lable at ScienceDirect

Journal of Structural Geology 70 (2015) 12e22

Contents lists avai

Journal of Structural Geology

journal homepage: www.elsevier .com/locate/ jsg

Lattice-preferred orientation of olivine found in diamond-bearinggarnet peridotites in Finsch, South Africa and implications for seismicanisotropy

Jaeseok Lee, Haemyeong Jung*

Tectonophysics Laboratory, School of Earth and Environmental Sciences, Seoul National University, Seoul 151-742, Republic of Korea

a r t i c l e i n f o

Article history:Received 17 June 2014Received in revised form14 October 2014Accepted 25 October 2014Available online 12 November 2014

Keywords:OlivineLattice preferred orientationPressure effectSeismic anisotropySubduction zoneFinsch

* Corresponding author. School of Earth and Envirotional University, 311 ho, 25-1 dong, 1 Gwanak-ro, Gwpublic of Korea. Tel.: þ82 2 880 6733; fax: þ82 2 871

E-mail address: hjung@snu.ac.kr (H. Jung).

http://dx.doi.org/10.1016/j.jsg.2014.10.0150191-8141/© 2014 The Authors. Published by Elsevier

a b s t r a c t

Seismic anisotropy in the upper mantle provides important constraints on mantle dynamics, continentalevolution and global tectonics and is believed to be produced by the flow-induced lattice-preferredorientation (LPO) of olivine. Recent experimental studies at high pressure and temperature have sug-gested that the LPO of olivine is affected by pressure in addition to water and stress. However, there hasbeen no report yet for the pressure-induced LPO of natural olivine because samples from the deep uppermantle are rare and often unsuitable for study due to ambiguous foliation and lineation. Here we showevidence of the pressure-induced LPO of natural olivine in diamond-bearing garnet peridotites fromFinsch, South Africa. We found that the [010] axes of olivine are aligned subnormal to foliation and thatthe [001] axes are aligned subparallel to lineation, which is known as B-type LPO of olivine. The equi-librium pressure of the samples, as estimated using geobarometer, was greater than 4 GPa, indicatingthat the samples originated from a depth greater than ~120 km. In addition, FTIR spectroscopy of theolivine showed that the samples are dry, with a water content of less than 90 ± 20 ppm H/Si(5.5 ± 1.2 ppm wt. H2O). These data suggest that the samples are the first natural examples of olivinedisplaying B-type LPOs produced due to high pressure under dry condition. Our data indicate that thetrench-parallel seismic anisotropy observed in many subduction zones in and below subducting slabs atdepths greater than ~90 km under dry condition may be attributed to the pressure-induced olivinefabrics (B-type LPO) and may be interpreted as the entrainment of the sub-lithospheric mantle in thedirection of subduction rather than anomalous trench-parallel flow.© 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

1. Introduction

Seismic anisotropy is a powerful tool for understanding theglobal tectonics of the upper mantle in the Earth (Becker et al.,2012; Fouch et al., 2000; Long, 2013; Long and Silver, 2008;Silver, 1996; Song and Kawakatsu, 2012) and has been observedin many subduction zones worldwide (Long, 2013; Park and Levin,2002; Russo and Silver, 1994; Savage, 1999; Wang and Zhao, 2013).Trench-parallel seismic anisotropy has been observed in themantlewedge above subducting slabs (Long, 2013; Long and Silver, 2008;Smith et al., 2001), as well as below subducting slabs at a deeperportion of the upper mantle (Long and Silver, 2008, 2009; Russo

nmental Sciences, Seoul Na-anak-gu, Seoul 151-742, Re-3269.

Ltd. This is an open access article u

and Silver, 1994; Tian and Zhao, 2012). Proposed mechanisms forthe source of this trench-parallel seismic anisotropy include water-induced B-type LPO of olivine in the mantle wedge (Jung, 2009;Jung and Karato, 2001a; Karato et al., 2008; Katayama and Karato,2006; Kneller et al., 2008; Mizukami et al., 2004), LPO of serpen-tine in serpentinite altered from peridotite (Ji et al., 2013; Jung,2011; Katayama et al., 2009; Soda and Wenk, 2014; Watanabeet al., 2011), trench-parallel mantle flow due to slab roll back(Long and Silver, 2008, 2009; Russo and Silver, 1994), rapid toroidalflowaround slab edge (Jadamec and Billen, 2010), pressure-inducedB-type LPO of olivine due to slip transition at high pressure greaterthan P¼ 3 GPa (Jung et al., 2009; Ohuchi et al., 2011; Raterron et al.,2011), aligned faults by hydration in subducting oceanic plate(Faccenda et al., 2008) and an effective orthorhombic symmetry forthe oceanic asthenosphere, which is translated to the depthbeneath the subducting slab (Song and Kawakatsu, 2012). However,the origin of seismic anisotropy remains poorly understood.

nder the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

J. Lee, H. Jung / Journal of Structural Geology 70 (2015) 12e22 13

Because olivine is the dominant mineral in the upper mantleand is elastically anisotropic (Abramson et al., 1997; Kumazawa andAnderson, 1969), seismic anisotropy may be attributable to the LPOof olivine (Ben Ismail and Mainprice, 1998; Long and Silver, 2008;Mainprice, 2007; Nicolas and Christensen, 1987). Previous studieshave shown that the LPO of olivine is affected by the physical andchemical conditions of deformation, including water, stress, andtemperature (Jung and Karato, 2001a; Jung et al., 2006; Karatoet al., 2008; Katayama and Karato, 2006). Various types of LPOshave been observed in olivine depending on water content (COH)and stress at pressures less than 2.3 GPa (Jung et al., 2006). Both A-and D-type LPOs were found under dry conditions(COH < 200 ppm H/Si), while B- C-, and E-type LPOs were foundunder wet conditions. Among these LPOs, the B-type is particularlyimportant to understand the trench-parallel seismic anisotropy ofthe mantle wedge, and it is characterized by the alignment of the[001] axes subparallel to the shear direction and of the [010] axessubnormal to the shear plane. Additionally, recent high pressureexperiments on olivine aggregates under dry conditions haverevealed that a change in the LPO of olivine fromA-type to B-type isinduced by high pressure greater than P ¼ 3 GPa (Jung et al., 2009)and P ¼ 5 GPa (Ohuchi et al., 2011). Although numerous studieshave been conducted on the LPO of natural olivine at shallowdepths (P � 2 GPa) (Jung, 2009; Jung et al., 2014; Kim and Jung,2014; Mizukami et al., 2004; Palasse et al., 2012; Park et al., inpress; Park and Jung, 2014; Skemer et al., 2010; Tasaka et al.,2008; Tommasi et al., 2008; Warren et al., 2008), studies on theLPO of olivine from the deep upper mantle (P > 3 GPa) have beenvery limited (Baptiste et al., 2012; Skemer and Karato, 2008; Wanget al., 2013a; Xu et al., 2006). In the present study, we demonstratethat pressure-induced B-type LPO of olivine does occur at highpressure in natural rocks, a finding with significant implications forseismic anisotropy and global tectonics.

