S21C-1833
Using SS precursors to investigate the lithosphere and the MTZ discontinuities beneath eastern Asia and the western Pacific
We attempt to use the large amount of seismic data recorded by the GRSN and the GRF stations in Germany to study the lithosphere and mantle transition zone in the continent-ocean transition with SS precursors. SS bounce points from events in the seismically very active region of the northwest Pacific sample a corridor from the Aleutians, Kamchatka and the Japan subduction zone through the North China craton to the Tibetan plateau. The corridor passes different tectonic units such as subduction zones, an old continental shield, a fold belt and a high plateau. We aim to get information about the lithosphere and the mantle transition beneath different geologic units along the corridor and infer geodynamic processes at depth. We aim to investigate the transition zone thickness beneath continents and oceans as it was suggested they might extend into the mantle transition zone. However, this correlation is not evident in other studies.High resolution images of SS precursors may also reveal the interaction of the subducted oceanic lithosphere with the mantle transition and serve to answer the often debated question about the scale and extent of the slab stagnation within the mantle transition zone beneath the NW Pacific subduction zone. We attempt to use the short period content of SS waveform data and experiment with methods like CMP stack and migration techniques to study the shallower structures such as the lithosphere-asthenosphere boundary with SS precursors.
S21C-1834
Plate boundary structure at the Junction of Japan and Kuril Trenches
The Japan and Kuril trenches are plate convergent zone where the Pacific Plate is subducting below the Japanese islands. Many earthquakes occur around the Japan and Kuril Trenches associated with the Pacific Plate subduction. It is not clear a relationship seismicity and palte boundary geometry. Therefore, we undertook an earthquake observation@to reveal the seismic activities of the northern Japan Trench fore- arc region region using long-term type ocean bottom seismographs (LTOBSs). During the earthquake observation, we conducted seismic refraction experiments using an airgun array and the LTOBSs in the study area. In order to investigate the plate boundary structure at the Junction of Japan and Kuril trenches and understand a relation seismic structure and seismic activity. To obtain the seismic structure is also important for analyses of the data recorded by the LTOBSs. In this study, we present two-dimensional seismic structures obtained by the seismic refraction experiments. The wide-angle seismic surveys were conducted in the northern Japan Trench fore-arc region in 2003 and 2005. 53 ocean bottom seismometers were deployed on 9 seismic lines. The airgun array had 4500 cubic-inches of a total capacity and was shot with a 200m shooting interval. Seismic structures beneath the profiles were estimated by forward modelind 2- D ray tracing (Zelt and Smith 1992). Our results show that the subducting Pacific plate increases its dip from about 5 degrees to 15 degrees approximately. This bending point in the oceanic plate correspond to the eastern edge of the trenchward rupture zone of 1968 Tokachi-oki earthquake(Mw=8.2).
S21C-1835
P- and S-Receiver Function Analysis of Borehole Broadband Ocean Bottom Seismic Data
We use P- (Prf) and S-receiver functions (Srf) techniques to decipher the lithospheric thickness of oceanic plates using the borehole broadband seismological data set. These stations are deployed under Japanese Ocean Hemisphere network Project (OHP): WP1 and WP2 in the Philippine sea and the northwestern Pacific ocean, respectively. Both the stations are located in a borehole of depth ~500m with water column of thickness ~5km that produces P-waves water reverberations. These water reverberations travel almost vertically downward and contaminate the vertical component of seismograms. To minimize the water multiples in case of Prf, we prefer the Radial receiver functions and in case of Srf, we used P-Sv-Sh components. Reflectivity synthetics show that the effect of water reverberations have no effect on Sp phases, as multiples arrive latter than the primary S-phase while the radial components are least affected in case of Prf. We found a consistent negative phase interpreted as Lithosphere-Asthenosphere Boundary (LAB) apart from a positive phase corresponding to Moho in both the receiver functions. For WP1 the LAB is at 71±1.8 km (for 49Ma lithospheric age, Western Philippine Sea plate) and 50±1.4 km (for 20Ma, Eastern Philippine Sea plate); for WP2 in the Pacific ocean the lithospheric thickness is 76±4.2 km (for 129Ma). For the observed Prf, the synthetic studies show that the observed amplitude of negative phases are not sufficient enough to match only with water demanding for the presence of a lvl. The thicknesses of the oceanic plates estimated above is consistent with the thermally controlled origin for the oceanic LAB and since it is a sharp boundary (~3s), indicates chemical or fabric origin (Kawakatsu et al., 2008, this meeting).
S21C-1836
Crustal Stress and Mantle Strain in the Tibet-Tarim border region.
