Journal of Petrology Advance Access published online on December 4, 2007
Journal of Petrology, doi:10.1093/petrology/egm074
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Partial Melting and Counterclockwise P–T Path of Subducted Oceanic Crust (Sierra del Convento Mélange, Cuba)
1Departamento De MineralogÍa Y PetrologÍa, Universidad De Granada, Avda. Fuentenueva SN, 18002 Granada, Spain
2Institut Für Geowissenschaften, Universität Mainz, 55099 Mainz, Germany
3Instituto De GeologÍa Y PaleontologÍa, Via Blanca Y Carretera Central, San Miguel Del Padrón, 11000 Ciudad Habana, Cuba
4Fachbereich Geographie, Geologie Und Mineralogie, Universität Salzburg, Hellbrunner Strasse 34, A-5020 Salzburg, Austria
5Departamento De GeologÍa, Instituto Superior Minero-Metalúrgico, Las Coloradas De Moa, HolguÍn, Cuba
Received February 8, 2007; Revised typescript accepted October 29, 2007
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONPartial melting of subducted oceanic crust has been identified in the Sierra del Convento mélange (Cuba). This serpentinite-matrix mélange contains blocks of mid-ocean ridge basalt (MORB)-derived plagioclase-lacking epidote ± garnet amphibolite intimately associated with peraluminous trondhjemitic–tonalitic rocks. Field relations, major element bulk-rock compositions, mineral assemblages, peak metamorphic conditions (c. 750°C, 14–16 kbar), experimental evidence, and theoretical phase relations support formation of the trondhjemitic–tonalitic rocks by wet melting of subducted amphibolites. Phase relations and mass-balance calculations indicate eutectic- and peritectic-like melting reactions characterized by large stoichiometric coefficients of reactant plagioclase and suggest that this phase was completely consumed upon melting. The magmatic assemblages of the trondhjemitic–tonalitic melts, consisting of plagioclase, quartz, epidote, ± paragonite, ± pargasite, and ± kyanite, crystallized at depth (14–15 kbar). The peraluminous composition of the melts is consistent with experimental evidence, explains the presence of magmatic paragonite and (relict) kyanite, and places important constraints on the interpretation of slab-derived magmatic rocks. Calculated P–T conditions indicate counterclockwise P–T paths during exhumation, when retrograde blueschist-facies overprints, composed of combinations of omphacite, glaucophane, actinolite, tremolite, paragonite, lawsonite, albite, (clino)zoisite, chlorite, pumpellyite and phengite, were formed in the amphibolites and trondhjemites. Partial melting of subducted oceanic crust in eastern Cuba is unique in the Caribbean realm and has important consequences for the plate-tectonic interpretation of the region, as it supports a scenario of onset of subduction of a young oceanic lithosphere during the early Cretaceous (c. 120 Ma). The counterclockwise P–T paths were caused by ensuing exhumation during continued subduction.
KEY WORDS: amphibolite; Cuba; exhumation; partial melting; trondhjemite; subduction
|
INTRODUCTION |
|---|
Formation and consumption of oceanic lithosphere at ridges and trenches, respectively, are the most important geodynamic processes that cause energy and material recycling (transfer) between the lithosphere and mantle. Subduction consumes lithosphere that sinks into the mantle, but material fractionated into fluids or/and melts formed at various depths in the subducting slab is transferred to the mantle wedge, eventually triggering partial melting of ultramafic material and the formation of volcanic arcs. Because material deeply subducted to sub-arc depths of 100–200 km rarely returns to the surface (Ernst, 1999
), deciding whether fluids or melts are formed at sub-arc depths is a matter of geochemical inference, experimental work, and thermal modelling. Fluid fluxing out of the subducting slab is generally considered to be the main control for material transport to, and melting of peridotite in, the mantle wedge (e.g. Schmidt & Poli, 1998; Stern, 2002; Grove et al., 2006, and references therein). However, geochemical data (e.g. Defant & Drummond, 1990; Drummond et al., 1996; Stolz et al., 1996; Elliott et al., 1997), thermal modelling (e.g. Peacock et al., 1994; van Keken et al. 2002; Gerya & Yuen, 2003; Peacock, 2003; Kincaid & Griffiths, 2004; Conder, 2005; Abers et al., 2006), and experimental studies (e.g. Johnson & Plank, 1999; Prouteau et al., 2001; Kessel et al., 2005; Hermann et al. 2006) have shown that partial melting of subducted sedimentary and/or mafic oceanic crust plays a major role in transferring matter from the slab to the mantle wedge. However, although partial melting of subducted slabs may be a common phenomenon, natural examples of partially molten subduction complexes returned to the Earth's surface are rare.
The amphibolite blocks of the ‘Catalina Schist’ mélange, California, constitute perhaps the best-known example of exhumed fragments of subducted oceanic crust where partial melting and metasomatic mass-transfer processes took place during subduction and accretion to the upper plate mantle (Sorensen & Barton, 1987; Sorensen, 1988; Sorensen & Grossman, 1989; Bebout & Barton, 2002). The study of these processes in this and other complexes is thus critical to unravel the geochemical evolution of the subduction factory (Bebout, 2007). Here we examine partial melting processes of subducted oceanic mafic crust in the Sierra del Convento serpentinite mélange, eastern Cuba, which represents a subduction channel related to Mesozoic subduction on the northern margin of the Caribbean plate (García-Casco et al., 2006). We present descriptions of field relations, mineral assemblages and textures, and major element compositions of rocks and minerals, as well as P–T estimates for the various stages of evolution of amphibolites and their partial melting trondhjemitic–tonalitic products. These data are used to derive petrogenetic and tectonic models of formation.
|
GEOLOGICAL OVERVIEW |
|---|
Geodynamic setting
Cuba forms part of the circum-Caribbean orogenic belt, which extends from Guatemala through the Greater and Lesser Antilles to northern South America (Fig. 1a). The belt encompasses the active volcanic arc of the Lesser Antilles, where the Atlantic lithosphere subducts below the Caribbean plate. However, Cretaceous to Tertiary volcanic-arc rocks all along the belt (Fig. 1a) document a long-lasting history of subduction on this margin of the Caribbean plate. Although the details of the Mesozoic evolution of this plate margin are debated (see Iturralde-Vinent & Lidiak, 2006
; Pindell et al., 2005, 2006), geological and geophysical data indicate that the arc probably occupied a position between the Americas, similar to that of the present-day Central American Arc, during the early Cretaceous. This implies that the Caribbean plate formed in the western Pacific and drifted eastwards relative to the Americas during the Mesozoic and Tertiary (Pindell et al., 2005). As a consequence of this drift, the volcanic arc finally collided with the continental margins of North America and South America during the latest Cretaceous–Tertiary.
|
Eastward drift of the Caribbean plate began in the Aptian (c. 120 Ma; Pindell et al. 2005
), when westward subduction of the Protocaribbean (i.e. Atlantic) also began. Ocean- and continent-derived high-pressure rocks found all along the edges of the Caribbean plate formed during the early Cretaceous to Tertiary and trace the history of subduction in the region (see review by Pindell et al., 2005; also García-Casco et al., 2002, 2006; Harlow et al., 2004; Escuder-Viruete & Pérez-Estaún, 2006; Krebs et al., 2007). In Cuba, metamorphic rocks documenting subduction and accretion of oceanic crust during the Cretaceous are found in serpentinite mélanges associated with ophiolitic tectonic units (collectively termed ‘northern ophiolite belt’; Iturralde-Vinent, 1996, 1998) distributed east–west along the >1000 km length of the island (Somin & Millán, 1981; Millán, 1996; García-Casco et al., 2002, 2006; Fig. 1b). Based on an analysis of ages and P–T paths of metamorphic blocks from these mélanges, García-Casco et al. (2006) proposed that hot subduction related to Aptian onset of subduction of the Protocaribbean is recorded in the eastern Cuba mélanges. The rocks described in this paper come from one of these mélanges located in the Sierra del Convento (Fig. 1c).
Geological setting
The geology of Eastern Cuba is characterized by large amounts of volcanic arc and ophiolitic materials. The ophiolite bodies (Mayarí–Cristal and Moa–Baracoa; Fig. 1b) were grouped in the ‘eastern Cuba ophiolites’ by Iturralde-Vinent et al. (2006) to emphasize their geological–petrological differences with respect to the ‘northern ophiolite belt’ of west–central Cuba. The former have geochemical supra-subduction signatures (Proenza et al., 2006; Marchesi et al., 2007) and override Cretaceous volcanic arc complexes along north-directed thrusts. Thrusting occurred during the late Cretaceous to earliest Paleocene (Cobiella et al., 1984; Iturralde-Vinent et al., 2006). This structural arrangement differs from that of west–central Cuba, where the volcanic arc units override the northern ophiolite belt units. Another important difference with respect to west–central Cuba is that the volcanic arc units of eastern Cuba are locally metamorphosed to the blueschist facies (Purial Complex; Boiteau et al., 1972; Somin & Millán, 1981; Cobiella et al., 1984; Millán & Somin, 1985). This type of metamorphism, dated as late Cretaceous (c. 75 Ma; Somin et al., 1992; Iturralde-Vinent et al., 2006), documents subduction of the volcanic arc terrane as the result of a complicated plate-tectonic configuration (García-Casco et al., 2006) or subduction erosion processes (J. Pindell, personal communication, 2006). Platform-like Mesozoic sedimentary rocks forming the Asuncion terrane were also metamorphosed under high-P low-T conditions (Millán et al., 1985), but the timing of metamorphism is unknown.
The Sierra del Convento and La Corea mélanges (Fig. 1b) occur at the base of the ophiolite bodies of eastern Cuba, overriding the volcanic arc complexes. The two mélanges are similar and contain subduction-related metamorphic blocks of blueschist and epidote-garnet amphibolite; eclogite is rare or absent. Available K–Ar mineral and whole-rock ages in these mélanges are 125–66 and 116–82 Ma in La Corea and Sierra del Convento, respectively (Somin et al., 1992; Iturralde-Vinent et al., 1996; Millán, 1996, and references therein). These data and our unpublished Ar/Ar and sensitive high-resolution ion microprobe (SHRIMP) zircon ages (ranging from 114 to 83 Ma) from the Sierra del Convento mélange demonstrate that the history of subduction in eastern Cuba can be traced from the early Cretaceous (Aptian) to the late Cretaceous.
The Sierra del Convento mélange
The Sierra del Convento mélange is located in the south of eastern Cuba (Fig. 1b). It is tectonically emplaced on top of the Purial metavolcanic arc complex. The geology of the region is poorly known, and detailed structural and petrologic analysis is lacking. Basic field and petrographic descriptions were provided by Boiteau et al. (1972), Somin & Millán (1981), Cobiella et al. (1984), Kulachkov & Leyva (1990), Hernández & Canedo (1995), Leyva (1996), and Millán (1996).
The field relationships of the mélange are obscured by intense weathering, tropical vegetation, and recent fracturing related to important post-Eocene activity along the sinistral transcurrent Oriente Fault that connects the Puerto Rico trench with the Cayman trough (Rojas-Agramonte et al., 2005, and references therein). The mélange is made of sheared serpentinite formed by hydration of harzburgite. It lacks exotic bocks in its central and topographically highest portion (maximum height 755 m). Towards its borders, however, exotic blocks occur, forming four submélange bodies of kilometre-scale, namely (Fig. 1c) Posango (to the east), El Palenque (to the north), Sabanalamar (to the west), and Macambo (to the south). Our field observations suggest that these submélange bodies occur at the base of the Sierra del Convento mélange, in contact with the underlying volcanic arc Purial Complex.
