Orientation Of Olivine Research Articles

Preferred Mineral Orientation by Recrystallization under non-Hydrostatic Stress and Geotectonic Stress in the Upper Mantle

Significant elastic anisotropy of olivine single crystal is well known. The velocity of P waves has a maximum in the direction of crystallographic a-axis (identical to optical-acoustic Z-axis). It is also known that dunite and peridotite which are found at the central core of an orogenic zone have a strong tendency to preferred orientation of olivine. The preferred orientation of olivine and elastic anisotropy of dunite in the tectonic zone of Japan are measured optically and acoustically. The latter is carried out using the ultrasonic wave at high pressures. The direction dependence of P wave velocity is expressed by V, = a + b cos2 8, where the direction of 8 = 0 is found to be nearly equal to the direction of lineation which coincides statistically with the direction of a-axis concentration. Recently the existence of anisotropy of the uppermoFt mantle came to be convincing by analysing the results of seismic refraction profiles at the ocean bottom. The P wave velocity can be expressed empirically as V, = a + b cos2 8 where 0 is the azimuth to the perpendicular to the ridge direction. Here the observed seismic anisotropy is attributed to the regional preferred orientation of constituent minerals of rocks composing the upper mantle. The mechanism to produce preferred orientation is known to be either a recrystallization due to elastic anisotropy of minerals or a non-elastic effect such as gliding. In olivine the translation gliding in (010) plane is known. However, it does not seem that such a pure translation gliding can give rise to the significant preferred lattice orientation under stress. Let the preferred orientation of olivine aggregates be due to recrystallization. The stable mineral orientation is specified by the minimum condition of a certain thermodynamic function 4. It must be consistent with the stress and strain continuities at the grain boundary. According to our theory (1) and (2) 4 is represented by a sum of: (a) A function dependent upon a grain shape (interface configuration between grains) (first order). It plays a significant role in mineral orientation of tectonite. However, it is insensitive to the elastic anisotropy. (b) A linear function of deviatric stress component (second order). (c) A quadratic function of stress component (second order). Terms (b) and (c) are related to the stored energy and potential due to deformation and are dependent upon the stress-strain relation, while (a) is independent of stressstrain relation. The expression of C#J is shown in Table 1. Under high hydrostatic pressures greater than lOkb, term (b) becomes predominant. In the actual mineral aggregates, the grain boundary is not a plane, and the surface integral expression of 4 is necessary. To find the condition C#J = min the variational problem for the relative angle between the axis of anisotropy and the axis of external

沈み込み帯での地震波異方性とかんらん石の格子選択配向

Seismic anisotropy is a powerful tool for detecting the geometry and style of deformation in the Earth's interior, as it primarily reflects the deformation-induced preferred orientation of anisotropic crystals. Although seismic anisotropy in the oceanic upper mantle is generally parallel to the plate motion, the propagation direction of the fast shear-wave is oriented subparallel to the trench axis in several subduction zones. In this study, the distribution of seismic anisotropy in the mantle wedge is inferred from deformation mechanism: a lattice-preferred orientation and seismic anisotropy are generated by deformation via dislocation creep, but not by diffusion creep or frictional sliding. Based on the thermal structure beneath northeast Japan, deformation throughout most of the mantle wedge is inferred to be controlled by diffusion creep, and the region of dislocation creep is limited to a thin layer of 10-20 km thickness within a region of relatively high stress and low temperature located above the subducting plate and beneath the island arc crust. The variation of the seismic anisotropy in subduction systems is probably due to the change of active slip-system in olivine as a consequence of chemical and physical properties in the mantle wedge.

A multiscale approach to model the anisotropic deformation of lithospheric plates

