T51B-0538
Tomography of the subducting Cocos plate in central Mexico: Images of a truncated slab
The location of the subducting slab beneath Mexico City and its relation to the Trans-Mexican Volcanic Belt (TMVB) has been unknown because of the absence of deep seismicity that could be used to define the Wadati-Benioff zone. We used data from a temporary seismic network to locate the slab using seismic tomography. A break is seen in the Cocos plate under the TMVB. The break is seen with both P-wave and S-wave tomography and in a constrained tomographic inversion that finds parameters for a simple slab temperature model. The data used are 172 teleseismic earthquakes recorded by the Middle American Subduction Experiment (MASE). MASE was made up of 100 broadband seismometers spaced every 5 km running from Acapulco north through Mexico City almost to the Gulf Coast. In order to determine arrival time differences, Dt, across the array, waveforms were cross correlated. When Dt is plotted with respect to the latitude of the seismometer at which it was recorded, a Dt minimum (early arrivals) is seen near the TMVB. This minimum is shifted northward for back azimuths from the south, and southward for back azimuths from the north. The shift in the Dt minimum is indicative of a fast structure at depth. If there were no break in the slab, the localized minimum would not be seen. Tomography reveals an approximately 50-80 km thick slab diving into the mantle at about 75° to approximately 550 km depth and 375 km inland from Acapulco. We speculate the absence of deep earthquakes is due to low stresses in a young plate that has been truncated at depth.
T51B-0539
Shear wave splitting measurements and interpretation beneath Acapulco-Tampico transect in Mexico
We have examined shear wave splitting in teleseismic shear waves from 100 broadband stations installed from Acapulco to Tampico in Mexico over a period of 1.5 years (2005-2007). The instruments were part of the MASE (Middle America Subduction Experiment) which has the objective to build a geodynamical model of the subduction process beneath the Middle America Trench. The stations had a 5-6 km spacing and provided a unique data set which allows examination of the variation in splitting in high detail. Tomographic and receiver function studies in this area (done by MASE colleagues) show the presence of a flat slab under the western part of the array, and a steeply dipping slab beneath its center. According to geochronological data, the onset of flat slab subduction took place ~15-20 Ma, after the Cocos plate broke off the Farallon plate. We observe large splitting delay times with, on average, a fast direction in the northeast-southwest direction, but with considerable variation along the network. We compare the splitting results with the three dimensional structure inferred from the geochemistry and seismic analyses.
T51B-0540
Body-Wave Attenuation Structure in Central Mexico
Velocity spectra from moderate-sized earthquakes are used to investigate the P-wave attenuation structure in southern Mexico. In particular, we include regional events with magnitudes in the range 4.5 < M < 6.1 recorded from February 2005 to March 2007 on the Meso American Subduction Experiment (MASE) array, which consists of 100 broadband sensors from Acapulco to Tampico in central Mexico. By assuming a Brune-type source, a path-averaged frequency-independent Q is obtained for each seismogram in the frequency band 2 to 30 Hz, depending on the signal quality. These measurements are then inverted for spatial variations in Q. The 1- D tomography result shows a pattern of Q qualitatively similar to other subduction zones, with low attenuation crust (Q ~~1100), and high attenuation in the mantle wedge beneath the Trans-Mexico-Volcanic-Belt (Q < 250). The 2-D inversion shows more detail, with a low-Q zone above the slab between the depth of 100 km and 120 km. The location of the low-Q region provides some constraints on the geometry of the subducting slab, or with the structure provided by other methods such as receiver functions, the Q estimates will be used to estimate variations in viscosity. We also find that there are some attenuation variations in the crust.
T51B-0541
Modeling Of The Seismic Response Of The Flat Slab In Central Mexico
The subducted Cocos plate beneath Central Mexico as imaged with receiver functions from the MASE (Meso- American Subduction Experiment) array underplates the continental crust for a distance of approximately 300 km from the trench. In this study we investigate the azimuthal dependence of this image and construct a velocity model to match the impedance changes of the interface between the crust and slab as well as mid-crustal features. The receiver functions are modeled with a 2D finite-difference code. The initial results a low-velocity zone (lower than normal oceanic crustal velocities) are necessary to reproduce the impedance contrasts. We also use an inverse method to estimate the velocity changes.
