| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology Volume 42 Number 12 Pages 2175-2195 2001
© Oxford University Press 2001
Differentiation Processes of Deccan Trap Basalts: Contribution from Geochemistry and Experimental Petrology
1EARTHQUAKE RESEARCH INSTITUTE, THE UNIVERSITY OF TOKYO, TOKYO 113-0032, JAPAN
2INSTITUTE FOR GEOTHERMAL SCIENCES, KYOTO UNIVERSITY, NOGUCHIBARU, BEPPU 874-0903, JAPAN
3PLOT NO. 89, FRIENDS HOUSING SOCIETY LAYOUT-4, DEENDAYALNAGAR, NAGPUR 400 022, INDIA
4FACULTY OF GEO-ENVIRONMENTAL SCIENCES, RISSHO UNIVERSITY, KUMAGAYA 360-0194, JAPAN
5DEPARTMENT OF EARTH SCIENCES, NIHON UNIVERSITY, TOKYO 156-0045, JAPAN
Received June 5, 2000; Revised typescript accepted May 18, 2001
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGICAL BACKGROUNDSELECTION OF THE LEAST-...EXPERIMENTAL PETROLOGYCONCLUSIONSAPPENDIXREFERENCES |
|---|
The Deccan Traps basalts can be divided into sub-groups based on the inferred type and/or amount of contamination. The elemental characteristics of Ba, Sr, TiO2 and Zr/Nb are used to classify the sub-groups; the least-contaminated group has Ba contents <100 ppm, Sr 190–240 ppm and TiO2 2·0–4·0 wt %, and the most-contaminated group has TiO2 contents <1·5 wt % and Zr/Nb
15. Analyses of 325 basalts, which were collected from 27 well-distributed sections through the Deccan Traps, demonstrate that the least- and most-contaminated groups are distributed widely. To understand the shallow-level fractionation of the Deccan Trap magmas, melting experiments were conducted at atmospheric pressure (100 kPa) at both fayalite–magnetite–quartz (FMQ ) and nickel–nickel oxide (NNO) oxygen fugacities for three Mg-rich basalts, one of which belongs to the least-contaminated group. The results indicate that the phenocryst assemblage and the chemical trend of the least-contaminated basalts can reasonably be explained by fractional crystallization in shallow chambers under FMQ-buffered conditions. The inferred fractional crystallization process also reproduces the chemical trend of the most-contaminated basalts, implying that crustal contamination was not accompanied by the shallow-level fractional crystallization. KEY WORDS: Deccan Traps; fractional crystallization; crustal contamination; melting experiment; tholeiitic magma
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL BACKGROUNDSELECTION OF THE LEAST-...EXPERIMENTAL PETROLOGYCONCLUSIONSAPPENDIXREFERENCES |
|---|
The Deccan Traps province, one of the most voluminous continental flood basalt provinces on Earth, consists of basalts with a wide range of compositions not only in isotopic ratios but also in major and trace element contents (e.g. Mahoney, 1988
). During ascent of a plume-related magma through thick continental crust, the variety of compositions would be produced by various differentiation processes such as fractional crystallization, crustal contamination and accumulation of phenocryst phases (e.g. Sen, 1986; Cox & Mitchell, 1988; Lightfoot et al., 1990). The range of the compositions could also be generated, in part, by addition of magmas derived from continental lithospheric mantle (e.g. Hooper, 1994; Turner & Hawkesworth, 1995). When we evaluate the effects of each differentiation process on the chemical compositions, the evolutionary history of Deccan Trap magmas can be understood.
The following studies were conducted to explore the differentiation processes of Deccan Trap magmas. The Ambenali Formation basalts, which form the thickest formation in the western Deccan, are regarded to be the least contaminated by continental crust and the least affected by the continental mantle lithosphere, on the basis of Nd, Sr, Pb and O isotopic signatures (e.g. Cox & Mitchell, 1988; Mahoney, 1988). The major and trace element compositions of the Ambenali basalts are controlled by the amount of gabbroic fractionation (Cox & Devey, 1987; Devey & Cox, 1987; Cox & Mitchell, 1988; Lightfoot et al., 1990; Aoki et al., 1992; Sen, 1995). The geochemical features of most other Deccan Trap basalts can be explained by mixing between the least-contaminated Ambenali-type magma and three types of material: broadly granitic, possibly upper crust; granulitic and amphibolitic, probably lower crust; and continental lithospheric mantle material (Mahoney et al., 1982; Cox & Hawkesworth, 1985; Matsuhisa et al., 1986; Lightfoot & Hawkesworth, 1988; Lightfoot et al., 1990; Hooper, 1994; Peng et al., 1994; Peng & Mahoney, 1995). The Bushe Formation basalts have the highest 87Sr/86Sr and lowest 
Nd among Deccan Trap basalts, and are considered to be the most contaminated by broadly granitic-type crust.
These conclusions are based on studies of basalts in the southwestern Deccan (i.e. the Western Ghats area), but should be extended to cover the basalt piles in the central and eastern Deccan (e.g. Subbarao et al., 1994; Peng et al., 1998; Mahoney et al., 2000). Peng et al. (1998) reported that lavas isotopically and chemically indistinguishable from the Ambenali basalts form parts of two sections in the eastern Deccan. Peng et al. (1988) and Mahoney et al. (2000) also found lavas chemically similar to the Bushe formation, although these lavas were isotopically slightly different from the Bushe basalts. However, the overall distribution of Ambenali-type and Bushe-type basalts in the central and eastern Deccan has not been worked out.
