Journal of Petrology Advance Access originally published online on July 20, 2005
Journal of Petrology 2005 46(12):2427-2464; doi:10.1093/petrology/egi060
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Hafnium Isotope and Trace Element Constraints on the Nature of Mantle Heterogeneity beneath the Central Southwest Indian Ridge (13°E to 47°E)
1 DEPARTMENT OF TERRESTRIAL MAGNETISM, CARNEGIE INSTITUTION OF WASHINGTON, 5241 BROAD BRANCH ROAD, NW, WASHINGTON, DC 20015, USA
2 DEPARTMENT OF GEOLOGICAL SCIENCES, UNIVERSITY OF CAPE TOWN, PRIVATE BAG, RONDEBOSCH 7701, SOUTH AFRICA
RECEIVED JULY 6, 2004; ACCEPTED JUNE 10, 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSAMPLES AND ANALYTICAL METHODSRESULTSDISCUSSION AND CONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Hafnium isotope and incompatible trace element data are presented for a suite of mid-ocean ridge basalts (MORB) from 13 to 47°E on the Southwest Indian Ridge (SWIR), one of the slowest spreading and most isotopically heterogeneous mid-ocean ridges. Variations in Nd–Hf isotope compositions and Lu/Hf ratios clearly distinguish an Atlantic–Pacific-type MORB source, present west of 26°E, characterized by relatively low
Hf values for a given Nd relative to the regression line through all Nd–Hf isotope data for oceanic basalts (termed the ‘Nd–Hf mantle array line’; the deviation from this line is termed 
Hf) and low Lu/Hf ratios, from an Indian Ocean-type MORB signature, present east of 32°E, characterized by relatively high Hf values and Lu/Hf ratios. Additionally, two localized, isotopically anomalous areas, at 13–15°E and 39–41°E, are characterized by distinctly low negative and high positive Hf values, respectively. The low Hf MORB from 13 to 15°E appear to reflect contamination by HIMU-type mantle from the nearby Bouvet mantle plume, whereas the trace element and isotopic compositions of MORB from 39 to 41°E are most consistent with contamination by metasomatized Archean continental lithospheric mantle. Relatively small source-melt fractionation of Lu/Hf relative to Sm/Nd, compared with MORB from faster-spreading ridges, argues against a significant role for garnet pyroxenite in the generation of most central SWIR MORB. Correlations between Hf and Sr and Pb isotopic and trace element ratios clearly delineate a high-Hf ‘Indian Ocean mantle component’ that can explain the isotope composition of most Indian Ocean MORB as mixtures between this component and a heterogeneous Atlantic–Pacific-type MORB source. The Hf, Nd and Sr isotope compositions of Indian Ocean MORB appear to be most consistent with the hypothesis that this component represents fragments of subduction-modified lithospheric mantle beneath Proterozoic orogenic belts that foundered into the nascent Indian Ocean upper mantle during the Mesozoic breakup of Gondwana. KEY WORDS: mid-ocean ridge basalt; isotopes; incompatible elements; Indian Ocean
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLES AND ANALYTICAL METHODSRESULTSDISCUSSION AND CONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The Southwest Indian Ridge (SWIR) is among the slowest-spreading of all mid-ocean ridges (spreading rate
18 mm/year; Dick et al., 2003
) and is arguably the most geochemically heterogeneous ridge system on Earth (e.g. Hamelin & Allègre, 1985; Mahoney et al., 1992). The high degree of tectonic segmentation and low magma supply at this spreading center (Fisher & Goodwillie, 1997; Cannat et al., 1999) appear to have hindered the dispersal of geochemical anomalies arising from nearby mantle plumes (e.g. le Roex et al., 1983, 1992) and/or passive heterogeneities possibly derived from deeply or shallowly recycled continental material (e.g. Mahoney et al., 1992).
The greatest geochemical variations observed in mid-ocean ridge basalt (MORB) along the SWIR occur in the central portion of the ridge, between the Shaka fracture zone at roughly 10°E and the Indomed Fracture Zone at roughly 47°E (Fig. 1). These variations include localized isotopic anomalies at 13–15°E, which have yielded the most radiogenic 206Pb/204Pb values yet measured for Indian Ocean MORB (le Roex et al., 1992), and the so-called ‘DUPAL’ segment at 39–41°E, which encompasses the lowest 206Pb/204Pb values and among the lowest 143Nd/144Nd values of any MORB yet reported (Hamelin & Allègre, 1985; Mahoney et al., 1992). Superimposed on these short-wavelength features is a broad isotopic transition across the region from Atlantic–Pacific-type normal MORB (N-MORB) compositions in the west, with relatively radiogenic 206Pb/204Pb, moderate 143Nd/144Nd and unradiogenic 87Sr/86Sr, to Indian Ocean-type depleted MORB compositions in the east, with unradiogenic 206Pb/204Pb (but relatively high 207Pb/204Pb and 208Pb/204Pb), and variable 143Nd/144Nd and 87Sr/86Sr (Mahoney et al., 1992).
|
Basalts generated at Indian Ocean spreading centers have long been known to have isotope compositions that are largely distinct from MORB produced in the Atlantic and Pacific oceans (e.g. Subbarao & Hedge, 1973
; Dupré & Allègre, 1983). The Indian Ocean mantle signature is characterized by 87Sr/86Sr values that are more radiogenic and 206Pb/204Pb values that are less radiogenic, particularly for a given 143Nd/144Nd, 207Pb/204Pb or 208Pb/204Pb value, than are encountered in Atlantic or Pacific MORB (e.g. Dosso et al., 1986; Price et al., 1986; Mahoney et al., 1989, 1992, 2002). The origin of this signature has long been the source of controversy, having been attributed to addition of plume material, deeply recycled pelagic sediments, subduction-modified mantle wedge material and continental lithosphere to the Indian Ocean upper mantle (e.g. Storey et al., 1989; Mahoney et al., 1992; Weis & Frey, 1996; Rehkamper & Hofmann, 1997; Kempton et al., 2002; Escrig et al., 2004; Hanan et al., 2004; Zhang et al., 2005).
The Lu–Hf isotope system provides a unique perspective on the origin of mantle heterogeneity (e.g. Patchett & Tatsumoto, 1980a; Salters & Hart, 1991). Lu is slightly more compatible in mantle phases than Hf and in most circumstances the Lu–Hf system closely mirrors the Sm–Nd isotope system, leading to a high degree of correlation between Hf and Nd isotope ratios in oceanic basalts (Patchett & Tatsumoto, 1980a). However, Lu/Hf ratios are more readily fractionated than Sm/Nd by several processes, including melting in the presence of garnet (in which Lu is highly compatible; Irving & Frey, 1978) and metasomatism by carbonated or hydrous fluids (in which Hf is relatively insoluble, e.g. Nelson et al., 1988; Ayers et al., 1997). Also, because of the slower diffusion of Hf relative to Nd in some mantle minerals, particularly garnet, metasomatized mantle peridotites can have radiogenic Hf—a signature of ancient melt depletion—combined with unradiogenic Nd contributed by the metasomatic agent (Bedini et al., 2004; Carlson et al., 2004). Additionally, pelagic sediments display significant decoupling between Hf and Nd isotopes and Lu/Hf and Sm/Nd ratios (e.g. Patchett et al., 1984; Vervoort et al., 1999).
Here, we present Hf isotope and incompatible trace element data for a suite of MORB samples from the central SWIR (Fig. 1) as well as for basalts from the neighboring Bouvet, Marion and Prince Edward islands, and alkalic basalts from the Madagascar large igneous province, which appear to have resulted from the Late Cretaceous activity of the Marion hotspot. This study has three main objectives. The first is to better characterize the geochemistry of MORB from the central SWIR using our extensive trace element and Hf isotope datasets. The second is to use these data in an attempt to constrain the mineralogy and composition of the mantle sources of these basalts and identify the materials and processes responsible for the extreme isotopic compositions present along this ridge, particularly the anomalous DUPAL segment at 39–41°E. The third is to use the geochemical contrast between the unusually clear Atlantic–Pacific and Indian Ocean-type mantle signatures juxtaposed along the central SWIR, as revealed by Hf isotopes, to evaluate the various models that have been proposed for the origin of the Indian Ocean MORB signature.
|
SAMPLES AND ANALYTICAL METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLES AND ANALYTICAL METHODSRESULTSDISCUSSION AND CONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Samples analyzed in this study are from the PROTEA 05 expedition of the R.V. Melville (indicated by prefix ‘P’ in the sample name), cruises 22 and 53 of the S.A. Agulhas (indicated by prefixes ‘AG22’ or ‘AG53’), cruise 34 of the R.V. Marion Dufresne (indicated by prefix ‘MD34’) and cruise ANT IV/4 of the F.S. Polarstern (indicated by prefix ‘PS4’). For the most part, these samples have been described previously for major and some trace element abundances and Sr, Nd and Pb isotope ratios (Hamelin & Allègre, 1985
; le Roex et al., 1989, 1992; Mahoney et al., 1992). Because not all of the MORB samples for which we present data have had major element data reported in the published literature, we provide a full set of major element data for these samples in an electronic appendix, which can be accessed at http://petrology.oupjournals.org/ or obtained from the first author upon request. The vast majority of sample materials analyzed were whole-rock powders made from fresh pillow interiors; a minority were fresh glasses from chilled pillow rinds (indicated with a ‘G’ suffix). All whole-rock samples selected for analysis have loss on ignition (LOI) values of <2·5 wt %, with most samples having LOI <1 wt %. The paucity of available glass is largely due to three factors. First, fresh glass appears to be less abundant at ultra-slow-spreading ridges than at other types of spreading centers because of the longer time intervals between eruptions. Second, the dredging strategy used, particularly on the PROTEA 5 cruise, focused on fracture zones and inside-corner highs to recover peridotites and lower crustal material, rather than focusing on recovery of glassy lavas on-axis. Third, much of the glass that was recovered has been consumed by previous geochemical investigations. However, we have analyzed trace element abundances in two glass–whole-rock pairs (P25-211–P25-211G and P37-2–P37-2G) and found excellent agreement (Table 1).
|
Trace element analyses were performed on dissolved sample solutions by inductively coupled plasma mass spectrometry (ICPMS) using a Perkin Elmer/SCIEX Elan 6000 quadrupole instrument at the Department of Geological Sciences, University of Cape Town. Sample solutions were run in peak-hopping mode, using In, Re and Bi as internal standards. Calibration was performed using multi-element standard solutions. Accuracy and precision were checked by repeated analysis of USGS basalt standard BHVO-1 and in-house MORB glass standard V40-56 (see Table 1 for measured and reference values for these standards).
Hafnium separation and isotope ratio measurements were performed at the Department of Terrestrial Magnetism of the Carnegie Institution of Washington (DTM). Hf separations were conducted using a three-column, cation–anion–cation separation procedure in a HCl–HF medium adapted from Patchett & Tatsumoto (1980b) and Blichert-Toft et al. (1997). Hf recoveries were >75% with Yb/Hf and Lu/Hf ratios of <0·0005 and <0·0002, respectively. Hf blanks are <25 pg and thus are negligible; no blank corrections were applied. Analysis of Hf isotope ratios was performed by ICPMS on a VG Plasma 54 multicollector system (equipped with nine Faraday collectors) in static mode, with 173Yb and 175Lu monitored to allow correction of 176Hf for isobaric interferences by176Yb and 176Lu. All Faraday cup efficiencies were set to unity and Hf isotope ratios were fractionation-corrected to a 179Hf/177Hf value of 0·7325. The Hf isotope data presented here were collected during two analysis periods within a 1 year interval (April 2000 to April 2001). Fifty-three analyses of the JMC 475 Hf standard, run as every third sample, yielded a mean 176Hf/177Hf ratio of 0·282152 with a 2
uncertainty of ±0·000012. To facilitate comparison, all sample data have been normalized to a JMC 475 176Hf/177Hf ratio of 0·28216. Two SWIR samples that we analyzed for Hf isotope composition (MD34-D5 and MD34-D6) were also analyzed by Chauvel & Blichert-Toft (2001). For both samples, our values agree within 2 analytical uncertainties.
