Metasomatism in the Lithospheric Mantle beneath Middle Atlas (Morocco) and the Origin of Fe- and Mg-rich Wehrlites
1CNR–Istituto di Geoscienze E Georisorse, Sezione di Pavia, Via Ferrata 1, I-27100 Pavia, Italy
2Dipartimento di Scienze Della Terra, Università di Pavia, Via Ferrata 1, I-27100 Pavia, Italy
3Institut Universitaire Européen de La Mer, UniversitÉ de Bretagne Occidentale, UMR 6538, Place Copernic, F-29280, Plouzane, France
4Laboratoire de Géologie—UMR 6524, Université Blaise Pascal, 5 Rue Kessler, F-63038, Clermont-Ferrand Cedex, France
RECEIVED JUNE 21, 2008; ACCEPTED NOVEMBER 21, 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND TEXTURESANALYTICAL TECHNIQUESMINERAL GEOCHEMISTRYDISCUSSIONCONCLUDING REMARKSSUPPLEMENTARY DATAREFERENCES |
|---|
Mantle xenoliths enclosed in the Plio-Quaternary alkaline basalts from Bou Ibalrhatene, Middle Atlas, Morocco, are characterized by a wide range of lithological and chemical heterogeneity, consistent with metasomatism of their lithospheric mantle source. Subordinate porphyroclastic to protogranular spinel-lherzolites associated with websterites exhibit major and trace element signatures, along with a depleted mantle isotopic affinity, that testify to ancient melt extraction processes, possibly during Neo- to Paleoproterozoic times. These samples show large ion lithophile element (LILE) enrichments that have been imparted a long time after the accretion of their protoliths to the lithosphere, induced by alkaline melts rising from upwelling HIMU-like asthenosphere since the late Cretaceous or Eocene. Extensive sectors of the percolated lherzolite and harzburgite mantle progressively approached chemical equilibrium with the migrating melts, forming porphyroblastic amphibole-rich ultramafic rocks with strong LILE enrichment, highly variable (enriched to strongly depleted) high field strength element (HFSE) abundances and HIMU-like isotopic signatures. As a result, the older depleted lithosphere was progressively refertilized. The presence among the Bou Ibalrhatene ultramafic xenoliths of both Fe- and Mg-rich wehrlites is a rare occurrence in xenolith suites. The Fe-rich wehrlites formed by the migration through dunitic channels, which opened during earlier reactive porous flow, of larger and larger volumes of rising alkaline melts, becoming progressively less evolved in composition and more similar to the erupted lavas. In contrast, subordinate Mg-rich wehrlites characterized by the presence of apatite, extreme LILE enrichment and HFSE depletion testify to the localized chemical effects of either carbonatite melts related to the oldest (i.e. Eocene) volcanic activity in the area or, more probably, to highly evolved SiO2-saturated silicate melts that vanished through interaction with the lithosphere. As a whole, the textural, petrographic and chemical characteristics of the Bou Ibalrhatene xenoliths suggest prolonged thermo-chemical erosion of an originally heterogeneous lithosphere with a dominant depleted mantle isotopic signature by melts rising from a HIMU-like asthenospheric mantle source in response to intraplate tectonic reactivation and rifting of the Pan-African basement. The overall scenario envisaged for lithosphere–asthenosphere interaction beneath the Middle Atlas is similar to that inferred for other Cenozoic Central–Northern African volcanic centres, such as Hoggar (Algeria) and Gharyan (Libya). The majority of the mantle sectors beneath the NE–SW volcanic alignment associated with the Trans-Moroccan fault system consist of rejuvenated lithosphere formed by small-scale, shallow-mantle upwelling and accretion of asthenospheric material since late Cretaceous or Eocene times.
KEY WORDS: lithospheric mantle; metasomatism; wehrlites; trace elements; Sr–Nd isotopes; Morocco; mantle xenoliths
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND TEXTURESANALYTICAL TECHNIQUESMINERAL GEOCHEMISTRYDISCUSSIONCONCLUDING REMARKSSUPPLEMENTARY DATAREFERENCES |
|---|
Mantle xenoliths brought to the surface by Cenozoic within-plate, anorogenic volcanism south of the Maghrebian orogenic system provide an important opportunity to constrain the composition and evolution of the North African lithospheric mantle. At present, detailed investigations of major xenolith occurrences are still scarce and more chemical information is needed to develop a comprehensive model for the interaction between the lithosphere and the underlying convecting mantle in this area, as well as to constrain the role of the lithospheric mantle as a magma source component.
Detailed studies of mantle xenolith lithologies have been reported only for the In Teria and Manzaz (Hoggar, Algeria) districts (Dautria et al., 1992
; Beccaluva et al., 2007) and the Gharyan (Libya) volcanic field (Beccaluva et al., 2008). Abundant peridotite–pyroxenite xenoliths are also entrained within Cenozoic volcanic rocks from the Middle Atlas (Morocco). Although their presence and importance for mantle studies has been revealed by preliminary studies (Moukadiri, 1983; Moukadiri & Kornprobst, 1984), complete chemical information, including trace element and isotope data, is not yet available for this mantle xenolith occurrence, thus preventing a comparison with the corresponding Algerian and Libyan mantle sectors.
At Bou Ibalrhatene, in the volcanic district of Azrou–Timhadite (Middle Atlas, Morocco), a 1 km wide maar provides an opportunity to sample metasomatized amphibole-rich (
3%, up to 15%) and amphibole-poor (
1%) peridotite xenoliths (lherzolites and harzburgites), the latter characterized by protogranular to porphyroclastic textures (Mercier & Nicolas, 1974); these are associated with abundant wehrlites and granular to coarse-granular spinel (± garnet) websterites and clinopyroxenites or hornblendites with cumulus textures. The extreme heterogeneity of the mantle lithologies sampled by the Bou Ibalrhatene xenoliths provides a unique opportunity to reconstruct the compositional evolution of the lithospheric mantle beneath the area and, particularly, to track metasomatism-induced mineralogical and chemical changes caused by pervasive infiltration and channelling of large volumes of Neogene magmas within the pre-existing lithosphere.
Further interest in the Bou Ibalrhatene mantle xenolith occurrence arises from the presence of coexisting high-Mg and high-Fe wehrlites, which are characterized by a wide range of compositional and textural variability. The presence of both high-Mg and high-Fe wehrlites in a single xenolith suite is uncommon. Mantle wehrlites have been ubiquitously interpreted as the result of the percolation of melts rising through refractory lithospheric peridotites (Yaxley et al., 1991; Hauri et al., 1993; Xu et al., 1996; Coltorti et al., 1999). Therefore, they can provide valuable information on the composition of the metasomatic melts circulating through the lithospheric mantle and on the physico-chemical parameters driving the percolation process.
Literature data show that wehrlites can be very variable in terms of textures and chemical compositions, suggesting that they can be produced from a large compositional spectrum of melts (alkaline and carbonatite melts, H2O–CO2-rich melts or fluids). Wehrlites with moderate to strong enrichment in Fe occur worldwide (Zinngrebe & Foley, 1995; Xu et al., 1996; Neumann & Wulff-Pedersen, 1997; Wulff-Pedersen et al., 1999; Ionov et al., 2005). Interestingly, the geochemical and mineralogical characteristics of these rocks testify to a complex evolution and their interpretation often involves multi-stage metasomatic processes (e.g. Xu et al., 1996). The main focus of research has been mainly on the characterization of magnesian wehrlites sensu lato (i.e. with FoOl >90), which have mainly been attributed to metasomatism caused by Na-carbonatite melts rising from a deep mantle source (Yaxley et al., 1991). On the other hand, recent modelling has demonstrated that melt–peridotite interaction can strongly modify the composition of the migrating melts (Bedini et al., 1997; Xu et al., 1998; Ionov et al., 2002a). In this context, alternative hypotheses have been proposed to explain the occurrence of magnesian wehrlites. These involve impregnation processes caused by: (1) highly differentiated H2O–CO2-rich silicate melts; the ‘end-product’ after extensive reactive porous flow within ambient peridotite mantle (Laurora et al., 2001); (2) carbonate rich-fluids separated by immiscibility from residual, volatile-rich, silicate melts that evolved via reactive porous flow through the host peridotite (Zanetti et al., 1999; Morishita et al., 2003); (3) low-aSiO2 silicate melts leading to orthopyroxene consumption and crystallization of clinopyroxene + olivine assemblages (Zinngrebe & Foley, 1995; Rivalenti et al., 2004).
In this study we present the results of combined electron microprobe analysis (EMPA), laser ablation inductively coupled plasma (LA-ICP-MS) and thermal ionization mass spectrometry (TIMS) investigations of carefully selected Bou Ibalrhatene xenoliths. The aim is to understand the sequence of upper mantle processes within the North African lithospheric mantle, as well as to constrain the mechanisms of metasomatism and the nature of the metasomatic components involved. Moreover, events pre-dating those related to the Neogene volcanic activity that brought the xenoliths to the surface are also addressed to provide a comprehensive evolutionary history for the lithospheric mantle beneath the Middle Atlas, involving processes of thermo-chemical erosion since late Cretaceous–early Tertiary times.
|
GEOLOGICAL SETTING |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND TEXTURESANALYTICAL TECHNIQUESMINERAL GEOCHEMISTRYDISCUSSIONCONCLUDING REMARKSSUPPLEMENTARY DATAREFERENCES |
|---|
The Middle Atlas constitutes part of the Atlas Mountains of North Africa, which extend
2000 km from the Atlantic coast of Morocco to Tunisia (Fig. 1). This area is delimited by the Southern Rif and the Taza-Oujda Pass to the north, by the central part of the High Atlas to the south, by the Eastern Meseta to the east and by the Western Meseta to the west (Moukadiri, 1983). The region is about 13^200 km2 in area with three distinct units recognized on the basis of morphological and geological features (Du Dresnay, 1988). The northernmost unit is an uplifted Liassic Calcareous Plateau, which bounds the Sais Plain to the south. The central part is formed by a folded range, which is separated from the Calcareous Plateau by the Northern Middle Atlas tectonic line. To the south, the Haute Moulouya valley separates the Middle from the High Atlas.
|
A geochronological survey (K–Ar) of the volcanic rocks of the Middle Atlas and Haute Moulouya valley indicates three stages of alkaline mafic magmatism (Harmand & Cantagrel, 1984
). The oldest volcanic rocks (35 Ma) are located along the northern boundary of the High Atlas mountains and mainly consist of melanephelinitic lavas that are thought to be related to the syenite–carbonatite complex of Bou-Agrao. During the middle and late Miocene (from 15 to 6 Ma), scattered volcanic activity occurred in the Middle Atlas, simultaneously with Rif orogenesis and the uplift of central Morocco. The lavas have nephelinitic to melanephelinitic compositions. Subsequently, during mid-Quaternary times (1·8–0·5 Ma), a volcanic chain (120 km long) developed along a NW–SE-trending (170°N) axis coincident with the Trans-Moroccan fault system (Harmand & Cantagrel, 1984).
The xenoliths which are the focus of this study were collected from maar deposits located near Ibalrhatene, in the Azrou–Timhadite volcanic district (the main part of the Middle Atlas), which formed during late Pliocene and Quaternary times and includes hundreds of small volcanic vents (mainly strombolian cones, scoria cones and maars; Harmand & Moukadiri, 1986). The composition of the lavas is exclusively alkaline with moderate to strong silica undersaturation (alkali basalts, basanites, nephelinites).
|
ROCK TYPES AND TEXTURES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND TEXTURESANALYTICAL TECHNIQUESMINERAL GEOCHEMISTRYDISCUSSIONCONCLUDING REMARKSSUPPLEMENTARY DATAREFERENCES |
|---|
The 36 samples selected for this study are part of a large collection of ultramafic xenoliths (
150 samples, from 5 to 20 cm in diameter) collected from the Bou Ibalrhatene maar deposits. These xenoliths show a wide range of variability in their mineral mode and textural characteristics. Among the samples selected for detailed petrographic and geochemical study, lherzolites and harzburgites dominate, whereas pyroxenites and wehrlites are subordinate. Dunites are also common in the field, but poorly represented in our sample collection, which is biased towards clinopyroxene- and amphibole-bearing samples.
In all the amphibole-bearing lithologies, melt pockets occur containing small euhedral phenocrysts of olivine, clinopyroxene and spinel, especially around amphibole and spinel, but sometimes even within large olivine and orthopyroxene porphyroclasts. Petrographic evidence indicates that these melts are related to closed-system incongruent melting of a mineralogical association modally dominated by amphibole (see also Laurora et al., 2001; Ban et al., 2005; Bali et al., 2008). An exception is represented by the glassy vein in wehrlite IBA 3; its characteristic features are outlined below.
About 36% of the studied samples are amphibole-free to -poor (amphibole 0–1 vol.%) spinel lherzolites and harzburgites. Amphibole-free protogranular spinel lherzolites (Mercier & Nicolas, 1974) are found in composite xenoliths, in association with granular spinel websterites (IBA 6 and IBA 7) and spinel ± garnet websterites (IBA 21). More frequently, amphibole-free to -poor peridotites show porphyroclastic (lherzolites IBA 8, IBA 29 and IBA 79; harzburgite IBA 19) and coarse-granular (lherzolites IBA 4, IBA 72, IBA 77, IBA 87 and IBA 89; harzburgites IBA 24 and IBA 71) textures (Table 1).
|
About 28% of the selected samples consists of strongly recrystallized amphibole-rich spinel lherzolites (IBA 10, IBA 57, IBA 70, IBA 80, IBA 82, IBA 88) and harzburgites (IBA 16, IBA 17, IBA 27, IBA 56, IBA 68), with amphibole contents in the range 5–10 vol.% and 5–13 vol.%, respectively, with porphyroblastic to poikiloblastic textures. They are characterized by large orthopyroxene (up to 9 mm, frequently exsolved) and amphibole (up to 6·5 mm) crystals, which include small rounded grains of all the other peridotite-forming minerals.
Discrete pyroxenite xenoliths have also been found (11% of the samples). They exhibit cumulus (spinel clinopyroxenite and hornblendite IBA 18 and websterite IBA 9) and coarse- to fine-granular (spinel websterites IBA 14 and IBA 28, with amphibole <1 to 2 vol.%, similar to composite samples) textures. The IBA 18 cumulate is characterized by the presence of banding, defined by the alternation of coarse (grain size up to 5 mm) and pegmatitic (grain size up to 1 cm) layers. Anhedral to subhedral amphibole and clinopyroxene are the cumulus minerals, whereas the interstices are occupied by anhedral amphibole, spinel and, occasionally, clinopyroxene. Websterite IBA 9 (Table 1b) exhibits relics of an early cumulus texture, which has been partially replaced by a porphyroclastic texture. Orthopyroxene, as the early cumulus phase, is often embedded in large, poikilitic, clinopyroxene. Interstitial clinopyroxene and subordinate amphibole and opaque minerals formed during the late stages of magmatic crystallization. Incipient subsolidus recrystallization is evidenced by the re-equilibration of magmatic pyroxene forming a pseudomorphic symplectitic intergrowth of two-pyroxenes ± olivine.
