Journal of Petrology Advance Access originally published online on April 29, 2005
Journal of Petrology 2005 46(10):1963-1996; doi:10.1093/petrology/egi044
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Geochronology and Petrogenesis of the Cretaceous Antampombato–Ambatovy Complex and Associated Dyke Swarm, Madagascar
1 DIPARTIMENTO DI SCIENZE DELLA TERRA, UNIVERSITÀ DI NAPOLI FEDERICO II, VIA MEZZOCANNONE 8, 80134 NAPOLI, ITALY
2 DIPARTIMENTO DI SCIENZE DELLA TERRA, UNIVERSITÀ DI FIRENZE, VIA G. LA PIRA 4, 50121 FIRENZE, ITALY
3 DIPARTIMENTO DI SCIENZE DELLA TERRA, UNIVERSITÀ DI PAVIA AND CNR–CENTRO DI GEOSCIENZE E GEORISORSE, VIA FERRATA 1, 27100 PAVIA, ITALY
4 COLLEGE OF OCEANIC AND ATMOSPHERIC SCIENCES, OREGON STATE UNIVERSITY, CORVALLIS, OR 97331-5503, USA
RECEIVED JULY 30, 2004; ACCEPTED MARCH 23, 2005
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONThe Antampombato–Ambatovy complex is the largest intrusion in the central–eastern part of the Cretaceous flood basalt province of Madagascar, with an exposed surface area of about 80 km2. It has an 40Ar/39Ar incremental heating age of 89·9 ± 0·4 Ma and a U–Pb age of 90 ± 2 Ma. The outcropping plutonic rocks range from dunite and wehrlite, through clinopyroxenite and gabbro, to sodic syenite. A dyke swarm cross-cutting some of the above lithologies (and the nearby Precambrian basement rocks) is formed of picritic basalts, alkali to transitional basalts, benmoreites and rhyolites; some of the latter are peralkaline. A few basaltic dykes have cumulate olivine textures, with up to 26 wt % MgO and 1200 ppm Ni, whereas others have characteristics more akin to those of primitive liquids (9 wt % MgO; Mg-number 0·61; 500 ppm Cr; 200 ppm Ni). These basalts have relatively high TiO2 (2·2 wt %) and total iron (14 wt % as Fe2O3), and moderate contents of Nb (10–11 ppm) and Zr (c. 100 ppm). Initial (at 90 Ma) Sr- and Nd-isotope ratios of the clinopyroxenites and basalt dykes are 0·7030–0·7037 and 0·51290–0·51283, respectively. Syenites and peralkaline rhyolites have Sr- and Nd-isotope ratios of 0·7037–0·7039 and 0·51271–0·51274, respectively. The data suggest derivation of the parental magmas from a time-integrated depleted mantle source, combined with small amounts of crustal contamination in the petrogenesis of the more evolved magmas. The isotopic compositions of the mafic–ultramafic rocks are most similar to those of the mid-ocean ridge basalt (MORB)-like igneous rocks of eastern Madagascar, and suggest the existence of an isotopically ‘depleted’ component in the source of the entire Madagascar province, even though the Antampombato basalts are chemically unlike the lavas and dykes with the same depleted isotopic signature found in western Madagascar. If this depleted component is plume-related, this suggests that the plume has a broadly MORB-source mantle composition. The existence of isotopically more enriched magma types in the Madagascan province has several possible petrogenetic explanations, one of which could be the interaction of plume-related melts with the deep lithospheric mantle beneath the island.
KEY WORDS: geochronology; flood basalts; Antampombato–Ambatovy intrusion; Cretaceous; Madagascar
|
INTRODUCTION |
|---|
The Late Cretaceous flood basalt province of Madagascar is one of the largest associated with Gondwana breakup. It covers most of the sedimentary basins around the island, and a few areas of Precambrian basement. It was erupted from 92 to 84 Ma (Storey et al., 1995
), and is, therefore, much younger than Africa–Madagascar separation, but slightly older than India–Madagascar separation and the opening of the Mascarene Basin (Storey et al., 1995). A number of researchers have suggested a genetic link between this province and the Marion hotspot track in the southwestern Indian Ocean (see Storey et al., 1995, 1997; Torsvik et al., 1998).
A review of the chemical and isotopic characteristics of the various basalt types in the northern part of the province has been published recently by Melluso et al. (2003), who recognized five compositional groups [western subprovince: low-Ti–Nb + low-Nb–high Ti basalts in the Antanimena region, transitional (TR) and tholeiitic (THO) basalts in the Mailaka section; eastern subprovince: low La/Yb and high La/Yb basalts] and pointed out major chemical differences between the basalts erupted in the two subprovinces. We refer to these chemical subdivisions in subsequent discussions. Further information on the characteristics of the igneous rocks from other parts of the province has been given by Mahoney et al. (1991), Dostal et al. (1992), Storey et al. (1997) and Bardintzeff et al. (2001). Dykes and rare lava flows in southern India (Torsvik et al., 2000; Anil Kumar et al., 2001; Pande et al., 2001) are the age (and approximate chemical) equivalents of the Madagascan basalts.
The lava flows and dykes throughout the province have been the subject of detailed studies in recent years. On the other hand, numerous intrusions, which appear to be approximately the same age as the volcanic rocks, and probably were the locus for intense magma differentiation, as well as the conduits for eruption of a significant part of the flood basalts, are relatively or completely unknown from a petrogenetic point of view. This is mostly because of the very difficult access. Most of the intrusions are located in northwestern Madagascar towards Cap St. André, an area far from the major roads; other intrusions are covered by dense forest. A number of igneous intrusions occur along the eastern coast, of which the Analalava complex (Fig. 1) is probably the largest. However, apart from one U/Pb age determination (Torsvik et al., 1998), there is nothing in the recent literature about the petrogenesis of this complex. The Antampombato–Ambatovy intrusion, which is the focus of this study, lies some 500 km to the south.
|
In this study we present high-resolution 40Ar/39Ar and U/Pb age determinations and mineral and whole-rock chemical and Nd–Sr isotopic data, aimed at addressing the petrogenesis of this intrusion in the framework of the volcanism of Madagascar. The intrusion is located roughly midway between the cities of Tananarive and Tamatave (Fig. 1). Many aspects make the intrusion interesting: (1) its age is unknown; some geological maps even suggest a Cenozoic (sensu lato) age, as a result of the lack of stratigraphic control; (2) according to the available geological maps, the rocks range from ultramafic to felsic, thus covering a complete spectrum of magma types, and perhaps represent the root of a volcanic complex, emplaced directly above the Precambrian basement; (3) ultramafic compositions with unmodified mantle-derived isotopic signatures are expected to be present; (4) the intrusion is not far from both the Tamatave dyke swarm and the Mananjary lava field (Storetvedt et al., 1992
; Storey et al., 1997; Melluso et al., 2002, 2003). Geochemical data for a few basalt dykes cross-cutting the Antampombato–Ambatovy intrusion were previously reported by Melluso et al. (2003). |
GEOLOGICAL SETTING AND SAMPLING STRATEGY |
|---|
The Antampombato–Ambatovy intrusion (hereafter Antampombato) is located not far from the eastern shoulder of the north–south-trending Moramanga rift zone, which developed in the late Cenozoic in response to extensional tectonics, and is associated with rare, strongly alkaline volcanism in its northern part (Alaotra lake; Fig. 1). The intrusion is formed of two outcrops of variably serpentinized ultramafic rocks (Ambatovy and Analamay), and of a third outcrop of syenitic rocks (Antampombato), all surrounded by gabbroic rocks (Fig. 2). The contacts between the various lithotypes are usually completely hidden by heavy vegetation and laterites, hence the actual emplacement sequence is uncertain. Therefore, we did not attempt detailed mapping of the various lithological units. The gabbros and ultramafic rocks and the Precambrian basement adjacent to the intrusion are cross-cut by dykes, which are particularly well exposed in the northwestern part of the intrusion. Individual dykes are up to 2 m wide, and can be followed for a few hundred metres. The Precambrian basement surrounding the intrusion is formed of orthogneisses, two-pyroxene–amphibole mafic granulites, and rare marbles [grouped together in the 700–2500 Ma Antananarivo block (Collins et al., 2003
); see also the review of de Wit (2003) for a detailed description of the Precambrian basement of the island]. The geological and structural setting of the Antampombato intrusion is summarized in the 1:100 000 Geological Map of Madagascar (sheets Moramanga–Lakato R47–S47 and Mandialaza–Fierenana R46–S46, 1962).
