Relationships between geochemistry and structure beneath a palaeo-spreading centre: a study of the mantle section in the Oman ophiolite

Earth and Planetary Science Letters 180 (2000) 133^148 www.elsevier.com/locate/epsl

Relationships between geochemistry and structure beneath a palaeo-spreading centre: a study of the mantle section in the Oman ophiolite Marguerite Godard *, David Jousselin 1 , Jean-Louis Bodinier Laboratoire de Tectonophysique, ISTEEM, CNRS, Universite¨ Montpellier 2, cc 49, Place Euge©ne Bataillon, 34095 Montpellier Cedex 5, France Received 8 September 1999; received in revised form 5 May 2000; accepted 9 May 2000

Abstract The Oman ophiolite exposes a large and well-preserved mantle section beneath a palaeo-spreading centre. The mantle section is mainly composed of extremely refractory harzburgites with relatively homogeneous modal and major element compositions. Nevertheless, our trace element data exhibit variations connected with the main mantle structures, which allow us to define three geochemical and structural domains. The main harzburgitic mantle section, mainly constituted of strongly refractory harzburgites characterised by chondrite-normalised REE patterns that are steadily depleted from HREE to LREE. These rocks are interpreted as mantle residues after s 15% melt extraction. Their REE signature can be explained by melt transport associated with partial melting. The diapir areas (mainly the Maqsad diapir), defined by plunging lineations. They are constituted of harzburgites with roughly the same modal composition as the main mantle section but distinct, concave-upward REE patterns. The regions of most active upwelling (characterised by sub-vertical lineations) are further distinguished by higher Al2 O3 /CaO ratios and TiO2 contents. This character is ascribed to focused partial melt upwelling. The diapirs are interpreted as local instabilities in upwelling mantle, possibly triggered by feedback mechanisms between deformation and melt percolation. The Maqsad diapir is topped by a thick, dunitic, mantle^crust transition zone (MTZ) that displays the same trace-element signature as the diapir. However, the dunites are distinguished by low Mg# values and Ni contents. Together with structural evidence, this allows us to interpret the MTZ dunites as diapir harzburgites that were strongly modified by olivine-forming melt^rock reactions at high melt/ rock ratios. The MTZ is thought to act as a major collecting zone for mantle melts. The cpx-harzburgites from the base of the mantle section. These rocks are distinguished by high clinopyroxene contents ( s 5%), low AL2 O3 /CaO and `spoon-shaped' REE patterns. They were individualised from the rest of the harzburgite mantle section by a cpxforming melt^rock reaction at decreasing malt mass. This reaction probably occurred at near-solidus conditions along the lithosphere^asthenosphere boundary. The formation of these three domains may be integrated in a geodynamic scenario involving the reactivation of an oceanic lithosphere, a process that would be related to the ridge propagator identified in the Oman ophiolite. ß 2000 Elsevier Science B.V. All rights reserved.

* Corresponding author. Tel.: +33 4 67 14 39 37; Fax: +33 4 67 14 36 03; E-mail: [email protected] 1

Present address: University of Oregon, 1272 Department of Geological Sciences, Eugene, OR 97403, USA.

0012-821X / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 1 2 - 8 2 1 X ( 0 0 ) 0 0 1 4 9 - 7

EPSL 5506 29-6-00

134

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

Keywords: ophiolite; oceanic crust; mantle; harzburgite; trace elements; melts; transport

1. Introduction Petrological and geochemical studies of midocean ridge basalts (MORB) and abyssal peridotites, together with geophysical studies, have revealed the complexity of melting and melt migration mechanisms associated with the accretion of oceanic lithosphere (see reviews [1^3]). Yet, the exact nature of melt processes in the upper mantle below spreading ridges is still debated [4^7]. Important issues are whether melting is fractional or at equilibrium (e.g., [8]) and whether melts are transported by pervasive porous £ow (e.g., [6,9]) or in chemically isolated conduits (porous £ow channels [5] or fractures [4]). Recent studies have shown that trace element variations in mantle peridotites, coupled with textural and petrographic investigations, and numerical modelling, may be used to evaluate the extent of processes such as melt transport and melt^rock interactions [5,6,10^15]. However, trace element data are scarce for abyssal peridotites because these rocks are rarely exposed and di¤cult to analyse by traditional analytical methods due to their extreme depletion. Until recently, geochemical data on oceanic lithosphere were mostly restricted to rare earth elements (REE) in clinopyroxenes from abyssal [8,16] and ophiolite [5,15] peridotites and rarely, in whole rock [17^20]. New perspectives have been opened by inductively coupled plasma^mass spectrometry (ICP^MS), which makes it possible to measure a number of trace elements (for instance, U, Th, Nb, Ta plus REE) at the very low concentration levels typical of oceanic peridotites [6,21,22]. Abyssal peridotites make it possible to study mantle processes in present-day oceanic settings (e.g., [6,8,21]), but their sampling is exceptional and limited to the uppermost limit of the oceanic mantle. In contrast, ophiolites expose large mantle sections as well as mantle^crust relationships, and make it possible to study large-scale variations in peridotites [5,15,23]. As such, their study can bring new insights into the understanding of

melt processes occurring in the oceanic mantle lithosphere. The Oman ophiolite contains the largest and one of the best-preserved sections of oceanic lithosphere [4,24,25]. In this paper, we present a systematic study of major and trace elements in the mantle peridotites of this ophiolite. Our purpose is to characterise the composition of the mantle section and to combine geochemical and structural data to place constraints on melt processes in the oceanic upper mantle. 2. Geological setting The mantle section of the Oman ophiolite is mainly composed of harzburgite tectonites with coarse-grained porphyroclastic textures inherited from high-temperature (T s 1200³C) deformations associated with asthenospheric mantle £ow [24,25]. It also contains a few centimetre- to metre-thick dunite bands, discordant to the harzburgite foliation. Near the base and top of the mantle section, the dunite bands become parallel to the foliation. As illustrated in Fig. 1, this study is focused on the south-eastern massifs of Nakhl, Sumail and Wadi Tayin, where the mantle section can reach a thickness of 12 km below the Moho (in the north-eastern part of Wadi Tayin [24,25]). In this region, structural studies have revealed the existence of a complex ridge system [26,27] and diapir areas characterised by concentric plunging peridotite lineations [28]. In the Sumail and Wadi Tayin massifs, diabase dyke directions de¢ne two ridge systems that are interpreted as the opening of a NW^SE propagator in a slightly older lithosphere, characterised by NE^SW diabase trajectories [26,27]. In the Sumail and Nakhl massifs, the peridotite foliation is generally parallel to the Moho, except in diapir areas. Diapirs are kilometre-scale concentric structures (V10U10 km2 ), characterised by steep vertical foliations ( s 45³) and lineations ( s 30³). They are observed

EPSL 5506 29-6-00

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

135

Fig. 1. (a) Synoptic map of the SE massifs of the Oman ophiolite (location in upper right inset) after Boudier et al. [27], showing the location of the analysed samples. The Nakhl massif was displaced toward the Sumail and Wadi Tayin massifs to account for their separation during post-obduction doming of the Jebel Akhdar. Thick dotted lines represent the limits of the NW^SE ridge propagator inferred from diabase dyke trajectories. Light grey areas represent the low-temperature deformation zones. Diapirs are indicated by light to dark grey patterns bounded by lineation isoplunge contours. (b) Enlargement of the Maqsad diapir area, also giving average isoplunge values for the diapir areas.

mostly along the NW^SE ridge system [27]. Thick dunite mantle^crust transition zones (MTZ ^ up to 800 m thick) are observed at the top of the diapiric structures [29]. MTZ contain di¡use segregations of clinopyroxene and plagioclase (`impregnated' dunites) as well as numerous gabbro sills. At the base of the ophiolite slab, the hightemperature deformation is overprinted by a lowtemperature (1100^900³C) deformation, responsible for a ¢ne-grained porphyroclastic texture in the harzburgites [24]. The low-temperature deformation is ascribed to the early detachment of the ophiolite [25,30]. 3. Sampling The 29 peridotites analysed for this study are

described in Table 1. Twenty-four peridotites were sampled within the limits of the NW trending propagator (Fig. 1). These include two harzburgites from the vicinity of the small Nakhl diapir (95OA77, 93OF101) and 16 samples from the Maqsad diapir area, in the western part of the Sumail massif. Among the Maqsad samples, 11 are harzburgites from the diapir area (de¢ned by lineations v 30³) and three are dunites (devoid of `impregnation' features) from the thick transition zone set above the diapir. The two other Maqsad peridotites represent the transition from the main harzburgitic sequence (96OF59B) to the dunitic MTZ (96OF59). The ¢ve remaining peridotites are from the north of the Wadi Tayin massif. They were sampled by Boudier and Coleman [24] and have already been studied by Kelemen and co-workers [5]. Two samples are dunites

EPSL 5506 29-6-00

Type

Harz.