The Tarim basin lies directly north of the Tibetan plateau, and is distinct from it by the apparently low degree of internal deformation. The border between Tibet and Tarim runs in an east-west direction, and is seismically active, with some researchers favoring southward subduction of the Tarim lithosphere. Relative motion between the lithospheres of Tibet and Tarim will likely result in coherent mantle deformation, with diagnostic signature of anisotropic seismic wave speed. We probe the nature of mantle texture along the border between Tarim and Tibet using data collected a decade ago by a network of portable broadband seismographs. We analyze core-refracted shear phases (SKS, PKS) for presence of birefringence that would be indicative of seismic anisotropy along their path. Preliminary results suggest fast shear-wave polarization with the NW-SE alignment, consistent with mantle flow along the Tibet-Tarim border. There is a certain degree of variability in measurements, which may reflect either lateral change in rock texture, or else its vertical stratification. We use multi-event inversions for models of various complexity (layered, tilted or both) to address possible sources of measurement scatter. Serendipitously, a recent earthquake sequence (Mar 20, 2008) due south of the location of our network offered evidence for the state of stress in the brittle part of the crust in this region. The normal faulting mechanism of the main shock (Mw=7.1, Depth=12km, from the Harvard CMT solution) shows a NW-SE extension. Focal mechanisms of the aftershocks also present a majority of E-W/NW-SE extension directions. The coincidence of the suggested regional extension in the crust and the deformation direction in the mantle is intriguing. However, some other regional focal mechanisms show low-angle nodal planes, and P-axes in the north-south direction, more consistent with the proposed collision between Tibetan and Tarim lithospheres. We compare the crustal stress directions inferred from focal mechanisms with our assessment of the mantle deformation directions, and explore the state of crust-mantle coupling at the border between Tibet and Tarim.
S21C-1837
Use of Surface Wave Reflections to Study Offshore Structures
It is understood that a lateral discontinuity of crustal structure can act as a reflector of surface waves. The reflection coefficient for surface waves at a realistic discontinuity is usually small (<0.1) that the reflected waves are weak, thus difficult to observe among surface wave coda. However, with the high spatial resolution and sensitivity of the Hi-net seismic network, surface waves reflections are observed by tiltmeters and seismometers at the Chugoku and Kinki region of Japan during various major earthquakes, including the Miyagi-ken-oki (2003, Mw7.0), Tokachi-oki (2003, Mw8.3) and Ibaraki-ken-oki (2008, Mw6.8) earthquakes. The reflected wave train has a duration of about 100 seconds and a predominant period of about 30 second. This long period implied that the reflector is likely a velocity anomaly located in the lower crsut. Beamforming analyses demonstrate that a northwestward extension of the Kyushu-Palau ridge southeast to Kyushu is a possible candidate of such a velocity anomaly that acts as a reflector. As the ridge is subducting northwest together with the Philippine Sea plate underneath the Eurasia plate, the reflector is possibly a subducted ridge or sea mount. With the location of reflectors assumed, particle motions show that the reflections are mainly Love waves. We intend to conduct numerical simulations to further quantify the location and nature of the reflectors. This study shows that surface wave reflection is a potential tool to identify subducted structures by land-based observations, supplementary to the orthodox but expensive body wave method involving the use of ocean-bottom seismometers or marine geophysical surveys. This new prospect may benefit researches concerning the relation between seismicity and bathymetry, or the so-called buoyancy hypothesis.
S21C-1838
Crustal Structure Of Taiwan Area From Radial Receiver Functions
We implement H-£e stacking (Zhu and Kanamori 2000) and three-layer media analysis (Tang et al. 2008) to investigate the crustal structure in Taiwan area. In this study, we use data recorded by the stations of Broadband Array in Taiwan for Seismology (BATS) and Central Weather Bureau (CWB) of Taiwan. Radial receiver functions (RRFs) computed from teleseismic recordings at 19 permanent broad-band seismic stations (including 6 surrounding-island stations). The Moho depth in Taiwan area (including surrounding- island area) varies from 6 to 40 km and Vp/Vs ratios of crustal average varies from 1.64 to 1.87. The deepest Moho depth in Taiwan locates at the Central Mountain Range (CMR) and consistent with the result obtained from gravity survey. In this study, we discovered that there are two discontinuities exist in the crust of some place in Taiwan area. The discontinuity 1 and the discontinuity 2 exist more clearly in the CMR than in other area. The depth of the discontinuity 1 varies from 5 to 8 km and while the depth of the discontinuity 2 varies from 11 to 22 km in our estimation. From our results, we imply that the discontinuity 1 and 2 are basement and the Conrad discontinuity respectively.
S21C-1839
Two Remarkable Depth Phases Observed in Southwest Japan for Intraslab Earthquakes Within the Philippine Sea Plate
Many distinct later phases are observed by local seismological networks for intraslab earthquakes that occur within the Philippine Sea plate, which is being subducted beneath southwest Japan. In this study, we analyze two remarkable later phases, identified between the P and S wave arrivals, which we refer as X1 and X2. We investigate NIED's Hi-net (High Sensitivity Seismograph Network, Japan) seismogram recordings of intraslab events which occurred beneath southwest Japan from October 2000 to May 2005 (focal depth from 30 to 70km, magnitudes between 4 and 7), and examine the features and origin of the two identified phases. The main characteristics of these phases are as follows: 1) they arrive about 5-15 seconds after the initial P-wave arrival and are observed at epicentral distances of about 150 km or larger, in a wide area in southwest Japan; 2) their apparent velocities are approximately 6.5 km/s, roughly coinciding with the P-wave velocity in the southwest Japan island-arc crust; 3) they can be clearly detected on both vertical and radial component seismograms and 4) their travel times depend on the focal depth of earthquakes. Based on these features, we interpret these X phases are depth phases arriving at stations as compressional waves. To clarify the origin of X1 and X2 phases, we calculated theoretical travel-times assuming a 1-D velocity model. A possible interpretation is that X1 is a pPxP (x indicates reflector) phase, i.e., a P-wave reflected at the free surface and at the island-arc Moho discontinuity or at a reflector within the island-arc crust, and X2 is a sPxP, i.e., a P- wave converted from S at the free surface and reflected at the island-arc Moho discontinuity or at a reflector within the island-arc crust. The two depth phases might be reflected either from Moho or from another reflector within the crust depending on the epicentral distance.