2. Sample descriptions

We studied garnet peridotites from Finsch, South Africa, whichis located at the western margin of the Kimberley block (Fig. 1)(Gibson et al., 2008). The garnet peridotites displayed a

Fig. 1. Geologic map of the study area, Finsch in So

porphyroclastic texture and are strongly foliated (Fig. 2a), with acomposition of primarily olivine with minor amounts of elongatedenstatite and garnet (Table 1, Fig. 2a). Samples showed composi-tional layering of olivine-rich and orthopyroxene-rich layer. Thethickness of olivine-rich and orthopyroxene-rich layer was~10e32 mm and ~6e21 mm, respectively. Olivine-rich layer con-sists of olivine ~70e90 % while orthopyroxene-rich layer consists oforthopyroxene ~40e60 %. The foliation of the rock specimens wasdetermined by the compositional layering of orthopyroxene andolivine with the variable ratios. Lineation(X) was determined by theshape preferred orientation of all grains on foliation (Panozzo,1984). Thin sections were made in 3 orthogonal planes (XZ, XY,and YZ) where Z is the direction normal to foliation and Y is thedirection normal to both X and Z, and the grain aspect ratios inthese 3 planes were determined and shown in Table 1. The aspectratio of all grains in the XZ plane was the largest. The aspect ratiosof olivine and orthopyroxene were also determined separately inspecimens and are shown in Table 2. It is found in general thatorthopyroxene has bigger aspect ratio than olivine.

3. Methods

3.1. Determination of the LPO of olivine and calculation of seismicanisotropy

The samples were cut parallel to their lineation and perpen-dicular to their foliation for the microstructure analysis. The LPO ofolivine was measured by electron back-scattered diffraction (EBSD;using the HKL system with Channel 5 software) at an acceleratingvoltage of 20 kV and a working distance of 15 mm using thescanning electron microscope JSM6308 of the School of Earth andEnvironmental Sciences (SEES) at Seoul National University (SNU),Korea. The EBSD pattern was indexed manually at each grain toobtain an accurate solution. We measured the LPO of ~230e700grains of olivine for each sample. The misorientation index (M-in-dex) (Skemer et al., 2005) was calculated to estimate the fabricstrength of the sample using the uncorrelated grain pairs deter-mined from the EBSD data. The seismic velocity and seismicanisotropy were calculated from the orientation data for olivine

uth Africa. Modified after Gibson et al. (2008).

Fig. 2. Diamond-bearing garnet peridotite in Finsch. (a) Photograph of sample 1512 showing the grain-shape preferred orientation of orthopyroxene and garnet. (b) Backscatteredelectron image showing a micro-diamond inside the garnet of sample 1512. (c) Microphotograph (reflected light) showing a large diamond (~3 mm) in sample 1517. (d) & (e) Ramanspectra of the diamonds shown in Fig. 2b and c, respectively. Ol: olivine, Grt: garnet, and Dia: diamond.

J. Lee, H. Jung / Journal of Structural Geology 70 (2015) 12e2214

using the elastic constant for olivine (Abramson et al., 1997) and asoftware program (Mainprice, 1990). Seismic anisotropy of wholerock including minor phases such as enstatite (orthopyroxene) andgarnet was also calculated.

3.2. Measurement of the mineral water content

The water contents of olivine and orthopyroxene weremeasured on doubly polished sections (~200 mm thick) withoutincluding the grain boundaries using a Nicolet Fourier-transforminfrared (FTIR) 6700 spectrometer with a continuum infrared mi-croscope at the Tectonophysics laboratory of the SEES. Each samplewas dried at 120 �C for over 24 h prior to the FTIR analysis. N2 gaswas flushed through both the infrared chamber and themicroscopeto remove atmospheric moisture. The IR beam size of50 mm� 50 mmwas used. Although new calibrations for calculatingwater content of olivine exist (Kov�acs et al., 2008; Withers et al.,2012), we used the unpolarized FTIR spectrum and Paterson's

Table 1Sample description and results.

Sample no. Olivine(%)

OPX(%)

Garnet(%)

LPO ofolivine

M-indexa Pressure(GPa)b

Temp.(�C)b

W(

1512 81 12 7 B-type 0.296 4.2 1030 61516 95 4 1 B-type 0.193 4.3 1010 51517 91 5 4 B-type 0.162 4.5 1000 9

N.D. denotes “not detected”. OPX: orthopyroxene (enstatite), LPO: lattice preferred oriena The M-index represents the fabric strength of olivine (Skemer et al., 2005).b Pressure and temperature were estimated using geothermobarometers (see Meth

geothermobarometer.c Water content was calculated using Paterson's calibration (Paterson, 1982).d The aspect ratios of all grains were determined in 3 dimensions (XY, XZ, and YZ plane

parallel to lineation, Z: normal to foliation, and Y: perpendicular to X and Z.

calibration (Paterson, 1982) for consistency with previous studieson the LPO of olivine (Jung and Karato, 2001a; Jung et al., 2009;Katayama et al., 2004). The same data using the Bell et al. (2003)calibration would yield a value ~3.5 times higher (Mosenfelderet al., 2006).

3.3. Estimation of pressure and temperature

The equilibrium pressure of the samples was estimated bygeobarometric method using the oxide mass percentages for Al inOpx (Brey and Kohler, 1990). Temperature was estimated using theFeeMg exchange between olivine and garnet (O'Neill and Wood,1979). A Jeol JXA-8900 electron probe micro-analyzer (EPMA) atthe National Center for Inter-university Research Facilities at theSNU was employed to measure the oxide mass percentages ofolivine, orthopyroxene, and garnet (Table 3), using an accelerationvoltage of 15 kV and a beam size of 5 mm. A total of 3 grains (coreand rim measurements) were analyzed for each specimen.

ater in olivineppm H/Si)c

Water in OPX(ppm H/Si)c

Aspect ratiod

(XYplane)Aspect ratiod

(XZplane)Aspect ratiod

(YZplane)

0 ± 15 N.D. 1.104 1.331 1.1450 ± 15 120 ± 30 1.054 1.309 1.1980 ± 20 350 ± 40 1.077 1.523 1.35

tation.

ods). Uncertainties of pressure and temperature are 0.5 GPa and 40 �C for the

) by drawing the grain-shape of up to 400 individual grains in thin sections where X:

Table 2Aspect ratio of olivine and orthopyroxene (Opx) in 3 orthogonal planes.

Sample no. Plane Aspect ratio of olivine Aspect ratio of Opx

1512 XY 1.099 1.267XZ 1.338 1.482YZ 1.127 1.463

1516 XY 1.066 1.069XZ 1.312 1.484YZ 1.184 1.297

1517 XY 1.055 1.21XZ 1.556 1.489YZ 1.359 1.479

X: parallel to lineation, Z: normal to foliation, and Y: normal to both X and Z.