Wehrlites and wehrlite-bearing composite xenoliths (18% of the samples) can be subdivided into Fe-rich and Mg-rich types (Table 1). Fe-rich wehrlites are both spinel-bearing (IBA 3, IBA 32, IBA 41, IBA 53 and IBA 69) and spinel-free (IBA 5). They are characterized by an early porphyroclastic texture (Fig. 2a) made up of olivine porphyroclasts that have been partially replaced by the pervasive (frequently multi-stage) crystallization of cpx + amph ± sp (which form, along with olivine, the mineral assemblage hereafter referred to as the ‘matrix assemblage’) with coarse-granular to equigranular textures (Fig. 2b).
|
In some samples, the late recrystallization stage is characterized by the development of olivine-free structures, such as globular clusters (up to 1 cm diameter: IBA 32 and IBA 53 samples, Fig. 2c) or veins (IBA 3, Fig. 2d). Clusters comprise poikilitic clinopyroxene full of small euhedral spinels, which are partially overgrown by amphibole (hereafter referred to as ‘cluster assemblages’). Petrographic evidence (e.g. the sharp boundary between clusters and matrix minerals) suggests that the globular clusters have grown isotropically through dissolution of the wehrlite ‘matrix mineral assemblage’. They are interpreted as original melt pockets interconnected by a network of thin interstitial veins; the latter now recorded by trails of very small grains of cluster minerals between the large matrix minerals. Clinopyroxene usually ranges from 15 to 20 vol.%, but it reaches 35 vol.% in the spinel-free IBA 5 wehrlite. Amphibole (3–15 vol.%) is anhedral to subhedral and mainly concentrated in the olivine-free clusters (IBA 41, IBA 32, IBA 53), as well as in veins (IBA 3, IBA 5). Generally, clinopyroxene and amphibole grown during the early, pervasive stages of the wehrlitization process are in textural equilibrium, whereas they show reaction relationships when occurring in clusters and veins. In wehrlite IBA 5, calcite globules (smaller than 1 mm) occur in the interstices or within olivine and clinopyroxene crystals affected by cracks. In IBA 3, the influx of exotic melts is documented by the presence of a vein, a few millimetres across, cross-cutting the peridotite and containing very fresh glass. In this vein, magmatic anhedral to subhedral amphibole is associated with abundant subhedral to euhedral clinopyroxene. Unlike melt pockets around amphiboles and spinels occurring in the other amphibole-bearing samples from Bou Ibalrhatene, no magmatic olivine or spinel was observed.
The second group of wehrlites (Mg-rich) is documented only by sample IBA 26, which is a composite xenolith consisting of an amphibole-bearing harzburgitic part (IBA 26a) and a wehrlitic part (IBA 26b). The two parts are connected by a transition zone in which both pyroxenes are present and clinopyroxene overgrows orthopyroxene. The harzburgitic part has a rather fine-grained foliated porphyroclastic texture. Porphyroclast relics are remnants of larger crystals that have been dismembered and dislocated by strong deformation. In thin cataclastic layers, the crystallization of secondary amphibole, orthopyroxene and euhedral apatite is observed. Locally, some secondary orthopyroxenes occur as large poikilitic crystals that include small grains of rounded olivine and amphibole. Amphibole shows evidence of extensive melting, which sometimes clearly involves also orthopyroxene. Apatite ranges from perfectly euhedral to vermiform when interstitial: generally, it appears to have behaved as a refractory phase, as previously reported by Chazot et al. (1996b). In the transition zone, clinopyroxene is poikiloblastic (including small grains of olivine and amphibole) to interstitial.
The wehrlitic part (IBA 26b) is mainly finer-grained than the harzburgitic part. The texture varies from porphyroclastic to poikiloblastic (Fig. 2e and f), in which clinopyroxene oikocrysts embed rounded olivine, amphibole, spinel and apatite. The latter also occur interstitially between secondary olivine with subhedral to euhedral shape. Sometimes, clinopyroxene oikocrysts surround relics of deformed olivine, with rounded to corroded boundaries. This observation strongly suggests that the peridotite–melt reaction resulted in olivine digestion at the time of clinopyroxene formation.
|
ANALYTICAL TECHNIQUES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND TEXTURESANALYTICAL TECHNIQUESMINERAL GEOCHEMISTRYDISCUSSIONCONCLUDING REMARKSSUPPLEMENTARY DATAREFERENCES |
|---|
The major element compositions of olivine, clinopyroxene, orthopyroxene, amphibole, spinel and apatite were determined using a CAMECA SX100 electron microprobe at the Laboratory of Geology of the University Blaise Pascal, Clermont-Ferrand (France) and a JEOL JXA-8600 installed at the CNR–Institute of Geosciences and Earth Resources, Firenze. Both electron microprobes are equipped with four wavelength-dispersive spectrometers, and energy-dispersive spectrometer(s). Analyses have been performed using a 15 kV accelerating voltage, 10–15 nA primary current and a 2–3 µm spot size. Standardization was done using natural minerals and glasses. Precision is better than 3% for many major elements and 5% for minor elements. The sum of the squares of the differences between observed and recommended values for the standards (R2) is typically lower than unity. Full details of the analytical procedures have been reported by Vaggelli et al. (1999
).
Trace elements in pyroxenes, amphibole and apatite were determined by LA-ICP-MS at the CNR–Institute of Geosciences and Earth Resources, Pavia using a double focusing sector field analyser (Finnigan Mat Element) coupled with a Q-switched Nd:YAG laser source (Quantel Brilliant) operating at 266 nm. Helium was used as the carrier gas and was mixed with Ar downstream of the ablation cell. Spot diameter was adjusted in the range 40–80 µm. Quantification was done using the NIST SRM 612 standard, with 44Ca as an internal standard for clinopyroxene, amphibole and apatite, 29Si for orthopyroxene and 25Mg for spinel. Detection limits were typically in the range 100–500 ppb for Sc, Ti and Cr, 10–100 ppb for V, Rb, Sr, Zr, Cs, Ba, Gd and Pb, 1–10 ppb for Y, Nb, La, Ce, Nd, Sm, Eu, Tb, Dy, Er, Yb, Hf and Ta, and usually <1 ppb for Pr, Ho, Tm, Lu, Th and U. Precision and accuracy (both better than 10%) were assessed from repeated analyses of the BCR-2g standard. Full details of the analytical parameters and quantification procedures have been given by Tiepolo et al. (2003).
Sr and Nd isotope compositions and Sm and Nd concentrations were measured in the Laboratoire de Géologie, Clermont-Ferrand (France), on clinopyroxene (seven samples) and amphibole (three samples) mineral separates. Selected samples were crushed using a stainless-steel mortar and pestle, sieved and split into coarse (>500 µm grain size), medium (500–250 µm) and fine (250–125 µm) fractions. Then, clinopyroxenes and amphiboles were concentrated from the medium and fine fractions with a Frantz magnetic separator, and finally hand-picked under a binocular microprobe. Prior to analysis, the mineral separates were acid-leached with 2M HCl for 30 min at 70°C, and 70–200 mg (depending on Nd concentration) were taken for analysis. Spiking with a mixed 149Sm–150Nd tracer, sample decomposition with HF, and chemical separation of the analytes by combined cation exchange and extraction chromatography with Sr Spec, TRU Spec and Ln Spec materials followed the methods described in detail by Pin et al. (1994) and Pin & Santos Zalduegui (1997). A fully automated, upgraded VG 54E mass spectrometer was used in dynamic triple collection mode for measuring 87Sr/86Sr and 143Nd/144Nd ratios. Sm isotope dilution measurements were performed with the same instrument operated in single collection mode. The average results and standard deviations obtained on pure element isotopic reference materials were: 87Sr/86Sr = 0·710253, SD = 15 x 10–6 (n = 7) for the SRM 987 Sr carbonate from NIST, 143Nd/144Nd = 0·511960, SD = 12 x 10–6 (n = 6) for the French AMES R Nd standard, and 143Nd/144Nd = 0·512110, SD = 10 x 10–6 (n = 5) for the Japanese JNdi-1 standard. This last value is equivalent to 143Nd/144Nd = 0·511853 for the La Jolla standard (Tanaka et al., 2000).
Representative mineral composition data are reported in Tables 1–12. The full datasets are provided as Supplementary Data Tables SD1–SD12 at http://petrology.oxfordjournals.org/.
|
|
|
|
|
|
|
|
|
|
|
|
MINERAL GEOCHEMISTRY |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND TEXTURESANALYTICAL TECHNIQUESMINERAL GEOCHEMISTRYDISCUSSIONCONCLUDING REMARKSSUPPLEMENTARY DATAREFERENCES |
|---|
Major elements
Olivine from the Ibalrhatene xenoliths (Table 2) shows a wide range of Fo contents [molar (100 x Mg/(Mg + FeT)] (Fig. 3a). These increase from 87·4–88·8 in Fe-rich wehrlites to 88·8–89·9 in pyroxenites and strongly recrystallized amphibole-rich peridotites, and to 90·2–91 in amphibole-poor spinel lherzolites and anhydrous spinel harzburgites. No significant variation was observed between the harzburgite and wehrlite parts of the composite xenolith IBA 26, which has high Fo (90·3 ± 0·1). The range of minor elements is similar in all the samples (i.e. NiO = 0·20–0·31 wt %). Microcrystalline olivine in the melt pockets has systematically higher Fo compared with that of the peridotite matrix, as previously observed elsewhere (e.g. Hauri et al., 1993
; Coltorti et al., 1999).
|
The variation of Mg-number [molar 100 x Mg/(Mg + FeT)] values in orthopyroxene (Table 3) broadly corresponds to that of the coexisting olivine. Orthopyroxene from the harzburgitic part of IBA 26 has the highest Mg-number (91·5). The Al content [from 5·3 (IBA 8) down to 0·5 (IBA 26) wt %] and Mg-number of peridotite orthopyroxene are inversely correlated; moreover, the Al content is inversely correlated with the Cr-number of spinel (Fig. 3b). In pyroxenite orthopyroxene, Al reaches the maximum values in the inner part of the spinel ± garnet websterite IBA 21 (6·4 wt %).
Based on the wollastonite index [Wo = molar (100 x Ca/(Ca + Mg + Fe2+ + Fe3+ + Mn); Morimoto (1988)], lherzolite and harzburgite clinopyroxene (Tables 4) ranges from augite to diopside (Wo = 44–49) in composition. Mg-number in lherzolite clinopyroxene ranges from 88 to 90; higher values are shown by clinopyroxene from amphibole-rich (90–91) and anhydrous (91–91·4) harzburgites (Fig. 3c). Ca contents increase from lherzolite (0·74–0·83 a.p.f.u.) to harzburgite (0·81–0·87 a.p.f.u.) xenoliths.
Clinopyroxene from amphibole-poor lherzolites has higher Ca, Na, Al and Ti, and lower Cr and Fe relative to that in amphibole-rich samples (Fig. 4a and b). In contrast, clinopyroxene from harzburgites has comparable Na, Al, Ti and FeT, regardless of the amphibole mode; Cr is higher in anhydrous than in hydrous harzburgites (Fig. 4a and b).
|
Matrix clinopyroxene in the Fe-rich wehrlites is diopside (Wo = 45–47) in samples IBA 5, IBA 41 and IBA 53, and augite (Wo = 43–44) in samples IBA 3, IBA 32 and IBA 69 (Table 4). Cluster clinopyroxene and euhedral microcrystalline grains in the IBA 3 glassy vein are diopside (Wo = 46–52). In IBA 3 and IBA 32, clinopyroxene shows chemical variations related to its textural position, as well as some core–rim variations (Fig. 5a and b). Clinopyroxene from the Mg-rich wehrlite is diopside and, relative to its analogue from the Fe-rich samples, it is enriched in Si and Cr, and depleted in Al and Ti (Fig. 5a and b). Also, FeT (<0·10 a.p.f.u.) is rather low and Mg-number is close to 0·90.
|
Clinopyroxene from pyroxenites is diopside (Wo = 45·1–48·5), except for IBA 9 (Wo = 43–44). It shows a wide range of Mg-number values, with the lowest occurring in cumulus samples, IBA 18 (Mg-number = 70·6–73) and IBA 9 (Mg-number
83), and the highest in granular websterites (Mg-number = 88·3–90·2). Clinopyroxene from IBA 18 is extremely depleted in Cr (0·001 a.p.f.u.). Al is up to 9 wt % in clinopyroxene from the inner part of the IBA 21 websterite. Euhedral clinopyroxene crystallized in the melt pockets has distinctively higher Ti, Cr and Al, and lower Na contents than that in the peridotite matrix (Fig. 5a and b).
According to the classification scheme of Leake et al. (1997), amphibole in the lherzolites and most harzburgites is a titanian pargasite, whereas it varies from titanian pargasite to edenite in the IBA 68 and IBA 56 amphibole-rich porphyroblastic harzburgites (Table 5), with a progressive decrease in Al, associated with an increase in Si and alkalis (Fig. 6). The variability of amphibole composition is clearly documented by the variation of Ti and Na vs Si (Fig. 7a and b). The amphiboles in the lherzolites and most harzburgites have comparable Na and lower Ti contents relative to amphiboles in peridotite xenoliths from Caussou in the Pyrenees metasomatized by alkaline melts (Fabriès et al., 1989). In contrast, amphiboles from amphibole-rich harzburgites IBA 56 and IBA 68 have Na and Ti contents overlapping those of amphiboles from Yemen apatite-bearing peridotites (Chazot et al., 1996b).
|
|
Amphibole in the Fe-rich wehrlites is titanian pargasite, but in IBA 3 some grain rims from the glassy vein are kaersutites (Ti >0·5 a.p.f.u.). Amphibole shows a large range in K2O content (1–2·7 wt %), which is negatively correlated with Na2O (1·7–3·5 wt %) (Table 5). In IBA 3, the matrix amphibole has high SiO2 (
42·9 wt %) and Na2O (2·8 wt %), and low TiO2 (2·8–3·2 wt %) concentrations (Fig. 7a and b), similar to the IBA 56 amphiboles. Similar SiO2 and Na2O concentrations are observed in amphibole cores from pseudo-vein, whereas rims are characterized by the highest TiO2 (4·6–5·1 wt %) and lowest SiO2 (40–41·8 wt %) contents (Fig. 7a and b).