Comparison of experimental melting data with natural-rock data is a powerful method to understand the differentiation process. For natural-rock data, we should use datasets for basalts from the whole Deccan Traps to help clarify the differentiation processes of the entire Deccan. In this paper, we select the least-contaminated (Ambenali-type) and most-contaminated (Bushe-type) basalts among widely distributed sections in the western, central and eastern Deccan Traps, and examine the distributions of lava flows of these two sub-groups. Next, the gabbroic fractionation of the least-contaminated basalts is evaluated on the basis of melting experiments at atmospheric pressure (100 kPa). Lastly, the chemical effects of crustal contamination are assessed by comparing variations in the most-contaminated basalts with those predicted by the melting experiments.
|
GEOLOGICAL BACKGROUND |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL BACKGROUNDSELECTION OF THE LEAST-...EXPERIMENTAL PETROLOGYCONCLUSIONSAPPENDIXREFERENCES |
|---|
The Deccan Traps province of India covers an area of 5 x 105 km2 (Fig. 1). The basalt flows are generally 10–50 m thick and tabular in form, dipping at <0·5° in various directions (e.g. Deshmukh, 1977
). The thickness of the lava pile is >2000 m in the Western Ghats area between Nasik and the southern edge of the Deccan Traps, and exceeds 1000 m in parts of the eastern Deccan along the Tapti and Narmada grabens (Fig. 1). In some southeastern and far-northern regions distant from both the Western Ghats and the eastern Deccan areas, the thickness of the lava pile is <100 m. In the Western Ghats and the eastern Deccan areas, a large number of dykes occur, in some cases as dyke swarms (Deshmukh & Sehgal, 1988; Hooper, 1990; Bhattacharji et al., 1994; Chatterjee & Nair, 1996; Chawade, 1996; Godbole & Ray, 1996; Sethna et al., 1996). These swarms probably formed under tensional stress fields associated with two main fault systems, the Panvel flexure and Tapti–Narmada grabens (Fig. 1).
|
The lava pile in the Western Ghats is divided into 11 stratigraphic formations (Fig. 1) based on both field observations and geochemical characteristics. The uppermost Panhala Formation appears on the southern edge of the Western Ghats and the lowermost Jawhar Formation exists in the Igatpuri section, indicating a southward dip of formational boundaries (e.g. Beane et al., 1986; Devey & Lightfoot, 1986; Mahoney, 1988).
The main eruption of the Deccan basalts took place in a geologically short period of time (probably <1 m.y.) at 66 Ma (e.g. Courtillot et al., 1988; Duncan & Pyle, 1988; Venkatesan et al., 1993; Baksi, 1994; Allègre et al., 1999), indicating high eruption rates of 
1 km3 per annum. Geological information, such as the absence of thick intra-flow sedimentary layers in any section and the lack of erosional profiles developed between successive eruptions (e.g. Deshmukh, 1977), also indicates a geologically short period of time for the main eruption. Plate tectonic reconstructions show that the Deccan eruptions took place when the Indian continent passed over a plume situated beneath the present site of Réunion (e.g. Müller et al., 1993).
|
SELECTION OF THE LEAST- AND MOST-CONTAMINATED GROUPS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL BACKGROUNDSELECTION OF THE LEAST-...EXPERIMENTAL PETROLOGYCONCLUSIONSAPPENDIXREFERENCES |
|---|
Samples
The samples chosen for this study came from throughout the present geographical distribution of the Deccan Traps (Fig. 1). The region between Igatpuri (IG) and Amboli (AB) is called the Western Ghats area in this study. The samples were collected from 10 formations (Panhala, Mahabaleshwar, Ambenali, Poladpur, Bushe, Khandala, Bhimashankar, Thakurvadi, Igatpuri and Jawhar). For convenience, we call the area between 74°30' and 77°E (BI, ST, BU, AJ, CK and LC) the Central area, and we call that east of 77°E (P, CH, AN, TL, MG, KV, NY, NP, NC, NJ and JB) the Eastern area (Fig. 1). To confirm intra-flow heterogeneity, one group of five samples was collected from the BI-13 flow in the Barwaha–Indore (BI) section and another group of 10 samples was collected from the MA-W flow in the Ambenali Formation of the Mahabaleshwar (MA) section. The intra-flow heterogeneity for each element is reported below.
The Deccan basalts are largely microporphyritic with phenocrysts of plagioclase, subordinate augite and rare olivine. Phenocrysts are set in a groundmass consisting of plagioclase, augite, rare Fe–Ti oxide minerals and glass. Some basalts contain glomero-porphyritic aggregates of augite and plagioclase crystals, occasionally in association with olivine.
Analytical methods
Analyses of major and trace elements of 325 samples were carried out by X-ray fluorescence spectrometry (XRF) with a Rigaku System 3080E3 instrument at the Earthquake Research Institute of the University of Tokyo, following the analytical procedures described by Kaneko (1995). The 1
value of a calibration line for each element is shown in the Appendix (Table A1).
Samples were ground to a fine powder in an agate mill. The powders were then dried for 24 h at 105°C. For major element analysis, 0·4000 ± 0·0004 g of powdered sample was mixed with 4·000 ± 0·004 g of anhydrous lithium tetraborate (Li2B4O7). The mixture was fused at 1100°C in a Pt95Au5 crucible and shaped into a glass bead, which was then used directly for the measurements. For trace elements (Rb, Sr, Ba, Y, Zr, Nb, V, Cr, Ni, Cu, Zn and Ga), 4·0 g of powdered sample was pressed into a pellet by a 10 ton force from a hydraulic press.