The association of experimental data showing that the plastic deformation of olivine, the main constituent of the upper mantle, is highly anisotropic and the ubiquitous seismic anisotropy in the upper mantle, which indicates that olivine crystals show coherent orientations over scales of tens to hundreds of kilometers, implies that the long‐term deformation in the upper mantle is anisotropic. We propose a multiscale approach, based on a combination of finite element and homogenization techniques, to model the deformation of a lithospheric plate while fully considering the mechanical anisotropy stemming from a strain‐induced orientation of olivine crystals in the mantle. This multiscale model explicitly takes into account the evolution of crystal preferred orientations (CPO) of olivine and of the mechanical anisotropy during the deformation. We performed a series of numerical experiments simulating the uniaxial extension of a homogeneous (100% olivine) but anisotropic plate to test the role of the olivine CPO on the plate mechanical behavior and the link between CPO and mechanical anisotropy evolution. Even for this simple solicitation, different orientations and intensity of the initial olivine CPO result in variable plate strengths and deformation regimes. A plate with an initial CPO where the olivine [100] and [010] axes are concentrated at 45° to the extension direction has high resolved shear stresses on the easy (010)[100] and (001)[100] slip systems of olivine. This results in low strength and in deformation by transtension. Plates with an initial CPO where the maximum of [100] axes is parallel or normal to the extension direction show a high initial strength. Isotropic plates have an intermediate behavior. The progressive rotation of olivine [100] axes toward the imposed stretching direction results in hardening in all models, except in those characterized by an initial concentration of olivine [100] axes normal to the imposed extension, in which softening is followed by hardening.

Peridotites from a ductile shear zone within back‐arc lithospheric mantle, southern Mariana Trench: Results of a Shinkai 6500 dive

Two N–S fault zones in the southern Mariana fore arc record at least 20 km of left‐lateral displacement. We examined the eastward facing slope of one of the fault zones (the West Santa Rosa Bank fault) from 6469 to 5957 m water depth using the submersible Shinkai 6500 (YK06‐12 Dive 973) as part of a cruise by the R/V Yokosuka in 2006. The dive recovered residual but still partly fertile lherzolite, residual lherzolite, and dunite; the samples show mylonitic, porphyroclastic, and coarse, moderately deformed secondary textures. Crystal‐preferred orientations of olivine within the peridotites show a typical [100](010) pattern, with the fabric intensity decreasing from rocks with coarse secondary texture to mylonites. The sampled peridotites therefore represent a ductile shear zone within the lithospheric mantle of the overriding plate. Peridotites were probably exposed in association with a tear in the subducting slab, previously inferred from bathymetry and seismicity. Furthermore, although the dive site is located in the fore arc close to the Mariana Trench, spinel compositions within the sampled peridotites are comparable to those from the Mariana Trough back arc, suggesting that back‐arc basin mantle is exposed along the West Santa Rosa Bank fault.

Bottom to top lithosphere structure and evolution of western Eger Rift (Central Europe)

We present model of the structure and development of the entire lithosphere beneath the western Eger Rift (ER). Its crustal architecture and paths of volcanic products are closely related to sutures/boundaries of uppermost mantle domains distinguished by different orientations of olivine fabric, derived from 3-D analysis of seismic anisotropy. Three different fabrics of the mantle lithosphere belong to the Saxothuringian (ST), Tepla-Barrandian (TB) and Moldanubian (MD) microplates assembled during the Variscan orogeny. Dipping fossil (pre-assembly) olivine orientations, consistent within each unit, do not support any voluminous mantle delamination. The variable rift structure and morphology depend on the character of the pre-rift suture between the northern ST unit and the TB/MD units in the southern rift flank. The proper rift with typical graben morphology has developed above the steep lithosphere-scale suture between the ST and TB units. This subduction-related boundary originated from the closure of the ST Ocean. Parts of the crust and mantle lithosphere were dragged there into asthenospheric depths and then rapidly uplifted. The suture is marked by abrupt change in the mantle fabric and sharp gradients in regional gravity field and in metamorphic grade. The secular TB-side-down normal movement is reflected in deep sedimentary basins, which developed since the Carboniferous to Cenozoic and in topography. The graben morphology of the ER terminates above the “triple junction” of the ST, TB and MD mantle lithospheres. The junction is characterized by offsets of surface boundaries of the tectonic units from their mantle counterparts indicating a detachment of the rigid upper crust from the mantle lithosphere. The southwest continuation of the rift features in Bavaria is expressed in occurrences of Cenozoic sediments and volcanics above an inclined broad transition zone between the ST and MD lithospheres. Schematic scenario of evolution of the region consists mainly of a subduction of the ST lithosphere to depths around 140 km, exhumation of HP-HT rocks and the post-tectonic granitoid plutonism.

