U53A-0047
Permeability anisotropy in marine mudstones in the Nankai Trough, SW Japan: Implications for hypothesized lateral fluid flow and chemical transport outboard of the trench
Characterizing dewatering pathways and chemical fluxes near and outboard of subduction trenches is important toward understanding early sediment dewatering and devolatilization. Quantifying fluid flow rates also constrains the hydraulic gradients driving flow, and thus ultimately hold implications for pore pressure distribution and fault mechanical strength. We focus on the well-studied Nankai Trough offshore SW Japan, where drilling has sampled the sedimentary section at several boreholes from ~11 km outboard of the trench to 3 km landward. At these drillsites, &δ37Cl data and correlation of distinct extrema in downhole chloride profiles have been interpreted to reflect substantial horizontal fluid flow to >10 km outboard of the trench within the ~400 m-thick, homogeneous Lower Shikoku Basin (LSB) facies mudstone. The estimated horizontal velocities are 13 ± 5 cm yr-1; the flow is presumably driven by loading during subduction, and mediated by either permeable conduits or strong anisotropy in permeability. However, the pressure gradients and sediment permeabilities necessary for such flow have not been quantified. Here, we address this problem by combining (1) laboratory measurement of horizontal and vertical sediment permeability from a combination of constant rate of strain (CRS) consolidation tests and flow-through measurements on core samples; and (2) numerical models of fluid flow within a cross section perpendicular to the trench. In our models, we assign hydrostatic pressure at the top and seaward edges, a no-flow condition at the base of the sediments, and pore pressures ranging from 40%-100% of lithostatic at the arcward model boundary. We assign sediment permeability on the basis of our laboratory measurements, and evaluate the possible role of thin permeable conduits as well as strong anisotropy in the incoming section. Our laboratory results define a systematic log-linear relationship between sediment permeability and porosity within the LSB mudstones. The overall variation in permeability for our suite of samples is ~1 order of magnitude. Notably, horizontal permeabilities fall within the range of measured vertical permeabilities, and indicate no significant anisotropy. Using laboratory-derived permeability values, simulated horizontal flow rates range from 10-4 to 10-1 cm yr-1, and decrease dramatically with distance seaward of the trench. With permeability anisotropy of 1000x (i.e. kh = 1000kv), simulated flow rates peak at 3 cm yr-1 at the trench, and decrease to 3x10-1 cm yr-1 by 10 km seaward. These flow rates are substantially lower than those inferred from the geochemical data and also lower than the plate convergence rate of 4 cm yr-1, such that net transport of fluids out of the subduction zone is not likely. If discrete conduits are included in our models, permeabilities of ~10-114m2 are required to sustain the inferred flow rates. However, no potential conduits in the LSB were observed by coring or logging- while-drilling. In contrast, net egress of fluids – and associated chemical transport and pressure translation – are plausible at margins where continuous permeable strata are subducting. Overall, our results highlight a major discrepancy between constraints on fluid flow derived from physical hydrogeology and inferences from geochemical data. In this case, we suggest that the chemical signals may be affected by other processes such as in situ clay dehydration and down-section chemical variations.
U53A-0048
Evidence of sequential deformation from peridotite to serpentinite: an implication for seismic properties in the trench side of the mantle wedge along a subduction zone
Hida Gaien belt, central Honshu, Japan, consists of high pressure metamorphic rocks and mafic/ultramafic rocks. Happo peridotite complex is the largest ultramafic complexes in Hida Gaien belt (e.g., Nozaka, 2006, JMG). Peridotites were classified into coarse-type peridotites and fine-type peridotites based on their microstructures, whereas serpentinites were classified into massive-type and foliated-type. The coarse-type peridotites, which occur with the massive-type serpentinites, show coarse granular textures. Crystal preferred orientations (CPOs) of the coarse-type peridotites were dominantly D-type and locally A- or E-type. The fine- type peridotites, which occur commonly with the foliated-type serpentinites, consist of remarkably elongated coarse olivine grains and fine olivine grains. Their olivine CPOs appears to be B-type that is characterized by [001](010) slip, although a- and c-axes of olivine grains were arranged on serpentine foliations. The distribution of peridotites and serpentinites in the Happo peridotite complex suggests that the original peridotite structures were overprinted by serpentinite structures. Since olivine CPOs within the fine-type peridotites were subparallel to the serpentinite foliations and lineations, it is likely that the foliated serpentinites would have subsequently occurred after the development of the fine-type peridotites. Seismic anisotropies of Happo peridotites have been estimated based on olivine CPOs, density and elastic constant of olivine (e.g., Tasaka et al., 2008, EPSL). As a result, mean velocities of P waves were 8.33-8.57 (km/s), whereas the mean velocities of P waves of serpentinites were reported 5.03-7.30 (km/s). These seismic anisotropies were compatible with seismic velocities of serpentinized area in the trench side of the mantle wedge, such as the Izu-Bonin subduction zone.