J. Lee, H. Jung / Journal of Structural Geology 70 (2015) 12e22 15

3.4. Identification of diamond in the specimen

A dispersive confocal DXR Raman microscope (Thermo Scienti-fic) housed at the Tectonophysics Laboratory of the SEES in SNU,Korea was used to identify diamond in the specimen. The Ramanmicroscope was equipped with a 532 nm laser (10 mW power) andan optical microscope (Olympus, 50� objective) with an automaticstage and had a resolution of 0.01 cm�1 over the wavenumberrange of 50e3550 cm�1 with a beam size of 0.67 mm, which is smallenough to analyze small inclusions. The Raman spectrum was ob-tained using a 32-s exposure time.

4. Results

4.1. Microstructures of samples

Microstructures of a typical olivine-rich layer (sample 1512) areshown in optical microphotograph (Fig. 3a). Porphyroclasts oforthopyroxene and garnet are observed. Porphyroclasts of ortho-pyroxene has aspect ratio up to 3.5:1 (Figs. 2a and 3a). Big olivinegrains are also elongated subparallel to lineation with aspect ratioup to 4:1. Garnets are less elongated with the size of 2e5 mm

Table 3Representative compositions of olivine, orthopyroxene, and garnet for each sample.

Sample 1512 1516

Olivine Opx Garnet Olivine

Oxide mass percentage (wt%)SiO2 40.58 56.43 42.22 40.25TiO2 0.02 0.07 0.44 0.02Al2O3 0.02 0.70 20.83 0.00Cr2O3 0.03 0.33 3.88 0.08FeO 7.06 4.35 6.58 7.11MnO 0.09 0.09 0.27 0.07MgO 51.14 36.24 22.22 53.45NiO 0.63 0.09 0.05 0.40CaO 0.01 0.61 4.71 0.00Na2O 0.03 0.20 0.05 0.01K2O 0.00 0.00 0.00 0.03Total 99.59 99.12 101.25 101.42Cation numbers in P.F.U.Si 0.99 1.95 2.97 0.97Ti 0.00 0.00 0.02 0.00Al 0.00 0.03 1.73 0.00Cr 0.00 0.01 0.22 0.00Fe 0.14 0.13 0.39 0.14Mn 0.00 0.00 0.02 0.00Mg 1.86 1.87 2.33 1.91Ni 0.01 0.00 0.00 0.01Ca 0.00 0.02 0.36 0.00Na 0.00 0.01 0.01 0.00K 0.00 0.00 0.00 0.00Total 3.01 4.03 8.04 3.03

(Fig. 3a), typically having a kelyphite rim with the thickness of~400 mm. We plotted the Flinn diagram in Fig. 4 using the aspectratios of all grains (Fig. 4a) and of each mineral (Fig. 4b) in XY andYZ plane to see the dominant shape of grains (Flinn, 1962). Allsamples were plotted under the k ¼ 1 line, showing that the shapeof grains are flattened rather than constricted. We found thatsamples contained diamonds as inclusions in garnet with the sizeof ~1 mm (Fig. 2b and c). Big diamond with the size of ~3 mm wasalso found in between grain boundaries of olivine (Fig. 2c). Therepresentative peak of diamond is shown in the micro-Ramanspectra (Fig. 2d and e). The existence of diamond in the samplesconfirms the deep origin of the specimens. Undulose extinctionsare frequently observed in olivine grains (Fig. 3b). To observedislocation microstructures of olivine, samples were decoratedwith oxygen in the air at 800 �C for 1 h (Jung and Karato, 2001b;Karato, 1987; Kohlstedt et al., 1976). Backscattered electron imageof sample showed that there are many free dislocations in olivine(Fig. 3c). It is also found that subgrain boundaries are well devel-oped in olivine (Fig. 3c).

4.2. Lattice preferred orientation of olivine

The LPO of olivine from the analysis of the electron back-scattered diffraction (EBSD) using a scanning electronmicroscope isplottedwith pole figures (Fig. 5). All of the samples showed a strongLPO of olivine. The [001] axes of olivine are aligned subparallel tothe lineation, and the [010] axes are aligned subnormal to thefoliation, which is known as the B-type LPO of olivine (Jung andKarato, 2001a). LPO of orthopyroxene is not reported because thenumber of grains of orthopyroxene was too small.

4.3. Water content of olivine and orthopyroxene

Because previous experimental studies had indicated that the B-type LPO of olivine could be formed by either water (>200 ppmH/Si) (Jung and Karato, 2001a) or high pressure (>3 GPa) under dry

1517

Opx Garnet Olivine Opx Garnet

57.60 41.77 39.81 57.30 41.290.00 0.25 0.00 0.02 0.040.64 18.25 0.00 0.51 18.580.38 6.65 0.04 0.28 6.244.37 6.07 7.86 4.68 6.820.06 0.29 0.06 0.13 0.29

37.42 20.65 52.21 37.17 19.940.19 0.02 0.56 0.14 0.040.62 6.35 0.03 0.75 6.210.14 0.00 0.02 0.07 0.040.02 0.01 0.00 0.02 0.00

101.45 100.31 100.59 101.05 99.49

1.95 3.00 0.97 1.95 3.000.00 0.01 0.00 0.00 0.000.03 1.55 0.00 0.02 1.590.01 0.38 0.00 0.01 0.360.12 0.36 0.16 0.13 0.410.00 0.02 0.00 0.00 0.021.89 2.21 1.89 1.89 2.160.01 0.00 0.01 0.00 0.000.02 0.49 0.00 0.03 0.480.01 0.00 0.00 0.00 0.010.00 0.00 0.00 0.00 0.004.04 8.02 3.03 4.04 8.03

Fig. 3. (a) Microphotograph (cross polarized light) of a thin section in XeZ direction (sample 1512). Porphyroclasts of garnet and orthopyroxene are shown. Olivine grains (Ol) areelongated subparallel to the lineation (X: EeWdirection). Diamond inclusion with the size of ~1 mm (Fig. 2b) was found inside the white square in garnet (indicated by white arrow)(b) Magnified view of the red rectangle in Fig. 3a. Undulose extinctions in olivine are indicated by white arrows. (c) Backscattered electron (BSE) image of olivine (sample 1512).Dislocations are shown as bright lines and dots. Free dislocations are visible. Subgrain boundaries are indicated by white arrows. The BSE image was taken at the acceleration voltageof 15 kV and working distance of 10 mm, and spot size of 60. Grt: garnet and Opx: orthopyroxene. (For interpretation of the references to colour in this figure legend, the reader isreferred to the web version of this article.)

J. Lee, H. Jung / Journal of Structural Geology 70 (2015) 12e2216

conditions (Jung et al., 2009; Ohuchi et al., 2011; Raterron et al.,2011), the water content of the olivine in each sample was deter-mined using FTIR spectroscopy (Fig. 6). No differences in watercontent were observed between the core and the rim of the olivinesamples. Seven grains of olivine were analyzed from each sample,and the average water content is shown in Table 1. The watercontent of the individual olivine samples ranged from 50 to120 ± 20 ppm H/Si (3e7.4 ± 1.2 ppm wt. H2O).