T51B-0542
3D geodynamic evolution of the Mexican Subduction Zone
A crucial feature of the Mexican Subduction Zone (MSZ) is a shallow subhorizontal plate interface in its central part. This particular configuration of the young subducting Cocos plate (~14 Ma) apparently creates distinct geodynamic consequences, such as thin continental lithosphere, relatively shallow intraslab seismicity, and a distal volcanic front. Despite small variations in convergence and plate age along the trench, MSZ shows large variations in slab geometry. To the North, the young (~10 Ma) Rivera plate subducts at ~45 deg similar to the South where the Cocos plate plunges into the mantle at ~45 deg. The cause of such strong along strike variations are investigated with 3D geodynamic models that incorporate a low viscosity wedge (LVW) that can vary in depth and viscosity reduction. Using GPlates, we include realistic paleoreconstructions of the North America, Caribbean, Cocos, Rivera and Pacific plates for the last 40 Ma. We show that in order to avoid complete slab flattening due to trench rollback the slab needs to be viscously weak. The weakness of the subducting slab is consistent with intense and irreversible internal deformation of the slab bending into subduction. Also we show that a flat slab segment in Central Mexico can be created by shallow LVW (~100-150 km), bounded by steep slabs to the North and South where the LVW extends to higher depths (~300 km). Also the models show a time- dependent LVW that evolves into a low viscosity layer above the flat slab, essentially decoupling the slab from the overriding plate. This result is consistent with the lack of in-plane compression in Central Mexico and the presence of periodically (3-7 years) and large slow slip events.
T51B-0543
Relative Importance of Trenchward Upper Plate Motion and Friction Along the Plate Interface for the Topographic Evolution of Subduction-Related Mountain Belts
We present finite-element models that investigate the relative importance of both trenchward motion of the upper plate and interplate coupling for the development of topography at convergent margins (Hampel and Pfiffner, 2005). Commonly, the role of a trenchward moving continental plate for the growth of topography is neglected in both modelling and field studies. Instead, forces exerted by the downgoing plate on the continental plate as well as interplate coupling are thought to be responsible for the deformation of the upper plate. Our model setup includes an oceanic plate, which is in contact with a continental plate along a frictional plate interface and driven by slab pull. Both lithospheres have an elasto-visco-plastic rheology. The models demonstrate that friction along the plate interface can only lead to a high topography if the upper plate is moving toward the trench. Without such a trenchward advance, no high topography is generated, as the upper plate subsides owing to the drag exerted by the subducting plate. Increasing the coefficient of friction only amplifies the drag and increases the amount of sub¬sidence. Our findings imply that trenchward motion of the continental plate plays a key role for the development of mountain belts at convergent margins; subduction of an oceanic plate even with high interplate coupling cannot explain the formation of Andean-type orogens.
T51B-0544
Earth Sphericity Effects on Subduction Morphology
We present here the first application in Geodynamics of a Multipole accelerated Boundary Element Method (FMM- BEM) for Stokes Flow. The approach offers the advantage of a reduced number of computational elements and linear scaling with the problem size. We show that this numerical mehod can be fruitfully applied to the simulation of several geodynamic systems at the planetary scale in spheical coordinates and we suggest a general appraoch for modeling combined mantle convection and plate tectonics. The potentialities of the approach are shown investigating the effect played by Earth sphericity on the subduction of a very wide oceanic lithosphere , comparing the morphology of the subducted lithosphere in a spherical and in flat setting. The results show a striking difference between the two models: while the slab on a "flat Earth" shows slight undulation, the same subducting plate on a spherical Earth-like setting presents a distinct folding below the trench far from the edges, with wavelength of (1000km-2000km) as Pacific trenches.