S21C-1840
Philippine Sea Plate and Serpentinized Mantle Wedge beneath Kii Peninsula
The deep low frequency events (DLFEs) are distributed widely from western Shikoku to central Tokai through southern Kii Peninsula (Obara, 2002). When we explain this phenomenon by the behavior of ewaterf dehydrated from the Philippine Sea Plate (PSP), it is important to know structure of the PSP and the mantle wedge beneath the Kii Peninsula. In order to predict accurate strong motions, it is very important to know large scale structure beneath the Kii Peninsula, through which seismic waves from the Tonankai and Nankai Earthquakes travel to Osaka region. The purpose of this study is to image the S wave velocity discontinuities beneath the Kii Peninsula by using receiver function (RF) analyses with waveforms from teleseismic events and to estimate the shapes of the PSP, the Moho and other discontinuities. RFs are calculated by deconvolving the vertical component of the P coda from teleseismic events from the corresponding radial component in order to remove source time functions and instrument impulse responses. We can convert the relative travel time between the Ps converted wave and the direct P wave to the depth of the S wave velocity discontinuity assuming a velocity structure. The continental Moho is located at about 30km beneath the northern half of the profile line and becomes shallower beneath the Kii Peninsula. The upper boundary and the slab Moho of the PSP are imaged descent toward the NW direction from the depths of 20 - 25 km at the SE edge of the image. The oceanic crust between them shows obvious low velocity anomalies to the depths where the DLFEs occur, and less low velocity anomalies beyond the depths. This suggests that the dehydration is active in the source region of the DLFEs. The discharged water serpentinizes the mantle wedge, which shows widespread low velocity anomalies. However, there are some problems in the imaging technique because we assume horizontal velocity structure. So, the improvement of the imaging technique is necessary for treating the dipping layers.
S21C-1841
Crustal Thickness and Seismic Anisotropy of the Rwenzori Region (Albertine Rift) From Receiver Functions and Shear-Wave Splitting
The Rwenzori mountain range is situated within the western branch of the East-African rift system. It is composed of metamorphic rocks and reaches altitudes of more than 5000 meters. Within the framework of the multidisciplinary RiftLink research group we have carried out a passive-source seismological study in western Uganda close to the Congo border. The project aims to constrain the development and uplift of the Rwenzori range and its relation to the formation of the rift zone. The temporary network consisted of 8 broad- band and 15 short-period seismic stations. The initial deployment started during April of 2006. The network was in operation until September 2007. Receiver functions computed for stations on the eastern rift shoulder reveal a rather simple crust. However, the Rwenzori range is characterized by a complex inter-crustal structure, which causes interference and masking of the Moho signal at several stations. Different techniques based on traveltime and waveform information are applied to derive Moho depths beneath the network. The mapping of crustal thickness provides evidence for the absence of a deep crustal root underneath the Rwenzori mountains. The receiver functions exhibit a pronounced negative phase at stations in the southeastern part of the network. This anomaly is interpreted as resulting from a low-velocity layer at roughly 15 km depth. The strong decrease of the S-wave velocity within this layer may be indicative for the presence of partial melt. The measured splitting parameters from teleseismic SKS phases exhibit fast-polarization directions that are parallel to the rift with delay times of about 1.2 seconds. The rift-parallel polarizations indicate that rifting in this region is probably assisted by magmatic intrusions. Indications for spatial changes in anisotropy come from the frequency dependence of the splitting parameters at some stations of the network.
S21C-1842
Crustal Structure in Cameroon From Receiver Function Analysis
Crustal structure under the Cameroon broadband seismic experiment has been investigated using P-wave receiver functions in order to determine the extent to which the extensional tectonism at the Cretaceous Benue Trough and the Tertiary-Recent volcanism along the Cameroon Volcanic Line (CVL) have modified the thickness and composition of the crust beneath Cameroon. The seismic experiment consisted of 8 stations deployed for 2 years (2005-2007) and 24 stations deployed for one year (2006-2007). Hκ stacking of receiver functions has been utilized to obtain crustal thicknesses and Vp/Vs ratios under the stations at both high (f < 1.2 Hz) and low (f < 0.5 Hz) frequencies. Data from over 300 earthquakes with magnitudes above 5.5 and in the 30 to 90 degree distance range were processed, from which 5 to 15 high quality receiver functions could be selected per station. Preliminary results indicate that crustal structure is quite variable beneath the Cameroon Volcanic line (CVL) and across much of Cameroon. Normal Precambrian crustal thicknesses and Vp/Vs ratios are found for the northern edge of the Congo Craton in southern Cameroon and the Proterozoic mobile belts between the craton and the CVL. Variable crustal thickness is found along the strike of the CVL, with thin crust near the coast and thicker crust inland. The crust beneath the Adamaoua plateau in the north appears to be normal, unaffected by the presence of the volcanic line. Finally, the crust beneath the Benue Trough appears to have been thinned compared to the CVL and Adamaoua plateau.