The calculation of orthopyroxene/olivine partition coefficientsof water can provide additional evidence that little water was lostfrom the olivine during uplift. Previous studies at high pressure andhigh temperature have established that the partition coefficient ofwater in orthopyroxene/olivine is DOpx/ol ¼ 10 ± 5, a value repre-senting the average of 12 experiments (Aubaud et al., 2004; Hauriet al., 2006). We measured the water content of orthopyroxeneusing a Nicolet 6700 FTIR spectrometer and found that orthopyr-oxene as well contained a small amount of water(115e380 ± 40 ppm H/Si). Our specimens displayed partition co-efficients of water of DOpx/ol ¼ 6 ± 3, which are in agreement withprevious results. No differences in water content were detectedbetween the core and the rim for orthopyroxene (Fig. 6). All of the

FTIR spectroscopy data showed that the olivine in our specimenswas originally dry.

4.4. Estimation of pressure and temperature of specimen

Because pressure is also an important factor for the LPO ofolivine (Couvy et al., 2004; Jung et al., 2009; Mainprice et al., 2005;Ohuchi et al., 2011) and the B-type LPO of olivine can be formed dueto high pressure (over 3 GPa), even under dry conditions (Jung et al.,2009; Ohuchi et al., 2011; Raterron et al., 2011), the pressure-temperature conditions of the samples were estimated using geo-thermobarometers (see methods). The samples were equilibratedat high pressure (4.3 ± 0.3 GPa) and a temperature of ~1050 �C(Table 1).

4.5. Seismic anisotropy

The LPOof olivineplays an important role in the seismic anisotropyof the upper mantle (Ben Ismail and Mainprice, 1998; Karato et al.,2008; Long, 2013; Long and Silver, 2008; Nicolas and Christensen,1987). The seismic anisotropy corresponding to the LPO of olivine

Fig. 4. Flinn diagram plotted using the aspect ratio of grains in XY and YZ plane to seethe dominant shape of grains. RXY and RYZ are the aspect ratio of (a) all grains and (b)each mineral in XY and YZ plane, respectively. Grain shape is in the field of constrictionfor k > 1, while it is in the field of flattening for k < 1. The k-value is defined ask ¼ (RXY e 1)/(RYZ e 1) (Flinn, 1962). X: lineation, Z: direction normal to foliation, andY: direction normal to both X and Z.

J. Lee, H. Jung / Journal of Structural Geology 70 (2015) 12e22 17

(Fig. 5) was calculated at P ¼ 4 GPa and T ¼ 1000 �C. The elastic stiff-nesses (Cij) of the VogiteReusseHill averagewere calculated using theLPO of olivine, and the density (r ¼ 3.237 g/cm3), and the Cij of singlecrystal olivine (Abramson et al., 1997). The elastic stiffnesses (Cij) of arepresentative olivine and whole rock (sample 1512) are shown inTable 4. Fig. 7 shows the seismic anisotropy of the P-wave (Vp), theS-wave velocity anisotropy (AVs) and the polarization direction of thefast shear wave (Vs1). The P-wave velocity was in the range of7.86e9.04km/s. TheP-waveanisotropy (11.0e13.9%)wassimilar to theS-wave anisotropy (10.45e14.24%). All of the samples showed that thepolarization directions of the fast shear waves (Vs1) were alignednearly normal to the lineation (flow direction) for the verticallypropagating S-waves (Fig. 7). The polarization directions of the fastS-waves were similar to those observed for the pressure-induced B-type LPOsof olivine inprevious experimental studies (Junget al., 2009;Ohuchi et al., 2011).

The seismic anisotropy of whole rock including minor phasessuch as orthopyroxene and garnet was also calculated. In additionto the LPO of olivine, we used the LPOs of orthopyroxene andgarnet, and the density of orthopyroxene (r ¼ 3.225 g/cm3) andgarnet (r¼ 3.602 g/cm3), and the Cij of single crystal orthopyroxene(Chai et al., 1997) and garnet (Bass, 1989) (see Fig. 8). We used theVogiteReusseHill average to calculate the seismic anisotropy forthe whole rocks. It is found that the contribution of these minorphases to the seismic anisotropy of whole rock was small. Forexample, P-wave velocity of the sample 1512was slightly decreasedto 0.12 km/s, and the P-wave anisotropy was decreased to 1.5% dueto the existence of orthopyroxene and garnet. S-wave anisotropywas also slightly decreased to 1.67%. However, the polarizationdirection of the fast shear waves (Vs1) was not changed comparedto that of olivine only (Fig. 7).

4.6. Estimation of stress in the specimen

The relationship between recrystallized grain size and stress ofolivine (Jung and Karato, 2001b; Karato et al., 1980; Van der Walet al., 1993; Zhang et al., 2000) was used to estimate the stress ofgarnet peridotites from Finsch. The linear intercept method (Jungand Karato, 2001b) was used to measure the recrystallized grain-size of olivine. Grain boundaries of recrystallized grains werecounted along the horizontal and vertical lines. Recrystallized grainsize in 2-dimension (2-D) was estimated from the total length oflines divided by the total number of grain boundaries. The averagerecrystallized grain size of olivine in the samples 1512, 1516, and1517 in 2-D was 709㎛, 2018㎛, and 1421㎛, respectively. We usedthe stereographic correction factor (C ¼ 1.5) for 3-dimensionalgrain size (Gifkins, 1970) and the recrystallized grain size in 3-Dwas 1063, 3028, and 2131 ㎛, respectively (Table 1). Stress ofspecimens was estimated for the dry condition of olivine (Jung andKarato, 2001b) and turned out to be small (5e11 ± 5 MPa).

5. Discussion and implications for trench-parallel seismicanisotropy

5.1. LPO of olivine

The natural olivine in diamond-bearing garnet (Fig. 2b, c) fromFinsch, South Africa displayed B-type LPO (Fig. 5). Although therewere previous studies on the olivine fabrics from the deep mantle atother localities (Baptiste et al., 2012; Jung et al., 2013; Skemer andKarato, 2008; Wang et al., 2013a, 2013b), this study represents thefirst observation of B-type LPO of olivine under dry condition innatural rocks from a depth of greater than 100 km (P > 3 GPa). Thisobservation supports previous experimental studies of the pressure-induced slip transition in olivine (Jung et al., 2009; Ohuchi et al.,2011, 2012; Raterron et al., 2011) and confirms that pressure-induced B-type LPO of olivine exist in nature. There was an initialstudy on the effect of pressure on LPO of olivine. Couvy et al. (2004)deformed forsterite (Fo100) aggregates at P¼ 11 GPa and T¼ 1400 �Cusing a multianvil press, finding different LPOs from A-type LPO.However, it was not certain whether the different LPOs were pro-duced due to pressure, stress, or water because (1) stress of specimenwas varied from ~1 GPa to ~100 MPa during experiment, (2) olivinecontained large amount of water (~2000 ppm H/Si). On the otherhand, C-type LPO of olivinewas reported from the garnet-peridotitesin Sulu terrane in China (Xu et al., 2006). Xu et al. (2006) reportedthat the samples came from high P and low T condition in a dryenvironment, arguing that C-type LPO of olivine was formed due tohigh P and low T deformation of olivine. However, there has been nosuch an experimental data yet supporting this.

