Journal of Petrology Advance Access originally published online on July 29, 2004
Journal of Petrology 2004 45(9):1877-1906; doi:10.1093/petrology/egh037
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology 45(9) © Oxford University Press 2004; all rights reserved
Mafic Pegmatites Intruding Oceanic Plateau Gabbros and Ultramafic Cumulates from Bolívar, Colombia: Evidence for a ‘Wet’ Mantle Plume?
1 DEPARTMENT OF EARTH SCIENCES, CARDIFF UNIVERSITY, PARK PLACE, CARDIFF CF10 3YE, UK
2 DEPARTMENT OF GEOLOGY, UNIVERSITY OF LEICESTER, UNIVERSITY ROAD, LEICESTER LE1 7RH, UK
3 NERC ISOTOPE GEOSCIENCES LABORATORY, c/o BRITISH GEOLOGICAL SURVEY, KEYWORTH, NOTTINGHAM NG12 5GG, UK
4 SCOTTISH UNIVERSITIES ENVIRONMENTAL RESEARCH CENTRE, EAST KILBRIDE, GLASGOW G75 0QF, UK
5 INGEOMINAS—REGIONAL PACIFICO, AA 9724, CALI, COLOMBIA
RECEIVED JULY 22, 2003; ACCEPTED APRIL 6, 2004
 |
ABSTRACT |
|---|
The fault-bounded Bolívar Ultramafic Complex (BUC) on the eastern fringes of the Western Cordillera of Colombia was tectonically accreted onto the western coast of South America in the late Cretaceous–early Tertiary, along with pillow basalts of the Caribbean–Colombian Oceanic Plateau (CCOP). The complex consists of a lower sequence of ultramafic cumulates, successively overlain by layered and isotropic gabbroic rocks. The gabbros grade into, and are intruded by, mafic pegmatites that consist of large magnesiohornblende and plagioclase crystals. These pegmatites yield a weighted mean 40Ar–39Ar step-heating age of 90·5 ± 0·9 Ma and thus coincide with the timing of peak CCOP volcanism. The chemistry of the BUC is not consistent with a subduction-related origin. However, the similarity in Sr–Nd–Pb–Hf isotopes between the CCOP and the BUC, in conjunction with their indistinguishable ages, suggests that the BUC is an integral part of the plume-derived CCOP. The parental magmas of the Bolívar complex were probably hydrous picrites that underwent 20–30% crystallization. The residual magmas from this fractionation contained
7·0 wt % MgO and enough water to permit fractionation of magnesiohornblende (and bytownite) on the sidewalls of veins, forming hornblende-rich pegmatites. As the residual magmas migrated upwards through the complex, they became more evolved and, thus, fractionated less magnesiohornblende along with a greater proportion of more sodic plagioclase. At the highest preserved levels of the complex the pegmatites consist predominantly of andesine plagioclase with quartz and biotite. During crystallization of the pegmatites some of the residual magma was tapped or erupted and intruded at higher levels to form breccias and dykes that are exposed nearby at Vijes. Our calculations suggest that the BUC resulted from melting of a mantle plume source region that contained at least 400 ppm H2O. We propose that mantle plumes, as well as being heterogeneous in terms of radiogenic isotopes and trace elements, are also heterogeneous with regard to their water content, and so can have portions that are more hydrous than others. This has fundamental implications for mantle rheology and the environmental impact of Large Igneous Province eruptions. KEY WORDS: mantle plume; Caribbean; Colombia; wet melting; oceanic plateau; pegmatite
|
INTRODUCTION |
|---|
The accretion of the Cretaceous Caribbean–Colombian Oceanic Plateau (CCOP) around the margins of the Caribbean and along the NW margin of South America has resulted in the CCOP being arguably one of the best exposed and understood oceanic plateaux (Donnelly et al., 1990
; Kerr et al., 1997b; Sinton et al., 1998). Detailed studies of these accreted sections have shed new light on the mantle source region, formation and structure of the CCOP in particular, and oceanic plateaux in general (e.g. Kerr et al., 1998; Walker et al., 1999; Hauff et al., 2000a).
The mid-Cretaceous period (85–125 Ma) is marked by a major phase of anomalous, mantle plume-related volcanism in the ocean basins (Larson, 1991; Eldholm & Coffin, 2000). The largest of these events are represented by the Ontong–Java and Manihiki plateaux (122 Ma and possibly 90 Ma; Mahoney et al., 1993), the southern Kerguelen Plateau (110–118 Ma; Coffin et al., 2002), the central Kerguelen Plateau–Broken Ridge (95–100 Ma; Coffin et al., 2002) and the Caribbean–Colombian Oceanic Plateau (88–93 Ma; Sinton et al., 1998). These oceanic plateaux have anomalously thick (>10 km) crust and cover areas in excess of 1·0 x 106 km2 (Coffin & Eldholm, 1994); their eruption in submarine conditions may well have contributed to anoxia and mass extinction events in the Cretaceous oceans (Vogt, 1989; Sinton & Duncan, 1997; Kerr, 1998).
It is now generally recognized that the CCOP formed in the Pacific at c. 90 Ma, on the eastward-moving Farallon plate, and possibly represents melting during the start-up phase of the Galapagos hotspot (Duncan & Hargraves, 1984; Richards et al., 1991; Thompson et al., 2004). Shortly after its formation, the northern part of the plateau appears to have collided with the eastward-dipping subduction zone of the Caribbean ‘Great Arc’ (Burke, 1988). The hot, thick and consequently buoyant plateau was relatively unsubductable and so clogged the subduction zone, resulting in a reversal in subduction polarity from east to west. This reversal allowed the northern portion of the CCOP to move into the seaway between North and South America along strike-slip faults to its present-day location (Burke, 1988).
The CCOP was drilled during Deep Sea Drilling Project (DSDP) Leg 15 in the Caribbean Sea (Bence et al., 1975), and portions of it are exposed around its margins as a result of tectonic processes (Kerr et al., 1997b; Sinton et al., 1998). In particular, the southern portion of the plateau collided with a continental subduction zone and, unable to subduct, was partially accreted in a series of tectonic slices along the northwestern margin of South America (Millward et al., 1984; Kerr et al., 1997a; Hauff et al., 2000b).
Although the majority of the accreted sections of the CCOP in northwestern South America consist of basaltic lavas and intrusive sheets (Kerr et al., 1997a), picrites and/or komatiites (sensu Kerr & Arndt, 2001) also occur; for example, Gorgona and in the Central Cordillera (Arndt et al., 1997; Kerr et al., 2002b). In addition to the extrusive and hypabyssal components, accreted gabbroic and ultramafic plutonic rocks interpreted as part of the plateau are also exposed in Colombia and Ecuador (Kerr et al., 1998, 2002a; Lapierre et al., 2000; Révillon et al., 2000).
One of the largest of these plutonic complexes is located west of the town of Bolívar (4°20'N, 76°15'W; Fig. 1); this complex predominantly comprises layered and isotropic gabbros and gabbronorites, underlain by more ultramafic rocks. The plutonic suite is known collectively as the Bolívar Ultramafic Complex (BUC) (Aspden, 1984). Intruding the BUC, and in places gradational to the gabbros, is a suite of predominantly mafic, hornblende-bearing pegmatite dykes, which are the main focus of this study. A limited amount of chemical data for gabbros from the BUC was reported by Kerr et al. (1998). In this paper we present a more extensive dataset for the gabbros and discuss 40Ar–39Ar ages and elemental and Sr–Nd–Pb–Hf isotope data for the pegmatites. We use these data to constrain models for the origin of the Bolívar pegmatites and their petrogenetic relationship to nearby felsic dykes and breccias. Our conclusions on the origin of the hydrous rocks in the Bolívar complex enable us to contribute to the important debate on the water content of mantle plumes and the associated implications for both mantle and magmatic processes.
|
|
ANALYTICAL METHODS |
|---|
Major element mineral compositions were determined using a JEOL 8600 electron microprobe at the University of Leicester. A 30 nA beam current, 15 kV accelerating voltage and 10 µm spot size were used for all analyses. A ZAF correction procedure was used to reduce the raw data. The major element data are reported in Tables 1 and 2 and summarized in Fig. 2, which also shows 2
error bars for the analyses. The trace element abundances in magnesiohornblende and plagioclase were analysed on mineral separates by solution inductively coupled plasma mass spectrometry (ICP-MS) at Cardiff (see below) after hand-picking of crystals (Table 3).
|
|
|
|
40Ar–39Ar dating was carried out at the Scottish Universities Environmental Research Centre, East Kilbride. Hand-picked magnesiohornblende grains were wrapped in copper foil packets and placed in quartz vials. Samples were irradiated for 18 h at 1 MW in the Cd-shielded CLICIT facility at the Oregon State University (OSU) TRIGA reactor. The flux monitor Taylor Creek Rhyolite sanidine 85G003 (28·34 ± 0·28 Ma, Renne et al., 1998
) was loaded into copper packets and placed every 7 mm in each of the vials; the neutron flux monitor J is conservatively known to better than 0·3% at any given position in the vials. On the basis of previous analyses of optical grade CaF2 and degassed, Fe-doped K-silicate glass irradiated in the CLICIT facility at OSU, corrections for interfering neutron-induced reactions on 40K and 40Ca are [40Ar/39Ar]K = 0·00086, [36Ar/37Ar]Ca = 0·000264 and [39Ar/37Ar]Ca = 0·000673.
Individual rock cores were loaded into a glass side-arm over a double vacuum furnace attached to a gas clean-up system. Incremental-heating experiments consisted of between five and 17 steps. Following a 15 min heating step and 10 min additional clean-up stage, the purified gas was analysed on an EMI secondary electron multiplier operated at a gain of 5000, with a MAP 215 rare gas mass spectrometer. For each analysis, the peaks over the mass range 36–40 were measured over nine cycles. Blanks were analysed at least every six steps, usually at the end of each experiment. Analyses were then corrected for blanks, which typically were 3·0 x 10–17 moles for 36Ar, 2 x 10–15 moles for 39Ar and 2·4 x 10–15 moles for 40Ar. Monitor samples were analysed and their J curves calculated from their positions within the vial. These curves were interpolated and J values were obtained for each sample, which was then used as a correction factor. The goodness of fit [MSWD, or SUMS/(n – 2) where n is the number of steps] indicates if the scatter about the isochron regression line is solely analytical [e.g. SUMS/(n – 2) <2·5] or implies geological disturbance of the measured argon isotopes (McDougall & Harrison, 1988). 40Ar–39Ar data are reported in Table 4.
|
Weathered surfaces were removed from the whole-rock samples using a clipper saw and the samples crushed using a fly press. In the case of the coarse-grained pegmatites a large volume (
1000 cm3) of sample was crushed and thoroughly mixed to ensure that the samples were representative of the whole rock. After crushing, the samples were powdered in an agate Tema® mill. The major elements along with Zr, Cr, Ni, Sc and V were analysed by X-ray fluorescence (XRF) at Leicester University using conventional techniques; further details have been given by Tarney & Marsh (1991) and Kerr et al. (1996a). The XRF data are reported in Tables 5–7.
|
|
|
Most samples have also been analysed by ICP-MS using a Thermo Elemental Series X at Cardiff for the elements Ba, Ga, Nb, Rb, Sr, Th, Y, Hf and rare earth elements (REE). Rh and Re were used as internal standards for drift correction, and the calibration was carried out using laboratory standards and selected international standards with similar matrices. The samples were prepared for analysis by dissolution of 0·1 g powder with HF and HNO3, spiked by internal standards and run in a nitric acid matrix at a 0·1 g/dm3 concentration of total dissolved solids. The correction procedure involves blank-subtraction, drift-monitoring and correction for oxide–hydroxide interferences and isotopic overlaps.
