Earth and Planetary Science Letters 437 (2016) 138–149

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

www.elsevier.com/locate/epsl

Major and trace element and Sr and Nd isotopic results from mantle

diapirs in the Oman ophiolite: Implications for off-axis magmatic

processes

Marie Nicolle a,∗, David Jousselin a, Laurie Reisberg a, Delphine Bosch b, Aurore Stephant a

a Université de Lorraine, CRPG, 54501, Vandoeuvre-lès-Nancy, Franceb Université Montpellier II, Géosciences Montpellier, UMR-CNRS 5243, 34095 Montpellier cedex 5, France

a r t i c l e i n f o a b s t r a c t

Article history:Received 5 August 2014Received in revised form 8 December 2015Accepted 14 December 2015Available online 20 January 2016Editor: T. Elliott

Keywords:mantleophioliteoff-axis magmatismpyroxenitemajor and trace elementsSr and Nd isotopes

The Oman ophiolite includes both a fossil fast spreading axis, defined by five mantle diapirs, and an off-axis mantle diapir emplaced 30 km from the axis, providing a natural laboratory for the study of off-axis magmatic processes. We compare field and petrological observations coupled with geochemical and isotopic analyses of samples from the off-axis diapir with those of the nearest on-axis diapir, with a particular focus on the Moho Transition Zone (MTZ). Both diapirs are defined by the presence of steeply plunging lineations, but in the on-axis case, these lineations rotate gradually into parallelism with the horizontal magmatic lineations of the overlying crust, while in the off-axis case, a shear zone separates the steeply plunging lineations from the horizontal lineations of the surrounding mantle. In the on-axis diapir, the MTZ is 50 to 500 m thick and composed of dunite with layered gabbro lenses whereas in the off-axis diapir, the MTZ is thicker and composed of dunite with massive (∼20% of MTZ) clinopyroxenite lenses and a notable absence of plagioclase. Moreover, the off-axis diapir is associated with amphibole-bearing intrusions, consisting of Mg-rich gabbroic sills in the mantle peripheral to the diapir, and microgabbroic lenses of broadly basaltic composition in the overlying crust. The εNd values of the pyroxenites in the MTZ of the off-axis diapir fully overlap with those of the intrusions in the surrounding mantle and crust, suggesting that they are genetically related. Calculated rare earth element (REE) abundances of liquids in equilibrium with clinopyroxene imply that the magmas that traversed the MTZ of the off-axis diapir were more depleted in highly incompatible elements than their counterparts in the MTZ of the on-axis diapir. On the other hand, Nd isotopic compositions of the off-axis samples (εNd = 6.2–7.9 in 18 of 19 samples) indicate derivation of their parental magmas from a less depleted source than that which produced the magma associated with the on-axis gabbro (εNd = 7.8–9.2, 10 analyses).To explain these observations, we suggest that the earliest magmas in the uprising off-axis diapir formed from the partial melting of pyroxenite veins with less radiogenic Nd isotopic compositions than those of the ambient peridotite. As the diapir traversed the cool, hydrated lithosphere these early melts interacted with depleted harzburgites, lowering the incompatible element contents of the melt products while having little effect on their Nd isotopic compositions. The great abundance of clinopyroxene in the off-axis MTZ might be explained by the high pyroxene component in the original melt but perhaps also by the presence of water in the lithosphere, which would favor the crystallization of clinopyroxene while inhibiting that of plagioclase. The intrusions in the overlying crust could represent, to first order, the secondary melts produced by the melt–harzburgite reaction, while the sills in the surrounding mantle may be cumulates from such secondary melts.These results shed light on processes occurring during interaction between rising off-axis material and depleted, hydrated lithospheric mantle. Furthermore, if our interpretation is correct, the low εNd values of the off-axis samples contribute to the growing body of evidence for the presence of pyroxenite veins in the MORB mantle source.

© 2015 Elsevier B.V. All rights reserved.

* Corresponding author.

E-mail address: mnicolle@crpg.cnrs-nancy.fr (M. Nicolle).

http://dx.doi.org/10.1016/j.epsl.2015.12.0050012-821X/© 2015 Elsevier B.V. All rights reserved.

1. Introduction

Increasing evidence suggests that magmatic activity beneath

mid-ocean ridges is not limited to the ridge axis. In the study of

M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149 139

Fig. 1. Map of the southern part of the Oman ophiolite showing the locations of the on-axis Maqsad and off-axis Mansah diapirs (open rectangles). As seen most clearly in the Sumail massif, a system of dikes formed in a NW-SE ridge system crosscuts an older dike system oriented NE-SW. Plunging lineations (indicated by zones contoured in deep yellow) define the mantle diapirs. I) Interpretative sketch of the ophiolite (circles: diapirs, dark gray: shear zones, thick blue lines: dike orientations, dashed line: ridge axis). II–III) Schematic cross-sections of the on-axis Maqsad diapir and off-axis Mansah diapir, modified from Jousselin and Nicolas (2000). Horizontal straight lines in the lithospheric harzburgite represent the frozen flow direction of the lithospheric mantle. Lines in the asthenospheric section represent flow orientations in each diapir. MTZ denotes Moho Transition Zone. Each small letter references an illustration in Fig. 2. Adapted from Boudier et al. (1997). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Toomey et al. (2007), mantle melt upwelling was found 5 to 20 kmfrom the axis over nearly half of a 200 km first order segment of the East Pacific Rise (EPR). The discovery of melt lenses in the lower crust, up to 20 km from the rise (Crawford and Webb, 2002;Crawford et al., 1999; Garmany, 1989), also shows that not all melt is focused beneath the ridge. Furthermore, seamount volcanism provides direct evidence of off-axis magmatic processes.

When melt is delivered off-axis, its composition may be mod-ified by interaction with cool, hydrothermally altered lithospheric mantle (Bosch et al., 2004). However, assimilation and fractional crystallization in the crust may obscure the effects of sub-Moho processes in erupted basalts. Off-axis magma may even be blocked by the cold crustal barrier and not reach the surface. Thus the ori-gin of off-axis magmatism may be more effectively investigated by examining rocks beneath the Moho, which are preserved from complications occurring at higher levels. An ideal target is the Moho Transition Zone (MTZ), the region immediately underlying the crust–mantle interface, which is accessible almost exclusively in ophiolites (Boudier and Nicolas, 1995).

The Oman ophiolite presents relics of a former fast-spreading ridge, delineated by five on-axis diapirs (Boudier et al., 1997), whose size and spacing match the distribution of melt upwelling centers on the EPR (Toomey et al., 2007). An off-axis upwelling center was also identified (Jousselin and Nicolas, 2000), providing a field laboratory to explore the contrasts between on and off-axis magmatism. Below, we compare major and trace element and Sr

and Nd isotope data for whole rocks and minerals from the MTZ in and around this off-axis diapir with similar data from the more classic MTZ section of a nearby on-axis diapir.

2. Geological setting and field description of the diapirs

The Oman ophiolite, about 500 km long and 50 to 100 km wide, was obducted on the Oman margin during the closure of the Neo-Tethys ocean 95 million years ago (Coleman, 1981; Hacker, 1994). Its continuous gabbroic crust indicates the presence of extensive, long lived magma chambers along the ridge, suggesting that it is a relic of a fast spreading center (Nicolas and Boudier, 1995), as confirmed by high resolution dating indicating a half spread-ing rate of 10 cm/yr (Rioux et al., 2012). The tectonic context of the Oman ophiolite remains controversial, the two main hypothe-ses being a Mid-Ocean Ridge (e.g. Boudier et al., 1988; Godard et al., 2006, 2003; Hacker, 1994; Nicolas and Boudier, 2003) and a Supra-Subduction Zone (SSZ) setting (e.g. Pearce and Cann, 1971;Pearce et al., 1981; Shervais, 2001).

In the southern part of the ophiolite (the Rustaq, Sumail and Wadi Tayin massifs), a NW-SE dike system is observed over a width of 20–50 km, separating into two parts a NE-SW dike sys-tem found in the outer parts of the massifs (Fig. 1). This ge-ometry shows that a paleo-spreading center is preserved in the center of the massifs (Boudier et al., 1997), along an alignment of five mantle diapirs (Fig. 1-I). Mantle diapirs, formed by local-

140 M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149

ized melt upwelling, are identified by steeply plunging lineations, the presence of numerous dikes in the harzburgite and a MTZ thicker than 50 m (Jousselin et al., 1998). In the on-axis case (Fig. 1-II), radial horizontal lineations in the mantle at the out-skirts of the diapirs and their parallelism with magmatic lineations in the lower crust indicate formation beneath a mature spread-ing center, when the gabbro was still in a magmatic state. The on-axis MTZ varies from a few meters to 450 m in thickness and contains mostly dunite, as well as a few horizons with pla-gioclase impregnated dunites and layered gabbro lenses (Boudier and Nicolas, 1995; Jousselin et al., 2012; Kelemen et al., 1997;Koga et al., 2001) (Fig. 2a). We studied the Maqsad diapir of the Sumail massif, a typical example of an on-axis diapir in-vestigated in several previous studies (e.g. Godard et al., 2000;Koga et al., 2001).