Harz.

Harz.

Harz.

94MA6T

95OD84

93OF21

EPSL 5506 29-6-00

Harz.

Harz.

Du.

Du.

Cpx-harz.

Cpx-harz.

Cpx-harz.

94OE23

94OE27

OMB32

OMB33

OMB14

OMB23

OMB27

Harz.

Harz.

Harz.

Harz.

95OA77

93OF101

90OA38

90OA37A

Nakhl

Harz.

91OA29

Wadi Tayin

Harz.

93OF23

Harz.

95OD107

93OF22

Harz.

95OD153

Harz.

Harz.

95OD131

95OD174

Harz.

95OD191

Du.

95OD78A

Du.

Du.

94MA5

Harz.

Du.

94MA5ALT

95OD92

Harz.

96OF59B

96OF59

Harz.

93OA09

Maqsad area (Sumail massif)

location

Sample

40

30

30

20

80

55

20

55

60

65

40

35

75

80

50

80

60

80

35

35

50

35

45

0

0

0

0

15

40

5

0

20

25

10

10

20

20

20

25

0

75

55

50

70

45

80

35

30

30

35

30

0

0

0

0

5

40

(³)

(³)

70

L

F

Structure

Diapir

Diapir

Diapir

Diapir

Diapir

Diapir

Diapir

Diapir

Diapir

Diapir

Diapir

MTZ

MTZ

MTZ

MTZ

73.4

75.4

81.9

78.9

71.1

69.6

75.8

98.6

97.9

72.3

80.4

71.0

71.8

72.3

74.5

74.2

77.0

72.1

72.7

72.9

70.9

74.3

75.0

95^99

95^99

95^99

95^99

78.7

72.4

23.4

21.0

17.0

19.3

21.4

22.3

17.2

^

^

23.5

18.2

26.4

23.4

23.1

21.0

21.5

17.5

23.1

24.5

24.6

26.3

23.3

20.5

18.7

25.4

olivine opx

1.4

2.2

2.6

3.0

1.0

1.8

6.5

6.7

5.3

0.1

0.5

3.8

1.2

2.6

3.5

3.0

2.8

2.8

3.4

3.3

2.8

2.5

2.8

2.1

4.3

63

63

63

63

cpx

1.2

0.6

0.7

0.2

1.0

1.4

1.7

1.3

1.6

0.4

0.2

6 0.1

1.3

1.6

1.7

1.5

2.1

1.5

6 0.1

6 0.1

6 0.1

0.3

0.2

62

62

62

62

sp.

Primary modal composition

40.81 0.76

40.29 0.52

39.97 0.43

39.63 0.41

39.87 0.90

40.89 1.07

39.76 1.01

35.00 0.33

35.00 0.37

42.37 0.77

38.50 0.49

40.69 0.57

39.80 0.97

42.34 1.09

40.10 1.03

40.12 0.99

40.66 1.11

40.62 1.02

40.28 0.47

40.28 0.54

41.22 0.64

41.24 0.67

40.13 0.62

36.20 0.87

35.86 0.71

35.48 0.71

36.19 0.72

39.94 0.81

4.40

4.09

5.23

3.70

3.89

5.06

3.92

3.40

2.68

5.59

3.81

4.13

4.85

5.87

5.72

4.57

4.71

4.32

4.29

3.86

4.39

3.47

4.77

4.77

4.27

6.82

5.33

4.13

4.11

Al2 O3 FeO

40.60 0.62

(wt%)

SiO2

3.01

3.80

3.64

4.23

3.72

2.58

3.86

4.34

5.35

2.47

4.25

3.41

3.12

2.11

2.48

3.45

3.11

3.65

3.70

4.41

3.30

4.28

3.17

5.73

4.48

3.60

5.47

3.56

3.61

0.12

0.12

0.12

0.12

0.12

0.12

0.12

0.11

0.12

0.13

0.07

0.12

0.12

0.13

0.12

0.12

0.12

0.12

0.11

0.11

0.10

0.11

0.11

0.16

0.11

0.17

0.16

0.12

0.12

CaO

41.53 0.85

40.38 0.62

43.82 0.42

41.83 0.63

39.70 1.70

40.64 1.80

41.16 1.46

42.90 0.02

40.41 0.15

42.56 1.16

41.86 0.49

40.87 0.96

40.76 1.03

42.60 0.98

41.87 0.88

41.52 0.88

42.52 1.04

41.06 1.00

40.47 0.84

40.78 0.80

41.32 0.89

42.03 0.73

41.92 1.22

42.23 0.45

42.30 0.32

42.26 0.34

41.18 0.39

42.36 0.56

41.16 0.76

Fe2 O3 MnO MgO

0.14

0.10

0.01

0.01

0.02

0.01

0.02

^

^

0.04

0.02

0.25

0.88

0.01

0.94

0.68

0.22

0.01

6 0.01

0.01

6 0.01

0.02

6 0.01

0.30

0.01

0.01

0.16

0.01

0.03

Na2 O

Table 1 Structural features, modal compositions, and major and trace element analyses of the studied Oman peridotites

0.27

0.06

0.01

^

^

^

^

^

^

0.03

0.03

0.25

0.14

0.08

0.10

0.34

0.12

0.01

6 0.01

6 0.01

6 0.01

6 0.01

6 0.01

6 0.01

0.01

0.02

0.08

0.02

0.01

K2 O

0.02

0.02

0.02

0.01

0.02

0.02

0.02

0.01

0.02

0.02

0.02

0.02

0.07

0.07

0.07

0.07

0.07

0.07

0.01

0.01

0.01

0.01

0.01

0.07

0.07

0.08

0.08

0.07

0.02

TiO2

0.01

0.02

0.01

0.02

^

^

^

^

^

0.02

0.02

0.02

0.02

0.01

0.03

0.01

0.02

0.01

6 0.01

6 0.01

6 0.01

6 0.01

6 0.01

0.01

0.01

0.01

0.01

0.01

0.03

P2 O5

8.06

9.70

6.36

9.50

9.60

6.50

9.14

13.14

16.50

4.82

10.34

8.03

7.99

4.94

6.97

7.91

7.00

7.97

8.86

8.88

7.16

7.76

6.68

9.37

11.50

9.79

10.62

8.95

8.40

LOI

Mg#

87.8

91.2

88.4 91.2

90.7

91.1

90.8

90.7 90.9

90.6

90.3

90.9

99.98 91.3

99.7

100.04

100.09

99.54 90.8

98.69 90.8

100.37

99.25 91.4

100.60

99.98 90.7

99.90 90.8

99.32 91.1

99.75 90.5

100.23

99.89 90.5

100.66

100.70

99.86 90.7

99.03 90.5

99.68 90.4

99.03 91.0

100.32

98.63 90.8

100.68

99.65 90.2

99.29 88.3

100.39

100.54

99.47 91.0

Total

Ni

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

2260

2239

2121

2037

1924

2138

n.a.

n.a.

n.a.

n.a.

n.a.

1477

1320

1451

1406

2146

n.a.