U53A-0049
Hydration induced strain softening in the fore-arc side of the mantle wedge: an example from Higashi-akaishi peridotites, the southwest Japan
Garnet-clinopyroxenites locally crop out in the Higashi-akaishi peridotite mass in the Cretaceous Sanbagawa metamorphic belt, central Shikoku. Garnet-clinopyroxenites contain 20-80% garnet, and these rocks occur within dunites as lenses, boudins or layers with thickness in the range from a few to around 50centimeters. Garnet-clinopyroxenites were plastically sheared, resulting in asymmetric boudins and rotation structures with a top-to-the-north sense of shear. Microstructural analyses were performed in optical microscope for clinopyroxene, whereas garnet microstructures were analysed by crystal-orientation maps made by the SEM- EBSD system. Both clinopyroxene and garnet grains show elongated shapes with somestructures. Grain sizes in garnet are comparable with those in clinopyroxene (0.22-0.3 mm) regard less of ether modal compositions. These indicate that both minerals could be deformed under similar plasticity. Crystal preferred orientations (CPO) of both clinopyroxene and garnet were analyzed by the SEM-EBSD system. [001] in clinopyroxene is subparallel to the stretching lineation; both [100] and [010] directions form girdle patterns perpendicular to the lineation (L-type). Garnet CPOs show no preferred orientation. Fabric strength (J-index) of clinopyroxene becomes intenser as increasing modal composition of garnet. The P-T condition of the garnet clinopyroxenites was estimated to be at 2.8 GPa /750-800 degrees by Mizukami and Wallis (2005). As a result, flow strength of garnet is similar to clinopyroxene, or weaker at this condition, as supported by flow laws of both minerals.
U53A-0050
Permeability-Porosity Relationships of Subduction Zone Sediments
Permeability-porosity relationships for sediments from Northern Barbados, Costa Rica, Nankai, and Peru subduction zones were examined based on their sediment type and grain size distribution. Greater correlation was observed between permeability and porosity for siliciclastic sediments, diatom oozes, and nannofossil chalk than for nannofossil oozes. For siliciclastic sediments, grouping of sediments by clay content yields relationships that are generally consistent with results from other marine settings and suggest decreasing permeability for a given porosity as clay content increases. Correction of measured porosities for smectite content generally improves the quality of permeability-porosity relationships. The relationship between permeability and porosity for diatom oozes may be controlled by the amount of clay present in the ooze, causing diatom oozes to behave similarly to siliciclastic sediments. For a given porosity the nannofossil oozes have higher permeability values by 1.5 orders of magnitude than the siliciclastic sediments. However, the use of a permeability-porosity relation may not be appropriate for unconsolidated carbonates such as nannofossil oozes. This study provided insight to the effects of porosity correction for smectite, variations in lithology and grain size in permeability-porosity relationships. However, further progress in delineating controls on permeability will require more careful and better documented permeability tests on characterized samples.
U53A-0051
Opal Cementation of Hemipelagic Sediment: Influence on Sediment Consolidation
Previous studies at ODP Sites 1173, 1174, and 1177 offshore southern Japan within Shikoku Basin suggest that a minor amount of opal cement inhibits consolidation of hemipelagic sediment approaching the Nankai Trough subduction zone. We examine additional ODP and DSDP sites offshore Japan (Sites 297, 442, 443, 444, 582), Alaska (Site 178), California (Site 1020), and Guatemala (Site 495) to determine if amorphous cementation of hemipelagic sediment is common and what effects this cement has on sediment strength and deformation. Within sediment from these sites, the presence of an amorphous silica cement is detected through secondary and backscattered electron (SEM and BSE) images. The cement appears as a coating at grain contacts, filling voids in zones of clastic material, and as altered material in contact with volcanic glass shards. The observed cement appears to be mobile and sourced from these volcanic glass shards, and may be the material strengthening the sediment. Using an alkaline leaching method, the measured amorphous silica content in hemipelagic sediments from Sites 297, 442, 443, 444, and 582 is higher in the Upper Shikoku Basin (USB) sediment than in the Lower Shikoku Basin (LSB), similar to results for Sites 1173, 1174, and 1177. Thus, amorphous silica cementation appears to be a common feature throughout Shikoku Basin. Results from Alaska, California, and Guatemala sites will help determine if this cementation is common where conditions are favorable for the formation of sediment strengthening amorphous silica cement.
U53A-0052
Compressional and Shear Wave Velocities of an Antigorite Rock at 1 GPa up to 550C
Serpentines play key roles in the water transportation in a subduction zone, the slab-mantle coupling and the generation of slab earthquakes. Geophysical mapping of serpentinized regions is important for understanding of subduction zone processes. There are three major forms of serpentine: antigorite, lizardite and chrysotile. Antigorite is stable up to 600°C at 1 GPa, while lizardite and chrysotile are stable below 300°C. Watanabe et al. (2007) showed that High-T type (containing antigorite) serpentinized peridotites have distinctive higher velocity and lower Poisson's ratio than Low-T type (containing lizardite and/or chrysotile) with the same density. Seismological observations on warm subduction zones like Costa Rica should be interpreted on the basis of antigorite properties. However, our knowledge on elastic properties of antigorite is still lacking. We have measured compressional and shear wave velocities on an antigorite rock at 1 GPa up to 550°C. The rock sample is mostly composed of antigorite (>95 vol.%) and grains are well aligned. The lineation is interpreted to be parallel to crystallographic b-axis, and cleavages normal to c-axis. Measuremens were made with various directions of propagation and oscillation. Reflecting the crystallographic structure, the sample shows strong anisotropy of velocity. The fastest direction of compressional wave is the b-axis direction (8.39±0.04 km/s at 550°C), and the slowest the c-axis direction (5.78±0.06 km/s at 550°C). No significant temperature dependence of Vp is observed in the b-axis direction, while Vp decreases by 3.3% from room temperature to 550°C in the c-axis direction. Shear wave velocity propagating in the c-axis direction at 550°C is 3.15±0.05 km/s and 3.44±0.03 km/s for oscillating parallel to the a- and b-axes, respectively. We can estimate elastic moduli of an antigorite single crystal from measured velocities on the assumption of orthorhombic symmetry of elasticity, and then calculate isotropic compressional and shear wave velocities of a randomly-oriented antigorite polycrystalline aggregate by Voigt-Reuss-Hill average. Compressional and shear wave velocities at 1GPa and 550°C are calculated to be 6.8±0.1 km/s and 3.46±0.02 km, resulting in Vp/Vs of 1.96±0.04. This will give a higher estimated degree of serpentinization than those based on Low-T type properties. Matsubara et al. (2008) reported a region of relatively high Vp/Vs (1.85) just above the subducting Philippine Sea plate beneath Southwestern Japan (warm subduction zone). The degree of serpentinization is estimated to be 42-61% based on our measurements.