Shear-wave splitting analysis of the upper mantle at the Niigata-Kobe Tectonic Zone with the data of the Joint Seismic Observations at NKTZ

AbstractWe conducted seismic observations with a spatially high density seismic network at the Niigata-Kobe Tectonic Zone, central Japan. The seismic network was used for the analysis of shear-wave splitting. Large lateral variations were found in the polarization direction data: the northern part of the research area yields polarization directions ofNW-SE (Region A), the central part of the research area with the polarization direction of NNE-SSW (Region B), the eastern part of the research area with the polarization direction of NE-SW (Region C), and the southern part of the research area with the polarization direction of E-W (Region D). The polarization directions in Regions B and C could be explained by the preferred orientation of olivine caused by the flow of the subducting Philippine Sea plate. However, the cause of anisotropic region which was related to the heterogeneous structure was also plausible. The polarization direction in Region A might be related to the flow caused by both of the subducting Philippine Sea and Pacific slabs. The polarization direction at the Region D could not be produced by the flow in the wedge and might be related to an anisotropic region beneath the slab. The lateral variation of the polarization direction does not support a model that the NKTZ is a collision region.

Upper mantle seismic anisotropy resulting from pressure-induced slip transition in olivine

Seismic anisotropy in the oceanic lithosphere results from flow-induced crystallographic preferred orientation of dry olivine during lithosphere creation. Recent experiments, however, showed that high water activity changes the flow mechanisms of olivine and hence the crystallographic preferred orientation, better explaining the seismic anisotropy in the mantle wedge above subduction zones. Whether changes in the crystallographic preferred orientation of olivine are unique to the effects of water has become controversial and is critical to resolve. Here we report low-stress, high-strain experiments on typical dry mantle rock at high pressures and temperatures, showing that at ∼3 GPa, pressure induces the same profound transition in olivine crystallographic preferred orientation that is produced by high water activity at lower pressure. One important consequence for global tectonics is that alignment of fast seismic waves parallel to trenches beneath subducting slabs probably reflects entrainment of sub-lithospheric mantle in the direction of subduction, rather than trench-parallel flow as interpreted at present. From the large variety of crystallographic preferred orientations in olivine in both experiments and natural rocks, we infer that in addition to the pressure-induced changes in olivine slip systems implied here, there are probably further changes in slip systems at higher pressure and temperature. Seismic anisotropy in Earth’s oceanic lithosphere and in the mantle wedge above subduction zones is associated with crystallographic preferred orientations of olivine. Experiments at high pressure and temperature suggest that a pressure of ∼3 GPa can induce the same changes in the crystal structure of olivine as high water activity at lower pressures.

Relationship between Raman spectral pattern and crystallographic orientation of a rock‐forming mineral: a case study of Fo89Fa11 olivine

AbstractThe relationship between Raman spectra and crystallographic orientation was examined for single crystals of Fo89Fa11 olivine [(Mg0.89Fe0.11)2SiO4]. Raman spectra were obtained for chemically homogeneous olivine grains with various orientations on a thin section of mantle‐derived rock (dunite) using micro‐Raman equipment and an unpolarized exciting laser. Crystallographic orientations of each olivine grain were determined using an electron backscattered diffraction (EBSD) method. Five apparent peaks at 822 (peak 1), 854 (peak 2), 880 (peak 3), 917 (peak 4), and 959 cm−1 (peak 5) were observed in the spectral range of 700–1050 cm−1. Intensity ratios of peak i to peak 2, Ii/I2, for i = 1, 4, and 5 were formulated empirically as functions of crystallographic orientations: Ii/I2 = (a1iθ2/π2 + a2iθ/π + a3i)sin2ϕ + (b1iθ2/π2 + b2iθ/π + b3i)sinϕ + ci where a1i, a2i, a3i, b1i, b2i, b3i, ci are constants. ϕ is the angle between the [100] axis and the incident direction of the laser, and θ is the angle between the [001] axis and the incident direction of laser projected on the {100} plane. These equations well describe the relationships between Ii/I2 and crystallographic orientation. The obtained empirical equations enable Raman spectroscopic determination of the crystallographic orientation of olivine. Copyright © 2008 John Wiley & Sons, Ltd.

Geodynamic Significance of Seismic Anisotropy of the Upper Mantle: New Insights from Laboratory Studies

Seismic anisotropy is caused mainly by the lattice-preferred orientation of anisotropic minerals. Major breakthroughs have occurred in the study of lattice-preferred orientation in olivine during the past ∼10 years through large-strain, shear deformation experiments at high pressures. The role of water as well as stress, temperature, pressure, and partial melting has been addressed. The influence of water is large, and new results require major modifications to the geodynamic interpretation of seismic anisotropy in tectonically active regions such as subduction zones, asthenosphere, and plumes. The main effect of partial melting on deformation fabrics is through the redistribution of water, not through a change in deformation geometry. A combination of new experimental results with seismological observations provides new insights into the distribution of water associated with plume-asthenosphere interactions, formation of the oceanic lithosphere, and subduction. However, large uncertainties remain regarding the role of pressure and the deformation fabrics at low stress conditions.