T51B-0545
4-D Subduction Models Incorporating an Upper Plate
Thus far, relatively simplistic models of free subduction have been employed in which the trench and plate kinematics are emergent features completely driven by the negative buoyancy of the slab. This has allowed us to build a fundamental understanding of subduction processes such as the kinematics of subduction zones, the strength of slabs, and mantle flow-plate coupling. Additionaly, these efforts have helped to develop appreciable insight into subduction processes when considering the energetics of subduction, in particular how energy is dissipated in various parts of the system such as generating mantle flow and bending the plate. We are now in a position to build upon this knowledge and shift our focus towards the dynamic controls of deformation in the upper plate (vertical motions, extension, shortening, and dynamic topography). Here, the state of stress in the overriding plate is the product of the delicate balance of large tectonic forces in a highly-coupled system, and must therefore include all components of the system: the subducting plate, the overriding plate, and the underlying mantle flow which couples everything together. We will present some initial results of the fully dynamic 3-D models of free subduction which incorporate an overriding plate and systematically investigate how variations in the style and strength of subduction are expressed by the tectonics of the overriding plate. Deformation is driven in the overriding plate by the forces generated from the subducting plate and the type of boundary condition on the non-subducting side of the overriding plate (either fixed or free). Ultimately, these new models will help to address a range of issues: how the overriding plate influences the plate and trench kinematics; the formation and evolution of back-arc basins; the variation of tractions on the base of the overriding plate; the nature of forces which drive plates; and the dynamics controls on seismic coupling at the plate boundary.
T51B-0546
Large Scale, High Resolution, Mantle Dynamics Modeling
To model the geodynamic evolution of plate convergence, subduction and collision and to allow for a connection to various types of observational data, geophysical, geodetical and geological, we developed a 4D (space-time) numerical mantle convection code. The model is based on a spherical 3D Eulerian fem model, with quadratic elements, on top of which we constructed a 3D Lagrangian particle in cell(PIC) method. We use the PIC method to transport material properties and to incorporate a viscoelastic rheology. Since capturing small scale processes associated with localization phenomena require a high resolution, we spend a considerable effort on implementing solvers suitable to solve for models with over 100 million degrees of freedom. We implemented Additive Schwartz type ILU based methods in combination with a Krylov solver, GMRES. However we found that for problems with over 500 thousend degrees of freedom the convergence of the solver degraded severely. This observation is known from the literature [Saad, 2003] and results from the local character of the ILU preconditioner resulting in a poor approximation of the inverse of A for large A. The size of A for which ILU is no longer usable depends on the condition of A and on the amount of fill in allowed for the ILU preconditioner. We found that for our problems with over 5×105 degrees of freedom convergence became to slow to solve the system within an acceptable amount of walltime, one minute, even when allowing for considerable amount of fill in. We also implemented MUMPS and found good scaling results for problems up to 107 degrees of freedom for up to 32 CPU¡¯s. For problems with over 100 million degrees of freedom we implemented Algebraic Multigrid type methods (AMG) from the ML library [Sala, 2006]. Since multigrid methods are most effective for single parameter problems, we rebuild our model to use the SIMPLE method in the Stokes solver [Patankar, 1980]. We present scaling results from these solvers for 3D spherical models. We also applied the above mentioned method to a high resolution (~ 1 km) 2D mantle convection model with temperature, pressure and phase dependent rheology including several phase transitions. We focus on a model of a subducting lithospheric slab which is subject to strong folding at the bottom of the mantle's D" region which includes the postperovskite phase boundary. For a detailed description of this model we refer to poster [Mantel convection models of the D" region, U17] [Saad, 2003] Saad, Y. (2003). Iterative methods for sparse linear systems. [Sala, 2006] Sala. M (2006) An Object-Oriented Framework for the Development of Scalable Parallel Multilevel Preconditioners. ACM Transactions on Mathematical Software, 32 (3), 2006 [Patankar, 1980] Patankar, S. V.(1980) Numerical Heat Transfer and Fluid Flow, Hemisphere, Washington.