The Mansah diapir is also located in the Sumail massif, but about 30 km northeast of the alignment defined by the five on-axis diapirs (Fig. 1). The study of Jousselin and Nicolas (2000) and our new field observations reveal characteristics contrasting with those of the on-axis diapirs. The Mansah diapir (Fig. 1-III) is bounded by shear zones that separate the steeply plunging lineations of the di-apir from the horizontal lineations of the surrounding harzburgites. This structure is interpreted to result from the impingement of the diapir on the off-axis horizontally laminated lithospheric mantle. The overlying Moho and lowermost crust are tectonized, as if dis-turbed by the uprising diapir (Fig. 2b). In places, the top of the MTZ forms a physical mixing zone between the MTZ pyroxenites and the gabbros from the pre-existing lower crust (Fig. 2c). The MTZ is very thick, up to 1 km, and is composed of dunite with py-roxenite lenses (Fig. 2d) varying in size from tens of centimeters to tens of meters. These lenses make up ∼20% of the MTZ section and contain clinopyroxene with variable amounts of olivine and al-most no orthopyroxene. They are less strongly deformed than the on-axis layered gabbro lenses, which are entirely absent in the off-axis case.

Within a five kilometer region around and above the off-axis Mansah diapir, but not inside the diapir itself, the lower crust and mantle are intruded by gabbroic and microgabbroic bod-ies (Fig. 1-III). Similar intrusions are not reported elsewhere in the ophiolite. In the mantle section, they consist of 10 to 50 cm thick gabbroic sills, with fine-grained centers and amphibole-rich, coarse-grained margins. These are referred to as “Wasit” sills (Fig. 2e), after their type locality. The intrusions in the crust and the upper 50 m of the mantle have microgabbroic textures, and are referred to as “Nidab” intrusions (Fig. 2f), after their type locality. These intrusions sometimes take the form of dikes, 10 to 200 cmthick, but more often form large (>10 m), chaotic lenses. Xenoliths of harzburgite and dunite, varying in size from a few centimeters to several meters, are incorporated into the Nidab intrusions. Their elongated shapes with rounded edges and the reaction zones at their boundaries suggest incomplete assimilation of these litholo-gies, which are most likely derived from the lithospheric mantle overlying the rising diapir before it reached the Moho (Fig. 2g–h). Both the Wasit sills and the Nidab lenses lack the deformation fabric of the mantle and crustal sections into which they intrude. Together with the fine-grained texture and the presence of amphi-bole, these characteristics are consistent with their injection into frozen and hydrated off-axis lithosphere.

3. Rock types investigated

Petrographic descriptions and geographic coordinates of sam-ples are given in supplementary Table A.1; locations are in Fig. 3. Lithologies include:

Fig. 2. (a) Dunite with layered gabbro lenses in the MTZ of the on-axis Maqsad diapir. (b) Tectonized Moho (highlighted by the dashed yellow line), with sets of westward dipping normal faults in the overlying lower crust, above the off-axis Mansah diapir. (c) Semi-solid melange of pyroxenite from the off-axis MTZ with the crustal gabbros at the Moho. (d) Dunite (red) with pyroxenite lenses (green) in the MTZ of the off-axis Mansah diapir. (e) Close-up of a Wasit sill, with fine-grained center and coarse-grained borders. (f) Nidab intrusions in the lower crust. (g) Elon-gated peridotite xenoliths in a Nidab intrusion. Contours highlighted by the dashed lines. (h) Close-up of a peridotite xenolith (5 cm long) in a Nidab intrusion show-ing a 0.5 cm wide pyroxenite reaction rim. Du – Dunite; Px – Pyroxenite; Cpx – Clinopyroxene; Gb – Gabbro. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

– Dunites and harzburgites from both the on and off-axis di-apirs. Some can be described as “impregnated”, i.e. they con-tain interstitial phases suggesting precipitation from or inter-action with melts. These phases consist of clinopyroxene and plagioclase in the on-axis samples, but only clinopyroxene in the off-axis samples.

– Gabbro lenses, found only in the on-axis MTZ, composed of olivine, clinopyroxene and plagioclase.

– Pyroxenites, found only in the off-axis MTZ, composed of clinopyroxene with varying amounts of olivine and extremely rare orthopyroxene. Though some samples approach wehrlitic compositions, all will be referred to as pyroxenites for simplic-ity. Two samples are from a mixing zone between pyroxenite and lower crustal gabbro at the Moho (Fig. 2c). One was an-alyzed in bulk while in the other, a dark layer (olivine and clinopyroxene with minor plagioclase) and a white layer (es-sentially plagioclase) were analyzed separately.

– The off-axis gabbroic (Wasit) and microgabbroic (Nidab) in-trusions, which rim the Mansah diapir, and contain olivine, pyroxene, plagioclase and amphibole.

M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149 141

Fig. 3. Simplified geologic maps of the Mansah and Maqsad areas, adapted from Jousselin and Nicolas (2000) and Jousselin et al. (1998), showing sample locations. Blue: Nidab intrusions; green: Wasit sills; black: pyroxenites; red: gabbros; orange: dunites; yellow: harzburgites; purple: gabbro–pyroxenite melange. Gb – gabbro; Px – pyroxenite. The dashed lines represent the wadis. The contours represent lin-eation isodips. (a) Map of the Mansah area. Dark gray – crust; light gray – dunite and pyroxenites; white – harzburgite. (b) Map of the Maqsad area. Dark gray – gab-bros in both the crust and the MTZ; light gray – MTZ dunites; white – harzburgite. (For interpretation of the references to color in this figure legend, the reader is re-ferred to the web version of this article.)

4. Results

Descriptions of all analytical techniques can be found in the electronic appendix. Data are presented in supplementary Ta-bles A.2 to A.6.

4.1. Whole rock major element compositions

Major element abundances are consistent with rock types and reflect modal variations (Fig. A.1), so only a few key observations are noted here. Harzburgites and dunites lacking melt impregna-tions have the same compositions on and off-axis. Nidab and Wasit off-axis intrusions differ from the on-axis MTZ gabbros, having higher FeO, MnO, TiO2 and Total Alkali (Na2O + K2O) contents. The Nidab intrusions have globally basaltic compositions, with higher Al2O3 and total alkalis and much lower MgO contents than the Wasit sills.

4.2. Whole rock trace element compositions

In both diapirs, harzburgites and dunites lacking melt impreg-nation features have extremely low Rare Earth Element (REE) and other incompatible trace element contents, with most values be-low detection limits (Fig. A.2a). Chondrite normalized values of the measurable HREE hint at progressive depletion towards the lighter REE (Fig. A.3a), as noted by Godard et al. (2000) in their study of Maqsad diapir peridotites. In contrast, impregnated harzburgites and dunites have nearly flat HREE patterns and moderate depletion in LREE (Fig. A.3a), with positive anomalies in Eu in the on-axis samples. This feature is consistent with the presence of plagioclase, which also may explain the strong positive Pb and Sr anomalies in the full trace element patterns of the on-axis rocks (Fig. A.2a), though hydrothermal alteration may also play a role.

The rocks of the off-axis gabbro–pyroxenite melange have REE compositions similar to those of off-axis pyroxenites and on-axis gabbros. The white layer displays strong positive Eu (Fig. A.3b) and Sr (Fig. A.2b) anomalies, characteristic of cumulus plagioclase.

The Wasit sills show REE patterns similar to those of the off-axis pyroxenites, but with concentrations 3 to 5 times higher (Fig. A.3c) and large positive Sr anomalies (Fig. A.4a). The Nidab intrusions have the highest REE concentrations and, in all but one case, display nearly flat REE patterns that contrast with the LREE depleted patterns of the Wasit sills (Fig. A.3d). They have positive anomalies in Eu and Sr along with negative anomalies in Zr and Th in some samples, but relative depletion in Nb and Ta is not evident (Fig. A.4b).