(ppm)

2870

2420

2110

3307

3030

2910

2440

5780

4100

2910

1660

3090

2530

3080

2290

2315

3070

3090

2550

2560

3230

2990

3120

4350

3120

2590

2590

2930

2690

Cr

136 M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

Rb

(ppb)

Sr

32.5

23.5

25.6

27.6

22.6

23.9

95OD153

95OD107 (2)

95OD174 (2)

93OF22 (2)

93OF23

EPSL 5506 29-6-00

62.4

34.8

40.5

34.7

139.6

94OE27

OMB32 (2)

OMB33

OMB14

OMB23

37.9

44.7

83.1

1.6

90OA38

90OA37A

PCC-1 (7)

RSD (%)

1.4

374

1 385

5 292

9 684

9 567

820

103

165

513

3 027

390

3 240

3 303

18 780

587

7 497

9 715

835

6 162

2 789

3 951

1 181

6 360

1 612

2 082

17 749

1.6

114.2

152.0

13.1

12.3

6.6

54.0

12.4

13.8

38.0

62.2

96.7

60.2

88.1

13.5

38.7

10.0

^

8.1

11.3

25.5

13.7

25.6

26.6

36.7

47.2

32.9

50.8

52.3

75.6

111.8

Zr

see overleaf for more information.

38.9

27.4

95OA77 (4)

93OF101

Nakhl

57.7

29.9

94OE23

OMB27 (2)

21.6

91OA29

Wadi Tayin

39.2

33.6

95OD191

95OD131 (2)

27.0

28.5

95OD92 (2)

93OF21

33.0

96OF59

95OD84

29.9

95OD78A (2)

23.1

9 835

25.9

94MA5 (2)

94MA6T (2)

1 190

31.1

94MA5ALT

5 755

96OF59B (2)

1 040

31.8

30.1

93OA09 (2)

Maqsad area (Sumail massif)

Sample

Table 1 (continued)

2.6

24.5

31.2

23.8

24.4

21.9

17.7

15.2

15.9

14.0

15.1

31.6

25.6

32.7

47.0

30.8

21.9

24.7

31.5

23.9

24.2

20.0

23.7

25.7

25.3

29.7

22.9

23.8

22.8

26.7

22.1

Nb

2.8

8.1

1.9

^

1.4

0.7

13.7

39.2

7.0

2.1

3.4

9.3

^

^

2.8

21.0

3.0

4.4

4.3

3.1

1.3

2.1

2.8

3.1

6.4

1.7

2.7

3.1

1.3

9.5

4.8

Cs

1.1

1033

304

170

386

219

340

77

89

167

229

75

102

178

895

71

167

221

58

97

148

235

168

178

191

645

314

142

838

133

391

Ba

1.8

37.6

5.9

4.0

4.2

6.1

3.1

1.2

1.6

3.7

6.0

^

3.1

^

3.4

2.9

^

^

^

3.4

1.7

3.1

4.1

^

4.1

^

^

^

3.3

3.6

La

9.2

^

7.3

2.1

^

2.5

1.9

5.0

3.6

5.8

9.4

7.9

7.8

3.1

4.5

6.2

8.2

1.5

56.2

18.6

11.8

11.0

4.8

10.6

4.5

2.7

4.2

10.7

22.1

10.4

13.1

Ce

2.2

7.6

^

^

^

^

^

^

0.4

^

2.0

^

^

^

^

^

^

^

^

^

^

0.6

^

^

^

^

^

^

^

^

1.5

Pr

2.4

29.5

25.4

10.0

7.7

3.0

8.7

4.6

4.5

4.8

9.1

25.1

11.8

18.1

^

4.6

3.0

^

3.3

2.3

4.3

2.6

5.1

6.5

4.6

8.1

4.5

5.1

6.3

8.1

7.2

Nd

6.1

7.6

14.0

5.2

3.1

7.0

5.1

6.8

^

4.9

15.1

^

7.5

^

2.5

^

1.3

^

^

^

^

2.1

^

^

^

^

^

3.3

4.5

^

Sm

7.1

2.1

4.5

2.0

1.1

0.7

3.4

2.3

3.8

0.8

1.9

5.8

^

3.3

0.6

1.5

0.5

0.6

^

0.6

^

0.5

0.7

^

^

^

^

1.4

1.4

1.8

^

Eu

5.2

9.5

17.6

7.0

2.9

2.9

18.7

17.8

23.3

3.0

6.5

21.1

9.7

11.1

2.0

4.7

1.8

1.5

2.3

2.6

2.2

^

^

^

3.6

6.6

^

5.0

5.1

6.5

5.6

Gd

4.0

1.45

3.11

1.62

0.54

0.50

4.65

5.34

6.26

0.77

1.50

4.50

1.92

1.94

0.67

1.04

0.41

0.47

0.59

0.51

0.65

0.59

0.65

0.57

0.80

1.07

0.76

1.09

1.00

1.07

1.02

Tb

9.5 8.9

4.1

12.1

29.3

14.9

6.9

5.1

50.1

54.8

57.1

6.0

13.6

36.0

17.0

16.7

8.8

10.3

6.5

6.8

9.3

6.8

6.7

7.2

7.2

6.2

9.2

11.2

7.7

12.4

10.9

Dy

1.6

3.11

8.34

4.90

1.84

1.75

13.67

15.66

15.50

1.75

3.74

9.77

4.78

4.73

2.89

3.14

2.67

2.50

3.32

2.69

2.32

2.50

3.06

2.59

3.25

3.08

2.67

3.91

3.63

2.17

3.17

Ho

3.9

13.5

29.5

19.9

9.6

7.6

47.7

51.8

48.7

7.0

14.5

34.0

16.6

17.7

12.6

14.4

12.3

12.3

14.5

13.8

10.7

10.9

13.0

13.9

14.6

11.6

11.8

16.6

15.0

9.7

13.4

Er

Yb

2.9

1.3

2.90 23.5

6.24 49.2

4.49 35.9

2.19 22.1

1.82 19.1

8.77 65.1

9.84 77.0

9.08 65.7

1.65 14.5

2.82 22.7

6.89 51.2

3.87 30.2

3.85 32.2

2.88 25.1

3.45 29.0

2.93 28.9

2.58 25.7

3.39 31.7

2.82 27.4

2.48 23.6

2.66 25.7

2.96 28.2

3.45 27.7

3.15 29.8

2.57 23.2

2.26 19.9

4.22 40.7

3.83 35.3

2.37 19.7

3.11 26.2

Tm

6.0

6.8

7.1

6.0

7.2

6.7

6.0

7.2

6.1

5.3

6.3

6.3

6.3

6.4

5.6

4.8

9.6

8.8

4.8

1.3

5.4

10.5

7.7

5.2

4.6

13.4

14.6

13.3

3.6

4.8

10.4

Lu

4.9

5.1

9.1

3.3

^

1.3

3.9

3.0

3.0

1.4

3.0

5.2

4.1

2.6

^

2.2

1.0

0.7

0.8

1.3

1.8

^

1.6

1.0

2.3

2.5

2.4

2.9

2.6

3.2

3.3

Hf

9.3

1.5

2.6

1.4

0.8

0.8

0.7

0.6

1.6

0.7

0.7

2.0

2.6

1.1

0.6

0.6

0.6

^

0.6

0.6

1.5

1.9

1.0

0.6

1.5

0.8

0.7

0.8

0.5

0.8

1.4

Ta

^

^

13

25

100

16

24

11

54

^

^

^

18

^

56

18

21

53

30

12

56

10

20

21

48

55

^

77

81

0.2

7997

Pb

U

0.55

0.85

0.48

0.6

0.44

^

^

1.33

0.8

2.3

10.7 5.21

1.50 ^

1.18 ^

0.6

^

1.36 0.92

0.49 0.3

^

^

1.25 1.03

1.40 ^

1.37 ^

1.04 ^

^

0.64 0.79

0.92 1.08

^

0.92 0.79

0.64 0.56

0.66 0.42

^

0.94 0.90

0.40 ^

0.87 ^

^

0.97 1.27

0.94 ^

0.63 0.7

1.06 0.8

0.83 0.43

Th

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148 137

138

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

Except for the MTZ dunites for which modes were estimated from thin section observations (in italics), the peridotite modes were calculated with a total inversion method [58] applied to a mass-balance equation relating whole rock and mineral compositions in the CFMAS system. Mineral compositions (unpublished) were obtained by electron microprobe at Service Commun `Microsonde-Sud', Universite¨ Montpellier 2. Major elements and Cr in whole rocks were analysed by ICP^AES at CEREGE, Universite¨ Aix-Marseille III. Ni and trace elements (REE, Rb, Sr, Zr, Nb, Cs, Ba, Hf, Ta, Pb Th and U) were analysed by solution ICP^MS at ISTEEM, Universite¨ Montpellier 2. The analytical procedure and the accuracy of the ICP^MS analyses are given in the EPSL Online Background Dataset2 . Results for which the analytical error is greater than 30% are reported in italics. Analyses below the detection level or for which errors are greater than 66% were eliminated. Duplicate analyses were performed for several samples and the standard PCC-1 (numbers of duplicates are indicated in parentheses). These analyses were reproducible within the analytical error. MTZ = mantle^crust transition zone; diapir = diapir areas de¢ned by structural mapping (Fig. 1); Harz. = harzburgite; Cpx-harz. = cpx-harzburgite; F = foliation; L = lineation; opx = orthopyroxene; cpx = clinopyroxene; sp = spinel; LOI = loss on ignition; Mg# = cationic ratio: 100 Mg/(Mg+Fetotal ); n.a. = not analysed; RSD (%) = relative standard deviation (in percent) on duplicate PCC-1 analyses. 6