S21C-1843
Shear Velocity Structure of the Cameroon Volcanic Line Region from Rayleigh Wave Phase Velocities
The Cameroon Volcanic Line (CVL) is an 1800 km long footprint of Cenozoic volcanism extending from the Gulf of Guinea into Central Africa that lacks the age progression anticipated for hotspot-related volcanism. We investigate upper mantle shear velocity structure beneath the CVL in an effort to gain insights about the current state of the uppermost mantle in the region and to evaluate possible models of the origin of the CVL. To estimate the shear velocity structure in the CVL region, we determined the phase velocities of Rayleigh waves from teleseimic earthquakes recorded by the Cameroon Seismic Experiment, a two year PASSCAL broadband seismometer deployment of 32 stations spaced roughly 50-200km apart, using the Forsyth and Lee (2005) two-plane wave technique. Data from more than 300 shallow earthquakes with magnitudes greater than 4.5 Ms and distances from 3000-12000 km were processed at a series of frequency bands between 0.005 and 0.05 Hz, providing resolution of shear velocities to depths of about 250km. Preliminary results from phase velocity maps show a clear continental signature and suggest significant lateral shear velocity structure in the CVL region. To the south of the Cameroon Line phase velocities are fast, a feature associated with cold mantle beneath the proterozoic mobile belts and the Congo craton. Directly under the CVL we observe a linear low-velocity feature over the period range of 20-140s which indicates that the source of the CVL anomaly extends from the uppermost mantle to depths of at least 150 km. Results at the longest periods are not conclusive due to lower resolution, but we attempt to bound the possible depth extent of the CVL anomaly and thus determine if the CVL is related to an entirely upper mantle process or the result of a deeper source.
S21C-1844
Modeling LPO and 3D shear-wave splitting at mid-ocean ridges
We present fluid-dynamical models of a stationary mid-oceanic spreading center to investigate the
development of upper-mantle anisotropy and its effect on 3D shear-wave propagation. At mid-ocean ridges
upwelling mantle material is generally constrained to vertical flow at the ridge axis and horizontal flow at great
distances from the spreading center. This is commonly described by analytical corner-flow solutions. In the
modeling presented here, we include effects of pressure-, temperature-, and stress-dependent viscosity at
fast and slow-spreading ridges which leads to modified patterns of strain-induced lattice-preferred orientation
(LPO). We use a combination of D-Rex (Kaminski et al., 2004) and FDCON (e.g., Schmeling and Marquart,
1991) for the calculation of complete elastic tensors in the mantle flow deformation field. The orientation of
the mineral aggregates (olivine and enstatite) causes seismic anisotropy and affects the propagation of
shear waves. In anisotropic media shear waves are split into two components with perpendicular polarizations
and different velocities. We simulate shear-wave splitting of teleseismic SKS waves in 3-dimensional models
of mid-ocean ridges and analyse variations of the splitting parameters as functions of backazimuth, incidence
angle, and distance from the ridge axis. For vertically incoming waves the polarization direction of the fast
shear wave is always parallel to the spreading direction. For realistic non-vertical wave incidence, fast-
polarizations vary with backazimuth and differ from the spreading direction. Those deviations reach values of
up to 13°. Delay times between fast and slow shear waves are highly variable near the ridge and
reach values close to 2.5 s at 350 km distance from the ridge axis. We find that the influence of local
anisotropic structures near the ridge axis, which are different for slow and fast-spreading ridges, is negligible
in view of uncertainties related to real splitting measurements. Calculated splitting parameters for a model of
the fast-spreading East Pacific Rise agree well with previously published splitting measurements.
Kaminski, E., N. M. Ribe, and J. T. Browaeys, 2004, D-Rex, a program for calculation of seismic anisotropy
due to crystal lattice preferred orientation in the convective upper mantle. Geophys. J. Int., 158, 744-752.
Schmeling, H., and G. Marquart, 1991, The influence of second scale convection on the thickness of the
continental lithosphere and crust. Tectonophysics, 189, 281-306.
S21C-1845
Numerical modeling of surface wave propagation effects along mid-ocean ridges: implications for ultra-fast and ultra-slow spreading ridges
Using simple geodynamic models converted to seismic properties, we calculate synthetic wave fields of Love and Rayleigh waves propagating along mid-ocean ridges and investigate various contributions of heterogeneous ridge structure to the wave propagation. We examine the effects of water depth, lithospheric thickness, anisotropy, and melt on the wave fields. For example: while most surface wave studies neglect lateral refraction of the wave field and assume energy travels along great circle arcs, in the presence of lateral velocity variations packets of surface wave energy propagate along paths that deviate from the great circle path, resulting in focusing and defocusing of energy. Beneath mid-ocean ridges, low seismic velocities create a waveguide that parallels the ridge; this waveguide can strongly refract and trap short-period surface wave energy (<30 s period). Due to the sensitivity of short-period waves to the waveguide, the waveguide effect can be exploited to constrain upper mantle structure using real data. To illustrate the potential of short-period waves to probe fine-scale structure of the upper mantle, we compare synthetics with data collected along the Southern East Pacific Rise (SEPR) and show that the SEPR data are not consistent with a simplified anisotropy model of vertical alignment of olivine a-axes in the ridge upwelling zone and horizontal alignment away from upwelling zone. We also examine the effect of ridge-axis lithospheric thickness on wave propagation along an ultra-slow spreading ridge and speculate as to the nature of mantle upwelling in this environment.