4.3. Mineral major element compositions

Major element zoning between core and rim is not observed in any of the analyzed phases. In all thin sections, different grains of each phase yield indistinguishable results (Table A.4). Molar forsterite (Fo) values in olivine vary from 0.92 to 0.94 in dunites and harzburgites from both diapirs (except for on-axis impreg-nated dunite 07-OD20D with Fo 0.90). In the off-axis case, average olivine Fo values were 0.91 and 0.92 in the two pyroxenites from within the diapir, and varied from 0.79 to 0.88 in the Wasit and Nidab intrusions surrounding the diapir. Eleven of the 12 gabbros analyzed from the on-axis diapir MTZ yielded olivine Fo values of 0.89–0.91. Clinopyroxene in harzburgites from both diapirs, and in pyroxenites from the off-axis diapir, have Mg# [Mg/(Mg + Fe)] ranging from 0.93 to 0.95, while those from the off-axis intrusions range from 0.76 to 0.90. Mg# of clinopyroxenes of all 10 analyzed on-axis gabbros display a limited range (0.89–0.93). Among all of our samples, clinopyroxene Mg# and olivine Fo are positively cor-related (Fig. A.5b). The highest fraction of molar anorthite (An%) in plagioclase (0.96) was found in the impregnated dunite 07-OD20D, while the An% values of plagioclase of on-axis gabbros vary from 0.81 to 0.96. In contrast, An% values of plagioclase from the off-axis intrusions vary greatly, from 0.51 to 0.92.

4.4. Trace element compositions of separated minerals

Trace element contents of individual phases (clinopyroxene, plagioclase, amphibole) from a given rock were quite homogeneous (Table A.5) so only average REE compositions for each sample are plotted in the figures cited below.

All clinopyroxenes from on-axis gabbros, off-axis pyroxenites and the gabbro–pyroxenite melange display flat HREE patterns and depletion in LREE, with those from the off-axis pyroxenites having globally lower REE contents (Fig. 4a) than those from the on-axis gabbros (this study and Koga et al., 2001). Clinopyroxenes from the Wasit sills (Fig. 4b) also have flat HREE patterns and show depletion in LREE, but this depletion is less marked than in the

142 M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149

Fig. 4. Chondrite normalized REE compositions of separated minerals. (a) Clinopy-roxene in off-axis pyroxenites and gabbro–pyroxenite melange, the red field corre-sponds to clinopyroxene analyzed in on-axis gabbros shown for comparison (data from our study and Koga et al., 2001). (b) Clinopyroxene in Wasit and Nidab in-trusions. (c) Plagioclase in on-axis gabbros and gabbro–pyroxenite melange. (d) Plagioclase in Wasit and Nidab intrusions. (e) Amphibole in Wasit and Nidab in-trusions. For each phase, the compositions from a given sample were homogeneous so only average values are plotted. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

pyroxenites, and overall REE contents are much higher. Wasit sill clinopyroxenes display negative anomalies in Sr, Pb and Ba (hosted in plagioclase) as well as Nb, Ta and Ti (incorporated in amphi-bole) (Fig. A.6). Unlike in the other lithologies, clinopyroxene in the Nidab intrusions have nearly flat REE patterns, with about ten times chondritic abundances (Fig. 4b).

Plagioclase from the on-axis gabbros (Fig. 4c) show moder-ate enrichment in LREE relative to HREE, with a strong posi-tive anomaly in Eu. In the plagioclase off-axis gabbro–pyroxenite melange, enrichment in LREE is much less evident. Plagioclase from the off-axis intrusions (Fig. 4d) have much higher overall REE abundances with greater LREE enrichment than the on-axis gab-bros.

Amphibole, found only in the off-axis intrusions, have REE spec-tra with forms very similar to those of the corresponding clinopy-roxene, but with 2 to 3 times higher overall abundances (Fig. 4e). The Nidab intrusions show nearly flat REE patterns with a hint of LREE depletion manifested by a small downturn in La, while the relative LREE depletion of the Wasit sills is more evident. Occa-sional minor negative Eu anomalies suggest some degree of equi-libration with plagioclase. Amphibole in both types of intrusions have high abundances of the High Field Strength Elements (HFSE) Nb, Ta and Ti, which complement their low abundances in the as-sociated clinopyroxene (Table A.5 and Fig. A.6).

4.5. Sr and Nd isotopic compositions

Sr and Nd isotopic analyses were performed on 37 samples from the off-axis diapir (pyroxenites, Wasit and Nidab intru-

sions, pyroxenite–gabbro melange) and the on-axis diapir (gabbros, dunites and harzburgites). As discussed below, both leached and unleached samples were analyzed. Some Nd analyses were un-successful because of very low Nd concentrations. Among the unleached samples, on-axis gabbros have high εNd(t) values (εNd(t) = 104 ×[(143Nd/144Nd)sample/(

143Nd/144Nd)CHUR −1], with t = 95 Ma, the age of the ophiolite and CHUR representing the chondritic ratio), between +8.5 and +9.2. In contrast, the un-leached off-axis samples have lower εNd values between +5.6 and +7.6, with one Wasit sill at +8.8. The dichotomy is even stronger in Sr isotopes. Unleached on-axis gabbros have MORB-like 87Sr/86Sr ratios, ranging from 0.703010 to 0.703438, with the exception of one sample (07-OD45; 0.703749), whereas the off-axis samples have more radiogenic values, ranging from 0.703524 to 0.706192 (Table A.6 and Fig. 5).

Sr isotopic compositions of samples from the ocean lithosphere are often perturbed by hydrothermal circulation of seawater (see Bosch et al., 2004, for an example from the Oman ophiolite). To decipher the importance of this effect, analyses of leached sam-ples were performed. The efficiency of our leaching procedure was first demonstrated by comparing results obtained on leached whole rocks with those from handpicked clinopyroxene sepa-rates (see electronic appendix for details). For Sr isotopes, the di-chotomy between on and off-axis results seen in unleached whole rocks nearly disappears in the leached samples (Table A.6 and Fig. 6). The on-axis leached samples have 87Sr/86Sr ratios ranging from 0.703004 to 0.703832 (except for the harzburgite 07-OD27B: 0.708932), while the off-axis leached samples have 87Sr/86Sr ra-tios between 0.703136 and 0.704390. Thus the highly radiogenic Sr compositions of the unleached off-axis samples result mainly from post-magmatic hydrothermal processes. The much smaller post-magmatic hydrothermal influence on the on-axis gabbros may reflect the relatively high Sr concentrations of these samples as well as the presence of a large on-axis magma chamber that could protect the underlying gabbros from hydrothermal circulation. Nev-ertheless, even after leaching to remove the hydrothermal compo-nent, off-axis samples have on average higher 87Sr/86Sr ratios.

Due to the extremely low Nd concentration of seawater, Nd isotopic compositions are expected to be essentially unaffected by hydrothermal circulation, so it is not surprising to find that the leached samples display the same distinction between on and off-axis Nd signatures seen in the unleached samples (Fig. 6). Among the on-axis gabbros, εNd ranges from +8 to +8.8, which puts them in the field of MORBs and unaltered ridge-related sam-ples from Oman (Godard et al., 2006; McCulloch et al., 1981;Rioux et al., 2012, 2013). The off-axis leached samples have globally lower and more heterogeneous εNd values (+6.2 to +7.9). With an εNd value of +8.8, the gabbro–pyroxenite mixture (10-OM39C) is an exception, but this high value may reflect that of the pre-existing gabbroic crust rather than the impinging off-axis diapir. The off-axis leached Wasit and Nidab intrusions have 87Sr/86Sr ratios and εNd values that fully overlap, but extend to higher 87Sr/86Sr, than the range of the leached pyroxenites.

5. Discussion

5.1. Contrasts between the on and off-axis diapirs

Field mapping, petrological and geochemical investigations re-vealed several striking features of the off-axis Mansah diapir that contrast markedly with the on-axis case (Fig. 1). In Mansah, a shear zone separates the steeply plunging lineations of the diapir from the horizontal foliations of the surrounding harzburgite, con-sistent with the uprise of hot material rooted in the asthenosphere through colder lithosphere. In Maqsad, the continuous and grad-ual rotation of lineations towards parallelism with the surrounding

M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149 143

Fig. 5. (a) εNd versus 87Sr/86Sr ratios of our unleached samples (solid triangles, circles, diamonds and squares). The black area represents the unaltered samples (Wadi Tayin massif) from McCulloch et al. (1981). The dark gray and light gray areas represent the volcanics V1 and V2 from Godard et al. (2006). MORB field: O’Nions et al. (1977), DePaolo and Wasserburg (1976), Mahoney et al. (1989), Staudigel et al. (1991), White et al. (1993). The isotopic compositions of Cretaceous seawater are from DePaolo and Wasserburg (1976). The seawater mixing curve was constructed assuming a Sr/Nd ratio of 2.6 × 106 in seawater and of 23 in the rocks, based on the Sr/Nd ratios in our pyroxenite samples. (b) Histogram of 87Sr/86Sr ratios of all the on and off-axis unleached samples. The light gray area represents the MORB field.

