constrained by seismic refraction data, while new data has enhanced resolution in Antarctica, South America, Africa, and Greenland. To a first approximation, the continents and their margins are outlined by the 30-km contour. The part of the continental interior enclosed by the 40-km contour and regions with crustal thickness of 45 to 50 km are found on all well-surveyed continents. Continental crust with thickness in excess of 50 km is exceedingly rare and accounts for less than 10% of surveyed continental crust. These observations, now available on a global basis, provide important information to be used for numerous scientific studies, including ongoing refinements to global crustal models.
http://quake.wr.usgs.gov/research/structure/CrustalStructure
S53B-0218 1340h
Constraining crustal S-wave velocities in southern California by deformable layer tomography
We are building a new crustal S-wave velocity model for southern California using more than twenty years of S-wave first arrival data from a dense distribution of local earthquakes and seismologic stations, as compiled by the Southern California Earthquake Center (SCEC) data center. Together with the P-wave model, the S-wave model plays a critical role for elastic modeling and tectonic interpretation for the crust. However, tomographic construction of crustal velocity models is challenged by the variable crustal thickness. Due to high velocity and low velocity gradient of the mantle below the Moho, the Moho-grazing first-arrival raypaths at regional distances rarely extend below the Moho. The depth-velocity ambiguity makes it difficulty to simultaneously constrain the lateral velocity variation and Moho depth variation. Despite of the dense distribution of local earthquakes and stations, the S-wave velocities in the lower crust have not been well constrained previously because there are less S-wave picks than P-wave picks, particularly with respect to the Moho-grazing S-wave. We are using a deformable layer tomography (DLT) that takes advantage of the fact that the 1-D S-wave velocity is well constrained for the crust and upper mantle at some locations from previous waveform studies. We are also using the well-constrained velocities of the shallow crust from previous studies for initial model values. Unlike most of the previous tomography methods that use fixed-in-space cells or nodes, DLT directly inverts for the depths of velocity interfaces. DLT can resolve velocity-defined units that may vary significantly in thickness and pinchout to yield robust solutions in regions of sub-parallel rays and to accurately represent basin and Moho geometries. The SCEC data center compiled over 800,000 S-wave picks from local events since 1983. The raw data were sorted according to a number of quality control thresholds to maintain the quality of hypocenters and number of picks per event, resulting in a data set of over 400,000 S-wave picks from nearly 80,000 local events. Synthetic tests of the DLT algorithm using real sources and station locations suggest that the geometry of major S-wave velocity contours can be resolved in the study area. At similar data fitness level, we expect the DLT method will achieve more reliable result of the S-wave velocity contours in the lower crust in comparison with the previous S-wave models.
S53B-0219 1340h
Moho Configuration from a Single Fold Reflection Profile obtained from a Seismic Refraction Survey in the Deccan Trap Region, India
Controlled Source Seismic (CSS) studies by the National Geophysical Research Institute (NGRI) were carried out during 2001 - 03 along nine select traverses (E-W and N-S) in the Narmada and Tapti river basin, India. A total of 705 km of refraction data, with record lengths of 20 seconds, were collected using a 120 channel Radio Frequency Telemetry (RFT) system (EAGLE 88) in SEG-D format. A 40 km long line from Jhagadia to Rajpipla (W-E direction) is here interpreted on the basis of both refracted and reflected phases. Processing of the first 6-8 sec focused on the basement configuration. Conventional ray tracing analysis of refracted and wide angle reflected phases constrains the velocity structure to about 10 km depth: from surface to depth the key units include a Recent formation with velocity of 1.9 to 2.3 km/s, a Quaternary unit with a velocity 2.9 - 3.3 km/s, traps with velocity varying from 4.8 to 5.2 km/s, low velocity Mesozoic sedimentary rocks with a velocity of 4.3 km/s, all underlain by granitic basement with velocity varying between 5.8 and 6.1 km/s. To investigate the lower crust the refraction data were reprocessed as a single fold seismic reflection profile extending to 20s travel time. Both mid-crustal and Moho reflections can be clearly identified. The Moho dips from east to west, with the Moho at a depth of 34 km below Jhagadia. The Moho depths, as well as the velocity structure derived from the reflection data, correlate well with the results from the Mehmadabad-Billimora 260 km long (N-S) deep seismic sounding profile (Kaila et al., 1981). This study demonstrates how reflection techniques can be used to extract additional information from previous refraction surveys in others parts of India.