Sr, Pb, Nd and Hf isotope ratios were measured on unleached samples and determined at the NERC Isotope Geosciences Laboratory, Keyworth (NIGL). The data are presented in Table 8. Procedures used in the analysis of Sr, Pb, Nd and Hf isotopes have been described by Royse et al. (1998), Kempton et al. (2000) and Kempton & McGill (2002). Sr and Nd were run as the metal species on single Ta filaments and double Re–Ta filaments respectively, using a Finnigan MAT 262 multicollector mass spectrometer in static mode. The effects of fractionation during runs were eliminated by normalizing 87Sr/86Sr to a 86Sr/88Sr value of 0·1194 and 143Nd/144Nd to a 146Nd/144Nd value of 0·7219. Sample values for 143Nd/144Nd are reported relative to an accepted value of 0·71024 for NBS 987 and 0·51186 for La Jolla. Minimum uncertainty is derived from the external precision of standard measurements that over the course of analysis average 29 ppm (2) for 143Nd/144Nd and 35 ppm (2) for 87Sr/86Sr. Blanks for Sr and Nd are typically less than 300 and 200 pg, respectively.
|
Pb isotopes were analysed by multicollector (MC)-ICP-MS on a VG P54 system. Mass fractionation was corrected for during the run using a 203Tl/205Tl value of 0·41876, which was determined empirically by cross calibration with NBS 981. All Pb isotope ratios have been corrected relative to the NBS 981 composition of Todt et al. (1996)
. On the basis of repeated runs of NBS 981, the reproducibility of whole-rock Pb isotope measurements is better than ±0·01% (2). Hf was also analysed on the VG P54. Within-run standard error for Hf isotope measurements is normally less than 22 ppm (2). Minimum uncertainties are derived from the external precision of standard measurements, which average 44 ppm (2). Replicate analysis of our internal rock standard, pk-G-D12, over the course of analysis yielded 0·283050 ± 12 (2, n = 45), which is indistinguishable from our previously reported values (Nowell et al., 1998; Kempton et al., 2000). The data are corrected for mass fractionation during the run by normalization to 179Hf/177Hf of 0·7325 and are reported relative to an accepted value of JMC 475 of 0·282160, as recommended by Nowell et al. (1998). In all cases, within-run precision is less than external reproducibility, and the procedural blank is less than 200 pg. Sr, Nd and Hf isotope ratios have been age corrected to 90 Ma.
|
FIELD RELATIONSHIPS, PETROGRAPHY AND MINERALOGY |
|---|
 TOPABSTRACTINTRODUCTIONANALYTICAL METHODS FIELD RELATIONSHIPS, PETROGRAPHY... 40Ar-39Ar AGES OF THE...GEOCHEMISTRYPETROGENESISPOSSIBLE MODELS FOR THE...IMPLICATIONS OF LOCALIZED HIGHER...CONCLUSIONSREFERENCES |
|---|
Field relationships in the Bolívar Ultramafic Complex
The BUC forms part of the eastern foothills of the Western Cordillera. The complex is bounded to the east by the Cauca–Patia Fault and to the west by the Roldanillo Fault (Fig. 1). The Roldanillo Fault marks the contact of the BUC with the basaltic, fault-bounded blocks of the CCOP (Kerr et al., 1997a
). These accreted basaltic slices alternate with fault blocks of deformed late Cretaceous marine shales and greywackes (Fig. 1). The detailed geology of the Bolívar ultramafic rocks has already been discussed in some detail (Barrero, 1979; Bourgois et al., 1982; Nivia et al., 1992; Nivia, 1996) and will be summarized only briefly here.
The ultramafic rocks consist of intercalations of 10 m thick layers of serpentinized dunite with 10–30 cm thick units of serpentinized olivine websterite, clinopyroxenite, wehrlite and olivine gabbronorite. The BUC also comprises an extensive sequence of layered and isotropic gabbros. The lower gabbro horizons are dominated by inter-banded (0·3–1·0 m thick) gabbro, olivine gabbro, gabbronorite and occasionally anorthosite, whereas the upper horizons consist of isotropic gabbro, which in places grades into anorthosite. A significant proportion of the BUC rocks are extensively foliated and sheared, and the clinopyroxene in some of the gabbro horizons has been replaced by secondary amphibole (Nivia, 1996).
Historically, the elucidation of the structure of the BUC has proved problematic, not least because of poor exposure. Barrero (1979) proposed that the BUC was a concentrically zoned body, whereas Bourgois et al. (1982) suggested that the general form of the complex was anticlinal. In the 1990s the construction of a new road through the complex opened up new road-cut sections that revealed structural evidence suggesting that the complex comprises a series of faulted–imbricated blocks rather than being a concentrically zoned body or an anticlinal structure (Nivia et al., 1992). On the basis of this new structural evidence, it is clear that the ultramafic rocks generally occupy the lower layers of the complex and grade upwards into the layered gabbro, with the isotropic gabbro occurring at high structural levels within the complex (Nivia et al., 1992).
The most conspicuous feature of the BUC is abundant plagioclase–hornblende-bearing pegmatite dykes and veins (Fig. 3), most of which trend east–west. These veins are, on average, 50 cm wide, although they range from a few centimetres to 70 cm wide. On the basis of field observations, three main pegmatite types can be identified: (1) mafic pegmatites, containing a large proportion of hornblende (>50%) with plagioclase; (2) intermediate pegmatites, containing between 50 and 10% hornblende with abundant plagioclase; (3) felsic pegmatites comprising predominantly plagioclase with <10% hornblende and small, but variable, amounts of quartz and biotite. Euhedral to subhedral hornblende crystals are up to 10 cm in length, and many grow perpendicular to the walls of the veins (Fig. 3).
|
Some of the veins were intruded after solidification of the host rocks, as evidenced by sharp contacts between the veins and the host rock. However, the absence of chilled margins suggests that the veins were intruded only shortly after the solidification of the host rock. In several places the pegmatites show a progressive gradation to more gabbroic compositions. Although many of the pegmatites are essentially undeformed, some are sheared and ptygmatically folded. The close temporal relationship between the BUC and the pegmatites is further evidenced by the occurrence of rounded blocks of leucrocratic gabbro within some of the pegmatites.
Observations in a quarry just outside Bolívar town (Nivia et al., 1992; Nivia, 1996) suggest that the mineralogy of the pegmatites is controlled, at least in part, by their level of emplacement within the complex; that is, the pegmatites found in the lower parts of the BUC are of the hornblende-rich variety, whereas at higher structural levels hornblende-poor, quartz and biotite-bearing pegmatites predominate. Furthermore, many of the veins possess a much greater modal percentage of hornblende at their margins than in their centre.
Field relationships of the Vijes felsites
Just north of the town of Vijes, and 50 km south of the BUC (Fig. 1), a series of felsites (rhyodacites) are associated with the oceanic plateau basalts of the Western Cordillera. The Vijes felsites occur along strike and in the same fault block as the BUC. The field relationships and possible origin of these felsites have previously been discussed by Millward (1983), Aspden & McCourt (1986) and Kerr et al. (1996b), and they were originally observed to comprise a sequence of plugs (up to 3·5 km2 in area), sheets and dykes (up to 3–4 m wide). Construction of a new road through the area has allowed access to deeper levels of this rock association and uncovered evidence of magma mixing in the form of volcanic rocks comprising felsic blebs in a basaltic matrix (Fig. 3c). Additionally, the new road cuts have exposed bedded deposits of felsitic volcanic ash and breccia, comprising felsite clasts (up to 5 cm in diameter) set in a fine-grained felsitic matrix (Fig. 3d). Some of the larger felsite breccia clasts have been sampled and analysed during the present study. For the purposes of this study, the Vijes rocks can be considered in terms of three broad groups: felsite dykes, felsite breccias and basalts–andesites from the mixed magma units.
Petrography and mineral chemistry of the BUC and Vijes felsites
In thin section the plagioclase and hornblende crystals in the pegmatites show virtually no evidence of zoning. Unlike many pegmatites, accessory minerals are uncommon, and apatite is the only accessory phase that has been positively identified. Many of the plagioclase crystals have undergone variable degrees of late-stage sericitization, particularly around their margins, and along cracks. Magnesiohornblende crystals have in places been altered to chlorite and an oxide mineral.
To fully characterize the chemistry of the pegmatite minerals, representative samples from the mafic, intermediate and felsic pegmatite groups were analysed by electron microprobe at Leicester University. Representative data are listed in Tables 1 and 2 and summarized graphically in Fig. 2. All amphiboles can be classified as magnesiohornblendes according to the revised amphibole classification scheme of the International Mineralogical Association (Leake et al., 1997); that is, they have CaB 
1·5, (Na + K)A < 0·5, Mg/(Mg + Fe2+) > 0·5, and Si in the formula ranges from 7·5 to 6·5 (Fig. 2). Magnesiohornblendes in the mafic pegmatites (those with a high modal content of amphibole) have generally higher Ca and Ti and lower Mn than magnesiohornblendes from the intermediate pegmatites (Fig. 2). Although there is no petrographic evidence of zoning, core vs rim analyses of several amphiboles indicate a small but systematic variation in chemistry; that is, Si, Ca, Mg and Mn are higher, whereas Al, Fe and K are lower in cores relative to rims of crystals (Fig. 2 and Table 1).
Plagioclase varies widely in composition from An84 to An34. Mafic pegmatites contain plagioclases that fall broadly into two compositional groups: An84–82 (i.e. bytownite) and An54–47 (i.e. labradorite). Intermediate and felsic pegmatites contain andesine plagioclases, which are much more restricted in composition (i.e. An50–41 and An41–34, respectively; Fig. 2 and Table 2). As a result of alteration it was difficult to analyse the rims of many of the plagioclase crystals. However, where this was possible, cores and rims were usually found to have very similar An contents (Table 2). Small decreases in An content (up to An3–4) were observed between core and rim in only a couple of intermediate pegmatites.