Fig. 6. (a) εNd versus 87Sr/86Sr ratios of on and off-axis leached whole rocks or clinopyroxene (cpx) separates. The εNd mean uncertainty (2σ ) including all identified sources of error is indicated, mean uncertainties for Sr isotopic ratios are roughly the size of the symbols. For one sample, no 87Sr/86Sr ratio is available, as indicated by the horizontal double arrow. Light gray area – Oman unaltered samples (McCulloch, 1981). References for MORB and Oman unaltered fields as in Fig. 5. (b) Inset showing all samples for which Nd data are available, leached and unleached.

144 M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149

harzburgite is the expected result of diapirism beneath a spreading axis, where lithosphere is absent. The off-axis MTZ is much thicker than the on-axis MTZ and totally free of gabbroic lenses, contain-ing instead massive accumulations (around 20% of MTZ) of weakly deformed clinopyroxenite lacking both plagioclase and orthopyrox-ene. The mantle and crust surrounding the off-axis diapir contain intrusions bearing magmatic amphibole, suggesting crystallization from a hydrated melt. Clinopyroxene from off-axis MTZ pyroxenites have lower incompatible element abundances than those of on-axis MTZ gabbros, and are more depleted in LREE. However the Nd isotopic compositions of all off-axis lithologies are less radiogenic than those of the on-axis MTZ, and on average their Sr isotopic compositions are slightly more radiogenic (Fig. 6).

5.2. Why is the off-axis diapir so rich in clinopyroxene?

The key to understanding the differences between the on and off-axis diapirs is to determine the origin of the abun-dant clinopyroxenite in the off-axis MTZ. Several processes can form clinopyroxene concentrations. Bédard (1991, 1993, 2001) andLissenberg and Dick (2008) noted that pyroxenite can form in the lower oceanic crust by the reaction of migrating melt with gab-bros. However this scenario is not applicable here because the mantle section traversed by the rising off-axis diapir was de-void of gabbro. Alternatively, the pyroxenites could be of cumu-late origin if clinopyroxene precipitates before plagioclase. During fractional crystallization, both greater pressures and higher water contents favor the formation of clinopyroxene (Feig et al., 2006;Koepke et al., 2009). As the MTZ of the on and off-axis di-apirs probably formed at similar depths (i.e., crustal thickness was roughly constant), a differential pressure effect seems unlikely. On the other hand, enhanced water contents in the off-axis diapir are consistent with geological constraints. As the hot diapir rose from the asthenosphere, it partially assimilated the intervening litho-spheric mantle, as shown by the deformed ultramafic xenoliths in the Nidab intrusions (Fig. 2g–h). The presence of amphiboles in the margins of the Wasit sills surrounding the diapir implies that their parental magmas were water rich, suggesting that the off-axis lithosphere incorporated into the rising diapir was hy-drated. Evidence for hydrothermal processes at Moho depth in the Oman ophiolite (Bosch et al., 2004; Nicolas and Mainprice, 2005;Nicolas et al., 2003) provides a likely hydration mechanism, assum-ing that such circulation can descend into the underlying mantle section.

However the presence of water, though favoring clinopyrox-ene crystallization while inhibiting that of plagioclase, may not suffice to explain the large volumes of clinopyroxenite found in the Mansah diapir MTZ. If these formed by crystallization from a magma, they should be conjugate to large masses of resid-ual melt. The volume of the Wasit and Nidab intrusions, which could in theory represent such residual melts, is much smaller than that of the crustal section conjugate to the thinner gabbroic MTZ of the on-axis diapir. Furthermore, even at water saturation, the temperature interval for clinopyroxene fractionation without plagioclase is limited in MORB type magmas (∼40 ◦C in experi-ments using hydrated tholeiitic compositions; Koepke et al., 2009;Feig et al., 2010). So if the pyroxenites were cumulates from even a water-rich MORB-type melt, it is surprising that they contain no trace of plagioclase.

Thus an additional cause besides the presence of water may be needed to explain the very large volumes of pyroxenite observed in the off-axis diapir. One possibility, developed below, is that the compositions of melts delivered to the off-axis MTZ differed from those delivered on-axis, being enriched in a clinopyroxene com-ponent. This hypothesis is supported by the difference in Sr, and especially Nd, isotopic composition between the MTZ of the two

diapirs, indicating that the on-axis and off-axis magmas were not derived from the same source.

5.3. Melting of a pyroxenite-bearing source and resulting melt–peridotite interaction

The less radiogenic Nd isotopic compositions of the Mansah py-roxenites, compared to those of rocks from the on-axis MTZ, may indicate that the Mansah magmas contained a melt-component derived from pyroxenite veins. Such veins may originate directly or indirectly from recycled oceanic crust (e.g. Allègre and Tur-cotte, 1986), though several other origins, including delamination of lower crustal material or metasomatic segregation could also be envisaged (Hirschmann and Stolper, 1996). The presence of such veins in the MORB source mantle has been postulated on the basis of direct observations in ultramafic massifs and of coupled iso-topic and trace element systematics in basalts (Hirschmann and Stolper, 1996; Niu et al., 1999; Prinzhofer et al., 1989; Stracke et al., 2000). However, the signatures of hypothesized pyroxenite vein components in MORB magmas are quite subtle, due to mixing and dilution with peridotite-derived melts. Rocks from the MTZ may approach the pre-mixing isotopic compositions of the mag-mas more closely and thus may more easily reveal the presence of pyroxenite veins.

Pyroxenite veins have lower melting points than the surround-ing peridotite and will thus be the first components to melt in an upwelling mantle diapir (Hirschmann and Stolper, 1996; Kogiso et al., 2004a, 2004b; Lambart et al., 2012; Stracke et al., 2000). Be-neath the ridge axis, melting rates are high, leading to a greater contribution from peridotite. In the MTZ of the on-axis diapir, the early pyroxenite-derived melts are overwhelmed by melts derived from the surrounding fertile peridotite, so the Nd and Sr iso-topic ratios have MORB-like values. Off-axis, extents of melting are lower, so the pyroxenite signature is more likely to be preserved. As the off-axis diapir ascends towards the Moho, it interacts with and incorporates the harzburgitic lithosphere. This lithosphere is highly depleted in incompatible elements and cannot contribute much Nd and Sr to the rising melts, so the magmas reaching the MTZ retain the Nd and Sr isotopic signature of the original pyrox-enite veins (see the electronic appendices for an example of how such a vein melting model might work).

Melts derived from pyroxenite veins are unlikely to be in equilibrium with the surrounding peridotite (e.g. Sobolev et al., 2007). Based on both experiments and theoretical modeling,Lambart et al. (2012) determined that the interaction of melts derived from pyroxenite veins with peridotite leads to the crys-tallization of clinopyroxene and, depending on the conditions and compositions of the melt, to the crystallization of olivine and the dissolution of orthopyroxene or vice versa. Although their experi-mental conditions are not the same as ours (they consider lherzo-lite instead of harzburgite, their modeling and experiments were performed at higher pressures and under anhydrous conditions), their study shows that it is possible to produce a large amount of clinopyroxene by melt interaction in the mantle over a wide range of pressure and temperature conditions. An analogous pro-cess could explain the clinopyroxenes found in the MTZ of the off-axis diapir.