In the IBA 26 sample, amphibole is edenite to sub-calcic edenite (Table 5); this latter has Ca <1·5 a.p.f.u., but Na <0·5 a.p.f.u., and therefore does not lie in any compositional field defined by Leake et al. (1997). Relative to amphibole from Fe-rich wehrlites, it possesses higher Mg-number (0·90), and is enriched in SiO2, Cr2O3 and Na2O, and depleted in K2O, Al2O3 and TiO2. The compositions of these amphiboles approach that of amphiboles from other apatite-bearing mantle occurrences, such as Yemen (Chazot et al., 1996b), Mt. Leura (SE Australia; Yaxley et al., 1998) and Finero (Zanetti et al., 1999).
Amphibole from IBA 18 cumulus pyroxenite has the lowest Cr2O3 (0·01–0·07 wt %) contents and Mg-number (0·65) values (Table 5).
Spinel from ultramafic xenoliths of Ibalrhatene is characterized by large compositional variability. According to Haggerty's (1991) classification, it ranges from spinel sensu stricto in granular websterites and amphibole-rich harzburgites, to Mg–Al chromite in anhydrous harzburgites and Fe-rich wehrlites, and to chromite in Mg-rich sample IB26 (Cr-number = 81–82). The composition of spinel from cumulus pyroxenite is titanomagnetite (Ti = 0·47–1·1 a.p.f.u.).
Spinel (Table 6) is Mg–Al chromite in Fe-rich wehrlites and chromite in the Mg-rich wehrlite sample. The Cr-number is positively correlated with the Fo content. Apatite (Table 7) from IBA 56 and IBA 6 Mg-rich wehrlite has low FeO (<0·3 wt %), but significant Na2O (up to 1·6 wt %) and MgO (0·85 wt %). Carbonate (Table 7) from the Fe-wehrlite IBA 5 is a calcite. Glass in veins of wehrlite IBA 3 has a composition straddling the basanitic–foiditic fields (Le Bas et al., 1986). Garnets in websterite IBA 21 are zoned, with increasing almandine in the rim. Their major element composition approaches that of their analogues from the Beni Bousera pyroxenites (Kornprobst et al., 1990).
Trace elements
Representative trace element abundances of analysed clinopyroxene and amphibole grains are reported in Tables 8 and 9. Their C1-normalized (Anders & Grevesse, 1989) trace element patterns are reported in Figs 8–18.
|
|
|
|
|
|
|
|
|
|
|
Clinopyroxenes from lherzolites, harzburgites, composite lherzolite–websterite samples and granular websterites can be grouped into two main types (Fig. 8a–f), similar to the subdivision defined by Ionov et al. (2002a
). The first type (Type-1) occurs in the amphibole-free porphyroclastic lherzolites, in the composite protogranular–granular lherzolite–websterite samples and in the granular websterites (Fig. 8a and b); it is characterized by spoon-shaped rare earth element (REE) patterns with progressive light REE (LREE) enrichment and an almost flat heavy REE (HREE) distribution about 10 x C1 abundance level. The increase of LREE content and fractionation is associated with an increase of Th, U and Sr, and a decrease of Nb and Ta contents (Fig. 8a). In the inner part of the websterite IBA 21, the growth of secondary garnet is associated with a selective decrease of HREE and Y in the associated clinopyroxene (Fig. 8b). The second clinopyroxene type (Type-2) is more common and occurs in lherzolites and harzburgites with some dependence on amphibole presence and mode (Fig. 8c–f). Relative to Type-1, Type-2 clinopyroxenes are strongly LREE (50–160 x C1) and middle REE (MREE) (13–80 x C1) enriched over HREE (5–16 x C1), thus displaying an overall larger REE fractionation (LaN/YbN = 5·5–32·8). LaN/YbN values increase from amphibole-rich porphyroblastic lherzolites (5·5–13·6) to amphibole-poor coarse-granular lherzolites (5·9–14·2), amphibole-rich harzburgites (13·2–20·2) and anhydrous harzburgites (up to 32·7).
The trace element patterns of the minerals exhibit steady changes between the different types of lherzolite. In particular, the clinopyroxenes from the porphyroblastic amphibole-rich lherzolites show the highest overall REE and Sr contents, but the lowest Th, U and Ti contents (Fig. 8c). Hf is only slightly higher than or coincident with that of Type-1 peridotite clinopyroxenes, whereas Nb, Ta and Zr vary by more than an order of magnitude, with consequent strong variation of the fractionation between adjacent elements (i.e. Nb and Ta, Zr and Hf) and with respect to the REE. In particular, (Nb,Ta,Zr)/LREE in IBA 57 and IBA 80 samples are relatively large; their values are similar to those expected to characterize clinopyroxenes in equilibrium with alkaline melts. Clinopyroxenes from coarse-granular lherzolites show progressive REE/LILE (large ion lithophile element) and REE/HFSE (high field strength element) fractionation trends relative to that of the porphyroblastic lherzolites. Zr and Hf are nearly completely buffered to the Type-1 clinopyroxene composition and define marked negative anomalies. Nb and Ta contents are lower than in the clinopyroxene from porphyroblastic lherzolites, but still higher than in that from porphyroclastic–protogranular lherzolites. Conversely U, Th and Ti contents increase, approaching the composition of Type-1 clinopyroxene.
Amphibole from granular websterite IBA 14 has a spoon-shaped REE pattern with Zr, Hf, Th and U contents similar to those observed in the clinopyroxene (Fig. 9a). This feature along with the highest Ti and Sr contents, which generate positive Ti and Sr anomalies, points to a substantial equilibrium between the two phases (Vannucci et al., 1995). In contrast, the relatively low U, Th and LREE contents of the amphibole from the composite IBA 6 sample indicate disequilibrium with respect to the composition of the associated clinopyroxene.
The majority of amphiboles from lherzolite and harzburgite samples have highly fractionated REE patterns (Type-2) with variable LREE/HREE fractionation (LaN/YbN = 5·50–35·9); the highest LaN/YbN values are recorded in amphibole-rich harzburgites.
Clinopyroxenes from Fe-rich wehrlites have REE compositions that markedly differ from those of Types-1 and -2 from lherzolites and harzburgites. They have LREE/MREE-enriched convex-upward patterns (Figs 10–12
) with modest overall REE fractionation (LaN/YbN = 1·5–2·5). Only IBA 5 and IBA 53 samples contain clinopyroxene with higher enrichment in La and Ce (LaN/YbN = 4·2–6·8) and steadier REE fractionation. LREE/MREE-enriched convex-upward patterns are also shown by cumulus pyroxenites. More heterogeneity appears when considering the other trace elements. Clinopyroxenes in wehrlite samples IBA 5 and IBA 41 have high Ta contents and weak negative Zr–Hf anomalies. In contrast, the other wehrlites contain cpx with low Ta contents and pronounced Zr–Hf negative anomalies. As also revealed by their major element variation, clinopyroxenes from IBA 3 and IBA 32 show marked texture-related compositional variations (Figs 11 and 12). In pyroxenites, clinopyroxene has contrasted cpx trace element compositions (Fig. 13), with high Zr and Hf and very low Th and U contents (clinopyroxenite IBA 18) or negative Zr and Hf anomalies and high Th and U contents (websterite IBA 9). Generally, the trace element signatures of clinopyroxenes from Fe-rich wehrlites and cumulate samples approach those found in Fe-wehrlites from Tok (Ionov et al., 2006), being close to the compositions expected for clinopyroxene crystallized from typical alkaline melts.
Clinopyroxene from the Mg-rich wehrlite (IBA 26) shows an extreme composition with high trace element contents (LaN up to 217; Fig. 14), marked LREE and MREE enrichment, higher overall REE fractionation (LaN/YbN = 6·6–7·8) relative to clinopyroxenes from Fe-rich wehrlites, and strongly negative Nb, Ta, Zr and Hf anomalies. Similar trace element characteristics have been observed in clinopyroxenes from various xenolith suites recording metasomatism by carbonatite melts (Hauri et al., 1993; Coltorti et al., 1999).
In most samples, the approach of chemical equilibrium between amphibole and the associated clinopyroxene is documented by the trace element abundance patterns (Figs 10, 13, 14 and 15) and confirmed by amphibole–clinopyroxene partition coefficients (Fig. 16), which are close to those previously published for similar samples (e.g. Witt-Eickschen & Harte, 1994; Vannucci et al., 1995; Chazot et al., 1996a). However, in some samples amphibole has a variable trace element composition depending on its textural position in the sample (Fig. 15), and equilibrium with the clinopyroxene is only partially achieved. These relationships will be discussed in more detail below.
Orthopyroxene (Table 10) shows variable trace element composition (Fig. 17), approaching chemical equilibrium with the associated clinopyroxene. Neoblasts with pronounced positive spikes in HFSE (Fig. 18) and orthopyroxene from amphibole-rich porphyroblastic harzburgites are the notable exception (see relevant discussion below: Fig. 19).
|
Spinel (Table 11) from Fe-rich wehrlites is characterized by variable HFSE contents (Nb = 0·28–0·92 ppm; Ta = 0·004–0·07 ppm; Zr = 0·15–1·48 ppm; Hf = 0·004–0·02 ppm; Ti = 0·11–0·66 ppm), significant V (255–850 ppm) and low Sc (0·12–1·92 ppm) contents (Fig. 20a). Spinel from the Mg-rich wehrlite is similar to its analogue from Fe-rich samples, except for the greater enrichment in Cr. Other trace elements are near or below the detection limits. Higher HFSE contents are observed in spinel (ulvöspinel) from cumulus pyroxenite IBA 18.
|
Calcite grains (Table 12) have very low REE abundances ranging from less than 1 to 20 x C1 chondrite values. These concentrations are substantially lower than those reported in the literature for carbonatitic calcites (e.g. Ionov & Harmer, 2002
), being similar to those of carbonates from Spitsbergen (Ionov, 1998) and Gobernador Gregores (Laurora et al., 2001).
Apatites analysed in IBA 26 and IBA 56 samples show extreme enrichment in U, Th, REE (LaN = 20000), Sr and Y, low Ba, Nb and Ta, and very low Zr, Hf and Ti (Table 12; Fig. 20b). Their composition is similar to that reported for apatite in metasomatized mantle xenoliths from Sava’i and Tubuai (Hauri et al., 1994), Yemen (Chazot et al., 1996a) and Mt. Leura (Yaxley & Kamenetsky, 1999).
Secondary garnet from the inner part of websterite IBA 21 shows a characteristic HREE-enriched pattern (HREEN 10), similar to that from Beni Bousera pyroxenites (Pearson et al., 1993; Fig. 20c).
Sr–Nd isotopes
TIMS investigations of Sr and Nd isotope compositions and precise determination of Sm/Nd by isotope dilution have been carried out on acid-leached clinopyroxene and amphibole separates from representative samples of the major lithologies, lherzolite, harburgite and wehrlite, as well as from IBA 28 websterite and IBA 18 cumulate (Table 13). With the exception of samples with Type-1 clinopyroxene that show a prevailing depleted mid-ocean ridge basalt mantle (DMM) signature (Fig. 21), the remaining samples fall in the ocean island basalt (OIB) field and overlap the composition of the Moroccan alkaline lavas (El Azzouzi et al., 1999), regardless of their mode and major and trace element chemistry. This feature strongly suggests that only few sectors of the older lithospheric mantle beneath the Middle Atlas escaped isotopic overprinting by mafic alkaline melt during Neogene events.
|
|
Among samples with Type-2 clinopyroxene the IBA 18 cumulate and the Mg-rich wehrlite have the most unradiogenic Sr and Nd isotopic ratios, respectively. Amphibole and clinopyroxene from the IBA 26 and IBA 41 wehrlite samples are in isotopic equilibrium, whereas their analogues in IBA 5 are not, because of the higher
Nd (
Nd >>1) and the less Sr radiogenic composition of the amphibole. This suggests the effects of multiple events and the possible role of older components more radiogenic in Sr. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND TEXTURESANALYTICAL TECHNIQUESMINERAL GEOCHEMISTRYDISCUSSIONCONCLUDING REMARKSSUPPLEMENTARY DATAREFERENCES |
|---|
The Ibou Ibalrhatene lithospheric mantle before metasomatim
The geochemical data presented above indicate that all the xenoliths studied here underwent variable enrichment in incompatible trace elements as a consequence of the interaction with migrating melts. Nevertheless, valuable information about the petrochemical characteristics of the mantle lithosphere beneath the Middle Atlas before the widespread metasomatism can be gained through the consideration of the textural, mineralogical and geochemical features of the Type-1 xenoliths, in which amphibole is absent or occurs exceptionally and the metasomatic fingerprint is limited to only the most incompatible trace elements, such as LREE, Sr, U and Th. Type-1 samples include amphibole-free porphyroclastic spinel lherzolites IBA 29 and IBA 79, the composite IBA 6, IBA 7 and IBA 21 samples consisting of protogranular spinel lherzolite and granular spinel ± garnet websterite, and the granular spinel websterites IBA 14 and IBA 28.
In addition, some Type-2 xenoliths, despite exhibiting strong enrichments in highly incompatible trace elements, preserve petrographic features and abundances in moderately incompatible elements that can be still ascribed to the pre-metasomatic evolutionary stages. These samples, which can provide complementary information on the more refractory sectors of the ancient lithosphere, include the amphibole-free spinel harzburgites IBA 19, IBA 24 and IBA 71, which are characterized by a coarse-granular to porphyroclastic texture. In these samples, the low modal clinopyroxene is significantly impoverished in fusible major elements and moderately incompatible trace elements with respect to Type-2 peridotites, pointing to larger degrees of partial melting rather than to mineralogical and chemical changes induced by melt–peridotite reaction (for a more detailed discussion see Rampone et al., 2004; Zanetti et al., 2006).
The greater enrichment in highly incompatible elements shown by the minerals from these harzburgites with respect to the more fertile Type-1 analogues most probably results from the percolation of small volumes of LILE-enriched melts, owing to the low capability of harzburgites to induce chemical buffering in the percolating melts (e.g. Bedini et al., 1997; Ionov et al., 2002a; Rivalenti et al., 2007) and the easier possibility that small melt volumes can migrate through pyroxene-depleted peridotites (Toramaru & Fujii, 1987; Faul, 1997; Zhu & Hirth, 2003).
In contrast, the strong enrichment in highly incompatible elements in Type-2 amphibole-poor spinel lherzolites requires a metasomatic process with a relatively large time-integrated melt/rock ratio (see also Xu et al., 1998), which drastically modified the original textural features to the present coarse-granular ones, as well as the modal and chemical composition.
The high Al and Ti contents in pyroxenes and spinel, the high Na content in clinopyroxene and the relatively low Mg-number value shown by all the Type-1 lherzolite minerals, as well as the large clinopyroxene modal content, collectively indicate that the Type-1 peridotites come from a very fertile mantle sector. Moreover, the moderately incompatible element abundances of the pyroxenes overlap estimates for spinel-facies depleted mantle (DM) reservoirs (Fig. 22).
|
The concentrations of fusible elements and moderately incompatible elements decrease in the anhydrous harzburgites. The overall decrease of Na, Ti and Al from lherzolite to harzburgite provides evidence for the progressive extraction of fusible components and the formation of variously refractory assemblages.