T51B-0547
Subduction Stability: Lithospheric Strength and Roll-back
In exploring the issue of subduction zone stability, we ran a series of simulations representing subduction systems consisting of simple 2D representations of oceanic lithosphere subducting beneath continental lithosphere. Our modelling software utilizes temperature dependent visco-elasto-plastic rheologies as well as a few proxies for significant chemical processes such as ecologitization and hydration. With externally imposed convergence rates, these models evolve from a contrived subduction initiation state to "normal-looking" subduction within approximately 10 million years. The simulations are then allowed to continue to evolve for up to 30 million more years. From our early results, we note that while most systems start with similar subduction geometries, they may deviate from each other over time. Notably, subduction initiated at "cooler" (and therefore stronger) junctures tend to form very stable subduction zones which maintain normal-looking geometries throughout the life of the simulation. However, subduction initiated at warmer margins tend to result in slab rollback relatively quickly. Systems with junctures of intermediate temperature also tend to subduct stably for a substantial amount of time, yet they too eventually result in rollback as the subducting slab entrains and removes some of the cooler lithosphere near the juncture, allowing hotter asthenospheric material into the contact region between the plates. The hot, low-viscosity material sharply reduces the fluid-dynamically derived suction force that partially supports the stable subduction geometry, facilitating the retreat of the subducting slab as well as the rifting of the over-riding slab. These simulations incorporate a variety of approximations and assumptions which may not reflect the actual conditions within the Earth. However, they do offer a chance to observe how a system that at least appears geometrically similar to observed Earth systems may behave when subjected to varying conditions. These early results suggest that the longevity of a stable subduction zone may be partly tied to the strength of the material of the over-riding plate with strong plates perhaps "holding" subduction zones in place while weak or weakened plates allow for slab retreat. Of course, other parameters may also contribute heavily to these outcomes, but they have not been systematically tested as of the time of this writing.
T51B-0548
Where Does Subducted Water Go? Estimates from Seismic Tomography of Serpentinite in the Forearc Mantle and Quartz in the Deep Crust
Much of the large amounts of water are carried down in subduction zones is driven upward into the overlying upper mantle and crust. The world oceans are consumed in approximately a billion years. Although this water is concluded to be responsible for the volatile-rich volcanic arc magmas, and reduced upper mantle viscosity and shallow convection in the backarc, the largest amount of water expulsion is estimated to be beneath the forearc, at least for subduction of young hot oceanic lithosphere such as Cascadia and SW Japan. The upward expelled fluids appear to be responsible for low-velocity serpentinite in the forearc mantle and in a concentration of silica deposited at the base of the forearc lower crust. High-resolution seismic tomography P- and S-wave velocities in northern Cascadia are now of adequate resolution to quantitatively estimate the total fluid incorporated into serpentinite hydration, and the rising silica-rich fluid required to deposit the silica-rich layer. The average serpentinization is ~30% from both P- and S-wave velocities (Vp=7.6, Vs=4.2). Just seaward of the arc where the forearc mantle wedge is thickest, this hydration approximately represents the total amount of water estimated to be released from the downgoing plate under this region in the most recent ~43 Ma subduction phase. Little fluid may reach the crust and surface. Similar results are found for a number other subduction zones. The ~10 km thick lower crustal layer of remarkable low Vp/Vs of 1.65 (Poisson's Ratio, σ, of 0.22) is concentrated mainly just seaward of the mantle corner. It may be explained by a ~20% volume quartz content removed from rising silica-saturated fluids. This represents a significant addition of silica to the continental crust, and to the average composition of some crustal accretion. The required total water flux again is similar to that estimated to be released from the underlying subducting slab. In this mid-forearc region, much of the fluid may reach the upper crust and surface. We conclude that within the forearc, arc, and backarc, we are now able to account for most of the total water carried into the earth by the subducted oceanic plates.
T51B-0549
Effects of fluid circulation in ocean crust on subduction zone temperature
Hydrothermal circulation plays an important role in redistributing heat within and extracting heat from high- permeability ocean crust. We examine the effect of hydrothermal circulation on subduction zone temperature by modeling fluid and heat transport within a vertical cross-section through a subduction zone. The effect of hydrothermal circulation is quantified by comparing results from simulations with coupled fluid and heat transport to those from simulations with only heat conduction (no fluid flow). We examine how convergence rate and subduction geometry influence the evolution of hydrothermal circulation within the subducting crust. Two end- member geometries are modeled, one mimics the Middle America subduction zone off Nicoya Peninsula, Costa Rica, the other simulates the Barbados margin. We simulate fluid and heat flow for systems with upper basaltic basement permeability ranging from 10-13 to 10-10 m2. Additionally, we examine the effect of permeability reduction within the basaltic basement as it is subducted. Models with fluid transport show suppressed temperatures along the subducting slab relative to models with no fluid transport. Faster convergence rates reduce the thermal effect of fluid circulation in subducting crust and temperatures are less suppressed. This reduced effect of hydrothermal circulation is also observed for smaller wedge tapers.