Evolution of olivine lattice preferred orientation during simple shear in the mantle

Understanding the variation of olivine lattice preferred orientation (LPO) as a function of shear strain is important for models that relate seismic anisotropy to the kinematics of deformation. We present results on the evolution of olivine orientation as a function of shear strain in samples from a shear zone in the Josephine Peridotite (southwest Oregon). We find that the LPO in harzburgites re-orients from a pre-existing LPO outside the shear zone to a new LPO with the olivine [100] maximum aligned sub-parallel to the shear direction between 168% and 258% shear strain. The strain at which [100] aligns with the shear plane is slightly higher than that observed in experimental samples, which do not have an initial LPO. While our observations broadly agree with the experimental observations, our results suggest that a pre-existing LPO influences the strain necessary for LPO alignment with the shear direction. In addition, olivine re-alignment appears to be dominated by slip on both (010)[100] and (001)[100], due to the orientation of the pre-existing LPO. Fabric strengths, quantified using both the J- and M-indices, do not increase with increasing shear strain. Unlike experimental observations, our natural samples do not have a secondary LPO peak. The lack of a secondary peak suggests that subgrain rotation recrystallization dominates over grain boundary migration during fabric re-alignment. Harzburgites exhibit girdle patterns among [010] and [001] axes, while a dunite has point maxima. Combined with the observation that harzburgites are finer grained than dunites, we speculate that additional phases (i.e., pyroxenes) limit olivine grain growth and promote grain boundary sliding. Grain boundary sliding may relax the requirement for slip on the hardest olivine system, enhancing activation of the two easiest olivine slip systems, resulting in the [010] and [001] girdle patterns. Overall, our results provide an improved framework for calibration of LPO evolution models.

Converted waves reveal a thick and layered tectosphere beneath the Kalahari super-craton

Thick and high-velocity roots are generally observed beneath the Archean cratons. Inside these high-velocity keels, velocity contrasts are detected neither by surface nor by body waves tomographies. We present here evidences based on the S-to-P and P-to-S converted waves for the existence of an irregularly stratified and thick keel beneath the Kalahari super-craton. To satisfy surface wave data and S-to-P conversions, the velocity model should have beneath the Moho a ∼ 160 km thick anisotropic structure with vertical slow axis and decreasing anisotropic parameters with depth. Such a structure may stem from the preferred orientation of olivine [100] axis in the horizontal plane under shearing deformation. This structure, together with the ∼ 100 km thick layer underlying it, forms a ∼ 300 km thick continental root beneath the cratons. Inside this root, the P and S velocities should be higher on average respectively by an amount of 6% and 4% than the AK135 velocity model. Beneath ∼ 300–350 km depth, a low velocity zone is clearly detected that may be either the remainder of large magma reservoirs related to cratonic flood basalts or a melted silicate layer created by the transformation, just above the 410-km deep discontinuity, of wadsleyite assembly to olivine assembly.

Continuation of the San Andreas fault system into the upper mantle: Evidence from spinel peridotite xenoliths in the Coyote Lake basalt, central California

The Coyote Lake basalt, located near the intersection of the Hayward and Calaveras faults in central California, contains spinel peridotite xenoliths from the mantle beneath the San Andreas fault system. Six upper mantle xenoliths were studied in detail by a combination of petrologic techniques. Temperature estimates, obtained from three two-pyroxene geothermometers and the Al-in-orthopyroxene geothermometer, indicate that the xenoliths equilibrated at 970–1100 °C. A thermal model was used to estimate the corresponding depth of equilibration for these xenoliths, resulting in depths between 38 and 43 km. The lattice preferred orientation of olivine measured in five of the xenolith samples show strong point distributions of olivine crystallographic axes suggesting that fabrics formed under high-temperature conditions. Calculated seismic anisotropy values indicate an average shear wave anisotropy of 6%, higher than the anisotropy calculated from xenoliths from other tectonic environments. Using this value, the anisotropic layer responsible for fault-parallel shear wave splitting in central California is less than 100 km thick. The strong fabric preserved in the xenoliths suggests that a mantle shear zone exists below the Calaveras fault to a depth of at least 40 km, and combining xenolith petrofabrics with shear wave splitting studies helps distinguish between different models for deformation at depth beneath the San Andrea fault system.