|
We collected 46 samples from those areas free of thick lateritic soil, primarily from streams or in small quarries, particularly from the areas with ultramafic rocks. We also obtained additional samples from the Precambrian basement underneath the intrusion (MIN2B, a felsic gneiss), or wall-rocks adjacent to dykes in the northwestern part of the intrusion (M490, another felsic gneiss, and M491, a two-pyroxene–amphibole mafic granulite; Table 1).
|
|
ANALYTICAL TECHNIQUES |
|---|
Bulk-rock and mineral chemistry
Clean rock chips, obtained using a steel jaw crusher and washed in distilled water, were ground in a low-blank agate mortar. Major and some trace elements (Sc, V, Cr, Ni, Cu, Zn, Rb, Sr, Y, Zr, Nb, Ba) were analysed by X-ray fluorescence (XRF) in Naples using a Philips PW1400 instrument, following the methods described by Melluso et al. (1997
, 2001, 2002, 2003); the data are reported in Table 1. Analytical precision is estimated to be within 1% for SiO2, TiO2, Al2O3, Fe2O3 and CaO, better than 6% for K2O, ±0·03 wt % for MnO and P2O5, generally better than 5–10% for all trace elements in the observed compositional ranges, and ±1 ppm for Rb and Nb below 10 ppm. No analytical bias was observed between these and the XRF analyses reported in our earlier studies. Na and Mg were analysed by atomic absorption spectrophotometry. Typical precision is better than 2% for Mg and better than 6% for Na. LOI (weight loss on ignition) was measured gravimetrically. Rare earth elements (REE) and other trace elements in Table 2 were analysed for a subset of samples by inductively coupled plasma–mass spectrometry (ICP-MS) at CRPG Nancy (France) and Pisa University (Italy). Mineral compositions (Tables 3–8) were analysed at the CNR laboratory of Padova, utilizing a CAMECA Camebax electron microprobe equipped with four wavelength-dispersive spectrometers and one energy-dispersive spectrometer. Silicates and oxides were used as standards. An acceleration voltage of 20 kV and beam current of 20 mA were used.
|
|
|
|
|
|
|
40Ar/39Ar dating
Feldspar separates were first washed in a mild acid solution (HF, HNO3), followed by ultrasonic cleaning in water. Samples were then wrapped in Cu foil, loaded into evacuated quartz tubes and irradiated for 6 h in the 1 MW TRIGA experimental reactor at Oregon State University. We monitored the neutron fluence of the reactor with the Fish Canyon Tuff biotite standard (28·03 ± 0·16 Ma, Renne et al., 1998
), spaced at regular intervals among the sample unknowns. Errors in monitor measurements and gradient fitting accumulated to about 0·5%. Following irradiation and initial decay of short-lived radionuclides, we loaded samples in a high-vacuum, gas extraction line, and subsequently heated each sample in increments from 400° to 1400°C. Gases released at each temperature increment were passed over hot Zr–Al getters to remove active gases, and admitted to a MAP-215/50 mass spectrometer for analysis of Ar isotopic compositions. The Ar data were acquired in a peak-hopping mode (for m/z = 35, 36, 37, 38, 39, 40) by computer (Koppers, 2002). Peak decay was linear and typically <10% over 12 sets of peaks and backgrounds. Mass discrimination on the MAP system was measured with zero age samples run in the same way as the samples, and was constant at 1·005 (for 2 a.m.u.). The sensitivity of the mass spectrometer is 4 x 10–14 mol/V and measured backgrounds were 1·5 x 10–18 mol at m/z = 36, 2 x 10–18 mol at m/z = 39 and 1·5 x 10–6 mol at m/z = 40. Procedure blanks for the resistance furnace ranged from 3·0 x 10–18 mol 36Ar and 9·0 x 1016 mol 40Ar at 600°C to 6·4 x 10–18 mol 36Ar and 1·9 x 10–15 mol 40Ar at 1400°C. Plateau and isochron ages are reported in Table 9.
|
U/Pb dating
Zircons were prepared as handpicked mineral separates from syenite M481, obtained using heavy liquids and a Frantz magnetic separator. About 25 grains devoid of cracks and inclusions were selected by further hand-picking. Zircons were mounted in epoxy and polished down to 20 µm. They are colourless to light yellow and translucent; in most cases they represent fragments of larger crystals. A CITL Mk3 cathodoluminescence system at the CNR–Istituto di Geoscienze e Georisorse, Sezione di Pavia was used to investigate the structure and morphology of the zircons. The samples revealed the presence of a thin optically brighter rim around a darker core, but no inherited cores were observed. Locations for analysis were selected to avoid fractures and the bright rim. Zircon geochronology was carried out by laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) at the CNR–Istituto di Geoscienze e Georisorse, Sezione di Pavia, following the method developed by Tiepolo (2003)
. The instrument couples a magnetic sector ICP-MS system with a Nd:YAG laser probe working at 213 nm. The signals of masses 202Hg, 204(Pb + Hg), 206Pb, 207Pb, 208Pb, 232Th, 235U and 238U were acquired in time-resolved mode. The laser was operated at a repetition rate of 10 Hz with a pulse energy of about 0·15 mJ and the spot diameter was 20 µm. U–Pb fractionation and mass bias effects were corrected by external standardization using zircon 91500 (Wiedenbeck et al., 1995). Each analytical run consisted of 20 analyses including four standards at the beginning and four standards at the end of the run. In each run zircon 02123 (Ketchum et al., 2001) was also analysed for quality control. Data reduction and age calculation was performed using the software package LamTrace (developed by S. Jackson initially at Memorial University of Newfoundland and later at Macquarie University, Sydney) using the measured 207Pb/235U ratio. Average ages were calculated using the Isoplot/Ex program (Ludwig, 2000); the data are reported in Table 10.
|
Nd–Sr isotopes
Sr and Nd isotope analyses were performed at the Department of Earth Sciences, University of Firenze. Sample powder (20 mg) was dissolved in a HF–HNO3–HCl mixture. Sr and Nd fractions were separated following standard chromatographic techniques using AG50x8 and PTFE–HDEHP resins with HCl as eluent. The total procedural blank was <200 pg for Sr and <100 pg for Nd, making blank correction negligible. Mass spectrometric analyses were performed on a Thermo Finnigan Triton-Ti thermal ionization mass spectrometer equipped with nine movable collectors. Sr and Nd isotope compositions were measured in dynamic mode and are reported normalized to 86Sr/88Sr = 0·1194 and 146Nd/144Nd = 0·7219, respectively. Exponential-law mass fractionation correction was used for all Sr and Nd isotopic data. Uncertainties in measured (m) and initial (0) isotopic ratios refer to the least significant digits and represent ±2
run precision and ±2 propagated error, respectively. The external precision of NIST SRM987 was 87Sr/86Sr = 0·710251 ± 10 (2, n = 39), and that of the La Jolla standard was 143Nd/144Nd = 0·511845 ± 4 (2, n = 11). The data are reported in Table 11.
|
|
CLASSIFICATION OF THE SAMPLES, PETROGRAPHY, AND MINERAL COMPOSITIONS |
|---|
Plutonic rocks
The Antampombato complex includes two cores of ultramafic rocks, of dunite (olivine + Cr-spinel ± rare cpx), wehrlite (olivine + cpx ± spinel) and olivine clinopyroxenite (cpx ± ol ± spinel). Dunites have cumulus olivine, sometimes with polygonal boundaries, and interstitial spinel; they are variably serpentinized. Clinopyroxene is absent in the dunites (Fig. 3a), but is a cumulus phase in the wehrlites (Fig. 3b) and clinopyroxenites (Fig. 3c). In the clinopyroxenites, olivine is rare and interstitial. Small veins or layers rich in plagioclase ± olivine ± clinopyroxene (gabbroic sensu lato) sometimes cross-cut the ultramafic rocks.
|
The ultramafic rocks are surrounded by, and grade into cumulate olivine gabbros and gabbros (pl + cpx ± ol ± amph ± Fe–Ti oxides). In these rocks plagioclase and clinopyroxene are the main cumulus minerals, whereas olivine is a cumulus phase only in some Mg-rich samples; it is more often found interstitially (Fig. 3d and e). Igneous layering is frequently observed, with plagioclase-rich layers alternating with clinopyroxene-rich layers. Other minerals such as amphibole and oxides are mostly confined to the interstices between the cumulus crystals or as rims on clinopyroxene.