Supplementary Sr, Nd and Pb isotope ratios were determined on a subset of samples. Nd and Pb isotope ratio measurements were performed on the VG Plasma 54 at DTM. Sr was separated using standard cation exchange techniques and Sr isotope ratios were measured in multi-dynamic mode by thermal ionization mass spectrometry on the VG 354 instrument at DTM. Sr isotope ratios were fractionation-corrected to 86Sr/88Sr = 0·1194. Repeat analyses of NBS 987 gave a mean 87Sr/86Sr value of 0·710275 ± 18 (2, n = 6), but all data reported are normalized to a 87Sr/86Sr value of 0·710250. Nd was separated from the other rare earth elements (REE) using EiChrom ‘Lanthanide’ resin and measured in static mode. 140Ce and 147Sm were monitored to correct for isobaric interferences of Ce and Sm on 142Nd and 144Nd, respectively. Nd isotope ratios were fractionation-corrected to a 146Nd/144Nd value of 0·7219. Repeat analysis of the La Jolla Nd standard yielded a mean 143Nd/144Nd value of 0·511835 ± 12 (2, n = 10), but sample values were normalized to La Jolla 143Nd/144Nd = 0·51185. Pb was separated by anion exchange chromatography in an HBr medium and Pb isotope ratio measurements were performed in static mode using the thallium normalization technique (Rehkämper & Halliday, 1998) for mass bias correction. All Tl normalized samples and standards were prepared with 208Pb/205Tl ratios of 0·2 ± 0·04. Pb isotope measurements were conducted by bracketing the analysis of each sample with analyses of the NBS 981 Pb standard. Measured Pb isotope ratios for NBS 981 using Tl normalization yielded 206Pb/204Pb and 207Pb/204Pb values within 0·015 and 208Pb/204Pb values within 0·030 of the reference values of Todt et al. (1996), to which the Pb isotope ratio values of all samples were normalized. Procedural blanks for Sr, Nd and Pb are less than 150 pg, 20 pg and 100 pg, respectively, and are negligible.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLES AND ANALYTICAL METHODSRESULTSDISCUSSION AND CONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Incompatible trace elements
MORB from the central SWIR display wide ranges of incompatible trace element concentrations (Table 1) and ratios that suggest derivation by variable degrees and depths of partial melting of a number of variably depleted and enriched mantle sources. Seawater alteration appears to have had little effect on the trace element compositions of the samples. Only the highly mobile elements Rb and Cs and, to a lesser extent, U show evidence for alteration-related disturbance relative to less mobile elements such as Th, Nb and the light rare earth elements (LREE) in the whole-rock samples (Fig. 2).
|
|
One characteristic shared by nearly all of the SWIR samples is a slight depletion in the heavy REE (HREE) relative to the middle REE (MREE), resulting in (Tb/Yb)N ratios (the subscript N indicates normalization to chondritic values; Evensen et al., 1978
) that are slightly higher [mean = 1·22 ± 0·08 (1) excluding samples with (La/Sm)N >1] than those of average N-MORB from slow- to fast-spreading centers [>2 mm/year full rate; mean = 1·08 ± 0·11(1) excluding samples with (La/Sm)N >1; PETDB MORB database, 2004, available at http://www.petrology.oupjournals.org]. This mild HREE depletion appears to be common to MORB from ultra-slow-spreading centers, as it is also found in MORB from the Mid-Cayman Rise, American–Antarctic Ridge and Gakkel Ridge (le Roex et al., 1985; Mühe et al., 1993, 1997; Elthon et al., 1995). Along axis, one of the most notable regional (as opposed to local) variations in incompatible element composition occurs between 26°E and 32°E, where there is a marked eastward increase in Lu/Hf ratio both overall (Fig. 3a) and for a given Sm/Nd ratio (Fig. 3b). This shift in Lu/Hf values also marks a major division in the Hf–Nd–Sr isotopic composition of central SWIR MORB, as will be shown in the next section. The regional Lu/Hf variations therefore appear to represent long-lived features of the mantle source, and are not simply caused by regional variations in depth of melting or source mineralogy. The incompatible element compositions of central SWIR MORB, divided into six sections, are described below.
|
13–15°E: enriched MORB
Dredging of the 13–15°E portion of the extremely slow oblique-spreading ridge segment at 10–16°E (8 mm/year effective spreading rate; Dick et al., 2003
) recovered lava types ranging from olivine tholeiites to moderately nepheline-normative alkali basalts. The four samples from this segment selected for trace element analysis encompass nearly the full range of incompatible element enrichment and silica saturation or undersaturation of the suite reported by le Roex et al. (1992). The samples are characterized by the highest (La/Sm)N and Nb/Zr ratios yet found for SWIR MORB (1·5–3·4 and 0·09–0·31, respectively; and primitive mantle-normalized incompatible element patterns with strong positive Nb anomalies (Fig. 2a). The two most incompatible-element enriched alkali basalts (PS4-3-14 and PS4-4-101) are nearly indistinguishable from Bouvet Island basalts in terms of their incompatible element compositions.
17–26°E: normal (N-type) MORB
Samples in this section come from the long orthogonal spreading segment at 16–25°E, as well as the DuToit Fracture Zone (hereafter termed ‘DuToit FZ’) and the western portion of the short 26–28°E spreading segment between the DuToit and Bain fracture zones (Fig. 1). Most of the lavas from this area are normal (N-type) MORB with (La/Sm)N (0·63–0·82), Ba/Nb (1·5–4·4) and Nb/Zr (0·02–0·04) ratios that vary from slightly below to slightly above average N-MORB values (see Hofmann, 1988; Sun & McDonough, 1989). The trace element patterns for the five on-axis samples from the 16–25°E segment and sample P9-1G from the 26–28°E segment are very similar, particularly for the mildly incompatible elements (i.e. Sm to Lu; Fig. 2b). As a result, these lavas display restricted ranges of (Sm/Yb)N and (Lu/Hf)N ratios (1·29–1·41 and 0·63–0·71, respectively). In contrast, the three MORB samples from the DuToit FZ show greater variability in their trace element patterns for this range of elements, and two of the three samples (P11-17 and P12-12) have mild negative Hf anomalies and slightly higher ratios of MREE to HREE, giving them relatively high (Sm/Yb)N and (Lu/Hf)N ratios (1·48–1·58 and 0·75–0·79, respectively; Fig. 2c). The cause of the negative Hf anomalies in these fracture zone samples is unknown, as their Zr and Ti contents are not similarly depleted.
In addition to the N-MORB lavas described above, a small number of vesicular basalts and andesites lacking glassy rinds were recovered in PROTEA Expedition dredges 14 and 9 (Fisher et al., 1985), located in the DuToit FZ and on the 26–28°E spreading segment, respectively. These samples include P14-89, P14-90, and P9-4 described by le Roex et al. (1989), Mahoney et al. (1992) and this study, and P9-7 described only by Mahoney et al. (1992). They are remarkable in that they have extreme depletions in Nb and very strong relative enrichments in Ba and Pb (e.g. Ba/Nb = 60–120, Ce/Pb = 3·1–4·7, le Roex et al., 1989; Mahoney et al., 1992; Fig. 2d). These traits are more extreme than those for any other MORB sample yet published (see Kamenetsky et al., 2001; Le Roux et al., 2002a; Fig. 2d). The combination of their unusual appearance and major element similarity to arc tholeiites (le Roex et al., 1989), led Mahoney et al. (1992) to speculate that these rocks are exotic lava fragments originating at the South Sandwich island arc and carried eastward as ice-rafted debris. Figure 2d clearly shows that the incompatible element patterns for these anomalous lavas display an extraordinary similarity to those for the dominantly tholeiitic, basaltic to andesitic lavas forming the South Sandwich Islands (Pearce et al., 1995). This strong trace element similarity, along with isotopic similarities between these dredge samples and South Sandwich Island lavas (Cohen & O'Nions, 1982; Mahoney et al., 1992; Pearce et al., 1995), appears to confirm the hypothesis of Mahoney et al. (1992) for the exotic origin of these lavas.
32–34°E: normal and transitional MORB
Five MORB samples were analyzed from this area of the ridge, obtained from one dredge in the northeastern part of the Bain Fracture Zone and three dredges on the short 32·5°–34°E spreading segment between the Bain and Prince Edward fracture zones. Although roughly half as many MORB samples were analyzed from this section as compared with 17–26°E, these samples display greater variation in their ratios of highly and moderately incompatible elements [e.g. (La/Sm)N = 0·62 to 0·86, Nb/Zr = 0·02 to 0·05; Ba/Nb = 3·1 to 7·6; Table 1]. As noted above, these samples also tend to display higher (Lu/Hf)N ratios than those from the 17–26°E section (0·74 to 1·0 vs 0·63 to 0·79, respectively; Fig. 3a).
36–39°E: Marion-influenced MORB
The segment at 36–39°E, located between the Prince Edward and Simpson fracture zones, is the portion of the SWIR that lies closest to Marion and Prince Edward Islands, and presumably to the present-day location of the Marion hotspot. Most lavas dredged from this segment are mildly enriched in incompatible elements, but one sample (P25-217) is moderately depleted (e.g. La/SmN = 0·7–1·5, Nb/Zr = 0·05–0·12, Ba/Nb = 5·0–7·2). All samples have primitive mantle-normalized incompatible element patterns with positive Nb anomalies. In this, they resemble lavas from Marion and Prince Edward Islands (Table 2, Fig. 2f), although they do not attain the same degree of incompatible element enrichment as these ocean island lavas (e.g. La/SmN = 2·1–2·8, Nb/Zr = 0·14–0·17, Ba/Nb = 7–8).
39–41°E: ‘Dupal-type’ MORB
Basalts from the 39–41°E segment are among the most isotopically unusual basalts recovered from the global mid-ocean ridge system (e.g. Hamelin & Allègre, 1985; Hart et al., 1986; Mahoney et al., 1992), and are particularly notable for their extremely unradiogenic 143Nd/144Nd and 206Pb/204Pb ratios, combined with relatively high 207Pb/204Pb and 208Pb/204Pb for a given 206Pb/204Pb value (i.e. high 7/4 and 8/4; Hart, 1984). These samples are characterized by mild enrichments in the LREE [(La/Sm)N = 1·1–1·6], and, relative to the LREE, strong enrichments in Ba (Ba/La = 13–19, Na/Nb = 12/24, Ba/Th = 172–312), variable enrichments in Pb (Ce/Pb = 13–28) and strong depletions in Th and U (Fig. 2g). Unlike the anomalous lavas from the DuToit Fracture Zone region, these samples are non-vesicular, have glassy chilled margins and are, from appearance, indistinguishable from other SWIR MORB. It should also be noted that these characteristics are displayed by both whole-rock lavas and fresh basaltic glass (e.g. sample P28-8G) and thus appear to be primary features and not the result of seawater alteration.
The strong enrichments in Ba and Pb and, in the case of the most isotopically extreme sample (MD34-D5), relative depletions in Zr and Hf, are suggestive of the addition of a subduction or continental component to the sources of these lavas. Ba and Pb enrichments and high field strength element depletions are characteristic of many arc lavas, as well as mean continental crust, pelagic clays and some types of metasomatized peridotite xenoliths from the continental lithospheric mantle (e.g. Rudnick & Fountain, 1995; Plank & Langmuir, 1998; Grégoire et al., 2003).
42–47°E: heterogeneous MORB
This portion of the ridge, extending from the Discovery II to the Indomed fracture zones, was only sparsely sampled by the PROTEA 5 and Marion Dufresne 34 cruises, but the recovered samples span a wide geochemical range. They include incompatible element-depleted normal MORB, relatively undepleted transitional MORB and incompatible element-enriched (E-type) MORB [e.g. Nb/La = 0·8–1·4; Nb/Zr = 0·02 to 0·23; (La/Sm)N = 0·63–3·14; Fig. 2h]. The E-MORB samples come from a single on-axis dredge (PROTEA 5 dredge 37) near 46°E, just west of the Indomed fracture zone and far from any known intraplate volcanic features. These samples display incompatible element patterns with negative slopes from Ba to Lu and mild positive Nb anomalies. The most enriched of these, P37-1, a mildly fractionated, nepheline-normative basalt, has an incompatible element pattern similar to that of average ocean island basalt (OIB) (Sun & McDonough, 1989). Also from this region is the sample MD34-D4, which has, along with positive Ba and Pb anomalies (Fig. 2h), a remarkably radiogenic Hf isotope composition (Chauvel & Blichert-Toft, 2001).