T51B-0550
Relation Between Fluid Release, Vein Formation and Earthquakes in Subduction Processes
Considering the localization of seismogenic zones during subduction processes, two factors are generally accepted for their determination: temperature and fluid pressure. Beneath large sedimentary accretion wedges the transition to aseismic stable sliding is temperature controlled. In this case the maximum temperature for seismic behavior in subducted crustal rocks is limited to ~350 °C. In addition, great earthquake ruptures initiated at less than this temperature may propagate with decreasing slip to where the temperature might reach ~450 °C. Although links between seismic activity and release of fluids in the subduction zones remain poorly understood, fluid pressure can strongly influence the mechanical behavior of rocks; such fluid may origin from in situ dehydration or may be derived externally. Experiments, conducted in numerous systems, have shown that rocks undergo sudden weakening and embrittlement (a change from ductile to brittle behavior) during their dehydration. In addition, conditions that support decreased rock permeability, will contribute to an increase of fluid pressure. In the presence of an externally derived fluid, embrittlement is independent of whether a rock undergoes a stable or unstable state of sliding. Abundance of synmetamorphic segregations in a variety of exhumed metamorphic rocks testifies to the importance of fluid-rock interaction at all depths. These segregations were created throughout the different stages of the tectono-metamorphic evolution. In subduction zones dehydration reactions during metamorphism represent a substantial source of fluid. We investigated the parameters that control fluids production and vein formation in a forearc wedge and in a slab by employing thermodynamic modeling. Our study confirms the hypothesis that fluids involved in vein formation originate locally from the surrounding rocks. No external fluids are involved. The quantity of released fluid is controlled by the thermal structure of the subduction zone as well as by the chemistry of the sediments involved in the wedge. We also show as already proposed by Delany & Helgeson (1978), that volume changes, which are accompanied by dehydration reactions during subduction contribute significantly to earthquake generation.
T51B-0551
Seismic Constraints on the Deeper-water Transportation in the Mantle - Reflected Seismic Waves from the Slab-mantle Interface Revisited
Based on the analysis of seismic receiver functions, Kawakatsu and Watada (2007, Science) recently reported the presence of a low-velocity layer atop of the subducting slab beneath northeastern Japan, and interpreted it as a pathway of water transportation into the deep mantle down to a depth of ~150km. They also suggested that this low-velocity layer may continue to the deeper mantle in a cold subduction zone environment such as beneath the central/southwestern Japan, where various reflected and/or converted seismic waves from the deep slab-mantle interface have been reported. In order to give further constraints on the structure of the slab-mantle interface, the seismic reflection phase reported by Fukao et al. (1978, Nature), and later analyzed by Helffrich and Stein (1993, GJI) is revisited. We collected the dense network (J-array, Hi-net, F-net) recordings of earthquakes occurred in the Ryukyu arc. The preliminary analysis for events occurred in 2007 shows clear secondary arrivals which share similar characteristics (arrival times and slowness) with those of Fukao et al.'s observation. For example, the data of the 2007/4/20 19:37 event (h=48km, Mw 5.8) exhibit secondary arrivals in a distant range between 11.5 and 15.5 degrees with an apparent velocity of ~0.1degree/sec, some of which originate from reflections at the slab-mantle interface deeper than 200km. By analyzing earthquakes with various epicentral distances, the dense array recording of recent Japanese networks should allow us to obtain a tight constraint on the structure of the slab- mantle interface. A mineralogical model which is consistent with the seismic observation will be also attempted.
