T53D-01 INVITED
Evolution of subducted slab morphology in the Western Pacific based on seismic tomography and paleogeographic reconstructions
Plate motions and subducting slab morphology are intricately connected. Through the integration of seismicity, tomographic images, and relative plate motions the evolution of mantle structure can be interpreted. Tomographic images of P-wave, shear wave-speed, and bulk sound speed perturbations of the Northwest Pacific region have been interpreted to define the extent and geometry of the subducting Pacific plate in the upper mantle. The morphology of the subducted Pacific plate along the Kurile-Japan-Izu-Bonin-Mariana arc system was found to vary both in geometry and dip along the entire length of the margin. To understand these differences and evolution to the current slab morphology a tectonic reconstruction for the Western Pacific was created, which describes the geologic history of the past 20 million years. The paleogeographic reconstruction illustrates the collision of the Japan and Kurile arcs, the opening of the Kurile Basin and Sea of Japan, change in motion of the Izu-Bonin arc, developing curvature of the Mariana arc, disparity in Pacific plate velocities along the convergent margin, and variation in rates of trench retreat along different segments of the arc system. The new plate motion model and interpretations of the physical properties of the mantle imaged with the P-wave and joint tomography are tools to assess the spatial and temporal evolution of the Pacific plate morphology from the mid-Miocene to the present and provide limitations in plausible plate motions for the region.
T53D-02
High-frequency seismic reflections from the slab-interface between the Cocos plate and the mantle wedge beneath Costa Rica.
The oceanic lithosphere and sediments subducting beneath the volcanic arc of Central America experience high- pressure and high-temperature metamorphism during which they gradually release aqueous fluids and hydrous melts. The devolatization process has been studies extensively using geochemical controls on the transport of elements, earthquake seismology and geodynamic modeling. Seismology is powerful tool to measure physical properties in situ, but local earthquake studies do not yet have the instrument density and the resolving capability to delineate the fine-scale structure of the slab-mantle interface. This nature of this thin boundary, however, is important for our understanding of geochemical mass transfers and for the geodynamic evolution of subduction zones. In this presentation we show S-P and S-S seismic reflections from the Cocos plate slab-interface beneath Costa Rica that were recorded during the 2005 TICOCAVA explosion seismology experiment. Seismic refractions from Shot point 21, which was located in the backarc 50 km behind the volcanic front of the Cordillera Central, were recorded by an array of 740 Reftek Texans that were deployed from the Herradura Peninsula at the Pacific coast, across the Cordillera Central to the Caribbean coast. Instruments located in front of the arc, at distances between 134 km and 57 km of the shot location, recorded both P and S waves that must have reflected from the top of the slab beneath the mantle wedge at depths between 35 km and 58 km. The slab reflections are unique in two aspects: We clearly observe the variation of S-P and S-S reflection amplitude with incidence angle at the slab-interface over a wide range of angles (13 to 53 degrees). In addition, we find that the frequency of these slab reflections is between 15 and 20 Hz, which is much higher than the dominant frequency of turning waves recorded from the same explosion. Modeling of these slab reflections shows that the slab interface may be a zone of just 100-200 meter thickness where the shear wave velocity is greatly reduced. The presence of pressurized fluids at the slab interface can explain the unusual reflection amplitudes for high-frequency seismic waves. Thermal models of the Central American subduction zone predict that large amounts of water are released from the subducting plate into the upper portion of the mantle wedge. The presence of a continuous, reflecting boundary at the top of the slab suggests that these slab fluids travel along the slab interface before they are absorbed into the overlying mantle. This form of lubrication of the shear zone may be essential for maintaining subduction of oceanic lithosphere into the deeper mantle.
T53D-03
The subduction zone flow field from seismic anisotropy: A global view
Understanding the flow field that accompanies subduction remains one of the important unsolved problems in the Earth sciences, since it has implications for mantle dynamics, the tectonics of the back-arc region, and the physical and chemical characteristics of arc volcanism. An important constraint on this flow field is provided by observations of seismic anisotropy, as manifested by shear-wave splitting. We have compiled a global splitting data set (including previously published studies and new measurements) for 12 subduction zones worldwide and have searched for trends by comparing splitting observations with tectonic parameters. We find systematic variations in both mantle-wedge and sub-slab anisotropy with trench-migration velocity, Vt, referenced to a hotspot reference frame. These variations are most simply explained by the creation of a three-dimensional trench-parallel flow field induced by this trench motion. In particular, we find that in the subslab region, trench- parallel flow dominates the flow field and its magnitude scales with Vt. In the mantle wedge, the trench- parallel flow field interacts with classical 2-D corner flow produced by the convergence of the two plates. The relative influence of these two flows is governed by the relative magnitude of Vt and the convergence velocity, Vc. Thus, trench migration constitutes a first-order property in controlling the flow field accompanying the subduction process.