The gabbros, olivine gabbros and gabbronorites of the BUC are equigranular and commonly have ophitic textures. In the lower layered part of the sequence, crystals are oriented with their long axis parallel to the compositional banding. Olivines are anhedral and are partially altered to serpentine, whereas clinopyroxenes are mainly subhedral to anhedral, show hypersthene exolution lamellae and are variably altered to amphibole. Orthopyroxenes are subhedral to euhedral with a stubby prismatic habit. Plagioclases are variably altered to sericite and in the cumulitic levels exhibit zonation. More detailed descriptions have been given by Nivia (1996).
Petrographically, the felsites consist of 5–10% plagioclase and (strain shadowed) quartz microphenocrysts in a fine-grained or glassy groundmass, comprising quartz, plagioclase and minor hornblende. The groundmass has been extensively modified by low-grade metamorphism to chlorite, clays, epidote and pumpellyite. A similar grade of alteration and metamorphism is observed in the oceanic plateau basalts in the Western Cordillera, suggesting that the felsites formed prior to continental accretion of the CCOP.
|
40Ar–39Ar AGES OF THE BOLÍVAR PEGMATITES |
|---|
Previous attempts at radiometric dating of the Bolívar pegmatites have applied the K–Ar method to the magnesiohornblendes. These have yielded ages of 106 ± 18, 102 ± 18, 78 ± 18 and 70 ± 14 Ma (Barrero, 1979
; Brook, 1984). However, the Bolívar pegmatites have undergone variable degrees of sub-solidus alteration and this, combined with their low contents of K2O, has resulted in the wide range of reported K–Ar ages with large errors. Our 40Ar–39Ar step-heating experiments on magnesiohornblendes from the Bolívar pegmatites have yielded more precise ages with much smaller errors than the K–Ar dating (Table 4); step-age spectra and inverse isochron correlation diagrams are given in Fig. 4. Although some experiments yielded (unreported) discordant ages, the four experiments reported in Table 4 all gave concordant ages with low SUMS/(n – 2) values, and near-atmospheric 36Ar/40Ar intercepts with between 68 and 98% of the released 39Ar comprising the age-determining plateau (Fig. 4). The four concordant step-heating age determinations give a weighted mean age of 90·5 ± 0·9 Ma.
|
, Steps that have been used in the calculation of the inverse isochron age (
, steps not used in the regression.Of the four concordant age determinations, two were for mafic pegmatites and two for intermediate pegmatites. The two mafic pegmatites, Bol 94-22 and JTBol-4, give slightly older step-heated plateau ages (weighted mean 92·5 ± 1·8 Ma) than the two intermediate pegmatites, Bol 94-18 and Bol 94-15 (weighted mean 89·7 ± 1·1 Ma). It is significant that the age range of the pegmatites, in particular the older ages, are consistent with the proposed connection between the Cenomanian–Turonian (93 Ma) anoxic event and CCOP volcanism (Sinton & Duncan, 1997
; Kerr, 1998). |
GEOCHEMISTRY |
|---|
Major elements
The Bolívar gabbros range in composition from 4·1 to 31·5 wt % MgO (Table 5; Fig. 5), largely reflecting the modal abundance of olivine. For samples with MgO <12 wt %, two groups of gabbros can be distinguished in terms of TiO2 contents: enriched (e-) gabbros (TiO2 >0·6 wt %) and depleted (d-) gabbros (TiO2 <0·2 wt %). These groups can be distinguished on the basis of incompatible trace element abundances (see below).
|
The petrographic division of the pegmatites into mafic, intermediate and felsic types can be more rigorously quantified on the basis of major element composition. Mafic pegmatites contain >7 wt % MgO, >11 wt % CaO and <53 wt % SiO2, whereas intermediate pegmatites range from 7 to
2 wt % MgO, from 11 to 5 wt % CaO and from 53 to 75 wt % SiO2. Felsic pegmatites contain <2 wt % MgO, <5 wt % CaO and >75 wt % SiO2. (Table 6; Fig. 5). These chemical differences are consistent with the mineralogical changes on going from mafic through intermediate to felsic pegmatites, i.e. increasing modal percentage of more Na-rich plagioclase at the expense of magnesiohornblende.
On the basis of their field relationships, the rocks from Vijes can be divided into three broad groups: felsite dykes, felsite breccias and basalts–andesites from the mixed magma bodies. Although the Vijes felsites and basalts are altered, containing abundant secondary clay minerals, these three groups can also be distinguished on the basis of major element chemistry (Table 7; Fig. 5). Felsite dykes generally have lower MgO (1 wt %) and Fe2O3(t) (4–5 wt %) than breccias [MgO 1–4 wt %; Fe2O3(t) 5–13 wt %]. Although some of the scatter in Na2O and K2O concentrations (Table 7; Fig. 5) may be due to sub-solidus alteration, it is noteworthy that the dykes all have higher Na2O and generally higher K2O than the breccia clasts. The basaltic–andesitic portion of the mixed magma unit possesses broadly higher MgO, TiO2, Fe2O3 and CaO and lower SiO2 than the more felsitic rocks (Table 7; Fig. 5).
Trace elements
The major element division of Bolívar gabbros into two groups can also be observed in trace elements, with the e-gabbros having higher contents of incompatible trace elements than the d-gabbros (Figs 6 and 7). The e-gabbros are characterized by essentially flat primitive-mantle-normalized trace element patterns and the similarity to the basalts of the Western Cordillera is striking (Fig. 7a). In contrast, the d-gabbros are more compositionally variable, with the majority showing depletion in the most incompatible trace elements (Figs 6 and 7). It is interesting to note that the d-gabbros are those from the middle of the complex west of Bolívar, which are most closely associated with the pegmatites; that is, these gabbros either grade into, or are intruded by, the pegmatites. In contrast, the e-gabbros are rarely associated with pegmatites and predominate to the south of the BUC.
|
|
Mafic, intermediate and felsic pegmatites have distinctive primitive-mantle-normalized trace element patterns (Fig. 7b). Mafic pegmatites are moderately enriched in incompatible trace elements (2–8 times primitive mantle); they show a marked depletion in the most incompatible trace elements and a small but distinct negative Ti anomaly (Fig. 7b). Unsurprisingly, the trace element pattern for magnesiohornblende (hand-picked and analysed by solution ICP-MS) is very similar to the pattern for the mafic pegmatites. In contrast, felsic pegmatites are markedly depleted in the heavy rare earth elements (HREE;
0·1 times primitive mantle) relative to the most incompatible trace elements, although the latter are similar to the values for mafic pegmatites (Fig. 7b). Felsic pegmatites also show negative Nb and Ti anomalies, combined with a positive Eu anomaly. The patterns of these felsic pegmatites are similar to the trace element pattern for plagioclase (andesine) in the Bolívar pegmatites (Fig. 7d). As with major elements, the trace element contents of the intermediate pegmatites lie between the compositions of the mafic and felsic pegmatites (Fig. 7b). The Vijes felsic dykes are slightly enriched in the most incompatible trace elements relative to the HREE and have negative Ti anomalies (Figs 6 and 7c). Of the three rock groups observed at Vijes, the felsite dykes have the highest abundances of incompatible trace elements, whereas these elements are significantly less abundant in felsite breccia clasts (Figs 6 and 7c). Although data are limited, breccia clasts also appear to show a progressive decrease in most incompatible trace element contents as major element compositions become more evolved (Fig. 6). The basaltic–andesitic part of the Vijes mixed magma has a flat to slightly light REE (LREE)-depleted primitive-mantle-normalized pattern, with trace element contents intermediate between those of the felsite dykes and the breccia clasts (Fig. 7c). As will be discussed below, the lack of a significant Nb anomaly, positive or negative, for all the Vijes rock types is significant.
Radiogenic isotopes
Given that all the samples collected during the present study have undergone moderate degrees of sub-solidus alteration, the most robust isotope systems are likely to be Sm–Nd and Lu–Hf. With the exception of one sample, all the pegmatites and Vijes rocks show only limited variation in 
Hfi and Ndi (calculated here and elsewhere for an age of 90 Ma), ranging from +12·8 to +14·4 and from +6·5 to +7·3, respectively (Fig. 8b). The exceptional sample, intermediate pegmatite Bol 94-19, has a similar Ndi to the other pegmatites but a higher Hfi value (+15·2). Two of the gabbros have slightly lower Hfi values (+12·4 to +12·5) than Vijes rocks and pegmatites but are similar to most rocks from the Western Cordillera. One of the e-gabbros (Bol 94-28) and one of the d-gabbros (Bol94-11) have a Hfi values of 16·1 and 14·6, respectively, but have Ndi values that are within the range of the other gabbros and pegmatites (Fig. 8b). Paradoxically, one of the e-gabbros (i.e. those which are more enriched in incompatible trace elements) has higher Ndi than one of the d-gabbros analysed (JTBol-8). Most basalts of the Western Cordillera (including Bol 94-25, which was collected near the margin of the BUC) range from +6·5 to +7·6 and from +12·1 to +12·6 for Ndi and Hfi, respectively. One exception is Cbu 94-13. This sample has an Ndi value similar to other basalts of the Western Cordillera, but its Hfi is significantly higher (+18·5). Its Hf isotope composition is comparable to that of Gorgona komatiites (Fig. 8b), although it is displaced to a significantly lower Ndi value.
|
The Bolívar pegmatites and gabbros display a relatively restricted range in (87Sr/86Sr)i from 0·7033 to 0·7038. In contrast, (87Sr/86Sr)i for the Vijes rocks ranges from 0·7045 to 0·7055. Given the advanced state of alteration of these rocks, and the fact that their
Ndi and Hfi values substantially overlap with the compositions of the pegmatites and gabbros, it is likely that these high 87Sr/86Sr values are the result of alteration. Some the CCOP basalts of the western Cordillera also have (87Sr/86Sr)i >0·704 (Fig. 8a), as do rocks from elsewhere in the CCOP (Kerr et al., 1996a, 1997a; Hauff et al., 2000b). These high (87Sr/86Sr)i signatures have, for the most part, been ascribed to interaction with seawater (e.g. Révillon et al., 2002; Thompson et al., 2004).