In the central–eastern part of the complex, a medium- to fine-grained (0·2–2 mm crystal size) leucogabbro (M476) is formed of plagioclase and poikilitic to euhedral amphibole, together with minor clinopyroxene with amphibole rims, and Fe–Ti oxides (Fig. 3f). Large outcrops of coarse syenite (Na-plag + cpx + amph ± opx ± qtz ± Fe–Ti oxides) form the western part of the intrusion. Biotite, titanite, apatite, and zircon are accessory minerals (Fig. 3g).
Dykes
The dykes were classified using the total alkalis–silica (TAS) diagram of Le Bas et al. (1986; Fig. 4) based upon their major element chemistry. Different petrographic types of basaltic dykes can be recognized. The most Mg-rich dykes (M486 and M495) are strongly porphyritic with large olivine phenocrysts set in an altered groundmass rich in secondary amphibole. Plagioclase is not optically visible or is a microlitic phase. Rare clinopyroxenite inclusions, less than 2 cm in size, have been found (Table 1). Samples M497a and M497b, parts of the same dyke collected at different places, are typical olivine (± clinopyroxene)-phyric basalts, with plagioclase, clinopyroxene, olivine and Fe–Ti oxides in the groundmass (Fig. 3h). Chromium spinel is found as an inclusion in olivine. The less Mg-rich basalt dykes are mostly aphyric, often with doleritic textures and sometimes completely aphanitic, with only plagioclase and/or clinopyroxene microphenocrysts (Fig. 3i and j). The benmoreite dyke (M475) is a fine-grained rock, with rare plagioclase phenocrysts, in a holocrystalline groundmass with plagioclase, amphibole, rare clinopyroxene, and Fe–Ti oxides (Fig. 3k). The rhyolitic dykes are aphyric, with rare, altered, feldspar microphenocrysts, and feldspar, quartz and altered mafic minerals in the groundmass (Fig. 3l). The rhyolites cross-cut the gabbros in the northern part of the intrusion (sample M494) or the Precambrian basement in the northwestern part of the intrusion (sample M487).
|
Mineral chemistry
A summary of the range of mineral chemical variations in the Antampombato rocks is given in Fig. 5 and Tables 3–8.
|
Olivine has a fairly restricted range of composition in the dunites and wehrlites (Fo87–80) and in the gabbros (Fo78–76), whereas that in the basalts overlaps both ranges (Fo87–75) (Tables 3 and 8; Fig. 5a). These compositions should have been in equilibrium with magmas having Mg/(Mg + Fe2+) ratios between 0·69 and 0·5 (based on KdFe/Mg = 0·33; Roeder & Emslie, 1970
); values in the lower part of the range of Mg/(Mg + Fe2+) are observed in the associated basalt dykes (0·85–0·48; Table 1). This feature indicates that the most Mg-rich dykes have excess olivine phenocrysts. Chromium-bearing spinels (Table 4) exhibit a wide range of compositions that peak towards Cr-rich compositions in the Al–Cr–Fe2+ + Ti diagram (Fig. 5b). The compositions broadly overlap the range in the Madagascan basalts (Mailaka).
The clinopyroxenes in the Antampombato rocks are diopsides, typical of mildly alkaline suites, with high CaO contents (Fig. 5c; Table 5) and Ti enrichment (up to 2·3 wt % TiO2) in the intermediate compositions (salites). Diopside–hedenbergite compositions are also observed in the transitional, mid-ocean ridge basalt (MORB)-like, basalts of the western part of the province (Melluso et al. 2001). The composition of the clinopyroxene in the basaltic dykes fully overlaps with that observed in the ultramafic and gabbroic rocks. The Antampombato syenites, which are silica-saturated or slightly oversaturated rocks, contain augite and rare Fe-rich orthopyroxene (Table 5).
Plagioclase is the dominant feldspar throughout the complex. The composition of this phase in the gabbros and basalt dykes completely overlaps (Fig. 5d; Table 6). Extremely sodic plagioclase in the evolved samples shows no sign of subsolidus re-equilibration.
Amphiboles are a common phase and also have a very wide range of composition, from kaersutite in the gabbros, typical of alkaline magmas, to Fe-edenite in the syenites (Table 7). The presence of hydrous minerals in the Antampombato gabbros is an unusual feature in the Madagascan Cretaceous igneous province, which is otherwise characterized by completely anhydrous parageneses (e.g. Melluso et al., 2001).
Ilmenites have no Al2O3, and very low MgO (<0·13 wt %), whereas Ti-magnetites range in ulvöspinel content from 1·5 to 74 mol %, and are low in Al2O3 (<1 wt %) (Table 4). The equilibrium magnetite–ilmenite pairs give maximum T and log fO2 of 1064°C and –10·67, respectively, plotting on, or very close to, the synthetic quartz–fayalite–magnetite (QFM) buffer. Those ilmenite–magnetite pairs that re-equilibrated in the subsolidus plot towards the Ni–NiO (NNO) buffer or slightly above it, suggesting exchange with external oxidizing fluids (see Fig. 5e).
Biotite is found in a few syenites as an Fe-rich variety.
|
RADIOMETRIC AGES |
|---|
Two 40Ar/39Ar age determinations have been performed, using the methods and facilities described by Duncan et al. (1997)
, on calcic plagioclase from gabbro M471 and sodic plagioclase from syenite M481 (Table 9; Fig. 6). We calculated ages in three ways. Individual temperature steps were plotted against temperature (i.e. per cent gas released) in age spectra plots. A sequence of consecutive, concordant step ages, called a plateau, may indicate the crystallization age of the rock. No such plateau age was observed for M471, but a four-step plateau age of 89·9 ± 0·4 Ma (weighted mean of the steps, 2 error) was found for M481. Next, step isotopic compositions (40Ar/36Ar vs 39Ar/36Ar) were plotted and examined for collinearity, for which the slope is proportional to age (isochron age). This is especially informative in the case of plutonic rocks, such as the gabbros, that may have crystallized without equilibration with atmospheric Ar (40Ar/36Aratm = 295·5). Initial 40Ar/36Ar may thus have been significantly greater than the atmospheric value (assumed in calculating the step ages and plateaux). In sample M471 five step compositions (700°–1100°C) are collinear, providing an isochron age of 88·6 ± 27·8 Ma. This age, although imprecise, is consistent with the plateau age of M481. The isochron age for M481 is, likewise, consistent at 89·8 ± 0·6 Ma. Finally, all step ages for a given sample can be combined to provide a total fusion age, which is comparable with a conventional K–Ar age. The total fusion ages, 566·8 ± 4·7 Ma for M471 and 97·4 ± 0·4 Ma for M481, are very different and indicate the presence of undegassed initial Ar (‘excess’). In summary, the most reliable ages are the isochron ages because they do not require any assumption about the initial composition of Ar in the feldspars. In the case of M471 there are apparently significant amounts of ‘excess Ar’, whereas for M481 there is much less, leading to a plateau age concordant with the isochron age at 89–90 Ma (Fig. 6).
|
In situ Pb isotope geochronology of zircons from syenite sample M481 has been carried out by LA-ICP-MS. U–Pb ratios and ages for M481 zircons are reported in Table 9. On a conventional concordia diagram (not shown), the data are sub-concordant and dispersed roughly along a line parallel to the x-axis, which might reflect a common-Pb contribution in the analysed samples. The common-Pb content in zircons is, however, extremely small and not distinguishable from the signal of 204Pb. To achieve a common-Pb corrected age, the regression of total 207Pb/206Pb against total 238U/206Pb for the collective data has been performed and results have been plotted on a Tera–Wasserburg concordia diagram (Fig. 7a). This procedure, according to Compston et al. (1992)
, yields an estimate of the age without explicit correction for common-Pb. The linear regression of the data gives a concordia intersection age of 89 ± 6 Ma (2). A cumulative probability diagram (Fig. 7b) of the 206Pb/238U ages reveals a normal distribution for the selected samples, confirming that they represent a single population with a weighted mean age of 90 ± 2 Ma (2, n = 17), which is statistically equivalent to the determined intercept age. This is the second U/Pb age determination of the Madagascan volcanics, and compares very well with that obtained by Torsvik et al. (1998) on a gabbronorite from the Analalava complex (91·6 ± 0·3 Ma).
|
The age determinations with the two different techniques are identical within error, and confirm that the Antampombato complex is contemporaneous with the Late Cretaceous volcanic activity.