U53A-0053
In situ experimental study of subduction zone fluids using diamond anvil cells
Experiments carried out in diamond anvil cells combined with in situ synchrotron light source measurements represent the only one issue to observe and study fluid equilibria in real time, at the pressure and temperature conditions of the subduction zones. We will present new results recently obtained at the DIFFABS beam line (SOLEIL Synchrotron) aiming at studying equilibria between silica-rich hydrous melts and aqueous fluids in the presence of U, Th, Pb, Ba and Br. We used synchrotron X-Ray fluorescence analysis performed in situ in Bassett-modified hydrothermal diamond anvil cells in order to monitor the chemical transfers of the studied elements between the phases in equilibrium at different pressures (up to 1.6 GPa) and temperatures (up to 900°C). We have calculated the partition coefficients for each studied element (i): Difluid/melt = Cifluid/Cimelt. Results show that U and Th exhibit more affinities for the silica-rich hydrous fluids in the presence or absence of Br, considered here such as an analogue for Cl, (i.e. 0.4 < DUfluid/melt < 0.7 depending on P,T conditions). Br partitioning shows that whereas this halogen element has very strong affinity to the aqueous fluid during magma degassing (DBrfluid/melt >> 10 after decompression) this coefficient decreases with pressure suggesting that Br would not be immediately washed out from the subducted plate during dehydration but may be recycled deeper in the mantle. These new data combined with previous ones obtained for Pb, Ba (Bureau et al., 2007, HPR vol 27, p. 235) and Rb, Sr, Zr (Bureau et al., 2004, Eos Trans. AGU, 85(47), V11C-05), allow us to propose a general outline of the fluid phase transfers through the subduction factory: (1) at shallow level: their nature and composition, the impact of the presence of halogens and the fertilizing role of such fluids in the mantle wedge, where the generation of arc magmas takes place (2) deeper in the mantle: where hydrous silica-rich supercritical fluids may also favour a deep recycling of a fraction of volatiles and trace elements present in the subducted oceanic crust.
U53A-0054
Spatial Heterogeneity of the Mantle Wedge Structure Corresponding to Interplate Coupling in NE Japan Forearc Region
The Japan Trench is a plate convergent zone where the Pacific Plate is subducting below the NE Japan arc. Interplate coupling along the plate interface is estimated to be strong by a backslip modeling of the land GPS observation in the middle to southern part of the arc [Suwa et al., 2006]. However, the off-Miyagi region and the off-Fukushima region show different characteristics of the interplate seismic activity. In the off Miyagi region, the large earthquakes with thrust mechanisms have occurred at an interval of about 40 years, and an interplate earthquake of M 7.2 occurred in this region on 16 August 2005. In the off-Fukushima region, few large interplate earthquakes have occurred while the background microseismicity is very high. In order to clarify differences in the seismic velocity structures, corresponding to the differences in the seismic activity between these regions, we estimated a 3D seismic velocity by the double-difference tomography method using both land station data and offshore station data. In our results, most of the relocated hypocenters are along the plate boundary. The subducting oceanic crust and the mantle wedge of the overriding plate were imaged as the landward dipping low velocity layer and the high velocity layer above it, respectively. In the mantle wedge, there are some velocity variations. Comparing spatial extents of the rupture areas of the 1978 and 2005 earthquakes [Yamanaka and Kikuchi, 2004; Yaginuma, 2006] and the velocity structure, we found that the location of high Vp, high Vs and low Vp/Vs anomaly corresponds to the rupture areas of the large interplate earthquakes. The high velocities and low Vp/Vs feature is interpreted as the non-serpentinized mantle wedge, and this may be the reason why M7 earthquakes repeatedly occurred in off-Miyagi region. In the off-Fukushima forearc region, high Vp/Vs area is found in the tip of this low velocity mantle wedge. This low Vp/Vs area is corresponding to the low backslip area revealed by GPS study [Iinuma et al. 2007]. We think that serpentinized mantle wedge delimited the extent of the seismogenic zone. The size of the seismogenic part is smaller in the off-Fukushima region with lower seismic coupling than in the off-Miyagi region.
U53A-0055
MT Measurements in the San-in Region, Japan, in Cooperation with Seafloor Electromagnetic Observations in the Sea of Japan
MT measurements have been made since 2006, together with Seafloor electromagnetic observations MT measurements in the Sea of Japan, in order to investigate the resistivity structure beneath the San-in region, Japan (the Tottori and surrounding region in the northern part of Chugoku district, southwestern Japan), where epicenters of seismic activities are remarkably distributing within a line belt with a width of about 4-9 km along the coastal line of the Sea of Japan. The depths of the hypocenters are located up to about 10km depths. In the seismic belt, several large earthquakes of M6.2-7.4 took place in 1943, 1983 and 2000. Moreover, quaternary volcanoes, such as Daisen and Oginosen volcanoes are also located in the seismic belt. Wide-band magnetotelluirc (MT) observations have been made along several survey profiles of almost N-S direction in the San-in region since 1998, to investigate heterogeneity in the crustal electrical resistivity structure. Through those wide-band MT measurements we found the low resistive region beneath seismogenic zone of the high seismicity belt along each MT profile, and that the upper resistive crust corresponds to the seismogenic zone in the Tottori and northern Hyogo region. The low resistive region found along each wide-band MT profile seems to form a conductive zone extending in the almost E-W direction beneath the seismic belt extending in the almost same direction of the conductor. This result strongly suggests the existence of crustal fluid beneath the seismogenic zone in the focal area. The survey lines should be extended toward the Sea of Japan and longer period MT data on the land area should be obtained, in order to investigate deeper resistivity structure and clarify the relationship between subducting Philippine Sea plate and the deeper resistivity structure beneath the San-in region. Therefore, we have carried out not only seafloor EM measurements but also longer period MT surveys on land since 2006 along two almost N-S profiles passing through lines of longitudes, 134.3E and 133.4E. It is important to analyze both data sets to obtain high accuracy resistivity structure. We will show the outlines of the observation and the preliminary results along the two survey lines in presentation.