T51B-0552
Compressional and shear wave velocities in serpentinized peridotites
Serpentinized peridotites in the wedge mantle play key roles in the transport of water and the slab-mantle coupling. Geophysical mapping of serpentinized regions is essential for good understanding of subduction zone processes. Tomographic studies have related low-velocity and high Poisson's ratio (high Vp/Vs) to serpentinized peridotites [e.g., Kamiya and Kobayashi (2000)]. Their interpretations are based on velocity measurements mostly on Low-T type (containing lizardite and/or chrysotile) serpentinized peridotites [e.g., Christensen (1996)]. However, it is questionable whether velocities of Low-T type are applicable to warm subduction zones like Costa Rica, where the other serpentine mineral antigorite is expected. From crystallographic studies, higher elastic stiffness is expected in antigorite. Watanabe et al. (2007) have shown that antigorite-bearing serpentinized peridotites have higher velocity and lower Poissonfs ratio than Low-T type serpentinized peridotites. However, their argument was based on the arithmetic mean of velocities in three mutually orthogonal directions. Anisotropic nature of rock samples should be properly taken into account in calculating the average velocity. The temperature dependence of velocities should also be investigated. In order to solve these problems, we are now conducting (1) determination of all elastic moduli from velocity measurements in various directions and (2) velocity measurements under high pressure and temperature conditions. Compressional and shear wave velocities are measured on octahedral specimens by the pulse transmission technique in various directions of propagation and polarization. Specimens were prepared from antigorite- bearing serpentinized peridotites collected in Hida outer belt. Measurements are conducted at the room temperature and under the confining pressure up to 180 MPa. All elastic moduli with orthorhombic symmetry are determined, and isotropic velocities and Poisson's ratio are calculated assuming that similar rocks are randomly packed. The higher velocity and lower Poisson's ratio of antigorite-bearing serpentinized peridotites has been confirmed. Velocities of an antigorite rock are measured up to 600°C at 1 GPa. The rock sample shows numerous parallel lines, which are interpreted to be parallel to b-axis of antigorite grains. The compressional wave velocity was 8.4 km/s and 6.6 km/s in the direction parallel and perpendicular to the lines, respectively. No significant temperature dependence of the velocity was observed in the direction parallel to the lines, while the compressional wave velocity decreases by 4% from the room temperature to 600°C in the direction perpendicular to the lines. Observed anisotropic behaviors are consistent with crystallographic natures of antigorite.
T51B-0553
Submarine landslides in subduction zones: insights from analogue and numerical models
Submarine landslides are common features at subduction zones. Detailed analysis of bathymetry around the Japan Trench found numerous slope failures along the inflection lines in the ocean floor slope, sub-parallel to the trench (Sasaki, 2003). Since the overriding plate at the Japan Trench has no accretion of ocean floor sediments but subduction erosion, these slope failures may have strongly controlled by the surface topographic relief of the subducting plate (Sasaki, 2003). In fact, several significant submarine landslide features are believed to be formed by seamount subduction to the trench. Such mega-scale submarine landslides, which also believed due to seamount subduction, have been recognized in Nankai Trough with typical accretionary prism development. The authors have been running analogue model experiments with granular materials and numerical simulation that approximate the geologic body as particles to model the deformation of accretionary prisms. The analogue model results are further analyzed by optical image correlation technique, PIV, to extract active deformation features in detail. By using the same tectonic model for subduction margin tectonics, we examined the surface deformation process by these two techniques. The model results produce two-types of slope failures: one related to seamount subduction and the other at the thrust front. The former examples are generally mega-scale and occur at the trench-side slope of the topographic high formed after the seamount subduction. The latter ones are minor but commonly found at the active frontal thrust. Such failures may be general features and expected to be found at drill holes of IODP (e.g. NanTroSEIZE). http://earth.kumst.kyoto- u.ac.jp/yamada/index_e.html
T51B-0554
The Effective Factors in the Deformation of Accretionary Wedges: Insight from Sandbox Experiments
Sandbox experiments are therefore frequently used to model the upper crust deformation in convergent plate margins and similar compressional tectonic regimes following the critical-taper Coulomb wedge theory. Several parameters are known tp affect the observed deformation in the sandbox experiments as well as actual accretionary wedges. In this study, control on wedge deformation from varying basal friction, rigid base tilt angle values, and presence of decollement levels simulated by thin glass microbead layers, are investigated through 2- D sandbox experiments The results obtained through the particle image velocimetry(PIV) analyses are useful as compareing phenomena displayed in the physical models and numerical simulations. By applying PIV analysis, two deformation cycles of frontal growth in sequence can be seen: a simple shear zone and a inverted triangular zone. The simple shear zone will become the thrust, and the inverted triangular zone will become the pop-up structure. In our sandbox results, the high basal friction models have more apparent uplift. The pop-up structures are always in low basal friction models, and these can be observed in the wider structural spaces. The higher angle of rigid base has more underthrusting at the displacement fields in the both of high and low basal friction models. In the experiments of the glass microbeads layer below 2 cm of the surface, the imbricate thrusts nucleate and propagate above the microbeads layer, the underthusting developments below the microbeads layer. The different thickness of microbeads layer and the depth below the surface are something we intend to experiment within the future. These parameters are valuable to determine the actual accretionary wedge.