A non-destructive instrumental neutron activation analysis (INAA) technique was used for the analyses of Th, Sc and rare-earth elements (REE; La, Ce, Nd, Sm, Eu, Tb, Yb and Lu) on seven samples (Table A1). The powders were activated with thermal neutrons for 1 h with 5·5 x 103 n/cm2 s at the ‘S pipe’ of the JRR-4 reactor of Japan Atomic Energy Research Institute. Simultaneously, JB-1, JR-2 (standard rocks from the Geological Survey of Japan) and BCR-1 (a standard rock from the US Geological Survey) were activated as standards. After a suitable time, the gamma-ray spectra of activated samples were counted in two different ways. The first involved use of a Ge detector coupled to a 2048 multi-channel analyser at the Faculty of Sciences, Gakushuin University. In the second method, use was made of another type of Ge detector, this time coupled to a 4096 multi-channel analyser. The details of the analytical procedure have been described by Fukuoka et al. (1987). The analytical error involved in the results is <±5%.
|
We collected one fresh basalt (NY-02-51) with fresh glass inclusions in the olivine phenocrysts, and the H2O contents in the glass inclusions were analysed by Fourier transform IR spectroscopy (FTIR-300E, JASCO Corporation) at the Earthquake Research Institute of the University of Tokyo, following the method of Yamashita et al. (1997)
. For the analysis, doubly polished olivine crystal plates (with inclusions intersected on both sides) were prepared by hand grinding on 1000-grit paper rinsed with H2O and then polished with 10 000-grit paper. The H2O contents of the glass were determined from the IR spectra, which were calibrated from manometric and spectroscopic measurements of natural and synthetic basaltic glass samples (Yasuda, in preparation). The intensities for the 3550 cm-1 band of the basaltic glass samples were calibrated. The accuracy of the measurements is <±10%, which is small compared with the intersample variation, as will be shown in the following section.
Analytical results
Major and trace element compositions of representative samples are shown in the Appendix (Table A1).
The complete dataset may be downloaded from the Journal of Petrology Web site at http://www.petrology.oupjournals.org. The 1 standard deviations were calculated for five samples from the BI-13 flow and 10 samples from the MA-W flow. These values are the same as or smaller than the 1 values of the calibration lines for the contents of incompatible trace elements (Rb, Sr, Ba, Zr, Y and Nb) as shown in Table A1. For some of the major and compatible trace elements, on the other hand, the intra-flow variations are greater than those of the calibration lines.
The H2O contents of glass inclusions in olivine phenocrysts from a fresh basalt (NY-02-51) range between 0·5 and 0·7 wt % (Table 1) and are within the range of those in mid-ocean ridge basalts (0·1–0·8 wt % H2O; e.g. Dixon et al., 1988; Jambon & Zimmermann, 1990; Johnson et al., 1994), and much smaller than those in arc basalts (up to 6 wt % H2O; e.g. Sisson & Layne, 1993; Stolper & Newman, 1994).
|
Identification of the least- and most-contaminated groups
In the Western Ghats area, each of the upper five formations has been defined based on Ba, Sr and TiO2 contents and Zr/Nb, 87Sr/86Sr and Nd (Cox & Hawkesworth, 1985; Devey & Lightfoot, 1986; Lightfoot & Hawkesworth, 1988; Lightfoot et al., 1990; Peng et al., 1994). The Ambenali formation is formed by basalts with restricted compositions of Ba < 100 ppm, Sr 190–240 ppm, TiO2 2·0–4·0 wt %, Zr/Nb = 10–18, 87Sr/86Sr < 0·7050 and Nd > 3·0, indicating that these basalts are less contaminated by continental crust and/or continental lithospheric mantle than basalts from any other formation. Figure 2 shows that the majority (>90%) of basalts with Ba < 100 ppm, Sr 190–240 ppm and TiO2 2·0–4·0 wt % have isotopic compositions of 87Sr/86Sr < 0·7050 and Nd > 3. It is thus possible in most cases to identify probable least-contaminated basalts on the basis of Ba, Sr and TiO2 when isotope data are not available.
|
Bushe Formation basalts are regarded as the most contaminated by broadly granitic-type crust because of the high 87Sr/86Sr (>0·7120), greater than values for basalts from any other formation in the Western Ghats (Figs
2 and 3). Figure 3 shows that the majority (>95%) of the Bushe basalts have TiO2 contents <1·5 wt % and Zr/Nb 15. We use TiO2 content and Zr/Nb to identify probable most-contaminated basalts when isotope data are lacking.
|
When we select the least-contaminated (Ambenali-type) and most-contaminated (Bushe-type) groups from the Western Ghats area on the basis of Ba, Sr, TiO2 and Zr/Nb, most Ambenali lavas lie within the least-contaminated group, and most Bushe lavas fall into the most-contaminated group (Fig. 4a). This fact demonstrates that the chosen chemical criteria are useful for identifying the least- and most-contaminated basalts of the Deccan Traps.
|
Distributions of the least- and most-contaminated groups in the Central and Eastern areas
In the Central and Eastern areas, the least-contaminated basalts are distributed over large regions (Fig. 4b). Most Ambenali-like lavas reported by Peng et al. (1998) from the Chikaldara (CH) and Nagpur–Jabalpur (NJ) sections are classified as least-contaminated basalts. Although we expected that the least-contaminated basalts in the other sections would also have geochemical affinities with the Ambenali lavas, some of the least-contaminated basalts do not have similar isotopic compositions to the Ambenali lavas. Mahoney et al. (2000) have reported that basalts with isotopic compositions similar to the Ambenali lavas were not present in the Shahada–Toranmal (ST) section, although we could not distinguish the least-contaminated basalts at the ST section from the Ambenali basalts based on major and trace element compositions (Fig. 4b). The least-contaminated basalts in the Central and Eastern areas do not constitute a thick (>500 m) lava pile such as the Ambenali Formation in the Western Ghats area (Fig. 4), which suggests that distribution of the least-contaminated basalts in the Central and Eastern areas is different from that in the Ambenali Formation.