S53B-0220 1340h
Teleseismic P-wave Velocity Tomography Beneath The Arabian Peninsula
We have imaged tomographically the three-dimensional velocity structure of the upper mantle beneath the Arabian Peninsula using teleseismic P-waves. The data came from the Saudi Arabian National Digital Seismic Network (SANDSN) operated by King Abdulaziz City for Science and Technology (KACST) and three permanent stations (RAYN, EIL and MRNI). The KACST network consists of 38 stations (27 broadband and 11 short-period) spread throughout most of western Saudi Arabia. P wave travel time residuals were obtained for 131 earthquakes in the distance range from 30\deg to 90\deg, resulting in 1716 rays paths. We find a pronounced low velocity anomaly beneath the southeastern Arabian Shield and southern Red Sea that likely represents a northward continuation of the Afar hotspot. We also image smaller low velocity anomalies beneath the Dead Sea Transform, the Gulf of Aqaba, and the northeastern corner of the Arabian Shield. The origin of these low velocity anomalies is uncertain.
S53B-0221 1340h
Geometry and P- and S- Velocity Structures of the "African Anomaly"
Seismic evidence shows that the "African anomaly", a prominent low-velocity anomaly in the lower mantle beneath southern Africa, has a very-low velocity province (VLVP) in the lowermost 200-300 km of the Earth's mantle with rapidly varying geometries and a strong S-velocity reduction gradient from -2% (top) to -9% - -12% (bottom). It is now clear that the VLVP is compositionally distinct and can be best explained by partial melting driven by a compositional change produced in the early Earth's history (Wen, 2001; Wen et. al, 2001). However, the nature of the anomaly in the lower mantle remains unknown and controversial. Here, we investigate its geometry, and both P- and S- velocity perturbations along a specific great arc from the Drake Passage to the Hindu Kush region. The southwestern side of the anomaly is sampled by direct S, ScS, and SKS waves from earthquakes occurring in the South Sandwich islands and the East Pacific rise. Direct S waves from the South Sandwich islands earthquakes exhibit no travel time delay with respect to the preliminary reference Earth model (PREM) from 44$^{\circ}$~~to 59$^{\circ}$~~across the Kaapvaal array, but they are increasingly delayed from 3 s at 67.5$^{\circ}$~~to 12 s at 74$^{\circ}$~~across the Tanzania array and from 12 s at 74.5$^{\circ}$~~to 19 s at 89$^{\circ}$~~across the Kenya array. The ScS residuals have a simple trend increasing from 4.5 s at 44$^{\circ}$~~to 20.5 s at 74$^{\circ}$. The SKS waves for earthquakes occurring in the East Pacific rise show a uniform delay of about 3 s before 101$^{\circ}$, steeply increasing to a uniform delay of 7 s after 107$^{\circ}$. These observations indicate that the "African anomaly" extends about 1300 km above the core-mantle boundary (CMB) with its southwestern edge dipping towards its center and its basal layer near the CMB extends further southwest. The northeastern side of the anomaly is constrained by the direct S and ScS waves from three events in the Hindu Kush region, the SKS waves from an earthquake in the Drake Passage, and the SKS and SKKS waves from an earthquake in the Japan sea. Both S and ScS waves from earthquakes occurring in the Hindu Kush region do not show travel time delays with respect to PREM, placing bounds on the geographic extent of the anomaly on the northeastern side. The SKS residuals of the Drake Passage earthquake shows a decreasing trend from 7 s at 87$^{\circ}$~~to about 1 s at 93.5$^{\circ}$. The SKKS wave from an earthquake in the Japan sea is only delayed after 140$^{\circ}$~~whereas the SKS residuals are about 4 s between 129$^{\circ}$~~and 140$^{\circ}$. These observations suggest that, on the northeastern side, the anomaly also dips toward its center beneath southern Africa. Therefore, the "African anomaly" exhibits a "cusp-like" geometry in the lower mantle. The magnitudes of these travel time residuals can be best explained by a shear velocity structure, with average reductions of -5% for the basal layer and -2% - -3% for the portion in the lower mantle. With the "cusp-like" geometry, the P-wave velocity structure is constrained by the travel time of direct P, PcP and their differentials from the same events. P-wave data provide good sampling coverage for both the basal layer and the portion in the lower mantle. We test a series of P velocity models with uniform P to S velocity perturbation ratios from 1:1 to 1:7. We find that a ratio of 1:3 best explains the P wave data. The geometry and the P to S velocity perturbation ratio of the anomaly indicate that the "African anomaly" in the lower mantle likely, like the VLVP at its base, is compositionally distinct and has an origin of partial melting.