S21C-1846
Intraplate seismicity and mantle hydration at the Nicaraguan trench outer rise
We examine micro-earthquake records from a dense temporary array of ocean bottom seismometers and hydrophones that has been installed for a two-month period at the trench outer rise offshore Nicaragua. Approximately 1.5 locatable earthquakes per day within the array (110x120 km) show the high seismic activity in this region. Seismicity is restricted to the upper ~15 km of the mantle and hence where temperatures reach 350-400° C, which is smaller than values observed for large mantle intraplate events (650° C). Determination of moment tensor solutions suggest a change of the stress region from tensional in the upper layers of the oceanic plate to compressional beneath. The neutral plane between both regimes is located at ~6-9 km beneath Moho and thus very shallow. Fluids, which are thought to travel through the tensional fault system into the upper mantle, may not be able to penetrate any deeper. The earthquake catalogue, which seems to be complete for magnitudes above Mw=1.6-1.8, suggests a strong change of the lithospheric rheology when approaching the trench. b-factors, i.e. the ratio between small and large earthquakes increase significantly in the closest 20 km to the trench axis, implying that the crust and upper mantle is massively weakened and hence ruptures more frequently but under less release of stress. We explain this with a partly serpentinized upper mantle.
S21C-1847
A search for Sp converted phases in MASE seismic data from the steeply dipping slab beneath central Mexico
The Mesoamerican Seismic Experiment (MASE) consisted of a set of one hundred broadband seismic stations deployed at a spacing of 5 km along a line in central Mexico, perpendicular to the trench during 2005-2007. The objective of this experiment was to obtain high resolution images of the lithosphere beneath Mexico and to better understand the subduction of the Cocos plate beneath the North American plate. The lack of instrumentation and the fact that the seismogenic zone stops at about 60 km depth, has hindered previous attempts to define the position of the interface between the slab and the surrounding mantle. The close separation of stations of the MASE network has allowed images to be constructed of scattered waves generated at the top of the slab using receiver function analyses (Perez-Campos et al., 2008) but coherent scattered arrivals from the steeply dipping portion have been difficult to recognize. The receiver functions indicate the slab undergoes flat slab subduction and underplates the crust from Acapulco to just south of Mexico City. Tomographic images of the slab (Husker 2008) indicate it then becomes steeply dipping, as well as aseismic. We have searched the MASE data for Sp converted phases coming from the upper part of the slab in its dipping section with the ultimate objective to check if eclogitization of the oceanic crust and serpentinization of the overlaying mantle, as has been inferred in Japan, can be used to trace the dehydration processes that lead to the volcanism.
S21C-1848
Broadband Seismic Investigations of the Upper Mantle Beneath the Vienna and Pannonian Basins
The Carpathian Basins Project (CBP) aims to understand the origin of the Miocene-age extensional basins contained within the compressional arc of the Alpine-Carpathian system. To test competing models for the recent geological evolution of the Carpathian-Pannonian lithosphere and upper mantle, we present a new determination of P-wave velocity structure to depths of 700 km beneath this region. This model is based on inversion of seismic travel-time residuals from 97 broadband seismic stations. We include CBP data from a 15-month deployment of a high resolution network of 46 stations deployed NW-SE across the Vienna and western Pannonian basins through Austria, Hungary and Serbia, together with 10 broadband stations spread across the Pannonian basin and a further 41 permanent broadband stations. We use P-wave arrival times from approximately 341 teleseismic events. The 3-D velocity variation obtained by tomographic inversion of the P-wave travel-time residuals shows an approximately linear belt of fast material of width about 100 km, orientated WNW-ESE beneath the western Pannonian Basin at sub-lithospheric depths. This feature is apparently continuous with structure beneath the Eastern Alps, but becomes more diffuse into the transition zone. Our initial interpretation of these fast velocities is in terms of mantle downwelling related to the early collision of Adria and Europe. We use receiver functions to assess crustal structure variations. We also determine SKS anisotropy; regionally SKS varies systematically in direction, with a delay time of about 1.0s. E-W fast directions above the fast tomographic anomaly change to NW-SE across the Great Hungarian Plane and the Vienna Basin.
S21C-1849
2-D Receiver Function Image of the Crust and Upper Mantle Beneath Western Turkey
Turkey is located in a seismically active region. Seismicity mainly governed by the interaction of African, Arabian and Eurasian plates and in result of these motions Turkey has main fault zones and main tectonic active areas that comprise the subduction, continental collision or extensional regime. Western Turkey represents the extensional regime and subduction where the African plate subducting beneath Anatolian plate. Destructive earthquakes generally occur in the upper crust and determination of the discontinuities and velocity structure has an important role for understanding of seismicity models. In the study region, we do not have enough study that reflects the crustal structure of the region. We determined the velocity discontinuities beneath the broadband seismic stations which were located in western Turkey using receiver function (RF) analysis of the P-wave coda portions. For this purpose, we selected 33 stations and 50 teleseismic earthquakes that were occurred between 2007 and 2008. We computed the radial receiver functions using extended time multi taper method using Gaussian low-pass filter 0.3 Hz for each station and we selected two profiles: the one is running from the northwest of Bozcaada to Konya city and the other is running from the northeast of Izmir to the northeast corner of Rhodes island. We produced depth-migrated RFs using one dimensional velocity models that based on previous studies. We find that radial receiver function waveforms indicate strong lateral variations in the crust and uppermost mantle. Receiver function 2- D images indicate that Conrad discontinuity settled about 15 km beneath some stations and Moho discontinuity should be lie between 30 and 40 km for the first profile and between 20 and 30 km for the second profile. Crustal thickness increase from west to east related with the tectonic process that Arabian plate collides with the Anatolian plate at the eastern part of Turkey and it pushes the Anatolian plate to the west along the North Anatolian and East Anatolian Fault Zones. Difference between Moho depths along these two profiles indicates that the region is exposed to different tectonic forces that second profile is affected by subduction of African plate under Anatolian plate and active tectonic of Aegean Sea and it seems thinner than first profile.