We suggest that in the Mansah region, magmas formed by the partial melting of pyroxenite veins within an ascending astheno-spheric diapir. In Fig. 7, it is further speculated that a random abundance of pyroxenite veins within the root of the on-axis up-welling zone may have caused enhanced melting, and triggered the off-axis diapirism. In the final stage of its ascent, the off-axis melt reacted with the hydrated lithospheric harzburgite to crystallize clinopyroxene and some olivine, while dissolving orthopyroxene. In the model of Lambart et al. (2012), the scarcity of orthopyroxene

M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149 145

Fig. 7. Schematic cross-section of the flow structure with a possible mechanism for the triggering of off-axis diapirism. Black arrows show large scale mantle flow, color shades and solid contours show melt fraction in the melting region, which in many classic models is triangular and centered on the ridge axis (e.g. Parmentier and Morgan, 1990). A local high abundance of pyroxenite veins would produce an un-usual amount of melting that could create a gravitational instability and trigger an off-axis diapiric uprise superimposed on the normal flow. Also, because of its low melting point and of the large base of the triangular melt production zone, it may be difficult for any such heterogeneity to reach the axis unless it rose right beneath the axis where the flow is strictly vertical.

places strong constraints on the composition of the original pyrox-enite veins, as they must be able to produce melts with low silica activity. However, these constraints may be relaxed in our case, since the Mansah MTZ formed at lower pressures than those con-sidered by Lambart et al. (2012), and at these shallow depths, even MORB melts are out of equilibrium with orthopyroxene (Kelemen et al., 1995). As discussed above, the presence of water derived from the lithosphere would be likely to enhance the crystallization of clinopyroxene and suppress that of plagioclase, normally ex-pected to occur at our lower pressure conditions (Feig et al., 2006;Koepke et al., 2009). The overall result may be a MTZ composed of dunite with clinopyroxenite lenses.

5.4. Depleted incompatible element signature of pyroxenites

As noted above, the off-axis pyroxenites have less radiogenic Nd isotopic compositions than the on-axis gabbros (Fig. 6), indicat-ing crystallization from melts derived from a source with a lesser degree of the time-integrated LREE depletion. However, clinopy-roxenes from the pyroxenites have more depleted incompatible trace element signatures than clinopyroxenes from the gabbros (Fig. 4). To try to understand this paradox, we calculated the REE contents of melts in equilibrium with the clinopyroxenes in the two settings. This approach inherently assumes that the clinopy-roxene compositions currently observed were not modified by post-crystallization processes. Re-equilibration with trapped melts can augment the incompatible element contents of clinopyroxenes, leading to spuriously high REE contents in the calculated equi-librium melts (Bédard, 1994). However, in our case, the scarcity of intergranular phases, the absence of mineralogical zoning and the uniformity of REE compositions from a given thin section ar-gue against the presence of significant amounts of trapped melt, while the preserved magmatic textures indicate that evidence for such a phase has not been erased by recrystallization (Jousselin et al., 2012). Thus the REE contents of clinopyroxenes are probably original and can therefore be used to calculate the REE contents of the melts from which they formed. To do this, we used the predictive model for REE partitioning between clinopyroxene and melt of Wood and Blundy (1997). This model takes into account pressure and clinopyroxene composition, thus permitting more re-liable estimates of melt compositions in different contexts than those obtained assuming constant partition coefficients. To enable

Fig. 8. Chondrite normalized REE compositions of melts in equilibrium with clinopy-roxenes calculated using partition coefficients based on the predictive model of Wood and Blundy (1997). (a) Compositions of liquids in equilibrium with clinopy-roxene in on-axis gabbros (data from our study and Koga et al., 2001) and off-axis pyroxenites. Trace element compositions of N-MORB from the EPR and IOR are shown for comparison (source from the GERM database: Hart et al., 1999 and ref-erences therein). (b) Liquids in equilibrium with cpx from Wasit sills and off-axis pyroxenites; whole rock REE compositions of the Nidab intrusions shown for com-parison. Chondrite normalizing values and element order from Sun and McDonough(1989).

comparison with the previous study of Koga et al. (2001), we also used this model to recalculate the liquids in equilibrium with the clinopyroxenes they analyzed from the gabbros of the on-axis MTZ.

The calculated melts in equilibrium with clinopyroxenes (Fig. 8a) in the on-axis gabbros have nearly flat REE patterns, with abun-dances mostly 5 to 15 times chondritic values, except for mi-nor relative depletion in the lightest REE. Our calculated melt compositions are consistent with those of Koga et al. (2001), also from the Maqsad diapir. As they note, the Maqsad MTZ melts are relatively uniform; their compositions are within the range observed in the Oman volcanics (Alabaster et al., 1982;Pallister and Knight, 1981) and similar to those of MORB.

The calculated melts in equilibrium with clinopyroxenes in the off-axis pyroxenites (Fig. 8a) display globally flat or slightly slop-ing HREE patterns, with depletion in LREE. Overall, the calculated melts differ between the MTZ of the two diapirs, with those from the off-axis MTZ being more depleted than those of the on-axis MTZ, as shown by the lower Nd/Yb and La/Sm ratios and mostly lower Yb and La contents of the off-axis melts in equilibrium with the pyroxenites (Figs. 9a and A.7). Thus, unlike the inferred on-axis melts, the calculated melts in equilibrium with the off-axis MTZ are more depleted in REE and other incompatible elements than the Oman volcanics or most MORB, despite the Nd isotopic evidence for their derivation from a source with a lower time-averaged Sm/Nd ratio.

146 M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149

Fig. 9. (a) Chondrite normalized Nd/Yb ratios in the calculated melts in equilibrium with the clinopyroxenes in our samples and in Koga et al. (2001). Partition coef-ficients used in the calculation based on the model of Wood and Blundy (1997). All of the analyses are included for each sample. (b) Primitive mantle normalized La/Nb ratios versus 143Nd/144Nd ratios in the calculated melts in equilibrium with the clinopyroxenes in the off-axis pyroxenites. Here the partition coefficients of Hart and Dunn (1993) were used to calculate liquid Nb and La contents, as values for Nb are not available from the study of Wood and Blundy (1997). Though the exact val-ues of the calculated La/Nb ratios will depend on the choice of these coefficients, the low La/Nb values obtained are consistent with the existence of negative Nb anomalies in the equilibrium liquids. Data for the Geotimes (V1) and Lasail (V2) lavas from Godard et al. (2006) and for N-MORB are shown for comparison (with the 143Nd/144Nd ratio recalculated at 95 Ma). Primitive mantle normalization values taken from Sun and McDonough (1989).

In the context of our model, the low incompatible element con-tents of the off-axis pyroxenites could result in part from effective dilution caused by interaction of the original pyroxenite-derived melts with lithospheric, depleted harzburgite (direct evidence for harzburgite assimilation, albeit at a higher level, is provided by the ultramafic xenoliths in the Nidab intrusions). Clinopyroxenes formed during melt interaction with depleted harzburgite are ex-pected to have more depleted REE spectra than those formed by simple crystallization from melts. This effect might be enhanced by direct melting of the hydrated harzburgites as the diapir ascends, diluting the incompatible element contents even further, while having very little effect on Nd isotopic compositions, which would reflect those of the original pyroxenite veins. Nevertheless, as our modeling shows (see electronic appendices), interaction with de-pleted harzburgite, while greatly lowering REE contents, can ex-plain only part of the relative LREE depletion. In other words, the pyroxenite veins in the source must themselves have been some-what LREE depleted.

5.5. Origin of the Wasit and Nidab intrusions

Lambart et al. (2012) show that reaction of pyroxenite-derived melts with peridotite generates a second melt, of quantity and composition dependent on the conditions. If the pyroxenites of the off-axis diapir formed by an analogous process, the resultant melt may have escaped above and on both sides of the ascending di-apir, forming melt lenses in the surrounding mantle (Fig. 10). Given their MgO and Ni rich compositions, the Wasit sills may represent cumulates from such melts, thus explaining why they share the same Sr and Nd isotopic compositions as the pyroxenites. Clinopy-roxenes from the Wasit sills display LREE depletion similar to, but slightly less marked, than that of clinopyroxenes from the pyrox-enites (Fig. 4). However they have much higher overall REE levels, suggesting that their parental melts experienced extensive olivine and minor clinopyroxene fractionation prior to crystallization of the sills. This is supported by the lower Fo contents of the Wa-sit olivines and the lower Mg# of their clinopyroxenes relative to

Fig. 10. Bottom: evolution of the off-axis Mansah diapir and formation of the Wasit and Nidab intrusions. Top: schematic diagrams of the REE compositions expected in the melts, clinopyroxene and/or whole rocks produced during each step of the model (blue: observed; green: inferred liquid compositions that cannot be directly observed). Circled letters in bottom panel correspond to REE evolution stages shown in top panel. First, melting (at unknown depth but above the stability limit of gar-net) of a pyroxenite vein-rich zone triggers the rise of a diapir; after this diapir reaches the base of the lithospheric mantle (depleted harzburgite), it continues to ascend, incorporating the harzburgite and imprinting steeply plunging lineations; melt interacts with harzburgite forming pyroxenite towards the top of the diapir (light yellow); melt from the diapir intrudes into the overlying lithosphere, forming the earliest Wasit–Nidab sills and lenses which may contain mantle xenoliths. In the second stage (center), a large part of the depleted lithospheric mantle has been traversed and transformed by the diapir; early Wasit–Nidab intrusions are cross-cut by the rising diapir, and the pyroxenite-bearing zone grows thicker. Finally, the di-apir reaches and tectonises the Moho and lowermost crust, and forms, in places, the crust–pyroxenite melange; the final Nidab intrusions are injected into the crust. Wa-sit sills represent cumulates from the liquids expelled from the diapir, while Nidab intrusions are approximate mixtures of liquids conjugate to the Wasit sills and to the pyroxenite rich zones in the diapir. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

those of the pyroxenites (Fig. A.5). Using the model of Wood and Blundy (1997), we calculated REE compositions of liquids in equi-librium with clinopyroxene from the Wasit sills (Fig. 8b). However these values are only approximate, as the complementary Ta–Nb anomalies of the clinopyroxenes and amphiboles (Fig. A.6) sug-gest post-crystallization re-equilibration occurred between these two phases.