S53B-0222 1340h
Thermal structure beneath Tanzania from attenuation measurements using teleseismic P wave spectra
Using P wave spectral amplitude ratios from deep-focus earthquakes recorded at broadband seismic stations of the Tanzania PASSCAL network, we estimate the regional variation of sublithospheric mantle attenuation beneath the Tanzania craton and the eastern branch of the East African Rift. To constrain the thermal anomaly beneath the eastern rift, we analyze P wave attenuation beneath the Tanzania seismic network and the adjacent rift system in combination with velocity anomalies determined from seismic tomography. We conclude that the observed variation in $t^*$ can be explained by variation in $Q_p$ in the sublithospheric mantle and obtain values of $Q_{p}\sim$175 beneath the cratonic lithosphere and $Q_{p}\sim$80 beneath the rifted lithosphere. By combining the $Q_{p}$ values and a model of P wave velocity perturbations, we estimate that the temperature beneath the rifted lithosphere (100-400km depth) is 140-280K higher than ambient mantle temperatures, consistent with the observation that the 410km discontinuity in this region is depressed by 30-40km.
S53B-0223 1340h
Simultaneous Inversion of Receiver Functions and Surface Wave Dispersion For Lithospheric Structure Beneath The Middle East, North African, Central Asia, And Parts of Europe
Numerous researchers have studied lithospheric structure of Africa, Europe, the Middle East, and Asia for decades. Although there are agreements on the general lithospheric structure, interesting discrepancies exist on the specific lithospheric structure in a number of areas. Our objective is the construction of shear-wave velocity profiles for regions surrounding broadband seismic stations throughout the Middle East, central and north Africa, central Asia, and parts of Europe. This survey provides an opportunity to confirm and if needed, revise, models of the crust and upper mantle structure throughout the region. To estimate lithospheric structure, we use a joint inversion of receiver function and surface wave dispersion observations. We collected seismic data from available permanent and temporary three-component broadband seismic stations throughout the region. Two hundred and thirty-two stations, to date, have been investigated; 37 stations in central and north Africa, 72 stations in Middle East, 57 stations in Europe, and 66 stations in central Asia. We have examined receiver functions for 138 of stations in the period of 1990-2003, and have inverted a number using dispersion measurements extracted from global and regional tomographic models. We also applied the receiver function stacking procedure of Zhu and Kanamori [2000] to estimate Vp/Vs and crustal thickness. This initial analysis provides reasonable constraints on thickness and Poisson's ratio (which trade-off) for each station and helps identify stations situated in complex structures, where simple plate-layered interpretations fail. For most stations crustal thickness and Poisson's ratio vary as a function of back azimuth, indicating non-isotropic plane-layered structure. Crustal thickness for Middle East ranges from 32-56 km with a mean of 43 km. The average Vp/Vs value for the region is 1.73. Sites in Africa have a crustal thickness range of 27-55 km with a mean of 37 km and average Vp/Vs 1.76. For station in Asia we find a crustal thickness range of 28-80 km with a mean value of 53 km. The average Vp/Vs for this region is 1.75. For stations in Europe crustal thickness estimates range from 24-44 km with a mean of 33 km and an average value of 1.73 for Vp/Vs. The majority (but not all) of our crustal thickness estimates are consistent with previous geophysical/seismological work. The mean differences between our estimates and those of Crust 2.0 is about 1 km with a standard deviation of the differences is about 5 km. A comparison of the crustal thickness and Poisson's ratio estimates for the crust beneath these stations will be presented; also we will show a comparison of our results with present global crustal models.