|
WHOLE-ROCK GEOCHEMISTRY |
|---|
Ultramafic rocks
Dunites, wehrlites and olivine clinopyroxenites are characterized by high whole-rock Mg numbers [molar Mg/(Mg + Fe); 0·88–0·81], and by variable TiO2 (0·05–0·77 wt %), Ni (2200–558 ppm), Cr (1440–4095 ppm), Cu (0–189 ppm), V (21–243 ppm) and Sc (3–80 ppm) contents (Table 1, Fig. 8). The Ni content decreases in the olivine clinopyroxenites, as a result of the decreasing content of olivine, whereas TiO2, V and, particularly, Sc, increase with clinopyroxene content. Cr contents are high and scattered, because of the simultaneous partitioning of this element in spinel (for the most part) and clinopyroxene. Large ion lithophile elements (LILE) and high field strength elements (HFSE) are always low to very low (e.g. Nb 0–3 ppm, Zr <30 ppm, Sr <34 ppm). The REE are below the detection limits of ICP-MS in the dunites, below 2·1 times chondrite in wehrlites and no more than 9·1 times chondrite in the olivine clinopyroxenite M456 (Table 10; Fig. 9).
|
|
Gabbroic rocks
Gabbros are characterized by significantly lower Mg numbers (0·80–0·71), Cr (843–210 ppm), Ni (437–211 ppm) and higher V (128–371 ppm) than the ultramafic rocks. Scandium is still relatively high (34–61 ppm), but has a smaller range of variation than in the ultramafic rocks. Higher contents of TiO2 (0·42–1·44 wt %), Sr (199–301 ppm), Zr (17–74 ppm) and Nb (1–7 ppm) than in the ultramafic rocks are also observed. The gabbro M471 has no detectable Rb, very low REE contents (Fig. 9), and a small positive europium anomaly (Eu/Eu* = 1·16).
Basalt dykes
The basalt dykes have markedly sodic affinity (Na2O/K2O >1), and vary from slightly hypersthene-normative (5%) to slightly nepheline-normative (0·2–5%); therefore, they are transitional or alkali basalts. They are characterized by a wide range of MgO contents (24–6 wt %) and Mg/(Mg + Fe2+) (0·85–0·48), with many samples, particularly those with more than 20% MgO, being too olivine-rich to have been liquid compositions (see also mineral chemistry). CaO contents first increase with decreasing MgO from 7·9 to 13·1 wt %, and then decrease to 10·3 wt %. There is a smooth increase in incompatible elements from the most to the least Mg-rich compositions (e.g. Nb from 3·8 to 16·2 ppm, Zr from 52 to 135 ppm, Ba from 46 to 382 ppm, etc.), with a concomitant drop in Cr and Ni contents (from 2128 to 19 ppm and from 1206 to 40 ppm, respectively; Fig. 8). Sc, TiO2 and V reach values as high as 43 ppm, 3·29 wt %, and 611 ppm, respectively. The REE patterns show moderate light REE (LREE) enrichment [(La/Yb)n = 4·3–5·2], and no or weak positive europium anomalies (Eu/Eu* = 1·04–1·18). Ratios such as Ti/V, Zr/Y, La/Nb and Zr/Nb vary in a narrow range (26–32, 4·1–5·5, 0·9–1·1, and 7–11, respectively), strongly suggesting cogeneticity of the samples.
Felsic plutonic and subvolcanic rocks
The leucogabbro (M476) and the benmoreite dyke (M475) have low MgO (3·7–2·8 wt %), Sc and V (11–5 ppm and 76–116 ppm, respectively), and high Na2O (5·3–5·9 wt %) and K2O (1–2 wt %). Their REE patterns have roughly the same (La/Yb)n as the basalts (5·7), with Eu/Eu* close to unity. The syenites (M483, M480, M481 and M484) have even higher Na2O (7–8·7 wt %) and Ba (1320–2480 ppm), and variable Sr (77–446 ppm), Zr (82–372 ppm) and Nb (23–81 ppm) contents. They exhibit REE patterns with marked positive europium anomalies (Eu/Eu* = 1·94–2·08), evidence of accumulation of sodic plagioclase, and moderate LREE fractionation [(La/Yb)n = 7·7–8·2; Table 2]. The ferromagnesian element contents are variable, but systematically lower than in the basalts.
The rhyolite dykes M494 and M487 have less Na2O and more K2O than the syenites. Sample M487 is actually a comendite, being relatively Fe-poor (2·3 wt % total iron as Fe2O3), Al-rich (9·7 wt % Al2O3), and slightly peralkaline [A.I. = molar (Na + K)/Al = 1·28]. The rhyolites have relatively high Zr and Nb (479–752 ppm and 69–76 ppm, respectively), with moderate REE fractionation [(La/Yb)n = 6·5; Table 10] and a strong negative europium anomaly (Eu/Eu* = 0·24) in M487 (Fig. 9). The La/Nb ratio of M487 (0·95) is well within the range observed in the mafic dykes. Compared with the leucogabbro, the rhyolite also has lower contents of middle REE, a feature that may be due to very small amounts of apatite and/or titanite fractionation.
Diagrams using MgO as a differentiation index (Fig. 8) show marked inflections at 
10 wt % MgO for CaO, and Sc, and kinks at lower MgO for SiO2, TiO2, V and Sr. Elements such as Nb, Y and Zr smoothly increase towards the rhyolites, whereas Ni smoothly decreases. The kinks are clearly due to the onset of fractionation of clinopyroxene, plagioclase and Fe–Ti oxides, as expected from the observed phenocryst mineralogy. It is also evident that the magmatic evolution from the least to the most evolved basalts is dominated by clinopyroxene and olivine fractionation. Iron-rich plutonic rocks, which could be the cumulates required to drop the Ti, Fe and V concentrations in the evolved liquids, were not observed.
|
Sr–Nd ISOTOPES |
|---|
The rocks of the Antampombato intrusion have a moderate range in initial 87Sr/86Sr and 143Nd/144Nd (0·70301–0·70390 and 0·51271–0·51293, respectively), when recalculated to an age of 90 Ma consistent with the new radiometric results (Table 11; Fig. 10a). The clinopyroxenite M456, the basalt dyke M497a and the gabbro M471 have the lowest 87Sr/86Sr (0·70301–0·70305) and the highest 143Nd/144Nd (0·51292–0·51293). These values could be considered as representative of the mantle-derived mafic magmas that filled the intrusion. The other two basaltic dykes analysed (M500, M488) have higher 87Sr/86Sr (0·70336–0·70367) at roughly similar 143Nd/144Nd (0·51293–0·51290). The more evolved samples (leucogabbros, benmoreites, syenites and rhyolites) have higher 87Sr/86Sr and lower 143Nd/144Nd (0·70351–0·70390 and 0·51274–0·51271, respectively), which could be consistent with some assimilation of crustal material during magmatic differentiation (see below). The Precambrian basement sample MIN2B has 87Sr/86Sr = 0·72033 and 143Nd/144Nd = 0·511353.
|
|
MAGMATIC EVOLUTION |
|---|
The crystallization sequence of the ultramafic and mafic plutonic rocks is olivine + chromite, followed by clinopyroxene. Calcic plagioclase crystallizes and becomes a major phase subsequently. Oxides, Fe-olivine, and minor phase crystallization (mainly amphibole) complete the paragenesis. Roughly the same crystallization sequence and compositional range of the phases is observed in the mafic dykes. Taking into account the very similar isotopic and mineral chemical compositions, we consider the ultramafic plutonic and mafic dyke facies as cogenetic. The calculated liquids that could have been in equilibrium with the composition of M456 clinopyroxenite (assuming this sample to be monomineralic) are very similar to the composition of the basaltic dykes (see bold line in Fig. 9). Based on a Kdcpx/liq for Sc = 3 (Niu et al., 1996
), the clinopyroxenite should have been in equilibrium with a melt with 27 ppm Sc, a value within the range of the associated basalt dykes (21–43 ppm). The transition from primitive to evolved basalts (from M497a to M489, M500, M468 or M488) was successfully modelled with moderate degrees of fractionation (f, fraction of residual liquid, variable from 0·86 to 0·71) of assemblages with major clinopyroxene (30–43% of the crystal removed), with smaller amounts of plagioclase (23–40%), olivine (18–34%), and oxides (0–4%). Trace element contents are broadly in agreement with closed-system fractionation of the observed phases [e.g. the bulk partition coefficient for Sr (DSr) ranges from 0·44 to 0·51, DCr from 6·5 to 10, DNi from 3·4 to 4·8; DSc from 1·3 to 1·7].