Measured 206Pb/204Pb ratios of the Bolívar pegmatites range from 18·6 to 18·9. These values overlap the 206Pb/204Pb ratios of the d-gabbros, whereas e-gabbros have higher 206Pb/204Pb ratios (19·1–19·4), similar to basalts of the Western Cordillera (Fig. 8c and d). Measured 208Pb/204Pb ratios show a similar spread of data between the different sample types (Fig. 8c). With the exception of one sample, the pegmatites show a relatively limited variation in 207Pb/204Pb (15·533–15·546); intermediate pegmatite Bol 94-19 has a higher ratio of 15·556, similar to the d-gabbros. The rocks from Vijes show much more variable Pb isotope ratios than either the Bolívar pegmatites or the gabbros. For example, the felsite dyke Pan 94-23 has the most radiogenic Pb isotope composition of any Vijes or Bolívar rock (Fig. 8c), with a 206Pb/204Pb of 19·8. However, there is evidence that Pb isotope systematics can be disturbed by sub-solidus alteration (Hauff et al., 2000a; Kempton et al., 2002; Thompson, 2002), particularly in older rocks. It is significant that Pan 94-23 is the most altered of the Vijes samples analysed for isotopes, containing as it does, highly sericitized feldspars set in a groundmass that has been largely altered to clay minerals. Thus it may be that the Pb isotope systematics of Pan 94-23 have been disturbed by interaction with a fluid with a relatively radiogenic Pb isotope composition. Although the source of such a fluid remains unclear, the elevated 87Sr/86Sr ratio of Pan 94-23 is consistent with this interpretation, as are Nd and Hf isotope ratios that are virtually indistinguishable from those of other Vijes rocks (Fig. 8b). Therefore, Pb and Sr isotope compositions of many of the rocks may be able to tell us only about the alteration history.
The Pb isotope ratios of the e-gabbro(s) (Bol 94-24 and Bol 94-28) and the Bolívar basalt (Bol 94-25) are similar to those of the basalts from the Western Cordillera (Fig. 8c and d). The Bolívar pegmatites and the d-gabbros generally have lower Pb isotope ratios than the basalts of the Western Cordillera and are, in fact, most similar to the picrites and komatiites from Gorgona Island. However, the Nd isotope compositions of Gorgona rocks indicate derivation from a significantly more depleted source mantle source region (Fig. 8b).
|
PETROGENESIS |
|---|
Negative Nb anomaly in the pegmatites—evidence of subduction?
Felsic and intermediate pegmatites have negative Nb anomalies on primitive-mantle-normalized multi-element plots (Fig. 7b). This feature, in a rock that represents an original liquid composition, is usually taken as evidence for a subduction-related origin or contamination by continental crust. However, as will be discussed in more detail below, it is unlikely that these pegmatites represent original liquid compositions. Alternatively, the negative Nb anomaly may be due to the accumulation of a particular mineral phase. We discuss this possibility for the Bolívar pegmatites below.
Figure 9a shows that (Nb/La)n correlates positively with MgO for the pegmatites. The samples with the lowest MgO and (Nb/La)n are the felsic and intermediate pegmatites. These rocks have a high modal abundance of plagioclase and thus it may be accumulation of this phase that is the source of the low (Nb/La)n. Felsic and intermediate pegmatites also have positive Eu anomalies—a result of their high plagioclase content. We can therefore use the size of the positive Eu anomaly as a proxy for the amount of plagioclase in the pegmatite. A plot of (Nb/La)n vs Eu* (i.e. a measure of the Eu anomaly; Fig. 9b), shows that, although a strong linear correlation is absent, samples with low (Nb/La)n also tend to have a larger Eu anomaly and hence more plagioclase. Thus, the samples with the largest negative Nb anomaly also contain the highest modal percentage of plagioclase. This occurs because the average Kd for Nb in plagioclase in a basaltic melt is 0·009, whereas the average Kd for La in plagioclase in a basaltic melt is 0·156. (Partition coefficient data from the GERM website http://www.earthref.org/databases/KDD/main.htm.) La will, therefore, be present in higher concentrations than Nb in plagioclase. Our two analyses of plagioclase from the Bolívar pegmatites (Fig. 7d) confirm this conclusion. Thus the low (Nb/La)n values for the more felsic Bolívar pegmatites simply reflect their high modal plagioclase content and do not indicate a subduction-related origin for these rocks.
|
Magmatic processes
To understand the petrogenesis of the Bolívar pegmatites, we must first assess whether they represent original, closed-system, magmatic compositions. Several lines of evidence suggest that they do not. First, the relative lack of compositional zoning in both the magnesiohornblende and plagioclase crystals argues against closed-system evolution. If crystallization occurred in a closed system, the liquid remaining after early crystallization would be depleted in CaO and MgO. This magma would precipitate more sodic feldspars and less MgO-rich hornblendes, both at the margins of the crystals and in the centre of the pegmatite veins, neither of which is observed. The more felsic pegmatites also tend to occur at higher stratigraphic levels in the BUC, whereas mafic pegmatites predominate nearer the base. This suggests that some form of differentiation has occurred within the pegmatites as melts migrated upwards. Thus, a possible model for the magmatic evolution of the pegmatites is differentiation by side-wall crystallization from mafic magma migrating upwards through the Bolívar complex. If the magma from which the crystals grew was constantly replenished from below, the pegmatite veins would consist of crystals of broadly uniform composition. In such a model the more evolved crystals would grow from magmas that had migrated to a higher level within the complex.
The lack of zoning in the most felsic pegmatites suggests that these rocks do not represent the final crystallization products of a residual melt and that hydrous magma also passed through these conduits en route to somewhere else. The similarity in Nd and Hf isotope ratios for most Bolívar pegmatites and Vijes felsites suggests that they may be genetically related. Thus the modelling presented below will test the proposal that the Vijes felsites (both dykes and breccia clasts) may represent the magmas residual from the crystallization of the Bolívar pegmatites. Vijes dykes and breccias display two distinct compositional groups on variation diagrams, particularly those involving trace elements (Fig. 6). When the compositions of magnesiohornblende and plagioclase (labradorite) from the mafic pegmatites are plotted on the variation diagrams, it becomes clear that the lower concentrations of Zr, Nb, LREE and Rb in the Vijes felsite breccia clasts could be due to the fractionation of magnesiohornblende from a basaltic magma.
The composition of the Vijes felsite breccia clasts can be successfully modelled by 85% fractionation of a mineral assemblage comprising 71% magnesiohornblende and 29% plagioclase, from a parental magma with the composition of the basaltic portion of the Vijes mixed magma (VBM) (Fig. 6; Table 9). In contrast, the higher concentrations of Zr, Nb and LREE in the Vijes felsite dykes (Fig. 6) can be modelled by 85% fractionation of a crystal assemblage that contains a higher proportion of plagioclase (42% plagioclase; 58% magnesiohornblende), again using VBM as a starting composition (Fig. 6; Table 9). As magnesiohornblende contains higher concentrations of incompatible trace elements than the postulated parental magma, removal of an assemblage comprising a significant proportion of magnesiohornblende will result in a reduction in the incompatible element trace element content of the felsitic daughter magma. Figure 10 is a primitive-mantle-normalized multi-element plot showing results of calculations, assuming 80% fractionation of assemblages with variable proportions of plagioclase and magnesiohornblende and VBM as the starting composition. Although these calculations give slightly different results relative to the mass balance modelling described above, they none the less illustrate the influence of fractionation of magnesiohornblende in reducing the concentration of the incompatible trace elements in the residual magma. This modelling clearly shows that the composition of the felsite breccia clasts requires removal of a crystal assemblage comprising a higher proportion of hornblende than is necessary to explain the composition of the less incompatible-element-depleted Vijes dykes.
|
|
In contrast, the major element chemistry of the felsite breccia clasts and dykes shows less compositional divergence at 70–80 wt % SiO2 than the trace elements (Fig. 5). There are several possible reasons for this. First, magnesiohornblende and plagioclase, although the main fractionating minerals, are not the only fractionating phases, as small amounts (<10%) of biotite and quartz are also found in the felsic pegmatites. The mismatch for TiO2 (Fig. 10) values (i.e. neither the magnesiohornblende nor the plagioclase contain enough TiO2 to explain the depletion with increasing SiO2; Fig. 5a) can be explained by invoking fractionation of Fe–Ti oxide from the felsite magma. Additionally, the small negative Nb anomalies observed in the primitive-mantle-normalized patterns for the felsite dykes (Fig. 7c) can also be explained by the removal of a small amount of Fe–Ti oxide, as Nb is moderately compatible in magnetite and ilmenite. Second, the major element compositions (particularly K2O and Na2O) of the felsites and, to a lesser extent, of the pegmatites may have been modified by sub-solidus alteration. Evidence for elemental mobility is also observed in the variable Rb contents of the Bolívar and Vijes rocks, in contrast to generally more immobile elements such as Zr and Nb (Fig. 6).
The d-gabbros of the BUC are in some cases transitional to the pegmatites, grading mineralogically from hornblende-bearing gabbro to pegmatite. This suggests that the pegmatites (and ultimately the felsites) may have crystallized from magmas similar to those that crystallized the gabbros. However, although the Nd isotope systematics suggest that the pegmatites and gabbros may be genetically related, two out of the four gabbros have lower Hfi values than the pegmatites and comparable Ndi values to most of the basalts of the Western Cordillera. As there are few (if any) feasible crustal contaminants within the oceanic lithosphere that would increase Hfi while leaving Ndi unaffected for the pegmatites and some of the gabbros, it is probable that this signature is derived from the mantle source region. Salters & Zindler (1995) have reported Hfi values of >20 for abyssal peridotites from the SW Indian Ridge. Thus, melting of a heterogeneous mantle source region containing such a high Hfi component may explain the difference in Hfi between the Bolívar pegmatites and some of the gabbros. In this scenario, the generally slightly younger pegmatites crystallized from heterogeneous magmas that had already undergone significant fractionation from a high-MgO parental magma (Fig. 11).
|
To assess the validity of this proposed petrogenetic scheme, we need to start with a suitable parental magma composition. Kerr et al. (1997a)
have shown that the basalts of the Western Cordillera have undergone 20% crystallization of an assemblage dominated by olivine and clinopyroxene. Therefore, none of the Western Cordillera basalts can be used directly as a parental magma in our modelling. We have therefore chosen to use a picrite from Curaçao (also a part of the CCOP; Kerr et al., 1996a) as the starting composition. We have used the MIXFRAC program of Nielsen (1990) to model the fractionation of the parental magma, and these results are shown in Figs 5 and 6. Although the MIXFRAC parameterization is primarily designed to model the fractionation of dry magmas, in the early stages of fractionation of a picritic magma olivine is likely to be the dominant fractionating phase regardless of whether the magma is wet or dry. The results show that the composition of the more mafic (6·7 wt % MgO) of the two VBM samples (Pan 94-28) can be modelled by 20–30% fractionation of a crystal assemblage dominated by olivine with smaller amounts of pyroxene and plagioclase (Figs. 5 and 6). Back-projection of the fractionation trend intersects the compositions of some d-gabbros (Figs 5 and 6), thus providing support for the contention that the d-gabbros represent the cumulates fractionated from the parental magmas that form the VBM magmas. In contrast, the more evolved (3·61 wt % MgO) VBM sample (Pan 94-29) has lower incompatible trace element concentrations (Figs 6 and 10) and so does not fit the modelled trend as well. These lower concentrations can be explained by fractionation of magnesiohornblende from the magma. The fact that Pan 94-29 fits the ‘dry’ modelled fractionation trend suggests that relatively little magnesiohornblende fractionated from this magma, thus validating the use of a dry magma modelling program. Magnesiohornblende obviously begins to fractionate from the VBM magmas at MgO contents <6·5 wt %.