U53A-0056
Electrical Structure Beneath the Back-arc Region of Southwest Japan: Results From a Seafloor Magnetotelluric Array Study
Seafloor electromagnetic (EM) observations were conducted off southwest Japan together with magnetotelluric (MT) measurements on land. The seafloor and land observations were to investigate the electrical conductivity structure in the back-arc region of southwest Japan. Epicenters of earthquakes in that region are known to distribute within a belt of about 4-9 km wide parallel to the coast line of the Japan Sea. Most of the focal depths are shallower than approximately 10km. In the seismic belt, several large earthquakes of M6.2-7.4 occurred in 1943, 1983 and 2000. Furthermore, quaternary volcanoes, such as Daisen and Oginosen Volcanoes are also located in the seismic belt. Wide-band MT observations have been made along a number of NS profiles on land since 1998 so as to reveal high conductivity regions beneath the seismic belt on each MT profile. It is noteworthy that the earthquakes seem to occur on the boundary between the upper resistive crust and the highly conductive body in the lower crust. The high conductivity regions found beneath each wide-band MT profile may form a conductive zone extending in an almost E-W direction. Coincidence of the hypocenter distribution with the upper surface of the conductive zone as well as the presence of deep low-frequency events suggests that crustal fluid must involve focal mechanism in the seismogenic zone. In order to clarify the relation among the mantle dynamics in the back-arc region, the lower crustal conductor found on land, and the volcanism in southwest Japan, we added two seafloor MT arrays, one traversing the non-volcanic region in the eastern part of southwest Japan and the other running through a volcanic ridge including the Oki Islands that are quaternary volcanoes as well. These seafloor arrays are indispensable to image the subducting Philippine Sea plate, a possible source of the crustal fluid. The electrical images of the back-arc region will be compared in terms of a fluid-driven part of the mantle dynamics beneath southwest Japan. A comparative synthesis of the arc dynamics beneath the Japanese Islands will be further discussed based on the electrical section of northeast Japan (Toh et al., 2006) in view of water circulation in the wedge mantle.
U53A-0057
Rayleigh-Wave Tomography in the Nicaragua-Costa Rica Subduction Zone
The goal of this study is to image crust and mantle structure in the Nicaragua-Costa Rica subduction zone by applying Rayleigh wave tomography to waveforms recorded by the TUCAN Broadband Seismometer Experiment. The 48-station TUCAN array included two dense station lines normal to the arc, one in Nicaragua and the other in Costa Rica, and two sparser lines along the fore-arc and in the back-arc. Stations were in the field from July 2004 until March 2006. Two-dimensional phase velocity maps of the fundamental mode from 15 to 167 s were inverted for three-dimensional shear-wave structure, using a starting model with an average crustal thickness based on receiver functions at TUCAN stations. The primary features in the shear wave models are fast velocities associated with the subducting slab and upper plate, and a slow mantle wedge. The subducting slab can be best seen at depths of 90-200 km with fast velocity anomalies of 4-7% compared to the regional average. The slow wedge region beneath Nicaragua occupies a larger volume and has lower velocities than are found in Costa Rica. Geochemical data contain strong along-arc variations between Nicaragua and Costa Rica that are consistent with a mantle wedge beneath Nicaragua that is more hydrated and may contain a greater extent and depth of melting. The shear velocities obtained from the Rayleigh wave inversions therefore correlate with the geochemical trends. They also match observations of slower velocities and greater shear attenuation in the mantle wedge obtained in local body wave tomography with TUCAN data. The average thickness of the upper plate in the back-arc in the Rayleigh wave inversions (~55 km) roughly corresponds to its thickness seen in shear attenuation models, a result that can be used to constrain thermal models for the subduction zone.