The transition from basalt to benmoreite (e.g. M468 to M475) involved c. 65% crystal removal, still dominated by clinopyroxene (46%), with lower amounts of plagioclase (21·4%) and olivine (15%) and a relatively high amount of Fe–Ti oxides (18%). The benmoreite–rhyolite transition was modelled with 35% fractionation of Na-plagioclase (45%), amphibole (53%) and apatite (2%). Major element mass balance calculations, therefore, indicate that the transition from basalt to rhyolite could be accounted for by 80–85% fractionation of gabbroic or Fe-gabbroic assemblages. Assuming Nb to be the most incompatible element, the transition between the basalt M497a and the rhyolite M487 could result from c. 87% fractionation. With this calculated amount of residual liquid, Th, Rb, and Pb are shown to be enriched in the liquids more than predicted from perfectly incompatible behaviour. As is well known, these are the trace elements typically enriched during crustal contamination processes.
Fractional crystallization models using the MELTS algorithm (Ghiorso & Sack, 1995) indicate that, at the QFM oxygen buffer and low pressure under anhydrous conditions, the transition from M497a to a ‘rhyolitic’ composition (74·3 wt % SiO2) is obtained through 94·6% fractionation of olivine (7·9%), clinopyroxene (38·6%), orthopyroxene (0·69%), plagioclase (36·65%), spinel (10·32%) and apatite (0·44%). Similar results have been obtained starting from the more evolved basalt M500 to obtain a 77·7 wt % SiO2 composition (94·3% total fractionation of 3·87% olivine, 39·44% clinopyroxene, 0·54% orthopyroxene, 38·49% plagioclase, 11·49% spinel, and 0·45% apatite). Starting from the leucogabbro M476, we can obtain a ‘rhyolitic’ composition (78·5 wt % SiO2) after 87·2% fractionation of plagioclase (63·93%), clinopyroxene (13·6%), spinel (7·78%), apatite (1·57%), ilmenite (0·24%), and negligible orthopyroxene (0·02%). However, we note that the silica-rich compositions obtained in the MELTS calculations are very different from those actually observed in the most evolved rocks of this study, and generally speaking, in natural magmas (e.g. MnO is always higher than 1·3 wt %, and in some cases is as high as 3·8 wt %; Al2O3 is always lower than 4 wt %). Moreover, we note a very strong increase of total iron (as Fe2O3) in the evolved basaltic compositions obtained from MELTS before Fe–Ti oxide crystallization (up to 23 wt %), leading to a decrease in the SiO2 content of the calculated melts. Consequently, a jump in the SiO2 content after Fe–Ti oxide fractionation is noted (see Fig. 4).
The temperature range during the transition from basalt to rhyolite was estimated as 1220°C to 780–800°C from MELTS calculations. The incoming of oxide crystallization starting from M497a (or M500) is at c. 1100° C, close to the maximum T obtained by magnetite–ilmenite equilibrium pairs (1064°C; see above).
The common feature between the major element mass balance calculations and the results from MELTS modelling is the relatively high clinopyroxene/plagioclase ratio in the fractionating assemblages. This feature is more typical of mildly alkaline and alkaline series, rather than tholeiitic series. The larger amount of clinopyroxene fractionation with respect to plagioclase has already been noted in the models for the differentiation of the Mailaka transitional basalts (Melluso et al., 2001). For instance, the transition from primitive to evolved transitional basalts required 23% olivine, 46% diopside and 31% plagioclase fractionation, whereas the transition from primitive to evolved tholeiitic basaltic andesites required 44% olivine, 18% augite, and 37% plagioclase.
It is worth noting that early clinopyroxene fractionation is typical of basalts at high pressure (>5–6 kbar; e.g. Thompson, 1987). However, simple geological reasons seem to exclude a lower crustal depth of intrusion for the Antampombato complex. In a tectonically stable island, average exhumation rates of 25–27 km in 90 Myr seem unrealistically high. Additionally, geobarometric estimates based on the composition of the most Mg-rich clinopyroxenes (Nimis, 1999) indicate very low (1 kbar) pressures of crystallization. We simply think that relatively high 
of the Antampombato magmas, as inferred by the presence of amphibole, together with their slightly alkaline nature, may have had a locally significant role.
|
RHYOLITES OF MADAGASCAR |
|---|
To model the change in Nd–Sr isotope composition and trace element characteristics in the transition from basalt to rhyolite, we used a simple assimilation–fractional crystallization (AFC) model (DePaolo, 1981
) involving a crustal contaminant with the trace element and isotopic composition of the felsic gneiss MIN2B. The results indicate that the range of Nd isotopic compositions and incompatible trace element (e.g. Nb) contents requires f (fraction of residual liquid) <0·15 and r (mass assimilated/mass fractionated) <0·15; thus the evolved rhyolites can be formed by 8–10% crustal contamination combined with 80–85% crystal fractionation (Fig. 10b). Using the EC-AFC model of Spera & Bohrson (2001), the results indicate broadly closed-system fractionation in the basaltic range of compositions. Subsequent very small degrees of contamination by upper crustal material during the benmoreite–comendite stage, as the heated basement crossed the solidus temperature, is required by the model (Fig. 10b). The rhyolites (comendites) in the Antampombato complex are thus the result of large amounts of fractionation of basaltic magmas, like those that formed the mafic dykes, coupled with small amounts of crustal contamination.
A comparison between the Antampombato felsic dyke rocks and Mailaka rhyodacites, which show strong petrographic and isotopic evidence of crustal contamination by underlying peraluminous lithologies (Melluso et al., 2001) reveals some interesting features. The REE contents are similar, and characterized by a shallow slope in the heavy REE (HREE) pattern and similar LREE to HREE fractionation (Fig. 11a). The main differences are related to the marked troughs at Sr, Eu, P and Ti (typical of pantellerites and comendites), and the absence of a Nb trough in the mantle-normalized pattern of the Antampombato rhyolite (Fig. 11a and b). The very marked influence of crustal contamination in the petrogenesis of the Mailaka felsic rocks is expecially noted in a diagram of MgO–87Sr/86Sr (Fig. 11c), as they plot near samples of the Precambrian basement, whereas the Antampombato rhyolites have isotopic composition much closer to that of the spatially associated basalts. It is clear that both Antampombato and Mailaka rhyolites are mainly the product of prolonged crystal fractionation from basaltic parents [see above and Melluso et al. (2001)], but more detailed petrological information is required on the significant amount of rhyolites found in the Mananjary area [these latter are completely unknown from a petrological point of view; see also Besairie (1964)] and in the Androy complex (Storey et al., 1997; Mahoney et al., in preparation), to better clarify the role that peraluminous or non-peraluminous contaminants may have in modifying the composition of evolved magmas of the Madagascan province. Finally, we recall that chemical correlations between rhyolitic units of Madagascar and those cropping out in southwestern India with roughly the same age (Torsvik et al., 2000; Pande et al., 2001) are a fascinating topic for future detailed work.
|
|
REGIONAL SIGNIFICANCE AND PETROGENESIS OF THE ANTAMPOMBATO MAFIC DYKES |
|---|
The mildly enriched geochemical characteristics of the Antampombato mafic dykes are evident from values of Zr/Nb lower than those of normal MORB (Zr/Nb >17). Ratios of elements highly susceptible to modification by crustal contamination, such as Ce/Pb, La/Nb and Ba/Nb, are close to, or within, the ranges of values typical of oceanic basalts (Table 10). From another point of view, we note that the most primitive basalts have relatively high iron (as Fe2O3) contents, unlike melts derived from depleted mantle sources such as MORB-source mantle. This feature, and the enriched geochemical characteristics, are not easily reconciled with shallow melting of suboceanic depleted mantle, but can be obtained from melting deeper and/or Fe-enriched sources (see Scarrow & Cox, 1995
; Gibson et al., 2000).