T53D-04
Trench-parallel flow in the mantle wedge: insights from integrating seismology and 3D subduction zone modeling
The mantle wedge in many subduction zones is characterized by a cold, low attenuation fore-arc and a hot, high attenuation arc and back-arc. The transition from fore-arc to arc mantle correlates in the Honshu, Nankai and possibly the Ryukyu subduction zones with a change in orientation of shear wave splitting from trench-parallel to trench-normal. For the Marianas, Andes and Central America subduction zones the trench-parallel splitting occurs further into the arc as well. The differences between these fundamental seismological observations suggest that multiple mechanisms can be responsible for the formation of trench-parallel anisotropy. Proposed mechanisms include fabric transitions in olivine due to changes in hydration and stress, melt alignment, and 3D flow. We use high resolution 2D and 3D dynamical models of these subduction zones to investigate the role of 3D flow and olivine fabric transitions in generating the observed patterns of anisotropy. The slab geometry is determined from available shallow geophysics and Benioff zone seismicity. For Ryukyu we obtain satisfactory splitting magnitudes for models that are based strictly on 2D corner flow with the B-type fabric in the cold fore-arc mantle. The widespread trench-parallel anisotropy in the Marianas and the 30-34S segment of the Andean subduction zone is better explained by 3D flow driven by along-trench pressure differences induced by changes in slab dip (Andes) or curvature (Marianas).Initial models for the Nicaragua-Costa Rica and Cascadia subduction zones incorporating present-day slab shape are dominated by 2D cornerflow and cannot explain the observed anisotropy. Future modeling will explore whether the time-dependent evolution of this subduction zone can create three-dimensional flow patterns with significant arc-parallel flow.
T53D-05
Comparative Study of Subduction Zone Thermal Structure: Implications for Slab Dehydration and Fluid Supply for Mantle Wedge Serpentinization and Arc Volcanism
Aqueous fluid from the dehydrating slab is critical to the processes of forearc mantle wedge serpentinization and arc volcanism. Its availability depends mainly on the thermal structure of the subducting slab, which is strongly controlled by the age of the slab and mantle wedge flow. In this study, we develop 2-D steady state numerical thermal models for a number of subduction zones to investigate how the thermal structure affects the fluid supply. Subduction zones investigated in this comparative study include Cascadia, Chile, Colombia-Ecuador, Costa Rica, Hikurangi, Kermadec, Mariana, Mexico, Nankai, NE Japan, and Sumatra. Geophysical and geological observations indicate that the shallow part of the forearc mantle wedge is decoupled from the subducting slab and does not participate in the wedge flow. The maximum depth of the slab-mantle wedge decoupling is one of the most important parameters controlling the subduction zone thermal structure. In our models, the depth of downdip transition from decoupling to coupling is constrained by surface heat flow and the location of the arc, beneath which the mantle wedge temperature is required to be greater than 1200°C. We find that the optimal transition depth for most subduction zones is in the range of 70 to 90 km; too shallow a transition will over-predict the forearc heat flow, and too deep a transition will under-predict the mantle temperature beneath the arc. The model results show that, for all subduction zones, the stagnant part of the forearc mantle wedge is sufficiently cold to allow serpentine to be stable, but the actual degree of its serpentinization should differ between different subduction zones depending on the availability of fluids. For subduction zones with a young and warm slab such as Cascadia and Nankai, dehydration of the subducting crust peaks at depths shallower than the decoupling-coupling transition depth and therefore provides ample fluid to serpentinize the overlying stagnant mantle wedge. Because of the shallow slab-crust dehydration, a warm slab becomes relatively anhydrous when it reaches the depth beneath the volcanic arc and provides little fluid for melt generation, leading to subdued arc volcanism. For subduction zones with an old and cold slab such as NE Japan and Hikurangi, slab-crust dehydration does not peak until around the depth of the decoupling-coupling transition, such that the cold slab provides large fluid flux into the hot and flowing part of the mantle wedge to promote melt production and arc volcanism. Little metamorphic fluid is available at shallower depths to serpentinize the stagnant part of the mantle wedge at cold-slab subduction zones except at ocean-ocean subduction zones such as Mariana and Kermadec, where the overriding mantle material at shallow depths beneath the thin oceanic curst is likely to be highly serpentinized by fluid released from porosity collapse in the subducted sediment and crust, not by metamorphic fluids from slab dehydration.