Hafnium isotope characteristics of central SWIR MORB
Hafnium isotope data for 30 central SWIR MORB samples investigated in this study are presented in Table 3 together with their Sr, Nd and Pb isotopic compositions. The data span a wide range of 176Hf/177Hf values (0·28279–0·28327; Hf = +0·6 to +17·7) encompassing the values of most OIB and MORB (e.g. Salters, 1996; Chauvel & Blichert-Toft, 2001). Plotted vs longitude (Fig. 4), Hf values display a roughly linear gradient from about +10 in the west to about +17 in the east of the study area, punctuated by localized regions of low Hf at 13–15°E, 39–41°E, and 46°E. This pattern is similar to that observed for Nd isotopes, which show a gradual eastward increase in Nd (from approximately +7·5 to +10; Mahoney et al., 1992), although the gradient in Hf isotopes is significantly greater relative to the total range of variation. The trend of Hf vs longitude is essentially the inverse of that for 206Pb/204Pb with the major exception of the 39–41°E segment, which produced MORB having both unradiogenic 206Pb/204Pb and 176Hf/177Hf ratios. No systematic gradient is present in 87Sr/86Sr values, but the three regions characterized by anomalously unradiogenic Hf are also characterized by radiogenic 87Sr/86Sr.
|
|
As with Nd isotope compositions (Fig. 5), Hf isotope compositions of central SWIR MORB correlate negatively with both 87Sr/86Sr and 206Pb/204Pb (Fig. 6). However, the
Hf values correlate markedly worse with 87Sr/86Sr (r2 = 0·55 vs 0·78) and better with 206Pb/204Pb (r2 = 0·79 vs 0·39, if the 39–41°E samples are excluded) than do Nd values for the same samples. The relatively poor correlation of Hf with 87Sr/86Sr is due to the fact that the data form two elongated, roughly parallel groupings in 87Sr/86Sr–Hf space (Fig. 6a), composed of samples from 13 to 26°E and 32 to 47°E, that are vertically offset from each other by 
5 Hf units. No such clear separation is apparent in the plot of 87Sr/86Sr vs Nd (Fig. 5a). The good correlation of Hf with 206Pb/204Pb (apart from the 39–41°E samples) is due to the strong, nearly linear east–west gradients in both of these isotopic tracers (Fig. 6b). In contrast, the milder and less well-defined east–west gradient in Nd isotope composition yields a flatter overall slope and a large spread in 206Pb/204Pb ratios (17·8–19·0) at intermediate Nd values of +7 to +9 (Fig. 5b), resulting in a relatively poor correlation.
|
|
Like nearly all data for oceanic basalts, the central SWIR MORB samples investigated fall within 5
Hf units of the ‘Nd–Hf mantle array’ line on a plot of Nd vs Hf (Fig. 7). This line is the linear regression through the oceanic basalt Nd–Hf isotope dataset, defined by the equation Hf = 1·33 x Nd + 3·19 (Vervoort et al., 1999). Small, regionally coherent deviations from this line do occur. As in the plot of 87Sr/86Sr vs Hf (Fig. 6a), samples from the western (13–26°E) and eastern (32–47°E) regions have systematically lower and higher Hf values, respectively, for a given Nd value. Whereas samples from the western region fall below the Nd–Hf mantle array line by 1–4 Hf units, those from the eastern region fall either near or above this line (i.e. between 1 Hf unit below and 3 Hf units above this line). Sample MD34-D4 from 43°E (Chauvel & Blichert-Toft; 2001) has the highest Hf value (+21) reported for an Indian Ocean MORB outside the Australian–Antarctic Discordance (AAD; see Hanan et al., 2004) and plots 6·7 Hf units above the Nd–Hf mantle array (Fig. 7).
|
The Nd–Hf isotopic data arrays for the 13–26°E and 32–47°E regions have slopes of 1·30 (r2 = 0·85) and 1·17 (r2 = 0·88), respectively, that vary only slightly from each other and that of the Nd–Hf mantle array line (1·33; Vervoort et al., 1999
). Without the 39–41°E segment data (which exert a strong control on the slope), the remaining samples from 32 to 47°E form a shallower but less coherent array with a slope of 0·99 (r2 = 0·54).
The large ranges in Hf isotope composition (6–10 Hf units) of SWIR MORB at both relatively depleted and enriched Sr and Nd isotope compositions (i.e. at 87Sr/86Sr <0·7027 and >0·7035, respectively, and Nd >+8 and <+5, respectively) suggests the presence of multiple depleted, as well as enriched mantle source components beneath the central SWIR, each with a distinct time-integrated decoupling of Lu/Hf from Sm/Nd relative to compositions lying on the Nd–Hf isotopic mantle array line. A useful way to constrain the isotopic and trace element characteristics of these end-members is to determine whether the variability in Hf isotope ratios of SWIR MORB at a given Nd isotopic composition varies systematically with other isotopic and trace element ratios. This can be accomplished through the use of the parameter Hf (Johnson & Beard, 1993), which is defined as the vertical displacement of a given sample in Hf units from the Nd–Hf isotopic mantle array line. Across the study area, Hf displays a roughly linear increase with longitude (Fig. 8). Figures 9 and 10 illustrate that the Hf values of most samples (apart from those from 13 to 15°E and 39 to 41°E) correlate weakly to very well with 87Sr/86Sr, 206Pb/204Pb, 7/4 and 8/4 and some trace element ratios, such as Ba/Nb and Ba/Th. Interestingly, most data from other Indian Ocean ridges lie on or near these same correlation trends (Salters, 1996; Chauvel & Blichert-Toft, 2001), indicating that the processes responsible for decoupling Lu/Hf from Sm/Nd affect other isotope and trace element ratios in a similar fashion throughout the Indian Ocean mantle domain. Significantly, samples from the 39–41°E ‘Dupal’ segment constitute vectors in these figures that are strongly oblique to the main data arrays.
|
|
|
Hafnium isotope characteristics of intraplate samples
Hf, and supplementary Sr, Nd and Pb isotope ratios were determined for basalts and hawaiites from Bouvet, Marion and Prince Edward Islands, volcanic islands presumably marking the present location of the Bouvet and Marion hotspots, as well as for Cretaceous alkali basalt and basanite dikes from southern Madagascar (Table 4). The age and location of the latter suggests that they are the products of Cretaceous activity of the Marion hotspot. They also have strong Sr, Nd and Pb isotopic similarities to 39–41°E SWIR MORB (Mahoney et al., 1991
).
|
In Nd–Hf isotope space (Fig. 7), the Bouvet samples fall slightly below the mantle array and within the field for St. Helena OIB (Salters & White, 1998
), consistent with its moderate HIMU Sr–Nd–Pb isotope signature (e.g. Sun, 1980). The Marion and Prince Edward Island lavas also fall below the mantle array, but at higher Nd values intermediate between St. Helena OIB and Atlantic–Pacific MORB. Interestingly, three samples analyzed from the 10–16°E oblique-spreading segment of the SWIR have Nd–Hf (as well as Sr and Pb) isotopic compositions similar to lavas from Bouvet Island, even though these MORB samples were emplaced over 600 km to the east. In contrast, MORB from the 36–39°E segment of the SWIR, emplaced less than 250 km from Marion and Prince Edward Islands, show little to no Nd–Hf isotopic similarity to OIB from these islands and lie at significantly higher Nd and Hf values (Fig. 7).
The alkaline Ejeda–Bekily dike samples analyzed are part of the 88 Ma Madagascar large igneous province (LIP; Mahoney et al., 1991; Dostal et al., 1992). The Madagascar LIP encompasses tholeiitic and alkalic lava sequences and dikes emplaced mainly in the eastern and western coastal regions of Madagascar, and the adjacent submarine Madagascar Plateau, a large basaltic edifice extending south of the island (Storey et al., 1998). This igneous province lies along the Late Cretaceous trace of the Marion hotspot (Hartnady & le Roex, 1985) and a significant portion of the lavas bear an isotopic resemblance to those of Marion Island (Mahoney et al., 1991; Storey et al., 1998). The Ejeda–Bekily dikes of southern Madagascar are composed of relatively primitive alkali basalts and basanites that are notable for having radiogenic Sr and unradiogenic Nd and Pb isotope ratios approaching or matching those of MORB from the 39–41°E ‘Dupal’ segment of the SWIR (e.g. Nd 
–2·5, 206Pb/204Pb 17·4; Mahoney et al., 1991). Like these unusual MORB, the dike samples tend to have high Hf for their relatively unradiogenic Nd isotope compositions (Hf = +0·3 to +9·9, Hf = +0·4 to +3·5; Fig. 7).
|
DISCUSSION AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLES AND ANALYTICAL METHODSRESULTSDISCUSSION AND CONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The Hf isotope and trace element data presented above allow us to place new constraints on the mineralogy, composition and origin of mantle sources present beneath the central SWIR. We use these data in the discussion below to address four questions. First, do the Sm–Nd and Lu–Hf isotope systematics of central SWIR MORB indicate that garnet pyroxenite was a regionally important source component in their genesis, as has been recently inferred for MORB from the western SWIR (Salters & Dick, 2002
)? Second, what is the most likely origin of the extreme geochemical signature expressed by MORB from 39 to 41°E? Third, what are the regionally important mantle source components responsible for the main isotopic variations of central SWIR MORB? Fourth, what constraints do Hf isotope data place on the origin of the distinct isotopic characteristics of Indian Ocean MORB?
Constraints on the role of garnet pyroxenite as a melt source in central SWIR MORB
An apparent inconsistency exists between the relatively shallow depths of magma generation necessary to produce the observed thickness of oceanic crust (typically 5–7 km) as inferred from major element-based melting models (e.g. Klein & Langmuir, 1987; Niu & Batiza, 1991) and the apparent requirement from trace element and isotope constraints (e.g. Salters & Hart, 1989; Bourdon et al., 1996) that significant amounts of MORB melting has occurred in the presence of garnet. This has prompted the hypothesis that pods or veins of highly fertile garnet pyroxenite are widespread in the upper mantle (e.g. Hirschmann & Stolper, 1996). Garnet in such pyroxenite veins would probably be stable to depths significantly shallower than the top of the garnet stability field in peridotite (70 km; O'Neill, 1981) and, because garnet pyroxenite has a lower solidus temperature than peridotite, these veins should contribute disproportionately to melts generated during MORB genesis (Perterman & Hirschmann, 2003). Although some recent studies of clinopyroxene partitioning (e.g. Blundy et al., 1998) and trace element variations (Chauvel & Blichert-Toft, 2001) have suggested that garnet need not be involved in the genesis of most MORB, this contention remains hotly debated (e.g. Salters & Longhi, 1999; Salters et al., 2002). Nevertheless, the notion of an upper mantle veined with garnet pyroxenite is an intriguing and frequently explored hypothesis for both the origin of the apparent garnet signature in MORB and as a source of compositional heterogeneity in the upper mantle (e.g. Lassiter et al., 2000; Le Roux et al., 2002b; Salters & Dick, 2002).
If garnet pyroxenite does indeed contribute a melt component during MORB genesis, then one might expect this component to be least diluted in lavas erupted at deep, ultra-slow-spreading centers such as the SWIR, the Mid-Cayman Rise, American–Antarctic Ridge and the Gakkel Ridge. Both the mean degree and rate of melting at such ridges are low, as indicated by lava chemistry and low crustal thickness (Bown & White, 1994; Robinson et al., 1996). Because of the lower solidus temperatures of pyroxenite relative to peridotite (e.g. Hirschmann & Stolper, 1996) melting at an ultra-slow-spreading center should generate melts with a larger contribution from a garnet pyroxenite component (if present), than melting beneath a faster-spreading center.
Globally, N-MORB from ultra-slow-spreading centers tend to have slightly greater depletions in HREE relative to MREE than N-MORB from slow- to fast-spreading ridges [with (Tb/Yb)N means of 1·22 ± 0·06 and 1·08 ± 0·11 (1), respectively, excluding enriched samples with (La/Sm)N >1; le Roex et al., 1985; Price et al., 1986; Mühe et al., 1993, 1997; Elthon et al., 1995; Müller et al., 1997; Lamont PETDB petrological database]. Although this difference is not statistically significant, it may indicate that MORB from ultra-slow-spreading centers received a greater contribution from melts of garnet-bearing mantle lithologies than MORB from faster-spreading ridges. Because melting is believed to commence at relatively shallow depths beneath these deep spreading centers (Klein & Langmuir, 1987), the apparent garnet signature in the REE patterns of basalts from ultra-slow-spreading ridges would be consistent with the presence of shallow garnet pyroxenite veins in the underlying mantle.
The 176Hf/177Hf and 143Nd/144Nd ratios of MORB allow us to estimate the time-integrated Lu/Hf and Sm/Nd ratios of each sample's source. This information, combined with measured Lu/Hf and Sm/Nd ratios, makes it possible to calculate the extent of fractionation of Lu/Hf and Sm/Nd between source and melt. These fractionations are termed 
(Lu/Hf) and (Sm/Nd), and are defined as
 | (1) |
(Sm/Nd) being similarly defined. The term 176Lu/177Lusource is determined as the 176Lu/177Hf value on a mantle isochron (formed by fractionating a chondritic reservoir at 2 Ga) that corresponds to the sample's 176Hf/177Hf ratio (Salters & Hart, 1989). The choice of a 2 Ga age for the MORB source in this formulation is not critical: the use of an older or younger age would affect the absolute, but not the relative values of (Lu/Hf) and (Sm/Nd).
Garnet fractionates Lu from Hf to a much greater extent than Sm from Nd (e.g. Irving & Frey, 1978). Therefore, garnet pyroxenite, which typically has substantially higher garnet modes than suboceanic garnet lherzolite (e.g. Bodinier et al., 1987), should yield melts having compositions falling near the low end of the range of Lu/Hf ratios measured in MORB. Garnet pyroxenite melts should also record significantly larger (Lu/Hf) values relative to a given (Sm/Nd) than other MORB (i.e. show a significantly greater decrease in Lu/Hf for a given decrease in Sm/Nd during melting; Hirschmann & Stolper, 1996).