The differences between mafic (basalts and basaltic andesites) samples of the western and eastern sides of the island are highlighted in Fig. 12. The samples of the northwestern part of the island have very high LILE/HFSE ratios (e.g. Ba/Nb = 68 ± 28; n = 85), systematically low HFSE contents, relatively low total iron (as Fe2O3) and TiO2, and high Al2O3. These features strongly imply melting of a shallow, depleted, MORB-like, mantle source, coupled with widely variable amounts of contamination by crustal materials (Melluso et al., 1997, 2001, 2003). Low HFSE contents and LILE/HFSE ratios are shown by some Mananjary basalts (MAN90-16, MAN90-22, MAN90-77, MAN90-15), which are MORB-like also on the basis of their Sr–Nd isotope composition (Storey et al., 1997, and see below). Storey et al. (1997) described low-Ti basalts with markedly high LILE/HFSE ratios in the Mananjary sector, but did not provide analyses.
|
, Antampombato mafic dykes. (a) Ba/Nb vs Nb; (b) Fe2O3 vs SiO2; (c) Zr vs Nb; (d) Al2O3 vs TiO2. A partial melting curve starting from depleted MORB mantle (The bulk of the northeastern province, the Antampombato and most Mananjary basalts appear to have been derived from a distinctly more enriched mantle source, with smaller effects of crustal contamination, as indicated, for instance, by generally high, although variable, TiO2, Fe2O3, HFSE contents, and by relatively low Ba/Nb (average 12·0 ± 3·4, n = 120) (Fig. 12a–d). The extreme end-member is represented by the low-SiO2, high-Mg–Ti Mananjary basalts (Storey et al., 1997
). These basalts have a paragenesis unusually rich in opaque microphenocrysts, in addition to olivine (and possibly clinopyroxene) phenocrysts (A. D. Saunders, personal communication, 2002), and hence are very different from the most mafic Antampombato dykes, in which, excluding chromite inclusions in olivine, Fe–Ti oxides are mostly found in the fine-grained groundmass. The high Mg–Ti sample MAN90-45 has been described as a high Fe–Ti picrite, and attributed to partial melting of an enriched plume-type composition by Gibson et al. (2000). Evidence of melting of sources more enriched than the N-MORB source is also suggested by the position of most samples from eastern Madagascar above the model melting curve in Fig. 12c (compare the high Mg–Ti Mananjary rocks); the high Nb contents would require unreasonably low degrees of partial melting (<1%) of a depleted mantle source.
The Sr–Nd isotope ratios of the mafic–ultramafic rocks of the Antampombato complex plot well within the Indian MORB field, possibly slightly displaced towards components more enriched than N-MORB; the Antampombato basalts also overlap isotopically with the range observed in the present-day Marion–Prince Edward archipelago (Fig. 14). These data indicate major differences in the Sr–Nd isotopic composition of the Antampombato mafic rocks and the associated mafic and ultramafic lithologies compared with the mafic volcanics of northeastern Madagascar [in particular, the low La/Yb and the high La/Yb types, as defined by Melluso et al. (2003)]. Major differences from tholeiites and alkali basalts found in the Ejeda–Bekily dyke swarm and in the southeasternmost outcrops (Mahoney et al., 1991; Dostal et al., 1992) are also noted. The Antampombato mafic dykes are isotopically almost identical to the transitional basalts of Mailaka [when the effects of interaction with seawater have been removed (see Melluso et al., 2001)], and to samples from the Mananjary transect (Storey et al., 1997). However, these similarities readily disappear when incompatible element contents are considered (Fig. 13a and b). The transitional basalts of Mailaka (samples M115, M111 of Fig. 13a), although having markedly lower incompatible element contents than the Antampombato basalts at roughly the same MgO content, have peaks at Ba, Sr and Pb, as a result of interaction with crustal materials, whereas the flat pattern of strongly incompatible elements and marked trough at Pb of the Antampombato basalts is distinctive, further evidence for negligible crustal contamination. The same evidence cannot be deduced for most Mananjary and northern Madagascar basalts, because of the peak at Pb in their mantle-normalized patterns (Fig. 13b).
|
|
Any attempt to constrain the depth and degree of partial melting of the Antampombato mantle-derived magmas suffers from the lack of reliable estimates of mantle source composition in addition to the typical limitations of inverse modelling procedures. In our case there is also the lack of different primary magmas with variable degrees of silica saturation. We attempted to match the REE contents of the most Mg-rich basalts of the province by mixing melts derived from non-modal fractional melting of mantle sources (Table 12), starting from peridotitic assemblages with spinel or garnet. The internal consistency of the model requires that the calculated melts had to fit both ratios and concentrations of the most primitive rock samples, because parallel REE patterns can give identical ratios between elements but very different absolute abundances. Modelling was also performed for a few low-MgO basalts, allowing <50% crystal fractionation of gabbroic (sensu lato) assemblages to reach their evolved chemical composition.
|
In Figs 12–15, the significant span of composition of the mafic rocks of the Madagascan province is evident, resulting from changes in the amount of partial melting, changes in the depth of melting, the composition of the mantle source, degree of crystal fractionation and crustal contamination. This last factor is likely to be the cause of the position in Fig. 15 of the low-Nb basalts and basaltic andesites of the Antanimena section (northwestern province), which plot at relatively high La/Yb, and relatively low Gd/Yb, and seem to be displaced from the roughly linear trend that includes most samples (see inset to Fig. 15).
|
, low La/Yb basalts; 
, high La/Yb basalts; +, Mananjary basalts; open crosses, MORB-like Mananjary basalts. Data are from Melluso et al. (1997The range of compositions of calculated melts from spinel- and garnet-bearing sources, starting from depleted MORB source mantle and primitive mantle, shown in Fig. 15, indicates that the Madagascan basalts could be derived from mixing of melts formed at different depths. Given the geochemical evidence of provinciality in the northern Madagascan province (Melluso et al., 2003
, and see Figs 12 and 13), we consider a range from incompatible element-depleted sources for the western Madagascar basalts to primitive mantle (PM) or enriched (1·5–2 x PM) mantle sources for the enriched Madagascar basalts.
We note that the Antampombato mafic dykes, the high La/Yb basalts of northeastern Madagascar, the high Ti–Mg basalts, and some other Mananjary lava flows plot far from the spinel-bearing mantle melting curves for either depleted MORB or primitive mantle (Fig. 15). The position of these samples implies that mixing of melts generated from the garnet facies contributed significantly to the total composition (see Table 12), even though similar results can be obtained from melting a more LREE-enriched source composition. Indeed, Melluso et al. (2003) provisionally ascribed the genesis of Antampombato primitive magmas to 3–5% partial melting of a slightly incompatible element-enriched, spinel-bearing, mantle source. The high La/Yb basalts of northern Madagascar are thought to have been generated by melting of mantle sources with significant amounts of garnet (Melluso et al., 2003). Here we add new information on a regional scale.
It is possible broadly to match the REE contents and ratios of the mafic Madagascan basalts ranging from compositions obtained by dominant or exclusive melting in the spinel field (M102, tholeiitic basalt, Mailaka; M422, picrite basalt, Antanimena; M162, picrite basalt, Mailaka; data from Melluso et al., 2001, 2003) to melts generated dominantly in the presence of garnet (MAN90-45, high Mg–Ti basalt, Mananjary; Storey et al., 1997) (Table 12; Fig. 15). A small, but significant, input of melts from garnet peridotite is needed to model the transitional basalts of the western subprovince, which thus may have formed by melting of depleted MORB mantle starting from depths where garnet is stable. Starting from a primitive mantle composition, the match of the REE pattern of the dyke M497a was obtained through mixing between a 9% partial melt from spinel peridotite and a 5% partial melt from garnet peridotite in the proportions 40% sp-melt to 60% gt-melt. Therefore, roughly 45% of the total melt fraction that formed M497a may have been derived from melts in equilibrium with garnet, with an apparent total melt fraction of 6·5% (Table 12). Using major element mass balance calculations, we obtained a similar value of 6% partial melting for M497a. We were not able to match the incompatible element composition of M497a, starting from a depleted mantle source, even assuming unreasonably low degrees of partial melting of spinel- or garnet-bearing mantle or a combination of both. In the Zr–Nb diagram (Fig. 12c), a fit for Antampombato basalts can be obtained through 1–1·5% melting of a depleted MORB mantle.
A melt segregation depth equivalent to a pressure of 25 kbar is obtained for the Antampombato sample M497a based on a SiO2–pressure regression line interpolating all the compositions obtained by Kushiro (1996) in his partial melting experiments on fertile peridotite PHN1611 from 5 to 30 kbar [P (kbar) = 163·51 – 2·9655 x SiO2 (wt %)]. This suggests that the last equilibration of M497a occurred close to the spinel–garnet peridotite transition (c. 28 kbar; Robinson & Wood, 1997); this is also the case for the transitional basalts of Mailaka (28–29 kbar). According to the above relationship, the high Mg–Ti sample MAN90-45 should have last equilibrated with a peridotite at c. 37 kbar (slightly less than 120 km depth), i.e. well within the garnet stability field.