from 7^8 km below the Moho, and three are harzburgites from the low-temperature deformation zone (10^12 km below the Moho). These last three harzburgites are clinopyroxene-rich (cpx s 5%), compared with the other Oman harzburgite. In the following, they will be referred to as cpx-harzburgites. 4. Geochemistry of the Oman ophiolite peridotites The samples were analysed as whole rocks for major, minor (Ni, Cr) and trace elements (REE, Cs, Rb, Ba, Th, U, Pb, Nb, Ta, Zr and Hf). Trace element concentrations were determined on a VGPQ2 Turbo+ ICP^MS (ISTEEM, Montpellier) following the method described in the EPSL Online Background Dataset2 . Analytical results are given in Table 1. 4.1. Major and minor elements The analysed samples plot near the most depleted end of the mantle fractionation array on the MgO/SiO2 vs. Al2 O3 /SiO2 diagram (Fig. 2a). Similar to the most refractory abyssal peridotites [6,21,31], the Oman peridotites display low Al2 O3 / SiO2 ratios (0.01^0.03). They also have high MgO/SiO2 ratios typical of refractory mantle rocks and, except for the addition of water, do not display any of the major element variations

2

http://www.elsevier.nl/locate/epsl; mirror site: http://www.elsevier.com/locate/epsl

occasionally associated with oceanic alteration (e.g., MgO depletion [6,32]). The harzburgite composition in Fig. 2a is consistent with the evolutionary trend ^ parallel to the mantle array ^ shown by Southwest Indian Ridge and Western Alps peridotites. Similar to other suites of mantle rocks, the Oman peridotites show a broad positive correlation between Al2 O3 and CaO (Fig. 2b). However, in contrast to dominantly lherzolitic suites (e.g., Western Alps peridotites) that show nearly constant Al2 O3 /CaO ratios close to primitive mantle estimates, the Oman peridotites show important variations of this ratio. Al2 O3 /CaO tends to decrease with increasing fertility (i.e., cpx content) in the main harzburgite sequence, down to very low values (0.53^0.69) in the cpx-harzburgites from the base of the mantle section. In contrast, the harzburgites from the most actively upwelling areas in the Maqsad diapir (inner zones, with lineation dip s 40³), as well as the MTZ dunites and the harzburgite 96OF59B collected close to the Maqsad MTZ, are distinguished by higher Al2 O3 /CaO ratios (0.94^1.17). These samples are also enriched in TiO2 (0.07^0.08) compared with the other peridotites (TiO2 6 0.02). With the exception of the MTZ dunites, the studied peridotites have a narrow range of MgO and FeO contents (Fig. 2c). The Mg# ratio (100U(Mg+Fetotal )) of these samples (90.8 þ 0.3) does not show signi¢cant variations as a function of peridotite fertility, from the dunites of the main mantle section to the cpx-harzburgites. In contrast, the MTZ dunites display variable MgO and FeO contents, and lower Mg# values (87.8^

EPSL 5506 29-6-00

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

139

Fig. 2. Whole rock major element compositions of the studied Oman peridotites illustrated in (a) MgO/SiO2 vs. Al2 O3 /SiO2 , (b) Al2 O3 vs. CaO and (c) FeO vs. MgO diagrams. Published data are also shown for comparison: (1) SWIR: abyssal peridotites from the Southwest Indian Ridge [32]; (2) EPR: abyssal peridotites from the East Paci¢c Rise (Garrett fault zone [6] and Terevaka fault zone [31]); (3) WAO: Western Alps orogenic lherzolites and ophiolites (Lanzo [19] and Internal Ligurides [20]); (4) IBM forearc: drilled peridotites from the Izu^Bonin^Mariana forearc [21]. Whole rock compositions are recalculated on a volatile-free basis and FeO stands for total Fe content. Solid lines in panels a and b represent the silicate Earth di¡erentiation trend and the bulk silicate Earth ( = primitive mantle) Al2 O3 /CaO ratio (PM), respectively [55,56]. Dotted lines in panel b show constant Al2 O3 / CaO values and ¢ne grey lines in panel c show constant Mg# values. The dotted grey line in panel c shows the variation of olivine composition constrained by FeO+MgO = 66.67 mol%. L = lineation dip.

90.1) than the other peridotites. The MTZ samples are further distinguished by lower Ni contents ( 6 1500 ppm) than the other Oman peridotites (1923^2530 ppm, this study and unpublished data). 4.2. Trace elements The normalised REE and trace element patterns of the analysed samples are shown in Figs. 3 and 4, respectively. Similar to other ophiolitic refractory peridotites (e.g., [18,22,33]), the Oman peridotites are strongly depleted in lithophile trace elements, with whole rock concentrations well be-

low chondritic and primitive mantle (PM) values (mostly 6 0.1UPM). The harzburgites from the main mantle section display relatively linear REE patterns (Fig. 3a), characterised by a steady decrease of normalised REE abundances from heavy (HREE) to light REE (LREE). The harzburgites from the Maqsad diapir area are distinguished by concave-upward REE patterns (Fig. 3b), with a steep HREE segment from Gd to Lu, and a relatively £at LREE segment from La to Gd. In addition, diapir harzburgites are characterised by strikingly homogeneous HREE contents. It is worth noting that the two harzburgites from the Nakhl diapir area

EPSL 5506 29-6-00

140

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

Fig. 3. Chondrite-normalised [36] REE patterns for the studied Oman peridotites: (a) harzburgites from the main harzburgite section and the Nakhl diapir area, and harzburgite sample 96OF59B from the vicinity of the MTZ in the Maqsad area (Fig. 1); (b) harzburgites from the Maqsad diapir; (c) dunites; (d) cpx-harzburgites.

and the harzburgite 96OF59B collected close to the Maqsad MTZ (Fig. 3a) also have concaveupward REE patterns, very similar in shape to those observed in the Maqsad diapir. The dunites have almost linear or slightly concave REE patterns (Fig. 3c), without signi¢cant di¡erences between dunites from the main mantle section and those from the MTZ. The cpx-harzburgites are clearly distinguished from the other peridotites by `spoon-shaped' REE patterns. They show a convex-upward distribution of middle REE (MREE) and HREE, from Nd to Lu, and a slight enrichment of La and Ce. These REE patterns are very similar in shape to those obtained by Kelemen et al. [5] on clinopyroxenes from the same samples (OMB14, OMB23 and OMB27 analysed by SIMS). In addition, the cpx-harzburgites display the highest HREE contents (Yb = 0.4Uchondrites), but almost the lowest LREE abundances (Ce = 0.004^0.015Uchondrites), among the studied peridotites.