These submélange bodies are similar and contain low-grade tectonic blocks of metabasite, metagreywacke, metapelite–semipelite, and pelitic gneiss metamorphosed to the blueschist facies with greenschist-facies overprints. The submélange bodies also contain blocks of metabasites that have been metamorphosed to the epidote–amphibolite facies. These are the ‘high-grade’ amphibolite blocks described below. Our unpublished trace element analyses of the high-grade blocks show normal mid-ocean ridge basalt (N-MORB) signatures, indicating subducted oceanic lithosphere, as opposed to other low-grade blocks of metabasite, metagreywacke and metapelite–semipelite, which represent metamorphosed volcanic-arc derived rocks of the Purial Complex.
Importantly, the high-grade amphibolite blocks are made of plagioclase-lacking peak mineral assemblages consisting of pargasite, epidote, ± quartz, ± garnet, ± clinopyroxene (the last is rare). The white material seen in the matrix of these rocks (e.g. Fig. 2a and b, except for the leucocratic segregations) is not plagioclase but epidote. Following the common usage and recent recommendations of the IUGS Subcommission on the Systematics of Metamorphic Rocks (Coutinho et al., 2007) the lack of peak metamorphic plagioclase prevents use of the term amphibolite to name these rocks. On the other hand, the abundance of epidote prevents use of the term hornblendite, which would also convey a wrong connotation. Thus, the term amphibolite is used here to name the studied rocks although their mineral assemblages do not conform to its classical use for rocks made of amphibole and plagioclase.
|
Another prominent petrological feature of the Sierra del Convento mélange is the intimate association of amphibolite with centimetre- to metre-sized layers, pockets and veins of trondhjemite–tonalite (Fig. 2). The nature and origin of these rocks has not been addressed in detail. Somin & Millán (1981
) and Millán (1996) suggested that they represent ophiolitic plagiogranite pods formed during differentiation of MOR gabbros, which were later subducted and metamorphosed together with the enclosing mafic rocks. In contrast, Hernández & Canedo (1995) suggested that the leucocratic material is post-metamorphic and identified it as material of the Cretaceous volcanic arc intruding the mélange. Based on field relations, Kulachkov & Leyva (1990) and Leyva (1996) conceded that there is a genetic relationship between the amphibolites and the leucocratic material but did not indicate the type of relationship. In this paper we concentrate on these high-grade amphibolite blocks and associated trondhjemitic–tonalitic rocks. |
FIELD RELATIONSHIPS |
|---|
The high-grade blocks of amphibolite are of metre to tens of metres size, medium- to coarse-grained, and massive to banded (Fig. 2a-c). In the Macambo region, however, the blocks appear to be larger. In the Palenque region the amphibolites form discrete blocks, commonly metre-sized. This region provides the largest amount of leucocratic material intimately associated with the amphibolites. In the Posango and Sabanalamar regions the blocks are best observed as detrital material forming part of recent alluvial deposits along the Yacabo and Sabanalamar rivers rather than as tectonic blocks within serpentinite.
The tonalitic–trondhjemitic bodies are of centimetre to metre size. They are exclusively associated with the amphibolites and not with other types of exotic blocks within the mélange, and do not appear to crosscut the serpentinite matrix. The structure of the bodies varies from concordant layers to crosscutting veins relative to the metamorphic foliation of the amphibolites. Banded, stromatic, vein-like, and agmatitic structures typical of migmatites are common (Fig. 2a–e).
The banded structure is characterized by the association of mesosome (Ep ± Grt amphibolite) and leucosome (trondhjemite–tonalite) parallel to the main syn-metamorphic foliation. Locally, melanosome material consisting of >90% modal pargasitic amphibole (plus small amounts of epidote, rutile, titanite, and apatite ± quartz) is spatially associated with banded mesosome–leucosome pairs resembling stromatic structure of migmatites (Fig. 2b). The texture of this melanosome is variable, with grain size averaging less than 1 mm in some samples to several centimetres in others. Because of the abundance of pargasite, the melanosome material is termed here hornblendite, following the recent recommendations of the IUGS Subcommission on the Systematics of Metamorphic Rocks (Coutinho et al., 2007). Our unpublished major and trace element and Rb/Sr and Sm/Nd isotope data provide evidence that these hornblendites do not represent true restites but a metasomatic rock formed by interaction (i.e. back-reaction) of the amphibolites and fluids evolved from associated melts.
Stretched, boudinaged and folded veins and pockets of trondhjemite–tonalite indicate that the amphibolites and associated segregations experienced ductile deformation (Fig. 2a–c). However, crosscutting veins and agmatitic (magmatic breccia-like) structures also document brittle behaviour of the amphibolites during melt extraction (Fig. 2d and e). In all types of structures, but most typically in the agmatitic ones, the segregations contain small (generally centimetre-size) nodules of amphibolite with sharp to diffuse boundaries, suggesting disaggregation of amphibolite within the silicic melt (Fig. 2f).
Below we present detailed information on six representative samples of plagioclase-lacking amphibolite and five representative samples of associated trondhjemite–tonalite. Samples CV53b-I and CV62b-I are quartz–epidote amphibolite, CV228d and CV228e are quartz–epidote–garnet amphibolite, CV230b is epidote-garnet amphibolite, and CV237j is epidote–garnet–clinopyroxene amphibolite. Samples CV62b-II, CV228c and SC21 are epidote–amphibole trondhjemite–tonalite, and CV53b-II and CV60a are epidote trondhjemite. The numbers in the sample labels refer to sample sites (see Fig. 1c; sample SC21 is a pebble collected in the bed of the Yacabo River). Closely associated amphibolite–trondhjemite pairs were sampled in sites CV62 and CV228. These samples are used below to examine melting relations.
|
ANALYTICAL TECHNIQUES |
|---|
Whole-rock compositions were determined on glass beads by X-ray fluorescence (XRF) using a Philips Magix Pro (PW-2440) spectrometer at the University of Granada (Table 1; trondhjemite sample CV53b-II could not be analysed because of its small size). Loss on ignition (LOI) was determined on pressed powder pellets. Mineral compositions were obtained by wavelength-dispersive spectrometry (WDS) with a Cameca SX-50 microprobe (University of Granada), operated at 20 kV and 20 nA, and synthetic SiO2, Al2O3, MnTiO3, Fe2O3, MgO and natural diopside, albite and sanidine as calibration standards, and by energy-dispersive spectrometry (EDS) with a Zeiss DSM 950 scanning microscope, equipped with a Link Isis series 300 Analytical Pentafet system, operated at 20 kV and 1–2 nA beam current, with counting times of 50–100 s, and the same calibration standards. All analyses listed in Tables 3–10
were obtained by WDS. Clinopyroxene and amphibole compositions were normalized following the schemes of Morimoto et al. (1988) and Leake et al. (1997), respectively. Garnet was normalized to eight cations and 12 oxygens, and Fe3+ was estimated by stoichiometry. Epidote, feldspar, lawsonite, and kyanite were normalized to 12·5, eight, eight, and five oxygens, respectively, and Fetotal = Fe3+. Mica, chlorite, and pumpellyite were normalized to 22, 28, and 24·5, oxygens, respectively, and Fetotal = Fe2+. Mineral and end-member abbreviations are after Kretz (1983), except for amphibole (Amp), silicate melt (L) and fluid (H2O), with end-members of phases written entirely in lower case. The atomic concentration of elements per formula units is abbreviated a.p.f.u.
|
Elemental XR images were obtained with the same Cameca SX-50 microprobe operated at 20 kV and 200–300 nA beam current, with step (pixel) size of 3–10 µm and counting time of 15–100 ms. We found that a high beam current combined with short counting time (milliseconds rather than seconds) avoids the problem of beam damage to silicates (see De Andrade et al., 2006
). The images were processed with Imager software (R. L. Torres-Roldán & A. García-Casco, unpublished) and consist of the XR signals of K
lines of the elements or element ratios (colour-coded; expressed in counts/nA per s), corrected for 3·5 µs deadtime and with voids, polish defects, and all other mineral phases masked out, overlain onto a grey-scale base-layer calculated with the expression 
[(counts/nA per s)i·Ai], (where A is atomic number, and i is Si, Ti, Al, Fe, Mn, Mg, Ca, Na, and K), which contains the basic textural information of the scanned areas. |
BULK-ROCK COMPOSITION |
|---|
A detailed assessment of the bulk-rock chemistry of the blocks in the mélange will be presented elsewhere. Here, we briefly comment on some aspects of their major element geochemistry, which help explain key features of their mineral assemblages.
The studied amphibolite samples have subalkaline, low-K (tholeiitic) basaltic composition (Fig. 3a and b). Their SiO2 content ranges from 43·25 to 48·02 wt %, with lower values in the quartz-free samples. The latter samples trend toward (apparent) picritic composition. Notably, all the samples have relatively low Na2O contents (1·29–1·85 wt %), so that the samples have a relatively high inverse agpaitic index as compared with MORB (Fig. 3d). The garnet-bearing samples have lower Mg-number [Mg/(Mg + 
)] atomic proportions; Table 1).
|
The composition of the tonalite–trondhjemite samples is varied. In the total alkalis–silica (TAS) diagram they range from andesite through trachyandesite, trachyte and trachydacite to dacite (Fig. 3a). They are rich in Na2O (5·58–7·55 wt %) and have very low K2O contents (0·08–0·26 wt %; Fig. 3b, Table 1), resembling rocks of the low-K tholeiite series. The FeO and MgO contents are low, and the Mg-number is relatively high, ranging from 0·38 to 0·67 (Fig. 3c). All the samples are classified as trondhjemite in the O’Connor–Barker (O’Connor, 1965
; Barker, 1979) An–Ab–Or classification scheme for granitic rocks (not shown), except for sample SC21, which is classified as tonalite. However, sample SC21 is in fact leucocratic, having a greater amount of epidote (as a result of higher CaO contents). For this reason all the samples are henceforth collectively termed trondhjemites. Importantly, the trondhjemites are Al-saturated (i.e. peraluminous; ASI = 1·023–1·142; Fig. 3d) and plot close to the alkali feldspar–white mica mixing line delineated in Fig. 3d. It should be noted that this line indicates albite–paragonite mixtures in K-poor rocks. |
TEXTURES AND MINERAL ASSEMBLAGES |
|---|
Amphibolites
Peak metamorphic mineral assemblages consist of pargasitic amphibole, epidote, ± garnet, ± quartz, ± diopsidic clinopyroxene (rare), rutile, titanite, and accessory apatite (Fig. 4a–f; Table 2). These assemblages define a crude foliation. Quartz appears as small dispersed grains and millimetre-sized ‘pockets’ elongated along the foliation, although quartz-lacking samples are also common. Syn- to post-kinematic garnet is abundant but not present in all samples (Fig. 4a–f; Table 2). Its occurrence is influenced by bulk composition, as indicated by the lower Mg-number of garnet-bearing samples (Table 1). If forms large porphyroblasts, 0·5–3 cm in diameter, containing inclusions of amphibole and epidote (Fig. 4a–f). Na-rich diopsidic clinopyroxene has been found in quartz-free samples of the Macambo region. It occurs as medium-grained millimetre-sized grains in the matrix (Fig. 4d–f). Titanite appears as idiomorphic peak metamorphic crystals elongated along the foliation, but it also replaces rutile.
|
|
Retrograde overprints consist of combinations of glaucophane, actinolite, albite, (clino)zoisite, chlorite, pumpellyite, and, less abundantly, phengitic mica and omphacite, the latter being observed only in samples containing peak metamorphic diopsidic clinopyroxene (Fig. 4a–f, Table 2). All these retrograde minerals are fine-grained and corrode the peak-metamorphic minerals, but they also are dispersed in the matrix or located in fractures. Retrograde glaucophane is typically aggregated with actinolite, chlorite and albite, and commonly overprints peak amphibole (Fig. 4c and f). Retrograde omphacite replaces peak diopsidic clinopyroxene and is also present in retrograde assemblages consisting of glaucophane + magnesiohornblende–actinolite + albite + epidote, which appear dispersed in the matrix and replacing pargasitic amphibole and garnet (Fig. 4f). Chlorite and pumpellyite replace garnet and pargasite. Scarce phengitic mica appears dispersed in the matrix of some samples. Retrograded crystals of pargasite commonly contain small needles of ‘exsolved’ rutile or titanite.