From the modelling described above, we conclude that significant melting in the presence of residual garnet is needed for many basalt types of the northeastern part of the Madagascan province (the high La/Yb basalts, the Antampombato basalts, and possibly the low La/Yb basalts), whereas the depth of initiation of melting is variable in the Mananjary sector. Relatively high mantle potential temperatures (>1400°C) are required to melt anhydrous mantle starting from the garnet stability field. This has already been stated in earlier works (Storey et al., 1997; Gibson et al., 2000; Melluso et al., 2001, 2003). The variable degree of incompatible element enrichment of the different basalt types throughout the eastern areas, together with the presence of more depleted magmas beneath western areas, indicates that the mantle source of the Madagascan basalts is not uniform in chemistry. As the basalts of the western subprovince do not seem to have any trace of a typical, plume-related, chemical composition, we could argue that this plume has the chemical composition of MORB mantle. Therefore, all the enriched components could derive from interaction of the rising magmas with the mantle lithosphere. Alternatively, the plume could have been chemically heterogeneous; MORB-like beneath some locations and more incompatible element-enriched elsewhere in the province.
It should be noted that the thickness of the Madagascan lithosphere and its variability in Cretaceous times are poorly known. A thick lithospheric root (130 km) is known to be present in the eastern part of the island (Rakotondraompiana et al., 1999). This root could be a remnant of the cratonic Antongil block, which occupies parts of eastern and northeastern Madagascar (see de Wit, 2003, fig. 5). In theory, the Antongil block may also have had a thick, cool lithospheric root in the Cretaceous, thus forcing deeper initiation of melting of hotter-than-normal sublithospheric mantle sources. We speculated earlier that the Madagascan lithosphere in the western areas could have been relatively thinned, as a result of the presence of well-developed sedimentary basins linked to the Karoo rift systems, allowing melting of generally shallower mantle sources, whereas the eastern Madagascan lithosphere could have been at least locally ‘cratonic’ (
100 km thick), before Madagascar–India separation (Melluso et al., 2002). Interaction of sublithospheric melts with such a thick lithosphere may have caused at least part of the chemical variability of the basalts of east Madagascar.
|
CONCLUSIONS |
|---|
The Antampombato–Ambatovy intrusion of central–eastern Madagascar has a tightly constrained 40Ar/39Ar and U/Pb age of 89–90 Ma, and includes a variety of high-level plutonic rocks, ranging from ultramafic to felsic in composition, which were intruded by a dyke swarm ranging in composition from transitional basalt to alkali rhyolite (comendite). All these rocks are broadly cogenetic, and were generated by low-pressure crystal accumulation or fractionation starting from a transitional basalt parental magma. The Sr and Nd isotopic compositions of the basalts plot well within the Indian MORB field, but also overlap with the products of the present-day Marion hotspot. Contents and fractionation of the least and the most incompatible elements suggest moderate degrees of partial melting (6–7%) of a primitive mantle source, starting from depths at which garnet is stable, and continuing up into the spinel facies.
There is sufficient evidence that the mafic magmas that crystallized to form the Antampombato intrusion did not have the chemical and isotopic composition of the two basalt groups found in northeastern Madagascar (see Melluso et al., 2002, 2003). The Antampombato basalts are, therefore, a distinct chemical type, sharing many geochemical characteristics with basalts of the eastern part of the Madagascan province but not having the extreme composition of the high Mg–Ti rocks of the Mananjary sector.
The Antampombato mafic and ultramafic rocks largely escaped crustal contamination before emplacement, thus giving one of the best ‘windows’ to the mantle sources of the Madagascan province, and to the chemical variability of flood basalt sequences with respect to the composition of the mantle sources and the depth of melting. We emphasize the evidence for distinct mantle sources on the two sides of the island, even though characterized by similar (MORB-like) Sr–Nd isotopic composition. Nevertheless, despite the addition of new data, distinguishing the role of ‘plume’, MORB mantle and continental lithosphere as sources of the Madagascan Cretaceous magmatism, and their relative role in space and time, is still a difficult and uncertain task.
|
ACKNOWLEDGEMENTS |
|---|
Massimo Tiepolo (Pavia), Raul Carampin (Padova), John Huard (Corvallis), Vincenzo Monetti and Antonio Canzanella (Napoli), Chiara Petrone (Firenze), Massimo D'Orazio and Samuele Agostini (Pisa) are thanked for their help and expertise in data collection, sample preparation and analysis. The personnel of the Antampombato mine, particularly Fidy Rabetsimamanga, are thanked for their invaluable help in the field. Edmee Tidahy and Volonirina Rasoamalala are thanked for their help in the field. Andrea Marzoli and Cedric Rapaille are thanked for MELTS calculations and useful suggestions. Andy Saunders very kindly provided petrographic information on the high Mg–Ti Mananjary samples. This work has been supported by MIUR-PRIN grants (2001) to L.M. The constructive comments of Andy Saunders and an anonymous reviewer, and the patient and skilled advice of Marjorie Wilson were very helpful in the preparation of a revised version.
* Corresponding author. E-mail: melluso{at}unina.it
|
REFERENCES |
|---|
Anil Kumar, G., Pande, K., Venkatesan, T. R. & Bhaskar Rao, Y. J. (2001). The Karnataka Late Cretaceous dikes as products of the Marion hot spot at the Madagascar–India breakup event: evidence from 40Ar/39Ar geochronology and geochemistry. Geophysical Research Letters 228, 2715–2718.[CrossRef]
Bardintzeff, J.-M., Bonin, B. & Rasamimanana, G. (2001). The Cretaceous Morondava volcanic province (West Madagascar): mineralogical, petrological and geochemical aspects. Journal of African Earth Sciences 32, 299–316.[CrossRef]
Besairie, H. (1964). Geological Map of Madagascar. Tananarive: Service Géologique de Madagascar.
Boynton, W. B. (1984). Cosmochemistry of rare earth elements: meteorite studies. In: Henderson, P. (ed.) Rare Earth Element Geochemistry. Amsterdam: Elsevier, pp. 63–114.
Collins, A. S., Fitzsimmons, I. C. W., Hulscher, B. & Razakamanana, T. (2003). Structure of the eastern margin of the East African Orogen in central Madagascar. Precambrian Research 123, 111–133.[CrossRef][Web of Science]
Compston, W., Williams, I. S., Kirschvink, J. L., Zhang, Z. & Guogan, M. A. (1992). Zircon U–Pb ages for the Early Cambrian time-scale. Journal of the Geological Society, London 149, 171–184.
DePaolo, D. J. (1981). Trace element and isotopic effects of coupled fractional crystallization and crustal contamination processes. Earth and Planetary Science Letters 53, 189–202.[CrossRef][Web of Science]
de Wit, M. J. (2003). Madagascar: heads it's a continent, tails it's an island. Annual Review of Earth and Planetary Sciences 31, 213–248.[CrossRef][Web of Science]
Dostal, J., Dupuy, C., Nicollet, C. & Cantagrel, J. M. (1992). Geochemistry and petrogenesis of upper Cretaceous basaltic rocks from southern Madagascar. Chemical Geology 97, 199–218.[CrossRef][Web of Science]
Duncan, R. A., Hooper, P. R., Rehacek, J., Marsh, J. S. & Duncan, A. R. (1997). The timing and duration of the Karoo igneous event, southern Gondwana. Journal of Geophysical Research 102, 18127–18138.[CrossRef]
Ghiorso, M. S. & Sack, R. O. (1995). Chemical mass transfer in magmatic processes. IV. A revised and internally consistent thermodynamic model for the interpolation and extrapolation of liquid–solid equilibria in magmatic systems at elevated temperatures and pressures. Contributions to Mineralogy and Petrology 119, 197–212.[Web of Science]
Gibson, S. A., Thompson, R. N. & Dickin, A. P. (2000). Ferropicrites: geochemical evidence of Fe-rich streaks in upwelling mantle plumes. Earth and Planetary Science Letters 174, 355–374.[CrossRef][Web of Science]
Hart, S. R. & Dunn, T. (1993). Experimental cpx/melt partitioning of 24 trace elements. Contributions to Mineralogy and Petrology 113, 1–8.[CrossRef][Web of Science]
Ito, E., White, W. M. & Göpel, C. (1987). The O, Sr, Nd and Pb isotope geochemistry of MORB. Chemical Geology 62, 157–176.[CrossRef][Web of Science]
Ketchum, J. W. F., Jackson, S. E., Culshaw, N. G. & Barr, S. M. (2001). Depositional and tectonic setting of the Paleoproterozoic Lower Aillik Group, Makkovik Province, Canada: evolution of a passive margin–foredeep sequence based on petrochemistry and U–Pb (TIMS and LAM-ICP-MS) geochronology. Precambrian Research 105, 331–356.[CrossRef]
Kinzler, R. J. (1997). Melting of mantle peridotite at pressures approaching the spinel to garnet transition: application to mid-ocean ridge basalt petrogenesis. Journal of Geophysical Research 102, 853–874.[CrossRef]
Koppers, A. A. P. (2002). ArArCALC software for 40Ar/39Ar age calculations. Computers and Geosciences 28, 605–619.[CrossRef]
Kostopoulos, D. K. & James, S. D. (1992). Parametrization of the melting regime of the shallow upper mantle and the effects of variable lithospheric stretching on the mantle modal stratification and trace element concentrations in magmas. Journal of Petrology 33, 665–691.