Deformation and melt transport in a highly depleted peridotite massif from the Canadian Cordillera: Implications to seismic anisotropy above subduction zones

Seismic anisotropy in subduction zones results from a combination of various processes. Although it depends primarily on the orientation of olivine in response to flow, the presence of water and melt in the wedge may modify the deformation of olivine. The melt distribution also influences anisotropy. Direct observations of the deformation and melt-rock interactions in a strongly depleted spinel-harzburgite massif from the Cache Creek terrane in the Canadian Cordillera allow evaluating the relative contribution of each process. Structural mapping shows that this massif has recorded high-temperature, low-stress deformation, high degrees of partial melting, and synkinematic melt-rock interaction at shallow depths (< 70 km) in the mantle, probably above an oblique subduction. Deformation, marked by shallow-dipping lineations and steep foliations, controlled melt distribution: reactive dunites and pyroxenite dykes are dominantly parallel to the foliation. Analysis of olivine crystal preferred orientations (CPO) indicates deformation by dislocation creep with dominant [100] glide. Glide planes are however different in harzburgites and dunites, suggesting that higher melt contents may favor glide on (001) relative to (010). Seismic properties, calculated by considering explicitly the large-scale structure of the massif, the olivine and pyroxene CPO, and possible melt distributions, show that the strain-induced olivine CPO results in up to 5% P- and S-wave anisotropy with fast seismic directions parallel to the lineation. Synkinematic melt transport by diffuse porous flow leading to melt pockets or dykes aligned in the foliation may significantly enhance this anisotropy, in particular for S-waves. In contrast, focused melt flow is not recorded by seismic anisotropy, unless associated with very high instantaneous melt fractions. Orientation of pyroxenite dykes suggests that the present orientation of the structures is representative of the pre-obduction situation, implying trench-parallel fast polarizations and high delay times as observed above the Kurils, Ryukyu, Taiwan, and Tonga subductions.

Electrical structure beneath the northern MELT line on the East Pacific Rise at 15°45′S

The electrical structure of the upper mantle beneath the East Pacific Rise (EPR) at 15°45′S is imaged by inverting seafloor magnetotelluric data obtained during the Mantle ELectromagnetic and Tomography (MELT) experiment. The electrical conductivity model shows no evidence for a conductive region immediately beneath the ridge, in contrast to the model previously obtained beneath the EPR at 17°S. This observation can be explained by differences in current melt production along the ridge, consistent with other observations. The mantle to the east of the ridge at 60 –100 km depth is anisotropic, with higher conductivity in the spreading direction compared to the along‐strike direction, similar to the 17°S region. The high conductivity in the spreading direction can be explained by a hydrated mantle with strain‐induced lattice preferred orientation of olivine or by partial melt preferentially connected in the spreading direction.

Midmantle deformation between the Australian continent and the Fiji‐Tonga subduction zone?

Shear wave splitting measurements are performed at temporary stations deployed across the Australian continent for direct S waves from events occurring in the neighboring subduction zones, especially Fiji‐Tonga. Unlike core refracted shear waves, direct S waves are not polarized at the core‐mantle boundary and are therefore subject to influences on the source side, on the receiver side, and along the ray path in between. Source side contamination due to the development of lattice preferred orientation of olivine can be minimized by considering deep events with an epicentral location below 410 km depth, the olivine‐spinel phase transition depth. Anisotropic contributions from the transition zone or lower mantle can only be reliably inferred in the case of high delays that cannot be generated solely by the upper mantle structure underneath a station. The analysis of a comprehensive data set for direct S waves from deep events leads to an average delay time of 0.9 s at stations deployed across the continent. Despite the limitations of such an arithmetic average, the result highlights the lack of evidence for midmantle deformation between the Fiji‐Tonga subduction zone and the Australian continent, unlike previously suggested in studies dealing with the permanent seismological stations only. A striking consistency between measurements performed across several stations belonging to an array in western Australia suggests a very coherent lithospheric structure underneath the Yilgarn craton.