S53B-0224 1340h
High-Resolution Tomography of the Kaapvaal Craton
We present the results of a tomographic study using the broadband waveforms recorded by the Kaapvaal Seismic Array deployed in Southern Africa between 1997 and 1999. The Kaapvaal array is composed of about 80 stations at an average spacing of ~100 km. We accurately measure the phase delays between the different stations using a waveform cross-correlation method. For each event, we compute and subtract the average of all the measurements to reduce the effect of structures outside the array. The dataset is composed of phase delays of broadband {\it{SH}} waves as well as narrowband Rayleigh waves at a number of frequencies between 0.01 Hz and 0.04 Hz. A crustal correction is applied to the phase-delay measurements before they are inverted using an LSQR algorithm. We use a normal-mode coupling method to compute the Fr{\'{e}}chet (sensitivity) kernels for the Rayleigh waves and a normal-mode-ray-theory hybrid approach to compute the kernels for the {\it{SH}} waves. This hybrid approach is adopted to ensure both the efficiency of the calculations with a 10-km grid spacing for the {\it{SH}} waves and the accuracy of the kernels in the near and intermediate fields from the seismic stations. The resulting shear-wave speed model has similar patterns to the shear-wave model for the Kaapvaal region (James et al. 2001) but covers a broader area. The wave-speed anomalies show good correlation with surface geology, in particular with the limits of the cratonic blocks. The upper mantle beneath the Kaapvaal craton is characterized by high wave speed reaching at least 300-km depth. One of the interesting features of our model is a low-speed region in the upper mantle to the east of the Kimberley Array which coincides with the suture between the Witwatersrand and the Kimberley blocks.
S53B-0225 1340h
Three-Dimensional Full-Wave Tomography on a Laptop
Recent advances in seismic tomography show that to resolve structures of sizes smaller than the first Fresnel zone width of the waves used, three-dimensional (3-D) Fr{\'{e}}chet kernels (a.k.a. the banana-doughnut kernels) must be used. Dahlen et al. (2000) proposed an efficient algorithm which made the 3-D kernels practical for global tomography (Montelli et al. 2004). However, ray-theory approximation in Dahlen et al. (2000) is only applicable to observations from far-field high-frequency body waves. We propose an alternative efficient approach to computing the 3-D kernels based on the normal-mode theory which provides accurate, full-wave solution to the wave equation. This aprroach comes from the realization that the heterogeneity-induced waveform perturbations only depend on the strain Green tensor (SGT) which is a function of the earth model only. Thus, a database of SGTs can be established for a reference Earth model such as AK135, which eliminates the need for repetitive evaluations of the SGTs in subsequent 3-D kernel calculations. The SGT database is composed of all the independent elements of the third-order SGT which requires a certain amount of CPU time and disk space depending on the size of the problem. For example, for a grid of 30 km in space and 2 sec in time, a complete SGT database for global tomography requires a few weeks of single processor CPU time and ~80 GBytes of disk space. Preliminary tests show that this modest amount of overhead work leads to two orders of magnitude increase in efficiency for 3-D kernel calculations, making it practical to conduct almost all global and regional tomography studies without making any high-frequency approximaiton. This approach is completely general and flexible. It can be used to compute 3-D kernels of any types of seismic data (traveltime, amplitude, splitting), for any phases on the seismogram, and for any model parameters. It can also be used for inversions of earthquake's centroid and even higher moments. Moreover, terabyte SGT databases can be established at large facilities for public use in global high-resolution and regional fine-resolution studies while individual researchers can create small, portable and customized databases for fast inversions using single workstations or even laptops.