Contrary to these expectations, the central SWIR lavas analyzed (excluding samples that are obviously the product of isotopically anomalous mantle sources, i.e. those from 13 to 15°E, 39 to 41°E, and 46°E) display a range of Lu/Hf values similar to that of MORB from the Atlantic and Pacific Oceans (Fig. 11). On the plot of 176Lu/177Hf vs 176Hf/177Hf (Fig. 11a) and (Lu/Hf) vs (Sm/Nd) (Fig. 11b) all but two of these ‘normal’ SWIR samples lie within the fields defined by data for MORB from the East Pacific Rise (EPR) and northern Mid-Atlantic Ridge (MAR; Salters, 1996; Nowell et al., 1998). More significantly, none of these samples have (Lu/Hf) or (Sm/Nd) values that fall significantly above the 40% contour (of Salters & Hart, 1989) for the percentage of melting occurring in the garnet stability field. This suggests that they have had even less contribution of melt from garnet-bearing lithologies than many samples from the northern MAR.
|
The fact that the central SWIR lavas record relatively small Lu/Hf fractionations during melting, and have
(Lu/Hf) and (Sm/Nd) values very similar to that of MORB from the EPR (which has substantially higher rates and mean degrees of melting), argues against an origin for the garnet signature in most SWIR lavas from shallow melting of garnet pyroxenite veins. Rather, it suggests that, like EPR MORB, these lavas were generated by melting that began within the field of garnet stability in peridotite. A pyroxenitic source component is more plausible for MORB from 13 to 15°E on the obliquely spreading section of the SWIR, which have both low Lu/Hf and particularly high (Lu/Hf) for their (Sm/Nd) values, most of which fall above the 40% contour for the percentage of melting in the presence of garnet (Fig. 11). Indeed, Salters & Dick (2002) found independent evidence for a garnet pyroxenite component in this region as a result of the lack of overlap in Nd isotopic composition and REE concentrations between peridotites (or calculated equilibrium melts of peridotites) and basalts from this segment. In contrast, the relatively high 176Lu/177Hf ratios (for their low 176Hf/177Hf ratios) and low (Lu/Hf) and (Sm/Nd) values for 39–41°E MORB appear to be inconsistent with a major source contribution from garnet pyroxenite.
A full discussion of how the low degrees of melting, thin crust (<4 km, Dick et al., 2003) and relatively great depths of melt generation inferred for the SWIR can be reconciled is outside the scope of this paper. However, we note that N-MORB from the central SWIR are characterized by lower Si8/Fe8 ratios (Si8 and Fe8 are SiO2 and FeO contents corrected for fractionation to 8 wt % MgO) than are typical for MORB from other ultra-slow-spreading centers (5 for the SWIR vs 6–7 for the American–Antarctic Ridge and Mid-Cayman Rise; le Roex et al., 1985, 1989; Elthon et al., 1995; Robinson et al., 1996), which are indicative of higher mean pressures of melting (Niu & Batiza, 1991). The extremely high degree of tectonic segmentation of most of the central SWIR may have resulted in unusually effective conductive cooling of the uppermost mantle, deepening the top of the mantle melting column beneath the SWIR (e.g. Bown & White, 1994). If mantle melting beneath ultra-slow-spreading ridges does typically commence within the garnet stability field (even slightly below the garnet–spinel transition), the suppression of melting in the upper part of the spinel stability field, normally the interval of highest melt productivity beneath spreading centers (Asimow et al., 1997), would result in a higher than normal proportion of melting occurring in the garnet stability field. MORB from this environment might have proportions of melts derived from garnet and spinel lherzolite similar to that of MORB produced at more vigorous, faster-spreading ridges, where the melting interval both starts at greater depth and ends at shallower depth than beneath the central SWIR. Comparisons of the (Lu/Hf) and (Sm/Nd) values of the central SWIR lavas with those of MORB from less segmented ultra-slow ridges, such as the Gakkel Ridge (which are not yet available), would be useful for testing this hypothesis.
The origin of 39–41°E MORB signature
The extremely unradiogenic Nd and Pb isotopic compositions of MORB from the 39–41°E segment require that their source contain a time-integrated LREE-enriched, low µ (238U/204Pb) component. On the basis of depleted mantle Nd model ages for these samples, which provide only minimum age constraints, this enriched component must have been isolated from the convecting mantle for at least 1·5 Gyr. Several different origins have been proposed to explain the extreme isotopic signature of the 39–41°E segment lavas. These include contamination of the upper mantle by (1) one or more mantle plumes in the region (e.g. Mahoney et al., 1989, 1992), (2) focused recycling into the upper mantle of ancient recycled pelagic sediments (Hamelin & Allègre, 1985; le Roex et al., 1989), (3) disrupted or delaminated ancient continental lithospheric mantle (Hamelin & Allègre, 1985; Mahoney et al., 1989, 1992) or (4) lower continental crust (Escrig et al., 2004). These hypotheses are not necessarily mutually exclusive, as mantle plumes are widely considered to be a likely agency of both disruption of continental lithosphere and transport of ancient, deeply recycled pelagic sediments into the upper mantle (e.g. Storey, 1995; Blichert-Toft et al., 1999).
The new trace element and Hf isotope data presented here offer the opportunity to place tighter constraints on the origin of the 39–41°E MORB source. More broadly, it is important to identify the exotic component responsible for the extreme isotopic composition of 39–41°E MORB because each of the candidate materials mentioned above has also been proposed to be responsible for the distinct isotopic signature of the Indian Ocean mantle domain as a whole (Storey et al., 1989; Mahoney et al., 1992; Weis & Frey, 1996; Rehkämper & Hofmann, 1997; Chauvel & Blichert-Toft, 2001; Escrig et al., 2004; Hanan et al., 2004). Establishing that one or more of these exotic materials was involved in producing the localized MORB signature at 39–41°E would make the involvement of that or similar materials in the entire Indian Ocean mantle domain more credible.
The isotopic signature of the 39–41°E MORB is unlikely to be wholly due to direct contamination by mantle plumes. The Nd and Pb isotopic compositions of basalts generated by all Indian Ocean plumes (e.g. Reunion, Marion, Crozet, Amsterdam–St. Paul and Kerguelen) are simply too radiogenic to account for the isotopic characteristics of these MORB (Mahoney et al., 1992). The location of the 39–41°E segment adjacent to the trace of the Marion hotspot does suggest, however, that the Marion plume may have played an important indirect role in mixing the enriched ‘DUPAL’ component into the local asthenosphere.
Over 1–2 Gyr, subducted pelagic sediments should evolve highly radiogenic 87Sr/86Sr and strongly unradiogenic Nd and Pb isotope ratios (e.g. Ben Othman et al., 1989; Rehkamper & Hofmann, 1997), making them plausible as the material responsible for the ‘DUPAL’ isotope signature of 39–41°E MORB. Because they are deficient in zircon (a major host of Hf in the continental crust), pelagic marine sediments also tend to have low Sm/Nd and relatively high Lu/Hf, which results in decoupled unradiogenic143Nd/144Nd and relatively radiogenic176Hf/177Hf ratios with time (e.g. Patchett et al., 1984; Vervoort et al., 1999). The addition of such material to the sources of oceanic basalts causes their compositions to plot off the mantle array toward low Nd values, resulting in Nd–Hf arrays displaced above the mantle array (e.g. Blichert-Toft et al., 1999), similar to that observed for the 39–41°E segment lavas.
Significant differences in the incompatible element compositions of pelagic sediment and 39–41° MORB, however, indicate a problem with ancient pelagic sediments as the ‘DUPAL’ component present at 39–41°E. Although the the 39–41°E MORB are characterized by relative enrichments in Ba and Pb, qualitatively similar to pelagic and global subducting sediment, these MORB are also characterized by pronounced depletions in U and Th relative to elements of similar incompatibility such as Nb and La. This results in 39–41°E MORB having Nb/U ratios similar to those of the depleted mantle (39–68). Pelagic sediments and altered oceanic crust, in contrast, are typically enriched in U and have low Nb/U ratios (typically 5–10; Staudigel et al., 1996; Plank & Langmuir, 1998). It is clear from a plot of Nb/U vs La/Th (Fig. 12) that mixtures of subducting sediment and altered oceanic crust cannot explain the trace element ratios of 39–41°E MORB. Moreover, the maximum contribution that subducting sediments could have made to the mantle sources of these MORB is very small, <0·1%. Thus, pelagic sediments do not provide an adequate explanation for the geochemical features of 39–41°E MORB.
|
Ancient metasomatized continental lithospheric mantle and lower continental crust are both plausible source materials for isotopically anomalous oceanic basalts in the South Atlantic and Indian Oceans (Hamelin & Allègre, 1985
; Mahoney et al., 1989, 1992; Milner & le Roex, 1996; Douglass et al., 1999; Kamenetsky et al., 2001). Isotopic studies of peridotite xenoliths from Archean cratons and lower crustal granulite xenoliths from Archean and Proterozoic terranes in the Gondwanan continents have shown that both of these materials may have sufficiently unradiogenic 143Nd/144Nd and 206Pb/204Pb ratios and sufficiently high 7/4, 8/4 and 87Sr/86Sr values to explain the isotope composition of 39–41°E SWIR MORB via mixing with depleted MORB-source mantle (e.g. Menzies & Murthy, 1980; Cohen et al., 1984; Rudnick et al., 1986; Walker et al., 1989; Huang et al., 1995; Vervoort et al., 2000). In contrast, mantle xenoliths from Proterozoic and younger continental terranes do not typically have sufficiently extreme Nd and Pb isotope compositions to constitute a credible mixing end-member for these lavas (e.g. McDonough & McCulloch, 1987; Lee et al., 1996; Carlson et al., 2004).
Distinguishing between ancient continental lithospheric mantle and lower continental crustal contamination of oceanic basalt sources has proven to be difficult and remains the subject of considerable controversy (e.g. Milner & le Roex, 1996; Douglass et al., 1999; Hanan et al., 2004), as these materials can have similar Sr, Nd and Pb isotope compositions (Cohen et al., 1984; Rudnick et al., 1986). Much of the isotopic variablility in kimberlite-borne cratonic peroditite xenoliths (constituting the vast majority of Archean peridotites analyzed), particularly in Pb isotope composition, may be due to infiltration of the xenolith by the host kimberlite (which compromises bulk analyses; Grégoire et al., 2003) or precipitation of metasomatic clinopyroxene by kimberlitic magmas shortly before eruption (Simon et al., 2003). These complications have made it very difficult to estimate the mean isotopic composition of cratonic Gondwanan lithospheric mantle.
Fortunately, recent trace element and isotopic studies have provided improved constraints on the incompatible element and isotopic composition of southern African cratonic lithospheric mantle and Precambrian lower crust. Grégoire et al. (2003) have presented an extensive set of in situ trace element data, obtained via laser ablation-ICPMS, of phases in a suite of representative peridotite xenoliths from the Archean Kaapvaal Craton of southern Africa. These data allow calculation of the bulk compositions of Archean southern African peridotites for a large suite of incompatible elements, free of contamination by the host kimberlite. The trace element patterns of these peridotites fall into two categories suggestive of different metasomatic agents or processes. Type 1 peridotites, as defined by Grégoire et al. (2003), display fairly smooth trace element patterns, similar to what might be expected for metasomatism by silicate melts derived by moderate degrees of melting of a depleted asthenospheric source. Type 2 peridotites have some characteristics in common with the 39–41°E SWIR lavas, including strong relative enrichments in Ba and the LREE, depletions in U, Th, Hf, Zr and Ti (Fig. 13) and relatively high (Lu/Hf)N/(Sm/Nd)N ratios (mean = 1·4), all of which may be attributable to metasomatism by a volatile-rich melt or fluid (Grégoire et al., 2003).
|
|
New high-precision whole-rock trace element concentration data have also been presented recently for a suite of southern African lower crustal granulite xenoliths (Schmitz et al., 2004
; M. D. Schmitz, unpublished data, 2004). These data indicate that each of three mineralogically defined granulite types (mafic, felsic and sedimentary) is also characterized by distinct incompatible element compositions (Fig. 13a). Felsic and sedimentary granulites have roughly similar incompatible element patterns, with strong enrichments in Ba, Th, U and Pb and sharp depletions in Nb. Mafic granulites show similarities to the 39–41°E MORB in that they have enrichments in Ba but depletions in Th and U, and are generally without Nb anomalies. Like the type 2 peridotities, the mafic granulites also have high (Lu/Hf)N/(Sm/Nd)N ratios (mean = 1·6).