The most-contaminated basalts are also widely distributed in the Central and Eastern areas (Mahoney et al., 2000); they are present in the ST, P and CH sections (Fig. 4b). The lava pile of the most-contaminated basalts overlies the least-contaminated lavas directly at the P section. At the CH section, the most-contaminated basalts are interbedded with the least-contaminated lavas. On the other hand, in the Western Ghats area, the most-contaminated lavas (Bushe Formation) are located under the least-contaminated lavas (Ambenali Formation). The stratigraphic discrepancy indicates that some or all of the formations in the Western Ghats area do not continue to the Central and Eastern areas (Peng et al., 1998; Mahoney et al., 2000). One possible explanation of this difference is that lava flows of the Central and Eastern areas did not erupt from a vent located in the Western Ghats area (Peng et al., 1998). The vents of the Eastern and Central lava flows may have been dyke swarms along Tapti–Narmada grabens.
|
EXPERIMENTAL PETROLOGY |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL BACKGROUNDSELECTION OF THE LEAST-...EXPERIMENTAL PETROLOGYCONCLUSIONSAPPENDIXREFERENCES |
|---|
A goal of the present study was to evaluate the effects of fractional crystallization and crustal contamination on Deccan magmas by comparison of natural-rock and experimental data. Before making the comparison, the effects of phenocryst accumulation recorded in the major element compositions of the least-contaminated basalts should be eliminated so as to evaluate fractional crystallization processes. The effects of phenocryst accumulation can be identified by the following data. The least-contaminated basalts include aphyric to highly porphyritic rocks whose phenocryst contents vary from 0 to 30 vol. %. The most voluminous phenocryst is plagioclase. Some porphyritic basalts have higher Al2O3 and lower FeO* contents than the aphyric basalts (Fig. 5), indicating accumulation of plagioclase phenocrysts in these porphyritic basalts (e.g. Cox & Mitchell, 1988
). We therefore selected only aphyric and sparsely porphyritic basalts among the least-contaminated basalts.
|
Melting experiments
Two previous studies conducted melting experiments on the Deccan Trap basalts (Krishnamurthy & Cox, 1977; Cohen & Sen, 1994). Krishnamurthy & Cox (1977) reported melting experiments at atmospheric pressure (100 kPa), but they used alkalic basalts from the northwestern Deccan Traps (e.g. Krishnamurthy & Cox, 1980; Mahoney et al., 1985; Melluso et al., 1995; Krishnamurthy et al., 2000) as the starting materials. Their results are not applicable to tholeiitic magmas, which form the majority of Deccan Trap basalts. Cohen & Sen (1994), however, conducted their experiments on tholeiites at 600 MPa, and concluded that the high-pressure liquid line of descent (LLD) is not adequate to explain the trend in chemical characteristics of the Ambenali basalts. We therefore conducted melting experiments at atmospheric pressure (100 kPa).
As the H2O content in one of our least-contaminated samples is negligibly small (Table 1), its fractional crystallization properties were studied under dry conditions.
Starting materials
For melting experiments at atmospheric pressure, we selected one of the most Mg-rich (MgO 6·7 wt %) tholeiitic basalts (MA-W-25 from the Ambenali Formation) among the least-contaminated group (Table A1). Two other Mg-rich (MgO >9 wt %) basalts were also selected as starting materials to investigate crystallization processes at high MgO content; one (BH-14 from the Khandala Formation) belongs to the most-contaminated group and the other (IG-02 from the Jawhar Formation) is classified as intermediate in contamination.
BH-14 and IG-02 are sparsely olivine-phyric basalts, and MA-W-25 is a sparsely olivine–plagioclase–augite-phyric basalt. The phenocryst contents of MA-W-25, BH-14 and IG-02 are 3, 2 and 3 vol. %, respectively (Table A1).
Experimental and analytical methods
Samples were crushed into millimetre-sized pieces with an iron mortar and pestle, and visibly altered pieces were removed by handpicking. The rock chips were placed in an agate ball mill with acetone and ground into powder. The powdered sample was dried and pellets of about 4 g in weight were formed with a hydraulic press. The pellets were then crushed into many small chips (200–400 mg in weight) which were bound up with Pt wire of diameter 0·050 mm. The sample was fused into a glass at a higher temperature (1300°C) than the liquidus under the oxidation conditions for the target experiments, and then quenched. The glass beads were then used as starting samples.
The glass beads on Pt loops were suspended in the hot spot of a quenching furnace. Temperature was measured during each run with a Pt/Pt87Rh13 thermocouple with accuracy of better than ±1·5°C. A mixture of CO2 and H2, flowing vertically upward through the furnace, was used to maintain oxygen fugacity (fO2) at fayalite–magnetite–quartz (FMQ) and nickel–nickel oxide (NNO) buffers. These conditions were selected because most tholeiitic basalts are considered to have crystallized at fO2 between these two buffers (e.g. Walker et al., 1979; Helz & Thornber, 1987). After a sufficient run time as described below, experiments were terminated by quenching in water. Experimental conditions are reported in Table 2.
|
|
The compositions of phases in run products were analysed with a JEOL JXA-8800R electron-probe microanalyser (EPMA) at the Earthquake Research Institute of the University of Tokyo, using a 15 kV accelerating voltage and a 12 nA beam current. The counting time was 10 s. The beam size was 10 µm for glass analysis and 1 µm for mineral phases. The data were reduced using the Bence & Albee (1968)
matrix correction and included the modification described by Nakamura & Kushiro (1970).
Experimental results
The proportions and compositions of phases are summarized in Tables 2 and 3. The phase proportions in Table 2 were estimated based on a least-squares mass-balance calculation of major element compositions for the phases (Table 3). For MA-W-25, olivine, plagioclase and augite crystallize simultaneously at 1155°C at the FMQ buffer. These three mineral phases crystallize even at temperatures as low as 1114°C (Table 2, Fig. 6). On the other hand, at the NNO buffer, plagioclase, augite and Fe–Ti oxide mineral crystallize at 1154°C. The plagioclase, augite and Fe–Ti oxide mineral are stable at <1144°C, and low-Ca pyroxene also appears at <1134°C (Table 2, Fig. 6).
|
Olivine is the liquidus phase for both BH-14 and IG-02 at 1220°C (Table 2), which is consistent with the petrographical observations. The crystallization sequence for BH-14 is olivine plagioclase augite Fe–Ti oxide mineral at both FMQ and NNO buffers (Table 2, Fig. 6). On the other hand, for IG-02, the second crystallizing phase is augite, which appears at 1190°C, followed by plagioclase at <1170°C. The higher crystallization temperature of augite for IG-02 may be due to its higher CaO/Al2O3 compared with BH-14 (Table A1). In addition to the above three mineral phases (olivine, plagioclase, augite), an Fe–Ti oxide also crystallizes at 
1184°C at the NNO buffer, and low-Ca pyroxene begins to crystallize at 1134°C at the FMQ buffer in the IG-02 composition.