Trondhjemitic segregations
The magmatic mineral assemblage of the trondhjemites is composed of medium-grained plagioclase and quartz with subordinate medium-grained paragonite, epidote, ± pargasite, plus accessory apatite, titanite, and rutile (Fig. 4g–m; Table 2). Kyanite is locally present as tiny relict inclusions within magmatic epidote (Fig. 4m). Epidote and pargasite are idiomorphic and medium-grained (1–3 mm in length). Pargasite frequently has a large aspect ratio of 5:1 (Fig. 4j). Paragonite has medium grain size (2–3 mm in length) and idiomorphic habit, and is frequently corroded by plagioclase and fine-grained quartz, ± K-feldspar ± phengite (Fig. 4h and l). Magmatic paragonite has been identified in samples SC21, CV228c and CV60a, whereas its presence is more uncertain in samples CV62b-II and CV53B-II.
Retrograde mineral assemblages overprint the magmatic assemblages. Magmatic plagioclase appears generally transformed to retrograde albite plus fine-grained (clino)zoisite, paragonite and, locally, lawsonite (Fig. 4g and k; Table 2). Magmatic epidote is overprinted by fine-grained overgrowths of (clino)zoisite + quartz symplectite, which extend outwards from the magmatic epidote crystal and invade nearby plagioclase (Fig. 4i, k, and m). Pargasitic amphibole is partly replaced by magnesiohornblende–tremolite, chlorite, and pumpellyite (Fig. 4j). Glaucophane is not present. Titanite replaces rutile. If present, small laths of retrograde phengitic mica are dispersed in the matrix, associated with retrograde paragonite–chlorite, and replacing magmatic plagioclase and paragonite (Fig. 4l). Very small amounts of retrograde K-feldspar, typically replacing magmatic paragonite, are present in some samples.
|
MINERAL COMPOSITION |
|---|
Amphibole
Amphibolites
Matrix amphibole is edenitic–pargasitic in composition (Fig. 5a). Zoning is faint (Fig. 4a and c), and appears either as patches or, occasionally, as faint ill-defined concentric bands. Concentric zoning is defined by cores of edenite and outer shells of pargasite, indicating prograde growth (Fig. 6). Peak metamorphic pargasitic compositions are rich in Na-in-A (maximum 0·85 a.p.f.u.), total Al (maximum 2·70 a.p.f.u.), and Ti (maximum 0·3 a.p.f.u.), and poor in Si (minimum 6·19 a.p.f.u.), Na-in-M4 (minimum 0·10 a.p.f.u.), and Mg-number (minimum 0·506; Figs 5a and 6; Table 3). As would be expected for high-variance mineral assemblages, the composition of amphibole is strongly controlled by bulk-rock composition: amphibole with higher Mg-number is present in rocks with higher bulk Mg-number (Fig. 5a). Notably, Na(M4) is relatively high in pre-peak (edenitic) compositions (Fig. 6), in particular in garnet-free samples, where it is up to 0·47 a.p.f.u. and almost reaches magnesiokatophorite compositions. This edenite-to-pargasite prograde zoning correlates with the composition of amphibole inclusions within garnet, which ranges from edenite to pargasite but does not reach peak pargasitic composition as recorded in matrix amphibole.
|
|
|
The outermost rims (50–100 µm in width) of matrix amphibole show zoning from pargasite through edenite–magnesiohornblende to actinolite (Figs 4b and 5a), indicating retrograde growth or overprint. Actinolitic compositions are poor in Na(A) (maximum 0·02 a.p.f.u.), total Al (maximum 0·26 a.p.f.u.), and Ti (maximum 0·001 a.p.f.u.), and rich in Si (minimum 7·88 a.p.f.u.) and Mg-number (minimum 0·78; Fig. 5a; Table 3). Na(M4) of retrograde edenite and magnesiohornblende reaches 0·49 a.p.f.u. (Fig. 6), and a few analyses attain magnesiokatophorite and barroisite compositions.
The patchy zoning of most matrix crystals reflects a compositional spectrum comprising pargasite–magnesiohornblende–actinolite and indicates that it developed during retrograde adjustment or growth. These retrograde adjustments are clearly developed in single crystals of pargasite containing small needles of exsolved rutile or titanite.
The compositional variability of retrograde glaucophane is relatively large [Si 7·67–7·99, Al 1·46–1·98, Ca 0·04–0·43, Na(M4) 1·54–1·91, Na(A) 0·01–0·17 a.p.f.u., and Mg-number 0·51–0·64; Fig. 5a; Table 4]. This compositional range reflects compositional heterogeneity within single samples as a result of the effects of local effective bulk-composition at reaction sites and the timing of formation during the retrograde path. The first effect is best illustrated by higher Mg-number and lower Mg-number glaucophane grains present in single samples and grown adjacent to matrix amphibole and garnet, respectively.
|
Trondhjemites
Magmatic amphibole shows no growth zoning. Its composition is pargasitic [Si down to 6·01, Al up to 2·87, Ti up to 0·26, Na(A) up to 0·72, K(A) up to 0·11, Na(M4)
0·35 a.p.f.u., Mg-number 0·75; Fig. 5b; Table 3], similar to that of peak metamorphic amphibole from the amphibolites except for having higher Mg-number (Fig. 5b). The composition of these grains is overprinted by patchy retrograde zoning, developed along fractures, exfoliation planes and crystal rims (Fig. 4j). The retrograde composition of amphibole ranges from pargasite through edenite, magnesiohornblende to tremolite (Fig. 5b; Table 3). Notably, Na(M4) first increases along this path (pargasite to edenite), reaching 0·44 a.p.f.u. and approaching sodic–calcic (magnesiokatophorite–magnesiotaramite) composition, then decreases almost to zero in the magnesiohornblende–tremolite compositions. This trend suggests a retrograde path with a first step of decreasing temperature at near-constant pressure, followed by decreasing temperature and pressure.
Garnet
Garnets in the amphibolites are relatively rich in almandine (Xalm = 0·50–0·55) and, to some extent, grossular (0·2–0·3), and poor in pyrope (0·15–0·20) and spessartine (0·02–0·10), comparable with type C (i.e. low-temperature) eclogitic garnet (Fig. 7a; Table 5). As in amphibole, the composition of garnet is influenced by bulk-rock composition; higher Mg-number is observed in garnets from Fe-poorer bulk-rock composition (Fig. 7b).
|
|
Garnet zoning is faint. In most samples the cores are near homogeneous. The rims and regions adjacent to fractures show retrograde readjustments denoted by an increase in spessartine and decrease in Mg-number (Figs 4a, d and 7b–d). However, garnet in sample CV230b shows spessartine-poor and high Mg-number rims denoting prograde growth (see García-Casco et al., 2006
). Faint prograde zoning and the lack of low-temperature phases included within garnet suggest a relatively late stage of garnet growth during prograde metamorphism (i.e. close to and during peak conditions). Thus, much of the compositional variation of garnet is the result of retrograde readjustments characterized by increase in Mn and decrease in Mg-number (Fig. 7b). Grossular contents may not be significantly affected by retrograde readjustments, but they increase sharply (to Xgrs 0·6) in fractures traversing garnet in sample CV237j (Fig. 7e; Table 5).
Epidote
Matrix epidote from amphibolite shows variable Fe3+ contents, with pistacite contents (Xps = Fe3+/[(Al – 2) + Fe3+)] ranging from 0·2 to 0·7, except in sample CV230b, where Xps = 0·1–0·2 (Table 6). Zoning is generally patchy, although some grains may show concentric zoning with lower Xps at the rims (Fig. 4b), probably reflecting retrograde readjustments or growth.
|
In the trondhjemitic segregations magmatic epidote is richer in Fe3+ than retrograde grains of (clino)zoisite (Xps up to 0·5 and 0·1, respectively; Fig. 4i and k; Table 6). The exception is sample CV60a, which contains magmatic epidote with low Xps (<0·1; Table 6). Magmatic crystals show patchy zoning, with irregular high-Fe3+ areas in the interior, which probably formed at higher temperature. Intermediate compositions of magmatic epidote are interpreted as the result of crystallization upon cooling and/or subsolidus retrograde readjustments.
Clinopyroxene
Peak metamorphic clinopyroxene from sample CV237j is diopsidic in composition (Fig. 8; Table 5), with Mg-number = 0·67–0·74, Fe3+/(Fe3+ + Fe2+) = 0·18–0·37, Ti up to 0·03 a.p.f.u., and Na = 0·07–0·09 a.p.f.u. (Xjd = 0·04–0·06). This suggests a relatively high pressure of formation at high temperature. The composition of retrograde omphacite is Ca = 0·51–0·63 a.p.f.u., Mg-number = 0·67–0·82, and Fe3+/(Fe3+ + Fe2+) = 0·23–0·62, which translates into Xjd = 0·32–0·39 and Xae = 0·06–0·14 (Fig. 8; Table 5).
|
Plagioclase
Plagioclase in the amphibolites is retrograde and almost pure albite in composition (Xab > 0·92, with most analyses reaching Xab > 0·99; Table 7).
|
The composition of magmatic plagioclase in the trondhjemites is uncertain, as it has been largely retrogressed to albite-rich plagioclase. The magmatic crystals display retrograde patchy zoning that preserves irregular calcic regions dispersed within albite-richer retrograde areas (Fig. 4g). The maximum Xan of these relict regions, identified with the aid of XR images, are 0·29 (CV228c), 0·18 (SC21), 0·18 (CV62b-II), and 0·17 (CV53b-II and CV60a) (Fig. 11b, Table 7). Although these figures do not necessarily represent the original magmatic composition of plagioclase, they suggest that it was sodic andesine to calcic oligoclase.
Paragonite
Magmatic crystals of paragonite from the trondhjemites are rich in the muscovite component (K up to 0·35 a.p.f.u., XK = K/(K + Na) = 0·18; Table 8). These crystals show patchy zoning with relict high-K areas in the interior of the grains with relatively high Ca contents (0·06–0·10 a.p.f.u.) and K-poorer areas distributed irregularly but typically along (001) planes and close to the rims, indicating retrograde readjustments upon cooling (Fig. 4h). The retrograded regions are somewhat richer in the margarite component (Fig. 9). Discrete grains of retrograde paragonite have Ca- and K-poor compositions and are of almost pure paragonite end-member composition (Figs 4h and 9, Table 8). A detailed assessment of the composition of paragonite from sample SC21 has been given by García-Casco (2007).
|
|
Phengite
Retrograde phengitic mica in the amphibolites (samples CV53b-I and CV62b-I) is very rich in celadonite content, with Si = 6·89–6·96 a.p.f.u., Mg = 0·77–0·86,
= 0·21–0·27, Na = 0·06–0·09, and Mg-number (Fe = ) = 0·76–0·79 (Fig. 9; Table 8), indicating high pressure during retrogression. The composition of phengitic mica in the trondhjemites is heterogeneous (Fig. 9; Table 8) and varies from high celadonite content (Si up to 6·97, Fe up to 0·21, Mg up to 0·83 a.p.f.u.) to almost pure muscovite (Si down to 6·05 a.p.f.u.). Mg-number ranges from 0·69 to 0·92 and shows a positive correlation with Si, which is best shown when analysing the chemical variation of phengite in single samples. Ba contents are low (0·06–0·01 a.p.f.u.). The contents of Na (0·27–0·05 a.p.f.u.) and Ti (0·014–0 a.p.f.u.) are low to very low, with lower contents in celadonite-rich compositions (Fig. 9; Table 8). These compositional variations are consistent with continuing growth or readjustment during retrogression, with the onset of growth at relatively high temperature being represented by the lower Si and higher Na and Ti compositions.