The Nidab intrusions, which have bulk major element composi-tions approximating those of basaltic liquids (Fig. A.1) and nearly flat whole rock and clinopyroxene REE spectra (Figs. A.3 and 4b), may represent mixtures between liquids residual to the pyroxen-ites and those residual to the Wasit sills. The Nidab intrusions contain abundant amphibole-rich veins and numerous ultramafic and crustal xenoliths, suggesting that they have assimilated litho-spheric material hydrated by hydrothermal circulation of seawater. This could explain why the Nidab intrusions, while having Sr and Nd isotopic compositions broadly similar to those of the pyroxen-ites and the Wasit sills, include samples with somewhat elevated 87Sr/86Sr ratios even after leaching. The full-overlap between the εNd values of the Wasit and Nidab intrusions surrounding the off-axis diapir, and those of the pyroxenites within the diapir, sug-gest that all of these rocks form a magmatic suite, derived from the same initial melts. The minor heterogeneity of the Nd and Sr isotopic signatures at the level of the MTZ could reflect isotopic variability among the pyroxenite veins that provided the original melts.

M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149 147

The structural positions of the Wasit sills and Nidab intrusions can be understood within the context of the off-axis diapir. The absence of sills within the diapir and the sheer zone separating the diapir from the Wasit sills of the surrounding mantle suggest that as the diapir ascended towards the Moho, it progressively en-gulfed the lithospheric mantle containing the Wasit sills and Nidab dikes that were produced above and at its periphery (Fig. 10). The presence of elongated and rounded mantle xenoliths in the Nidab intrusions would not be possible if the Nidab intrusions were pro-duced only after the diapir had reached the Moho. These intrusions are only present in the last few meters of the mantle and in the lowermost crust. We suggest that to first order, the Nidab lenses represent the liquid produced by melt–rock reaction in the diapir, mixed with residual liquids from the Wasit sills and modified by interaction with the pre-existing crust and uppermost mantle.

5.6. Could the Mansah diapir result from SSZ magmatism?

As discussed above, we think that our off-axis results are most simply explained by interaction of pyroxenite-derived melts with hydrated, depleted harzburgites in the lithospheric section tra-versed by an upwelling diapir. Nevertheless, we must consider the alternate possibility that the unusual features of the Mansah diapir reflect late-stage magmatic processes in an SSZ setting. Though the absence of arc volcanism or of any volcaniclastic sediments, and the lack of arc-like trace element signatures (Godard et al., 2006)argue against the existence of a fully developed subduction envi-ronment, evidence exists from throughout the ophiolite for higher fluid contents in the magmatic sources relative to those of MORB (MacLeod et al., 2013). Given the tectonic complexity that surely existed prior to obduction, such hydrous conditions could perhaps be reconciled with several different models.

The REE spectra of clinopyroxenes from the Mansah pyroxen-ites resemble those of clinopyroxenes from lower crustal wehrlites of the Haylayn block described by Koepke et al. (2009), which they related to the Lasail magmas (also called V2) found in the northern part of the Oman ophiolite (Ernewein et al., 1988). These late-stage magmas are thought to result from fluid-enhanced melting of previously depleted lithosphere, with the fluids derived from dehydration of sediments and altered crust of the underlying slab during incipient (Boudier et al., 1988) or more well-developed (Pearce et al., 1981) subduction. However several observations ar-gue against a subduction related origin for the Mansah diapir. The negative Nb–Ta anomalies characteristic of lavas from SSZ settings are lacking in Mansah, notably among the rocks with composi-tions most closely resembling those of liquids (Nidab intrusions, Fig. A.4). Moreover, although the presence of abundant clinopyrox-ene is considered to be an argument in favor of SSZ settings, the liquids in equilibrium with our off-axis pyroxenites lack the nega-tive Nb anomalies that characterize lavas from SSZ (Fig. 9b; ratios La/Nb <0.5). V2 magmas have never been found in the Sumail block, while in the northern blocks of the ophiolite, where these magmas are located, no structures resembling the Mansah diapir exist. The V2 magmas have Nd isotopic compositions very similar to those of the V1 (i.e., ridge-related) magmas (V1: εNd = 7.9–8.8; V2: εNd = 7.3–8.7; Godard et al., 2006) and of the on-axis Maqsad diapir (εNd = 7.8–8.8). In contrast, samples related to the off-axis Mansah diapir have mostly less radiogenic values (εNd = 6.2–7.9). This contrast in Nd isotopic composition suggests that the Mansah diapir is not genetically related to the V2 magmas, so even if the latter have a subduction-related origin this does not imply a simi-lar origin for the Mansah magmas.

Rioux et al. (2012, 2013) found that post-ridge magmatic crustal intrusions from the Sumail, Wadi Tayn and Rustaq–Haylayn massifs have younger U–Pb zircon ages and slightly lower εNd val-ues than ridge related magmas. They therefore drew a parallel with

V2 magmatism (somewhat tenuous in our view, since as noted above, V2 and V1 magmas have indistinguishable εNd values), ar-guing that both formed in a proto-arc setting. Since the samples analyzed by Rioux et al. (2013) were collected far from the Mansah diapir and in the upper crust rather than the MTZ, their pertinence to the age of this structure is not obvious. In the shallow subduc-tion model of Boudier et al. (1988), oceanic detachment is inferred to occur along the asthenosphere–lithosphere boundary. Since the Mansah diapir shows asthenospheric flow structures, if this model is correct, emplacement of the diapir must predate detachment and thus be roughly concomitant with magmatism at the ridge axis. Therefore, clarification of the tectonic context of the off-axis diapirism requires precise determination of its timing.

We conclude that there is no compelling evidence to suggest that development of the Mansah structure was related to fluids rising in a supra-subduction zone setting. Nevertheless, more work, in particular U–Pb zircon dating (e.g. Rioux et al., 2013) of rocks associated with the Mansah diapir, is needed to better define the tectonic setting in which the diapir ascended.

6. Summary and conclusion

Our study of the Mansah diapir of the Oman ophiolite pro-vides new insights into off-axis magmatic processes, particularly those occurring in the critical crust–mantle boundary region. The Moho Transition Zone (MTZ) of this off-axis diapir displays plung-ing lineations that cut at high angle the horizontal foliation of the surrounding mantle, contrasting with the continuity from vertical to horizontal mantle flow observed in on-axis diapirs. The MTZ of the off-axis diapir is even thicker than that of the nearby on-axis Maqsad diapir and contains massive clinopyroxenite lenses in dunite, rather than the gabbroic lenses in dunite found in the on-axis case. Gabbroic sills intrude the mantle around the off-axis di-apir and microgabbroic intrusions are found in the overlying crust. Clinopyroxenes from the off-axis pyroxenites were derived from melts more depleted in incompatible elements than their on-axis equivalents. However, the Nd isotopic composition of the off-axis melts indicates a less depleted source (εNd = 6.2–7.9 in 18 of 19 analyses) than that of the on-axis melts (εNd = 7.8–9.2, 10 analy-ses), which are isotopically similar to Indian Ocean MORB.

To explain these unusual features, we suggest that in the off-axis setting, a diapir rooted in the asthenosphere ascended through the depleted, hydrated lithosphere. The first melts to form were derived from pyroxenite veins at depth, as suggested by the less ra-diogenic εNd values of the off-axis samples. During its ascent, the diapir assimilated parts of the hydrated and depleted lithosphere in its path. As the melt rose, it interacted with the lithospheric harzburgite, forming clinopyroxene and olivine at the expense of the orthopyroxene, and acquiring a more depleted trace element signature. As harzburgite contains very little Nd, this process had little effect on Nd isotopic compositions. The water derived from this hydrothermally altered lithosphere enhanced the crystalliza-tion of clinopyroxene and suppressed that of plagioclase. The gab-broic sills found in the mantle around the diapir are cumulates formed from the fractionated daughters of liquids produced by the reaction between the harzburgite and pyroxenitic melts, while the microgabbroic intrusions, found essentially in the crust, represent the approximate corresponding liquid compositions.