Volatilization of Fe and Na from the experimental charge is a serious problem in the atmospheric pressure experiments (e.g. Tormey et al., 1987). In the present experiments, however, serious volatilization was not confirmed because the duration of the experiments was limited (61 h). The calculated per cent FeO loss and per cent Na2O loss in the melting experiments were within the precision of the EPMA technique (Table 2).
Attainment of equilibrium
The experiments were run as long as practically possible (Table 2) to achieve homogeneity in the products, but were limited by Na loss from the samples. Longer duration experiments than those carried out here (>61 h) are required to achieve total homogeneity of the phases (e.g. Grove & Bryan, 1983; Tormey et al., 1987). This is because the homogenization scale of SiO2 in melt is 0·15 mm in 61 h, using an interdiffusion coefficient D of 10-9 cm/s at 1100–1200°C (e.g. Koyaguchi, 1989); this scale is distinctly smaller than the size of the sample (2–3 mm in diameter). However, the following facts show that the experiments at temperatures >1120°C closely approach equilibrium.
First, quenched liquid, olivine, plagioclase and augite in high-temperature experiments (>1120°C) are chemically homogeneous within the precision of the EPMA technique. Because some augite crystals in low-temperature experiments (<1120°C) were found to be chemically heterogeneous, these compositions were not used for evaluating fractional crystallization processes.
Second, the Fe–Mg partitioning between magnesian phases (olivine and augite) and melts observed in the present experiments at high temperature (>1120°C) is constant. Under the conditions of the FMQ buffer, the Fe–Mg partitioning coefficients (KD) are 0·28–0·34 for olivine–melt and 0·26–0·32 for augite–melt. The KD values were calculated by using an Fe3+ content appropriate for FMQ and NNO buffers (Sack et al., 1980), and are in good agreement with those reported from previous equilibrium experiments (e.g. Tormey et al., 1987; Yang et al., 1996). The KD values under the NNO buffer are also constant (0·24–0·30 for olivine and 0·23–0·36 for augite), although with a slightly greater range than those under the FMQ buffer.
Third, almost all of the experimental temperatures coincide with predicted temperatures calculated using the geothermometer of Ford et al. (1983). Those workers reported that temperatures could be obtained within errors of ±10°C by using major element compositions of magmas equilibrated with olivine crystals. When we consider the accuracy of temperature measurements in our experiments (±1·5°C), the majority of the calculated temperatures are all within 10°C of the experimental temperature (Fig. 7). However, calculated temperatures for three runs (24-2, 26-2 and 29-2) are distinctly lower (by >15°C) than the experimental temperatures, indicating a thermocouple problem or modification during quenching in the glasses from these runs. We do not use results of these three runs in the following discussion.
|
Comparison of natural-rock and experimental data
Evaluation of fractional crystallization
The least-contaminated basalts contain phenocrysts of plagioclase, augite and olivine, and do not possess any Fe–Ti oxide mineral as a phenocryst phase. The phenocryst assemblage in the least-contaminated basalts is identical to that crystallizing during the melting of MA-W-25, BH-14 and IG-02 at the FMQ buffer. We suggest that the FMQ-buffered conditions are more representative of Deccan Trap magmas than are those equivalent to the fO2 of the NNO buffers. This is consistent with the estimated fO2 for Deccan Trap basalts based on the compositions of Fe–Ti oxides (Sen, 1986; Sethna & Sethna, 1988).
To check the validity of the fractional crystallization processes, compositions of phenocryst phases in the least-contaminated basalts are compared in Fig. 8 with those from experiments on a least-contaminated basalt (MA-W-25). The minerals obtained from the experiments and the natural phenocrysts have nearly identical compositions for olivine, plagioclase and augite. Because natural olivine compositions are known for one least-contaminated basalt (NY-02-51 in Table 1), the natural phenocrysts have restricted compositions compared with the minerals obtained by the experiments. Figure 8 shows that augite crystals obtained from the experiments have lower CaO contents than augite phenocrysts in the least-contaminated basalts. This result is simply due to CaO depletion of the starting composition (Fig. 9).
|
|
The major element compositions of the least-contaminated basalts define certain fundamental chemical trends: SiO2 is constant, and FeO* and K2O show a gradual increase with decreasing MgO content. In addition, Al2O3 and CaO show a progressive decrease with decreasing MgO content (Fig. 9). The chemical trends observed in the least-contaminated basalts can be approximately reproduced by the LLD obtained from the experiments at both FMQ- and NNO-buffered conditions, indicating that the least-contaminated basalts experienced fractional crystallization at low pressure (100 kPa) under dry conditions. Figure 9 shows that the FeO* content in the experimental melts at NNO-buffered conditions progressively decreases, whereas that in the least-contaminated basalts increases with decreasing MgO content. Therefore, the LLD under FMQ-buffered conditions is more favourable than that under NNO-buffered conditions to explain the chemical trend of the least-contaminated basalts.
The results of the experiments (Fig. 6) show that one of the most evolved magmas, with 5·0 wt % MgO, can be produced by subtraction of 8 wt % of olivine, 20 wt % of plagioclase and 7 wt % of augite from one of the most Mg-rich basalts (MA-W-25). This result is consistent with the earlier conclusions (e.g. Cox, 1980; Cox & Devey, 1987; Devey & Cox, 1987; Cox & Mitchell, 1988; Lightfoot et al., 1990; Sen, 1995) that the least-contaminated magma evolved its chemical composition by gabbroic fractionation (olivine, plagioclase and augite) in shallow magma chambers and/or dykes.