We have carried out mixing–melting models to determine how well mixtures of depleted MORB-source mantle with southern African lower crust or metasomatized cratonic peridotite can approximate the incompatible element compositions of SWIR MORB from 39 to 41°E. In these simple models, compositions were calculated for bulk mixtures of depleted mantle (Workman & Hart, 2005) with the mean compositions of each of the two types of metasomatized southern African peridotite (Grégoire et al., 2003; Table 5) or each of the three types of southern African lower crustal granulite (Schmitz et al., 2004; M. D. Schmitz, unpublished data, 2004; Table 5). Non-modal partial melt compositions of these mixed source compositions were then calculated and compared with the incompatible element abundances of each of the three most primitive MORB samples from 39 to 41°E (i.e. MD34-D5, P27-2 and P28-9; Fig. 13b, c and d). The free parameters in these models are the mixing proportions of the various source materials, the degree of partial melting and the spinel/garnet ratio in the source. Fixed parameters are the modal and melt contribution proportions for olivine, orthopyroxene, clinopyroxene and spinel + garnet and the trace element partition coefficients (Table 6). For one sample that displays trace element evidence suggesting contamination by Marion plume material (sample P27-2, with a slight positive Nb anomaly), we also included a Marion plume source composition (see Table 5 legend for explanation) as a possible mixing component in the models. Residuals, representing the difference between the concentration generated by the model and that of the sample's measured concentration, were calculated for each element using the following equation:
 | (2) |
in many peridotite phases, Pb is not included in the residual calculations for the depleted mantle-metasomatized peridotite models, nor (to keep the comparisons on an equal basis) for the depleted mantle–lower crustal models, although it is shown in Fig. 13 for illustrative purposes. Our simple modeling shows that melts of mixtures of depleted mantle with between 7 and 18 wt % metasomatized cratonic peridotite (dominantly Type 2 peridotite), or between 1 and 2 wt % lower continental crust (dominantly mafic granulite) can both closely simulate the incompatible element patterns of 39–41°E SWIR MORB. In one case each, metasomatized cratonic peridotite and continental lower crust provide the best fit to a sample's incompatible element pattern (to samples MD34-D5 and P28-9, respectively), and in the third case, both provide roughly equal fits (to sample P27-2). Although the new trace element data do not allow us to exclude either continental lithospheric mantle or lower crust as likely contributors to the source of 39–41°E MORB, our modeling does indicate that addition of either of these materials to the Indian Ocean upper mantle can account for the incompatible element characteristics of these unusual basalts. This strongly supports a role for continental material (crust and/or mantle) in the genesis of the 39–41°E MORB.
An important additional constraint on the possible role of lithospheric peridotite and continental lower crust in the genesis of the 39–41°E MORB is provided by new Nd–Hf isotope data obtained on Gondwanan cratonic peridotites and lower crustal granulite xenoliths. The Nd–Hf isotope compositions of clinopyroxene and garnet mineral separates of southern African cratonic peridotites (Simon et al., 2002; Bedini et al., 2004; Fig. 14) tend to fall well above the Nd–Hf mantle array with mostly positive and extremely variable Hf (–5 to +470) and mostly negative Nd values. These data indicate that mixtures of metasomatized Archean cratonic mantle with depleted MORB-source mantle could have the strong negative Nd and moderate positive Hf values necessary to account for the 39–41°E MORB samples. However, such mixtures would need to contain a large proportion of the lithospheric mantle component, between roughly 10 and 25%, to explain the Nd isotopic compositions of the 39–41°E samples (assuming Nd concentrations and Nd values of 0·9 and 4·0 ppm and +10 and –12 units, respectively, for depleted mantle and Archean lithospheric mantle; Bedini et al., 2004).
|
Nd–Hf isotope data have also recently been reported for lower crustal granulite xenoliths from eastern Australia (Vervoort et al., 2000
) and southern Africa (Schmitz et al., 2004). The Australian crustal xenolith data have Nd isotopic compositions ranging between Nd = +9·5 and –18 and, with the exception of one sample, plot within 8 Hf units of the Nd–Hf isotopic mantle array line (mean Hf = +2; Vervoort et al., 2000). The southern African lower crustal xenolith data, which are more relevant to the SWIR given their proximity, are much more isotopically heterogeneous (Fig. 14). These granulite xenoliths span large ranges of present-day Nd and Hf isotopic compositions (Nd = –25 to +134, Hf = –26 to +38, mean Hf = –22) falling between 45 Hf units above and 218 Hf units below the mantle array (Schmitz et al., 2004; Fig. 14). Of the three granulite xenolith types defined by Schmitz et al. (2004), only the mafic granulites have isotopic compositions that typically plot above the mantle array (Hf = –3 to +45) at unradiogenic Nd isotopic compositions (Nd = –7 to –12; Fig. 14). Thus, mixtures of depleted mantle with 5–10% of continental lower crust can also explain the Nd and Hf isotopic composition of 39–41°E samples, but only if this material was dominantly composed of mafic granulite (Schmitz et al., 2004).
One piece of evidence that appears to support the involvement of lithospheric mantle over lower continental crust in the genesis of the 39–41°E samples is their isotopic resemblance to the Ejeda–Bekily dikes from southern Madagascar (Mahoney et al., 1991). These alkalic and basanitic dikes are relatively unfractionated and show little to no major or trace element evidence for crustal contamination (Dostal et al., 1992). On the basis of their incompatible element and isotopic compositions (including their isotopic dissimilarity to Marion hotspot lavas), the Ejeda–Bekily dikes have been interpreted by Mahoney et al. (1991) and Storey et al. (1998) to represent melts of Archean or Early Proterozoic Madagascan lithospheric mantle. The dikes and the 39–41°E SWIR lavas have similar Hf isotopic compositions, both falling above the mantle array in Nd–Hf isotope space (Fig. 7), and the dike samples plot closely to the 39–41°E MORB trends in most diagrams of Hf vs isotope and trace element ratios (Figs 9 and 10). The resemblance in isotope and trace element ratios between the Ejeda–Bekily dikes and 39–41°E SWIR MORB strongly suggests that both shared similar mantle sources. Because these silica-undersaturated dikes have major and trace element compositions consistent with being derived from an enriched continental lithospheric mantle source (Storey et al., 1998) their strong isotopic resemblance to 39–41°E MORB would therefore seem to support a major role for continental lithospheric mantle in the sources of these MORB as well.
Escrig et al. (2004) came to a different conclusion and inferred that lower continental crust, rather than continental lithospheric mantle, was the most plausible contaminant responsible for the extreme isotope compositions of the 39–41°E MORB. This inference was based on measurement of a radiogenic 187Os/188Os ratio (0·335) in sample MD34-D5, the most isotopically extreme of the 39–41°E MORB samples. Although such a radiogenic Os isotope composition would indeed argue for a source dominated by a crustal, rather than a mantle component, the amount of crust needed to reach this radiogenic an Os isotopic composition is extremely large (>50%), given possible Re and Os concentrations and Os isotopic composition estimates for lower crust and MORB-source peridotite. Assuming 187Os/188Os values of 0·80 and 0·129 and Os concentrations of 0·1 and 2 ppb for lower continental crust and MORB-source mantle, respectively (Meisel et al., 1996; Saal et al., 1998), it would require a mixture with nearly 90% lower crust to explain the Os isotope composition of sample MD34-D5. It seems unlikely that a source that contained such a large amount of crustal material could produce a basalt with the major and trace element composition of this sample. An alternative possibility is that this sample is contaminated by seawater Os, which is very radiogenic (Williams & Turekian, 2004), and, given the low Os concentration of this and most other MORB (6 ppt Os for MD34-D5; Escrig et al., 2004), is a significant concern with Os isotope measurements in MORB.
A small but non-trivial portion of the continental rifting that dispersed central Gondwana occurred within Archean cratonic lithosphere (in East Antarctica, southern Africa, Madagascar and Southern India; e.g. Groenewald et al., 1991; Tucker et al., 1999; Fig. 15). If it can be assumed that the mantle portion of this lithosphere had chemical and isotopic traits similar to that of South African cratonic peroditite, then the separation and mixing of portions of this lithospheric mantle into the adjacent asthenosphere (probably via thermal erosion by the Marion mantle plume; Mahoney et al., 1992; Storey et al., 1998) appears to offer the simplest explanation for the geochemical signature of 39–41°E MORB and its location adjacent to the trace of the Marion hotspot.
|
Distribution of depleted components beneath the SWIR
Two main isotopic components dominate the Sr, Nd and Pb isotopic compositions of N-MORB from the central SWIR (Mahoney et al., 1992
). One, an Atlantic–Pacific-type N-MORB-source component, characterized by moderately radiogenic 206Pb/204Pb, low 7/4 and 8/4 values and unradiogenic 87Sr/86Sr, is dominant from 17 to 26°E along the western SWIR (Mahoney et al., 1992). It also appears to be present, although variably overprinted by contamination from the Bouvet plume, westward to the Bouvet triple junction (le Roex et al., 1983, 1992). The other component, an Indian Ocean-type N-MORB source, is characterized by relatively unradiogenic 206Pb/204Pb and high 7/4 and 8/4 values and by moderately elevated 87Sr/86Sr relative to most Atlantic and Pacific MORB. This latter component is present east of 32°E, but its most extreme expression is in the 42–47°E region, from the Discovery II to the Indomed fracture zone (Figs 1 and 4).
The data presented here clearly show that the Indian Ocean and Atlantic–Pacific N-MORB signatures in SWIR MORB are also distinct in terms of their Hf isotope compositions (Fig. 4), with samples dominated by the Indian Ocean MORB source tending to lie above the Nd–Hf mantle array at a wide range of Nd values (+7 to +11), and those dominated by the Atlantic–Pacific N-MORB source tending to lie below it at a narrower range of Nd values (+7 to +9). The distinct Hf–Nd isotopic compositions of these two MORB sources provide important clues to their origins.
The Atlantic–Pacific mantle signature
From about 43°E to 10°E on the SWIR (excluding the 39–41°E segment) there is a gradual decrease in Nd and Hf and a gradual increase in 206Pb/204Pb ratios ranging from values typical of Indian Ocean-type N-MORB to those similar to Bouvet OIB (Fig. 4). Sr isotopes reveal that this gradual isotopic transition is not simply the result of binary mixing between an Indian Ocean-type N-MORB source and material from the Bouvet plume, but rather is the result of both of these components mixing with an Atlantic–Pacific-type N-MORB source component with unradiogenic Sr, but intermediate Nd, Hf and Pb isotope ratios (Mahoney et al., 1992). The Hf isotope ratios of SWIR lavas dominated by the Atlantic–Pacific signature are lower than those of most MORB from the northern MAR or EPR (Salters, 1996; Nowell et al., 1998; Chauvel & Blichert-Toft, 2001). However, Andres et al. (2002) found that there is a general southward decrease in Hf from 40° to 55°S along the southern MAR. Further, MORB from the few portions of the MAR south of 49°S uncontaminated by the ‘LOMU’ and Shona geochemical anomalies (Douglass et al., 1999; Le Roux et al., 2002a) have moderate Nd and negative Hf values that are nearly identical to those of Atlantic–Pacific-type MORB from 17° to 26°E on the SWIR (Andres et al., 2002; Fig. 7). This Nd–Hf isotopic similarity strongly suggests that the western SWIR N-MORB are derived from MORB-source mantle similar to that beneath the southernmost MAR and this, along with the gradual eastward increase in Hf values from the western to the eastern SWIR, confirms the finding of Mahoney et al. (1992) that the western boundary of the Indian Ocean mantle domain is largely gradational in nature and is located east of 26°E on the SWIR. On the basis of Hf–Nd and Hf–87Sr/86Sr isotopic and Lu/Hf ratio variations, we propose that the boundary is more precisely located between 26°E and 32°E.
The Indian Ocean mantle signature
In most respects, MORB from the 42–47°E section of the SWIR lie at the opposite end of the isotopic spectrum from the Atlantic–Pacific-type MORB source expressed at 17–26°E (Fig. 4). But normal and transitional MORB from this region, which includes the Discovery II and Indomed fracture zones and the spreading segments between them, are unusually heterogeneous, particularly in Sr and Nd isotope ratios (e.g. 87Sr/86Sr = 0·7023–0·7033, Nd = +7 to +11·2; Mahoney et al., 1992; Figs 5 and 6). Samples from this region encompass the most radiogenic Nd and Hf isotope compositions reported thus far for the Indian Ocean outside the AAD (Mahoney et al., 1992; Chauvel & Blichert-Toft, 2001), as well as among the lowest 206Pb/204Pb values reported for the SWIR outside the 39–41°E segment (Price et al., 1986; Mahoney et al., 1992). However, Nd–Hf–Sr isotope systematics reveal that there are at least two depleted end-members sampled by MORB at 42–47°E, indicated by the isotopic differences between the samples with the most radiogenic Nd and Hf isotope ratios (e.g. P29-34 with Nd = +11·2, Hf = +17, 87Sr/86Sr = 0·7026 vs MD34-D4 with Nd = +8·4, Hf = +21 and 87Sr/86Sr = 0·7030; Mahoney et al., 1992; Chauvel & Blichert-Toft, 2001; Fig. 7).