European mantle lithosphere assembled from rigid microplates with inherited seismic anisotropy

We have modelled three-dimensional seismic anisotropy of the mantle lithosphere from anisotropic parameters of teleseismic body waves. We invert jointly shear-wave splitting parameters and P residual spheres based on data from dense networks of temporary and permanent stations in four European regions ranging from the Variscan belt to the Baltic Shield. Changes in orientation of the large-scale anisotropy, caused by systematic preferred orientation of olivine, identify boundaries of domains of mantle lithosphere. Individual domains are characterized by a consistent large-scale orientation of anisotropy approximated by hexagonal or orthorhombic symmetry with generally inclined symmetry axes. The domains are separated by mapped tectonic boundaries (sutures), which cut the entire lithosphere. Besides the change of anisotropy orientation at domain boundaries, we often observe a change of the lithosphere and/or crust thicknesses. We do not detect any fabric of the mantle lithosphere, which could have been produced by a collision of microcontinents in a volume detectable by large-scale seismic anisotropy. The observations of consistent anisotropy within individual blocks of the mantle lithosphere reflect frozen-in olivine preferred orientation, most probably formed prior to the assembly of microcontinents that created the modern European landmass. Therefore, our findings support a plate tectonic view of the continental lithosphere as a mosaic of rigid blocks of the mantle lithosphere with complicated but relatively sharp contact zones. These contacts are blurred by the easily deformed overlying crust terranes.

Flow and electrical anisotropy in the upper mantle: Finite-element models constraints on the effects of olivine crystal preferred orientation and microstructure

Long-period magnetotelluric (MT) data shows that electrical conductivity in the upper mantle is highly anisotropic. Agreement between high electrical conductivity directions and seismic anisotropy fast directions suggests that anisotropic diffusion of hydrogen along oriented olivine crystals controls the anisotropy of electrical conductivity in the upper mantle, since both seismic waves and hydrogen diffusion are faster along the [1 0 0] axis of olivine. Thus, MT electrical anisotropy data, like seismic anisotropy, may be used to map flow patterns in the upper mantle. However, observed electrical anisotropies are significantly higher than seismic ones. To quantify the influence of strain-induced crystal preferred orientations of olivine on upper mantle bulk electrical conductivities, we calculate the macroscopic electrical conductivity anisotropy of a series of naturally and experimentally deformed peridotites using an anisotropic finite-element model. These models, which fully take into account the microstructure: crystal and shape preferred orientations, based on orientation maps obtained by indexation of electron back-scattered diffraction (EBSD) patterns, show macroscopic electrical anisotropy factors ranging from 3 to 16. The intensity of electrical anisotropy depends to first order on the intensity of the olivine crystal preferred orientations, but the relation saturates for strong crystal preferred orientations. In addition, the spatial distribution of the various crystal orientations may significantly enhance influence the anisotropy in strongly textured mantle rocks. The strongest anisotropy factors (>10) occur in mantle rocks in which deformation by dislocation creep has produced not only crystal but also strong shape preferred orientations, even if the latter is masked by recrystallization. Higher anisotropy factors (>100) observed in a few MT experiments imply however an additional, as yet unknown, mechanism controlling electrical conduction at asthenospheric depths in these regions.