Chlorite
Retrograde chlorite in the amphibolites has Al = 4·34–5·17 a.p.f.u., Mn = 0·015–0·127 a.p.f.u., and Mg-number (Fe2+ = Fetotal) = 0·50–0·71 (Table 9). Part of this compositional variability is due to the effect of bulk composition, with high Mg-number compositions from high Mg-number samples CV53b-I and CV62b-I and low Mg-number compositions from low Mg-number sample CV237j. The Al- and Mn-richer compositions are found in chlorite from Fe-richer samples bearing garnet. In these samples retrograde chlorite occurs adjacent to garnet and within the amphibolite matrix, and its composition mimics the site of growth, with higher Fe, Al, and Mn contents in the former and higher Mg contents in the latter.
|
Retrograde chlorite in the trondhjemites is similar to that of the amphibolites (Al = 4·49–5·11 a.p.f.u.), except for being more magnesian (Mg-number = 0·68–0·87; Table 9).
Other minerals
Retrograde pumpellyite in the amphibolite samples has Al = 4·81–4·92 a.p.f.u., Mg = 0·62–0·78 a.p.f.u., = 0·34–0·53 a.p.f.u., and Mg-number (Fe2+ = Fetotal) = 0·56–0·70, whereas in the trondhjemite samples it has Al = 4·47–4·93, Mg = 0·82–1·09 a.p.f.u., = 0·20–0·40 a.p.f.u., and Mg-number (Fe2+ = Fetotal) = 0·70–0·83 (Table 10). This again reflects the effect of bulk-rock composition. Titanite has up to 0·072 Al (per five oxygens). Retrograde K-feldspar from the trondhjemites is almost pure orthoclase in composition, with Xab = 0·004–0·025 and trace amounts of Ba (0·001–0·003 a.p.f.u.; Table 7). Relict magmatic kyanite and retrograde lawsonite from the trondhjemites are almost pure in composition (Table 10).
|
|
PEAK METAMORPHIC AND MAGMATIC PHASE RELATIONS |
|---|
Amphibolites
The peak metamorphic assemblages of the amphibolites are systematized in the ACF and AFN diagrams of Fig. 10. These diagrams are projected from coexisting phases and appropriate exchange vectors, which allow condensation of the composition space. To give an indication of the chemistry of the peak metamorphic minerals these phase diagrams also show the projection of relevant end-members of the solid solutions of interest. Because the ACF diagram is constructed after projection from quartz, the lack of quartz in sample CV237j helps explain the apparent reaction relation between the peak metamorphic minerals plotted in Fig. 10c.
|
The ACF diagrams (Fig. 10a–c) show that epidote is paragenetic with peak pargasitic amphibole ± garnet ± clinopyroxene, indicating epidote-amphibolite facies conditions. These diagrams also help interpret the lack of peak metamorphic plagioclase in the studied samples. This is best illustrated by the quartz–epidote amphibolite samples CV53b-I and CV62b-I (Fig. 10a). The bulk compositions of these samples plot within the tie-lines connecting peak metamorphic epidote and pargasite, implying that no plagioclase was stable in these bulk compositions at peak metamorphic conditions. The same conclusion is reached for the garnet-bearing samples after inspection of the ACF diagram (Fig. 10b and c). However, the lack of peak metamorphic plagioclase in these samples is best appreciated in the AFN diagram projected from epidote (Fig. 10d). In this diagram, the bulk compositions of the epidote–garnet amphibolite samples CV228e and CV228e plot within the tie-lines connecting garnet and pargasite. It should be noted that plagioclase would have been stable only in bulk compositions richer in Na2O and/or Al2O3, such as those of average MORB plotted in Fig. 10d. The relations depicted in the AFN diagram of Fig. 10d strongly suggest that plagioclase consumption in, and alkali subtraction from, a MORB-like amphibolite by means of partial melting and subsequent melt extraction, respectively, are mechanisms that could account for the development of the studied bulk compositions and mineral assemblages.
Trondhjemitic segregations
The magmatic assemblages of the trondhjemites are systematized in the ACF and ACN diagrams of Fig. 11a and b, respectively, projected from coexisting phases and appropriate exchange vectors. It should be noted that the compositions of the trondhjemite samples plot in the respective peraluminous fields of both diagrams. The tie-line crosscutting relations depicted in the ACF diagram (Fig. 11a) do not necessarily represent reaction between magmatic pargasite, epidote, plagioclase and paragonite. These apparent reaction relations are best explained as a result of condensation of the composition space by projection from exchange vectors. A better representation of the magmatic assemblages is provided by the ACN diagram projected from pargasite (Fig. 11b). In this diagram no tie-line crosscutting relations exist between the magmatic minerals. Also, it clearly shows that the bulk composition of the trondhjemite magmas is appropriate for crystallization of magmatic paragonite from the melt.
|
The ACN diagram of Fig. 11b shows variable shapes of the plagioclase–epidote–paragonite tie-triangles. This is a consequence of the variable composition of magmatic plagioclase, which in turn is a result of variable bulk composition of the samples (i.e. magmas). This is indicated by the fact that Na-richer magmatic plagioclase is found in Na-richer samples (CV62b-II). However, as noted above, plagioclase is overprinted by albite, paragonite, and (clino)zoisite, with relict regions having higher Ca contents (Fig. 4g). Although these regions were identified with the aid of XR images, it is probable that they do not retain peak Ca contents. Consequently, the variable shapes of the plagioclase–epidote–paragonite tie-triangles shown in Fig. 11b may be flawed. This has important consequences for thermobarometry, as discussed below.
|
P–T CONDITIONS AND P–T PATHS |
|---|
Temperatures and pressures were estimated following the optimal P–T method of Powell & Holland (1994
) using THERMOCALC (Holland & Powell, 1998, version 3.25, dataset 5.5, 12 November 2004). Calculations were performed using different combinations of phases grown during pre-peak, peak, and post-peak conditions, as described below for each sample. An H2O-fluid was included in all assemblages. All calculations are based on at least four independent reactions. The activities and (1
) activity uncertainties of each end-member included in the calculations were obtained with software AX (T. Holland & R. Powell, unpublished). As indicated by Powell & Holland (1994), the uncertainties on the calculated P and T are not altered by a proportional change (e.g. doubling) of all the activity uncertainties, as long as the averaging is statistically consistent. The P–T uncertainties (T and P) given represent ±1 (95% confidence). The correlation between T and P for a given calculation is given below as ‘corr’. High correlation indicates that with a T (or P) the other value is well constrained. Thus, large T and P imply well-constrained P–T data if the former are highly correlated. The T and P uncertainties and correlations are appropriately incorporated into the uncertainty ellipses of Fig. 12 calculated following Powell & Holland (1994). To keep the calculated P–T error ellipses small and improve the statistics of the calculations, phase components with very low activities were neglected in the calculations, and some phase components with errant behaviour (outliers) were occasionally rejected [for justification see Powell & Holland (1994)]. The calculations of pre-peak and peak conditions passed the ‘sigfit’ test for statistical consistency, but some of the calculations of magmatic and retrograde conditions did not pass this test, suggesting equilibrium problems. The test implies that the ‘sigfit’ values (quoted below for each calculation) should approach unity [for further details see Powell & Holland (1994)].
|
Amphibolites
For thermobarometry, it is not possible to use the high-variance mineral assemblage amphibole + epidote + quartz of the garnet-free amphibolites because no set of linearly independent reactions can be found. Consequently, P–T calculations were performed only for the lower-variance (garnet-bearing) samples (Fig. 12a–c).
The calculated peak P–T conditions are based on the matrix assemblages Grt + Amp + Ep + Qtz (samples CV228d and CV228e), Grt + Amp + Ep (CV230b), and Grt + Amp + Ep + diopsidic Cpx (CV237j). The assemblages used for P–T calculations are shown in the ACF and AFN phase diagrams of Fig. 10 (see also Fig. 14). Pre-peak conditions were calculated using the composition of inclusions within garnet and the assemblages Grt + Amp + Ep + Qtz (sample CV228d) and Grt + Amp + Ep (CV230b). Retrograde conditions were calculated using actinolitic Amp + Gl + Chl + Ep + Grt (retrograded rims) + Qtz + Ab (samples CV228d and CV228e), actinolitic Amp + Gl + Chl + Ep + Grt (retrograded rims) + Ab (CV230b), and actinolitic Amp + Gl + Chl + Ep ± Grt (retrograded rims) + Omp + Ab (CV237j).
The calculated peak conditions are 673 ± 49°C, 14·6 ± 2·2 kbar (corr 0·455, sigfit 1·13), 683 ± 44°C, 15·7 ± 2·1 kbar (0·476, 1·05), and 708 ± 69°C, 13·8 ± 2·2 kbar (–0·251, 0·81) for samples CV228d, CV228d, and CV230b, respectively, suggesting that these blocks underwent similar P–T conditions in the range of 675–700°C and 14–16 kbar. However, peak conditions for CV273j are more uncertain. Using the assemblage indicated above with the composition of diopsidic clinopyroxene with the lowest, average, and highest Na contents, the calculated conditions are 859 ± 46°C, 13 ± 2·2 kbar (0·181, 1·04), 811 ± 56°C, 14·4 ± 2·7 kbar (0·134, 1·31) and 768 ± 56°C 15 ± 2·7 kbar (0·107, 1·36), respectively, suggesting higher temperatures and similar to somewhat lower pressures than the above three samples. As discussed below, the results calculated with the highest Na content of diopsidic clinopyroxene are consistent with expected phase relations.
The calculated pre-peak conditions for samples CV228d and CV230b are slightly lower than those for peak conditions, down to 644 ± 42°C, 13·1 ± 1·9 kbar (0·514, 0·88) and 610 ± 62°C, 13·9 ± 2 kbar (–0·218, 0·96), respectively.
The calculated retrograde conditions are 525 ± 28°C, 11·2 ± 1·6 kbar (0·29, 2·35), 520 ± 17°C, 11·2 ± 1 kbar (0·345, 1·49), and 472 ± 21°C, 10 ± 1 kbar (0·473, 1·48) for samples CV228d, CV228d, and CV230b, respectively, suggesting that these blocks underwent similar retrograde P–T paths. The retrograde conditions for sample CV237j are 512 ± 33°C, 11·9 ± 1·3 kbar (0·155, 2·64) and 444 ± 29°C, 10·6 ± 1·1 kbar (0·437, 2·19) with or without retrograded garnet, respectively, included in the calculations.
These P–T calculations suggest counterclockwise P–T paths characterized by pre-peak prograde paths with increasing P and T within the epidote–amphibolite facies, peak conditions within the epidote–amphibolite facies, and retrogression within the blueschist facies (Fig. 12a–c).
Trondhjemitic segregations
As for the amphibolites, P–T calculations were performed using the lower-variance pargasite-bearing assemblages of samples CV228c, SC21, and CV62b-II (Fig. 12d). The calculated magmatic P–T conditions are based on the assemblages Amp + Ep + high-K Pa + Qtz + high-Ca Pl. The assemblages used for P–T calculations are shown in the ACF and ACN phase diagrams of Fig. 11. Retrograde conditions were calculated using tremolitic Amp + Chl + Ep + low-K Pa + Lws + Qtz + Ab (sample CV228c) and tremolitic Amp + Chl + Ep + low-K Pa + Qtz + Ab (samples SC21 and CV62b-II).