Similar to refractory peridotites from other mantle suites [14,21,22,34,35], the Oman peridotites are characterised by U-shaped normalised trace element patterns (Fig. 4). These patterns re£ect the depletion of LREE, Zr and Hf relative to HREE, and the selective enrichment of highly incompatible elements (Cs to Ta) relative to LREE (e.g., Th/Ce = 0.9^9.1UPM). However, the trace element patterns show prominent Sr spikes which are re£ected in extremely high Sr/Ce ratios (up to 390UPM). In addition, U, Nb and Ta are enriched relative to the neighbouring elements Th and LREE. The U/Th ratio, for instance, is in the range of 1.1^5.3UPM. The Nb/Ta ratio is generally close to PM estimates (20.7 þ 2.6, compared with PM = 17.6 [36]), except for the inner zone of the Maqsad diapir and the two samples from the Nakhl diapir area. Being strongly depleted in Ta relative to Nb, these samples are distinguished by high Nb/Ta values, up to 78 (Figs. 4 and 5). Zr shows variable, but systematic,

EPSL 5506 29-6-00

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

141

depletion relative to MREE and Hf. The Zr/Hf ratio (10.8^25.3) is lower than PM estimates (36.25 [36]). 5. Discussion The mantle unit of the Oman ophiolite is composed of refractory peridotites with extremely low trace element contents. However, the samples are much less depleted in LREE relative to MREE and HREE than generally expected for mantle residues after partial melting (e.g., [17,18,33,37]). Some samples from the diapir areas and the cpxharzburgites from the base of the mantle unit even show subtle LREE enrichment on normal-

Fig. 5. Variations of (a) Al2 O3 /CaO, (b) TiO2 content, (c) chondrite-normalised Ho/Yb and (d) Nb/Ta vs. lineation dip in the analysed Oman peridotites. PM = primitive mantle values [36,55,56].

Fig. 4. PM-normalised [36] trace element patterns for the studied Oman peridotites: (a) harzburgites and dunites from the main harzburgite section, harzburgites from the outer zone of the Maqsad diapir and cpx-harzburgites; (b) harzburgites from the inner zone of the Maqsad diapir and from the Nakhl diapir area, MTZ dunites and harzburgite sample 96OF59B from the vicinity of the MTZ in the Maqsad area (Fig. 1).

ised diagrams (Fig. 3). Furthermore, all samples are enriched in highly incompatible elements (Rb to Ta) relative to LREE (Fig. 4). LREE enrichment in ophiolitic peridotites is generally attributed to secondary processes such as serpentinisation, oceanic alteration or contamination by crustal £uids during ophiolite obduc-

EPSL 5506 29-6-00

142

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

tion [33,37]. However, similar to geochemical parameters hardly a¡ected by alteration (e.g., MREE/HREE ratio, and Al2 O3 and TiO2 concentrations), LREE/MREE and LREE/HREE ratios in Oman peridotites show variations related to mantle structures. The (Ce/Sm)N ratio, for instance, is in the range 1.35^1.43 in the Maqsad diapir area, compared with 0.65^1.11 in the main harzburgite section. Clearly, such a relationship is not consistent with secondary contamination. In addition, trace element enrichment involves elements such as Th, Nb and Ta, which are reputedly immobile during alteration. In fact, the paradox of incompatible trace element enrichment in refractory peridotites is not speci¢c to the ophiolites. Virtually all harzburgitic mantle rocks for which trace element data are available show similar enrichment marked by U-shaped trace element patterns ([34] and references therein, [14,21,22]). In several cases, the enrichment is observed in peridotites that are totally devoid of alteration, as well as in acid-leached mineral separates [35,38]. For these reasons, we consider that the U-shaped trace element patterns of the analysed samples (Fig. 4) represent a mantle signature. In detail, however, the distribution of the alkaline and alkaline-earth elements (Cs, Rb, Ba and Sr), and particularly the positive Sr anomaly, might result from secondary processes. Yet, no correlation is observed between the serpentinisation degree and, for instance, Sr/Ce. Based on the distribution of the less mobile elements, including REE, our data de¢ne three geochemical domains, broadly coinciding with major structural features: (1) a main harzburgite section, (2) a diapiric domain, including the Maqsad and Nakhl diapir harzburgites as well as the MTZ dunites, and (3) a basal section constituted of cpx-harzburgites. In the following, we discuss the role of melt processes in the individualisation of these three domains, as well as their geodynamic signi¢cance. 5.1. The main harzburgite section: melt extraction versus melt transport Away from diapir areas and at some distance ( s 2 km) above the base of the ophiolite slab, the

Fig. 6. Modal compositions of the analysed Oman peridotites plotted on a cpx/opx vs. olivine diagram. The symbols are the same as in Fig. 5. Published melting models are also shown for comparison: (1) polybaric melting after Niu [40], with (1a) or without (1b) `excess olivine'; (2) isobaric melting after Walter et al. [39], at 11 (2a), 16 (2b) and 17 kbar (2c). The initial modal composition is given by Niu [40] for polybaric melting and was ¢xed as follows for isobaric melting: 55% olivine, 28% opx, 15% cpx and 2% spinel. Numbers refer to percent melting degrees.

main harzburgite section is composed of refractory peridotites with relatively homogeneous modal and major element compositions. To account for these compositions, melting of a pristine mantle source requires a high degree of melt extraction, in the range 15^30% (Fig. 6 [39,40]). Such values are signi¢cantly higher than melting degrees inferred for MORB (5^15% [2]) and highTi magmas such as those forming the dyke complex and the MORB-like Geotimes lava sequence in Oman [25]. Present-day intermediate and fastspreading ridges show a similar discrepancy between melting degree inferred from residual mantle and that inferred from lava (e.g., [6,16]). This paradox is generally solved by considering that abyssal peridotites represent only the shallowest part of the mantle column a¡ected by partial melting and thus record the highest melting degrees. In contrast, MORB are thought to represent integrated melts from the whole melting column and thus record average melting degrees. In

EPSL 5506 29-6-00

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

Oman, however, the main harzburgite section does not show any change that would indicate decreasing melting degree at depth, down to about 10 km from the Moho. In terms of cpx content, the basal section of the mantle unit is more fertile than the main harzburgite section. However, as discussed below, this feature results from secondary fertilisation rather than from a lesser degree of melt extraction. In fact, melt extraction fails to account for two notable features of the Oman harzburgites : (1) their relatively linear REE patterns (Fig. 3), with no trace of the strong LREE depletion predicted by the standard melting models (e.g., [8]) and (2) the high olivine proportion (75^85%) observed in several samples (Fig. 6). The relatively unfractionated REE distribution may simply result from melt transport through molten peridotites (e.g., [10^12,41]), a process which was also referred to as `open-system melting' [42]. During this process, the most incompatible elements (LREE) tend to be `bu¡ered' by percolating melt while the less incompatible elements (HREE), similar to major elements, are governed by melting reactions [13,41,43]. As noted by Vernie©res et al. [41], this process would account for negative correlations between LREE/HREE ratios and peridotite fertility, as commonly observed in heterogeneous peridotite suites ([34] and references therein). However, melt transport alone cannot explain olivine proportions higher than predicted by melting reactions [39,40]. This feature is indicative of olivine-forming reactions between residual peridotites and in¢ltrated melts ([3,44] and references therein). With respect to incompatible trace elements, reactive porous £ow at increasing melt mass is virtually indistinguishable from partial melting coupled with melt transport. On the other hand, these two processes can be readily distinguished from standard melting models, as well as from reactions at decreasing melt mass [41]. In Fig. 7 are compared two numerical experiments performed with the plate model of Vernie©res et al. [41]. The ¢rst one (Fig. 7a) simulates standard incremental melting [8] and produces strongly LREE-depleted peridotite residues, quite di¡erent in shape from the Oman harzburgites. In this ex-