Kushiro, I. (1996). Partial melting of a fertile mantle peridotite at high pressures: an experimental study using aggregates of diamond. In: Basu, A. & Hart, S. R. (eds) Earth Processes: Reading the Isotopic Code. Geophysical Monograph, American Geophysical Union 95, 109–122.
Le Bas, M. J., Le Maitre, R. W., Streckeisen, A. & Zanettin, P. (1986). A chemical classification of volcanic rocks based on the total alkali–silica diagram. Journal of Petrology 27, 745–750.
Ludwig, K. R. (2000). Isoplot—a Geochronological Toolkit for Microsoft Excel. Berkeley Geochronological Center, Special Publications 1a, 53.
Mahoney, J. J., Nicollet, C. & Dupuy, C. (1991). Madagascar basalts: tracking oceanic and continental sources. Earth and Planetary Science Letters 104, 350–363.[CrossRef][Web of Science]
Mahoney, J. J., le Roex, A. P., Peng, Z. X., 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.
McKenzie, D. P. & O'Nions, R. K. (1991). Partial melt distributions from inversion of rare earth element concentrations. Journal of Petrology 32, 1021–1092.
Melluso, L., Morra, V., Brotzu, P., Razafiniparany, A., Ratrimo, V. & Razafimahatratra, D. (1997). Geochemistry and Sr-isotopic composition of the Cretaceous flood basalt sequence of northern Madagascar: petrogenetic and geodynamic implications. Journal of African Earth Sciences 24, 371–390.[CrossRef]
Melluso, L., Morra, V., Brotzu, P. & Mahoney, J. J. (2001). The Cretaceous igneous province of Madagascar: geochemistry and petrogenesis of lavas and dikes from the central–western sector. Journal of Petrology 42, 1249–1278.
Melluso, L., Morra, V., Brotzu, P., D'Antonio, M. & Bennio, L. (2002). Petrogenesis of the Late Cretaceous tholeiitic magmatism in the passive margins of northeastern Madagascar. In: Menzies, M. A., Klemperer, S. L., Ebinger, C. J. & Baker, J. (eds) Volcanic Rifted Margins. Geological Society of America, Special Papers 362, 83–98.
Melluso, L., Morra, V., Brotzu, P., Franciosi, L., Petteruti Lieberknecht, A. M. & Bennio, L. (2003). Geochemical provinciality in the Cretaceous magmatism of northern Madagascar, and mantle source implications. Journal of the Geological Society, London 160, 477–488.
Nimis, P. (1999). Clinopyroxene geobarometry of magmatic rocks. Part 2. Structural geobarometers for basic to acid, tholeiitic and mildly alkaline magmatic systems. Contributions to Mineralogy and Petrology 135, 62–74.[CrossRef][Web of Science]
Niu, Y., Waggoner, D. G., Sinton, J. M. & Mahoney, J. J. (1996). Mantle source heterogeneity and melting processes beneath seafloor spreading centers: the East Pacific Rise, 18°–19°S. Journal of Geophysical Research 101, 27711–27733.[CrossRef]
Pande, K., Sheth, H. C. & Bhutani, R. (2001). 40Ar/39Ar age of the St. Mary's Islands volcanics, southern India: record of India–Madagascar break-up on the Indian subcontinent. Earth and Planetary Science Letters 193, 39–46.[CrossRef][Web of Science]
Rakotondraompiana, S. A., Albouy, Y. & Piqué, A. (1999). Modèle de lithosphère pour l'île de Madagascar (océan Indien occidental): nouvelle interprétation des données gravimétriques. Journal of African Earth Sciences 28, 961–973.[CrossRef]
Rehkämper, M. & Hofmann, A. W. (1997). Recycled ocean crust sediment in Indian Ocean MORB. Earth and Planetary Science Letters 147, 93–106.[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 40Ar/39Ar dating. Chemical Geology (Isotope Geosciences Section) 145, 117–152.
Robinson, J. A. C. & Wood, B. J. (1997). The depth of the spinel to garnet transition at the peridotite solidus. Earth and Planetary Science Letters 164, 277–284.
Roeder, P. L. & Emslie, R. F. (1970). Olivine–liquid equilibrium. Contributions to Mineralogy and Petrology 29, 275–289.[CrossRef][Web of Science]
Scarrow, J. H. & Cox, K. G. (1995). Basalts generated by decompressive adiabatic melting of a mantle plume: a case study from the Isle of Skye, NW Scotland. Journal of Petrology 36, 3–22.
Spera, F. J. & Bohrson, W. A. (2001). Energy-constrained open-system magmatic processes, 1: general model and energy-constrained assimilation and fractional crystallization (EC-AFC) formulation. Journal of Petrology 42, 999–1018.
Storetvedt, M., Mitchell, J. G., Abranches, M. C., Maaloe, S. & Robin, G. (1992). The coast-parallel dikes of east Madagascar: age of intrusion, remagnetization and tectonic aspects. Journal of African Earth Sciences 15, 237–249.[CrossRef]
Storey, M., Mahoney, J. J., Saunders, A. D., Duncan, R. A., Kelley, S. P. & Coffin, M. F. (1995). Timing of hot spot-related vulcanism and the breakup of Madagascar and India. Science 267, 852–855.
Storey, M., Mahoney, J. J. & Saunders, A. D. (1997). Cretaceous basalts in Madagascar and the transition between plume and continental lithosphere mantle sources. In: Mahoney, J. J. & Coffin, M. F. (eds) Large Igneous Provinces: Continental, Oceanic and Planetary Flood Volcanism. American Geophysical Union, Monograph 100, 95–122.
Sun, S.-s. & McDonough, W. F. (1989). Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. In: Saunders, A. D. & Norry, M. J. (eds) Magmatism in the Ocean Basins. Geological Society, London, Special Publications 42, 313–345.
Thompson, R. N. (1987). Phase-equilibria constraints on the genesis and magmatic evolution of oceanic basalts. Earth-Science Reviews 24, 161–210.
Tiepolo, M. (2003). In situ Pb geochronology of zircon with laser ablation–inductively coupled plasma–sector field mass spectrometry. Chemical Geology 199, 159–177.[CrossRef][Web of Science]
Torsvik, T. H., Tucker, R. D., Ashwal, L. D., Eide, E. A., Rakotosolofo, N. A. & de Wit, M. J. (1998). Late Cretaceous magmatism of Madagascar: paleomagnetic evidence for a stationary hotspot. Earth and Planetary Science Letters 164, 221–232.[CrossRef][Web of Science]
Torsvik, T. H., Tucker, R. D., Ashwal, L. D., Carter, L. M., Jamtveit, B., Vidyadharan, K. T. & Venkataramana, P. (2000). Late Cretaceous India–Madagascar fit and timing of break-up related magmatism. Terra Nova 12, 220–224.[CrossRef][Web of Science]
Wiedenbeck, M., Alle, P., Corfu, F., Griffin, W. L., Meier, M., Oberli, F., Von Quadt, A., Roddick, J. C. & Spiegel, W. (1995). Three natural zircon standards for U–Th–Pb, Lu–Hf, trace elements and REE analysis. Geostandards Newsletter 19, 1–23.[Web of Science]
Workman, R. K. & Hart, S. R. (2005). Major and trace element composition of the depleted MORB mantle (DMM). Earth and Planetary Science Letters 231, 53–72.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Melluso, H. C. Sheth, J. J. Mahoney, V. Morra, C. M. Petrone, and M. Storey Correlations between silicic volcanic rocks of the St Mary's Islands (southwestern India) and eastern Madagascar: implications for Late Cretaceous India-Madagascar reconstructions Journal of the Geological Society, March 1, 2009; 166(2): 283 - 294. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. J. Mahoney, A. D. Saunders, M. Storey, and A. Randriamanantenasoa Geochemistry of the Volcan de l' Androy Basalt-Rhyolite Complex, Madagascar Cretaceous Igneous Province J. Petrology, June 1, 2008; 49(6): 1069 - 1096. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||