The calculated magmatic conditions are 747 ± 88°C, 14·7 ± 3·2 kbar (0·988, 0·95), 793 ± 96°C, 19·5 ± 4 kbar (0·97, 0·65), and 730 ± 81°C, 16·8 ± 3·6 kbar (0·983, 0·51) for samples CV228c, SC21 and CV62b-II, respectively. The average calculated temperatures (c. 750°C) are within error of calculated temperatures for the amphibolites. The pressure for CV228c overlaps with those for the common (Cpx-lacking) amphibolites. However, the pressures for samples SC21 and CV62b-II are higher. This is a consequence of the higher albite contents of plagioclase of these samples (Fig. 11b), and suggests that the equilibrium composition of magmatic plagioclase is probably not recorded in the analyses performed, as indicated above. Indeed, the P–T conditions calculated for SC21 using the composition of plagioclase from sample CV228c (probably closer to the original composition of magmatic plagioclase) are 749 ± 89°C at 15·4 ± 3·3 kbar (0·988, 0·55), similar to those calculated for CV228c and approaching the pressures calculated for peak conditions in the amphibolites. Consequently, these latter P–T conditions for sample SC21 are plotted in Fig. 12d (ellipse with slightly higher peak P).
The conditions calculated for retrogression are 356 ± 18°C, 4·7 ± 1 kbar (0·963, 0·31), 478 ± 43°C, 9·9 ± 2·2 kbar (0·638, 2·15), and 550 ± 21°C, 7·1 ± 0·9 kbar (0·697, 0·79) for samples CV228c, SC21 and CV62b-II, respectively. These conditions are consistent with the retrograde conditions calculated for the amphibolites and indicate cooling at high pressure.
|
DISCUSSION |
|---|
A subduction-related migmatitic complex
Field relations in the Sierra del Convento mélange show that the trondhjemitic–tonalitic bodies are closely related to blocks of amphibolite (Fig. 2) and do not crosscut other types of block or the serpentinite matrix of the mélange. This implies that the trondhjemitic–tonalitic melts do not represent exotic intrusions of volcanic-arc magmas and that these melts formed prior to incorporation of the amphibolite blocks into the mélange. On the other hand, the agmatitic structures and veins crosscutting the peak-metamorphic foliation of the amphibolite blocks (Fig. 2d and e) indicate that the melts did not form prior to subduction (i.e. they are not oceanic plagiogranites). The structures show that melt formation and segregation took place during and shortly after ductile deformation at near-peak metamorphic conditions, implying formation by partial melting of amphibolite in the subduction environment.
These inferences are in agreement with the major element composition of the trondhjemites. The peraluminous character of these rocks suggests that they do not represent oceanic plagiogranites or adakitic magmas, which are typically metaluminous (Fig. 3d). The low K2O contents of the studied rocks also argue against typical adakite magma (Fig. 3b and d). Although less silicic, however, the trondhjemites of the Sierra del Convento mélange are comparable with peraluminous melts of the Catalina Schist formed during partial melting and metasomatism of subducted oceanic crust (Sorensen & Barton, 1987; Sorensen, 1988; Sorensen & Grossman, 1989; Bebout & Barton, 2002; Fig. 3d).
The major element bulk composition of the amphibolites, on the other hand, underlines their residual character attained after extraction of melt. The amphibolites have lower SiO2 (trending toward apparent picritic compositions) and alkali components, notably Na2O, and higher molar Al2O3/(Na2O + K2O) than average MORB (Fig. 3). As discussed below, melting of amphibolite consumed relative large amounts of plagioclase up to its total consumption, a process consistent with these geochemical features if the primary tonalitic–trondhjemitic liquid was extracted after melting (see Fig. 10d). Not surprisingly, the above-mentioned geochemical characteristics are similar to those of residual amphibolite from the Catalina Schist (Fig. 3).
Further evidence for partial melting of amphibolite is given by the calculated peak P–T conditions, which lie above the H2O-saturated basaltic solidus (Figs 12a, b and 13a). These conditions suggest amphibolite melting at relatively low temperature, below the maximum stability of epidote in the basaltic system (Figs 12a, b and 13). This is in agreement with epidote forming part of the peak metamorphic supersolidus assemblage of the studied amphibolite samples (Fig. 10). In addition, such mineral assemblages and melting conditions indicate that melting occurred under H2O-present conditions, and that the dry solidus of amphibolite was not intersected (Fig. 12a and b) except perhaps in the Cpx-bearing sample (Fig. 12c).
|
Wet melting of amphibolite
H2O-fluid strongly reduces the stability of plagioclase upon melting in the basaltic system, in particular at moderate to high pressure. This effect can be appreciated in the H2O-saturated and 5 wt % added H2O ‘experimental pseudosections’ compiled by Green (1982
) for a standard tholeiitic basaltic composition (Fig. 13). The relations shown in these pseudosections indicate that the stable assemblage appropriate for the calculated peak P–T conditions is Ep + Amp + Cpx + Grt + L ± Qtz, and that the studied amphibolite samples should have contained plagioclase before partial melting.
The pseudosections of Fig. 13 also suggest that garnet and clinopyroxene are expected after wet melting of amphibolite. As discussed above, the lack of garnet in the higher Mg-number samples can be conceptualized as the result of a bulk composition effect. The lack of clinopyroxene in most samples, however, suggests that this phase is not a necessary product of fluid-present melting of amphibolite. Indeed, whereas experiments indicate that clinopyroxene is a systematic product of melting of rocks of basaltic composition under H2O-deficient conditions, they also suggest that this phase does not normally form upon wet melting near the solidus (Yoder & Tilley, 1962; Lambert & Wyllie, 1972; Heltz, 1973, 1976; Beard & Lofgren, 1991; Winther & Newton, 1991; Selbekk & Skjerlie, 2002). Consequently, the studied rocks demonstrate that clinopyroxene is not a necessary product of wet melting of amphibolite near the solidus at intermediate pressure, in agreement with proposals by Ellis & Thompson (1986) and Thompson & Ellis (1994).
The calculated conditions for the Cpx-bearing amphibolite sample CV237j (750–850°C, 13–15 kbar) are above the maximum thermal stability of epidote in the water-saturated basaltic system (Figs 12c and 13a). This is not in agreement with textural evidence in this sample, where epidote forms part of the peak assemblage. This conflict is resolved if the peak conditions correspond to those calculated with the maximum Na content of diopsidic clinopyroxene (Fig. 12c) and water availability was more limited (Fig. 13b).
Melting reactions
Establishing the precise nature of the wet-melting reactions undergone by the studied amphibolite samples is hampered by (1) the lack of indication of the composition of peak metamorphic plagioclase present in the samples before melting and (2) the possibility of post-melting processes, such as back-reaction between melt and amphibolite (e.g. Kriegsman, 2001) and/or magmatic differentiation, which would have modified the composition of pristine primary melts. However, the phase relations depicted in the AFM-like diagrams of Fig. 14, combined with the analysis of the respective reaction space, allow us to constrain the nature of the melting reactions for each type of amphibolite. The analysis of reaction space has been performed in the nine-component system SiO2–TiO2–Al2O3–FeO–MgO–CaO–Na2O–K2O–P2O5 using the software CSpace (Torres-Roldán et al., 2000). Component H2O and, consequently, the fluid phase, were excluded from consideration because the H2O contents of the melts are unknown, although an H2O-fluid should be considered part of the reactant assemblages. Also, Fe was treated as FeOtotal and the exchange vector KNa–1 was included in the calculation because of the lack of K-bearing phases during partial melting. With these and other constraints related to system degeneracy discussed below, the reaction space for each case discussed in the following paragraphs is uni-dimensional, and the resulting single reaction represents a mass balance that can be potentially identified as a melting reaction. All the reactions are expressed in oxy-equivalent units.
|
Garnet-lacking amphibolites
The mineral assemblage of the garnet-lacking amphibolite samples is represented in the AFM-like diagram of Fig. 14a by no more than peak pargasite (note that the phase relations are projected from epidote). This allows deduction of the composition of coexisting plagioclase before melting, provided that the composition of the melt is known. As an example, Fig. 14a shows the relations for sample site CV62b. If (1) trondhjemite vein CV62b-II represents the primary melt evolved from adjacent amphibolite CV62b-I, and (2) plagioclase was totally consumed upon melting, the composition of peak metamorphic plagioclase is constrained to be collinear with peak pargasite and melt, implying a degenerate relation (Fig. 14a). The corresponding calculated composition of plagioclase is Xab = 0·78 and the associated reaction is
|
|
The melting reaction deduced above is near-eutectic (i.e. Pl + Qtz + Amp + Ep = L). This general form remains unchanged even if the primary melt has a different composition from that of trondhjemite vein CV62b-II. Mass-balance constraints imply collinearity between amphibole, primary melt and plagioclase, and large relative amounts of reactant plagioclase in the associated degenerate reaction for any other tonalitic–trondhjemitic primary melt composition rich in plagioclase component. Consequently, we consider that the eutectic melting reaction Pl + Qtz + Amp + Ep = L is a good approximation to the wet-melting process undergone by the garnet-lacking amphibolites.
Garnet-bearing amphibolites
As an example, we discuss the relations for sample site CV228, where amphibolite CV228e is closely associated with trondhjemite segregate CV228c (Fig. 14b). The latter sample is taken as a preliminary guess for the primary melt. Fixing the composition of the primary melt does not uniquely constrain the composition of plagioclase coexisting with the mineral assemblage of amphibolite CV228e. Three reactions are possible depending on the composition of plagioclase. Plagioclase with Xab = 0·69 is collinear with amphibole and melt, implying a degenerate reaction involving no garnet (i.e. Pl + Qtz + Amp + Ep = L, as above). For plagioclase with Xab > 0·69 the melt plots to the left of the tie-line pargasite–plagioclase in the AFM-like diagram, as shown in Fig. 14b for plagioclase with Xab = 0·72. This type of topology implies a peritectic AFM reaction of the form Amp + Pl + Ep + Qtz = L + Grt. For plagioclase with Xab <0·69 the melt plots to the right of the tie-line pargasite–plagioclase and within the tie-triangle pargasite–garnet–plagioclase in the AFM-like diagram, implying an AFM melting reaction of the form Amp + Pl + Grt + Qtz = L + Ep. The corresponding mass-balance reactions calculated for Xab = 0·72, Xab = 0·69 and Xab = 0·66, respectively, are
|
|
Similar reactions would be deduced if trondhjemite CV228c does not represent a pristine primary melt. For other tonalitic–trondhjemitic primary melt compositions rich in plagioclase component it is possible to evaluate the melting relations fixing the composition of plagioclase. Our results indicate reactions with the same general forms as those deduced above (i.e. Pl + Qtz + Amp + Ep = L + Grt, Pl + Qtz + Amp + Ep = L, and Pl + Qtz + Amp + Grt = L + Ep). Therefore, melting of the studied garnet-bearing amphibolites should conform to one of these reactions.
The degenerate eutectic-like reaction Pl + Qtz + Amp + Ep = L involving no garnet is considered unlikely because there is no fundamental reason for a general collinear relation between plagioclase, amphibole and melt in garnet-bearing assemblages. The other two reactions are peritectic. We favour reaction Pl + Qtz + Amp + Ep = L + Grt (Fig. 14b) because garnet is idiomorphic in sample CV228e (Fig. 4a), a feature that conflicts with reactant garnet as predicted by reaction Amp + Pl + Grt = L + Ep. Furthermore, the latter reaction has epidote in the product, a result that is inconsistent with theory and experiments (e.g. Ellis & Thompson, 1986; Thompson & Ellis, 1994; Quirion & Jenkins, 1998).
A variety of melting reactions are possible for the garnet + clinopyroxene-bearing amphibolite sample CV237j. The number of possible reactions is larger in this case because the presence of quartz in the pre-melting assemblage of this quartz-lacking sample is uncertain. Garnet, clinopyroxene, amphibole and epidote may appear either as reactants or products in these reactions, depending on the composition of plagioclase and melt and the presence or absence of quartz. However, in all the possible reactions the amount of plagioclase involved is large, in agreement with the behaviour of this phase in other types of amphibolite studied.