When the trace element mineral chemistry is considered, only moderately incompatible elements still record melt extraction processes. An example is provided by the progressive decrease of Yb in cpx with the increasing Mg-number of olivine. An estimate of the extent of ancient melting episodes within the Middle Atlas mantle can be made considering only the less incompatible elements of the REE spectra. Partial melting models (Johnson et al., 1990) applied to amphibole-free harzburgites indicate that they can represent refractory residues after 10% and 20% partial melting of spinel-facies DM source, respectively, if alternatively batch or fractional partial melting are considered (Fig. 22a and b). Unfortunately, the metasomatic modification of the clinopyroxene LREE budget does not allow us to discriminate between the two different typologies of partial melting.
It is worth noting that the clinopyroxene mode in harzburgites ranges between 3 and 5%. According to commonly adopted melting models, these values would imply melting degrees of the order of 20% (e.g. 18–23% according to Niu, 1997), an estimate more akin to batch melting processes. Alternatively, the observed relationships between modal clinopyroxene and HREE abundances level could be ascribed to either trace element re-enrichment during metasomatism and/or decrease of the clinopyroxene mode as a consequence of melt–peridotite reaction.
Regardless of the precise extent of melting processes in the mantle beneath Bou Ibalhratene, the modelling provides convincing evidence for the presence of a variously depleted mantle protolith in this region, well before its sampling by rising alkaline basalts since Pliocene times.
From the textural and chemical evidence based on Type-1 samples and amphibole-free, refractory, IBA 19, IBA 24 and IBA 71 harzburgites, we argue that the lithospheric mantle beneath Bou Ibalrhatene was originally composed of sectors of very fertile, protogranular, lherzolites intercalated with olivine websterite, associated with more refractory parts, made of porphyroclastic to coarse-granular harzburgite.
Equilibrium temperature estimates for Type-1 xenoliths based on the two-pyroxene solvus (i.e. Wells, 1977; Brey & Köhler, 1990) are in the range 830–1030°C. Although no clear systematic variation is apparent as a function of mode or texture, the highest temperatures are recorded by the protogranular–granular ultramafic part of the composite xenolith IBA 6, whereas the lowest temperatures are given by the porphyroclastic lherzolite IBA 29. A range of 770–1030°C is estimated for the amphibole-free (Type 2) harzburgites, and once again the lowest temperatures are given by a porphyroclastic sample (IBA 24). Generally, it is apparent that the porphyroclastic peridotites experienced recrystallization at lower T conditions.
If Mercier et al. (1984) geobarometer, which can reproduce the pressure of experimental runs with an uncertainty of ±0·5 GPa (Yang et al., 1998), is adopted, P estimates are mostly in the range of 1·5–2·7 GPa, being nearly coincident with the Mercier (1980) estimates (Table 1). The application of the Nickel & Green (1985) geobarometer to the garnet–clinopyroxene pairs from websterite IBA 21 results in an equilibration pressure of 2·5 GPa, which is consistent with the 2·4 GPa value estimated using Mercier et al. (1984) geobarometer. These data indicate that the very fertile sectors probably extend down to 75 km depth.
The fO2 of the Type-1 xenoliths is significantly higher than that of abyssal peridotites (log fO2QFM = 0·4–1·3, where QFM is the quartz–fayalite–magnetite buffer), possibly related to the oxidizing activities of migrating melts or fluids. In the amphibole-free Type-2 harzburgites, log fO2QFM is 1·3–2, reflecting the larger metasomatic overprint.
Unfortunately, the present data do not allow us to constrain with much certainty the origin of the olivine websterites. In principle, they could be layers resulting from metamorphic banding that induced pyroxene concentration or from recycling of ancient oceanic crust (Morishita et al., 2003, and references therein); alternatively, they may simply have had magmatic origin. It is commonly observed that magmatic intrusions show large chemical gradients in contrast with the ambient peridotite. In the websterites of this study only the IBA 21 sample shows some zoning (e.g. Al) coupled with REE enrichments in the minerals of the adjacent peridotite. These observations are consistent with early magmatic-induced heterogeneity. According to the MREE–HREE concentration and fractionation of IBA 21 clinopyroxenes, the parent melt had an overall tholeiitic affinity. In general, the nearly complete chemical and textural equilibration with the wall-rock peridotite shown by the composite samples indicates that the formation of the pyroxenites was an older event that occurred largely before the alkaline magmatic activity. This inference is unequivocally supported by the highly radiogenic 143Nd/144Nd and the strongly unradiogenic 87Sr/86Sr of the IBA 28 clinopyroxene, which can be referred to a DM reservoir. The higher 87Sr/86Sr values of clinopyroxene from the IBA 79 lherzolite clearly point to some metasomatic Sr enrichment, as also evidenced by the normalized multi-element patterns.
Age constraints on the ancient melting event can be derived from the Nd isotope data. In this frame, metasomatic enrichment in Nd would imply a lowering of Sm/Nd, increasing the calculated model ages. This is not the case for the IBA 28 clinopyroxenes, because inspection of their REE patterns shows that metasomatism affected only trace elements more incompatible than Ce. The calculated ages range from 2·44 to 1·19 Ga for the IBA 28 websterite and from 1·37 to 0·41 Ga for the IBA 29 lherzolite, depending on the starting mantle reservoir used in the calculation [i.e. primitive mantle (PM) or DM]. Apart from the absolute values, these estimates provide evidence for an old age (possibly Neo- to Paleoproterozoic) of the depletion event(s), much earlier than the Neogene basaltic activity that brought the xenoliths to the surface and than the earlier carbonatite and alkaline magmatism of late Cretaceous to Early Tertiary age known in this region (Charlot et al., 1964; Tisserant et al., 1976; Harmand & Cantagrel, 1984; Wagner et al., 2003).
The occurrence of similar lherzolite–pyroxenite associations is common in mantle sequences involved in the Paleozoic Pan-African orogeny; they have been documented in many mantle xenolith occurrences within the African plate, such as Zabargad (Brueckner et al., 1995), Hoggar (Beccaluva et al., 2007), NE Brazil (Rivalenti et al., 2007) and the Adria Plate (Rampone et al., 1996).
The record of metasomatism in Ibalhratene lherzolites and harzburgites
Type-1 and Type-2 metasomatic imprints recorded by clinopyroxenes and orthopyroxenes from Bou Ibalrhatene lherzolites and harzburgites, as well as by amphiboles when present, were apparently established after both the partial melting events described in the previous section and the emplacement of the granular websterites.
The metasomatic enrichment is stronger in minerals from porphyroblastic lherzolites, porphyroblastic harzburgites and coarse-granular lherzolites. These lithologies are also characterized by extensive textural recrystallization, which in the case of the porphyroblastic samples is accompanied by formation of abundant amphibole. These features combine to indicate that such peridotites experienced metasomatic processes characterized by high time-integrated melt–rock ratios (see Xu et al., 1998). Moreover, clinopyroxenes from Type-2 lithologies have similar Nd and Sr isotopic compositions, which are consistent with those of the Middle Atlas alkaline basalts erupted in Tertiary–Quaternary times (El Azzouzi et al., 1999). This evidence points to a derivation from a common mantle reservoir for both metasomatic melts and the erupted volcanic rocks.
The observation that all the Type-2 trace-element enrichments occur in pervasively metasomatized peridotites without websterite relics, and that no pyroxenite with a Type-2 geochemical signature has been found so far, leads us to infer that the melt migration responsible for Type-2 metasomatism mainly developed through pervasive porous-flow and probably caused the dissolution of pristine websteritic layers. However, the porphyroblastic and coarse-granular peridotites display a range of petrographic features, mineral assemblages, equilibrium temperatures, redox conditions and major- and trace-element compositions, thus indicating some variability in the parameters controlling the metasomatic processes and, in turn, the composition of the migrating melts.
The recognition of petrochemical gradients that may constrain the physico-chemical parameters controlling melt–peridotite interaction highlighted the presence of a steady change in trace element characteristics from the different types of lherzolites (namely, porphyroclastic–protogranular, coarse-granular and porphyroblastic) to harzburgites. These latter show similar Type-2 trace element signatures regardless of their structure. Relative to their analogues from porphyroblastic lherzolites, harzburgite clinopyroxenes are characterized by broadly similar LREE, Sr, U and Th contents, lower HREE abundances and more negative HFSE anomalies. Noticeably, LREE are more constant and do not exhibit chromatographic gradients.
Several features shared by porphyroblastic harzburgites and lherzolites suggest that they underwent a similar chemical modification imparted by a common metasomatic melt; specifically, the low Fo and Cr-number values of olivine and spinel, respectively, the abundance of amphibole porphyroblasts, and the enrichment in highly incompatible trace elements recorded by both clinopyroxene and amphibole. Peculiarly, the melt–peridotite reaction in the porphyroblastic amphibole-rich harzburgites resulted in an increase in modal olivine (in addition to amphibole) and a significant decrease in orthopyroxene. On the other hand, a common more refractory composition of their protholiths is apparent for both anhydrous and porphyroblastic harzburgites from the high Mg-number values (associated with relatively low Al and Na) of pyroxenes and the coherent compositional trends they show in terms of major element chemistry (Fig. 3). A strongly refractory protolith can also account well for the lower HREE contents and the more constant LREE abundances of harzburgite clinopyroxenes, owing to its less efficient buffering control on the percolating melt.
The widespread occurrence of chemical disequilibrium between the minerals of the porphyroblastic harzburgites is worth noting. It occurs between pyroxenes and coexisting olivine and spinel, and is also highlighted by the trace element fingerprint of coexisting pyroxenes and amphibole. In particular, orthopyroxene is strongly enriched in LREE, U and Th relative to its analogue from the porphyroblastic and coarse-granular lherzolites; as a result, it exhibits an overall U-shaped pattern that strictly resembles those of pyroxenes from Type-1 samples (Fig. 17e). These features are tentatively regarded as a record of an early metasomatic stage related to the percolation of strongly evolved melts, probably the same event that imparted the metasomatic fingerprint to Type-1 samples. Conversely, amphiboles may have crystallized from slightly different melts. These latter were possibly also more variable in composition reflected in the range of Si, Al and alkali contents of harzburgite amphiboles, particularly from IBA 56 sample (Figs 6 and 7). T estimates based on Fe–Mg exchange between olivine and spinel (Ballhaus et al., 1991a, 1991b) (i.e. the minerals largely equilibrated with the metasomatic agent) are significantly higher (1020–1190°C) than those calculated for pyroxenes (900–970°C, Brey & Köhler, 1990), which possibly still partially record earlier equilibration conditions. In turn, this implies that heating of mantle assemblages was concurrent with metasomatism.
The progressive change in REE abundance and fractionation exhibited by minerals from Ibalrhatene lherzolites and harzburgites is common in world-wide xenolith occurrences. Among many others, the similarity with Spitsbergen (Ionov et al., 2002a, 2002b) is noticeable. In particular, Type-1 and Type-2 from this latter occurrence almost completely overlap their analogues from the Ibalrhatene xenoliths.
Ionov et al. (2002a) convincingly demonstrated that the whole range of REE patterns in Type-1 and Type-2 samples can be accounted for by a single mechanism of metasomatism that creates different trace element signatures in the peridotite mantle as a function of distance in the percolation systems.
Type-1 REE patterns match reasonably well those produced by geochemical models describing the chromatographic effects of melt percolation (Navon & Stolper, 1987; Bedini et al., 1997; Vernières et al., 1997), whereas Type-2 REE patterns require that xenoliths progressively approach chemical equilibrium with the migrating liquid. In other words, a Type-1 pattern would represent transient metasomatic zoning with variable enrichment in the most incompatible elements formed by an advancing metasomatic percolation front at an increasing distance from the magma source; in contrast, Type-2 patterns are related to an advanced stage of metasomatism when progressive re-equilibration with the migrating liquid results in a marked enrichment of less incompatible elements. Type-1 enrichments are best fitted by considering a small amount of melt entrapment or fractional crystallization in addition to chromatographic chemical exchange (Ionov et al., 2002a). The absence of trace-element zoning in the large clinopyroxenes from low-T Type-1 porphyroclastic lherzolites, reasonably enriched through solid diffusion, indicates that metasomatism was older than 106 years (see also Ionov et al., 2002a).
A cognate origin for the metasomatic melts inducing Type-1 and Type-2 geochemical signatures in the Ibalrhatene xenoliths is supported by the Sr isotopic composition of the clinopyroxenes from porphyroclastic lherzolites IBA 29 and IBA 79, which rapidly converge to typical Type-2 values with an increase in the metasomatic Sr overprint. The apparent delay in the resetting of the Nd isotope composition with respect to the Sr isotope composition is consistent with the lower metasomatic contribution of Nd, determined by the relatively greater compatibility of Nd in clinopyroxene-rich systems (Ionov et al., 2002a, 2002b; Bodinier et al., 2004; Rivalenti et al., 2007).
Given the close similarity of the trace element signatures of the Ibalrhatene xenoliths to those from Spitsbergen and the exhaustive numerical modelling presented by Ionov et al. (2002a), we do not propose here further modelling results to match the geochemical signatures of Ibalhratene lherzolite and harzburgite clinopyroxenes. A more detailed numerical simulation will be devoted to the interpretation of Fe-rich wehrlites.