S21C-1850
SOME PECULIARITIES OF THE GEOLOGICAL STRUCTURE OF STRONG EARTHQUAKE SOURCE ZONES OF ARMENIA AND THEIR SEISMISITY
This article describes the geological structure of some of the source-zones of strong earthquakes which have often occurred on the territory of the Republic of Armenia. On the basis of the analysis of the reached results and their comparison with more than 30 earthquake zones of the Prehistorical, historical and recent periods, which have occurred in the adjacent regions. We have presented the geological structural model of the source zones of strong earthquakes of the Armenian Upland. Armenia Upland is located in the collision zone of the Arabian and Eurasian plates and encompasses a part of the territory of Southern Caucasus, Eastern Anatolia and North-West Iran. This area is characterized by high seismicity. Numerous sources of strong (M>6.0) Prehistorical and historical earthquakes are exerted. About 20 strong earthquakes occurred only in 20th century here. The present paper is dedicated to the geological-structural peculiarities of the sources zone is a part physical-geological model of strong earthquakes sources of Armenian Upland . The Armenian Upland has composite mosaic-bloc frame which has been formed beginning form the Pre- Cambrian period up to the Holocene included. The sources of strong earthquakes are dated for different geological-structural conditions. At the same time definite regularities of the geological structure and tectonic position are observed. Assemblage of these regularities will allow us to create the source zone geological "image." This can be used for seismic hazard assessment and earthquake prediction in this region. In the contemporary structure the juncture zones are represented by sub parallel long live, modern active faults included. Their intensely folded, scaly structure is clearly pronounced. An expansion of ophiolits, olistostrom, as well as a linear (according to a general structure) stretch of intrusive bodies is typical for this zones. Sources of the other earthquakes represent zones of deep faults. These zones are either buried (unseen) or pronounced on the day surface only by individual fault fragments of that zone. While studying the structure of some source zones of strong earthquakes we can notice that went active faults are situated in the place where tectonic mélange and olistostrome level of Mesozoic and Cenozoic age spread. In this work described young ruptures, probably seismogenic natures. Connection of modern geodynamics with geodynamic active structures of the geological past is considered.
S21C-1851
Coda Q estimation in the Hellenic subduction zone
S-coda waves from 248 local earthquakes that occurred near the Hellenic subduction zone and recorded by the EGELADOS network are used to estimate the seismic attenuation. The single backscattering model has been applied for the estimation of coda Q values. Using the moving window Fourier analysis and squared envelope analysis to measure the power spectrum of the seismograms, we calculate the coda Q values. Coda Q estimations were made for five center frequency bands centered at 1 Hz, 2 Hz, 4 Hz, 8 Hz and 16 Hz. Coda Q values obtained are dependent on the corresponding frequencies. We separate the scattering Q values and the intrinsic Q values from the total coda Q values using the estimated lapse time and quality factor of direct S wave (Wennerberg, 1993). We consider the focal depth of the events and sample the three dimensional coda volume based on the Aki & Chouet's model (1975) and correlate to the heterogeneity of the crustal structure with estimated scattering coefficients. We separate the region into three regions based on the Hellenic subduction zone and a comparative investigations are made on the characteristics of the attenuation of each region. The attenuation of the high seismicity area around the subduction zone is higher and more frequency dependent. This result shows the agreement with studies in other seismic regions.
S21C-1852
The Seismic component of the IBERARRAY: Placing constraints on the Lithosphere and Mantle.
TOPOIBERIA, is a multidisciplinary large scale research project which aims to study the links between the deep and superficial processes within the Iberian Peninsula.One of its main experimental components is the deployment of the IBERARRAY seismic network. This is a dense array (60x60 km) of new generation dataloggers equipped with broad-band seismometers which will cover Iberia and North Morocco in three successive deployments, each lasting for about 18 months. The first leg, deployed since late 2007, covers the southern part of Iberia (35 stations) and northern Morocco (20 stations). Two data centers have been established one at the CSIC-Institute of Earth Sciences (CSIC-Barcelona) and a second at the Geologic and Mining Insititute (IGME-Madrid) the data follows a standard-conventional flow from recovery to archival. The field teams collect the recorded hard disk on the field and send data and metadata to a processing center, where raw data is collected and stored and a quality control checking is performed. This include a systematic inspection of the experimental parameters (batteries charge, thermal insulation, time adjustments, geophone leveling etc), the visual verification of the seismic waveforms and the analysis, using power density spectra (PSD), of the noise level of each station. All this information is disseminated between the research teams involved in the project using a dedicated website and the continuous seismic data is made accessible through FTP and CWQ servers. Some of the nodes of the theoretical network are covered by permanent stations of the national broad-band network (IGN) or other networks operating in the region (IAG-UGR, ROA). Data from those stations will also be integrated to the Iberarray database. This Iberarray network will provide a large database of both waveform and catalogued events, with an unprecedented resolution. Earthquake data at local, regional and teleseismic scales will be analyzed using different methodologies. The first result would be an increase in the accuracy of the location of regional seismicity and the termination of focal mechanisms. A special emphasis will be attributed to seismic tomographic techniques using travel times and waveforms of P and S arrivals at different scales as well as surface waves, using dispersion measurements as well as studies dealing with background/environmental noise. In addition, receiver function analysis for seismic imaging of deep lithospheric features and splitting analysis of shear-wave arrivals will also be developed.