The behaviour of incompatible trace elements during the fractional crystallization process is further examined assuming Rayleigh fractionation (e.g. Shaw, 1970). The element contents expected from fractional crystallization are shown in Fig. 10 for Ba, Nb, Zr and Y (see Table 4 for the fractional crystallization model used). The trace element variations observed in the least-contaminated basalts are reproduced by the modelled fractional crystallization pathways.
|
|
Because the MgO content of the majority of the least-contaminated basalts is 5·0–7·0 wt % (Fig. 9), the temperature range during the fractional crystallization is assumed to be 1150–1170°C (Fig. 6). Such a temperature range is hotter than previous estimates of <1145°C and 1025°C based on pigeonite and Fe–Ti oxide geothermometers, respectively (Sen, 1986; Sethna & Sethna, 1988). The appearance of pigeonite and Fe–Ti oxides solely as groundmass phases in the least-contaminated basalts may account for this.
In summary, the comparison of natural-rock and experimental data suggests that the petrography and the chemistry of the least-contaminated basalts can be reasonably explained by low-pressure (100 kPa) fractional crystallization under dry conditions.
Evaluation of crustal contamination
Crustal contamination might be accompanied by fractional crystallization for Deccan Trap magmas because contamination with wall-rock material could be aided by heat released by partial crystallization of magma. If so, positive correlations between the amount of crustal contamination and the degree of fractional crystallization in magmas erupted through continental crust are expected. Previous workers (e.g. DePaolo, 1981; Kinzler et al., 2000) reported that differentiated magmas with low mg-number [100 x Mg/(Mg + Fe2+)] and low MgO content often can be more contaminated than primitive magmas (e.g. 87Sr/86Sr of a differentiated magma may be higher than that of a primitive magma). The contamination (assimilation) accompanied by fractional crystallization is termed assimilation–fractional crystallization (AFC; e.g. DePaolo, 1981). For the Deccan Traps, however, AFC appears not to apply, because there is no correlation between the mg-number and 87Sr/86Sr in Deccan Trap basalts (Devey & Cox, 1987); rather, there is a rough negative correlation of Nd and mg-number (Mahoney, 1988; Mahoney et al., 2000).
To examine AFC from the aspect of experimental petrology, trace element variations of the most-contaminated Deccan basalts are compared with trends expected from closed-system fractional crystallization. The trace element contents (Ba, Nb, Zr and Y) in the most-contaminated basalts gradually increase with decreasing MgO content, whereas the variation of Ba contents is slightly scattered (Fig. 10). Figure 10 also shows the trace element paths expected from low-pressure fractional crystallization under dry conditions (see Table 4 for fractional crystallization model). The trace element variations in the most-contaminated basalts are generally explained by the fractional crystallization model.
The most-contaminated basalts have higher Ba and lower Nb and Zr contents than the least-contaminated basalts, and many likely continental crustal contaminants have low MgO (<4 wt %), higher Ba, and lower Nb and Zr contents compared with Deccan Trap basalts. If the most-contaminated basalts contain higher Ba and lower Nb and Zr contents than expected for fractional crystallization, then an AFC model might apply. However, AFC is not required for the variations of the most-contaminated basalts, which suggests the following scenario (Devey & Cox, 1987; Mahoney, 1988; Lightfoot et al., 1990; Mahoney et al., 2000): the chemical variations observed in the most-contaminated basalts were generated by fractional crystallization of a gabbroic assemblage in a chamber, without accompanying contamination. Most of the contamination may have taken place before most shallow-level fractional crystallization (e.g. Lightfoot et al., 1990).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL BACKGROUNDSELECTION OF THE LEAST-...EXPERIMENTAL PETROLOGYCONCLUSIONSAPPENDIXREFERENCES |
|---|
On the basis of Ba (<100 ppm), Sr (190–240 ppm) and TiO2 (2·0–4·0 wt %) contents, an inferred least-contaminated basalt group was selected from samples covering the whole area of the Deccan Traps. The least-contaminated basalts are widely distributed in the central and eastern Deccan and in the western Deccan, where they appear as the Ambenali Formation. The most-contaminated basalts that form the Bushe Formation in the western Deccan have chemical signatures of TiO2 < 1·5 and Zr/Nb
15. The most-contaminated group is located under the least-contaminated group in the western Deccan. On the other hand, the most-contaminated group caps or is interbedded with the least-contaminated group in parts of the central and eastern Deccan, as was also concluded by Peng et al. (1998) and Mahoney et al. (2000), indicating that formations in the western Deccan do not necessarily continue to the central and eastern Deccan. Phenocryst phases of the least-contaminated basalts (olivine, plagioclase and augite) are reproduced by melting experiments at atmospheric pressure (100 kPa) under FMQ-buffered conditions, and the liquid line of descent under these conditions agrees with the major and trace element trends of the least-contaminated basalts. The experimental results also show that the temperature range during fractional crystallization was 1150–1170°C. The chemical trends of the most-contaminated basalts are also reproduced by fractional crystallization at atmospheric pressure, which suggests that crustal contamination was not coupled to fractional crystallization in shallow magma chambers.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL BACKGROUNDSELECTION OF THE LEAST-...EXPERIMENTAL PETROLOGYCONCLUSIONSAPPENDIXREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
We thank A. Yasuda for his help in melting experiments and electron probe analyses, and K. K. K. Nair and D. B. Yedeker for providing the geological information. We are grateful to R. J. Arculus and T. Falloon for their careful and critical reviews. We are also grateful to J. J. Mahoney, G. Sen and Y. Tatsumi for their careful and constructive comments on an earlier version of this manuscript. K. Kaneko, I. Kaneoka, T. Koyaguchi, S. Nakada and H. Nagahara are thanked for constructive discussions. This work was supported by Research Fellowships from the Japan Society for the Promotion of Science for Young Scientists.
|
FOOTNOTES |
|---|
Extended dataset can be found at http://www.petrology.oupjournals.org
*Corresponding author. Present address: College of Environment and Disaster Research, Fuji Tokoha University, Ohbuchi 325, Fuji 417-0801, Japan. Telephone: 81-545-37-2007. Fax: 81-545-36-2651. e-mail: sano{at}fuji-tokoha-u.ac.jp
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL BACKGROUNDSELECTION OF THE LEAST-...EXPERIMENTAL PETROLOGYCONCLUSIONSAPPENDIXREFERENCES |
|---|
Allègre, C. J., Birck, J. L., Capmas, F. & Courtillot, V. (1999). Age of Deccan traps using 187Re–187Os systematics. Earth and Planetary Science Letters 170, 197–204.