Nevertheless, two points deserve some additional comment; specifically, the extent of Th and U enrichment and REE/HFSE fractionation. The concentration level of U and Th in clinopyroxenes (and amphiboles) is at least one order of magnitude higher than expected based on experimentally determined Cpx/LD (Hauri et al., 1994; Lundstrom et al., 1994) and OIB composition. Strong U and Th enrichments, along with REE enrichments comparable with those of Type-2 Ibalhratene xenoliths, are frequently documented in minerals from various mantle occurrences (e.g. Spitsbergen, Gobernador Gregores, West Eifel, Kerguelen Archipelago, Lherz) that underwent either anhydrous cryptic and/or hydrous patent metasomatism in different geodynamic settings (Grégoire et al., 2000; Ionov et al., 2002a; Witt-Eickschen et al., 2003; Bodinier et al., 2004; Rivalenti et al., 2004). This suggests that the high U and Th contents of the metasomatic agents are a consequence of melt differentiation during percolation through peridotite mantle, rather than a primitive signature inherited from the source, as is the case for the LREE. Moreover, the positive correlation between U and Th content and LREE fractionation shown by the Ibalrhatene peridotite minerals (and their analogues from Spitsbergen; Ionov et al., 2002a) suggests that they are ruled by the same chromatographic-type melt–peridotite chemical exchange (Fig. 23a). To account for the strong increase of incompatible elements in the percolating melts, melt migration must have been accompanied by fractional crystallization (see Bedini et al., 1997; Vernières et al. 1997; Xu et al., 1998; Bodinier et al., 2004; Rivalenti et al., 2007).
|
As for HFSE/REE fractionation, it is apparent that the range of Nb, Ta, Zr and Hf variation in clinopyroxenes and amphiboles from porphyroblastic lherzolites is substantially greater than that of the adjacent REE (Fig. 23b). A large variability of HFSE abundances, frequently coupled with a relative depletion relative to the adjacent REE, is very frequently documented in metasomatized mantle xenoliths. In contrast, such geochemical features are rare in erupted alkaline volcanic rocks sensu lato. Possible explanations include changes in the relative Sol/LiqD values during the transition from silicate melt to supercritical hydrous fluids (Moine et al., 2001
), HFSE buffering by titanian pargasite to kaersutite amphibole (Oberti et al., 2000; Tiepolo et al., 2000, 2001; Ionov et al., 2002a), reaction with carbonatite melts (Ionov et al., 1994, 2002a; Yaxley et al., 1998; Coltorti et al., 1999; Grégoire et al., 2000) and precipitation of Ti- and Zr-oxides during metasomatism (e.g. Bodinier et al., 1996; Bedini et al., 1997; Rivalenti et al., 2004). In the literature, the last two are regarded as the most plausible hypotheses, in particular when marked negative Zr and Hf anomalies occur in clinopyroxenes and amphiboles, as in this study. Several lines of evidence indicate that in the Ibalrhatene mantle HFSE/REE fractionation in metasomatized lherzolites has to be ascribed to precipitation of Ti- and Zr-oxides. Petrographic evidence shows that orthopyroxene is largely recrystallized rather than dissolved in porphyroblastic lherzolites, thus making it highly unlikely that the original melt was a carbonatite or a carbonate-rich melt. More importantly, some LA-ICP-MS analyses detected extreme HFSE enrichments (Ti up to 100 x C1, Zr up to 10 x C1 and Ta up to 3 x C1) in some neoblastic orthopyroxenes in both lherzolite and harzburgite (Fig. 18) that are revealed in equilibrium with the coexisting clinopyroxene as for the other trace elements. These spikes in HFSE can be attributed to the presence of micrometre-scale Ti- and Zr-minerals that randomly crystallize throughout the mantle column, giving rise to variable fractionation effects (Bodinier et al., 1996; Bedini et al., 1997).
According to available experimental °x/LiqD values, ilmenite and rutile precipitation can efficiently reduce the Ti, Nb and Ta concentrations in a rising melt (Bedini et al. 1997). These elements can also be buffered by armalcolite and minerals of the crichtonite series (i.e. loveringite), which have been recently found in equilibrium with strongly alkaline melts in mantle xenoliths (Grégoire et al., 2000; Kalfoun et al., 2002). Among these oxides, rutile and loveringite exhibit high Zr and Hf concentrations; their crystallization may thus induce dramatic Zr and Hf depletion and (Zr,Hf)/REE fractionation in percolating melts (Rivalenti et al., 2004).
The precipitation of these minerals is probably concentrated in the strongest metasomatized sectors, as supported by the very large variation in Nb, Ta, Hf and Zr content, as well as of the Nb/Ta and Zr/Hf ratios, in clinopyroxenes and amphiboles from porphyroblastic lherzolites (e.g. see Ta and Nb/Ta variation in the amphiboles in Fig. 9b).
Noticeably, Ti-oxide saturation is more feasible in a harzburgitic matrix, because an alkaline melt percolating through it becomes rapidly silica-saturated (e.g. Vannucci et al., 1998); moreover, the high volatile charge of this melt allows it to saturate in titanium oxides even for relatively low Ti concentrations (Ryerson & Watson, 1987; Foley & Wheller, 1990; Bodinier et al., 1996). A similar reaction has been proposed by Grégoire et al. (2000) to explain the precipitation of armalcolite in alkaline veins cutting harzburgitic xenoliths. Again, the appearance of deep negative HFSE anomalies is favoured by the low HFSE content in the original harzburgitic matrix.
Information on the geochemical characteristics of the metasomatic agents that produced Type-2 metasomatism can be derived from the composition of the putative melts in equilibrium with clinopyroxene. Owing to their high HFSE contents, clinopyroxenes from porphyroblastic IBA 57 and IBA 80 lherzolites are the most suitable for this calculation. Possible effects of T and system composition have been considered using two different sets of Cpx/LD values. The former is valid for high-T and -P basaltic systems (i.e. Hart & Dunn, 1993) and accounts for the trace element composition of clinopyroxene from anhydrous mantle xenoliths metasomatized by OIB melts (Rivalenti et al., 2007), whereas the latter is characterized by higher Cpx/LD with compatible HREE behaviour and applies to hydrous metasomatism at lower T (Rivalenti et al., 2004). Similar data have been used to calculate the composition of hypothetical melts in equilibrium with clinopyroxenes segregated by kimberlites (Grégoire et al., 2002). As a result, the calculated liquids (Fig. 24) show a pronounced alkaline affinity with strongly enriched LREE, (Nb,Ta)N/LREEN 1 and no significant Zr anomaly. The U, Th enrichment and the Hf, Ti depletion are thus regarded as products of the early stages of melt–peridotite reaction. Putative liquids calculated with the Hart & Dunn (1993) data have LREE contents and overall REE fractionation higher than those of Moroccan alkaline basalts and nephelinites, approaching the values shown by carbonatites. In contrast, liquids calculated using the second dataset (table 6 of Rivalenti et al., 2004) match the composition of the Moroccan alkali basalts in the MREE–HREE region. Generally, from Fig. 24 it follows that metasomatism was induced by a melt different from the host basalt and that the metasomatic agents were more close in composition to the products of late Cretaceous to early Tertiary carbonatite and alkaline magmatism in the High and Middle Atlas (El Azzouzi et al., 1999; Wagner et al., 2003, and references therein). Similar geochemical signatures are possessed by carbonate-rich alkaline melts that metasomatized mantle sectors in various geodynamic regions, such as Kerguelen (Grégoire et al., 2000), West Eifel (Witt-Eichschen et al., 2003) and Spitsbergen (Ionov et al., 2002a).
|
The origin of Fe-rich wehrlites
Fe-rich wehrlites do not show any clear textural or geochemical characteristics intermediate between Type-1 and Type-2 xenoliths. The most striking observation is the absence in their mineral assemblage of orthopyroxene, which, in contrast, ranges from 15 to 30 vol.% in lherzolitic and harzburgitic xenoliths. This highlights that the protoliths of the wehrlites underwent a very large time-integrated flux of SiO2-undersaturated melts (Kelemen et al., 1995
; Xu et al., 1998) that caused their almost complete thermochemical erosion.
The ubiquitous presence of an early porphyroclastic texture (basically defined by large Fe-rich, kinked olivine) suggests that the Fe-rich wehrlites developed in mantle regions that had previously undergone significant deformation, presumably at the time of the Tertiary–Quaternary rifting. Strongly deformed highly metasomatized mantle sectors are believed to be located at the boundary between an ascending asthenospheric mantle plume and thinning lithosphere (Bedini et al., 1997; Zangana et al., 1997; Rivalenti et al., 2007).
The textural and chemical heterogeneity of the Fe-rich wehrlites suggests that they record the injection of multiple melt inputs, often in chemical disequilibrium with one another. From a petrographic point of view, this is documented by the dissolution of porphyroclastic olivines, which display large embayments and concave boundaries towards clinopyroxene ± spinel ± amphibole mineral assemblages. Olivine dissolution is particularly apparent subsequently, during the development of the globular clusters. The absence of secondary olivine in the clusters points to an evolved composition of the percolating melts, which possibly became progressively SiO2-saturated.
It is thus proposed that the Ibalrhatene wehrlites are not the product of a single-stage interaction process between ambient peridotite and SiO2-undersaturated alkaline melts as described and modelled elsewhere (Zinngrebe & Foley, 1995; Peslier et al., 2002; Ionov et al., 2005). Rather, the petrographic characteristics point to an early recrystallization stage dominated by pyroxene-dissolving reactions, which formed dunitic sectors made up of olivine and spinel, followed by a further percolation event that led to the crystallization of clinopyroxene-bearing assemblages from evolved melts. The scenario envisaged here is closely similar to that documented for a xenolith suite from the Kerguelen archipelago affected by alkaline metasomatism (Grégoire et al., 2000). In orogenic massifs, similar occurrences are documented by the late crystallization of clinopyroxene in sectors exploited by large volumes of melts ascending towards the surface, similar to the wehrlitic bands or patches in dunite channels like those found in the Horoman (Takahashi, 1992; Takazawa et al., 1992) and Lanzo mantle sequences (Piccardo et al., 2007).
Evidence for multi-stage evolution and constraints on the fractionation trends of the melt are conclusively provided by the large, texture-related compositional variability displayed by IBA 32 and IBA 3 wehrlites (Figs 11, 12 and 15). Assuming constant partition coefficients, the composition of IBA 32 clinopyroxenes would indicate that the early pervasive stage of wehrlitization was triggered by percolation of melts strongly enriched in REE and Sr, but impoverished in HFSE4+, which were followed by melts progressively less enriched in REE and Sr, and with higher HFSE4+; namely, with more pronounced basaltic affinity. In IBA 32 wehrlite, this compositional trend is almost completely documented by the last interstitial amphiboles, which show trace element concentrations approaching those of IBA 18 cumulitic amphiboles.
Consistent trace element zoning and HFSE/REE fractionation variability is exhibited by clinopyroxenes and amphiboles from IBA 3. These samples also show some additional complexities; for example: (1) matrix-to-vein heterogeneity is also accompanied by core-to-rim zoning of matrix minerals; (2) the core compositions of the largest clinopyroxene and amphibole crystals in the glassy vein are similar to the rim compositions of the matrix grains, pointing to derivation from the same melt. Conversely, the rim compositions of vein clinopyroxene and amphibole are clearly in equilibrium with the associated glass, which shows distinct characteristics with respect to the putative liquids in equilibrium with the matrix minerals and the cores of the vein minerals, thus marking the percolation of a new melt batch.
In principle, it is possible that the compositional heterogeneity documented by IBA 3 and IBA 32 was related to injection of melts with distinct geochemical signatures reflecting both the source composition and partial melting processes. However, the steady evolution of the geochemical characteristics of the metasomatic minerals is more consistent with progressive fractionation of the alkaline melts during porous flow percolation, as previously envisaged by Laurora et al. (2001) and Ionov et al. (2006). Compositional zoning in terms of REE abundances and REE/HFSE fractionation of amphiboles has been described by Moine et al. (2001) for amphibole in veined, spinel-bearing dunite xenoliths from the Kerguelen Islands (Indian Ocean). These xenoliths are cross-cut by connected thick and thin veins. Amphibole from the thin veins has lower Ti, and higher Nb, Ta, Zr and Hf abundances relative to that from the thick veins. The observed chemical variations have been related by Moine et al. (2001) to the combination of two processes: (1) the changing composition of fluids from thick to thin veins as a result of crystallization of amphibole, ilmenite and phlogopite before crack propagation into the dunite wall; (2) the change in intrinsic amphibole–fluid partition coefficients for HFSE as a function of both aH2O and aTiO2 in evolved fluids.
Some changes in Sol/LiqD during the fractionation process are actually supported in our wehrlites by the observation that the Nb and Ta enrichments shown by matrix amphibole are not matched by the composition of the texturally associated clinopyroxene. This decoupling is so systematic that it can hardly be ascribed to the crystallization of the two minerals from different melts. Alternatively, it can result from a marked increase of Amph/LiqD(Nb,Ta) values, which can become compatible in SiO2-rich and H2O–TiO2-poor systems, as a consequence of the need of the amphibole structure to incorporate high-charge elements to balance the excess of the negative charge caused by the O2– 
OH– substitution (Tiepolo et al., 2000).
Among the four wehrlites with more homogenous trace element signatures (samples IBA 5, IBA 41, IBA 53 and IBA 69), IBA 41 and IBA 69 contain clinopyroxene and amphibole with pronounced convex-upward trace element patterns and LILE abundances similar to those of cluster minerals from IBA 3 and IBA 32, as well as to those of cumulate minerals from IBA 18 and IBA 9 pyroxenites. In contrast, minerals from IBA 5 and IBA 53 wehrlites display relative enrichment in LREE, U, Th, Nb, Ta, and MREE–HREE, Y, Zr, Hf and Ti contents, closely approaching those of minerals from Fe-wehrlites IBA 41 and IBA 69. Thus, it is plausible that the enrichment in highly incompatible elements recorded by IBA 5 and IBA 53 samples could be ascribed to the entrapment of very small volumes of migrating melts rather than to superimposed or distinct metasomatic events, as demonstrated by simple numerical simulations (Ionov et al., 2002a, 2006).
Putative liquids in equilibrium with clinopyroxenes from cumulus pyroxenites IBA 18 and IBA 9 calculated using the dataset for high-T basaltic systems (partition coefficients after Hart & Dunn, 1993) show REE abundance patterns very similar to those of the erupted Moroccan nephelinites and alkali basalts (Fig. 25a). The same is true of the hypothetical liquids in equilibrium with clinopyroxenes from IBA 41 and IBA 69, IBA 3 and IBA 32, although their absolute concentrations are significantly higher than those of the alkaline basalts. The basaltic nature of the metasomatic agents is also supported by the high Nb, Ta, Hf and Zr abundances in the clinopyroxenes and amphiboles in IBA 5 (and IBA 53) samples.
|
The mismatch of the REE concentration levels between the calculated liquids and the erupted volcanic rocks could be indicative of the migration of strongly REE-enriched alkaline basalts (Ionov et al., 2006
). If we assume that the Fe-rich wehrlites formed from percolating melts similar in composition to the erupted alkali basalts, the mismatch for REE could reflect an increase of the Sol/LiqD (see Fig. 25b) during low-T crystallization, as has been suggested elsewhere for the crystallization of alkaline (e.g. Rivalenti et al., 2004) and tholeiitic melts (Piccardo et al., 2007). This possibility is supported by the large Cpx/LD and Amph/LD values determined for mineral–glass pairs from the IBA 3 vein (the latter strictly consistent with those determined for amphibole-bearing glassy veins in xenoliths from Antarctica; Coltorti et al., 2004). Alternatively, the mismatch could provide evidence for a progressive trace element concentration in the percolating melts during their migration through the host peridotite.