Thus, it is concluded that the amount of plagioclase present in amphibolite prior to melting is a major control on the extent of wet melting. This inference applies to amphibolite with different mineral assemblages undergoing different melting reactions. Because of the relatively high pressure of melting (c. 15 kbar) the amount of plagioclase present in the pre-melting assemblages is inferred to have been small and, consequently, low melt fractions should have been produced. Low melt fractions are consistent with the Fe + Mg-poor nature of the studied trondhjemitic rocks. These inferences strongly suggest that plagioclase was completely consumed in the amphibolites upon melting and that the trondhjemites are (near-) primary melts.
Peraluminosity of primary slab melts
The studied rocks confirm that melts formed upon partial melting of metabasite at intermediate pressure under fluid-present conditions are peraluminous (see Ellis & Thompson, 1986; Thompson & Ellis, 1994) and that this type of melt can form in a subducting slab. However, peraluminosity is not a typical characteristic of magmas thought to have formed upon partial melting of subducted slabs (e.g. Cenozoic adakites, Archaean tonalite–trondhjemite complexes; Defant & Drummond, 1990; Martin, 1999). This observation should be taken into consideration when the origin of slab-derived magmas is addressed.
Experimental work on melting of natural and synthetic metabasites has shown that peraluminous melts form at low-temperature conditions close to the wet basaltic solidus, and metaluminous melts form under conditions that deviate from the wet basaltic solidus, typically at >850°C (Yoder & Tilley, 1962; Holloway & Burnham, 1972; Heltz, 1976; Ellis & Thompson, 1986; Beard & Lofgren, 1989; Rapp et al., 1991; Winther & Newton, 1991; Gaetani et al., 1993; Thompson & Ellis, 1994; Kawamoto, 1996; Springer & Seck, 1997; Nakajima & Arima, 1998; Prouteau et al., 2001; Selbekk & Skjerlie, 2002; Koepke et al., 2004). Under conditions close to the solidus, amphibole is abundant and plagioclase is scarce or not present in the residua, small melt fractions are formed, and the resulting melt is acid, in all aspects similar to characteristic amphibolite–trondhjemite associations of the Sierra del Convento mélange. Under conditions that deviate from the solidus, amphibole is scarce or not present, large melt fractions are formed, and the resulting melt is relatively basic (andesitic). It follows that metaluminous slab magmas, if formed by melting of mafic rocks close to the wet solidus, are not pristine slab melts. Metaluminosity of such magmas should be identified, instead, as a consequence of post-melting processes such as interaction with the mantle wedge and/or the base of the crust (see Kepezhinskas et al., 1995; Stern & Kilian, 1996; Martin, 1999; Prouteau et al., 2001; Yogodzinski et al., 2001).
Magmatic paragonite and kyanite
Magmatic paragonite is expected in peraluminous melts that crystallized at moderate to high pressure. As shown in Fig. 12d, magmatic paragonite is stable in the model NASH system above 8 kbar. Similar relations are predicted in the more complex NCASH system (García-Casco, 2007). Using a NCASH pseudosection approach, García-Casco (2007) calculated that paragonite is stable above the solidus of sample SC21 in the range 680–730°C at c. 14 kbar. This is consistent with the high K content of magmatic paragonite (Fig. 9), which indicates crystallization at high temperature (García-Casco, 2007). Crystallization of trondhjemitic melt at depth within the stability field of paragonite is also consistent with the occurrence of magmatic epidote in the studied rocks (Fig. 4i and k), a feature that is normally taken as an indication of moderate to high pressure of crystallization of magmatic rocks (Schmidt & Poli, 2004, and references therein). The model phase relations calculated by García-Casco (2007) show that magmatic epidote is not stable at <13 kbar in bulk composition SC21, in agreement with the calculated crystallization pressure of 14–16 kbar (Fig. 12d).
The presence of kyanite relicts within magmatic epidote in some samples (Fig. 4m) also points to crystallization at high pressure. The NCASH pseudosection for sample SC21 shows stable kyanite + epidote above the solidus at >720°C, 14–15 kbar and total consumption of kyanite upon reaction with melt to produce paragonite and more epidote during near-isobaric cooling (García-Casco, 2007). These relations explain the relict nature of kyanite within epidote in the studied rocks.
The consistency between natural mineral assemblages and theoretical phase relations allows the conclusion that paragonite is an expected product of crystallization of peraluminous trondhjemitic melt at intermediate to high pressure. Therefore, relatively low pressure (<8 kbar) of crystallization explains the lack of paragonite in natural peraluminous trondhjemites (e.g. Johnson et al., 1997).
Tectonic implications
Recent thermal models incorporating appropriate temperature dependence of mantle rheology suggest that melting of subducted oceanic crust may occur at shallower depths and be more general over a range of plate velocities and ages, subduction angles, and other geophysical variables (i.e. not restricted to special cases of subduction) than previously assumed based on the relatively cool slabs calculated in earlier models (e.g. van Keken et al., 2002; Gerya & Yuen, 2003; Conder, 2005; Abers et al., 2006). Following these recent results, melting in the Sierra del Convento could be interpreted as the result of normal subduction. Thermal models, however, also point towards special subduction environments where melting is possible at relatively shallow depths, including the edge of a slab (e.g. Kincaid & Griffiths, 2004), subhorizontal subduction (e.g. Manea et al., 2005), subduction retreat (e.g. Kincaid & Griffiths, 2003, 2004), onset of subduction (e.g. Gerya et al., 2002), or subduction of a very young slab or a ridge (e.g. Okudaira & Yoshitake, 2004; Uehara & Aoya, 2005). Though it is difficult to decipher the precise subduction environment of formation of the studied rocks, regional arguments favour a scenario of onset of subduction of young oceanic lithosphere.
Regional geological data for the Caribbean realm point to initiation of subduction during the Aptian (c. 120 Ma; Pindell et al., 2005, 2006). Pindell et al. have developed a tectonic model for the evolution of the Caribbean realm that incorporates the birth of a SW-dipping subduction system during the Aptian that consumed oceanic lithosphere of the Protocaribbean (Atlantic) basin, which was opening at that time, and, consequently, subduction of a hot, young oceanic lithosphere and/or a ridge was possible. Indirect evidence for the existence of a (near-orthogonal) subducting Protocaribbean ridge during the Cretaceous is given by geochemical studies of magmatic rocks from the overriding Caribbean plate (e.g. Jolly & Lidiak, 2006; Escuder-Viruete et al., 2007; see also discussion by Pindell et al., 2005, 2006). Thus, because age data indicate that the earliest stage of subduction in eastern Cuba is Aptian (c. 120 Ma; see García-Casco et al., 2006), the studied rocks probably formed in a scenario of onset of subduction of young lithosphere (Fig. 15).
|
After a thorough consideration of P–T–t data and thermal models, Krebs et al. (2007
) recently developed a similar scenario of onset of Aptian subduction of young (c. 20 Ma) Protocaribbean lithosphere for eclogite blocks of the Río San Juan mélange (Dominican Republic). The descriptions and interpretations of those workers suggest that the Río San Juan and the Sierra del Convento mélanges are geologically correlated. The Río San Juan mélange contains eclogites formed during the early Cretaceous (103·6 ± 2·7 Ma) and blueschist formed during the late Cretaceous. The eclogites attained similar peak temperature (c. 750°C) to the Sierra del Convento amphibolites, although the former attained substantially higher peak pressure (c. 23 kbar) and did not apparently undergo partial melting. These differences in the earliest products of subduction point to strike-parallel variations in the thermal structure (i.e. age) of the subducting Protocaribbean during the early Cretaceous. Plate-tectonic reconstructions locate eastern Cuba to the NNW of Hispaniola by Aptian times (Pindell et al., 2005), implying that the subducting slab was hotter (i.e. younger) to the NNW (relative to Hispaniola) and, consequently, that the ridge should have been located close to eastern Cuba by Aptian times.
This scenario for the onset of subduction explains the observed counterclockwise P–T paths of the Sierra del Convento amphibolite blocks. Onset of subduction produces a transient geothermal gradient at the slab–mantle interface, which cools upon continued subduction (Gerya et al., 2002). Our calculated P–T paths suggest that, after accretion to the overriding plate, the blocks of amphibolite and associated trondhjemites were refrigerated at depth, in agreement with the expected effects of continued subduction (Fig. 12). At this stage in the Sierra del Convento mélange the trondhjemite magmas formed by partial melting of amphibolite, cooled and crystallized. The subduction models of Gerya et al. (2002) show that continued subduction leads to the formation of a serpentinitic layer (subduction channel) by hydration of the overriding mantle some Ma after the onset of subduction. Gerya et al. also indicated that formation of such a serpentinite layer allows formation of serpentinite mélanges and the onset of exhumation of early accreted blocks along the subduction channel. In the Sierra del Convento mélange this stage is recorded by the blueschist-facies overprints developed as the amphibolite and trondhjemite blocks followed counterclockwise P–T paths (Fig. 15).
As subduction proceeds, more material normally of lower grade can be accreted to the subduction channel (Gerya et al., 2002). Blocks of blueschist not studied here record this stage in the Sierra del Convento. A similar picture evolves for the Río San Juan mélange (Krebs et al., 2007). Thus, both mélanges constitute fragments of a subduction channel documenting a long-lasting history of subduction, accretion, mélange formation, and uplift, and both give direct evidence for Aptian onset of subduction of the Protocaribbean (Fig. 15).
|
CONCLUSIONS |
|---|
The Sierra del Convento mélange (eastern Cuba) contains high-grade tectonic blocks of plagioclase-lacking epidote ± garnet amphibolite and associated tonalite–trondhjemite, which formed during subduction of oceanic lithosphere. Field relations, major element bulk-rock compositions, textures, peak metamorphic, magmatic and retrograde mineral assemblages, and P–T conditions, and the agreement between the observed mineral assemblages and those predicted by experimental and theoretical studies indicate that tonalite–trondhjemite formed by partial melting of amphibolite in the subduction environment and that both shared a common P–T history during exhumation. Thus, these tectonic blocks represent a rare example of oceanic subduction-related migmatites that have returned to the Earth's surface. Regional arguments suggest a scenario of onset of subduction of young oceanic lithosphere (during the Aptian). Partial melting at c. 15 kbar, 750°C was characterized by low melt fractions, was fluid-assisted, occurred close to the metabasite solidus, consumed large relative amounts of plagioclase in the amphibolites, and formed plagioclase-lacking residual amphibolite and peraluminous, K-poor, high Mg-number leucocratic tonalitic–trondhjemitic melts that segregated into veins, layers and agmatitic structures. Shortly after melt formation amphibolite + melt blocks were accreted to the overriding plate, where they began to cool as subduction proceeded, allowing the partial melts to crystallize at depth. This favoured the formation of magmatic paragonite, kyanite, epidote, plagioclase, pargasite, and quartz during crystallization of the melts. Later, syn-subduction exhumation in the mélange caused counterclockwise P–T paths, forming retrograde blueschist-facies assemblages in all types of rock. Blueschist-facies conditions prevailed during the ensuing history of subduction, when other blocks were incorporated into the mélange, attesting to a long-lasting history of accretion, mélange formation and uplift.
|
ACKNOWLEDGEMENTS |
|---|
The authors thank Walter Maresch, an anonymous reviewer, and editor Ron Frost for their perceptive comments and suggestions, which substantially improved this paper. Walter Maresch is also thanked for providing a preprint manuscript by Krebs and co-workers on the Río San Juan mélange. Manuel Iturralde-Vinent kindly reviewed an early version. This is a contribution to IGCP-546 ‘Subduction zones of the Caribbean’ and is Mainz Geocycles contribution 318. We appreciate financial support from MEC project CGL2006-08527/BTE.