143

periment, the presence of trapped melt in the residue ^ due to ine¤cient melt extraction ^ does not provide a better ¢t to the data (Fig. 7c). The second experiment (Fig. 7b) simulates reactive porous £ow at increasing melt mass and produces peridotite residues that are only moderately depleted in LREE. Moreover, with a small amount of trapped melt (0.5^1%), this experiment generates linear REE patterns comparable to those of the main harzburgite section. These results suggest that the Oman harzburgites were pervasively traversed by di¡use melt £ow. Porous £ow associated with partial melting would account for the observed REE patterns. However, the harzburgites probably underwent more reactive melt £ow at a shallow level, which contributed to their olivine enrichment. Crosscutting dunites are widespread throughout the main harzburgite section and would represent end products of the inferred olivine-forming reaction [44]. The dunites were probably individualised by channelled percolation [3,5] at the top of the melting column. It should nevertheless be noted that interactions of the harzburgites with MORB melts could hardly explain the selective enrichment of highly incompatible elements in these rocks (Fig. 4). The enrichment of Th, Nb and Ta relative to LREE, for instance, suggests that the mantle section was belatedly percolated by small fractions of volatilerich £uids (see below). 5.2. Diapir domains : high melt £ow areas 5.2.1. Mantle diapirs: instabilities created by partial melt upwelling The Maqsad diapir area is distinguished from the main harzburgite section by concave-upward REE patterns, and low and homogeneous HREE content (Fig. 5a). Moreover, the inner zone of the diapir (lineation dip s 40³) is characterised by elevated Al2 O3 /CaO and Nb/Ta ratios, and high TiO2 contents (Fig. 5b^d). However, while being closer to PM values in terms of Al2 O3 /CaO, the diapir shows a signi¢cant departure from PM with respect to Nb/Ta. Their homogeneous REE composition indicates that the diapir harzburgites were extensively re-

EPSL 5506 29-6-00

144

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

Fig. 7. (Top) `Plate model' of Vernie©res et al. [41] applied to the simulation of REE variations in a peridotite a¡ected by partial melting with (a) or without (b) melt in¢ltration. The chondrite-normalised REE patterns of the Oman harzburgites (main harzburgite section) are shown for comparison. Model (a) simulates standard incremental melting [8] whereas in model (b) the molten peridotites are percolated by a melt of ¢xed, N-MORB composition [36]. In this respect, model (b) is comparable to reactive melt £ow at increasing melt mass [41]. The initial modal composition was ¢xed as follows (spinel neglected): 57% olivine, 28% opx and 15% cpx. The melting reaction was taken from Walter et al. [39] (model 2a in Fig. 6), but other reactions provide similar results. Mineral/melt partition coe¤cients are the same as those selected by Bedini and Bodinier [57]. Numbers on the chondrite-normalised REE patterns indicate olivine proportion (in percentage) in residual peridotites. Thicker lines indicate the REE patterns of the less residual peridotites. In model (a), the most residual peridotite (76% olivine) is produced after 21.1% melt extraction. In model (b), the ratio of in¢ltrated melt to peridotite varies from 0.02 to 0.19. (Bottom) Modi¢cations of the REE patterns of residual peridotites due to the presence of equilibrium, trapped melt. Models (c) and (d) show the e¡ect of trapped melt on the most residual peridotites of models (a) and (b), respectively (thicker solid lines). Numbers on the REE patterns indicate the proportions of trapped melt (in percentage).

equilibrated with a pervasive melt [10,13,41]. Likewise, the distinctly higher Al2 O3 /CaO ratio and TiO2 content of the inner zone are indicative of relatively pristine melt composition, suggesting that the diapir has focused partial melt upwelling. A similar interpretation was recently proposed for refractory peridotites characterised by a high Ti content in minerals [23]. Nb/Ta fractionation is not well understood in this scheme. A possible interpretation is to consider that the melt that £owed through the diapir was enriched in volatiles [22,45^47].

Compositional di¡erences between the Maqsad diapir and the main harzburgite section suggest that the diapir intruded a pre-existing oceanic lithosphere, diapiric upwelling being probably associated with the opening of the NW^SE ridge system. Although radial lineations around the Maqsad diapir suggest a widespread radial solid £ow outwards [48], only the MTZ dunites and three harzburgite samples collected in the vicinity of the Maqsad and Nakhl diapirs, close to the MTZ, have geochemical compositions akin to the inner zone of the Maqsad diapir (i.e., high

EPSL 5506 29-6-00

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

Al2 O3 /CaO ratio and TiO2 content, and fractionated Nb/Ta ratio and HREE segment). Therefore, outward £ow from diapirs was probably restricted to the shallowest part of the mantle section (mostly the MTZ), with limited extension away from diapir areas. Dynamic models of mantle upwelling have pointed to the major role of feedback relationships between melt and solid £ows (e.g., [49]). The distinctive geochemical composition of the diapirs suggests that they represent instabilities triggered by partial melting and were probably rooted in a larger molten region that is not seen in the ophiolite. 5.2.2. Mantle^crust transition zone: melt accumulation at the Moho The MTZ dunites are distinguished from the other Oman peridotites by lower Ni contents, and lower and more variable Mg# ratios (87.8^ 90.1, compared with 90.3^91.3 ^ Fig. 2). In addition, they often show cpx and plagioclase impregnation textures (although not in the studied samples ^ Table 1). Because of these features, the MTZ was ¢rst interpreted as the lowest part of the crustal cumulate sequence [25]. However, the MTZ olivine crystals display the same high-temperature deformation as in the other Oman peridotites. Moreover, the low Ni and Mg# values, as well as the impregnation textures, can be explained by peridotite^melt interaction at a high melt/rock ratio (e.g., [50]). Melt^rock reactions would involve olivine precipitation (to form the dunites), followed by melt consumption (to form the impregnated dunites). Our data indicate that the MTZ dunites share with the Maqsad diapir several major and trace element characteristics, such as high Al2 O3 /CaO and Nb/Ta ratios, and high TiO2 contents. Together with the structural observations of Boudier and Nicolas [29], this suggests that the MTZ originates from the inner part of the diapirs, strongly modi¢ed by low-pressure melt^rock reactions. Melt accumulation at the Moho would focalise outward solid £ow from the diapirs against the mantle^crust boundary. In turn, deformation and/or melt^rock reaction might enhance the

145

role of the MTZ as a collecting layer for partial melts tapped from mantle diapirs. 5.3. Cpx-harzburgites: melt freezing at the asthenosphere^lithosphere boundary In addition to their location at the base of the ophiolite slab, the cpx-harzburgites are distinguished by high cpx contents ( s 5% ^ Table 1) and distinctive `spoon-shaped' REE patterns (Fig. 3). The presence of less refractory peridotites in the lower mantle section is relatively common in harzburgitic ophiolites. It was described in the eastern Mediterranean Tethyan ophiolites [51] and in the Bay of Islands ophiolite, in Newfoundland, for instance [15,23]. These rocks are generally ascribed to a lower degree of melt extraction, the deeper peridotites being supposedly less affected by pressure-release partial melting [15,23]. However, the Oman harzburgites do not show a gradual decrease of their refractory character from the top to the base of the mantle section, as would be expected in this model. Instead, the mantle sequence shows a rapid transition (within less than a few kilometres) from the main harzburgite section to the basal cpx-harzburgites. Besides, the cpx-harzburgites are virtually indistinguishable from the other Oman harzburgites with respect to their Mg# ratios and olivine proportions (Figs. 2 and 6). In fact, their more `fertile' character is mainly re£ected in higher cpx/opx ratios (Fig. 6), suggesting that they were individualised from the other harzburgites by a melt^ rock reaction involving precipitation of cpx at the expense of opx [52]. A similar reaction was recently described in the Ronda massif and interpreted as a near-solidus melt^freezing reaction occurring at the boundary of a partial melting domain developed at the expense of the lithospheric mantle [53]. A similar scenario may be envisioned for the formation of the cpx-harzburgites at the base of the Oman ophiolite, which is thought to have preserved the base of the oceanic lithosphere [30]. In this scheme, cpx precipitation in the Wadi Tayin basal harzburgites would underline the lower boundary of pre-existing oceanic lithosphere thermally eroded by upwelling, partially molten

EPSL 5506 29-6-00

146

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

asthenosphere. This process would be related to the opening of the NW^SE propagator identi¢ed in the Oman ophiolite [27]. Several studies have proposed that melt-consuming reactions occurring at the base of the lithosphere would release volatile- and LILE-enriched small melt fractions that may pervasively in¢ltrate large volumes of relatively cold lithospheric mantle [38,54]. Although this process was generally invoked for subcontinental lithosphere, it is strongly suggested in Oman by the existence of fertilised harzburgites at the base of the ophiolite and the pervasive enrichment of highly incompatible elements observed throughout the mantle section. 6. Conclusion The mantle section of the Oman ophiolite exposes refractory peridotites characterised by relatively homogeneous modal and major element compositions. However, trace elements display signi¢cant variations related to geological structures (revealed by high-temperature mantle deformation) and distance from the Moho. Our data distinguish three structural and geochemical domains and provide constraints on their origin : 1. The main harzburgitic mantle section is composed of strongly refractory harzburgites interpreted as mantle residues after s 15% melt extraction. Their REE signature can be explained by melt transport associated with partial melting. Their olivine-rich composition may partly result from an olivine-forming melt^rock reaction at shallow level. 2. The diapir areas (mainly the Maqsad diapir) show distinctive trace element compositions ascribed to focused upwelling of partial melts. They are interpreted as instabilities in upwelling mantle, triggered by partial melting and probably enhanced by feedback relationships between deformation and melt percolation. The diapirs are topped by a thick, dunitic, MTZ that displays the same trace element signature as the diapirs. The MTZ is considered to be a major collecting zone for mantle melts.