Aoki, K., Yoshida, T., Aramaki, S. & Kurasawa, H. (1992). Low-pressure fractional crystallization origin of the tholeiitic basalts of the Deccan plateau, India. Journal of Mineralogy, Petrology and Economic Geology 87, 375–387.
Baksi, A. K. (1994). Geochronological studies on whole-rock basalts, Deccan Traps, India: evaluation of the timing of volcanism relative to the K–T boundary. Earth and Planetary Science Letters 121, 43–56.[ISI]
Beane, J. E., Turner, C. A., Hooper, P. R., Subbarao, K. V. & Walsh, J. N. (1986). Stratigraphy, composition and form of the Deccan basalts, Western Ghats, India. Bulletin of Volcanology 48, 61–83.
Bence, A. E. & Albee, A. L. (1968). Empirical correction factors for the electron microanalysis of silicates and oxides. Journal of Geology 76, 382–403.[ISI]
Bhattacharji, S., Chatterjee, N., Wampler, J. M. & Gazi, M. (1994). Mafic dikes in Deccan volcanics: indicator of India intraplate rifting, crustal extension and Deccan flood basalt volcanism. In: Subbarao, K. V. (ed.) Volcanism . New Delhi: Wiley Eastern, pp. 253–276.
Bodas, M. S., Khadri, S. F. R. & Subbarao, K. V. (1988). Stratigraphy of the Jawhar and Igatpuri Formations, western Deccan basalt province. Memoirs of the Geological Society of India 10, 235–252.
Chatterjee, A. K. & Nair, K. K. K. (1996). Petrographic and geochemical studies of Narmada–Tapti–Satpura dyke system between 74° and 76° east longitudes, Central India. Gondwana Geological Magazine Special Volume 2, 251–266.
Chawade, M. P. (1996). The petrology and geochemistry of dykes in Deccan basalts in parts of Lower Narmada valley, around Chhaktala, Jabua distinct, M.P. Gondwana Geological Magazine Special Volume 2, 185–200.
Cohen, T. H. & Sen, G. (1994). Fractionation and ascent of Deccan Trap magmas: an experimental study at 6 kilobar pressure. In: Subbarao, K. V. (ed.) Volcanism . New Delhi: Wiley Eastern, pp. 173–186.
Courtillot, V., Féraud, G., Maluski, H., Vandamme, D., Moreau, M. G. & Besse, J. (1988). Deccan flood basalts and the Cretaceous/Tertiary boundary. Nature 333, 843–846.
Cox, K. G. (1980). A model for flood basalt volcanism. Journal of Petrology
21, 629–650.
Cox, K. G. & Devey, C. W. (1987). Fractionation processes in Deccan Traps magmas: comments on the paper by G. Sen—Mineralogy and petrogenesis of the Deccan Trap lava flows around Mahabaleshwar, India. Journal of Petrology
28, 235–238.
Cox, K. G. & Hawkesworth, C. J. (1984). Relative contribution of crust and mantle to flood basalt magmatism, Mahabaleshwar area, Deccan Traps. Philosophical Transactions of the Royal Society of London, Series A 310, 627–641.
Cox, K. G. & Hawkesworth, C. J. (1985). Geochemical stratigraphy of the Deccan Traps at Mahabaleshwar, Western Ghats, India, with implications for open system magmatic processes. Journal of Petrology
26, 355–377.
Cox, K. G. & Mitchell, C. (1988). Importance of crystal settling in the differentiation of Deccan Trap basaltic magmas. Nature 333, 447–449.
DePaolo, D. J. (1981). Trace element and isotopic effects of combined wallrock assimilation and fractional crystallization. Earth and Planetary Science Letters 53, 189–202.[ISI]
Deshmukh, S. S. (1977). A critical petrological study of the Deccan basalts and associated high level laterites in parts of the Western Ghats, Maharashtra state. Ph.D. thesis, Nagpur University, 306 pp.
Deshmukh, S. S. & Sehgal, M. N. (1988). Mafic dyke swarms in Deccan volcanic province of Madhya Pradesh and Maharashtra, in Deccan flood basalts. Memoirs of the Geological Society of India 10, 323–340.
Deshmukh, S. S., Sano, T., Fujii, T., Nair, K. K. K., Yedekar, D. B., Umino, S., Iwamori, H. & Aramaki, S. (1996a). Chemical stratigraphy and geochemistry of the basalts flows from the central and eastern parts of the Deccan Volcanic Province of India. Gondwana Geological Magazine Special Volume 2, 145–170.
Deshmukh, S. S., Sano, T. & Nair, K. K. K. (1996b). Geology and chemical stratigraphy of the Deccan basalts of Chikaldara and Behramghat sections from the eastern part of the Deccan Traps Province, India. Gondwana Geological Magazine Special Volume 2, 1–22.
Devey, C. W. & Cox, K. G. (1987). Relationships between crustal contamination and crystallization in continental flood basalt magmas with special reference to the Deccan Traps of the Western Ghats, India. Earth and Planetary Science Letters 84, 59–68.
Devey, C. W. & Lightfoot, P. C. (1986). Volcanological and tectonic control of stratigraphy and structure in the western Deccan Traps. Bulletin of Volcanology 48, 195–207.
Dixon, J. E., Stolper, E. & Delaney, J. R. (1988). Infrared spectroscopic measurements of CO2 and H2O in Juan de Fuca Ridge basaltic glasses. Earth and Planetary Science Letters 90, 87–104.