The less radiogenic Nd compositions of the off-axis samples add to the growing evidence for the presence of pyroxenite veins in the MORB mantle source. This component was probably present in the source of both diapirs, but in on-axis magmas its signature was muted by a larger relative contribution from peridotite melts. Our results show that off-axis melt delivery can have a major impact on the mineralogy and geochemistry of the magmatic products. As most of the geophysical evidence for off-axis deep magmatic

148 M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149

activity at fast spreading centers is recent, the importance of such processes may have been underappreciated.

Acknowledgements

This work was supported by the INSU-SYSTER (CNRS) program. The authors are grateful to Cédric Demeurie for the polished thin sections, Catherine Zimmermann and Christiane Parmentier for as-sistance with Sr and Nd chemistry and analyses, Olivier Bruguier for assistance with in-situ ICP-MS analyses and Lydéric France for helpful discussions. We thank Jean Bédard and Matthew Rioux for their thorough and constructive reviews, which greatly improved the quality of the manuscript. The patient and careful editorial handling of Tim Elliott is greatly appreciated. This is CRPG con-tribution number 2420.

Appendix A. Supplementary material

Supplementary material related to this article can be found on-line at http://dx.doi.org/10.1016/j.epsl.2015.12.005.

References

Alabaster, T., Pearce, J.A., Malpas, J., 1982. The volcanic stratigraphy and petrogenesis of the Oman ophiolite complex. Contrib. Mineral. Petrol. 81, 168–183.

Allègre, C.J., Turcotte, D.L., 1986. Implications of a two-component marble-cake mantle. Nature (ISSN 0028-0836) 323, 123–127. http://dx.doi.org/10.1038/323123a0.

Bédard, J.H., 1991. Cumulate recycling and crustal evolution in the bay of islands ophiolite. J. Geol. 99, 225–249. http://dx.doi.org/10.1086/629486.

Bédard, J.H., 1993. Oceanic crust as a reactive filter: synkinematic intrusion, hybridization, and assimilation in an ophiolitic magma chamber, western Newfoundland. Geology 21, 77. http://dx.doi.org/10.1130/0091-7613(1993)021<0077:OCAARF>2.3.CO;2.

Bédard, J.H., 1994. A procedure for calculating the equilibrium distribution of trace elements among the minerals of cumulate rocks, and the concentration of trace elements in the coexisting liquids. Chem. Geol. 118, 143–153.

Bédard, J.H., 2001. Parental magmas of the Nain Plutonic Suite anorthosites and mafic cumulates: a trace element modelling approach. Contrib. Mineral. Petrol. 141, 747–771. http://dx.doi.org/10.1007/s004100100268.

Bosch, D., Jamais, M., Boudier, F., Nicolas, A., Dautria, J.M., Agrinier, P., 2004. Deep and high-temperature hydrothermal circulation in the Oman Ophiolite—petrological and isotopic evidence. J. Petrol. 45, 1181–1208. http://dx.doi.org/10.1093/petrology/egh010.

Boudier, F., Ceuleneer, G., Nicolas, A., 1988. Shear zones, thrusts and related magma-tism in the Oman ophiolite: initiation of thrusting on an oceanic ridge. Tectono-physics 151, 275–296.

Boudier, F., Nicolas, A., 1995. Nature of the Moho transition zone in the Oman ophi-olite. J. Petrol. 36, 777.

Boudier, F., Nicolas, A., Ildefonse, B., Jousselin, D., 1997. EPR microplates, a model for the Oman ophiolite. Terra Nova 9, 79–82.

Coleman, R.G., 1981. Tectonic setting for ophiolite obduction in Oman. J. Geophys. Res. 86, 2497–2508.

Crawford, W.C., Webb, S.C., 2002. Variations in the distribution of magma in the lower crust and at the Moho beneath the East Pacific rise at 9–10◦N. Earth Planet. Sci. Lett. 203, 117–130.

Crawford, W.C., Webb, S.C., Hildebrand, J.A., 1999. Constraints on melt in the lower crust and Moho at the East Pacific Rise, 948′N, using seafloor compliance mea-surements. J. Geophys. Res., Solid Earth 1978–2012 (104), 2923–2939.

DePaolo, D.J., Wasserburg, G.J., 1976. Inferences about magma sources and mantle structure from variations of 143Nd/144Nd. Geophys. Res. Lett. 3, 743–746.

Ernewein, M., Pflumio, C., Whitechurch, H., 1988. The death of an accretion zone as evidenced by the magmatic history of the Sumail ophiolite (Oman). Tectono-physics 151, 247–274.

Feig, S.T., Koepke, J., Snow, J.E., 2006. Effect of water on tholeiitic basalt phase equilibria: an experimental study under oxidizing conditions. Contrib. Mineral. Petrol. 152, 611–638. http://dx.doi.org/10.1007/s00410-006-0123-2.

Feig, S.T., Koepke, J., Snow, J.E., 2010. Effect of oxygen fugacity and water on phase equilibria of a hydrous tholeiitic basalt. Contrib. Mineral. Petrol. 160, 551–568. http://dx.doi.org/10.1007/s00410-010-0493-3.

Garmany, J., 1989. Accumulations of melt at the base of young oceanic crust. Na-ture 340, 628–632. http://dx.doi.org/10.1038/340628a0.

Godard, M., Bosch, D., Einaudi, F., 2006. A MORB source for low-Ti magmatism in the Semail ophiolite. Chem. Geol. 234, 58–78. http://dx.doi.org/10.1016/j.chemgeo.2006.04.005.

Godard, M., Dautria, J.-M., Perrin, M., 2003. Geochemical variability of the Oman ophiolite lavas: relationship with spatial distribution and paleomagnetic direc-tions. Geochem. Geophys. Geosyst. 4. http://dx.doi.org/10.1029/2002GC000452.

Godard, M., Jousselin, D., Bodinier, J.L., 2000. Relationships between geochemistry and structure beneath a palaeo-spreading centre: a study of the mantle section in the Oman ophiolite. Earth Planet. Sci. Lett. 180, 133–148.

Hacker, B.R., 1994. Rapid emplacement of young oceanic lithosphere: argon geochronology of the Oman ophiolite. Science 265, 1563–1565. http://dx.doi.org/10.1126/science.265.5178.1563.

Hart, S.R., Blusztajn, J., Dick, H.J., Meyer, P.S., Muehlenbachs, K., 1999. The fingerprint of seawater circulation in a 500-meter section of ocean crust gabbros. Geochim. Cosmochim. Acta 63, 4059–4080.

Hart, S.R., Dunn, T., 1993. Experimental cpx/melt partitioning of 24 trace elements. Contrib. Mineral. Petrol. 113, 1–8.

Hirschmann, M.M., Stolper, E.M., 1996. A possible role for garnet pyroxenite in the origin of the “garnet signature” in MORB. Contrib. Mineral. Petrol. 124, 185–208.

Jousselin, D., Morales, L.F.G., Nicolle, M., Stephant, A., 2012. Gabbro layering induced by simple shear in the Oman ophiolite Moho transition zone. Earth Planet. Sci. Lett. 331–332, 55–66. http://dx.doi.org/10.1016/j.epsl.2012.02.022.

Jousselin, D., Nicolas, A., 2000. Oceanic ridge off-axis deep structure in the Mansah region (Sumail massif, Oman ophiolite). Mar. Geophys. Res. 21, 243–257.

Jousselin, D., Nicolas, A., Boudier, F., 1998. Detailed mapping of a mantle diapir below a paleo-spreading center in the Oman ophiolite. J. Geophys. Res. 103, 18153–18170.

Kelemen, P.B., Koga, K., Shimizu, N., 1997. Geochemistry of gabbro sills in the crust–mantle transition zone of the Oman ophiolite: implications for the origin of the oceanic lower crust. Earth Planet. Sci. Lett. 146, 475–488.

Kelemen, P.B., Shlmlzu, N., Saltersi, V.J., 1995. Extraction of mid-ocean-ridge basalt from the upwelling mantle by focused flow of. Nature 375, 29.