In conclusion, a multi-stage reaction process between rising melt and ambient peridotite is regarded as the most plausible scenario to explain the variation in trace element abundance and fractionation in clinopyroxene and amphibole depending on their textural occurrence (matrix, cluster, euhedral crystals), as observed in the IBA 32 and IBA 3 samples. We thus propose that the lithospheric mantle section in which the wherlites formed beneath the Middle Atlas interacted through time with larger and larger volumes of rising alkaline melts, which were progressively less evolved in composition. In this context, the early reactions of the host mantle with a decreasing mass of evolved alkaline melt has produced matrix clinopyroxene with the highest REE concentrations and the most marked HFSE negative anomalies. The successive injection of larger volumes of less evolved alkaline melt has generated cluster clinopyroxenes, whereas the last interaction with large volumes of more primitive alkaline melts similar to the surface lavas (El Azzouzi et al., 1999) has led to the precipitation of euhedral crystals that occur in melt pockets around partially melted amphiboles and in glass patches. We used the texture-related trace element composition of clinopyroxene to test this interpretation in the context of the plate model established by Vernières et al. (1997). The progressive increase of the overall REE content in the most differentiated melts indicates that the differentiation process was mainly governed by fractional crystallization. In addition, the lack of significant changes in REE fractionation suggests a subordinate amount of modal pyroxene.
Chemical and reaction parameters used in numerical modelling are reported in Table 14. The melt–rock reaction scheme is a simplified version of that proposed in previous studies (Bedini et al., 1997; Xu et al., 1998; Ionov et al., 2002a), as only one reaction zone, subdivided into 20 cells, has been considered (Table 14) to describe the pyroxene- and amphibole-forming reactions (i.e. wehrlitization process) in an already refractory mantle sector. The results (Fig. 26) show that most of the REE spectra of the Fe-rich wehrlites can be reproduced by the progressive interaction of alkaline melts with a strongly depleted peridotite. Clinopyroxene in the first cells shows convex-upward REE patterns and is very close to equilibrium with the infiltrated melt. The composition of clinopyroxenes from IBA 41, IBA 69, IBA 3 and IBA 32 wehrlites (cells from 9 to 13) is progressively accounted for by their crystallization from decreasing, although still large, volumes of more evolved melts.
|
|
As already observed, available Nd and Sr isotope data for separate minerals from Type-2 metasomatized xenoliths, Mg- and Fe-rich wehrlites, and the IBA 18 pyroxenite display a HIMU-like affinity and overlap those of the Moroccan alkaline lavas (El Azzouzi et al., 1999
), thus suggesting a common source for both these and the metasomatic agents. Unfortunately, inferences on the age of the metasomatic events are largely questionable. If the available Nd–Sm data are plotted on an isochron diagram it is not possible to define an errorchron. This simply reflects the overprinting of older lithospheric components during younger metasomatic events that possibly operated since late Cretaceous times. In this respect, it seems plausible that the samples that appear more extensively re-equilibrated with large volumes of metasomatic melts, such as the Fe-rich wehrlites, more closely approach the age of metasomatism. In contrast, the less re-equilibrated samples furnish meaningless ages, which are intermediate between the time of melt extraction and that of the closure after the metasomatic overprint.
When only clinopyroxene and amphibole separates from IBA 18 cumulate and Fe-rich wehrlite samples that display convex-upward patterns suggestive of equilibrium with alkaline melts are considered (all but IBA 5 and IBA 53 samples), an errorchron of 156 ± 16 Ma is obtained. Although we cannot attach any geochronological significance to this value, we observe that it is not much in excess of that for the carbonatite and alkaline magmatism in the region (late Cretaceous–Early Tertiary; Wagner et al., 2003) and that it provides only a rough indication of a young age for the metasomatic events.
The origin of Mg-rich wehrlites
It is generally believed that several petrochemical characteristics of the IBA 26 Mg-rich wehrlite can be considered as evidence of ‘carbonatite’ metasomatism in the mantle beneath Ibalrhatene. These comprise the strong REE/HFSE fractionation and the Si, Na and Cr enrichment shown by amphibole and clinopyroxene, the high Fo and XCr content in olivine and spinel, respectively, and the occurrence of apatite (e.g. Yaxley et al., 1991; Dautria et al., 1992; Rudnick et al., 1993). Moreover, it has been demonstrated experimentally that the relatively high Zr/Hf, Ta/Nb and U/Th values shown by both amphibole and clinopyroxene from IBA 26 can be considered as geochemical ‘markers’ of equilibrium with carbonatite melts (Green, 1995; Foley et al., 2001).
Overall, these observations suggest that Mg-rich wehrlite could be the product of a peculiar carbonate-rich melt, distinct from the liquids that formed the Fe-rich wehrlites, as well as the metasomatized lherzolite–harzburgite sequence. The presence of carbonatite melts of Cretaceous–Early Tertiary age in the Taourirt area (Wagner et al., 2003) and related to the oldest volcanic activity of Eocene age in the Bou-Agrao massif (Harmand & Cantagrel, 1984) and the Tamazert complex (Mourtada, 1997), provides further support in favour of separate (e.g. carbonatite) melt sources. Thus, if the Mg-rich wehrlites were produced by the infiltration of carbonatite melts, this probably happened during the early stages of metasomatic alteration of the lithosphere.
However, the texture-related overall variation of the major and trace element mineral chemistry of amphibole and clinopyroxene from the Fe-rich wehrlites (in particular, the IBA 3 sample) and of amphiboles from the apatite-bearing IBA 56 porphyroblastic harzburgite, along with their overall similarity to the isotopic signatures of erupted alkaline lavas, allows us to speculate that the Mg-rich wehrlites could represent the result of the interaction in localized mantle sectors of melts that were extremely evolved in composition after very large degrees of interaction with lithospheric peridotites.
In the literature, similar compositional trends have been ascribed to distinct melts or fluids that formed as the ultimate derivative of an initial volatile-bearing silicate liquid. In particular, the immiscibility and exsolution of carbonate- and P2O5-rich liquids after extensive crystallization of clinopyroxene, amphibole, and, possibly, phlogopite has been sometimes invoked (Zanetti et al., 1999; Ionov et al., 2006). This separate metasomatic fluid, similar to a primary carbonatite, is expected to produce disseminated phosphate and carbonate phases, as well as to possess extreme REE enrichment and HFSE depletion, because of the low solubility of HFSE in carbonate-rich melts coexisting with silicate melts (Ionov et al., 2002a; Bodinier et al., 2004).
A further possibility is that the metasomatic agent close to the percolation front was a SiO2-saturated silicate melt which became highly evolved by reactive porous flow at decreasing volume. This possibility is supported by general petrological considerations, from which is apparent that the fate of alkaline melts vanishing through interaction with lithospheric (shallow) peridotites is to become saturated in SiO2, as the initial reaction with peridotite largely involves dissolution of orthopyroxene (Rivalenti et al., 2004, and references therein). A silica-saturated composition of the melts would account for the occurrence of orthopyroxene-rich layers in the IBA 26 sample, a feature not consistent with strong metasomatic alteration induced by a silica-undersaturated primary carbonatitic melt. In addition, the major element characteristics of clinopyroxenes crystallized from alkaline melts that progressively become more SiO2-rich after increasing reaction with ambient peridotites are an Al decrease and Si, Cr and Mg-number increase (Vannucci et al., 1998), thus producing chemical trends similar to those defined by the IBA 56, IBA 3 and IBA 26 samples.
In this context, the percolation model proposed for the origin of Fe-rich wehrlites could, in principle, be extended to the Mg-rich sample, assuming that IBA 26 equilibrated with the more extreme derivatives of the initial melt. The model developed in the previous section is able to reproduce the LREE and MREE enrichment of IBA 26 clinopyroxene, but does not account for its high HREE abundance (Fig. 26), which can be only better approached, but not matched, by increasing the mode of newly formed olivine by a factor of two with respect to the value in Table 13. Moreover, the model does not account for the marked negative HFSE anomalies characterizing IBA 26 clinopyroxene and amphibole. The combined LILE enrichment and HFSE depletion of these latter can be successfully modelled only by invoking strong sink–source effects caused by HFSE-rich phases and/or significant decoupling between mineral/meltD values for REE and HFSE in response to T- and composition-dependent crystal-chemical constraints.
As discussed above, owing to the low Al content of clinopyroxenes and amphiboles from IBA 26, low mineral/meltD values for HFSE (see Hill et al., 2000) are believed to have governed the metasomatic process. The low Al content should have also induced relatively low mineral/meltD values for REE (Gaetani & Grove, 1995), unless there were significant changes in the melt composition. In this respect, Vannucci et al. (1998) demonstrated that if SiO2-undersaturated alkaline basaltic melts migrate through mantle peridotite following differentiation trends towards silica-rich compositions, the Cpx/meltD values progressively vary, ranging from values typical for basaltic systems (e.g. Hart & Dunn, 1993) to very large values for REE, but low to very low values for Sr and HFSE. The orthopyroxene-rich layers of IBA 26 sample indicate that the parent liquid could have had a significant silica content, possibly corresponding to the Fo–Cpx–SiO2 peritectic. These petrochemical observations suggest that the complex features characterising IBA 26 could result from the interaction with extreme silicate melt differentiates derived by interaction of alkaline basalt with the host peridotites.
Given that the occurrence of xenoliths along the NE–SW volcanic alignment parallel to the Trans-Moroccan fault system precludes sampling of mantle sectors from a laterally wide region, we speculate that Mg-rich wehrlite was most probably formed in proximity to the percolation front at a shallower depth with respect to the highly refractory region.
|
CONCLUDING REMARKS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND TEXTURESANALYTICAL TECHNIQUESMINERAL GEOCHEMISTRYDISCUSSIONCONCLUDING REMARKSSUPPLEMENTARY DATAREFERENCES |
|---|
The study of peridotite xenoliths brought to the surface by alkaline basaltic melts near Ibalrhatene (Azrou–Timahdite volcanic district, Middle Atlas, Morocco) has revealed extreme lithological and chemical heterogeneity in the lithospheric mantle beneath the NE–SW volcanic alignment parallel to the Trans-Moroccan fault system. A small number of samples (mostly lherzolites and websterites) still record old melt extraction and migration processes, possibly Neo- to Paleoproterozoic in age, documented by DMM-like isotopic signatures. After their accretion to the lithosphere, these mantle sectors cooled for a long time before the upward migration during late Cretaceous times of asthenospheric melts with HIMU-like isotopic signatures melts that imparted in the overlying lithospheric mantle a strong metasomatic overprint. The metasomatic agents reacted with variously depleted mantle sectors, thus producing a range of trace element signatures. Some lherzolites and websterites recording the chromatographic effects of melt percolation (Type-1 samples) were only partially affected by metasomatism, because of their location far (i.e. either laterally or at shallower depth) from the melt source. Other lherzolites and more refractory mantle sectors (i.e. harzburgites) progressively approached chemical equilibrium with the migrating melt and were extensively re-fertilized (Type-2 samples). During these episodes of thermo-chemical erosion of the lithosphere, it seems plausible that rising melts reacted with the host peridotite, forming dunitic channels like those commonly observed in orogenic peridotite massifs (Piccardo et al., 2007
). Further migration through these channels of larger and larger volumes of alkaline melts, becoming progressively less evolved in composition and more similar to the surface lavas, may have produced the Fe-rich wehrlites. Locally, the migration of metasomatic agents extremely enriched in incompatible elements, such as carbonatite melts related to the oldest (i.e. Eocene) volcanic activity (Harmand & Cantagrel, 1984; Wagner et al., 2003), or, more probably, of highly evolved SiO2-saturated silicate melts vanishing through interaction with the lithosphere, produced in localized mantle sectors unusual enrichments of highly incompatible elements coupled with marked HFSE depletions. Similar characteristics are observed in the Mg-wehrlite sample and its harzburgite host, as well as in the apatite-bearing IBA 56 porphyroblastic harzburgite and the IBA 3 Fe-rich wehrlite. The overall scenario envisaged here for the lithosphere–asthenosphere interaction beneath the Middle Atlas is illustrated in simplified form in Fig. 27.
|
The process of thermo-chemical erosion of the lithosphere is also documented by the isotopic signatures of the investigated xenoliths. With the exception of those subordinate lherzolite and websterite samples that have a DMM isotopic affinity, the great majority of the xenoliths have a clear HIMU fingerprint, similar to that of the erupted alkaline lavas. This situation is common to other Central–Northern Africa volcanic occurrences such as Hoggar (Algeria) and Gharyan (Libya), and indicates the ubiquitous presence of a common HIMU sub-lithospheric mantle component (Beccaluva et al., 2007
, 2008). The abundance of Type-2 samples and Fe-rich wehrlites relative to Type-1 samples suggests that wide sectors of the mantle beneath the NE–SW volcanic alignment parallel to the Trans-Moroccan fault system consist of rejuvenated lithosphere formed by intra-cratonic upwelling and accretion of asthenospheric material, The limited sampling of older cratonic mantle can be regarded as evidence in favour of small-scale shallow-mantle upwelling (see Fig. 27), triggered by intraplate tectonic reactivation and rifting of the Pan-African basement (Beccaluva et al., 2007, 2008). |
SUPPLEMENTARY DATA |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND TEXTURESANALYTICAL TECHNIQUESMINERAL GEOCHEMISTRYDISCUSSIONCONCLUDING REMARKSSUPPLEMENTARY DATAREFERENCES |
|---|
Supplementary data for this paper are available at Journal of Petrology online.
|
ACKNOWLEDGEMENTS |
|---|
G.C. and R.V. gratefully thank Lahcen Kabiri and Larbi Boudad as well as the ‘Faculté des Sciences et Techniques’ of Errachidia for logistic support in the Azrou region. We wish to thank the Executive Editor M. Wilson and J.-L. Bodinier, M. Coltorti, E.-R. Neumann and an anonymous reviewer for their constructive comments and suggestions, which greatly improved an earlier version of the manuscript. This work was made possible by PRIN-MIUR, Pavia University and CNR–IGG grants. N.R. acknowledges support from the Laboratoire de Géologie de l’Université Blaise Pascal and funding by Pavia University and CNR–IGG for Post-Doc Fellowships.