The proposed petrogenetic scheme for the Bolívar–Vijes complex, discussed in this section, is summarized schematically in Fig. 11.
Water content of the primary (parental) magma
Mantle melt modelling has been carried out for many CCOP magmas having very similar trace element chemistry to the Bolívar rocks (Kerr et al., 1996a, 1997a, 2002b; Arndt et al., 1997; Révillon et al., 2000; Herzberg & O'Hara, 2002); consequently, similar modelling will not be repeated here. Instead, our focus is on assessing the amount of water in the parental magma(s) and the implications this has for their petrogenesis. Mantle melt modelling for magmas of the CCOP shows that the MgO content of the parental magmas was on average 15 wt % MgO (e.g. Kerr et al., 1996a, 1996b; Révillon et al., 2000). Therefore, if we can estimate the amount of H2O a magma must contain before it begins to fractionate hornblende, we can calculate the water content of the parental magma.
The concentration of H2O necessary for amphibole to fractionate from a magma has been estimated by several studies. The threshold H2O contents obtained by these studies range from 6 wt % (Merzbacher & Eggler, 1984) through 5 wt % (Naney, 1983) to 2 wt % (Luhr, 1992). As two of these studies suggest that the threshold above which amphibole will fractionate is 5–6 wt % H2O, the average water content at which amphibole will begin to fractionate is likely to be >4 wt %. Therefore, the magma that VBM sample Pan 94-28 crystallized from (6·7 wt % MgO), probably originally contained between 4 and 6 wt % H2O. In contrast to VBM sample Pan 94-29, Pan 94-28 shows little chemical evidence for hornblende fractionation and was probably just saturated in water. Assuming (1) 25% fractional crystallization, (2) a partition coefficient (D) for water of 0·01 (in a magma not fractionating amphibole; Michael, 1995) and (3) a parental magma containing 15 wt % MgO, the amount of water in the parental magma can be calculated from the assumed water content of Pan 94-28. These calculations suggest that if the magma that Pan 94-28 crystallized from contained between 4 and 6 wt % H2O, then the parental magma would have contained between 3·0 and 4·5 wt % water.
In addition to water content, the temperature of the magma will also control whether amphibole will fractionate from a basic magma. Experiments by Helz (1982) and Allen & Boettcher (1983) have shown that amphibole is unlikely to be stable in a magma at temperatures in excess of 1000°C, even at high water contents. This, therefore, supports the results of our geochemical modelling, which suggest that hornblende did not fractionate from the Bolívar magmas at high MgO contents (i.e. high temperatures).
|
POSSIBLE MODELS FOR THE TECTONIC SETTING OF THE BUC |
|---|
Arc-related setting
The most obvious tectonic environment of formation for igneous rocks containing a large proportion of amphibole (implying a volatile-rich mantle source region) is in a subduction-related setting. This is a conclusion reached by some previous workers on the BUC (e.g. Barrero, 1979
; Bourgois et al., 1982). However, as we have shown above, the trace element patterns of the Bolívar rocks do not support a subduction-related origin. Furthermore, the residual felsitic magmas (the Vijes felsites) derived from crystallization of the Bolívar ultramafic rocks, gabbros and pegmatites, show no evidence of a subduction-related origin, i.e. they do not have high La/Nb ratios. Another line of evidence that argues against a subduction-related origin for the BUC is weighted mean age of the Bolívar pegmatites. This mean age almost exactly coincides with the main peak of ages for the CCOP (weighted mean age 90·6 ± 0·3; Fig. 12). These statistically indistinguishable ages mean that if the BUC is subduction-related, the arc had to form at the edge of the plateau at the same time as the plateau lavas were erupting. Although this is geodynamically possible, it is unlikely and, when combined with the lack of evidence for a subduction signature, makes this scenario highly improbable.
|
Melting during accretion of the CCOP
On the basis of the limited geochemical data then available, Nivia (1996)
and Kerr et al. (1998) proposed that the Bolívar dunites and gabbros were generated in a magma chamber within the CCOP. In their models, the pegmatites were proposed to represent fluid-aided melting at the base of the accreting plateau as slices of it were obducted onto the margin of NW South America in the late Cretaceous. Our new dates and radiogenic isotope data mean that we are now able to test this model more rigorously. The lack of an obvious geochemical input from sediment, although not precluding the obduction model, does make it less likely. However, it is the close similarity in ages between the CCOP and the Bolívar pegmatites that firmly rules out an origin for the pegmatites by melting during obduction. In short, the ages and chemistry of the Bolívar pegmatites mean that they had to have formed at the same time as the plateau, in a non-subduction-related environment, well away from a continental margin.
Derivation from a mantle plume
The geochemical and geochronological data presented in this paper suggest that the BUC is an integral part of the CCOP. This being the case, the question that arises is: ‘Where does the water come from?’ Wet melting of lithosphere heated from below by a mantle plume has been invoked as a model for the generation of some continental flood basalt provinces (Gallagher & Hawkesworth, 1992; Turner et al., 1996). Therefore in such provinces the continental lithosphere could act as a source region for ‘wet’ melts. However, the hydrous or metasomatized lithosphere that may be present below continental flood basalt provinces is relatively ancient cratonic lithosphere that has existed for time periods in excess of 0·5 Gyr. In contrast, the oldest in situ oceanic crust at the present day is Jurassic in age. The oceanic crust of the Farallon plate on which the CCOP was built is believed to be Early Cretaceous, or possibly Jurassic, in age (Nakanishi & Winterer, 1998). This oceanic lithosphere was therefore probably too young to be a realistic source of the higher water contents in the Bolívar magmas. Furthermore, oceanic lithosphere that has been altered and hydrated by circulating seawater will possess elevated 87Sr/86Sr ratios, because Jurassic and Early Cretaceous seawater has an average 87Sr/86Sr in excess of 0·707 (Jones et al., 1994). Thus contamination of plume-derived magmas by partial melting of altered oceanic crust will yield magmas with significantly higher 87Sr/86Sr. However, none of the BUC rocks possess (87Sr/86Sr)i values >0·7039, suggesting that they have not significantly interacted with altered oceanic crust. The Vijes rocks have elevated (87Sr/86Sr)i (0·7045–0·7055); however, they are much more altered than the BUC rocks and so these elevated values are more likely to be the result of seawater alteration and not contamination by a previously altered oceanic crust.
Therefore, if it is unlikely that the water in the BUC magmas is derived from the lithosphere, we need to assess the possibility that the water in these magmas had its ultimate origin as an integral part of the mantle plume source region of the CCOP. Over the last few years the issue of whether mantle plumes, the source regions of large igneous provinces (LIPs), contain significant (>300 ppm) quantities of water (and/or other volatiles) has been extensively discussed (Arndt et al., 1998; Michael, 1999; Jamtveit et al., 2001; Dixon et al., 2002; Nichols et al., 2002; Wallace, 2002).
Water in mantle plumes can ultimately come from one of two sources: it may be either juvenile (i.e. left over from accretion of the Earth) or recycled (i.e. derived from subducted slabs; Dixon et al., 2002). A key issue is the amount of water that survives subduction and is transported to the deep mantle. Dixon et al. (2002) have calculated that water is extracted from the lithosphere during subduction with >92% efficiency.
High-pressure experiments have shown that dense hydrous magnesian silicate phases can exist at high pressures in the mantle (Thompson, 1992; Kohlstedt et al., 1996; Ohtani et al., 1997). Data from mid-ocean ridge basalt (MORB), ocean islands and LIPs reveal that the water content of mantle plume source regions is generally higher than that of MORB. For example, the Sala y Gomez plume source region contains 750 ± 210 ppm water, whereas nearby depleted MORB-source mantle contains 120 ppm water (Simons et al., 2002). The same appears to be true of the Iceland hotspot, which has been calculated to contain between 620 and 920 ppm water. In contrast, farther south along the Reykyanes Ridge, the water content of depleted MORB mantle is 165 ppm (Nichols et al., 2002). Jamtveit et al. (2001) have analysed the water content of olivines from the North Atlantic Igneous Province and have used these to infer that the plume source region has (and had) a water content of >300 ppm.
The mantle melt modelling carried out by Kerr et al. (1996a, 1996b, 1997a, 2002b), Arndt et al. (1997), Hauff et al. (1997), Herzberg & O'Hara (2002) and Révillon et al. (2000) has shown that the extent of partial mantle melting responsible for the CCOP magmas was of the order of 20%. Assuming fractional melting of a mantle source region and a partition coefficient of water of 0·01 during mantle melting, we have calculated the water content of the mantle source region at variable degrees of melting for primary magmas containing between 1·5 and 4·5 wt % water. These results are summarized in Fig. 13, and reveal that at 20% melting the amount of water in the source varies between 188 and 565 ppm when the primary melt contains 1·5 and 4·5 wt % water, respectively. However, as was discussed above, the average threshold water contents for the crystallization of amphibole are greater than 4 wt %, which equates to a primary magma water content of 3 wt %. Thus, the water content of the mantle source region of the Bolívar and Vijes magmas is likely to have been >380 ppm. The estimated water content of the plume source region would be higher if we assumed a greater degree of mantle melting, a higher primary magma water content, a lower Kd for water in the mantle, or if we had used a pooled fractional melting model.
|
The evidence from volcanic rocks erupted in different tectonic settings seems to indicate that the mantle sources of ocean island basalts and LIPs contain higher levels of water than the depleted MORB-source mantle. However, as discussed by Dixon et al. (2002)
, this evidence suggests that mantle plume source regions, particularly those that are ultimately derived from subducted slabs, are more ‘damp’ than ‘wet’. Certainly, this is supported by the calculated levels of water in some mantle plumes and the general lack of evidence for the eruption of hydrous magmas in LIPs in general, and oceanic plateaux in particular. Nevertheless, our evidence implies that the plume head source region of the CCOP did contain a region (or regions) wet enough to produce localized hydrous melts.