Previous experimental studies also showed that B-type LPO ofolivine can be produced by the deformation of olivine under water-rich conditions (COH � 200 ppm H/Si) at the pressures (P� 2.2 GPa)(Jung and Karato, 2001a; Jung et al., 2006; Katayama and Karato,2006). Our samples contained only small amount of water inolivine (�120 ppmH/Si), indicating that the B-type LPO of olivine inthis study was not formed due to water. There are now many ex-amples of B-type LPO of olivine found in natural rocks, where B-type LPO of olivine was interpreted to have been produced in thepresence of water. Examples are the B-type LPOs of olivine from theHigashi-akaishi, southwest Japan (Mizukami et al., 2004), from theCima di Gagnone, southern Switzerland (Skemer et al., 2006), fromthe Val Malenco, Italy (Jung, 2009), from the Bergen Arc, south-western Norway (Jung et al., 2014), from the southern MarianaTrench (Michibayashi et al., 2007), from the mantle xenoliths inShanwang, eastern China (Park and Jung, 2014), from the Samba-gawa belt in southwestern Japan (Tasaka et al., 2008), and from the

Fig. 6. Representative unpolarized FTIR spectra of olivine and orthopyroxene. Aninfrared beam size of 50 mm � 50 mm was used in the transmission mode. These dataindicate that the olivine in all of the samples is dry, while orthopyroxene contains asmall amount of water (see Table 1 and text).

Fig. 5. Pole figures of olivine in equal-area and upper-hemisphere projection. L represents the lineation, and S represents the foliation. N is the number of grains analyzed. The fabricstrength (M) is represented as the M-index. The contours in the pole figures and the numbers in the legend correspond to the multiples of uniform distribution. The color codingrefers to the density of the data points. A half scatter-width of 30� was used to draw the pole figures. All of the samples showed a strong B-type LPO of olivine. (For interpretation ofthe references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 4Elastic stiffnesses (Cij) of olivine and whole rock for the representative sample 1512.

i＼j 1 2 3 4 5 6

olivine1 262.64 77.99 78.98 0.61 �6.02 �1.512 232.84 82.82 �5.72 0.68 �1.313 203.71 �4.35 �3.24 0.674 67.26 �0.34 �3.15 73.26 �4.256 86.45Density 3.237 g/cm3

Whole1 259.71 78.43 78.93 0.62 �4.84 �1.212 235.27 82.21 �4.98 0.86 �0.963 210.79 �3.78 �2.96 0.184 69.99 �0.33 �2.515 74.91 �3.566 86.16Density 3.262 g/cm3

Elastic stiffnesses, Cij (GPa) of olivine andwhole rock of sample 1512 at foliation (XY)at 4 GPa and 1000 �C.Reference axes defined as 2: lineation, 3: normal to foliation, and 1: perpendicular toboth 2 and 3 directions (eg., Y ¼ 1, X ¼ 2, Z ¼ 3).

J. Lee, H. Jung / Journal of Structural Geology 70 (2015) 12e2218

Fig. 7. Seismic velocity and anisotropy calculated from the LPOs of olivine (Fig. 5). The east-west direction corresponds to the lineation (L). The center of the circle corresponds to thedirection normal to the foliation (this orientation was chosen to show clearly the polarization direction of the fast S-waves for vertically arriving seismic waves and is different fromthat used in Fig. 5). The azimuthal anisotropy of the P-waves (Vp) and the polarization anisotropy of the S-waves (AVs) are shown. Vs1 is a plot of the polarization direction of thefast S-waves along different orientations of propagation; the center of the figure corresponds to the vertical propagation. Black square and open circle represent maximum andminimum value, respectively.

J. Lee, H. Jung / Journal of Structural Geology 70 (2015) 12e22 19

chlorite peridotites in Almklovdalen, southwestern Norway (Kimand Jung, 2014).

Another mechanism to produce the B-type LPO is the defor-mation of olivine in diffusion creep. Recent experimental studyshowed that LPO of olivine can be formed in diffusion creep(Miyazaki et al., 2013; Sundberg and Cooper, 2008). By conductingdeformation experiment of the mixture of olivine and orthopyr-oxene in diffusion creep, Sundberg and Cooper (2008) reported theformation of B-type LPO of olivine for the samples containingorthopyroxene (>35%) in diffusion creep. However, our samplescontained less than 10% of orthopyroxene. In addition, strong LPOof olivine (Fig. 5), numerous undulose extinctions (Fig. 3b) andmany dislocations (Fig. 3c) in olivine in our samples indicate thatthe olivine in our samples were deformed in dislocation creep andthe B-type LPO of olivine was not produced by diffusion creep withhigh content of orthopyroxene.

5.2. Water content of minerals

Water content of olivine in our samples showed50e120 ± 20 ppm H/Si (Table 1). This result indicates that all ofthe samples were dry compared to those of previous experi-mental studies (Jung and Karato, 2001a; Jung et al., 2006;

Katayama and Karato, 2006). This result indicates that theB-type LPO of olivine observed in this study was produced underdry condition (<200 ppm H/Si). FTIR spectroscopy of orthopyr-oxene also showed that orthopyroxene contained a small amountof water (115e380 ± 40 ppm H/Si). Our specimens displayedpartition coefficients of water of DOpx/ol ¼ 6 ± 3, which are inagreement with previous experimental studies (Aubaud et al.,2004; Hauri et al., 2006). This result indicates strongly thatboth olivine and orthopyroxene were deformed under dry con-dition and the B-type LPO of olivine was not produced due towater.

5.3. Implications for seismic anisotropy

The observation of B-type LPO of olivine in natural rocks origi-nating from a depth greater than ~90 km under dry condition hassignificant implications for the interpretation of seismic anisotropyand the understanding of global tectonics (i.e., mantle flow). Thepolarization anisotropy of the fast shear wave below Finsch wasreported in previous studies (Fouch et al., 2004; Silver et al., 2001).They found that the angle between absolute plate motion of Africacontinent and the polarization direction of fast shear wave belowthe studied area is about 71�. This polarization anisotropy of the fast

Fig. 8. Seismic velocity and anisotropy of whole rock calculated from the LPOs of olivine and minor phases such as orthopyroxene and garnet. The east-west direction correspondsto the lineation (L). The center of the circle corresponds to the direction normal to the foliation (this orientation was chosen to show clearly the polarization direction of the fast S-waves for vertically arriving seismic waves and is different from that used in Fig. 5). The azimuthal anisotropy of the P-waves (Vp) and the polarization anisotropy of the S-waves(AVs) are shown. Vs1 is a plot of the polarization direction of the fast S-waves along different orientations of propagation; the center of the figure corresponds to the verticalpropagation. Black square and open circle represent maximum and minimum value, respectively.