Duncan, R. A. & Pyle, D. G. (1988). Rapid eruption of the Deccan flood basalts, western India. Nature 333, 841–843.
Ford, C. E., Russell, D. G., Craven, J. A. & Fisk, M. R. (1983). Olivine–liquid equilibria: temperature, pressure and composition dependence of the crystal/liquid cation partition coefficients for Mg, Fe2+, Ca and Mn. Journal of Petrology 24, 256–265.
Fukuoka, T., Arai, F. & Nishio, F. (1987). Correlation of tephra layers in Antarctic ice by trace element abundances and refractive indices of glass shards. Bulletin of the Volcanological Society of Japan 32, 103–118.
Godbole, S. M. & Ray, B. (1996). Intrusive rocks of coastal Maharashtra. Gondwana Geological Magazine Special Volume 2, 233–250.
Godbole, S. M., Deshmukh, S. S. & Chatterjee, A. K. (1996). Geology and chemical stratigraphy of the basalt flows of Akot–Harisal section from Satpura ranges in the eastern part of the Deccan volcanic province. Gondwana Geological Magazine Special Volume 2, 115–124.
Grove, T. L. & Bryan, W. B. (1983). Fractionation of pyroxene-phyric MORB at low pressure: an experimental study. Contributions to Mineralogy and Petrology 84, 293–309.
Helz, R. T. & Thornber, C. R. (1987). Geothermometry of Kilauea Iki lava lake, Hawaii. Bulletin of Volcanology 49, 651–668.
Hooper, P. R. (1990). The timing of crustal extension and the eruption of continental flood basalts. Nature 345, 246–249.
Hooper, P. R. (1994). Sources of continental flood basalts: the lithospheric component. In: Subbarao, K. V. (ed.) Volcanism . New Delhi: Wiley Eastern, pp. 29–53.
Jambon, A. & Zimmermann, J. L. (1990). Water in oceanic basalts: evidence for dehydration of recycled crust. Earth and Planetary Science Letters 101, 323–331.
Johnson, M. J., Anderson, A. T., Jr & Rutherford, M. J. (1994). Preeruptive volatile contents of magmas. In: Carroll, M. R. & Holloway, J. R. (eds) Volatiles in Magma. Mineralogical Society of America, Reviews in Mineralogy 30, 281–330.
Kaneko, T. (1995). Geochemistry of Quaternary basaltic lavas in the Norikura area, central Japan: influence of the subcontinental upper mantle on the trace elements and Sr isotope compositions. Journal of Volcanology and Geothermal Research 64, 61–83.
Khadri, S. F. R., Subbarao, K. V., Hooper, P. R. & Walsh, J. N. (1988). Stratigraphy of Thakurvadi Formation, Western Deccan basalt province, India, in Deccan flood basalts. Memoirs of the Geological Society of India 10, 281–304.
Kinzler, R. J., Donnelly-Nolan, J. M. & Grove, T. L. (2000). Late Holocene hydrous mafic magmatism at the Paint Pot Crater and Callahan flows, Medicine Lake Volcano, N. California and the influence of H2O in the genesis of silicic magmas. Contributions to Mineralogy and Petrology 138, 1–16.
Koyaguchi, T. (1989). Chemical gradient at diffusive interfaces in magma chambers. Contributions to Mineralogy and Petrology 103, 143–152.
Krishnamurthy, P. & Cox, K. G. (1977). Picrite basalts and related lavas from the Deccan Traps of Western India. Contributions to Mineralogy and Petrology 62, 53–75.
Krishnamurthy, P. & Cox, K. G. (1980). A potassium-rich alkalic suite from the Deccan Traps, Rajpipla, India. Contributions to Mineralogy and Petrology 73, 179–189.
Krishnamurthy, P., Gopalan, K. & Macdougall, J. D. (2000). Olivine compositions in picrite basalts and the Deccan volcanic cycle. Journal of Petrology
41, 1057–1069.
Lightfoot, P. C. & Hawkesworth, C. J. (1988). Origin of Deccan Trap lavas: evidence from combined trace element and Sr-, Nd-, and Pb-isotope studies. Earth and Planetary Science Letters 91, 89–104.
Lightfoot, P. C., Hawkesworth, C. J., Devey, C. W., Rogers, N. W. & Van Calsteren, P. W. C. (1990). Source and differentiation of Deccan Trap lavas: implications of geochemical and mineral chemical variations. Journal of Petrology
31, 1165–1200.
Mahoney, J. J. (1988). Deccan traps. In: Macdougall, J. D. (ed.) Continental Flood Basalts . Dordrecht: Kluwer Academic, pp. 151–194.
Mahoney, J. J., Macdougall, D., Lugmair, G. W., Murali, A. V., SankarDas, M. & Gopalan, K. (1982). Origin of the Deccan Trap flows at Mahabaleshwar inferred from Nd and Sr isotopic, and chemical evidence. Earth and Planetary Science Letters 60, 47–60.
Mahoney, J. J., Macdougall, J. D., Lugmair, G. W., Gopalan, K. & Krishnamurthy, P. (1985). Origin of contemporaneous tholeiitic and K-rich alkalic lavas: a case study from the northern Deccan plateau, India. Earth and Planetary Science Letters 72, 39–53.
Mahoney, J. J., Sheth, H. C., Chandrasekharam, D. & Peng, Z. H. (2000). Geochemistry of flood basalts of the Toranmal section, northern Deccan Traps, India: implications for regional Deccan stratigraphy. Journal of Petrology
41, 1099–1120.
Matsuhisa, Y., Bhattacharya, S. K., Gopalan, K., Mahoney, J. J. & Macdougall, J. D. (1986). Oxygen isotope evidence for crustal contamination in Deccan basalts. Terra Cognita 6, 181.
McKenzie, D. & O’Nions, R. K. (1995). The source regions of ocean island basalts. Journal of Petrology
36, 133–159.