3. The cpx-rich harzburgites from the base of the mantle section were individualised from the rest of the harzburgite mantle section by a cpx-forming, melt^rock reaction. This reaction probably occurred as a near-solidus melt^freezing reaction at the boundary between oceanic lithosphere and partially molten, upwelling mantle. These three domains may be integrated in a geodynamic scenario involving the reactivation of an oceanic lithosphere, a process that would be related to the ridge propagator identi¢ed in the Oman ophiolite. The main harzburgite section would represent pre-existing oceanic lithosphere whereas the diapir areas would represent newly accreted oceanic lithosphere. In this scheme, the cpx-harzburgites would underline the base of the pre-existing lithosphere thermally eroded by upwelling mantle and fertilised by partial melts via a melt-consuming reaction. Small-volume melts, residual after this reaction, probably percolated throughout the oceanic lithosphere and would account for pervasive enrichment of highly incompatible elements in the whole mantle section. Acknowledgements This work has bene¢ted from discussions with R. Bedini, F. Boudier, G. Ceuleneer, C. Dupuy, Ph. Gouze and A. Nicolas. Financial support was provided by INSU-CNRS Project 97IT45. We thank S. Pourtales and L. Savoyant for technical assistance on ICP^MS, Y. Niu, P. Kelemen and an anonymous reviewer for helpful reviews and F. Albare©de for editorial handling.[FA]

References [1] D.L. Turcotte, J. Phipps Morgan, The physics of magma migration and mantle £ow beneath a mid-ocean ridge, in: J. Phipps Morgan, D.K. Blackman, J.M. Sinton (Eds.), Mantle Flow and Melt Generation at Mid-Ocean Ridges, American Geophysical Union Monograph 71, AGU, Washington, DC, 1992, pp. 155^182. [2] C.H. Langmuir, E.M. Klein, T. Plank, Petrological systematics of mid-ocean ridge basalts: constraints on melt

EPSL 5506 29-6-00

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148

[3]

[4] [5]

[6]

[7] [8]

[9]

[10] [11]

[12] [13]

[14]

[15]

[16]

generation beneath ocean ridges, in: J. Phipps Morgan, D.K. Blackman, J.M. Sinton (Eds.), Mantle Flow and Melt Generation at Mid-Ocean Ridges, American Geophysical Union Monograph 71, AGU, Washington, DC, 1992, pp. 183^280. P.B. Kelemen, J.A. Whitehead, E. Aharonov, K.A. Jordhal, A review of melt migration processes in the adiabatically upwelling mantle beneath oceanic spreading ridges, Phil. Trans. R. Soc. London A 355 (1997) 283^ 318. A. Nicolas, Structure of Ophiolites and Dynamics of Oceanic Lithosphere, Kluwer Academic Press, Boston, MA, 1989. P.B. Kelemen, N. Shimizu, V.J.M. Salters, Extraction of mid-ocean-ridge basalt from the upwelling mantle by focused £ow of melt in dunite channels, Nature 375 (1995) 747^753. Y. Niu, R. Hekinian, Basaltic liquids and harzburgitic residues in the Garrett Transform; a case study at fastspreading ridges, Earth Planet. Sci. Lett. 146 (1997) 243^ 258. P.D. Asimow, A model that reconciles major- and traceelement data from abyssal peridotites, Earth Planet. Sci. Lett. 169 (1999) 303^319. K.T.M. Johnson, H.J.B. Dick, N. Shimizu, Melting in the oceanic upper mantle: an ion microprobe study of diopsides in abyssal peridotites, J. Geophys. Res. 95 (1990) 2661^2678. R.T. Gregory, Melt percolation beneath a spreading ridge: evidence from the Semail peridotite, Oman, in: I.G. Gass, S.J. Lippard, A.W. Shelton (Eds.), Ophiolites and Oceanic Lithosphere, 13, Geological Society Special Publications, 1984. O. Navon, E. Stolper, Geochemical consequences of melt percolation: The upper mantle as a chromatographic column, J. Geol. 95 (1987) 285^307. J.-L. Bodinier, G. Vasseur, J. Vernie©res, C. Dupuy, J. Fabrie©s, Mechanisms of mantle metasomatism: Geochemical evidence from the Lherz Orogenic peridotite, J. Petrol. 31 (1990) 597^628. E. Takazawa, F.A. Frey, N. Shimizu, M. Obata, J.-L. Bodinier, Geochemical evidence for melt migration and reaction in the upper mantle, Nature 359 (1992) 55^58. M. Godard, J.-L. Bodinier, G. Vasseur, E¡ects of mineralogical reactions on trace element redistributions in mantle rocks during percolation processes: a chromatographic approach, Earth Planet. Sci. Lett. 133 (1995) 449^461. D. Van der Wal, J.-L. Bodinier, Origin of the recrystallisation front in the Ronda peridotite by km-scale pervasive porous melt £ow, Contrib. Mineral. Petrol. 122 (1996) 387^405. V.G. Batanova, G. Suhr, A.V. Sobolev, Origin of geochemical heterogeneity in the mantle peridotites from the Bay of Islands Ophiolite (Newfoundland, Canada): Ion probe study of clinopyroxenes, Geochim. Cosmochim. Acta 62 (1998) 853^866. H.J.B. Dick, J.H. Natland, Late-stage melt evolution and

[17] [18] [19] [20]

[21]

[22]

[23]

[24] [25] [26] [27] [28] [29] [30]

[31] [32]

147

transport in the shallow mantle beneath the East Paci¢c Rise, in: C. Mevel, K.M. Gillis, J.F. Allan et al. (Eds.), Hess Deep rift valley; Leg 147, sites 894^895, Proceedings of the Ocean Drilling Program, Scienti¢c Results 147, College Station, TX, 1996, pp. 103^134. J. Pallister, R.J. Knight, Rare earth element geochemistry of the Samail ophiolite near Ibra, Oman, J. Geophys. Res. 86 (1981) 2673^2697. A. Prinzhofer, C.J. Alle©gre, Residual peridotites and the mechanisms of partial melting, Earth Planet. Sci. Lett. 74 (1985) 251^265. J.-L. Bodinier, Geochemistry and petrogenesis of the Lanzo peridotite body, Western Alps, Tectonophysics 149 (1988). E. Rampone, A.W. Hofmann, G.B. Piccardo, R. Vannucci, P. Bottazzi, L. Ottolini, Trace element and isotope geochemistry of depleted peridotites from an N-MORB type ophiolite (Internal Liguride, N. Italy), Contrib. Mineral. Petrol. 123 (1996) 61^76. I.J. Parkinson, J.A. Pearce, Peridotites from the Izu-Bonin-Mariana Forearc (ODP Leg 125): Evidence for mantle melting and melt-mantle interaction in a supra-subduction zone setting, J. Petrol. 39 (1998) 1577^1618. J. Proenza, F. Gervilla, J.C. Melgarejo, J.-L. Bodinier, Alrich and Cr-rich chromitites from the Mayari-Baracoa ophiolitic belt (Eastern Cuba) as the consequence of interaction between volatile-rich melts and peridotites in suprasubduction mantle, Econ. Geol. 94 (1999) 547^566. G. Suhr, P. Robinson, Origin of mineral chemical strati¢cation in the mantle section of the Table Mountain massif (Bay of Islands Ophiolite, Newfoundland, Canada), Lithos 31 (1994) 81^102. F. Boudier, R.G. Coleman, Cross section through the peridotite in the Samail ophiolite, southeastern Oman mountains, J. Geophys. Res. 86 (1981) 2573^2592. S.J. Lippard, A.W. Shelton, I.G. Gass, The ophiolite of northern Oman, in: The Geological Society Memoirs 11, Backwell Scienti¢c Publications, London, 1986, 178 pp. A. Nicolas, F. Boudier, Mapping oceanic ridge segments in Oman ophiolites, J. Geophys. Res. 100 (1995) 6179^ 6197. F. Boudier, A. Nicolas, B. Ildefonse, D. Jousselin, EPR microplates, a model for the Oman ophiolite, Terra Nova 9 (1997) 79^82. G. Ceuleneer, A. Nicolas, F. Boudier, Mantle £ow patterns at an oceanic spreading centre; the Oman peridotites record, Tectonophysics 151 (1988) 1^26. F. Boudier, A. Nicolas, Nature of the Moho transition zone in the Oman ophiolite, J. Petrol. 36 (1995) 777^796. F. Boudier, G. Ceuleneer, A. Nicolas, Shear zones, thrusts and related magmatism in the Oman Ophiolite; initiation of thrusting on an oceanic ridge, Tectonophysics 151 (1988) 275^296. M. Constantin, Pe¨trologie des gabbros et pe¨ridotites de la dorsale Est-Paci¢que: La transition croute-manteau aux dorsales rapides, PhD Thesis, Bretagne Occidentale, 1995. J.E. Snow, H.J.B. Dick, Pervasive magnesium loss by ma-