The mantle plume source region of oceanic plateaux, particularly for well-exposed plateaux such as the CCOP, appears to be markedly heterogeneous in terms of its trace element and radiogenic isotope characteristics (Arndt et al., 1997; Hauff et al., 2000a; Kerr et al., 2002b; Révillon et al., 2002; Mamberti et al., 2003). This heterogeneity is not surprising when one considers the nature of the subducted materials believed to descend into the deep mantle. Although the calculations of Dixon et al. (2002) suggest that >92% of the water is extracted from slabs during subduction, 8% of the water is transported to depth. It is therefore not inconceivable that mantle plume source regions are also heterogeneous in terms of water content.
|
IMPLICATIONS OF LOCALIZED HIGHER WATER CONTENTS IN MANTLE PLUMES |
|---|
|
TOPABSTRACTINTRODUCTIONANALYTICAL METHODSFIELD RELATIONSHIPS, PETROGRAPHY...40Ar-39Ar AGES OF THE...GEOCHEMISTRYPETROGENESISPOSSIBLE MODELS FOR THE...IMPLICATIONS OF LOCALIZED HIGHER...CONCLUSIONSREFERENCES |
|---|
There are several important implications of localized higher water contents in mantle plumes. The first of these will be a lowering of the solidus temperature, which will result in either a higher percentage of melt generation (at a fixed mantle potential temperature) or, if the percentage of mantle melting remains the same, a lowering of the mantle temperature. These effects have been quantified by the hydrous melting experiments of Hirose & Kawamoto (1995)
on KLB-1 at 1 Ga, which revealed that at a mantle temperature of 1350°C the amount of melting increased from 20 to 30% upon addition of 1000 ppm water to the source. Likewise, the experiments showed that addition of 1000 ppm water to the source led to a reduction in the temperature needed to form the same amount of melt as an anhydrous source. At 20% melting this temperature was reduced by 35°C from 1350°C to 1315°C. The slightly more incompatible trace element-depleted nature of the Vijes basalt Pan 94-28(i.e. the one that has not fractionated magnesiohornblende) in comparison with other basalts of equivalent Mg-number from the Western Cordillera may be a reflection of higher degrees of melting. An obvious implication is that if all mantle plumes contain significant hydrous patches, then calculations that estimate the potential temperature (Tp) of mantle plumes based on the volume and chemistry of the erupted melt, which assume an anhydrous source (e.g. McKenzie & Bickle, 1988), may well be overestimates and thus in error.
Mantle plumes that contain wet patches also have implications for the rheology of the mantle. Experimental results reported by Hirth & Kohlstedt (1996) have shown that water, if contained in mantle olivine, can decrease the viscosity of the mantle by at least two orders of magnitude relative to anhydrous mantle. They also estimated that partial melting causes a large increase in the viscosity of residual mantle after melt extraction, as a result of dehydration. This reduction in viscosity has been proposed to create a viscosity discontinuity, which occurs at depths of 60–70 km below mid-ocean ridges (Hirth & Kohlstedt, 1996) or between 70 and 100 km depths for mantle plumes (Wallace, 1998). This viscosity discontinuity may be important in focusing melts from a broad melting region towards a localized eruptive centre (Wallace, 1998). The reduced viscosity may also mean that hydrous plumes (or hydrous patches in essentially anhydrous plumes) can rise through the mantle at a faster rate than totally anhydrous plumes.
Finally, if mantle plumes are more hydrous than has hitherto been realized, this has important ramifications for the environmental impact of LIPs. A key component of many models of the climatic impact of LIP eruption is of sulphuric acid aerosols and ash in blocking sunlight from reaching the Earth, so causing a ‘volcanic winter’ (e.g. Courtillot et al., 1996; Wignall, 2001). For these aerosols and ash to be distributed around the globe, and so have a global rather than a local influence on climate, they must be injected into the stratosphere (Strothers et al., 1986). The tropopause, the boundary between the troposphere and the stratosphere, is 9·5 km above sea level at polar latitudes and 17 km above sea level at equatorial latitudes. Therefore, eruptions from LIPs need to be explosive enough to enable them to inject ash and aerosols to heights of between 10 and 17 km, if LIP eruptions are to have an effect on global climate. Obviously, if mantle plume source regions are more hydrous than has previously been thought, this will mean that at least some of their eruptive episodes will be more explosive, thus making it more likely that ash and aerosols will reach the stratosphere. The presence of ash and volcanic breccias at Vijes is testament to explosive eruptions during the formation of the CCOP, as a result of higher than usual mantle water contents.
|
CONCLUSIONS |
|---|
(1)The Bolívar mafic pegmatites are indistinguishable in age from, and overlap in terms of radiogenic isotope composition with, the basalts of the CCOP. The chemistry of the Bolívar mafic pegmatites is inconsistent with a subduction-related origin for the Bolívar Ultramafic Complex.
(2)The Bolívar ultramafic rocks and gabbros represent the olivine and pyroxene crystal cumulates from the fractionation of a hydrous plume-derived magma. In the later stages of crystallization, when the MgO content reached 7 wt % MgO, the magma began to crystallize magnesiohornblende on the sidewalls of thin conduits as it migrated upwards. Continual replenishment of magma from below resulted in the formation of pegmatites with a high modal proportion of unzoned magnesiohornblende in the lower levels of the complex.
(3)At higher levels in the complex, felsic residual magmas crystallized hornblende-poor pegmatites with a high modal proportion of essentially unzoned plagioclase along with quartz and biotite.
(4)Tapping of the magma at different stages during pegmatite crystallization resulted in the emplacement and eruption of variably incompatible trace element-depleted felsites. The extent of incompatible trace element depletion is dependent on the amount of magnesiohornblende, relative to plagioclase, that is removed from the magma during pegmatite crystallization.
(5)A hydrous component appears to be an integral part of the mantle plume source region of the CCOP, and modelling calculations suggest that the mantle plume source region of the BUC contained at least 400 ppm water.
(6)We propose that, in addition to their heterogeneity in trace elements and radiogenic isotopes, mantle plumes are also heterogeneous in terms of their water content and may well possess localized hydrous regions.
(7)Mantle plumes that are wetter than previously realized probably melted at lower temperatures (or to greater extents at comparable temperatures), possessed lower viscosity and thus rose more quickly, and produced magmas with more explosive eruptions than equivalent dry compositions. The last factor has implications in particular for links between LIPs and mass extinction events.
|
ACKNOWLEDGEMENTS |
|---|
Discussions with Andy Saunders, Giz Marriner, Pat Thompson, Roz White, Bernard Leake, John Aspden and Dave Millward helped clarify many of the ideas contained in this paper. The analytical support of Rob Wilson and Nick Marsh at Leicester and Iain McDonald at Cardiff is greatly appreciated. The paper has benefited considerably from constructive reviews from Bob Duncan, Dominique Weiss and Henriette Lapierre, and the editorial advice of Marjorie Wilson. These studies were supported by the Natural Environment Research Council (UK) through Grants GR3/8984 and GR9/583A (to J.T.). Additional ICP-MS trace element analyses were supported by an NERC–HEFCE Joint Infrastructure Fund grant (NER/H/S/200/00862) to J. A. Pearce et al. Isotope analyses were funded by a block grant from the NERC to NIGL. We wish to thank INGEOMINAS for logistical support during fieldwork in Colombia. This paper represents NERC Isotope Geosciences Laboratory publication number 635.
|
FOOTNOTES |
|---|
* Corresponding author. E-mail: kerra{at}cf.ac.uk
Present address: Natural Environment Research Council, Polaris House, North Star Avenue, Swindon SN2 1EU, UK. 
|
REFERENCES |
|---|
Aitken, B. G. & Echeverría, L. M. (1984). Petrology and geochemistry of komatiites and tholeiites from Gorgona Island, Colombia. Contributions to Mineralogy and Petrology 86, 94–105.[CrossRef][Web of Science]
Allen, J. C. & Boettcher, A. L. (1983). The stability of amphibole in andesite and basalt at high pressures. American Mineralogist 68, 307–314.[Abstract]
Arndt, N., Ginibre, C., Chauvel, C., Albarède, F., Cheadle, M., Herzberg, C., Jenner, G. & Lahaye, Y. (1998). Were komatiites wet? Geology 26, 739–742.
Arndt, N. T., Kerr, A. C. & Tarney, J. (1997). Dynamic melting in plume heads: the formation of Gorgona komatiites and basalts. Earth and Planetary Science Letters 146, 289–301.[CrossRef][Web of Science]
Aspden, J. A. (1984). The geology of the Western Cordillera, Department of Valle, Colombia (sheets 261, 278, 280 and 299). Cali: INGEOMINAS–Mission Britanica (British Geological Survey).
Aspden, J. A. & McCourt, W. J. (1986). Low-K granitoids from the western Cordillera of Colombia. Geologia Norandina 10, 19–27.
Barrero, D. (1979). Geology of the Central western Cordillera, west of Buga and Roldanillo, Colombia. Publication Especial INGEOMINAS 4, 1–75.
Bence, A. E., Papike, J. J. & Ayuso, R. A. (1975). Petrology of submarine basalts from the Central Caribbean: DSDP Leg 15. Journal of Geophysical Research 80, 4775–4804.
Blichert-Toft, J. & White, W. M. (2001). Hf isotope geochemistry of the Galapagos Islands. Geochemistry, Geophysics, Geosystems 2, paper number 2000GC000138.
Bourgois, J., Calle, B., Tournon, J., Azema, J. & Toussaint, J. F. (1982). The Andean ophiolitic megastructure in the Buga–Buenaventura traverse (Western Cordillera, Valle, Colombia). Tectonophysics 82, 207–229.[CrossRef][Web of Science]
Brook, M. (1984). New radiometric age data from SW Colombia. Cali: INGEOMINAS–Mision Britanica (British Geological Survey), Report 10.
Burke, K. (1988). Tectonic evolution of the Caribbean. Annual Review of Earth and Planetary Sciences 16, 201–230.[CrossRef][Web of Science]
Castillo, P. R., Klein, E. M., Bender, J. F., Langmuir, C. H., Shirey, S. B., Batiza, R. & White, W. M. (2000). Petrology and Sr, Nd, and Pb isotope geochemistry of mid-ocean ridge basalt glasses from the 11°45' to 15°00' segment of the East Pacific Rise. Geochemistry, Geophysics, Geosystems 1, paper number 1999GC000024.
Coffin, M. F. & Eldholm, O. (1994). Large igneous provinces: crustal structure, dimensions, and external consequences. Reviews of Geophysics 32, 1–36.[CrossRef][Web of Science]
Coffin, M. F., Pringle, M. S., Duncan, R. A., Gladczenko, T. P., Storey, M., Muller, R. D. & Gahagan, L. A. (2002). Kerguelen hotspot magma output since 130 Ma. Journal of Petrology 43, 1121–1139.