The range of isotopic compositions present in MORB from the 42–47°E area provides important clues to mixing processes occurring beneath the east–central SWIR and has implications for the way in which the Indian Ocean mantle signature is produced. The seven samples from six dredge hauls with reported Sr, Nd, Pb and Hf isotope data from this region fall into three distinct isotopic categories. Samples P29-34 and P30-75 have isotope compositions that fall within the fields for Atlantic and Pacific MORB in Sr–Nd–Hf–Pb and Pb–Pb isotope space, although near the low 206Pb/204Pb–Nd and high 208Pb/204Pb edges of these fields (Table 3, Figs 5 and 6). These samples also have Hf values between +0·5 and –1. All of these traits suggest an isotopic affinity with depleted Atlantic–Pacific-type MORB sources. Samples P37-1 and P37-2 display anomalous isotope compositions indicative of long-term large ion lithophile and LREE enrichment and intermediate U/Pb ratios (e.g. 87Sr/86Sr = 0·7033–0·7038, Nd = +2·9 to +6·8, 206Pb/204Pb = 18·04–18·35; Mahoney et al., 1992). These samples fall above the Nd–Hf mantle array line, but only slightly (Hf +1·7). Finally, samples P41-49, P42-11 and MD34-D4 have Sr, Nd and (except for MD34-D4) 206Pb/204Pb isotope ratios intermediate between the depleted and enriched groups (87Sr/86Sr = 0·7029–0·7032, Nd = +7·0 to +8·4, 206Pb/204Pb = 17·8–18·2; Hamelin & Allègre, 1985), but have Hf values higher than either group (+2·3 to +6·7). These latter samples also have 7/4 and 8/4 values (of 5·1–5·4 and 37–65, respectively) significantly higher than those of the depleted P29-34 and P30-75 samples. It is this last group, characterized by Nd and Sr isotope ratios intermediate between Atlantic–Pacific MORB and OIB, low to moderate 206Pb/204Pb with elevated 7/4 and 8/4 and moderate to high positive Hf values, that is most characteristic of Indian Ocean MORB as a whole (e.g. Michard et al., 1986; Mahoney et al., 1989; Salters, 1996; Chauvel & Blichert-Toft, 2001; Hanan et al., 2004).
The fact that depleted MORB similar to that from the Atlantic and Pacific Oceans exists in a region of the SWIR otherwise characterized by Indian Ocean-type isotope compositions (Figs 4 –8) strongly supports the idea that the Indian Ocean MORB source is the product of an ‘Indian Ocean component’ (or a range of components) that has variably contaminated an upper mantle originally having an isotopic character similar to that beneath Atlantic and Pacific ridges (Storey et al., 1989; Mahoney et al., 1989, 1992). The preservation of relatively uncontaminated Atlantic–Pacific-type MORB source mantle in this region may be attributable to inefficient mixing beneath the east–central SWIR as a result of the cold upper mantle temperatures inferred for this region from seismic tomography (Debayle & Leveque, 1997). The good correlations between Hf and 87Sr/86Sr, 7/4, 8/4, Ba/Nb and Ba/Th displayed by SWIR MORB from east of 32°E (excluding 39–41°E), as well as most other Indian Ocean MORB (Figs 9 and 10), are clear indications that high Hf values are as intrinsic to this component as elevated 87Sr/86Sr, high 7/4 and 8/4 and enrichment in large ion lithophile elements. The lack of good correlations between Hf and 206Pb/204Pb (Fig. 9b) and between Nd and Hf isotope ratios (Fig. 7) in Indian Ocean MORB suggests either that the Indian Ocean component spans a wide range of Nd and Pb isotope compositions or, perhaps more likely, that the Atlantic–Pacific-type MORB source was already heterogeneous in these isotope ratios before being contaminated.
Hf isotope constraints on the origin of the Indian Ocean MORB isotopic signature
Several possible scenarios have been proposed for the origin of a component responsible for the isotopic differences between Indian Ocean and Atlantic–Pacific MORB. These include: (1) contamination of the Indian Ocean upper mantle by plume material derived either from the Kerguelen plume alone (Storey et al., 1989; Weis & Frey, 1996), or from a combination of central Indian Ocean plumes (e.g. Marion, Crozet and Kerguelen) mixed with delaminated continental lithospheric mantle (Mahoney et al., 1992); (2) focused recycling of pelagic sediments and subducted oceanic crust into the Indian Ocean upper mantle (Dupré & Allègre, 1983; Rehkämper & Hofmann, 1997); (3) foundering and mixing of continental lower crust into the Indian Ocean upper mantle during or after the opening of the Indian Ocean (e.g. Escrig et al., 2004; Hanan et al., 2004); (4) mobilization or delamination of subduction-modified mantle wedge material formed during Paleozoic subduction beneath eastern Gondwana (Kempton et al., 2002).
As pointed out by other workers (Mahoney et al., 1992; Hanan et al., 2004), Kerguelen and all other known Indian Ocean plumes have 206Pb/204Pb values that are too radiogenic (18 to 19; Salters & White, 1998; Yang et al., 1998) to explain the bulk of the Indian Ocean MORB data by mixing with an Atlantic–Pacific-type MORB source. A mixture of plume material with continental lithospheric mantle, on the other hand, might be able to produce an Indian Ocean component with an appropriate isotopic composition to account for Indian Ocean MORB (Mahoney et al., 1992). However, this seems unlikely for two reasons. First, mixing relations require that for an Indian Ocean component to have a plausible 206Pb/204Pb value of about 17·25 or less, either unrealistically large proportions of lithospheric mantle would be required (>50%) or this lithospheric mantle material must have had an extremely low mean 206Pb/204Pb value of 16·5 or less [assuming Pb abundances of 0·08 and 0·16 ppm for plume material (from Table 5) and lithospheric mantle (McDonough, 1990), respectively, and a 206Pb/204Pb ratio of 18·5 for the plume material]. Such unradiogenic Pb isotope compositions have been reported for Archean cratonic peridotites (e.g. Cohen et al., 1984) but not for peridotites of Proterozoic age (see Lee et al., 1996; Carlson et al., 2004). However, the great majority of the continental rifting that resulted in the breakup of central and eastern Gondwana to form the Indian Ocean involved Proterozoic lithosphere rather than Archean cratons (Meert, 2003; Fig. 15). Thus, unless there was major delamination of the Archean keels beneath Gondwanan cratons in response to Mesozoic plume-head impact, which is not supported by seismic and heat flow studies (e.g. Gupta, 1991; James et al., 2001), there was little opportunity for widespread mixing of plume material and/or Indian Ocean asthenosphere with low-206Pb/204Pb Archean mantle.
Second, the arrays defined by the 39–41°E MORB samples on plots of Hf versus other isotope ratios, which we attribute mainly to mixing between Indian Ocean asthenosphere and cratonic lithospheric mantle, have slopes that are highly oblique to the trends defined by the Indian Ocean-type SWIR MORB outside 39–41°E and other Indian Ocean MORB (Fig. 16). The 39–41°E samples appear to be pointing toward a component with markedly lower 206Pb/204Pb and higher 87Sr/86Sr, 7/4 and 8/4 for a given Hf value than the main trend of Indian Ocean MORB data. The extreme Hf isotope heterogeneity of mineral separates from Archean cratonic peridotite (Fig. 14) certainly permits this material to have a wide range of Hf values. However, the isotopic similarity between the Ejeda–Bekily dikes from southern Madagascar and 39–41°E MORB appears to support our contention that the high 87Sr/86Sr, low 206Pb/204Pb component tapped by these MORB represents Gondwanan Archean lithospheric mantle probably disrupted by the Marion plume in the late Cretaceous. If this inference is correct, it suggests that no mixture of this ancient, low 206Pb/204Pb lithospheric mantle with the Kerguelen, Marion or other Indian Ocean plumes would have sufficiently high Hf values to explain the highest-Hf SWIR MORB, such as MD34-D4 (Fig. 16). Moreover, because of the relatively high 87Sr/86Sr ratios of Gondwanan Archean lithospheric mantle (e.g. Menzies & Murthy, 1980; Walker et al., 1989) and those Indian Ocean plumes with positive Hf values (e.g. Yang et al., 1998; Doucet et al., 2004), it seems unlikely that a mixture of these materials would have the mildly radiogenic 87Sr/86Sr (<0·7036) required for the Indian Ocean component by the 87Sr/86Sr–Hf array (Fig. 16), even if the Hf value of the cratonic lithosphere involved was significantly higher than indicated by the 39–41°E MORB and Ejeda–Bekily dikes.
|
Addition of ancient pelagic sediments to the Indian Ocean upper mantle seems to offer a more plausible explanation for the differences between Atlantic–Pacific and Indian Ocean MORB than plume contamination, at least from a geochemical perspective. Although we cannot address why pelagic sediments should be preferentially recycled into the upper mantle beneath the Indian, rather than the Atlantic or Pacific Oceans, the estimated Sr, Nd and Pb isotope ratios of ancient (1–3 Ga) subducted pelagic sediments agree well with those required for the Indian Ocean mantle component, particularly if the sediments are mixed with variable quantities of subducted oceanic crust (Rehkämper & Hofmann, 1997
). The moderately decoupled Nd and Hf isotopic ratios in pelagic sediments (Fig. 14), and their Nd/Hf ratios, which are up to 10 times higher than mantle values (e.g. Ben Othman et al., 1989; Vervoort et al., 1999), also seem to offer an explanation for the high Hf values observed in Indian Ocean MORB.
However, as pointed out by Kempton et al. (2002), the Nd and Hf isotope compositions of pelagic sediments are too unradiogenic, and the decoupling between Hf and Nd isotope ratios and concentrations is too limited, to explain the Nd–Hf isotopic compositions of at least one-third of analyzed Indian Ocean MORB samples, those having both Nd > +7 and Hf > +2 (Fig. 7). Even for the optimal case in which pelagic sediment is added without any accompanying recycled oceanic crust, mixtures of pelagic sediments and Atlantic–Pacific-type MORB-source mantle will lie two or more Hf units above the Nd–Hf array only at Nd values less than about +5. Thus, the Indian Ocean mantle component must consist of a material with an even greater decoupling of its Nd and Hf concentrations and isotope ratios and/or a more radiogenic Hf isotope composition.
Hanan et al. (2004) have proposed that just such a material formed in the lower continental crust adjacent to continental rift zones. Recent analysis of lavas from zone B4 of the AAD (125–126°E) showed that these Indian Ocean-type MORB (according to their low 206Pb/204Pb and high 208Pb/204Pb ratios) have extremely radiogenic Nd and Hf isotopic compositions (Nd = +9·8 to +12·9, Hf = +22·2 to +28·8; Hanan et al., 2004), giving them the most radiogenic Hf isotope ratios yet published for any MORB, as well as the highest Hf values in the Indian Ocean (up to +8·5). To explain the long-term high Lu/Hf ratios required by these radiogenic Hf isotopic compositions, Hanan et al. (2004) proposed that their source is the product of mixing Pacific-type upper mantle with a restite of garnet-bearing granulitic lower continental crust that underwent extensive melting during continental rifting. This melting would have had the effect of increasing Sm/Nd and dramatically raising Lu/Hf, as a result of the compatibility of Lu in garnet. Extremely high (Lu/Hf)N ratios, up to 38, have been measured in lower crustal granulite xenoliths from other continental rift environments (Scherer et al., 1997). Over the 80 Myr since the separation of Australia from Antarctica (Storey, 1995), this lower crustal restite could develop high Nd and extremely high Hf and Hf values suitable as a mixing end-member for zone B4 AAD MORB and high Hf Indian Ocean MORB elsewhere, whereas less melt-depleted lower crustal granulite could constitute a low Nd–Hf end-member to explain more enriched Indian Ocean MORB (Hanan et al., 2004). The low U/Pb ratios of lower continental crust would also fit with the unradiogenic 206Pb/204Pb ratios measured in Indian Ocean MORB (Rudnick & Fountain, 1995; Hanan et al., 2004).