J. Lee, H. Jung / Journal of Structural Geology 70 (2015) 12e2220

shear wave below Finsch may be explained by the existence of thefossil B-type LPO of olivine. Although the samples used in this paperwas not deformed below the subducting slab, our samples are richin olivine and they were deformed at high pressure and hightemperature condition similar to the environment below the slab,we think that the results of LPO and seismic anisotropy of olivine inthis study may be applied to the area below the slab. The trench-parallel seismic anisotropy observed in and below the subductingslab in many subduction zones at depths greater than ~90 km (Longand Silver, 2008, 2009; Tian and Zhao, 2012) may be attributed tothe B-type LPO of olivine induced by pressure and may be inter-preted as the entrainment of the sub-lithospheric mantle in thedirection of subduction rather than anomalous trench-parallel flow.This conclusion is supported by a recent seismological observationin Japan where trench-parallel seismic anisotropy was observed inthe subducting Pacific slab under Tohoku and the Philippine Seaslab under Kyushu at depths greater than 90 km (Wang and Zhao,2013).

6. Conclusions

Our petrofabric analysis of olivine in garnet peridotites fromFinsch, South Africa have led to the following main conclusions:

i) SEM/EBSD study of olivine showed that [001] axes of olivineare aligned subparallel to the lineation and [010] axes arealigned subnormal to lineation, which is a B-type LPO ofolivine. FTIR study of olivine and orthopyrone showed thatwater content of olivine and orthopyroxene was low enoughcompared to previous experimental studies (Jung andKarato, 2001a; Jung et al., 2006; Katayama and Karato,2006), indicating that samples were deformed under drycondition. Diamonds were found in the specimens andequilibrium pressures of the specimen obtained by geo-barometry were greater than 4 GPa, indicating that samplesoriginated from a deep interior of the Earth and deformedunder high pressure. Based on these results, it is concludedthat olivine in diamond-bearing garnet peridotites from theFinsch was deformed under dry condition at high pressureenvironment and the B-type LPO of olivine was produceddue to high pressure.

ii) This is the first report of the pressure-induced B-type LPO ofolivine found in nature. Our findings are consistent with theprevious experimental results which showed the formationof B-type LPO of olivine under dry condition at high pres-sures above P ¼ 3.1 GPa (Jung et al., 2009) and 5 GPa (Ohuchiet al., 2011).

J. Lee, H. Jung / Journal of Structural Geology 70 (2015) 12e22 21

iii) Finding the B-type LPO of olivine at high pressure above3 GPa has significant implications for the interpretation ofseismic anisotropy in many subduction zones. Trench-parallel seismic anisotropy at depths greater than ~90 kmin and below the subducting slab in many subduction zones(Long, 2013; Long and Silver, 2008; Russo and Silver, 1994;Tian and Zhao, 2012; Wang and Zhao, 2013) may be attrib-utable to the pressure-induced B-type LPO of olivine withoutinvoking anomalous trench-parallel flow, affecting mantledynamics and global tectonics.

Further research on the petrofabrics of natural rocks from thedeep interior of the Earth, as well as continued study of high res-olution seismic anisotropy in subduction zones, is needed to betterunderstand the source of seismic anisotropy and to constrain themantle dynamics of subduction zones.

Acknowledgments

H.J. thanks to the late Dr. Joe Boyd for providing with the sam-ples while H.J. stayed at the DTM, Carnegie Institution of Wash-ington, D.C. We also thank H.W. Green, S. Karato, Y. W. Kil, J. Rhie, Y.Cao, S. Jung, and M. Park for valuable comments and discussions.Two anonymous reviewers and the editor Dr. T. Takeshita aregreatly appreciated for helpful and valuable comments. H.J. wassupported by NRF grant funded by the MEST (3345-20130011,3345-20140009) in Korea.

References

Abramson, E.H., Brown, J.M., Slutsky, L.J., Zaug, J.M., 1997. The elastic constants ofSan Carlos olivine to 17 GPa. J. Geophys. Research-Solid Earth 102,12253e12263.

Aubaud, C., Hauri, E.H., Hirschmann, M.M., 2004. Hydrogen partition coefficientsbetween nominally anhydrous minerals and basaltic melts. Geophys. Res. Lett.31.

Baptiste, V., Tommasi, A., Demouchy, S., 2012. Deformation and hydration of thelithospheric mantle beneath the Kaapvaal craton, South Africa. Lithos 149,31e50.

Bass, J.D., 1989. Elasticity of grossular and spessartite garnets by Brillouin spec-troscopy. J. Geophys. Research-Solid Earth Planets 94, 7621e7628.

Becker, T.W., Lebedev, S., Long, M.D., 2012. On the relationship between azimuthalanisotropy from shear wave splitting and surface wave tomography. J. Geophys.Research-Solid Earth 117.

Bell, D.R., Rossman, G.R., Maldener, J., Endisch, D., Rauch, F., 2003. Hydroxide inolivine: a quantitative determination of the absolute amount and calibration ofthe IR spectrum. J. Geophys. Research-Solid Earth 108.

Ben Ismail, W., Mainprice, D., 1998. An olivine fabric database: an overview of uppermantle fabrics and seismic anisotropy. Tectonophysics 296, 145e157.

Brey, G.P., Kohler, T., 1990. Geothermobarometry in 4-phase lherzolites .2. newthermobarometers, and practical assessment of existing thermobarometers.J. Petrol. 31, 1353e1378.

Chai, M., Brown, J.M., Slutsky, L.J., 1997. The elastic constants of an aluminousorthopyroxene to 12.5 GPa. J. Geophys. Research-Solid Earth 102, 14779e14785.

Couvy, H., Frost, D.J., Heidelbach, F., Nyilas, K., Ungar, T., Mackwell, S., Cordier, P.,2004. Shear deformation experiments of forsterite at 11GPa-1400 degrees C inthe multianvil apparatus. Eur. J. Mineral. 16, 877e889.

Faccenda, M., Burlini, L., Gerya, T.V., Mainprice, D., 2008. Fault-induced seismicanisotropy by hydration in subducting oceanic plates. Nature 455, 1097e1110.

Flinn, D., 1962. On folding during three-dimensional progressive deformation. Q. J.Geol. Soc. Lond. 118, 385e428.

Fouch, M.J., Fischer, K.M., Parmentier, E.M., Wysession, M.E., Clarke, T.J., 2000. Shearwave splitting, continental keels, and patterns of mantle flow. J. Geophys.Research-Solid Earth 105, 6255e6275.

Fouch, M.J., Silver, P.G., Bell, D.R., Lee, J.N., 2004. Small-scale variations in seismicanisotropy near Kimberley, South Africa. Geophys. J. Int. 157, 764e774.

Gibson, S.A., Malarkey, J., Day, J.A., 2008. Melt depletion and enrichment beneaththe Western kaapvaal craton: evidence from Finsch Peridotite Xenoliths.J. Petrol. 49, 1817e1852.

Gifkins, R.C., 1970. Transitions in creep behaviour. J. Mater. Sci. 5, 156e165.Hauri, E.H., Gaetani, G.A., Green, T.H., 2006. Partitioning of water during melting of

the Earth's upper mantle at H2O-undersaturated conditions. Earth Planet. Sci.Lett. 248, 715e734.

Jadamec, M.A., Billen, M.I., 2010. Reconciling surface plate motions with rapid three-dimensional mantle flow around a slab edge. Nature 465, 338e341.