Courtillot, V., Jaeger, J. J., Yang, Z. Z., Feraud, G. & Hofmann, C. K. B. (1996). The influence of continental flood basalts on mass extinctions; where do we stand? In: Ryder, G., Fastovsky, D. & Gartner, S. (eds) The Cretaceous–Tertiary Event and Other Catastrophes in Earth History. Geological Society of America, Special Papers 307, 513–525.
Dixon, J. E., Leist, L., Langmuir, C. & Schilling, J. G. (2002). Recycled dehydrated lithosphere observed in plume-influenced mid-ocean-ridge basalt. Nature 420, 385–389.[CrossRef][Medline]
Donnelly, T. W., Beets, D., Carr, M. J., Jackson, T., Klaver, G., Lewis, J., Maury, R., Schellenkens, H., Smith, A. L., Wadge, G. & Westercamp, D. (1990). History and tectonic setting of Caribbean magmatism. In: Dengo, G. & Case, J. E. (eds) The Geology of North America, Volume H, The Caribbean Region. Boulder, CO: Geological Society of America, pp. 339–374.
Duncan, R. A. & Hargraves, R. B. (1984). Plate tectonic evolution of the Caribbean region in the mantle reference frame. In: Bonini, W. E., Hargraves, R. B. & Shagan, R. (eds) The Caribbean–South America Plate Boundary and Regional Tectonics. Geological Society of America, Memoir 162, 81–93.
Dupré, B. & Echeverría, L. M. (1984). Pb isotopes of Gorgona Island (Colombia): isotopic variations correlated with magma type. Earth and Planetary Science Letters 67, 186–190.[CrossRef][Web of Science]
Eldholm, O. & Coffin, M. F. (2000). Large igneous provinces and plate tectonics. In: Richards, M. A., Gordon, R. G. & van der Hilst, R. D. (eds) The History and Dynamics of Global Plate Motions. American Geophysical Union Monograph 121, 309–326.
Gallagher, K. & Hawkesworth, C. J. (1992). Dehydration melting and the generation of continental flood basalts. Nature 358, 57–59.[CrossRef]
Hauff, F., Hoernle, K., Schmincke, H.-U. & Werner, R. (1997). A mid Cretaceous origin for the Galapagos hotspot: volcanological, petrological and geochemical evidence from Costa Rican oceanic crustal segments. Geologische Rundschau 86, 141–155.
Hauff, F., Hoernle, K., Tilton, G., Graham, D. W. & Kerr, A. C. (2000a). Large volume recycling of oceanic lithosphere over short time scales: geochemical constraints from the Caribbean Large Igneous Province. Earth and Planetary Science Letters 174, 247–263.[CrossRef][Web of Science]
Hauff, F., Hoernle, K., Van Den Bogaard, P., Alvarado, G. E. & Garbe-Schonberg, C. D. (2000b). Age and geochemistry of basaltic complexes in western Costa Rica: contributions to the geotectonic evolution of Central America. Geochemistry, Geophysics, Geosystems 1, paper number 1999GC000020.
Helz, R. T. (1982). Experimental studies of amphibole stability; phase relations and compositions of amphiboles produced in studies of the melting behavior of rocks. In: Veblen, D. R. & Ribbe, P. H. (eds) Amphiboles; Petrology and Experimental Phase Relations. Mineralogical Society of America, Reviews in Mineralogy 9B, 279–353.
Herzberg, C. & O'Hara, M. J. (2002). Plume-associated ultramafic magmas of Phanerozoic age. Journal of Petrology 43, 1857–1883.
Hirose, K. & Kawamoto, T. (1995). Hydrous partial melting of lherzolite at 1 GPa: the effect of H2O on the genesis of basaltic magmas. Earth and Planetary Science Letters 133, 463–473.[CrossRef][Web of Science]
Hirth, G. & Kohlstedt, D. L. (1996). Water in the oceanic upper mantle: implications for rheology, melt extraction and the evolution of the lithosphere. Earth and Planetary Science Letters 144, 93–108.[CrossRef][Web of Science]
Jamtveit, B., Brooker, R., Brooks, K., Larsen, L. M. & Pedersen, T. (2001). The water content of olivines from the North Atlantic Volcanic Province. Earth and Planetary Science Letters 186, 401–415.[CrossRef][Web of Science]
Jones, C. E., Jenkyns, H. C., Coe, A. L. & Hesselbo, S. P. (1994). Strontium isotopic variations in Jurassic and Cretaceous seawater. Geochimica et Cosmochimica Acta 58, 3061–3074.[CrossRef][Web of Science]
Kempton, P. D.& McGill, R. (2002). Procedures for the analysis of common lead at the NERC Isotope Geosciences Laboratory and an assessment of data quality. NIGL Report Series No. 178, 60 pp.
Kempton, P. D., Fitton, J. G., Saunders, A. D., Nowell, G. M., Taylor, R. N., Hardarson, B. S. & Pearson, G. (2000). The Iceland plume in space and time: a Sr–Nd–Pb–Hf study of the North Atlantic rifted margin. Earth and Planetary Science Letters 177, 255–271.[CrossRef][Web of Science]
Kempton, P. D., Pearce, J. A., Barry, T. L., Fitton, J. G., Langmuir, C. & Christie, D. M. (2002). Nd vs. Hf as a geochemical discriminant between Indian and Pacific MORB-source mantle domains: results from ODP Leg 187 to the Australian–Antarctic Discordance. Geolochemistry, Geophysics, Geosystems DOI 10.1029/2002GC000320.
Kerr, A. C. (1998). Oceanic plateau formation: a cause of mass extinction and black shale deposition around the Cenomanian–Turonian boundary. Journal of the Geological Society, London 155, 619–626.
Kerr, A. C. & Arndt, N. T. (2001). A note on the IUGS reclassification of the high-Mg and picritic volcanic rocks. Journal of Petrology 42, 2169–2171.
Kerr, A. C., Tarney, J., Marriner, G. F., Klaver, G. T., Saunders, A. D. & Thirlwall, M. F. (1996a). The geochemistry and petrogenesis of the late-Cretaceous picrites and basalts of Curaçao, Netherlands Antilles: a remnant of an oceanic plateau. Contributions to Mineralogy and Petrology 124, 29–43.[CrossRef][Web of Science]
Kerr, A. C., Tarney, J., Marriner, G. F., Nivia, A., Saunders, A. D. & Klaver, G. T. (1996b). The geochemistry and tectonic setting of late Cretaceous Caribbean and Colombian volcanism. Journal of South American Earth Sciences 9, 111–120.[CrossRef][Web of Science]
Kerr, A. C., Marriner, G. F., Tarney, J., Nivia, A., Saunders, A. D., Thirlwall, M. F. & Sinton, C. W. (1997a). Cretaceous basaltic terranes in western Colombia: elemental, chronological and Sr–Nd constraints on petrogenesis. Journal of Petrology 38, 677–702.[CrossRef][Web of Science]
Kerr, A. C., Tarney, J., Marriner, G. F., Nivia, A. & Saunders, A. D. (1997b). The Caribbean–Colombian Cretaceous igneous province: the internal anatomy of an oceanic plateau. In: Mahoney, J. J. & Coffin, M. (eds) Large Igneous Provinces; Continental, Oceanic and Planetary Flood Volcanism. American Geophysical Union Monograph 100, 45–93.
Kerr, A. C., Tarney, J., Nivia, A., Marriner, G. F. & Saunders, A. D. (1998). The internal structure of oceanic plateaus: inferences from obducted Cretaceous terranes in western Colombia and the Caribbean. Tectonophysics 292, 173–188.[CrossRef][Web of Science]
Kerr, A. C., Aspden, J. A., Tarney, J. & Pilatasig, L. F. (2002a). The nature and provenance of accreted oceanic terranes in western Ecuador: geochemical and tectonic constraints. Journal of the Geological Society, London 159, 577–594.
Kerr, A. C., Tarney, J., Kempton, P. D., Spadea, P., Nivia, A., Marriner, G. F. & Duncan, R. A. (2002b). Pervasive mantle plume head heterogeneity: evidence from the late Cretaceous Caribbean–Colombian oceanic plateau. Journal of Geophysical Research—Solid Earth 107, DOI 10.1029, 2001JB000790.
Kohlstedt, D. L., Keppler, H. & Rubie, D. C. (1996). Solubility of water in the alpha, beta and gamma phases of (Mg,Fe)2SiO4. Contributions to Mineralogy and Petrology 123, 345–357.[CrossRef][Web of Science]
Lapierre, H., Dupuis, V., De Lepinay, B. M., Bosch, D., Monie, P., Tardy, M., Maury, R. C., Hernandez, J., Polve, M., Yeghicheyan, D. & Cotten, J. (1999). Late Jurassic oceanic crust and Upper Cretaceous Caribbean plateau picritic basalts exposed in the Duarte igneous complex, Hispaniola. Journal of Geology 107, 193–207.[CrossRef][Web of Science]
Lapierre, H., Bosch, D., Dupuis, V., Polve, M., Maury, R. C., Hernandez, J., Monie, P., Yeghicheyan, D., Jaillard, E., Tardy, M., De Lepinay, B. M., Mamberti, M., Desmet, A., Keller, F. & Senebier, F. (2000). Multiple plume events in the genesis of the peri-Caribbean Cretaceous oceanic plateau province. Journal of Geophysical Research—Solid Earth 105, 8403–8421.
Larson, R. L. (1991). Latest pulse of Earth: evidence for a mid-Cretaceous superplume. Geology 19, 547–550.
Leake, B. E., Woolley, A. R., Arps, C. E. S., et al. (1997). Nomenclature of amphiboles: Report of the Subcommittee on Amphiboles of the International Mineralogical Association Commission on New Minerals and Mineral Names. Mineralogical Magazine 61/62, 295–321.
Luhr, J. F. (1992). Slab-derived fluids and partial melting in subduction zones; insights from two contrasting Mexican volcanoes (Colima and Ceboruco). Journal of Volcanology and Geothermal Research 54, 1–18.[CrossRef][Web of Science]
Mahoney, J. J., Natland, J. H., White, W. M., Poreda, R., Fisher, R. L. & Baxter, A. N. (1989). Isotopic and geochemical provinces of the western Indian Ocean spreading centers. Journal of Geophysical Research 94, 4033–4053.
Mahoney, J. J., Leroex, A. P., Peng, Z., Fisher, R. L. & Natland, J. H. (1992). Southwestern limits of Indian Ocean Ridge mantle and the origin of low 206Pb/204Pb mid-ocean ridge basalt: isotope systematics of the Central Southwest Indian Ridge (17°–50°E). Journal of Geophysical Research 97, 19771–19790.[CrossRef]
Mahoney, J. J., Storey, M., Duncan, R. A., Spencer, K. J. & Pringle, M. (1993). Geochemistry and age of the Ontong Java Plateau. In: Pringle, M. S. (ed.) The Mesozoic Pacific: Geology, Tectonics and Volcanism. Geophysical Monograph, American Geophysical Union 77, 233–261.