*Corresponding author: Nicola Raffone, Dipartimento di Scienze della Terra, Università di Pavia and CNR–IGG, Sezione di Pavia, via Ferrata 1, I-27100, Pavia, Italy. Telephone: 0039 0382 985842. Fax: 0039 0382 985890. E-mail: raffone{at}crystal.unipv.it
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND TEXTURESANALYTICAL TECHNIQUESMINERAL GEOCHEMISTRYDISCUSSIONCONCLUDING REMARKSSUPPLEMENTARY DATAREFERENCES |
|---|
Anders E, Grevesse N. Abundances of the elements: Meteoritic and solar. Geochimica et Cosmochimica Acta (1989) 53:197–214.[CrossRef][Web of Science]
Arai S. Characterization of spinel peridotites by olivine–spinel compositional relationships: review and interpretation. Chemical Geology (1994) 113:191–204.[CrossRef][Web of Science]
Bali E, Zanetti A, Szabò C, Peate DW, Waight TE. A micro-scale investigation of melt production and extraction in the upper mantle based on silicate melt pockets in ultramafic xenoliths from the Bakony–Balaton Highland Volcanic Field (Western Hungary). Contributions to Mineralogy and Petrology (2008) 155:165–179.[CrossRef][Web of Science]
Ballhaus C, Berry RF, Green DH. High pressure experimental calibration of the olivine–orthopyroxene–spinel oxygen geobarometer: implications for the oxidation state of the upper mantle. Contributions to Mineralogy and Petrology (1991a) 107:27–40.[CrossRef][Web of Science]
Ballhaus C, Berry RF, Green DH. High pressure experimental calibration of the olivine–orthopyroxene–spinel oxygen geobarometer: implications for the oxidation state of the upper mantle (Erratum). Contributions to Mineralogy and Petrology (1991b) 108:384.[CrossRef][Web of Science]
Ban M, Witt-Eickschen G, Klein M, Seck HA. The origin of glasses in hydrous mantle xenoliths from the West Eifel, Germany: incongruent breakdown of amphibole. Contributions to Mineralogy and Petrology (2005) 148:511–523.[CrossRef][Web of Science]
Beccaluva L, Azzouni-Sekkal A, Benhallou A, Bianchini G, Ellam RM, Marzola M, Siena F, Stuart FM. Intracratonic asthenosphere upwelling and lithosphere rejuvenation beneath the Hoggar swell (Algeria): evidence from HIMU metasomatised lherzolite mantle xenoliths. Earth and Planetary Science Letters (2007) 260:482–494.[CrossRef][Web of Science]
Beccaluva L, Bianchini G, Ellam RM, Marzola M, Oun KM, Siena F, Stuart FM. The role of HIMU metasomatic components in the North African lithospheric mantle: petrological evidence from the Gharyan lherzolite xenoliths, NW Libya. In: Metasomatism in oceanic and continental lithospheric mantle. Geological Society, London, Special Publications —Coltorti M, Gregoire M, eds. (2008) 293:253–277.
Bedini RM, Bodinier J.-L, Dautria J.-M, Morten L. Evolution of LILE-enriched small melt fractions in the lithospheric mantle: a case study from the East African Rift. Earth and Planetary Science Letters (1997) 153:67–83.[CrossRef][Web of Science]
Bodinier J.-L, Merlet C, Bedini RM, Simien F, Remaïdi M. Distribution of niobium, tantalum and other highly incompatible trace elements in the lithospheric mantle: the spinel paradox. Geochimica et Cosmochimica Acta (1996) 60:545–550.[CrossRef][Web of Science]
Bodinier J.-L, Menzies MA, Shimizu N, Frey FA, McPherson E. Silicate, hydrous and carbonate metasomatism at Lherz, France: contemporaneous derivatives of silicate melt–harzburgite reaction. Journal of Petrology (2004) 45:299–320.
Brey GP, Köhler T. Geothermobarometry in four-phase lherzolites II. New thermobarometers, and practical assessment of existing thermobarometers. Journal of Petrology (1990) 31:1353–1378.
Brueckner HK, Elhaddad MA, Hamelin B, Hemming SR, Kroner A, Reisberg L, Seyler M. A Pan African origin and uplift for the gneisses and peridotites of Zabargad Island, Red Sea; a Nd, Sr, Pb and Os isotope study. Journal of Geophysical Research (1995) 100:22283–22297.[CrossRef]
Bryndzia LT, Wood BJ. Oxygen thermobarometry of abyssal peridotites: The redox state and C–O–H volatile composition of the Earth's sub-oceanic upper mantle. American Journal of Science (1990) 290:1093–1116.
Charlot R, Choubert G, Faure-Muret A, Hamel C. Ages des aïoinites du Maroc Nord-Oriental. Compte Rendu Sommaire de séances de la Société Géologique de France (1964) 41:309–323.
Chazot G, Menzies MA, Harte B. Determination of partition coefficients between apatite, clinopyroxene, amphibole, and melt in natural spinel lherzolites from Yemen: implications for wet melting of the lithospheric mantle. Geochimica et Cosmochimica Acta (1996a) 60:423–437.[CrossRef][Web of Science]
Chazot G, Menzies MA, Harte B. Silicate glasses in spinel lherzolites from Yemen: origin and chemical composition. Chemical Geology (1996b) 134:159–179.[CrossRef][Web of Science]
Coltorti M, Bonadiman C, Hinton RW, Siena F, Upton B. GJ. Carbonatite metasomatism of the oceanic upper mantle: evidence from clinopyroxenes and glasses in ultramafic xenoliths of Grande Comore, Indian Ocean. Journal of Petrology (1999) 40:133–165.
Coltorti M, Beccaluva L, Bonadiman C, Faccini B, Ntaflos T, Siena F. Amphibole genesis via metasomatic reaction with clinopyroxene in mantle xenoliths from Victoria Land, Antarctica. Lithos (2004) 75:115–139.[CrossRef][Web of Science]
Dautria JM, Dupuy C, Takherist D, Dostal J. Carbonate metasomatism in the lithospheric mantle: peridotitic xenoliths from a melilititic district of the Sahara basin. Contributions to Mineralogy and Petrology (1992) 111:37–52.[CrossRef][Web of Science]
Du Dresnay R. Recent data on the geology of the Middle Atlas (Morocco). In: The Atlas System of Morocco—Jacobshagen H, ed. (1988) Berlin: Springer. 293–320.
El Azzab D, El Wartiti M. Paléomagnétisme des laves du Moyen Atlas (Maroc): rotations récentes. Comptes Rendus de l’Académie des Sciences (1998) 327:509–512.
El Azzouzi M, Bernard- Griffiths J, Bellon H, Maury RC, Piqué A, Fourcade S, Cotten J, Hernandez J. Évolution des sources du volcanisme marocain au cours du Néogène. Comptes Rendus de l’Académie des Sciences de Paris, Sciences de la Terre et des Planètes (1999) 329:95–102.
Fabriès J, Bodinier JL, Dupuy C, Lorand JP, Benkerrou C. Evidence for modal metasomatism in the orogenic spinel lherzolite body from Caussou (Northeastern Pyrenees, France). Journal of Petrology (1989) 30:199–228.
Faul UH. Permeability of partially molten upper mantle rocks from experiments and percolation theory. Journal of Geophysical Research (1997) 102:10299–10311.[CrossRef]
Foley SF, Wheller GE. Parallels in the origin of the geochemical signature of island arc volcanics and continental potassic igneous rocks: the role of residual titanates. Chemical Geology (1990) 85:1–18.[CrossRef][Web of Science]
Foley SF, Petibon CM, Jenner GA, Kjarsgaard BA. High U/Th partitioning by clinopyroxene from alkali silicate and carbonatite metasomatism: an origin for Th/U disequilibrium in mantle melts? Terra Nova (2001) 13:104–109.[CrossRef][Web of Science]
Gaetani GA, Grove TL. Partitioning of rare earth elements between clinopyroxene and silicate melt: Crystal-chemical controls. Geochimica et Cosmochimica Acta (1995) 59:1951–1962.[CrossRef][Web of Science]
Green TH. Experimental versus natural two-mineral partition coefficients: a ‘high-tech’ controversy. International Geology Review (1995) 37:851–865.
Grégoire M, Moine BN, O’Reilly SY, Cottin JY, Giret A. Trace element residence and partitioning in mantle xenoliths metasomatized by high alkaline silicate and carbonate-rich melts (Kerguelen Islands, Indian Ocean). Journal of Petrology (2000) 41:477–509.
Grégoire M, Bell DR, Le Roex AP. Mafic and ultramafic xenoliths from the Kaapvaal Craton (South Africa); trace element evidence from mantle magmatic and metasomatic processes. Geochimica et Cosmochimica Acta (2002) 66. Supplement 1, August 15, 290.
Haggerty SE. Oxide mineralogy of the upper mantle. Reviews in Mineralogy (1991) 25:355–416.[Abstract]
Harmand C, Cantagrel JM. Le volcanisme alcalin Tertiaire et Quaternaire du Moyen Atlas (Maroc): chronologie K/Ar et cadre géodynamique. Journal of African Earth Sciences (1984) 2:51–55.[CrossRef]
Harmand C, Moukadiri A. Synchronisme entre tectonique compressive et volcanisme alcalin: exemple de la province Quaternaire du Moyen Atlas (Maroc). Bulletin de la Société Géologique de France (1986) 4:595–603.
Hart SR, Dunn T. Experimental cpx/melt partitioning of 24 trace elements. Contributions to Mineralogy and Petrology (1993) 113:1–8.[CrossRef][Web of Science]
Hauri EH, Shimizu N, Dieu JJ, Hart SR. Evidence for hotspot-related carbonatite metasomatism in the oceanic upper mantle. Nature (1993) 365:221–227.[CrossRef]
Hauri EH, Wagner TP, Grove TL. Experimental and natural partitioning of Th, U, Pb and other trace elements between garnet, clinopyroxene and basaltic melts. Chemical Geology (1994) 117:149–166.[CrossRef][Web of Science]
Hellebrand E, Snow JE, Hoppe P, Hofmann AW. Garnet-field melting and late-stage refertilization in ‘residual’ abyssal peridotites from the Central Indian Ridge. Journal of Petrology (2002) 43:2305–2338.
Hill E, Wood BJ, Blundy JD. The effect of Ca-Tschermaks component on trace element partitioning between clinopyroxene and silicate melt. Lithos (2000) 52:203–215.
Hirschmann MM, Stolper EM. A possible role for garnet pyroxenite in the origin of the ‘garnet signature’ in MORB. Contributions to Mineralogy and Petrology (1996) 124:185–208.[CrossRef][Web of Science]
Hofmann AW. Mantle geochemistry: the message from oceanic volcanism. Nature (1997) 385:219–229.[CrossRef]
Ionov DA. Trace element composition of mantle-derived carbonates and coexisting phases in peridotite xenoliths from alkali basalts. Journal of Petrology (1998) 39:1931–1941.
Ionov DA, Harmer RE. Trace element distribution in calcite–dolomite carbonatites from Skitskop: inferences for differentiation of carbonatite magmas and the origin of carbonates in mantle xenoliths. Earth and Planetary Science Letters (2002) 198:495–510.[CrossRef][Web of Science]
Ionov DA, Hofmann AW, Shimizu N. Metasomatism-induced melting in mantle xenoliths from Mongolia. Journal of Petrology (1994) 35:753–785.
Ionov DA, Bodinier JL, Mukasa SB, Zanetti A. Mechanisms and sources of mantle metasomatism: major and trace element compositions of peridotite xenoliths from Spitsbergen in the context of numerical modelling. Journal of Petrology (2002a) 43:2219–2259.
Ionov DA, Mukasa SB, Bodinier JL. Sr–Nd–Pb isotopic composition of peridotite xenoliths from Spitsbergen: numerical modelling indicates Sr–Nd decoupling in the mantle by melt percolation metasomatism. Journal of Petrology (2002b) 43:2261–2278.
Ionov DA, Chanefo I, Bodinier J.-L. Origin of Fe-rich lherzolites and wehrlites from Tok, SE Siberia by reactive melt percolation in refractory mantle peridotites. Contributions to Mineralogy and Petrology (2005) 150:335–353.[CrossRef][Web of Science]
Ionov DA, Chazot G, Chauvel C, Merlet C, Bodinier J.-L. Trace element distribution in peridotite xenoliths from Tok, SE Siberian craton: a record of pervasive, multi-stage metasomatism in shallow refractory mantle. Geochimica et Cosmochimica Acta (2006) 70:1231–1260.[CrossRef][Web of Science]
Jacobsen SB, Wasserburg GJ. A two-reservoir recycling model for mantle-crust evolution. Proceedings of the National Academy of Sciences (1980) 77:6298–6302.
Johnson K. TM, Dick H. JB. Open system melting and temporal and spatial variation of peridotite and basalt at the Atlantis II Fracture Zone. Journal of Geophysical Research (1992) 97:9219–9241.
Johnson K. TM, Dick H. JB, Shimizu N. Melting in the oceanic upper mantle: an ion microprobe study of diopsides in abyssal peridotites. Journal of Geophysical Research (1990) 95:2661–2678.
Kalfoun F, Ionov DA, Merlet C. HFSE residence and Nb/Ta ratios in metasomatized, rutile-bearing mantle peridotites. Earth and Planetary Science Letters (2002) 199:49–65.[CrossRef][Web of Science]
Kelemen PB, Shimizu N, Salters V. JM. Extraction of mid-ocean-ridge basalt from the upwelling mantle by focused flow of melt in dunite channels. Nature (1995) 375:747–753.[CrossRef]
Kinzler RJ. Melting of mantle peridotite at pressures approaching the spinel to garnet transition: application to midocean ridge basalt petrogenesis. Journal of Geophysical Research (1997) 102:853–874.[CrossRef]
Kornprobst J, Piboule M, Roden M, Tabit A. Corundum-bearing garnet clinopyroxenites at Beni Bousera (Morocco): original plagioclase-rich gabbros recrystallized at depth within the mantle? Journal of Petrology (1990) 31:717–745.
Laurora A, Mazzucchelli M, Rivalenti G, Vannucci R, Zanetti A, Barbieri AM, Cingolani CA. Metasomatism and melting in carbonated peridotite xenoliths from mantle wedge: The Gobernador Gregores case (Southern Patagonia). Journal of Petrology (2001) 42:69–87.
Leake BE, Maresch WV, Nickel EH, Rock N. MS, Schumacher JC, Smith DC, Stephenson N. CN, Ungaretti L, Whittaker E. JW, Youzhi G. Nomenclature of amphiboles. Commission on new mineral names, International Mineralogical Association. Canadian Mineralogist (1997) 35:219–246.[Web of Science]
Le Bas MJ, Le Maitre RW, Streckeisen A, Zanettin B. A chemical classification of volcanic rocks based on the total alkali–silica diagram. Journal of Petrology (1986) 27:745–750.
Lundstrom CC, Shaw HF, Ryerson FJ, Phinney DL, Gill JB, Williams Q. Compositional controls on the partitioning of U, Th, Ba, Pb, Sr and Zr between clinopyroxene and haplobasaltic melts: implications for uranium series disequilibria in basalts. Earth and Planetary Science Letters (1994) 128:407–423.[CrossRef][Web of Science]
Maza M, Briqueu L, Rivalenti G, Dautria JM, Bosch D. Le complexe annulaire d’âge Oligocène de l’Achkal (Hoggar Centaral, Sud Algérie): témoin de la transition au Cénozoïque entre les magmatismes tholéitique et alcalin. Évidences par les isotopes du Sr, Nd et Pb. Comptes Rendus de l’Académie des Sciences, Sciences de la Terre et des Planètes (1998) 327:167–172.