Ji, S.C., Li, A., Wang, Q., Long, C.X., Wang, H.C., Marcotte, D., Salisbury, M., 2013.Seismic velocities, anisotropy, and shear-wave splitting of antigorite serpen-tinites and tectonic implications for subduction zones. J. Geophys. Research-Solid Earth 118, 1015e1037.

Jung, H., 2009. Deformation fabrics of olivine in Val Malenco peridotite found inItaly and implications for the seismic anisotropy in the upper mantle. Lithos109, 341e349.

Jung, H., 2011. Seismic anisotropy produced by serpentine in mantle wedge. EarthPlanet. Sci. Lett. 307, 535e543.

Jung, H., Karato, S., 2001a. Water-induced fabric transitions in olivine. Science 293,1460e1463.

Jung, H., Karato, S.I., 2001b. Effects of water on dynamically recrystallized grain-sizeof olivine. J. Struct. Geol. 23, 1337e1344.

Jung, H., Katayama, I., Jiang, Z., Hiraga, I., Karato, S., 2006. Effect of water and stresson the lattice-preferred orientation of olivine. Tectonophysics 421, 1e22.

Jung, H., Lee, J., Ko, B., Jung, S., Park, M., Cao, Y., Song, S.G., 2013. Natural type-Colivine fabrics in garnet peridotites in North Qaidam UHP collision belt, NWChina. Tectonophysics 594, 91e102.

Jung, H., Mo, W., Green, H.W., 2009. Upper mantle seismic anisotropy resulting frompressure-induced slip transition in olivine. Nat. Geosci. 2, 73e77.

Jung, S., Jung, H., Austrheim, H., 2014. Characterization of olivine fabrics andmylonite in the presence of fluid and implications for seismic anisotropy andshear localization. Earth Planets Space 66.

Karato, S., 1987. Scanning electron-microscope observation of dislocations inolivine. Phys. Chem. Min. 14, 245e248.

Karato, S., Jung, H., Katayama, I., Skemer, P., 2008. Geodynamic significance ofseismic anisotropy of the upper mantle: new insights from laboratory studies.Annu. Rev. Earth Planet. Sci. 36, 59e95.

Karato, S.I., Toriumi, M., Fujii, T., 1980. Dynamic recrystallization of olivine-singlecrystals during high-temperature creep. Geophys. Res. Lett. 7, 649e652.

Katayama, I., Hirauchi, H., Michibayashi, K., Ando, J., 2009. Trench-parallel anisot-ropy produced by serpentine deformation in the hydrated mantle wedge. Na-ture 461, 1114e1117.

Katayama, I., Jung, H., Karato, S.I., 2004. New type of olivine fabric from deformationexperiments at modest water content and low stress. Geology 32, 1045e1048.

Katayama, I., Karato, S., 2006. Effect of temperature on the B- to C-type olivine fabrictransition and implication for flow pattern in subduction zones. Phys. EarthPlanet. Inter. 157, 33e45.

Kim, D., Jung, H., 2014. Deformation microstructures of olivine and chlorite inchlorite peridotites from Almklovdalen in the Western Gneiss Region, SWNorway and implications for seismic anisotropy. Int. Geol. Rev. http://dx.doi.org/10.1080/00206814.00202014.00936054.

Kneller, E.A., Long, M.D., van Keken, P.E., 2008. Olivine fabric transitions and shearwave anisotropy in the Ryukyu subduction system. Earth Planet. Sci. Lett. 268,268e282.

Kohlstedt, D.L., Goetze, C., Durham, W.B., Vandersande, J., 1976. New technique fordecorating dislocations in olivine. Science 191, 1045e1046.

Kov�acs, I., Hermann, J., O'Neill, H.S.C., Gerald, J.F., Sambridge, M., Horvath, G., 2008.Quantitative absorbance spectroscopy with unpolarized light: part II. Experi-mental evaluation and development of a protocol for quantitative analysis ofmineral IR spectra. Am. Mineral. 93, 765e778.

Kumazawa, M., Anderson, O.L., 1969. Elastic moduli, pressure derivatives andtemperature derivatives of single-crystal olivine and single-crystal forsterite.J. Geophys. Res. 74, 5961e5972.

Long, M.D., 2013. Constraints on subduction geodynamics from seismic anisotropy.Rev. Geophys. 51, 76e112.

Long, M.D., Silver, P.G., 2008. The subduction zone flow field from seismic anisot-ropy: a global view. Science 319, 315e318.

Long, M.D., Silver, P.G., 2009. Mantle flow in subduction systems: the subslab flowfield and implications for mantle dynamics. J. Geophys. Research-Solid Earth114.

Mainprice, D., 1990. A Frotran program to calculate seismic anisotropy from thelattice preferred orientation of minerals. Comput. Geosci. 16, 385e393.

Mainprice, D., 2007. Seismic anisotropy of the deep earth from a Mineral and rockphysics perspective. In: Schubert, G. (Ed.), Treatise on Geophysics. Elsevier,Amsterdam, pp. 437e491.

Mainprice, D., Tommasi, A., Couvy, H., Cordier, P., Frost, D.J., 2005. Pressure sensi-tivity of olivine slip systems and seismic anisotropy of Earth's upper mantle.Nature 433, 731e733.

Michibayashi, K., Tasaka, M., Ohara, Y., Ishii, T., Okamoto, A., Fryer, P., 2007. Variablemicrostructure of peridotite samples from the southern Mariana Trench: evi-dence of a complex tectonic evolution. Tectonophysics 444, 111e118.

Miyazaki, T., Sueyoshi, K., Hiraga, T., 2013. Olivine crystals align during diffusioncreep of Earth's upper mantle. Nature 502, 321e325.

Mizukami, T., Wallis, S.R., Yamamoto, J., 2004. Natural examples of olivine latticepreferred orientation patterns with a flow normal a-axis maximum. Nature 427,432e436.

Mosenfelder, J.L., Deligne, N.I., Asimow, P.D., Rossman, G.R., 2006. Hydrogenincorporation in olivine from 2-12 GPa. Am. Mineral. 91, 285e294.

Nicolas, A., Christensen, N.I., 1987. Formation of anisotropy in Upper Mantle Peri-dotites - a review. In: Fuchs, K., Froidevaux, C. (Eds.), Composition, Structure

J. Lee, H. Jung / Journal of Structural Geology 70 (2015) 12e2222

and Dynamics of the Lithosphere-Asthenosphere System. American Geophys-ical Union, Washington, D. C., pp. 111e123.

O'Neill, H.S.C., Wood, B.J., 1979. Experimental-study of Fe-Mg partitioning betweengarnet and its calibration as a geothermometer. Contrib. Mineral. Petrol. 70,59e70.

Ohuchi, T., Kawazoe, T., Nishihara, Y., Nishiyama, N., Irifune, T., 2011. High pressureand temperature fabric transitions in olivine and variations in upper mantleseismic anisotropy. Earth Planet. Sci. Lett. 304, 55e63.