Mahoney, J. J., Sinton, J. M., MacDougall, J. D., Spencer, K. J. & Lugmair, G. W. (1994). Isotope and trace element characteristics of a super-fast spreading ridge: East Pacific Rise, 13–23°S. Earth and Planetary Science Letters 121, 173–193.[CrossRef][Web of Science]
Mamberti, M., Lapierre, H., Bosch, D., Jaillard, E., Ethien, R., Hernandez, J. & Polve, M. (2003). Accreted fragments of the Late Cretaceous Caribbean–Colombian Plateau in Ecuador. Lithos 66, 173–199.[CrossRef][Web of Science]
McDougall, I. & Harrison, T. M. (1988). Geochronology and Thermochronology by the 40Ar/39Ar Method. New York: Oxford University Press.
McKenzie, D. P. & Bickle, M. J. (1988). The volume and composition of melt generated by extension of the lithosphere. Journal of Petrology 29, 625–679.
Merzbacher, C. & Eggler, D. H. (1984). A magmatic geohygrometer; application to Mount St. Helens and other dacitic magmas. Geology 12, 587–590.
Michael, P. (1995). Regionally distinctive sources of MORB: evidence from trace elements and H2O. Earth and Planetary Science Letters 131, 301–320.[CrossRef][Web of Science]
Michael, P. (1999). Implications for magmatic processes at Ontong Java Plateau from volatile and major element contents of Cretaceous basalt glasses. Geochemistry, Geophysics, Geosystems 1, paper number 1999GC000025.
Millward, D. (1983). The Vijes felsites: tectonically controlled intrusions associated with the Romeral–Cauca fault system Valle. Bogota: INGEOMINAS.
Millward, D., Marriner, G. F. & Saunders, A. D. (1984). Cretaceous tholeiitic volcanic rocks from the Western Cordillera of Colombia. Journal of the Geological Society, London 141, 847–860.
Nakanishi, M. & Winterer, E. L. (1998). Tectonic history of the Pacific–Farallon–Phoenix triple junction from late Jurassic to early Cretaceous: an abandoned Mesozoic spreading system in the central Pacific basin. Journal of Geophysical Research—Solid Earth 103, 12453–12468.[CrossRef]
Naney, M. T. (1983). Phase equilibria of rock-forming ferromagnesian silicates in granitic systems. American Journal of Science 283, 993–1033.
Nichols, A. R. L., Carroll, M. R. & Hoskuldsson, A. (2002). Is the Iceland hot spot also wet? Evidence from the water contents of undegassed submarine and subglacial pillow basalts. Earth and Planetary Science Letters 202, 77–87.[CrossRef][Web of Science]
Nielsen, R. L. (1988). A model for the simulation of combined major and trace element liquid lines of descent. Contributions to Mineralogy and Petrology 52, 27–38.
Nielsen, R. L. (1990). Simulation of igneous differentiation processes. In: Nicholls, J. & Russell, J. K. (eds) Modern Methods of Igneous Petrology; Understanding Magmatic Processes. Mineralogical Society of Ameria, Reviews in Mineralogy 24, 65–105.
Nivia, A. (1996). The Bolívar mafic–ultramafic complex, SW Colombia: the base of an obducted oceanic plateau. Journal of South American Earth Sciences 9, 59–68.[CrossRef][Web of Science]
Nivia, A., Galvis, N. & Maya, M. (1992). Geologia de la Plancha 242—Zarzal. Bogota: INGEOMINAS.
Nowell, G. M., Kempton, P. D. & Noble, S. R. (1998). High precision Hf isotope measurements of MORB and OIB by thermal ionisation mass spectrometry: insights into the depleted mantle. Chemical Geology 149, 211–233.[CrossRef][Web of Science]
Ohtani, E., Mizobata, H., Kudoh, Y., Nagase, T., Arashi, H., Yurimoto, H. & Miyagi, I. (1997). A new hydrous silicate, a water reservoir, in the upper part of the lower mantle. Geophysical Research Letters 24, 1047–1050.[CrossRef][Web of Science]
Regelous, M., Niu, Y., Wendt, J. I., Batiza, R., Greig, A. & Collerson, K. D. (1999). Variations in the geochemistry of magmatism on the East Pacific Rise at 10°30'N since 800 ka. Earth and Planetary Science Letters 168, 45–63.[CrossRef][Web of Science]
Renne, P. R., Swisher, C. C., Deino, A. L., Karner, D. B., Owens, T. L. & Depaolo, D. J. (1998). Intercalibration of standards, absolute ages and uncertainties in Ar-40/Ar-39 dating. Chemical Geology 145, 117–152.[CrossRef][Web of Science]
Révillon, S., Arndt, N. T., Chauvel, C. & Hallot, E. (2000). Geochemical study of ultramafic volcanic and plutonic rocks from Gorgona Island, Colombia: the plumbing system of an oceanic plateau. Journal of Petrology 41, 1127–1153.
Révillon, S., Chauvel, C., Arndt, N. T., Pik, R., Martineau, F., Fourcade, S. & Marty, B. (2002). Heterogeneity of the Caribbean plateau mantle source: heterogeneity of the Caribbean plateau mantle source: Sr, O and He isotopic compositions of olivine and clinopyroxene from Gorgona Island. Earth and Planetary Science Letters 205, 91–106.[CrossRef][Web of Science]
Richards, M. A., Jones, D. L., Duncan, R. A. & DePaolo, D. J. (1991). A mantle plume initiation model for the formation of Wrangellia and other oceanic flood basalt plateaus. Science 254, 263–267.
Royse, K. R., Kempton, P. D. & Darbyshire, D. P. F. (1998). Procedure for the Analysis for Rubidium–Strontium and Samarium–Neodymium Isotopes at the NERC Isotope Geosciences Laboratory. Keyworth: NERC Isotope Geoscience Laboratories, 28 pp.
Salters, V. J. M. & Zindler, A. (1995). Extreme 176Hf/177Hf in the sub-oceanic mantle. Earth and Planetary Science Letters, 129, 13–30.[CrossRef][Web of Science]
Simons, K., Dixon, J., Schilling, J.-G., Kingsley, R. & Poreda, R. (2002). Volatiles in basaltic glasses form the Easter–Salas y Gomez seamount chain and Easter Microplate: implications for geochemical cycling of volatile elements. Geochemistry, Geophysics, Geosystems 3, DOI 10.1029/2001GC000173.
Sinton, C. W. & Duncan, R. A. (1997). Potential links between ocean plateau volcanism and global ocean anoxia at the Cenomanian–Turonian boundary. Economic Geology 92, 836–842.
Sinton, C. W., Duncan, R. A., Storey, M., Lewis, J. & Estrada, J. J. (1998). An oceanic flood basalt province within the Caribbean plate. Earth and Planetary Science Letters 155, 221–235.[CrossRef][Web of Science]
Strothers, R. B., Wolff, J. A., Self, S. & Rampino, M. R. (1986). Basaltic fissure eruptions, plume heights, and atmospheric aerosols. Geophysical Research Letters 13, 725–728.[Web of Science]
Sun, S.-S. & McDonough, W. F. (1989). Chemical and isotope systematics of oceanic basalts: implications for mantle composition and processes. In: Saunders, A. D. & Norry, M. J. (eds) Magmatism in the Ocean Basin. Geological Society, London, Special Publications 42, 313–345.
Tarney, J. & Marsh, N. G. (1991). Major and trace element geochemistry of Holes CY-1 and CY-4: implications for petrogenetic models. In: Gibson, I. L., Malpas, J., Robinson, P. T. & Xenophontos, C. (eds) Initial Reports Cyprus Crustal Study: Geological Survey of Canada 90-20, 133–176.
Thompson, A. B. (1992). Water in the Earth's upper mantle. Nature 358, 295–302.[CrossRef]
Thompson, P. M. E. (2002). Petrology and geochronology of an arc sequence, Bonaire, Dutch Antilles, and its relationship to the Caribbean Plateau. Ph.D. thesis, University of Leicester, 189 pp.
Thompson, P. M. E., Kempton, P. D., White, R. V., Kerr, A. C., Tarney, J., Saunders, A. D., Fitton, J. G. & McBirney, A. (2004). Hf–Nd isotope constraints on the origin of the Cretaceous Caribbean plateau and its relationship to the Galapagos plume. Earth and Planetary Science Letters 217, 59–75.[CrossRef][Web of Science]
Todt, W., Cliff, R. A., Hanser, A. & Hofmann, A. W. (1996). Evaluation of 202Pb–205Pb double spike for high-precision lead isotope analysis. In: Hart, S. R. & Basu, A. (eds) Earth Processes: Reading the Isotopic Code. Geophysical Monograph, American Geophysical Union 95, 429–437.
Turner, S., Hawkesworth, C., Gallagher, K., Stewart, K., Peate, D. & Mantovani, M. (1996). Mantle plumes, flood basalts, and thermal models for melt generation beneath continents: assessment of a conductive heating model and application to the Parana. Journal of Geophysical Research—Solid Earth 101, 11503–11518.[CrossRef]
Vogt, P. R. (1989). Volcanogenic upwelling of anoxic, nutrient-rich water: a possible factor in carbonate-bank/reef demise and benthic faunal extinctions. Geological Society of America Bulletin 101, 1225–1245.
Walker, R. J., Storey, M. J., Kerr, A. C., Tarney, J. & Arndt, N. T. (1999). Implications of 187Os isotopic heterogeneities in a mantle plume: evidence from Gorgona Island and Curaçao. Geochimica et Cosmochimica Acta 63, 713–728.[CrossRef][Web of Science]
Wallace, P. J. (1998). Water and partial melting in mantle plumes: inferences from the dissolved H2O concentrations of Hawaiian basaltic magmas. Geophysical Research Letters 25, 3639–3642.[CrossRef][Web of Science]
Wallace, P. J. (2002). Volatiles in submarine basaltic glasses from the Northern Kerguelen Plateau (ODP Site 1140): implications for source region compositions, magmatic processes, and plateau subsidence. Journal of Petrology 43, 1311–1326.
Wignall, P. B. (2001). Large igneous provinces and mass extinctions. Earth-Science Reviews 53, 1–33.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. A. BRYANT, G. M. YOGODZINSKI, M. L. HALL, J. L. LEWICKI, and D. G. BAILEY Geochemical Constraints on the Origin of Volcanic Rocks from the Andean Northern Volcanic Zone, Ecuador J. Petrology, June 1, 2006; 47(6): 1147 - 1175. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||