S21C-1853
Seismic Anisotropy Beneath the Southern Puna Plateau
The central Andean plateau is a prime region to study mantle flow above an active plateau margin, where it has been suggested that there is a link between plateau uplift and removal of the lower crust and lithospheric mantle. The southern Puna plateau (25 S to 28 S) is characterized by a number of anomalous features with respect to the rest of the Puna-Altiplano plateau including a distinctive spatial and geochemical pattern of mafic lavas and giant ignimbrites, a high topography with a large deficit in crustal shortening, an underlying slab with a gap in teleseismic intermediate depth seismicity, and a transitional dip between a steeper segment to the north and a flat-slab to the south. To investigate mantle deformation patterns across this region a total of 43 US and 30 German broadband seismic stations were deployed across the southern Puna plateau. The region of study has the advantage of deep seismicity and intermediate depth seismicity at the edge of the array which will help to constrain the depth of an anisotropic layer(s) responsible for any shear-wave splitting. Using observations of both teleseismic and local shear-wave splitting, the depth dependence of azimuthal anisotropy beneath the Puna plateau can be constrained. The preliminary results of shear-wave splitting measurements from PKS, SKS, and SKKS phases show a fairly complex pattern of shear wave splitting throughout the southern Andean plateau. In general we observe trench parallel subparallel splitting extending from the back arc to the foreland fold and thrust belt. We also observe several small regions of apparent no splitting which might suggest the presence of vertical asthenospheric flow. There is also a region of complex splitting where there is a rotation of the fast directions that are consistent with a radial flow pattern.
S21C-1854
Subduction Geometry at the Southeastern Caribbean Plate Boundary Inferred from BOLIVAR Receiver-function Images
The eastward motion of the Caribbean plate since late Paleocene has resulted in a progressive detachment of the oceanic lithosphere from the continent South American plate. The detachment process is, however, poorly understood. At least two very different models have been proposed. The tensile tear model invokes a break-off of the northward descending slab following the collision of the two plates while the shear tear model suggests a near vertical dip-slip detachment at the two plates? boundary, the El-Pilar fault. Mapping the subducted oceanic lithosphere beneath the southeastern Caribbean is thus crucial to understanding the dominant process controlling the regional tectonics. In this study we mapped the presence of the cold subducted oceanic lithosphere in the transition zone by investigating the topography of the 410-km and 660- km seismic discontinuities. The two discontinuities are believed to be caused by temperature sensitive phase transitions of mantle minerals. We generated 1662 receiver functions from seismograms of 112 earthquakes recorded by the BOLVIAR (Broadband Ocean-Land Investigations of Venezuela and the Antilles arc Region) seismic array. The array consists of 35 temporary broadband stations, 13 temporary broadband ocean bottom seismometers, and 35 permanent stations of the national seismic network of Venezuela. We applied the common-conversion-point (CCP) stacking technique to the receiver-function data to image the P to S conversion events and their lateral variations beneath the array. P to S time moveout were calculated with 3D crustal and mantle velocity models. A 4th root stacking technique was employed to boost coherent signals in the data. Beneath the southeastern Caribbean, the 410-km is featured by a narrow (~200 km laterally) 20 km uplift with a NS trending centering at 64° west, while the 660-km is depressed broadly (> ~400 km) with a moderate amount of ~15 km, a scenario that is more consistent with westward descending of the oceanic South American plate. We also found a thick transition zone beneath the Falcon region in northwestern Venezuela, which probably is associated with the subducted Cocos plate. A flat 410-km was observed beneath the Guayana shield, suggesting that the shield has a stable moderate deep keel which has little effect on the underlying transition zone.
S21C-1855
Surface expressed subduction earthquake segment boundary and its verification in seismological data
One of the key questions in seismotectonics is what determines the size of an earthquake rupture and whether the geological and tectonic structure has an influence on rupture dynamics. We have found evidence for a subduction segment boundary on Mejillones Peninsula in Northern Chile expressed in topographic features which form an E-W transect over the peninsula in the area around 23.3°S. This transect subdivides the peninsula in a northern and southern part exhibiting differences in geological and tectonic parameters. Stratigraphic data, morphotectonic structures, fault patterns and age of deformation as well as the coastal uplift on both parts have been examined and showed that the transect might act as a "hinge" line or segment boundary provided that it is a persistent feature over various seismic cycles. Confirmation for this interpretation is coming from seismological data particularly from the intensive study of the aftershock sequence of the M8.0, 1995, Antofagasta earthquake. The main shock hypocenter calculations reflecting the start of the rupture are all located in the area of the proposed segment boundary. Several E-W aligned aftershock hypocentres with strike slip focal mechanism are also congruent with the "hinge"-line over Mejillones Peninsula. Furthermore, a number of seismological parameters, like the seimic b-value, do change at the segment boundary. A very strong support for our hypothesis is coming from the recent M7.8, 2007, Tocopilla earthquake which ruptured the adjacent part of the seismogenic interface north of the Antofagasta earthquake fault plane. Preliminary hypocenter determinations of some aftershocks suggest that the Tocopilla fault plane ends where the Antofagasta fault plane starts, which is again congruent with the proposed segment boundary. In our presentation we would like to summarize the geological evidences and give some new results from the seismological studies of the Tocopilla earthquake.