T53D-06
Dehydration, Kinetics and Stress - Numerical Modelling of Gypsum as a Guide to Dehydration Reactions Under Tectonic Stress
In subduction zones, rocks of the oceanic crust and hydrated upper mantle are dehydrating whilst under stress. The feedbacks between deformation and dehydration will govern the rheology and the rate of dehydration and hence exert a considerable influence on subduction zone behaviour. The interaction of dehydration reactions with deformation is, however, far from fully understood. A key issue relates to the thermodynamics of dehydration. Such reactions are pressure dependent. However, during deformation the material is under deviatoric stress – so which is the right value of "pressure" to use? Moreover, if the experiment (or natural system) is drained, fluid pressure is different to confining pressure – the same question must be asked again. Our assertion is that the local driving force for dehydration is different at each solid/solid interface and each pore surface. The large scale behaviour is the net result of different local processes coupled to the kinetics of reaction, and the fluid pressure evolution due to reaction and fluid flow. We aim to use the behaviour of gypsum during dehydration and deformation to test this assertion and as a guide to more general aspects of such interactions. Laboratory experiments are being coupled to numerical modelling. To separate the rates of different steps in the process we use available data on the gypsum to bassanite and/or anhydrite transformations, plus our own experiments. As a foundation for understanding the complex interactions we have: 1. Characterized and modelled the reaction rate for gypsum dehydration under 1 bar in the absence of deformation. The latter provide key information on the rate and mechanism of dehydration, using the Johnson- Mehl-Avrami approach. 2. Investigated the rheology of gypsum in the absence of dehydration, i.e. at low temperatures. 3. Produced preliminary numerical models for porosity evolution during dehydration, where fluid pressure is coupled to reaction progress, but no deviatoric stress is applied. A thorough understanding of these separate aspects is required to interpret more complicated experiments which we have begun, and to model the full behaviour of the system under tectonic stress.
T53D-07
Effects of fluid circulation in subducting crust on temperatures along the plate interface in Nankai margin
Temperatures within subduction zones have been suggested as controls on the updip and downdip limits of the megathrust seismogenic zone due to their influence on diagenetic and metamorphic reaction progress. We suggest that fluid circulation in the subducting crust may be an important control on the temperature distribution in the Nankai margin. Most thermal models of subduction zones have not included hydrothermal circulation within ocean crust. While heat advection by fluid flow through the wedge, décollement, and underthrusting sediment is likely negligible, fluid flow in high permeability fractured basaltic basement may redistribute large quantities of heat and affect temperature distribution along the megathrust. We simulate temperatures in the Nankai margin with a model that approximates the effect of fluid circulation homogenizing temperatures within the basement aquifer of subducting crust. Fluid circulation in subducting crust may help explain two previously unexplained thermal anomalies on the Nankai margin, namely anomalously high heat flux in Nankai Trough and an anomalously steep drop in heat flux with distance landward on the margin wedge.
T53D-08
Seismic imaging of dehydration reactions in the subducted oceanic crust
The combined analysis of high-resolution seismic images of the Alaska and Cascadia subduction zones reveals where metamorphic fluids are released in the system. Both images, which are produced by 2-D migration of scattered teleseismic P-waves, show the subducted oceanic crust as a dipping low-velocity layer with a clear termination depth. In Alaska, the crust is thicker (15-20 km compared to 8 km) and terminates at greater depth (120 km compared to 40 km) than in Cascadia. Based on metamorphic reaction estimates and geodynamic models, we demonstrate that the termination depth corresponds to eclogitization of the crust triggered by dehydration of water-bearing minerals, and that the location of this reaction is dependent on the thermal structure of the subducted slab.