U53A-0058
Volume Distribution of Volcanic Edifices at the Southern Volcanic Zone of the Andes Between 33°S and 46°S
At convergent plate margins, lithospheric material is transported into the Earth's mantle by subduction processes; magmas form at depth and ascend to the surface, where they are erupted by arc volcanoes. Relations between in- and output are poorly understood because mass budgets are difficult to constrain. The quantification of magmatic output is therefore a first step towards understanding the subduction system of an arc segment, providing insights into volatile cycles, magmatic evolution and hazards. In order to constrain output fluxes at the Chilean Southern Volcanic Zone (SVZ), a first step is to estimate the volumes of volcanic edifices. We applied a Digital Elevation Model based on topographic data of the Shuttle Radar Topography Mission (SRTM) to determine the volume of each individual volcano or volcanic cluster, assuming a hypothetical flat volcano basis. This assumption is straightforward for isolated stratovolcanoes, e.g. Lanin, Villarrica or Tromen. Intrinsic uncertainties arise where volcanic edifices rest upon, or mold into, pre-existing rugged terrains of non-volcanic nature such as the Central Cordillera, in the case of calderas and complex multi-phase eruptive centers, where boundaries are obscured, and in the case of heavily eroded edifices. Accordingly, we offer a quality index to the calculated volumes and a discussion of error sources and ranges. Compared to the well-documented volcanoes at the Central American Volcanic Arc (CAVA), there seem to be several striking differences. While at the CAVA the average edifice volume is relatively small (50 km3) at a high volcano density (one volcano per 13 km of arc length), the average volume of > 200 km3 per volcano at the SVZ is allotted to a much smaller volcano abundance (1 per 25 km arc length). Individual edifices volumes are typically larger at the SVZ - the largest being the Cordon Caulle-Puyehue-Mencheca complex with a volume > 1200 km3 - than at the CAVA where Irazu displays the maximum of 242 km3. The wider scope of our study is to relate such surface expression of volcanic activity to the deep dynamics of the subduction system; i.e. to the magma productivity in response to subduction input and the controlling mantle processes. The Chilean subduction zone shows a pronounced and coupled along-strike segmentation in terms of topography, basin distribution, seismicity, and subduction geometry. We try to relate the spatial distribution of the volcanic edifices to these parameters in order to understand their along-arc systematics, which could hold clues to the melting processes.
U53A-0059
Crustal Structure and Deformation of the Incoming and Overriding Plates of the North Chilean Subduction Zone, 21-23.5ºS
We present MCS images of the crustal structure of the subduction zone of north Chile. The 50 Ma oceanic Nazca Plate subducts sub-orthogonally below the South American Plate at ~80-90 mm/yr. Here we focus on three reflection lines from Sonne 104 cruise that run perpendicular to the coast for ~450 km, imaging the overriding plate and some ~350 km of the oceanic incoming plate. The ocean plate displays well-defined top of the igneous crust reflections and fairly continuous Moho reflections 2-3 seconds (TWT) deeper. The deepest Moho reflections occur across the Iquique Ridge. The seismic data shows the deformation of the incoming oceanic crust as it approaches the outer rise bulge and bends into the trench. The top of the igneous crust shows clear development of faulting and prominent trenchward dipping reflections appears in the mantle, clearly below the Moho reflection. The bending-related deformation of the incoming plate forming horst-and-graben structures is observed underthursting below the margin. The inter- plate contact is observed about 50 km landward from the deformation front. The trench axis is largely devoid of stratified turbidites. But the three seismic lines show abundant debris from the continental slope accumulates at the slope toe forming a 5-10 km wide sediment prism. The prism is also observable in multibeam bathymetry maps. The landward segment of the frontal prism appears to be partially underthrusting the margin, providing clastic, fluid-rich material to the subduction channel. Thus the amount of fluid-rich sediment in this apparently starved trench seems to be considerable.
U53A-0060
Weak Interplate Coupling Induced by Seamount Subduction; Involvement of Fluid Migration in Determining Seismogenic Character
Seamounts are one of the obvious topographic features, and have been considered to constitute source regions of large earthquakes by raising the interplate mechanical coupling when subducted. However, recent seismic surveys have shown examples of erosion at the base of the overriding plate by seamount subduction while the seamounts have retained their original shapes to depths of at least ~ 8 km. Such observations may well propose weak interplate coupling over the subducted seamounts. The role of subducted seamounts in determining seismogenic character along the plate interface has still been in debate. M ~ 7 earthquakes have repeatedly occurred with a fairly constant recurrence period of ~ 20 years in a confined region along the Japan Trench where the Pacific Plate subducts beneath northeast Japan. A group of seamounts of various dimensions exist on the Pacific Plate, and a subducted seamount has been proposed as a source of the repeating large earthquakes. We conducted a marine seismic survey across the source region, and found a subducted seamount (~ 50 km in diameter, ~ 3 km in height) at ~ 10 km depth. However, the seamount is substantially offset from the source region, and therefore, it does not appear to constitute the source region by itself. We also conducted a passive seismic observation over the region. No earthquakes have been observed along the plate interface over the seamount. We analyzed the seismic waveforms of the 1982 M7.0 event along the series of the repeating M ~ 7 earthquakes to obtain its rupture process. The result shows that a large slip occurred in the frontal area of the seamount. These observations are in accord with weak interplate coupling along the plate interface over the seamount. Physical modeling of seamount subduction as convex upward geometry alone cannot explain such observations. Fluid migration may play an important role to determine seismogenic character around the subducted seamount.