*Corresponding author. Telephone: +34 958 246613. Fax: +34 958 243368. E-mail: agcasco{at}ugr.es
|
REFERENCES |
|---|
Abers GA, van Keken PE, Kneller EA, Ferris A, Stachnik JC. The thermal structure of subduction zones constrained by seismic imaging: Implications for slab dehydration and wedge flow. Earth and Planetary Science Letters (2006) 241:387–397.[CrossRef][Web of Science]
Barker F. Trondhjemites: Definition, environment and hypotheses of origin. In: Trondhjemites, Dacites and Related Rocks—Barker F, ed. (1979) Amsterdam: Elsevier. 1–12.
Beard JS, Lofgren GE. Effect of water on the composition of melts of greenstone and amphibolites. Science (1989) 244:195–197.
Beard JS, Lofgren GE. Dehydration melting and water-saturated melting of basaltic and andesitic greenstones and amphibolites at 1, 3, and 6·9 kb. Journal of Petrology (1991) 32:365–401.
Bebout GE. Metamorphic chemical geodynamics of subduction zones. Earth and Planetary Science Letters (2007) 260:373–393.[CrossRef][Web of Science]
Bebout GE, Barton MD. Tectonic and metasomatic mixing in a high-T, subduction-zone mélange—insights into the geochemical evolution of the slab–mantle interface. Chemical Geology (2002) 187:79–106.[CrossRef][Web of Science]
Boiteau A, Michard A, Saliot P. Métamporphisme de haute pression dans le complexe ophiolitique du Purial (Oriente, Cuba). Comptes Rendus de l’Académie des Sciences, Série D (1972) 274:2137–2140.
Cobiella J, Quintas F, Campos M. Geología de la Región Central y Suroriental de la Provincia de Guantánamo. Santiago de Cuba (1984) 125 pp: Santiago de Cuba: Editorial Oriente.
Coleman RG, Lee DE, Beatty LB, Brannock WW. Eclogites and eclogites: their differences and similarities. Geological Society of America Bulletin (1965) 76:483–508.
Conder JA. A case for hot slab surface temperatures in numerical viscous flow models of subduction zones with an improved fault zone parameterization. Physics of the Earth and Planetary Interiors (2005) 149:155–164.[CrossRef][Web of Science]
Coutinho JMV, Kräutner HG, Sassi F, Schmid R. A systematic nomenclature for metamorphic rocks: 8. Amphibolite and granulite. In: Recommendations by the IUGS Subcommission on the Systematics of Metamorphic Rocks. Web version of 01.02.2007 (2007) (http://www.bgs.ac.uk/SCMR/docs/papers/paper_8.pdf).
De Andrade V, Vidal O, Lewin E, O’Brien P, Aagard P. Quantification of electron microprobe compositional maps of rock thin sections: an optimized method and examples. Journal of Metamorphic Geology (2006) 24:655–668.[Web of Science]
Defant MJ, Drummond MS. Derivation of some modern arc magmas by melting of young subducted lithosphere. Nature (1990) 347:662–665.[CrossRef]
Drummond MS, Defant MJ, Kepezhinskas PK. Petrogenesis of slab-derived trondhjemite–tonalite–dacite/adakite magmas. Transactions of the Royal Society of Edinburgh: Earth Sciences (1996) 87:205–215.[Web of Science]
Elliott T, Plank T, Zindler A, White W, Bourdon B. Element transport from slab to volcanic front at the Mariana arc. Journal of Geophysical Research (1997) 102(B7):14991–15019.[CrossRef]
Ellis DJ, Thompson AB. Subsolidus and partial melting reactions in the quartz-excess CaO + MgO + Al2O3 + SiO2 + H2O system under water-excess and water-deficient conditions to 10 kb: Some implications for the origin of peraluminous melts from mafic rocks. Journal of Petrology (1986) 27:91–121.
Ernst WG. Metamorphism, partial preservation, and exhumation of ultrahigh-pressure belts. The Island Arc (1999) 8:125–153.[CrossRef]
Escuder-Viruete J, Pérez-Estaún A. Subduction-related P–T path for eclogites and garnet glaucophanites from the Samaná Peninsula basement complex, northern Hispaniola. International Journal of Earth Sciences (2006) 95:995–1017.[CrossRef][Web of Science]
Escuder-Viruete J, Contreras F, Stein G, Urien P, Joubert M, Pérez-Estaún A, Friedman R. Magmatic relationships and ages between adakites, magnesian andesites and Nb-enriched basalt–andesites from Hispaniola: Record of a major change in the Caribbean island arc magma sources. Lithos (2007) 99:151–177.[CrossRef][Web of Science]
Gaetani GA, Grove TL, Bryan WB. The influence of water on the petrogenesis of subduction-related igneous rocks. Nature (1993) 365:332–335.[CrossRef]
García-Casco A. Magmatic paragonite in trondhjemites from the Sierra del Convento mélange, Cuba. American Mineralogist (2007) 92:1232–1237.
García-Casco A, Torres-Roldán RL, Millán G, Monié P, Schneider J. Oscillatory zoning in eclogitic garnet and amphibole, Northern Serpentinite Mélange, Cuba: a record of tectonic instability during subduction? Journal of Metamorphic Geology (2002) 20:581–598.[CrossRef][Web of Science]
García-Casco A, Torres-Roldán RL, Iturralde-Vinent MA, Millán G, Núñez Cambra K, Lázaro C, Rodríguez Vega A. High pressure metamorphism of ophiolites in Cuba. Geologica Acta (2006) 4:63–88.
Gerya TV, Yuen DA. Rayleigh–Taylor instabilities from hydration and melting propel ‘cold plumes’ at subduction zones. Earth and Planetary Science Letters (2003) 212:47–62.[CrossRef][Web of Science]
Gerya TV, Stoeckhert B, Perchuk AL. Exhumation of high-pressure metamorphic rocks in a subduction channel—a numerical simulation. Tectonics (2002) 21:6-1–6-19.
Green TH. Anatexis of mafic crust and high pressure crystallization of andesite. In: Andesites—Thorpe RS, ed. (1982) Chichester: John Wiley. 465–487.
Grove TL, Chatterjee N, Parman SW, Médard E. The influence of H2O on mantle wedge melting. Earth and Planetary Science Letters (2006) 249:74–89.[CrossRef][Web of Science]
Harlow GE, Hemming SR, Avé Lallemant HG, Sisson VB, Sorensen SS. Two high-pressure–low-temperature serpentinite-matrix mélange belts, Motagua fault zone, Guatemala: A record of Aptian and Maastrichtian collisions. Geology (2004) 32:17–20.
Heltz RT. Phase relations of basalt in their melting ranges at PH2O = 5 kb as a function of oxygen fugacity. Part I. Mafic phases. Journal of Petrology (1973) 14:249–302.
Heltz RT. Phase relations of basalt in their melting ranges at PH2O = 5 kb. Part II. Melt compositions. Journal of Petrology (1976) 17:139–193.
Hermann J, Spandler C, Hack A, Korsakov AV. Aqueous fluids and hydrous melts in high-pressure and ultra-high pressure rocks: Implications for element transfer in subduction zones. Lithos (2006) 92:399–417.[CrossRef][Web of Science]
Hernández M, Canedo Z. Geoquímica de las ofiolitas meridionales de Cuba oriental. Revista Minería y Geología (1995) 3:3–9.
Hofmann AW. Chemical differentiation of the Earth: the relationship between mantle, continental crust, and oceanic crust. Earth and Planetary Science Letters (1988) 90:297–314.[CrossRef][Web of Science]
Holland TJB, Powell R. An internally consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology (1998) 16:309–343.[CrossRef][Web of Science]
Holloway JR, Burnham CW. Melting relations of basalt with equilibrium water pressure less than total pressure. Journal of Petrology (1972) 13:1–29.
Iturralde-Vinent MA. Introduction to Cuban Geology and Geophysics. Ofiolitas y Arcos Volcánicos de Cuba. IGCP Project 364 Special Contribution—Iturralde-Vinent MA, ed. (1996) 1:3–35.
Iturralde-Vinent MA. Sinopsis de la constitución Geológica de Cuba. Acta Geologica Hispanica (1998) 33:9–56.
Iturralde-Vinent MA, Lidiak EG. Caribbean tectonic, magmatic, metamorphic and stratigraphic events. Implications for plate tectonics. Geologica Acta (2006) 4:1–5.[Medline]
Iturralde-Vinent MA, Millán G, Korkas L, Nagy E. Geological interpretation of the Cuban K–Ar data base. Ofiolitas y Arcos Volcánicos de Cuba. IGCP Project 364 Special Contribution—Iturralde-Vinent MA, ed. (1996) 1:48–69.
Iturralde-Vinent MA, Díaz Otero C, Rodríguez Vega A, Díaz Martínez R. Tectonic implications of paleontologic dating of Cretaceous–Danian sections of Eastern Cuba. Geologica Acta (2006) 4:89–102.
Johnson K, Barnes CG, Miller CA. Petrology, geochemistry, and genesis of high-Al tonalite and trondhjemites of the Cornucopia Stock, Blue Mountains, northeastern Oregon. Journal of Petrology (1997) 38:1585–1611.[CrossRef][Web of Science]
Johnson MC, Plank T. Dehydration and melting experiments constrain the fate of subducted sediments. Geochemistry, Geophysics, Geosystems (1999) 1. paper number 1999GC000014.
Jolly WT, Lidiak EG. Role of crustal melting in petrogenesis of the Cretaceous Water Island Formation (Virgin Islands, northeast Antilles Island arc). Geologica Acta (2006) 4:7–33.
Kawamoto T. Experimental constraints on differentiation and H2O abundance of calc-alkaline magmas. Earth and Planetary Science Letters (1996) 144:577–589.[CrossRef][Web of Science]
Kelemen PB, Hanghoj K, & Greene A. One view of the geochemistry of subduction-related magmatic arcs with an emphasis on primitive andesite and lower crust. In: The Crust. [Treatise on Geochemistry, Volume 3 (eds H. D. Holland & K. K. Turekian).]—Rudnick RL, ed. (2003) Oxford: Elsevier–Pergamon. 593–659.
Kepezhinskas PK, Defant MJ, Drummond MS. Na metasomatism in the island-arc mantle by slab melt–peridotite interaction: evidence from mantle xenoliths in the north Kamchatka arc. Journal of Petrology (1995) 36:1505–1527.
Kessel R, Ulmer P, Pettke T, Schmidt MW, Thompson AB. The water–basalt system at 4 to 6 GPa: Phase relations and second critical endpoint in a K-free eclogite at 700 to 1400°C. Earth and Planetary Science Letters (2005) 237:873–892.[CrossRef][Web of Science]
Kincaid C, Griffiths RW. Laboratory models of the thermal evolution of the mantle during rollback subduction. Nature (2003) 425:58–62.[CrossRef]
Kincaid C, Griffiths RW. Variability in flow and temperatures within mantle subduction zones. Geochemistry, Geophysics, Geosystems (2004) 5. Q06002, doi:10.1029/2003GC000666.
Koepke J, Feig ST, Snow J, Freise M. Petrogenesis of oceanic plagiogranites by partial melting of gabbros: an experimental study. Contributions to Mineralogy and Petrology (2004) 146:414–432.[CrossRef][Web of Science]
Krebs M, Maresch WV, Schertl H-P, Baumann A, Draper G, Idleman B, Münker C. The dynamics of intra-oceanic subduction zones: A direct comparison between fossil petrological evidence (Rio San Juan Complex, Dominican Republic) and numerical simulation. In: Lithos (2007) (in press).