Mercier J.-CC. Single-pyroxene thermobarometry. Tectonophysics (1980) 70:1–37.[CrossRef][Web of Science]
Mercier J. CC, Benoit V, Girardeau J. Equilibrium state of diopside-bearing harzburgites from ophiolites: geobarometric and geodynamic implications. Contributions to Mineralogy and Petrology (1984) 85:391–403.[CrossRef][Web of Science]
Mercier J.-C, Nicolas A. Textures and fabrics of upper-mantle peridotites as illustrated by xenoliths from basalts. Journal of Petrology (1974) 16:454–487.[Web of Science]
Moine BN, Grégoire M, O’Reilly SY, Sheppard MF, Cottin JY. High field strength element fractionation in the upper mantle: evidence from amphibole-rich composite mantle xenoliths from Kerguelen Islands (Indian Ocean). Journal of Petrology (2001) 42:2145–2168.
Morimoto M. Nomenclature of pyroxenes. Commission on new mineral names, International Mineralogical Association. American Mineralogist (1988) 73:1123–1133.[Abstract]
Morishita T, Arai S, Tamura A. Petrology of an apatite-rich layer in the Finero phlogopite-peridotite, Italian Western Alps; implications for evolution of a metasomatising agent. Lithos (2003) 69:37–49.[CrossRef][Web of Science]
Moukadiri A. Les enclaves ultrabasiques assocées aux basaltes alcalins dans le district volcanique d’Azrou–Timahdit (Moyen Atlas, Maroc). (1983) 150. Thèse 3ème cycle, Université Blaise-Pascal, Clermont-Ferrand.
Moukadiri A, Kornprobst J. Garnet and/or spinel bearing pyroxenites in alkali basalts near Azrou (Middle Atlas, Morocco): mantle derived alumina-rich xenoliths related to the ‘ariegite–grospydite’ trend. In: Kimberlites II: The Mantle and Crust Relationship—Kornprobst J, ed. (1984) Amsterdam: Elsevier. 179–189.
Mourtada S. Pétrogènese des carbonatites et contribution à l’étude des minéralisations associées: exemple du complexe alcalin de Tamazert (Haut-Atlas Marocain). (1997) Thèse de Doctorat, Université Blaise-Pascal, Clermont-Ferrand.
Navon O, Stolper E. Geochemical consequences of melt percolation: upper mantle as a chromatographic column. Journal of Geology (1987) 95:285–307.[Web of Science]
Neumann ER, Wulff-Pedersen E. The origin of highly silicic melts in mantle xenoliths from Canary Islands. Journal of Petrology (1997) 38:1513–1539.
Nickel KG, Green DH. Empirical geothermobarometry for garnet peridotites and implications for the nature of the lithosphere, kimberlites and diamonds. Earth and Planetary Science Letters (1985) 73:423–437.
Niu Y. Mantle melting and mantle extraction processes beneath ocean ridges: evidence from abyssal peridotites. Journal of Petrology (1997) 38:1047–1074.
Oberti R, Vannucci R, Zanetti A, Tiepolo M, Brumm RC. A crystal chemical re-evaluation of amphibole/melt and amphibole/clinopyroxene DTi in petrogenetic studies. American Mineralogist (2000) 85:401–419.
O’Reilly SY, Griffin WL. Apatite in the mantle: implications for metasomatic processes and high heat production in Phanerozoic mantle. Lithos (2000) 53:217–232.[CrossRef][Web of Science]
Pearson DG, Davies GR, Nixon PH. Geochemical constraints on the petrogenesis of diamond facies pyroxenites from Beni Bousera peridotite massif, North Morocco. Journal of Petrology (1993) 34:125–172.
Peslier AH, Francis D, Ludden J. The lithospheric mantle beneath continental margins: melting and melt–rock reaction in Canadian Cordillera xenoliths. Journal of Petrology (2002) 43:2013–2047.
Piccardo GB, Zanetti A, Müntener O. Melt/peridotite interaction in the Southern Lanzo peridotite: field, textural and geochemical evidence. Lithos (2007) 94:181–209.[CrossRef][Web of Science]
Pin C, Santos Zalduegui JF. Sequential separation of light rare-earth elements, thorium and uranium by miniaturised extraction chromatography: Application to isotopic analyses of silicate rocks. Analytica Chimica Acta (1997) 339:79–89.[CrossRef][Web of Science]
Pin C, Briot D, Bassin C, Poitrasson F. Concomitant separation of strontium and samarium–neodymium for isotopic analysis in silicate samples, based on specific extraction chromatography. Analytica Chimica Acta (1994) 298:209–217.[CrossRef][Web of Science]
Putirka K, Mikaelian H, Ryerson F, Shaw H. New clinopyroxene–liquid thermometers for mafic, evolved, and volatile-bearing lava compositions, with applications to lavas from Tibet and the Snake River Plain, Idaho. American Mineralogist (2003) 88:1542–1554.
Rampone E, Hofmann AW, Piccardo GB, Vannucci R, Bottazzi P, Ottolini L. Trace element and isotope geochemistry of depleted peridotites from an N-MORB type ophiolite (Internal Ligurides, N Italy). Contributions to Mineralogy and Petrology (1996) 123:61–76.[CrossRef][Web of Science]
Rampone E, Romairone A, Hofmann AW. Contrasting bulk and mineral chemistry in depleted peridotites: evidence for reactive porous flow. Earth and Planetary Science Letters (2004) 218:491–506.[CrossRef][Web of Science]
Rivalenti G, Zanetti A, Mazzucchelli M, Vannucci R, Cingolani CA. Equivocal carbonate markers in the mantle xenoliths of the Patagonia backarc: the Gobernador Gregores case (Santa Cruz Province, Argentina). Contributions to Mineralogy and Petrology (2004) 147:647–670.[Web of Science]
Rivalenti G, Zanetti A, Girardi V. AV, Mazzucchelli M, Tassinari C. CG, Bertotto GW. The effect of the Fernando de Noronha plume on the mantle lithosphere in north-eastern Brazil. Lithos (2007) 94:111–131.[CrossRef][Web of Science]
Rudnick RL, McDonough WF, Chappel BC. Carbonatitic metasomatism in the northern Tanzanian mantle. Earth and Planetary Science Letters (1993) 114:463–475.[CrossRef][Web of Science]
Ryerson FJ, Watson EB. Rutile saturation in magmas: implications for Ti–Nb–Ta depletion in orogenic rock series. Earth and Planetary Science Letters (1987) 86:225–239.[CrossRef][Web of Science]
Takahashi N. Evidence for melt segregation towards in the Horoman mantle peridotite complex. Nature (1992) 359:52–55.[CrossRef]
Takazawa E, Frey FA, Shimizu N, Obata M, Bodinier JL. Geochemical evidence for melt migration and reaction in the upper mantle. Nature (1992) 359:55–58.[CrossRef]
Tanaka T, Togashi S, Kamioka H, Amakawa H, Kagami H, Hamamoto T, Yuhara M, et al. JNdi-1: a neodymium isotopic reference in consistency with La Jolla neodymium. Chemical Geology (2000) 168:279–281.[CrossRef][Web of Science]
Tiepolo M, Vannucci R, Oberti R, Foley SF, Bottazzi P, Zanetti A. Nb and Ta incorporation and fractionation in titanian pargasite and kaersutite: crystal-chemical constraints and implications for natural systems. Earth and Planetary Science Letters (2000) 176:185–201.[CrossRef][Web of Science]
Tiepolo M, Bottazzi P, Foley SF, Oberti R, Vannucci R, Zanetti A. Fractionation of Nb and Ta from Zr and Hf at mantle depth; the role of titanian pargasite and kaersutite. Journal of Petrology (2001) 42:221–232.
Tiepolo M, Bottazzi P, Palenzona M, Vannucci R. A laser probe coupled with ICP–double-focusing sector-field mass spectrometer for in situ analysis of geological samples and U–Pb dating of zircon. Canadian Mineralogist (2003) 41:259–272.[CrossRef][Web of Science]
Tisserant D, Thizat R, Agard J. Données géochronologiques sur le complexe de roches alcalines du Tamazeght (Haut-Atlas-de-Midelt, Maroc). Bullettin du BRGM, Section II (1976) 3:279–283.
Toramaru A, Fujii N. Connectivity of melt phase in a partially molten peridotite. Journal of Geophysical Research (1986) 91:9239–9252.
Vaggelli G, Olmi F, Conticelli S. Quantitative electron microprobe analysis of reference silicate mineral and glass samples. Acta Vulcanologica (1999) 11:297–303.
Vannucci R, Piccardo GB, Rivalenti G, Zanetti A, Rampone E, Ottolini L, Oberti R, Mazzucchelli M, Bottazzi P. Origin of LREE-depleted amphiboles in the subcontinental mantle. Geochimica et Cosmochimica Acta (1995) 59:1763–1771.[CrossRef][Web of Science]
Vannucci R, Bottazzi P, Wulff-Pedersen E, Neumann ER. Partitioning of REE, Y, Sr, Zr and Ti between clinopyroxene and silicate melts in the mantle under La Palma (Canary Islands): implications for the nature of the metasomatic agents. Earth and Planetary Science Letters (1998) 158:39–51.[CrossRef][Web of Science]
Vernières J, Godard M, Bodinier JL. A plate model for the simulation of trace element fractionation during partial melting and magma transport in the Earth's upper mantle. Journal of Geophysical Research (1997) 102(B11):24771–24784.[CrossRef]
Wagner C, Mokhtari A, Deloule E, Chabaux F. Carbonatite and alkaline magmatism in Taourirt (Morocco): petrological, geochemical and Sr–Nd-isotope characteristics. Journal of Petrology (2003) 44:937–965.
Wells P. RA. Pyroxene thermometry in simple and complex systems. Contributions to Mineralogy and Petrology (1977) 62:129–139.[CrossRef][Web of Science]
Witt-Eickschen G, Harte B. Distribution of trace elements between amphibole and clinopyroxene from mantle peridotites of the Eifel (western Germany): an ion-microprobe study. Chemical Geology (1994) 117:235–250.[CrossRef][Web of Science]
Witt-Eickschen G, Seck HA, Reys C. Multiple enrichment process and their relationship in the subcrustal lithosphere beneath the Eifel (Germany). Journal of Petrology (1993) 34:1–22.
Witt-Eickschen G, Seck HA, Mezger K, Eggins SM, Altherr R. Lithospheric mantle evolution beneath the Eifel (Germany): constraints from Sr–Nd–Pb isotopes and trace element abundances in spinel peridotite and pyroxenite xenoliths. Journal of Petrology (2003) 44:1077–1095.
Wulff-Pedersen E, Neumann ER, Vannucci R, Bottazzi P, Ottolini L. Silicic melts produced by reaction between peridotite and infiltrating basaltic melts: ion probe data on glasses and minerals in veined xenoliths from La Palma, Canary Islands. Contributions to Mineralogy and Petrology (1999) 137:59–82.[CrossRef][Web of Science]
Xu Y, Mercier J. CC, Menzies MA, Ross JV, Harte B, Lin C, Shi L. K-rich glass-bearing wehrlite xenoliths from Yitong, Northeastern China: petrological and chemical evidence for mantle metasomatism. Contributions to Mineralogy and Petrology (1996) 125:406–420.[CrossRef][Web of Science]
Xu YG, Menzies MA, Bodinier JL, Bedini RM, Vroon P, Mercier J. CC. Melt percolation and reaction atop a plume: evidence from the poikiloblastic peridotite xenoliths from Borée (Massif Central, France). Contributions to Mineralogy and Petrology (1998) 132:65–84.[CrossRef][Web of Science]
Yang H.-J, Frey FA, Weis D, Giret A, Pyle D, Michon G. Petrogenesis of the flood basalts forming the northern Kerguelan archipelago: implications for the Kerguelen plume. Journal of Petrology (1998) 39:711–748.
Yaxley GM, Kamenetsky V. In situ origin for glass in mantle xenoliths from southeastern Australia: insights from trace element compositions of glasses and metasomatic phases. Earth and Planetary Science Letters (1999) 172:97–109.[CrossRef][Web of Science]
Yaxley GM, Crawford AJ, Green DH. Evidence for carbonatite metasomatism in spinel peridotite xenoliths from western Victoria, Australia. Earth and Planetary Science Letters (1991) 107:305–317.[CrossRef][Web of Science]
Yaxley GM, Kamenetsky V, Green DH, Falloon TJ. Glasses in mantle xenoliths from Western Victoria, Australia, and their relevance to mantle processes. Earth and Planetary Science Letters (1997) 148:433–446.[CrossRef][Web of Science]
Yaxley GM, Green DH, Kamenetsky V. Carbonate metasomatism in the southeastern Australian lithosphere. Journal of Petrology (1998) 39:1917–1931.
Zanetti A, Mazzucchelli M, Rivalenti G, Vannucci R. The Finero phlogopite-peridotite massif: an example of subduction-related metasomatism. Contributions to Mineralogy and Petrology (1999) 134:107–122.[CrossRef][Web of Science]
Zanetti A, Spadea P, D’Antonio M, Raffone N, Vannucci R, Bruguier NJ. Evidence for garnet- to spinel-facies, high-degree, fractional mantle melting in supra-subduction peridotites from South Chamorro, Conical and Torishima Seamounts in the Izu–Bonin–Mariana forearc. Ofioliti (2006) 33:183–200.
Zangana NA, Downes H, Thirlwall MF, Hegner E. Relationship between deformation, equilibration temperatures, REE and radiogenic isotopes in mantle xenoliths (Ray Pic, Massif Central, France): an example of plume–lithosphere interaction? Contributions to Mineralogy and Petrology (1997) 127:187–203.[CrossRef][Web of Science]
Zhu W, Hirth G. A network model for permeability in partially molten rocks. Earth and Planetary Science Letters (2003) 212:407–416.[CrossRef][Web of Science]
Zindler A, Hart S. Chemical geodynamics. Annual Review of Earth and Planetary Sciences (1986) 14:493–571.[CrossRef][Web of Science]
Zinngrebe E, Foley SF. Metasomatism in mantle xenoliths from Gees, West Eifel, Germany: evidence from genesis of calc-alkaline glasses and metasomatic Ca-enrichment. Contributions to Mineralogy and Petrology (1995) 122:79–96.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Berger, J.-P. Liegeois, N. Ennih, and B. Bonin Flow of Canary mantle plume material through a subcontinental lithospheric corridor beneath Africa to the Mediterranean: COMMENT Geology, January 1, 2010; 38(1): e202 - e202. [Full Text] [PDF]
|
|
|
|
 |
|
 E. Rampone, R. L. M. Vissers, M. Poggio, M. Scambelluri, and A. Zanetti Melt Migration and Intrusion during Exhumation of the Alboran Lithosphere: the Tallante Mantle Xenolith Record (Betic Cordillera, SE Spain) J. Petrology, January 1, 2010; 51(1-2): 295 - 325. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||