EPSL 5506 29-6-00

148

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40] [41]

[42]

[43]

[44]

M. Godard et al. / Earth and Planetary Science Letters 180 (2000) 133^148 rine weathering of peridotite, Geochim. Cosmochim. Acta 59 (1995) 4219^4235. M. Sharma, G.J. Wasserburg, The neodymium isotopic compositions and rare earth patterns in highly depleted ultrama¢c rocks, Geochim. Cosmochim. Acta 60 (1996) 4537^4550. W.F. McDonough, F.A. Frey, Rare earth elements in upper mantle rocks, in: B.R. Lipin, G.A. McKay (Eds.), Geochemistry and Mineralogy of Rare Earth Elements, Reviews in Mineralogy 21, The Mineralogical Society of America, Washington, DC, 1989, pp. 99^145. O. Alard, J.-L. Bodinier, X. Lenoir, J.-M. Dautria, Uranium enrichment in the lithospheric mantle: Case studies from French Massif Central, in: 8th Annual V.M. Goldschmidt Conference 62A, Mineral. Mag, Toulouse (France), September 1998, 1998, pp. 31^32. S.S. Sun, W.F. McDonough, Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes, in: A.D. Saunders, M.J. Norry (Eds.), Magmatism in the Ocean Basins, 42, Geological Society of London, London, 1989, pp. 313^345. G. Gruau, J. Bernard-Gri¤ths, C. Lecuyer, The origin of U-shaped rare earth patterns in ophiolite peridotites: assessing the role of secondary alteration and melt/rock reaction, Geochim. Cosmochim. Acta 62 (1998) 3545^3560. R.-M. Bedini, J.-L. Bodinier, J.M. Dautria, L. Morten, Evolution of LILE-enriched small melts fractions in the lithospheric mantle: a case study from the East African Rift, Earth Planet. Sci. Lett. 153 (1997) 67^83. M.J. Walter, T.W. Sisson, D.C. Presnall, A mass proportion method for calculating melting reactions and application to melting of model upper mantle lherzolite, Earth Planet. Sci. Lett. 135 (1995) 77^90. Y. Niu, Mantle melting and melt extraction processes beneath ocean ridges; evidence from abyssal peridotites, J. Petrol. 38 (1997) 1047^1074. J. Vernie©res, M. Godard, J.-L. Bodinier, A plate model for the simulation of trace element fractionation during partial melting and magma transport in the Earth's upper mantle, J. Geophys. Res. 102 (1997) 24771^24784. K. Ozawa, N. Shimizu, Open-system melting in the upper mantle: constraints from the Hayachine-Miyamori ophiolite, northeastern Japan, J. Geophys. Res. 100 (1995) 22315^22335. J.-L. Bodinier, M.A. Menzies, M.F. Thirlwall, Continental to oceanic mantle transition ^ REE and Sr-Nd isotopic geochemistry of the Lanzo lherzolite massif, J. Petrol. special volume `Orogenic Lherzolites and Mantle Processes' (1991) 191^210. P.B. Kelemen, Reaction between ultrama¢c rock and fractionating basaltic magma; I, Phase relations, the origin of calc-alkaline magma series, and the formation of discordant dunite, J. Petrol. 31 (1990) 51^98.

[45] D.H. Eggler, Solubility of major and trace elements in mantle metasomatic £uids: Experimental constraints, in: M.A. Menzies, C.J. Hawkesworth (Eds.), Mantle Metasomatism, Geology Series, Academic Press London, London, 1987, pp. 21^41. [46] P. Schiano, R. Clocchiatti, Worldwide occurrence of silica-rich melts in sub-continental and sub-oceanic mantle minerals, Nature 368 (1994) 621^624. [47] J.-L. Bodinier, C. Merlet, R.M. Bedini, F. Simien, M. Rema|«di, C.J. Garrido, Distribution of Nb, Ta and other highly incompatible trace elements in the lithospheric mantle: the spinel paradox, Geochim. Cosmochim. Acta 60 (1996) 545^550. [48] D. Jousselin, A. Nicolas, F. Boudier, Detailed mapping of a mantle diapir below a paleo-spreading center in the Oman ophiolite, J. Geophys. Res. 103 (1998) 153^170. [49] W.R. Buck, W. Su, Focused mantle upwelling below midocean ridges due to feedback between viscosity and melting, Earth Planet. Sci. Lett. 16 (1989) 641^644. [50] G. Suhr, Melt migration under oceanic ridges: inference from reactive transport modeling of upper mantle hosted dunites, J. Petrol. 40 (1999) 575^600. [51] M. Menzies, C. Allen, Plagioclase lherzolite-residual mantle relashionships within two eastern Mediterranean ophiolites, Contrib. Mineral. Petrol. 45 (1974) 197^213. [52] J. Fabrie©s, J.-P. Lorand, J.-L. Bodinier, C. Dupuy, Evolution of the upper mantle beneath the Pyrenees: Evidence from orogenic spinel lherzolite massifs, J. Petrol. special volume `Orogenic lherzolites and mantle processes', 1991. [53] X. Lenoir, C. Garrido, J.-L. Bodinier, J.-M. Dautria, F. Gervilla, The recrystallization front of the Ronda peridotite: evidence for melting and thermal erosion of lithospheric mantle beneath the Alboran basin, J. Petrol., in press. [54] P.B. Kelemen, J.A. Whitehead, E. Aharonov, K.A. Jordahl, Experiments on £ow focusing in soluble porous media, with applications to melt extraction from the mantle, J. Geophys. Res. 100 (1995) 475^496. [55] E. Jagoutz, H. Palme, H. Baddenhausen, K. Blum, M. Cendales, G. Dreibus, B. Spettel, V. Lorenz, H. Vanke, The abundance of major, minor and trace elements in the earth's mantle as derived from primitive ultrama¢c nodules, Geochim. Cosmochim. Acta 11 (1979) 2031^2050. [56] S.R. Hart, A. Zindler, In search of a bulk-Earth composition, Chem. Geol. 57 (1986) 247^267. [57] R.-M. Bedini, J.-L. Bodinier, Distribution of incompatible trace elements between the constituents of spinel peridotite xenoliths: ICP-MS data from the East African Rift, Geochim. Cosmochim. Acta 63 (1999) 3883^3900. [58] A. Tarantola, B. Valette, Generalized non-linear inverse problems solved using least-squares criterion, Rev. Geophys. Space Phys. 20 (1982) 219^232.

EPSL 5506 29-6-00