U53A-0061
Quantitative Approach To Seamount Volumes And Eruption Rates For Serpentinite Mud Volcanoes
Serpentinite mud volcanoes in the Mariana forearc are formed by the hydration of mantle peridotite with slab- derived fluids. We calculated the volumes of five seamounts (Pacman, Celestial, Conical, Turquoise, and Big Blue Seamounts) using bathymetric, and (where available) multi-channel seismic reflection data. We interpolated the underlying pre-emplacent surfaces of each seamount using three methods (kriging, harmonic, and multiquadratic radial function), and three estimation techniques to define the pre-eruption seafloor surface beneath the seamounts (perimeter outline, wide area grid, and a geologically interpreted subsurface). We compared the percent difference between each. The interpolation methods produced volume results that differ as follows: 1.9% between kriging and the multiquadratic radial function, 3.6% between kriging and the harmonic function, and 3.8% between multiquadratic radial and the harmonic function. The techniques for determining the subsurface had larger percent differences as follows: 22% between the wide area grid and the geologic interpretation, 20% between geologic interpretation and the perimeter technique, and 18% between wide area grid and the perimeter technique. Based on linear regression results with a high R2, we conclude that there is a strong correlation between the results provided by all three methods. Based on the similarity of all of the regression slopes and their proximity to 1, we can conclude that no method will consistently over- or underestimate the volumes. The geologic interpretation technique should be used when subsurface data (seismic, drill cores, etc.) is available, or when workers have a strong geologic understanding and/or experience in the area. The other two techniques (perimeter and wide area grid) may be useful for large scale comparison studies that include many tens or hundreds of seamounts (or other features), where time constraints and a need for a systematic and repeatable approach is required, or for locations where workers do not have a strong geologic familiarity or experience. The oldest serpentinite muds date back to the initiation of subduction in the Mariana region in the Eocene. As a first approximation, dividing the volume of each seamount by 50 Ma yields a long-term eruption rate of 2.4 × 10-5 km3/year for Big Blue (largest seamount), and 3.26 ×10-5 km3/year for Conical (smallest seamount). As a comparison, these rates are approximately three to four orders of magnitude smaller than the igneous eruption rates for the Emperor Seamounts (1.0×10-2 km3/year). These rates are probably underestimates because some of the seamounts may have formed more recently, and because they erupt episodically. Expanding these calculations to include the volumes of all known serpentinite volcanoes in the Mariana forearc would give us a minimum estimate for the serpentinite production rate since the inception of subduction. Finally, we estimate the volume of the source protolith (peridotite) required to form the seamounts, using both a simple cylindrical model and a more complex cylindrical/cone model, assuming that the volume of the seamount approximates the volume of the protolith serpentinized minus the increase in volume (25-30%) caused by serpentinization. The calculated volumes of all of the seamounts needed a source conduit larger than 2km in diameter to produce enough serpentinite, requiring a more complex cone/cylinder model or an additional source of the protolith.
U53A-0062
The Impact of Splay Faults on the Fluid Budgets of Subduction Zones: A Modeling Approach
In subduction zones, permeability architecture acts as a primary control on forearc dewatering, devolatilization, and chemical transport. In addition, flow pathways govern the distribution of fluid pressure driven by sediment consolidation and dehydration, and thus impact the slip behavior and mechanical strength of faults. In particular, splay faults are ubiquitous features at convergent margins, although little is known about their role in distributing fluid pressures, transporting volatiles, and their overall effects on the fluid budget. Geochemical anomalies at seafloor seeps and shallow boreholes (Cl, B, and thermogenic hydrocarbons) suggest that these faults provide a hydraulic connection from deep within the subduction zone to the seafloor, and therefore play an important role in the flux of volatiles in the forearc. It is also likely that these faults affect fluid pressures along the décollement. Here, we report the results of a two-dimensional numerical model of coupled fluid-flow and transport designed to quantify the role of splay faults on fluid budgets, the distribution of fluid egress at the seafloor, the migration of chemically distinct fluids, and pore pressures within the forearc of an erosional subduction zone, using the well-studied Costa Rican margin as an example. Our model consists of a cross section perpendicular to the trench, extending from 10 km seaward of the deformation front to 40 km landward. In our models, we assign fluid sources within the subducted sediment section to represent both compaction and clay dehydration, in order to evaluate the migration of fresh (dehydration derived) water from deep in the subduction zone. We assign sediment porosity following an exponential decrease with depth, and define permeability using a relation to porosity derived from laboratory data. Because it is not composed of accreted sediment, we assign a uniform permeability to the overriding margin wedge. We evaluate the impact of splay faults by first establishing a baseline model with wedge and décollement permeabilities of 10-19 m2 and 10-15 m2 respectively, without including other faults. In subsequent model runs, we varied the splay fault and décollement permeability from 10- 17 m2 to 10-13 m2 in order to examine the effect of these changes on the fluid budget through comparison with the baseline model. Our results show that with a décollement permeability of 10- 15 m2, increasing splay fault permeability from 10-16 m2 to 10-14 m2 decreases the flux of fluid exiting the décollement at the seafloor from 52% to 28% of the total fluid budget. This represents a decrease of the overall budget in the baseline model, from 70% to 28% when considering a fault permeability of 10-14 m2. Inclusion of splay faults also substantially reduces fluid pressures on the décollement by focusing drainage toward the seafloor, especially in cases where fault permeability is greater than décollement permeability.
U53A-0063
Effects of Fluid Circulation in Subducting Crust on Megathrust Temperatures in Nankai Margin
Vigorous fluid circulation maintained in newly subducted ocean crust significantly affects subduction zone temperatures on the Nankai margin, Japan. The shallow part of the igneous ocean crust is pervasively fractured and thus highly permeable, allowing vigorous hydrothermal circulation. This circulation has been recognized as an important control on the thermal budget and evolution of ocean crust worldwide. However, previous subduction zone thermal models either do not include the effects of the hydrothermal circulation in ocean crust or assume that the circulation abruptly stops upon subduction. In a new subduction zone thermal model, we use a conductive proxy to incorporate the thermal effects of high Nusselt number fluid circulation in subducting crust. Results of this parameterized modeling indicate that hydrothermal circulation can reduce temperatures in the seismogenic zone of the Nankai margin plate boundary fault by ~20 °C at the updip limit of seismicity and ~100 °C at the downdip limit. With improved thermal models for subduction zones that include the effects of hydrothermal circulation in subducting crust, estimates of metamorphic reaction progress and interpretations of fault zone processes on various margins may need to be revisited.