S21C-1856
Unusally Large Amplitude Core Phases Recoded In Oahu, Hawaii at Distances near 360°
We explored the core-mantle boundary region using data recorded in the 'Synpodal' distance range of 350° ≤ Δ ≤ 360°. Broadband data was collected from shallow and medium depth earthquakes of magnitude ≥5.7 mb in Hawaii, Japan, Iceland, Jan Mayen Island, and Southern California. Body waves converge at the synpode after having traveled once around the Earth with higher amplitudes than similar waves recorded outside synpodal or the antipodal region (178° ≤ Δ ≤180° (Rial and Cormier, 1980) making core phases such as SK¬mS and SKmSSK¬nS for m + n > 3 potentially observable. High amplitude phases arriving in the 50-60 minute time interval in the data are seen in horizontal component records from station KIP in Oahu, Hawaii. This interval is populated by multiple under-side reflections from the core-mantle boundary, beginning with SKKKS and ending with SK∞S. Our search found nothing similar in any of the other source regions. Synthetic seismograms produced by normal-mode summation using anisotropic PREM also have low amplitude arrivals in this time window. The absence of destructive interference from ellipticity and great-circle averaged heterogeneity in the synthetics accentuates the unusual nature of the high-amplitude arrivals at Hawaii. Having ruled out local sources of the anomalous energy, we are left to conclude that structure beneath Hawaii produces large amplification of these usually low-energy phases. Possibilities include extensive low-velocites at the base of the mantle or possible "funneling" of downward propagating energy along the mid- Pacific superplume. Further research will require 3D synthetic seismogram modeling. Rial, J.A. and V.F. Cormier, Seismic waves at the epicenter's antipode, J. Geophys. Res., 85, 2661–2668, 1980.
S21C-1857
P Wave Velocity Structure in The Inner Core Constrained by PKP(BC)-PKP(DF) Travel Time Differences Measured at Dense Seismic Array in Japan
A complete understanding about the velocity structure in the inner core is very important to determine the precise inner core rotation rate and study the dynamic processes in it. The measurement of differential travel times of two core phases, such as PKP(BC) and PKP(DF), is useful to study the inner core velocity structure because it is less sensitive to heterogeneities in the mantle and the possible mislocations of the events, as well. To detect the small scale variations in P wave velocity in the inner core, we measured differential travel times of PKP(BC) and PKP(DF) for events occurred in South America and recorded at dense broadband seismic network (F-net) in Japan. Their epicentral ranges are between about 145° and 155° and they are appropriate to detect PKP(BC) and PKP(DF) phases simultaneously. Our preliminary results over several years of the data show that the gradient of observed differential travel times with epicentral distance is clearly less steep than theoretical one based on 1D reference Earth model (IASP91). We systematically analyze these differences in an effort to determine the precise P wave velocity structure in the inner core and more differential travel time data for other events and station pairs will be used to give more constraints on the possible small scale velocity changes in the inner core.
S21C-1858
Sdiff waves guided by ULVZ beneath Hawaii
Prominent postcursors to Sdiff waves, which sample the northern edge of the deepest slow anomaly beneath Hawaii have been reported (To and Fukao, 2007). The events and stations of these data are located in Papua New Guinea and in southern USA, respectively. These Sdiff waveforms have the following features. 1) When waveforms of one event, which is recorded at many stations, are aligned, the arrival times of both main phase and postcursor depend strongly on the station azimuth. The main phase progressively delays to the south. On the other hand, the postcursor arrives earliest at an azimuth of around 60 degrees, which corresponds to the azimuth from New Guinea to HKT station in Texas. The postcursors arrive later at stations located to the north and south of these stations. The time differences between the two phases are about 25 sec at the azimuth of 60 degrees and about 45 sec at northern station at the azimuth of 50 degrees. 2) When waveforms from many events, which are recorded at one station, are aligned, the postcursors show a consistent delay of 30~40 sec with respect to Sdiff travel time predicted from a 1D earth model (PREM). They do not show the strong azimuthal dependence in contrast to the case where traces from one event are aligned. These and other observations suggest that the structure generating the postcursors is located off great circle path on its station side rather than on the event side. By assuming that a single scattering causes travel time delays of the postcursor, we conducted a grid search to find a location of the scatterer. Observed travel time differences between the two phases are compared with synthetics calculated from 1D ray tracing. Grid search is made for the location of a scatterer by making an assumption that the postcursor is the trapped wave: trapped into a thin slow velocity layer at the bottom of the mantle. In the process of the grid search, we assigned 10% velocity reduction at the CMB for the secondary waves. On the other hand, the main phases are insensitive to the presence of the thin slow layer; therefore the velocity reduction is not assigned. The result shows that the scatterer is located at latitude and longitude of 25 and 225 degrees on the CMB. The location corresponds to the northeastern corner of the Pacific superplume image in recent tomographic models. To and Fukao(2007), Eos Trans. AGU, 88(52), Fall Meet. Suppl., Abstract S14A-03