Although the lower crustal contamination model of Hanan et al. (2004) offers a potential explanation for the Nd, Hf and Pb isotope differences between Indian Ocean-type and Atlantic–Pacific-type MORB, this hypothesis has one major problem. The 87Sr/86Sr values of Gondwanan lower crustal granulites tend to be rather high, even when corrected for radiogenic growth since the time of continental rifting (e.g. mean 87Sr/86Sr80 Ma = 0·7072 for Australian granulite xenoliths, mean 87Sr/86Sr150 Ma = 0·7068 for all southern African granulite xenoliths and 0·7060 for southern African mafic granulite xenoliths only; Rudnick et al. 1986; Rudnick, 1990; Huang et al., 1995; M. D. Schmitz, unpublished data, 2004). Because melt depletion cannot lower Sr isotope ratios, the low 87Sr/86Sr ratio of the Indian Ocean component inferred from the main 87Sr/86Sr–Hf correlation (Fig. 16) seriously limits the degree to which lower crustal granulite could be incorporated into sources of Indian Ocean MORB. Figure 17 shows that for mixtures of depleted mantle with mean Australian granulite or mean southern African mafic granulite, addition of <1% of the granulite end-member is required to explain the vast majority of Indian Ocean MORB with 87Sr/86Sr ratios of <0·7035. Such small proportions of lower crustal granulite are not nearly sufficient to explain the Nd and Hf isotopic ranges of Indian Ocean MORB (Fig. 17a), regardless of whether one uses a mean Australian or southern African granulite composition or the hypothetical high-Hf and -Nd ‘lower crustal granulite’ (LCG) end-member of Hanan et al. (2004). It is true that 87Sr/86Sr ratios in lower crustal granulite can extend to values as low as 0·7024 in Australian and 0·7020 in southern African granulites (Rudnick et al., 1986; M. D. Schmitz, unpublished data, 2004); however, these values are extremely unrepresentative. Such low 87Sr/86Sr ratios could potentially explain the unradiogenic Sr isotope compositions in a localized area, such as zone B4 of the AAD, but it is highly implausible that low 87Sr/86Sr ratios would be a general characteristic of a lower crustal granulite contaminant present throughout the Indian Ocean mantle domain.
|
Another environment where extreme decoupling of Lu/Hf and Sm/Nd ratios is expected to occur is in the mantle wedge of a subduction zone. During subduction, the mantle wedge is subjected to pervasive infiltration by hydrous, slab-derived fluids, and possibly sediment melts, and experiences partial melting due to the solidus-lowering effect of these fluids. Experimental evidence suggests that aqueous fluids driven off a subducting slab will have elevated Nd/Hf ratios as a result of the relative immobility of Hf (e.g. Ayers et al., 1997
). Recent Hf isotope data from ‘fluid-dominated’ volcanic arcs do suggest that slab-derived fluids can transport significant quantities of Hf into the sub-arc mantle (Woodhead et al., 2001), but data for the Mariana arc system, which has been most studied for Hf isotope systematics (Pearce et al., 1999; Woodhead et al., 2001), suggest that the Nd/Hf ratio of the fluid component added to the sub-arc mantle may be up to three times higher than that of the local subducting sediments. Thus, because of the high Nd/Hf ratios and relatively low Nd and Hf values of fluids derived from subducting sediments and oceanic crust, the likely net effect of fluid addition with respect to the Nd and Hf isotopes systems is a significant decrease in 143Nd/144Nd and Sm/Nd ratio, a mild decrease in 176Hf/177Hf and a significant increase in Lu/Hf ratio, relative to mantle melting residues that had not experienced fluid addition. Fluid-mobile elements such as Sr and especially Pb and Rb would also be added, resulting in decreased U/Pb and increased Rb/Sr ratios in the mantle wedge (e.g. Aizawa et al., 1999; Green & Adam, 2003).
Over hundreds of millions of years, the elemental and isotopic fractionations produced in the mantle wedge would result in it having radiogenic 176Hf/177Hf and 87Sr/86Sr and relatively unradiogenic 143Nd/144Nd and 206Pb/204Pb ratios, qualitatively similar to the contaminant inferred to be responsible for the Indian Ocean MORB source (Kempton & Pearce, 2003; Zhang et al., 2005). Indeed, positive Hf values are characteristic of suprasubduction zone magmas, particularly from long-lived, fluid-dominated arc systems (e.g. the Mariana and Tonga–Kermadec arcs; Pearce et al., 1999; Kempton et al., 2000). Kempton et al. (2002) advocated addition of such fluid-modified, melt-depleted mantle wedge material to the Indian Ocean upper mantle to explain the Nd–Hf isotopic differences between Indian Ocean- and Atlantic–Pacific-type MORB, both for the AAD (their study area) and for the greater Indian Ocean mantle domain. Their hypothesis is consistent with recent evidence (Gurnis et al., 1998) that the AAD overlies a stranded, westward-dipping slab from a subduction zone that was active on the Pacific margin of Gondwana from the Late Paleozoic to the Early Cretaceous.
Although there is no evidence for ancient subduction zones beneath spreading centers elsewhere in the Indian Ocean, it is important to note that the continental rifting that separated Africa, Antarctica, Indo-Madagascar and Australia to form the Indian Ocean occurred mostly within Meso- and Neoproterozoic (550–1500 Ma) orogenic belts (e.g. Groenewald et al., 1991; Meert, 2003; Fig. 15). These belts consist of high-grade metamorphic and igneous terranes amalgamated during the assembly of Gondwana and its precursor, Rodinia (e.g. Jacobs et al., 1998; Dalziel et al., 2000). Many of these terranes contain igneous and meta-igneous rocks with clear supra-subduction zone geochemical signatures (e.g. Cornell et al., 1996; Kröner et al., 1997, 2001; Paulsson & Austrheim, 2003), which appear to have formed in intra-oceanic arc and Andean-type continental margin environments (e.g. Stern, 1994; Handke et al., 1999). It is likely that slab-derived fluids and/or sediment melts were introduced to significant portions of the Proterozoic lithospheric mantle beneath these orogenic belts during the periods of subduction with which these magmatic rocks are associated.
Continental rifting during the Early Cretaceous may have resulted in the destabilization and delamination of some of this subduction-modified Meso- and Neoproterozoic lithospheric mantle, which could have foundered in and mixed with the upper mantle beneath the nascent Indian Ocean. If the subduction modification took place a sufficiently long time ago, addition of such mantle wedge material could explain the high Hf values and strong Nd–Hf isotopic decoupling of samples such as MD34-D4 from the 42–47°E section of the SWIR (Chauvel & Blichert-Toft, 2001), and the MORB from zone B4 of the AAD (Hanan et al., 2004) and could also explain the mildly elevated 87Sr/86Sr ratios of these samples.
We present a set of models (Fig. 18) for the development of a range of subduction-modified mantle (SMM) compositions, formed at times from 250 to 1500 Ma, in Nd–Hf and Sr–Hf isotope space. In these models, various amounts of a subduction component (either mean global subducting sediments or a slab-derived fluid) are added to an age-corrected Atlantic–Pacific-type depleted mantle composition (Table 5). The resulting mixture is then immediately subjected to partial melt extraction and the present-day isotope composition of the melt residue is calculated. The assumptions of these models are that: (1) the REE and Hf contents of the fluid are entirely derived from subducting sediments (Table 5; these elements appear to be essentially immobile in igneous oceanic crust during dehydration; Green & Adam, 2003); (2) the mobility of Hf from subducting sediments into a fluid phase is 50% lower than that of Lu, similar to the mobilities of Ti and Nb determined in sediment dehydration experiments (Aizawa et al., 1999); (3) the Sr isotopic composition of the fluid can be estimated by assuming that 30% of the Sr in the fluid is derived from igneous oceanic crust (with a depleted mantle isotopic composition) and 70% is derived from mean pelagic sediments (this is an approximation based on the lower 87Sr/86Sr ratios in lavas from ‘fluid-dominated’ relative to ‘sediment-dominated’ arcs; e.g. Davidson, 1986; Lin et al., 1990); (4) following sediment or fluid addition, the resulting mixture is subjected to removal of a 6% non-modal batch partial melt (in the spinel stability field) to simulate extraction of arc magmas (see Table 6 for partition coefficients and melting assemblage). This degree of melting is at the lower end of the range estimated for most arc magmas (e.g. Baker et al., 1994), but was chosen to approximate the mean degree of melting experienced by the entire mantle wedge. Increasing the degree of melting by as much as a factor of two will not significantly affect the model results. The details of the model calculations are presented in the caption to Fig. 18 and in Table 6.
|
, Hamelin & Allègre, 1985
|
The subduction-modified mantle (SMM) compositions produced by addition of bulk subducting sediments yield
Nd and Hf values that lie relatively close to the mantle array (Hf less than +3 between Nd = +5 and Nd = +15). This is because the effects of sediment addition and partial melting on SMM act in opposite directions and maintain good coupling between Sm/Nd and Lu/Hf ratios. The combination of slab fluid addition and partial melting, on the other hand, sharply lowers Sm/Nd relative to Lu/Hf. This results in the fluid-modified SMM trajectories lying above the Nd–Hf mantle array and having much shallower slopes. About 25% of the Indian Ocean MORB samples in Fig. 18a have Hf values that are too high to be explained by involvement of sediment-dominated SMM. In contrast, all high-Hf Indian Ocean MORB can be explained by mixing between Atlantic–Pacific-type MORB source mantle and fluid-modified SMM, although the samples with the highest Hf values require mixing with SMM components that formed at least 1000–1500 Myr ago.
The SMM trajectories in Sr–Hf isotope space do not provide equally strong constraints on the age or nature of a possible SMM component involved in the genesis of Indian Ocean MORB. Both sediment- and fluid-modified SMM compositions can explain the mildly elevated 87Sr/86Sr and radiogenic Hf values of the highest Hf Indian MORB. However, the fluid-modified SMM compositions in both Nd–Hf and Sr–Hf isotope space give more consistent results when attempting to reproduce the compositions of particular samples by mixing SMM with depleted mantle (Fig. 18). For example, in both Nd–Hf and Sr–Hf isotope ratio space, the high-Hf SWIR MORB sample MD34-D4 can be approximated by mixing depleted mantle with 30% of a 1500 Ma fluid-modified SMM component and, similarly, AAD zone B4 MORB sample MW8801 23-1 (one of only two with Sr isotope data; Pyle et al., 1992) can also be reproduced by mixing depleted mantle with 40% of a 1500 Ma SMM component, but one having experienced somewhat less fluid addition.
Mixing between an Atlantic–Pacific-type MORB source and ancient SMM variably modified by slab-derived fluids can simply explain why Indian Ocean MORB is characterized by high Hf values and a diverse range of Nd isotope compositions. That this hypothesis also explains why Indian MORB tend to have mildly, but not highly, elevated Sr isotopic compositions is also a point in its favor. Determining whether SMM is the most appropriate material to explain the Pb isotope compositions of Indian Ocean MORB will require better constraints on the mean relative contributions of U, Th and Pb from oceanic crust and sediments to slab-derived fluids, as well as more precisely determined partition coefficients for these elements during melting of hydrous peridotite, than are currently available in the published literature.
The fact that MORB from 43°E on the SWIR (sample MD34-D4; Chauvel & Blichert-Toft, 2001) and from zone B4 of the AAD (Hanan et al., 2004) lie on or above the curve of 500 Ma fluid-modified SMM compositions indicates that the Nd–Hf isotopic characteristics of all high-Hf Indian Ocean MORB cannot be explained by westward advection of mantle wedge material from the Pacific margin of Gondwana (see Kempton et al., 2002). Subduction at this margin appears to have begun only in the Early Paleozoic (500 Ma; Grunow et al., 1996), which would not allow sufficient time to develop the necessary decoupling of Nd and Hf isotopic compositions. It seems likely that, at least in some locations, this isotopic signature was acquired via destabilization and foundering of lithospheric mantle beneath the Proterozoic orogenic belts that border the Indian Ocean. We cannot, however, rule out a role for advected Paleozoic-age SMM in the genesis of Indian Ocean lavas with only mildly positive Hf values.
If the hypothesis is correct that the distinct isotope characteristics of Indian Ocean MORB are derived from contamination of the Indian Ocean upper mantle by delaminated or detached mantle lithosphere underlying Proterozoic orogenic belts, an important question remains: why does MORB from the Atlantic Ocean bear a closer isotopic resemblance to MORB of the Pacific rather than that of the Indian Ocean? Like the Indian Ocean, the Atlantic also formed as the result of Mesozoic continental breakup. A potential clue comes from the fact that geochemical signatures of contamination by continental and/or subduction components do exist in portions of the Mid-Atlantic Ridge, but these tend to be fairly localized (i.e. less than 500 km in extent along ridge; Shirey et al., 1987; Douglass et al., 1999; Kamenetsky et al., 2001; Le Roux et al., 2002a). Similar geochemical anomalies are absent along the East Pacific Rise (e.g. White et al., 1987; Mahoney et al., 1994). We speculate that the isotopic differences between MORB from the Indian and Atlantic Oceans is due two main factors. First, continental breakup leading to the formation of the Indian Ocean was much more complex, and involved a larger number of continents (and micro-continents) than that leading to the formation of the Atlantic Ocean (Fig. 15). Hence, there was a much longer zone of contact between rifted continental lithosphere and asthenosphere where delamination of continental mantle lithosphere could occur. Second, the Indian Ocean is bordered mainly by Proterozoic orogenic belts, whereas the central and South Atlantic is bordered by a much higher proportion of Archean cratons (e.g. Bernasconi, 1983; Black & Liegeois, 1993; Hanson, 2003; Fig. 15), which are probably less dense and more rigid than orogenic belts (e.g. Lenardic et al., 2003), and thus may be less susceptible to delamination.
|
SUPPLEMENTARY DATA |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLES AND ANALYTICAL METHODSRESULTSDISCUSSION AND CONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Supplementary data for this paper are available at Journal of Petrology online.
|
ACKNOWLEDGEMENTS |
|---|
We are grateful to Andreas Späth for technical assistance with the ICPMS system at UCT and for his hospitality to the first author on two trips to Cape Town to do the trace element work. Mary Horan and Tim Mock generously provided critical laboratory and mass spectrometry assistance at DTM. Mark Schmitz generously provided unpublished trace element and isotope data for his lower crustal xenolith suite. Robert L. Fisher shared his knowledge of the geology of the western Indian Ocean and provided the Marion Island and Prince Edward Island samples. John Mahoney is thanked for supplying several key samples from the SWIR and Madagascar, and for useful comments on an early draft of this paper. We also express sincere thanks to Warren Smith and the Scripps Institution of Oceanography Geological Collections for assistance in obtaining several MORB samples used in this study. This paper benefited from the detailed and insightful reviews of Catherine Chauvel, Dominique Weis and Tiffany Barry, as well as the Editor. This work was supported by US NSF grants OCE 9907173 to P.E.J. and R.W.C., OCE 0221368 to P.E.J. and by funding from the South African National Research Foundation, to A.L.R.