Ohuchi, T., Nishihara, Y., Kawazoe, T., Spengler, D., Shiraishi, R., Suzuki, A.,Kikegawa, T., Ohtani, E., 2012. Superplasticity in hydrous melt-bearing dunite:implications for shear localization in Earth's upper mantle. Earth Planet. Sci.Lett. 335, 59e71.

Palasse, L.N., Vissers, R.L.M., Paulssen, H., Basu, A.R., Drury, M.R., 2012. Micro-structural and seismic properties of the upper mantle underneath a riftedcontinental terrane (Baja California): an example of sub-crustal mechanicalasthenosphere? Earth Planet. Sci. Lett. 345, 60e71.

Panozzo, R., 1984. Two-dimensional strain from the orientation of lines in a plane.J. Struct. Geol. 6, 215e221.

Park, J., Levin, V., 2002. Seismic anisotropy: tracing plate dynamics in the mantle.Science 296, 485e489.

Park, M., Jung, H., Kil, Y., 2014. Petrofabrics of olivine in a rift axis and rift shoulderand their implications for seismic anisotropy beneath the Rio Grande rift. Isl.Arc. http://dx.doi.org/10.1111/iar.12089 in press.

Park, Y., Jung, H., 2014. Deformation microstructures of olivine and pyroxene inmantle xenoliths in Shanwang, eastern China near the convergent plate margin,and implications for seismic anisotropy. Int. Geol. Rev. http://dx.doi.org/10.1080/00206814.00202014.00928240.

Paterson, M.S., 1982. The determination of hydroxyl by infrared-absorption inquartz, silicate-glasses and similar materials. Bull. De. Mineral. 105, 20e29.

Raterron, P., Chen, J.H., Geenen, T., Girard, J., 2011. Pressure effect on forsteritedislocation slip systems: implications for upper-mantle LPO and low viscosityzone. Phys. Earth Planet. Inter. 188, 26e36.

Russo, R.M., Silver, P.G., 1994. Trench-parallel flow meneath the Nazca plate fromseismic anisotropy. Science 263, 1105e1111.

Savage, M.K., 1999. Seismic anisotropy and mantle deformation: what have welearned from shear wave splitting? Rev. Geophys. 37, 65e106.

Silver, P.G., 1996. Seismic anisotropy beneath the continents: probing the depths ofgeology. Annu. Rev. Earth Planet. Sci. 24, 375e432.

Silver, P.G., Gao, S.S., Liu, K.H., Kaapvaal Seismic, G., 2001. Mantle deformationbeneath southern Africa. Geophys. Res. Lett. 28, 2493e2496.

Skemer, P., Karato, S.I., 2008. Sheared lherzolite xenoliths revisited. J. Geophys.Research-Solid Earth 113.

Skemer, P., Katayama, B., Jiang, Z.T., Karato, S., 2005. The misorientation index:development of a new method for calculating the strength of lattice-preferredorientation. Tectonophysics 411, 157e167.

Skemer, P., Katayama, I., Karato, S.I., 2006. Deformation fabrics of the Cima diGagnone peridotite massif, Central Alps, Switzerland: evidence of deformationat low temperatures in the presence of water. Contrib. Mineral. Petrol. 152,43e51.

Skemer, P., Warren, J.M., Kelemen, P.B., Hirth, G., 2010. Microstructural and rheo-logical evolution of a Mantle Shear Zone. J. Petrol. 51, 43e53.

Smith, G.P., Wiens, D.A., Fischer, K.M., Dorman, L.M., Webb, S.C., Hildebrand, J.A.,2001. A complex pattern of mantle flow in the Lau backarc. Science 292,713e716.

Soda, Y., Wenk, H.R., 2014. Antigorite crystallographic preferred orientations inserpentinites from Japan. Tectonophysics 615, 199e212.

Song, T.R.A., Kawakatsu, H., 2012. Subduction of oceanic asthenosphere: evidencefrom sub-slab seismic anisotropy. Geophys. Res. Lett. 39.

Sundberg, M., Cooper, R.F., 2008. Crystallographic preferred orientation producedby diffusional creep of harzburgite: effects of chemical interactions amongphases during plastic flow. J. Geophys. Research-Solid Earth 113.

Tasaka, M., Michibayashi, K., Mainprice, D., 2008. B-type olivine fabrics developed inthe fore-arc side of the mantle wedge along a subducting slab. Earth Planet. Sci.Lett. 272, 747e757.

Tian, Y., Zhao, D.P., 2012. Seismic anisotropy and heterogeneity in the Alaska sub-duction zone. Geophys. J. Int. 190, 629e649.

Tommasi, A., Vauchez, A., Ionov, D.A., 2008. Deformation, static recrystallization,and reactive melt transport in shallow subcontinental mantle xenoliths (TokCenozoic volcanic field, SE Siberia). Earth Planet. Sci. Lett. 272, 65e77.

Van der Wal, D., Chopra, P., Drury, M., Fitz Gerald, J.D., 1993. Relationships betweendynamically recrystallized grain-size and deformation conditions in experi-mentally deformed olivine rocks. Geophys. Res. Lett. 20, 1479e1482.

Wang, J., Zhao, D.P., 2013. P-wave tomography for 3-D radial and azimuthalanisotropy of Tohoku and Kyushu subduction zones. Geophys. J. Int. 193,1166e1181.

Wang, Q., Xia, Q.K., O'Reilly, S.Y., Griffin, W.L., Beyer, E.E., Brueckner, H.K., 2013a.Pressure- and stress-induced fabric transition in olivine from peridotites in theWestern Gneiss Region (Norway): implications for mantle seismic anisotropy.J. Metamorph. Geol. 31, 93e111.

Wang, Y.F., Zhang, J.F., Shi, F., 2013b. The origin and geophysical implications of aweak C-type olivine fabric in the Xugou ultrahigh pressure garnet peridotite.Earth Planet. Sci. Lett. 376, 63e73.

Warren, J.M., Hirth, G., Kelemen, P.B., 2008. Evolution of olivine lattice preferredorientation during simple shear in the mantle. Earth Planet. Sci. Lett. 272,501e512.

Watanabe, T., Shirasugi, Y., Yano, H., Michibayashi, K., 2011. Seismic Velocity inAntigorite-bearing Serpentinite Mylonites. In: Geological Society, London,Special Publications, vol. 360, pp. 97e112.

Withers, A.C., Bureau, H., Raepsaet, C., Hirschmann, M.M., 2012. Calibration ofinfrared spectroscopy by elastic recoil detection analysis of H in syntheticolivine. Chem. Geol. 334, 92e98.

Xu, Z.Q., Wang, Q., Ji, S.C., Chen, J., Zeng, L.S., Yang, J.S., Chen, F.Y., Liang, F.H.,Wenk, H.R., 2006. Petrofabrics and seismic properties of garnet peridotite fromthe UHP Sulu terrane (China): implications for olivine deformation mechanismin a cold and dry subducting continental slab. Tectonophysics 421, 111e127.

Zhang, S.Q., Karato, S., Fitz Gerald, J.D., Faul, U.H., Zhou, Y., 2000. Simple sheardeformation of olivine aggregates. Tectonophysics 316, 133e152.