T51B-0555
Seismic velocity and attenuation in Izu-Bonin subduction zone inferred from BBOBS data
Seismic velocity and attenuation in the upper mantle and the subduction slab surrounding Izu-Bonin subduction zone is measured by using the waveform data recorded by Broad Band Ocean Bottom Seismometer (=BBOBS). As a part of Stagnant Slab Project, we deployed 12 BBOBS from October 2005 to October 2006 in and around the Philippine Sea. The sites were equipped with three components Guralp CMG-3T sensors recorded at 200 Hz. The BBOBS waveform data provide us much more information on physical property of subduction zone than previous on-land data. We analyze regional (< 15 degree) earthquakes that sample the Pacific slab, mantle above the Pacific slab (=mantle wedge), and the mantle beneath the Pacific slab. The waveform data from events in Izu-Bonin slab were recorded by the BBOBS array with adequate signal to noise ratio. We measured travel time residual dt and path integrated attenuation t*. The frequency dependent t* is measured using high frequency spectral decay and then path averaged attenuation Q is calculated. Time domain signals are windowed with the window length of 15 s; before 2.5 s and after 12.5 s onset of manually picked P wave. This window includes main pulse of the waveform and excludes later phases. Then amplitude spectra and the slope are calculated over the frequency range from 1.0 to 9.0 Hz. The results clearly show low velocity and high attenuation in the mantle wedge above the Pacific slab and high velocity and low attenuation in the Pacific slab. The estimated Q of P-wave at 1 Hz are 50-200 in the mantle wedge above the Pacific slab, 170-500 in the mantle beneath the Pacific slab, and 500-1000 in the Pacific slab. For all regions, data from shallower events have lower Q values. The intensity of P-wave low velocity anomaly in the mantle wedge above the Pacific slab is up to 10%. The high velocity anomaly in the Pacific slab is up to 8%. The mantle beneath the Pacific slab seems to have no velocity anomaly. The S-wave to P-wave velocity anomaly ratio is 1.1-1.3 in both mantle wedge and the Pacific slab. The observed low velocity anomaly in the mantle wedge above the Pacific slab is too large to be explained only by thermal anomaly. The temperature anomaly estimated from attenuation exceeds both wet and dry solidus of pyrolite. We may expect the existence of certain amount of partial melt in the mantle wedge above the Pacific slab.