U53A-0064
Volatiles and melting: Advanced models of fluid flow in subduction systems
Geodynamic models that incorporate theoretical and experimental constraints are critical for synthesizing the disparate geophysical and geochemical observations from subduction zones. Their application allows for quantitative inferences about the properties and processes occurring at these complex plate boundaries. Considerable work has developed high-resolution solid flow and thermal models for subduction zones, but there has been considerably less modeling of the explicit flow of hydrous fluids and magmas through the slab and wedge. Because many of the critical observations, particularly those of geochemistry, depend on the sources and pathways of fluids; considerable questions remain as how to relate these observations to the underlying dynamics. We present new models for calculating and exploring the generation and flow of fluids in subduction zones. Key features of these models include the ability to reuse existing high-resolution thermal/solid flow models on unstructured grids that can be generated for specific arc geometries, together with thermodynamic calculations of phase diagrams that can be used to calculate the rate and distribution of fluid production in the slab. Given these inputs, the models solve for the explicit flow of fluid using the magma-dynamics formulation of McKenzie (1984). The first generation of models address where and at what rates fluids are generated in the slab and explore the consequences of physical parameters such as permeability and solid rheology for affecting the potential fluid-flow paths through the slab and wedge. In particular, initial results suggest that flow paths are particularly sensitive to mantle rheology, with the possibility of considerable up- dip slab flow if the slabs are relatively strong.
U53A-0065
The Global Range of Subduction Zone Thermal Models
Two-dimensional thermal models provide insight to the dehydration and melting processes that occur in subduction zones. The wide range of slab geometries, ages, convergence velocities and upper plates result in an array of thermal structures that share many common features. We model forty-six arc sections in two dimensions using kinematically-defined slabs based on updated geometries from Syracuse and Abers [2006] to obtain a comprehensive suite of thermal models for the global subduction system. All models feature partial coupling between the slab and the overriding plate directly downdip of the thrust zone, invoked to replicate the cold nose observed in heat flow measurements and seismic attenuation. Four separate cases are tested with four separate sets of assumptions about the causes of the partial coupling: the downdip end of the partial coupling is at a constant depth, it is at constant distance trenchward from the arc, is defined by a critical surface slab temperature, or is adjusted such that the hottest part of the mantle wedge beneath the arc is at a constant temperature for all subduction zones. In all of these models, slabs reach temperatures where the top of the crust and sediments dehydrate before they reach sub-arc depths, and the overlying mantle wedge is too hot for hydrous minerals to be stable at sub-arc depths. The exceptions are cases where subduction is fast and coupling is controlled by a critical slab surface temperature (550°C); these cases also do not produce wedges hot enough to generate primitive arc magmas so probably underestimate temperatures. By contrast, the mantle within the downgoing plate remains cold enough for serpentine to be stable beyond the arc in all but the hottest subduction zones, allowing water to be carried beyond the arc in the slab.
U53A-0066
Chlorine and Fluorine Partition Coefficients Between Basalt and Dry Harzburgite at Mantle Wedge Conditions.
Volatiles released from the subducting slab play a fundamental role during the formation of arc magmas and their geochemical cycle in the wedge mantle, and recent advances of melt inclusion studies suggest that it may be possible to characterize the F and Cl content in arc primary magmas. Here, we present results of experimental determination of the Cl and F partition coefficients between orthopyroxene, olivine, and basaltic liquid. We achieved more than 110 measurements in 12 (nearly) dry melting experiments conducted over a range of Cl and F contents (0.13 to 0.61 wt% Cl and 100 ppm to 1.2 wt% F) at 1.2 GPa, 1330°C. The starting composition is a 30:70 mixture of PUM peridotite and basalt (similar to 82-72f, Gaetani and Grove, 1998, CMP). The low abundance F, Cl measurements in minerals were done by Cameca IMF 1280 at WHOI using the Cs+ primary and negative secondary ion mode. Our preliminary data show that the fluorine partition coefficients do not vary with its abundance, demonstrating the Henrian behavior up to 1.2 wt% F in melt. In contrast, we found that the Cl partition coefficients between orthopyroxene and basalt increase from 0.012 to 0.035 and correlate linearly with DMgO (ranging from 2.36 to 2.79). We also found that Cl and F are less incompatible in orthopyroxene than in olivine (in average: DClOpx/melt=0.02±0.008 and DClOl/melt=0.004±0.001 ; DFOpx/melt=0.04±0.004 and DFOl/melt=0.008±0.002 ). Our result show that Cl/F in a melt will be higher than its source.
U53A-0067
Continental Subduction Settings: Reduced Volatile Transport of Cl and S to the Surface?
A research focus of the Collaborative Research Center SFB 574 is the magmatic output of chlorine and sulfur as determined by melt inclusion analyses of primitive and evolved tephra deposits from the Central American (CAVA) and Southern Chilean (SVZ) arcs. Here we integrate our comprehensive geochemical data sets from the CAVA, consisting of 1200 melt inclusion and 300 whole-rock analyses from 71 volcanic centers, with our preliminary data from the SVZ and literature data from the Izu-Bonin-Marianas system (IBM). The highest chlorine contents in melt inclusions are reported from the IBM, an oceanic subduction zone endmember, whereas lower overall concentrations occur at the transitional CAVA. Melt-inclusion Cl-contents of both mafic and felsic CAVA rocks reach highest values in Nicaragua, where slab fluids dominate and the crust is thinnest, and gradually decrease towards the more continental Guatemalan segment of the arc. An exception to this trend are the high Cl contents in central Costa Rican tephras where the influence of the subducted, compositionally OIB-like Cocos Ridge is high. Preliminary analyses from the SVZ continental endmember reveal low melt Cl concentrations. Cl contents inversely correlate with Nd isotope ratios along the SVZ, with very low Cl and the highest Nd isotope ratios occurring at Llaima Volcano. Possible causes of this inverse relation are Cl-increase by crustal contamination reflected in lower Nd isotope ratios as well as lower degrees of partial melting of more enriched (pyroxenitic/eclogitic) material beneath thicker crust. The overall correlation, however, is modulated by local tectonic and compositional controls. Sulfur concentrations are also significantly lower in SVZ eruptives than at the CAVA, possibly reflecting absence of a S-rich source component at the SVZ. S-contents in mafic melt inclusions gradually decrease from ca. 2500 ppm in the southern SVZ to 500 ppm further north where the crust is thicker. This may reflect northward increasing S degassing during the melt's prolonged passage through the continually thickening continental crust. Although the total magma fluxes and magma production rates of the SVZ are still being determined, these first constraints on eruptive volatile output as recorded by melt inclusion geochemistry, integrated with comprehensive findings from the CAVA and IBM, may point to a broad picture of more limited volatile turnover in the subduction segments characterized by thick continental crust and thus a control on volatile contents based on the nature and thickness of the overriding plate.