Effect of water and stress on the lattice-preferred orientation of olivine

The influence of water and stress on the lattice-preferred orientation (LPO) of olivine aggregates was investigated through large strain, shear deformation experiments at high pressures and temperatures ( P = 0.5–2.1 GPa, T = 1470–1570 K) under both water-poor and water-rich conditions. The specimens are hot-pressed synthetic olivine aggregates or single crystals of olivine. Water was supplied to the sample by decomposition of a mixture of talc and brucite. Deformation experiments were conducted up to γ (shear strain) ∼ 6 using the Griggs apparatus where water fugacity was up to ∼ 13 GPa at the pressure of 2 GPa. The water content in olivine saturated with water increases with increasing pressure and the solubility of water in olivine at P = 0.5–2 GPa was ∼ 400–1200 ppm H/Si. Several new types of LPO in olivine are found depending on water content and stress. Samples deformed in water-poor conditions show a conventional LPO of olivine where the olivine [100] axis is subparallel to the shear direction, the (010) plane subparallel to the shear plane (type-A). However, we identified three new types (type-B, C, and E) of LPO of olivine depending on the water content and stress. The type-B LPO of olivine which was found at relatively high stress and/or under moderate to high water content conditions is characterized by the olivine [001] axis subparallel to the shear direction, the (010) plane subparallel to the shear plane. The type-C LPO which was found at low stress and under water-rich conditions is characterized by the olivine [001] axis subparallel to the shear direction, the (100) plane subparallel to the shear plane. The type-E LPO which was found under low stress and moderate water content is characterized by the olivine [100] axis subparallel to the shear direction, the (001) plane subparallel to the shear plane. Observations by transmission electron microscopy (TEM) and scanning electron microscopy (SEM) show that the dislocations in water-poor samples (type-A) are curved and both b = [100] and b = [001] dislocations have a similar population. Numerous subgrains are seen in water-poor samples in backscattered electron images. In contrast, water-rich samples (both type-B and type-C) contain mostly b = [001] dislocations and dislocations are straight and sub-grain boundaries are rare compared to those in water-poor samples. These observations suggest that (1) dominant slip systems in olivine change with water fugacity (and stress) and (2) grain boundary migration is enhanced in the presence of water. Seismic anisotropy corresponding to the fabrics under water-rich condition is significantly different from that under water-poor condition. Consequently, the relationship between seismic anisotropy and flow geometry in water-rich regions is expected to be different from that in water-poor regions in which type-A fabric dominates (i.e., the lithosphere). A few cases are discussed including anisotropy in the subduction zone and in the deep upper mantle.

Mantle dynamics beneath the East Pacific Rise at 17°S: Insights from the Mantle Electromagnetic and Tomography (MELT) experiment

The electromagnetic data from the Mantle Electromagnetic and Tomography (MELT) experiment are inverted for a two‐dimensional transversely anisotropic conductivity structure that incorporates a correction for three‐dimensional topographic effects on the magnetotelluric responses. The model space allows for different conductivity values in the along‐strike, cross‐strike, and vertical directions, along with imposed constraints of model smoothness and closeness among the three directions. Anisotropic models provide a slightly better fit to the data for a given level of model smoothness and are more consistent with other geophysical and laboratory data. The preferred anisotropic model displays a resistive uppermost 60‐km‐thick mantle independent of plate age, except in the vicinity of the ridge crest. In most inversions, a vertically aligned sheet‐like conductor at the ridge crest is especially prominent in the vertical conductivity. Its presence suggests that the melt is more highly concentrated and connected in the vertical direction immediately beneath the rise axis. The melt zone is at least 100 km wide and is asymmetric, having a greater extent to the west. Off‐axis, and to the east of the ridge, the mantle is more conductive in the direction of plate spreading at depths greater than 60 km. The flat resistive‐conductive boundary at 60 km agrees well with the inferred depth of the dry solidus of peridotite, and the deeper conductive region is consistent with the preferred orientation of olivine inferred from seismic observations. This suggests that the uppermost 60 km represents the region of mantle that has undergone melting at the ridge and has been depleted of water (dissolved hydrogen). By contrast, the underlying mantle has retained a significant amount of water.

北海道および東北日本沈み込み帯におけるS波偏向異方性とマントルウェッジ内の二次対流

We analyzed shear-wave splitting beneath Hokkaido and Tohoku, Japan. Waveforms of 912 intermediate-depth earthquakes recorded at 239 seismic stations were used, and 2276 splitting parameters, the leading shear-wave polarization direction (fast direction) and delay time between two split waves, were observed. Obtained results show that most of fast directions observed at stations in the back-arc side are perpendicular to the trench-axis, whereas those at stations in the fore-arc side are sub-parallel to it. Average delay times observed at stations in the back-arc side are 0.1-0.5 s, and those at stations in the fore-arc side are smaller (0.05-0.1s). This systematic spatial variation is observed for the whole study area, from the southern part of Tohoku to the eastern part of Hokkaido. We infer that the anisotropy caused by lattice-preferred orientation of olivine, which is probably attributable to the upwelling flow portion of the mechanically-induced convection in the mantle wedge, is a likely candidate for the shear-wave splitting in the back-arc mantle wedge. If this is the case, the present observations for the east of the arc-arc junction may indicate that the upwelling flow direction is sub-parallel not to the relative plate motion direction but to themaximum dip-direction of the subductiog slab, in contrast with that in the south of the arc-arc junction where the upwelling flow sub-parallel to the relative plate motion is expected. The anisotropy structure revealed in this study plays a crucial role on understanding ongoing mantle dynamics in subduction zones.