T51B-0556
3-D Geometry of the Philippine Sea Plate Slab just beneath the Tokyo Metropolitan Area and its Vicinity, Japan
We present a new view of three-dimensional geometry of slab(s) subducted just beneath the Tokyo metropolitan area, central Japan. Previously, several different models of the surface geometry of the subducted Philippine sea plate slab (PH slab) have been published using seismological data. First, we discriminate previously unknown seismic slab (called slab SG, or seismic slab SG) above the downgoing Pacific plate slab (PC slab), second propose possible models of the slab SG internal structure, and third discuss and demonstrate the tectonic background of slab SG. It is clear that the recent surface models of PH slab corresponds to the shallowest part of slab SG. Most of previous studies paid little attention to the tectonic characteristics of the vertical extent and/or the bottom geometry of slab SG with variable thickness. The bottom extent of slab SG beneath the Metropolitan area reaches 36.5N at least. The horizontal extent of seismic slab SG covers most of the Kanto Plain Lowland. The bottom depth of slab SG is approximately 120km near 36.5N and 139.0E, being the same as of the surface depth of PC slab there. Below the Sagami trough axis near 34.5N and 140.0E, slab SG locates at a depth of 80 to 90km. We confirmed that southernmost part of slab SG extends to areas around 34.3N and 140.3E at least, where seismic observation network density is not sufficient to locate correctly seismic events. The deepest portion of slab SG generally strikes the NNW-SSE direction, being approximately parallel to the volcanic front. The thickness of slab SG decreases gradually at areas west of the NNW-SSE striking deepest corner. We propose four basic internal models of slab SG as follows. (1) Slab SG consists of both PH slab at shallower depth and a deeper underlain slab (slab SL). (2) Bookshelf-like configuration of northwardly inclined multi-slabs on PC slab due to the intermittent southward shift of accumulation sites of short slab tips with episodic subduction at just south of the previously active paleo-Sagami trough(s). The evolutional bookshelf model is partly similar to accretion process near the deep trench system, but dynamic situation is not the same. (3) A structure combining models (1) and (2). (4) Slab SG is merely the eastern part, having been cooled by the downgoing PC slab, of 65 to 70km thick lithosphere of IOB (Izu Outer Block). IOB is the northeastern margin of the Philippine sea plate, ranging from the volcanic front along the Izu-Bonin Island arc to the Izu-Bonin trench.
T51B-0557
Toward Adjoint Tomography of the Japan Subduction Zone
In order to obtain better 3-D seismic velocity models of the Japan subduction zone and the neighboring region beneath Eastern China, we use the adjoint tomography technique to iteratively minimize the misfit between the synthetics and data from Hi-net, F-net and Global Seismographic Network (GSN) stations. We use Zhao et al.\/'s (1994) 3-D model embedded in Lebedev and Nolet's (2003) model as the initial model in the tomographic inversion. According to finite-frequency theory, the sensitive region along the ray path is given by a 3-D `banana-doughnut' kernel, and the overall spatial distribution of the sum of all available event-station kernels determines the resolvable volume in the inversion. We select a total of 300 events with Mw from 4.5--8 to obtain maximum coverage of this region while avoiding redundancy. We processed the data and synthetics using two types of bandpass filters: 6--30~s for all the records and 12--150~s for F-net and GSN records. The adjoint sources are constructed based upon the frequency-dependent traveltime misfit between synthetics and data. Given the adjoint sources, we use the adjoint spectral-element method to calculate banana-doughnut kernels for P, S and surface waves for the selected records. The weighted sums of the banana-doughnut kernels for all event-station pairs, with weights determined by the traveltime measurements, can be used to construct misfit kernels. These gradients will be used in a non-linear conjugate gradient algorithm to further refine the existing 3-D model.
T51B-0558
Imaging the Southern Hellenic Subduction Zone Through Migration of Scattered Teleseismic Body Waves
The active Hellenic subduction system located in the east-central Mediterranean region offers an excellent opportunity to study the dynamic interaction between subducting lithosphere and the surrounding mantle. An interdisciplinary effort to study these complex processes is currently undertaken by the MEDUSA ("Multi-disciplinary Experiments for Dynamic Understanding of Subduction under the Aegean Sea") project. As part of this initiative a dense array consisting of 40 broadband seismometers from the IRIS-PASSCAL pool has been deployed in southern Greece. The design of the array is optimized to both provide high-resolution for teleseismic migration techniques and be versatile towards application of other imaging techniques. This is achieved by a fan-shaped setup with an almost linear station- spacing in the regions where the slab is still shallow and a continuous broadening of the array as the slab descends deeper into the mantle. At this stage, the data have been used for receiver function analysis and migration of teleseismic scattered waves to obtain a two-dimensional image of the subduction zone. Preliminary results indicate the presence of two robust features: (1) a sharp Moho of variable thickness with an apparent thinning occurring at the eastern and western margin of the study area, and (2) a clearly defined subducted crust of the Hellenic slab, which plunges at an angle of 25° and exhibits low-velocities (relative to the surrounding mantle) to depths of at least 100 km.