S53B-0226 1340h
Evaluation of Philippine Sea Plate Reconstructions Based on Seismic Tomography of Subducted Slabs in the Western Pacific.
Mantle seismic tomography of the Western Pacific region (Yoshio Fukao and Sri Widiyantoro, personal communication) shows the distribution of subducted oceanic slabs. Plate reconstruction models for the Philippine Sea Plate are evaluated based on the correlation between past locations of subduction zones in the reconstructions and the distribution of subducted slabs in the mantle. Massive high velocity anomalies at transition zone depths and deeper beneath the Japan Sea, adjacent areas and the northern Philippine Sea are best explained by past northward subduction of the Philippine Sea Plate and westward subduction of the Pacific Plate. Greater than 700 km deep high velocity anomalies beneath the Philippine arc/Eastern Indonesia region are best associated with northward Indian Ocean Plate subduction. Other high velocity anomalies correlate well with present day Philippine Sea Plate boundaries. Deep high velocity anomalies are lacking beneath the central and southern Philippine Sea Plate. Mantle tomography supports neither southward subduction along a past northern boundary or large clockwise rotation of the Philippine Sea Plate.
S53B-0227 1340h
Analysis of Core-Mantle Boundary Structure Using P and S Diffracted Waves
Existing tomographic studies of the base of the mantle have been limited by poor coverage in the Southern Hemisphere. We have investigated the possibility of filling in the gaps by using diffracted P and S waves, which sample the base of the mantle. Using a new picking technique that allows us to process a large amount of data, we have created a new data set of diffracted P and S arrival times from long period seismograms. We used events from 1986 to 2001, obtaining approximately 15,000 Pdiff and 6,000 Sdiff measurements, with a distance range extending to 160 degrees. Coverage in the Southern Hemisphere is greatly improved, with Pdiff and Sdiff respectively sampling 93% and 81% of the core-mantle boundary, as parameterised by 4 degree equal area blocks. Resolution is also greatly improved in the Northern Hemisphere. Models dervied from our data are similar to those of previous studies, with improved definition of fast velocity anomalies in the Southern Hemisphere.
S53B-0228 1340h
Small Scale Lower-Mantle Structure as Revealed by SPdKS Back-Azimuth and Slowness Deviations
Back azimuth and slowness deviations of the seismic phase SPdKS are used to investigate small to large-scale velocity heterogeneity in the lower mantle and at the core-mantle boundary (CMB). SPdKS is the result of an SKS wave with an angle of incidence to the CMB equal to the critical angle for ScP, and initiates near 110 degrees. SPdKS has a small segment of P-wave diffraction along the mantle-side of the CMB, at the SKS core entry and exit location. SKS and SPdKS have very similar paths through the mantle. Global data were collected from the IRIS database for deep events at broadband networks with apertures of a few hundreds of kilometers. The array approach allows enhancement of resolution of key wavefield information. For example, the horizontal slowness and back azimuth of SKS and SPdKS can be accurately measured. Recent studies have presented models that should result in deviations from the theoretical great-circle plane containing the stations and earthquake. Here, we present horizontal slowness and back azimuth anomalies of SKS and SPdKS, which offer valuable insight into deep mantle structure. Observed back azimuth anomalies can be significant; preliminary analyses suggest the SPdKS back azimuth can be off by as much as 5 deg. We will present results for several key regions, including anomalies associated with superplume provinces.