Kretz R. Symbols for rock-forming minerals. American Mineralogist (1983) 68:277–279.[Abstract]
Kriegsman LM. Partial melting, partial melt extraction and partial back reaction in anatectic migmatites. Lithos (2001) 56:75–96.[CrossRef][Web of Science]
Kulachkov LV, Leyva RC. Informe sobre los resultados de los trabajos de reconocimiento geológico para cuarzo filoniano en la parte oriental de Cuba. (1990) Unpublished report of the Instituto Superior Minero-Metalúrgico de Moa (Cuba).
Lambert IB, Wyllie PJ. Melting of gabbro (quartz-eclogite) with excess water to 35 kilobars, with geological applications. Journal of Geology (1972) 80:693–708.[Web of Science]
Leake BE, Woolley AR, Arps CES, et al. Nomenclature of amphiboles: Report of the Subcommittee on Amphiboles of the International Mineralogical Association, Commission on New Minerals and Mineral Names. American Mineralogist (1997) 82:1019–1037.[Abstract]
Le Maitre RW, Streckeisen A, Zanettin B, et al, eds. Igneous Rocks: A Classification and Glossary of Terms (2002) 2nd. Cambridge: Cambridge University Press. 236.
Leyva RC. Características geológicas, regularidades de distribución y perspectivas de utilización del cuarzo filoniano de la región oriental de Cuba. In: Unpublished report of the Instituto Superior Minero-Metalúrgico de Moa (Cuba) (1996) 90.
Manea VC, Manea M, Kostoglodov V, Sewell G. Thermo-mechanical model of the mantle wedge in Central Mexican subduction zone and a blob tracing approach for the magma transport. Physics of the Earth and Planetary Interiors (2005) 149:165–186.[CrossRef][Web of Science]
Marchesi C, Garrido CJ, Bosch D, Proenza JA, Gervilla F, Monié P, Rodríguez-Vega A. Geochemistry of Cretaceous magmatism in eastern Cuba: recycling of North American continental sediments and implications for subduction polarity in the Greater Antilles Paleo-arc. Journal of Petrology (2007) 48:1813–1840.
Martin H. Adakitic magmas: modern analogues of Archaean granitoids. Lithos (1999) 46:411–429.[CrossRef][Web of Science]
Millán G. Metamorfitas de la asociación ofiolítica de Cuba. In: Ofiolitas y Arcos Volcánicos de Cuba. IGCP Project 364 Special Contribution—Iturralde-Vinent MA, ed. (1996) 1:131–146.
Millán G, Somin ML. Contribución al conocimiento geológico de las metamorfitas del Escambray y Purial. Reportes de Investigación de la Academia de Ciencias de Cuba (1985) 2:1–74.
Millán G, Somin ML, Díaz C. Nuevos datos sobre la geología del macizo montañoso de la Sierra del Purial, Cuba Oriental. Reporte de Investigación del Instituto de Geología y Paleontología (1985) 2:52–74.
Morimoto N, Fabries J, Ferguson AK, Ginzburg IV, Ross M, Seifert FA, Zussman J, Aoki K, Gottardi G. Nomenclature of pyroxenes: Report of the Subcommittee on Pyroxenes of the International Mineralogical Association, Commission on New Minerals and Mineral Names. American Mineralogist (1988) 73:1123–1133.[Abstract]
Nakajima K, Arima M. Melting experiments on hydrous low-K tholeiite: Implications for the genesis of tonalitic crust in the Izu–Bonin–Mariana arc. The Island Arc (1998) 7:359–373.
O’Connor JT. A classification of quartz-rich igneous rocks based on feldspar ratios. US Geological Survey Professional Papers (1965) 525B:B79–B84.
Okudaira T, Yoshitake Y. Thermal consequences of the formation of a slab window beneath the Mid-Cretaceous southwest Japan arc: A 2-D numerical analysis. The Island Arc (2004) 13:520–532.[CrossRef]
Peacock SM. Thermal structure and metamorphic evolution of subducting slabs. In: Eiler, J. (ed.). Inside the Subduction Factory. Geophysical Monograph, American Geophysical Union (2003) 138:7–22.
Peacock SM, Rushmer T, Thompson AB. Partial melting of subducting oceanic crust. Earth and Planetary Science Letters (1994) 121:227–244.[CrossRef][Web of Science]
Pindell JL, Maresch WV, Stanek KP, Draper G, Higgs R. Plate-kinematics and crustal dynamics of circum-Caribbean arc–continent interactions: Tectonic controls on basin development in Proto-Caribbean margins. Caribbean–South American Plate Interactions, Venezuela. Geological Society of America, Special Papers—Avé Lallemant HG, Sisson VB, eds. (2005) 394:7–52.[CrossRef]
Pindell JL, Kennan L, Stanek KP, Maresch WV, Draper G. Foundations of Gulf of Mexico and Caribbean evolution: eight controversies resolved. Geologica Acta (2006) 4:303–341.
Powell R, Holland TJB. Optimal geothermometry and geobarometry. American Mineralogist (1994) 79:120–133.[Abstract]
Proenza JA, Díaz-Martínez R, Iriondo A, Marchesi C, Melgarejo JC, Gervilla F, Garrido CJ, Rodríguez-Vega A, Lozano-Santacruz R, Blanco-Moreno JA. Primitive Cretaceous island-arc volcanic rocks in eastern Cuba: the Téneme Formation. Geologica Acta (2006) 4:103–121.
Prouteau G, Scaillet B, Pichavant M, Maury RC. Evidence for mantle metasomatism by hydrous silicic melts derived from subducted oceanic crust. Nature (2001) 410:197–200.[CrossRef]
Quirion DM, Jenkins DM. Dehydration and partial melting of tremolitic amphibole coexisting with zoisite, quartz, anorthite, diopside and water in the system H2O–CaO–MgO–Al2O3–SiO2. Contributions to Mineralogy and Petrology (1998) 130:379–389.[CrossRef][Web of Science]
Rapp RP, Watson EB, Miller CF. Partial melting of amphibolite/eclogite and the origin of Archean trondhjemites and tonalites. Precambrian Research (1991) 51:1–25.[CrossRef][Web of Science]
Rojas-Agramonte Y, Neubauer F, Handler R, García-Delgado DE, Friedl G, Delgado-Damas R. Variation of paleostress patterns along the Oriente Transform Fault, Cuba: Significance for Neogene–Quaternary tectonics of the Caribbean realm. Tectonophysics (2005) 396:161–180.[CrossRef][Web of Science]
Schmidt MW, Poli S. Experimentally based water budgets for dehydrating slabs and consequences for arc magma generation. Earth and Planetary Science Letters (1998) 163:361–379.[CrossRef][Web of Science]
Schmidt MW, Poli S. Magmatic epidote. In: Epidotes. Mineralogical Society of America, Reviews in Mineralogy and Geochemistry—Liebscher A, Franz G, eds. (2004) 56:399–430.
Selbekk RS, Skjerlie KP. Petrogenesis of the anorthosite dyke swarm of Tromsø, North Norway: Experimental evidence for hydrous anatexis of an alkaline mafic complex. Journal of Petrology (2002) 43:943–962.
Smithies RH. The Archaean tonalite–trondhjemite–granodiorite (TTG) series is not an analogue of Cenozoic adakite. Earth and Planetary Science Letters (2000) 182:115–125.[CrossRef][Web of Science]
Somin ML, Millán G. Geology of the Metamorphic Complexes of Cuba (1981) Nauka: Moscow. 219. p. (in Russian).
Somin ML, Arakelyants MM, Kolesnikov EM. Age and tectonic significance of high-pressure metamorphic rocks in Cuba. International Geology Review (1992) 34:105–118.
Sorensen SS. Petrology of amphibolite-facies mafic and ultramafic rocks from the Catalina Schist, Southern California: metasomatism and migmatization in a subduction zone metamorphic setting. Journal of Metamorphic Geology (1988) 6:405–435.[Web of Science]
Sorensen SS, Barton MD. Metasomatism and partial melting in a subduction complex: Catalina Schist, southern California. Geology (1987) 15:115–118.
Sorensen SS, Grossman JN. Enrichment in trace elements in garnet amphibolites from a paleo-subduction zone: Catalina Schist, southern California. Geochimica et Cosmochimica Acta (1989) 53:3155–3177.[CrossRef][Web of Science]
Spear FS. Metamorphic Phase Equilibria and Pressure–Temperature–Time Paths. Mineralogical Society of America, Monograph (1993) 1:799.
Springer W, Seck HA. Partial fusion of basic granulites at 5 to 15 kbar: implications for the origin of TTG magmas. Contributions to Mineralogy and Petrology (1997) 127:30–45.[CrossRef][Web of Science]
Staudigel H, Plank T, White B, & Schmincke H. Geochemical fluxes during seafloor alteration of the upper oceanic crust: DSDP sites 417–418. Subduction Top to Bottom. Geophysical Monograph, American Geophysical Union—Bebout GE, Scholl DW, Kirby SH, Platt JP, eds. (1996) 96:19–38.
Stern CR, Kilian R. Role of the subducted slab, mantle wedge and continental crust in the generation of adakites from the Andean Austral Volcanic Zone. Contributions to Mineralogy and Petrology (1996) 123:263–281.[CrossRef][Web of Science]
Stern RJ. Subduction zones. Reviews of Geophysics (2002) 40. article number 1012, doi:10.1029/2001RG000108.
Stolz AJ, Jochum KP, Spettel B, Hofman AW. Fluid- and melt-related enrichment in the subarc mantle: evidence from Nb/Ta variations in island-arc basalts. Geology (1996) 24:587–590.
Thompson AB, Ellis D. CaO + MgO + Al2O3 + SiO2 + H2O to 35 kb: Amphibole, talc, and zoisite dehydration and melting reactions in the silica-excess part of the system and their possible significance in subduction zones, amphibole melting, and magma fractionation. American Journal of Science (1994) 294:1229–1289.
Torres-Roldán RL, García-Casco A, García-Sánchez PA. CSpace: An integrated workplace for the graphical and algebraic analysis of phase assemblages on 32-bit Wintel platforms. Computers and Geosciences (2000) 26:779–793.[CrossRef]
Uehara S, Aoya M. Thermal model for approach of a spreading ridge to subduction zones and its implications for high-P/high-T metamorphism: Importance of subduction versus ridge approach ratio. Tectonics (2005) 24. TC4007, doi:10.1029/2004TC001715.
van Keken PE, Kiefer B, Peacock SM. High-resolution models of subduction zones: Implications for mineral dehydration reactions and the transport of water into the deep mantle. Geochemistry, Geophysics, Geosystems (2002) 3. 1056, doi:10.1029/2001GC000256.
Vielzeuf D, Schmidt MW. Melting reactions in hydrous systems revisited: applications to metapelites, metagreywackes and metabasalts. Contributions to Mineralogy and Petrology (2001) 141:251–267.[Web of Science]
Winther KT, Newton RC. Experimental melting of hydrous low-K tholeiite: evidence on the origin of Archaean cratons. Bulletin of the Geological Society of Denmark (1991) 39:213–228.[Web of Science]
Yoder HS Jr, Tilley CE. Origin of basalt magmas; an experimental study of natural and synthetic rock systems. Journal of Petrology (1962) 3:342–532.
Yogodzinski GM, Lees JM, Churikova TG, Dorendorf F, Wörner G, Volynets ON. Geochemical evidence for the melting of subducting oceanic lithosphere at plate edges. Nature (2001) 409:500–504.[CrossRef][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  B. E. Leake Hawthorne, F.C., Oberti, R., Della Ventura, G. and Mottana, A. (editors). Amphiboles: Crystal Chemistry, Occurrence and Health Issues.: Mineralogical Society of America, Geochemical Society and Accademia Nazionale dei Lincei, Chantilly, VA, USA. Reviews in Mineralogy and Geochemistry, 67, 2007. Price US$ 45.00, 545 pp., ISBN 978-0-939950-79-9. Mineralogical Magazine, October 1, 2007; 71(5): 591 - 592. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||