Koepke, J., Schoenborn, S., Oelze, M., Wittmann, H., Feig, S.T., Hellebrand, E., Boudier, F., Schoenberg, R., 2009. Petrogenesis of crustal wehrlites in the Oman ophiolite: experiments and natural rocks. Geochem. Geophys. Geosyst. 10. http://dx.doi.org/10.1029/2009GC002488.

Koga, K.T., Kelemen, P.B., Shimizu, N., 2001. Petrogenesis of the crust–mantle tran-sition zone and the origin of lower crustal wehrlite in the Oman ophiolite. Geochem. Geophys. Geosyst. 2, 1038.

Kogiso, T., Hirschmann, M.M., Pertermann, M., 2004a. High-pressure partial melting of mafic lithologies in the mantle. J. Petrol. 45, 2407–2422. http://dx.doi.org/10.1093/petrology/egh057.

Kogiso, T., Hirschmann, M.M., Reiners, P.W., 2004b. Length scales of man-tle heterogeneities and their relationship to ocean island basalt geochem-istry. Geochim. Cosmochim. Acta 68, 345–360. http://dx.doi.org/10.1016/S0016-7037(03)00419-8.

Lambart, S., Laporte, D., Provost, A., Schiano, P., 2012. Fate of pyroxenite-derived melts in the peridotitic mantle: thermodynamic and experimental constraints. J. Petrol. 53, 451–476. http://dx.doi.org/10.1093/petrology/egr068.

Lissenberg, C.J., Dick, H.J.B., 2008. Melt–rock reaction in the lower oceanic crust and its implications for the genesis of mid-ocean ridge basalt. Earth Planet. Sci. Lett. 271, 311–325. http://dx.doi.org/10.1016/j.epsl.2008.04.023.

MacLeod, C.J., Johan Lissenberg, C., Bibby, L.E., 2013. “Moist MORB” axial magmatism in the Oman ophiolite: the evidence against a mid-ocean ridge origin. Geol-ogy 41, 459–462. http://dx.doi.org/10.1130/G33904.1.

Mahoney, J.J., Natland, J.H., White, W.M., Poreda, R., Bloomer, S.H., Fisher, R.L., Baxter, A.N., 1989. Isotopic and geochemical provinces of the western In-dian Ocean Spreading Centers. J. Geophys. Res., Solid Earth 94, 4033–4052. http://dx.doi.org/10.1029/JB094iB04p04033.

McCulloch, M.T., Gregory, R.T., Wasserburg, G.J., Taylor, H.P., 1981. Sm–Nd, Rb–Sr, and 18O/16O isotopic systematics in an oceanic crustal section: evidence from the Samail Ophiolite. J. Geophys. Res., Solid Earth 1978–2012 (86), 2721–2735.

Nicolas, A., Boudier, F., 1995. Mapping oceanic ridge segments in Oman ophiolite. J. Geophys. Res. 100, 6179–6197.

Nicolas, A., Boudier, F., 2003. Where ophiolites come from and what they tell us. Spec. Pap., Geol. Soc. Am. 373, 137–152. http://dx.doi.org/10.1130/0-8137-2373-6.137.

Nicolas, A., Mainprice, D., 2005. Burst of high-temperature seawater injection throughout accreting oceanic crust: a case study in Oman ophiolite. Terra Nova 17, 326–330. http://dx.doi.org/10.1111/j.1365-3121.2005.00617.x.

Nicolas, A., Mainprice, D., Boudier, F., 2003. High-temperature seawater circulation throughout crust of oceanic ridges: a model derived from the Oman ophio-lites. J. Geophys. Res., Solid Earth (1978–2012) 108. http://dx.doi.org/10.1029/2002JB002094.

Niu, Y., Collerson, K.D., Batiza, R., Wendt, J.I., Regelous, M., 1999. Origin of enriched-type mid-ocean ridge basalt at ridges far from mantle plumes: the East Pacific Rise at 11◦20′N. J. Geophys. Res., Solid Earth 104, 7067–7087. http://dx.doi.org/10.1029/1998JB900037.

O’Nions, R.K., Hamilton, P.J., Evensen, N.M., 1977. Variations in 143Nd/144Nd and 87Sr/86Sr ratios in oceanic basalts. Earth Planet. Sci. Lett. 34, 13–22.

M. Nicolle et al. / Earth and Planetary Science Letters 437 (2016) 138–149 149

Pallister, J.S., Knight, R.J., 1981. Rare-earth element geochemistry of the Samail Ophiolite near Ibra, Oman. J. Geophys. Res., Solid Earth 1978–2012 (86), 2673–2697.

Parmentier, E.M., Morgan, J.P., 1990. Spreading rate dependence of three-dimensional structure in oceanic spreading centres. Nature 348, 325–328. http://dx.doi.org/10.1038/348325a0.

Pearce, J., Cann, J., 1971. Ophiolite origin investigated by discriminant analysis using Ti, Zr and Y. Earth Planet. Sci. Lett. 12, 339–349.

Pearce, J.A., Alabaster, T., Shelton, A.W., Searle, M.P., 1981. The Oman Ophiolite as a Cretaceous Arc-Basin Complex: evidence and implications. Philos. Trans. R. Soc. Lond. Ser. A 300, 299–317. http://dx.doi.org/10.1098/rsta.1981.0066.

Prinzhofer, A., Lewin, E., Allègre, C.J., 1989. Stochastic melting of the marble cake mantle: evidence from local study of the East Pacific Rise at 12◦50′N. Earth Planet. Sci. Lett. 92, 189–206.

Rioux, M., Bowring, S., Kelemen, P., Gordon, S., Dudás, F., Miller, R., 2012. Rapid crustal accretion and magma assimilation in the Oman-U.A.E. ophiolite: high precision U–Pb zircon geochronology of the gabbroic crust. J. Geophys. Res. 117, B07201. http://dx.doi.org/10.1029/2012JB009273.

Rioux, M., Bowring, S., Kelemen, P., Gordon, S., Miller, R., Dudás, F., 2013. Tectonic development of the Samail ophiolite: high-precision U–Pb zircon geochronology and Sm–Nd isotopic constraints on crustal growth and emplacement. J. Geophys. Res., Solid Earth 118, 2085–2101. http://dx.doi.org/10.1002/jgrb.50139.

Shervais, J.W., 2001. Birth, death, and resurrection: the life cycle of suprasubduction zone ophiolites. Geochem. Geophys. Geosyst. 2 (20), 925.

Sobolev, A.V., Hofmann, A.W., Kuzmin, D.V., Yaxley, G.M., Arndt, N.T., Chung, S.-L., Danyushevsky, L.V., Elliott, T., Frey, F.A., Garcia, M.O., Gurenko, A.A., Kamenetsky, V.S., Kerr, A.C., Krivolutskaya, N.A., Matvienkov, V.V., Nikogosian, I.K., Rocholl, A., Sigurdsson, I.A., Sushchevskaya, N.M., Teklay, M., 2007. The amount of recycled crust in sources of mantle-derived melts. Science 316, 412. http://dx.doi.org/10.1126/science.1138113.

Staudigel, H., Park, K.H., Pringle, M., Rubenstone, J.L., Smith, W.H.F., Zindler, A., 1991. The longevity of the South Pacific isotopic and thermal anomaly. Earth Planet. Sci. Lett. 102, 24–44. http://dx.doi.org/10.1016/0012-821X(91)90015-A.

Stracke, A., Salters, V.J., Sims, K.W., 2000. Assessing the presence of garnet-pyroxenite in the mantle sources of basalts through combined hafnium–neodymium–thorium isotope systematics. Geochem. Geophys. Geosyst. 1.

Sun, S.S., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. Geol. Soc. (Lond.) Spec. Publ. 42, 313–345. http://dx.doi.org/10.1144/GSL.SP.1989.042.01.19.

Toomey, D.R., Jousselin, D., Dunn, R.A., Wilcock, W.S.D., Detrick, R.S., 2007. Skew of mantle upwelling beneath the East Pacific Rise governs segmentation. Na-ture 446, 409–414. http://dx.doi.org/10.1038/nature05679.

White, W.M., McBirney, A.R., Duncan, R.A., 1993. Petrology and geochemistry of the Galapagos Islands: portrait of a pathological mantle plume. J. Geophys. Res. (ISSN 0148-0227) 98, 19533. http://dx.doi.org/10.1029/93JB02018.

Wood, B.J., Blundy, J.D., 1997. A predictive model for rare earth element parti-tioning between clinopyroxene and anhydrous silicate melt. Contrib. Mineral. Petrol. 129, 166–181. http://dx.doi.org/10.1007/s004100050330.