U53A-0068
Heavy Halogen (Br, I) Injections into the Stratosphere from Large Explosive Volcanic Eruptions: Information from Melt Inclusions
Large explosive volcanic eruptions inject gases, aerosols and ash into the stratosphere, thus influencing stratospheric chemistry and the Earth´s radiation budget. Such periodic injections cause turbulent mixing which enhance chemical reactions. Reactive species responsible for catalytic ozone depletion following large eruptions include not only chlorine (Cl), but also the heavy halogens bromine (Br) and iodine (I) occurring in trace amounts in volcanic plumes. Due to the higher catalytic potential for ozone destruction of Br and I than of Cl, even trace amounts of these elements in volcanic emissions are relevant for stratospheric chemistry. We have analysed halogens in volcanic glasses and in glass inclusions in phenocrysts using electron microprobe and Synchrotron-XRF microprobe methods. Halogens from bulk glass samples were extracted using pyrohydrolysis, and analysed by ICP-MS. Eruptions investigated include Baitoushan, China /North Korea (ca. 969 AD), Mt. Hudson, Chile (1991), and several dacitic Quaternary eruptions from Nicaragua. Chlorine concentrations in glass inclusions are typically 2000 to 4000 ppm, which is on average about 50 percent higher than the concentrations in the matrix glasses. Br concentrations in glass inclusions are typically in the range of 2 to 20 ppm. This gives an average Cl/ Br ratios of about 300:1. Using the petrologic method, involving the concentration differences of halogens between the glass inclusion and those retained in the matrix glass, the average Cl/ Br ratio of the volcanic emissions were about 200:1 Typical I concentrations of Nicaraguan glass inclusions range between 1 and 3 ppm. The resulting Cl/ I ratio of eruptive emissions is about 1100:1. Depending on eruption size, each large event injected between several kt and several hundred kt Br and I into the atmosphere. As a first approach to estimate global Br and I fluxes from subduction zones, although affected by several sources of uncertainty, we combine these halogen ratios of 200:1 and 1100:1 for Cl/Br and Cl/I, respectively, with the global eruptive Cl flux of 1.5 E8 Kg/m/Ma (Wallace 2005, JVGR), leading to fluxes of 7.5 E5 kg/m/Ma for Br and 1.4 E5 kg/m/Ma for I.
U53A-0069
Fluid Sources on the North Anatolian Fault in the Sea of Marmara From He Isotope Measurements
The Northern branch of the North Anatolian Fault system in the Sea of Marmara accommodates 20-25mm
yr-1 of dextral strike-slip, most of the movement between the Eurasian and Anatolian plates, and is
expected to be the locus of a major earthquake in the near future. Fluid expulsion sites are common along
fault scarps as well as on splays (Zitter et al., 2008; Geli et al., 2008). In order to investigate the origin of
these fluids, the He isotope composition and He/Ne ratios have been determined in fluids sampled from cold
seeps in the Sea of Marmara during the MARNAUT cruise (May 2007). Helium isotopes in combination with
the He/Ne ratio are exceptional tracers for distinguishing mantle-derived fluids from shallow fluids (either
crustal or surface derived); the ratio 3He/4He varies by more than 3 orders of magnitude in this
system and typically it is possible to identify trace mantle inputs to fluids using this method. The results of the
analyses show that mantle-derived He is readily detectable in all but one of the fluids sampled during the
MARNAUT cruise. The majority of samples contain between 15 and 20% mantle helium; two samples from
the westernmost cold seep, located on a secondary fault branch, are clearly different, being dominated by
mantle He with over 70% mantle He. Local magmatic sources for this mantle He are unlikely, therefore we
conclude that the fault itself serves as a conduit for mantle-derived fluids, consistent with seismic studies that
indicate a deep origin for this part of the fault [3]. Similar quantities of mantle He were measured from fluid
emissions on the San Andreas Fault [4], consequently it appears that strike-slip faults commonly (or at least
the only two strike-slip faults studied to date) permit some fluid transport across the mantle-crust boundary.
Around 0.05 mol 3He is injected into the Sea of Marmara by the NAF every year, compared to the global
3He degassing flux of 1000 mol 3He y-1. Exactly how fluid transport across the lithological Moho occurs
is a matter of debate. Normally, the topmost part of the mantle and base of the crust are considered too
ductile to support sufficient permeability for fluid advection. Possibly fault coupling between brittle and lower
crust can increase permeability via vertical splays. Modelling is underway to asses the permeability required
to sustain the fluid flux through the lower crust to produce the observed 3He flux and to determine if this level
of permeability can be explained by fault coupling through to the base of the crust. [1] Zitter et al Deep Sea
Research 55, p552 (2008); [2] Geli et al, EPSL in press; [3] Tary et al, this issue; [4] Kennedy et al, Science
278, p1278 (1997)
http://cdf.u-
3mrs.fr/~henry/marmara/index.html