* Corresponding author. Present address: Department of Geology, The Field Museum of Natural History, 1400 S. Lake Shore Drive, Chicago, IL 60605, USA. Telephone: (312) 665-7099. Fax: (312) 665-7641. E-mail: pjanney{at}fieldmuseum.org
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLES AND ANALYTICAL METHODSRESULTSDISCUSSION AND CONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Aizawa, Y., Tatsumi, Y. & Yamada, H. (1999). Element transport by dehydration of subducted sediment: implications for arc and ocean island magmatism. The Island Arc 8, 39–46.
Andres, M., Blichert-Toft, J. & Schilling, J.-G. (2002). Hafnium isotopes in basalts from the southern Mid-Atlantic Ridge from 40°S to 55°S: Discovery and Shona plume–ridge interactions and the role of recycled sediments. Geochemistry, Geophysics, Geosystems 3, doi: 10.1029/2002GC000324.
Asimow, P. D., Hirschmann, M. M. & Stolper, E. M. (1997). An analysis of variations in isentropic melt productivity. Philosophical Transactions of the Royal Society, Series A 355, 255–281.[CrossRef]
Ayers, J. C., Dittmer, S. K. & Layne, G. D. (1997). Partitioning of elements between peridotite and H2O at 2·0–3·0 GPa and 900–1100°C, and application to models of subduction zone processes. Earth and Planetary Science Letters 150, 381–398.[CrossRef][Web of Science]
Baker, M. B., Grove, T. L. & Price, R. (1994). Primitive basalts and andesites from the Mt. Shasta region, N. California: products of varying melt fraction and water content. Contributions to Mineralogy and Petrology 118, 111–129.[CrossRef][Web of Science]
Beattie, P. (1994). Systematics and energetics of trace element partitioning between olivine and silicate melts: implications for the nature of mineral–melt partitioning. Chemical Geology 117, 57–71.[CrossRef][Web of Science]
Bedini, R. M., Blichert-Toft, J., Boyet, M. & Albarède, F. (2004). Isotopic constraints on the cooling of the continental lithosphere. Earth and Planetary Science Letters 223, 99–111.[CrossRef][Web of Science]
Ben Othman, D., White, W. M. & Patchett, J. P. (1989). The geochemistry of marine sediments, island arc magma genesis and crust–mantle recycling. Earth and Planetary Science Letters 94, 1–21.[CrossRef][Web of Science]
Bernasconi, A. (1983). The Archean terranes of central eastern Brazil—a review. Precambrian Research 23, 107–131.[CrossRef][Web of Science]
Black, R. & Liegeois, J.-P. (1993). Cratons, mobile belts, alkaline rocks and continental lithospheric mantle: the Pan-Africa testimony. Journal of the Geological Society, London 150, 89–98.
Blichert-Toft, J. & Albarède, F. (1997). The Lu–Hf isotope geochemistry of chondrites and the evolution of the mantle–crust system. Earth and Planetary Science Letters 148, 243–258.[CrossRef][Web of Science]
Blichert-Toft, J., Chauvel, C. & Albarède, F. (1997). Separation of Hf and Lu for high-precision isotope analysis of rock samples by magnetic sector-multiple collector ICP-MS. Contributions to Mineralogy and Petrology 127, 248–260.[CrossRef][Web of Science]
Blichert-Toft, J., Frey, F. A. & Albarède, F. (1999). Hf isotope evidence for pelagic sediments in the source of Hawaiian basalts. Science 285, 879–882.[CrossRef][Web of Science][Medline]
Blundy, J. D., Robinson, J. A. C. & Wood, B. J. (1998). Heavy REE are compatible on the spinel lherzolite solidus. Earth and Planetary Science Letters 160, 493–504.[CrossRef][Web of Science]
Bodinier, J. L., Guiraud, M., Fabries, J., Dostal, J. & Dupuy, C. (1987). Petrogenesis of layered pyroxenites from the Lherz, Freychinede and Prades ultramafic bodies (Ariege, French Pyrenees). Geochimica et Cosmochimica Acta 51, 279–290.[CrossRef][Web of Science]
Bourdon, B., Zindler, A., Elliott, T. & Langmuir, C. H. (1996). Constraints on mantle melting at mid-ocean ridges from global 238U–230Th disequilibrium data. Nature 384, 231–235.[CrossRef]
Bown, J. W. & White, R. S. (1994). Variation with spreading rate of oceanic crustal thickness and geochemistry. Earth and Planetary Science Letters 121, 435–449.[CrossRef][Web of Science]
Cannat, M., Rommevaux-Jestin, C., Sauter, D., Deplus, C. & Mendel, V. (1999). Formation of the axial relief at the very slow spreading Southwest Indian Ridge (49–69°E). Journal of Geophysical Research 104, 22825–22843.[CrossRef]
Carlson, R. W., Irving, A. J., Schulze, D. J. & Hearn, B. C. (2004). Timing of Precambrian melt depletion and Phanerozoic refertilization events in the lithospheric mantle of the Wyoming Craton and adjacent Central Plains Orogen. Lithos 77, 453–472.[CrossRef][Web of Science]
Chauvel, C. & Blichert-Toft, J. (2001). A hafnium isotope and trace element perspective on melting of the depleted mantle. Earth and Planetary Science Letters 190, 137–151.[CrossRef][Web of Science]
Cohen, R. S. & O'Nions, R. K. (1982). Identification of recycled continental material in the mantle from Sr, Nd and Pb isotope investigations. Earth and Planetary Science Letters 61, 73–84.[CrossRef][Web of Science]
Cohen, R. S., O'Nions, R. K. & Dawson, J. B. (1984). Isotope geochemistry of xenoliths from East Africa: implications for development of mantle reservoirs and their interaction. Earth and Planetary Science Letters 68, 209–220.[CrossRef][Web of Science]
Cornell, D. H., Thomas, R. J., Bowring, S. A., Armstrong, R. A. & Grantham, G. H. (1996). Protolith interpretation in metamorphic terranes: a back-arc environment with Besshi-type base metal potential for the Quha Formation, Natal Province, South Africa. Precambrian Research 77, 243–271.[CrossRef][Web of Science]
Dalziel, I. W. D., Mosher, S. & Gahagan, L. M. (2000). Laurentia–Kalahari collision and the assembly of Rodinia. Journal of Geology 108, 499–513.[CrossRef][Web of Science]
Davidson, J. P. (1986). Isotopic and trace element constraints on the petrogenesis of subduction-related lavas from Martinique, Lesser Antilles. Journal of Geophysical Research 91, 5343–5962.
Debayle, E. & Leveque, J. J. (1997). Upper mantle heterogeneities in the Indian Ocean from waveform inversion. Geophysical Research Letters 24, 245–248.[CrossRef][Web of Science]
Dick, H. J. B., Lin, J. & Schouten, H. (2003). An ultraslow-spreading class of ocean ridge. Nature 426, 405–412.[CrossRef]
Dosso, L., Bougault, H., Beuzart, P., Calvez, J.-Y. & Joron, J.-L. (1988). The geochemical structure of the South-East Indian Ridge. Earth and Planetary Science Letters 88, 47–59.[CrossRef][Web of Science]
Dostal, J., Dupuy, C., Nicollet, C. & Cantagrel, J. M. (1992). Geochemistry and petrogenesis of Upper Cretaceous basaltic rocks from southern Malagasy. Chemical Geology 97, 199–218.[CrossRef][Web of Science]
Doucet, S., Weis, D., Scoates, J., Debaille, V. & Giret, A. (2004). Geochemical and Hf–Pb–Sr–Nd isotopic constraints on the origin of the Amsterdam–St. Paul (Indian Ocean) hotspot basalts. Earth and Planetary Science Letters 218, 179–195.[CrossRef][Web of Science]
Douglass, J., Schilling, J. G. & Fontignie, D. (1999). Plume–ridge interactions of the Discovery and Shona mantle plumes with the southern Mid-Atlantic Ridge (40–55°S). Journal of Geophysical Research 104, 2941–2962.[CrossRef]
Dupré, B. & Allègre, C. J. (1983). Pb–Sr isotope variations in Indian Ocean basalts and mixing phenomena. Nature 303, 142–146.[CrossRef]
Eggins, S. M., Woodhead, J. D., Kinsley, L. P. J., Mortimer, G. E., Sylvester, P. J., McCulloch, M. T., Hergt, J. M. & Handler, M. R. (1997). A simple method for the precise determination of forty trace elements in geological samples by ICPMS using enriched isotope internal standardisation. Chemical Geology 134, 311–326.[CrossRef][Web of Science]
Elthon, D., Ross, D. K. & Meen, J. K. (1995). Compositional variations of basaltic glasses from the Mid-Cayman Rise spreading center. Journal of Geophysical Research 100, 12497–12512.[CrossRef]
Escrig, S., Capmas, F., Dupré, B. & Allègre, C. J. (2004). Osmium isotopic constraints on the nature of the DUPAL anomaly from Indian mid-ocean-ridge basalts. Nature 431, 59–63.[CrossRef]
Evensen, N. M., Hamilton, P. J. & O'Nions, R. K. (1978). Rare-earth abundances in chondritic meteorites. Geochimica et Cosmochimica Acta 42, 1199–1212.[CrossRef][Web of Science]
Fisher, R. L. & Goodwillie, A. M. (1997). The physiography of the Southwest Indian Ridge. Marine Geophysical Researches 19, 451–455.[CrossRef][Web of Science]
Fisher, R. L., Natland, J. H., Dick, H. J. B. & Meyer, P. S. (1985). Shipboard Description and Inventory of Rocks Dredged on PROTEA Expedition, Leg 5, in the Southwest Indian Ocean. Scripps Institution of Oceanography Reference Series, reference number 85-15, 223 pp.
Green, T. H. & Adam, J. (2003). Experimentally-determined trace element characteristics of aqueous fluid from partially dehydrated mafic oceanic crust at 3·0 GPa, 650–700°C. European Journal of Mineralogy 15, 815–830.
Grégoire, M., Bell, D. R. & le Roex, A. P. (2003). Garnet lherzolites from the Kaapvaal craton (South Africa): trace element evidence for a metasomatic history. Journal of Petrology 44, 629–657.
Groenewald, P. B., Grantham, G. H. & Watkeys, M. K. (1991). Geological evidence for a Proterozoic to Mesozoic link between southeastern Africa and Dronning Maud Land, Antarctica. Journal of the Geological Society, London 148, 1115–1123.
Grunow, A., Hanson, R. & Wilson, T. (1996). Were aspects of Pan-African deformation linked to Iapetus opening? Geology 24, 1063–1066.
Gupta, M. L. (1991). Heat-flow and heat generation in the Archean Dharwar cratons and implications for the southern Indian shield geotherm and lithospheric thickness. Tectonophysics 194, 107–122.[CrossRef][Web of Science]
Gurnis, M., Müller, R. D. & Moresi, L. (1998). Cretaceous vertical motion of Australia and the Australian–Antarctic Discordance. Science 279, 1499–1504.[CrossRef][Web of Science][Medline]
Haggerty, S. E. (1999). Earth and planetary sciences—a diamond trilogy: superplumes, supercontinents, and supernovae. Science 285, 851–860.[CrossRef][Web of Science][Medline]
Hamelin, B. & Allègre, C. J. (1985). Large-scale regional units in the depleted upper mantle revealed by an isotope study of the South-West Indian Ridge. Nature 315, 196–199.[CrossRef]
Hanan, B. B., Blichert-Toft, J., Pyle, D. G. & Christie, D. M. (2004). Contrasting origins of the upper mantle revealed by hafnium and lead isotopes from the Southeast Indian Ridge. Nature 432, 91–94.[CrossRef]
Handke, M. J., Tucker, R. D. & Ashwal, L. D. (1999). Neoproterozoic continental arc magmatism in west–central Madagascar. Geology 27, 351–354.
Hanson, R. E. (2003). Proterozoic geochronology and tectonic evolution of southern Africa. In: Yoshida, M., Windley, B. F. & Dasgupta, S. (eds) Proterozoic East Gondwana: Supercontinent Assembly and Breakup. Geological Society, London, Special Publications 206, 427–463.
Hart, S. R. (1984). A large-scale anomaly in the Southern Hemisphere mantle. Nature 309, 753–757.[CrossRef]
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]
Hart, S. R., Gerlach, D. C. & White, D. M. (1986). A possible new Sr–Nd–Pb mantle array and consequences for mixing. Geochimica et Cosmochimica Acta 50, 1551–1557.[CrossRef][Web of Science]
Hartnady, C. J. H. & le Roex, A. P. (1985). Southern Ocean hotspot tracks and the Cenozoic absolute motion of the African, Antarctic, and South American plates. Earth and Planetary Science Letters 75, 245–257.[CrossRef][Web of Science]
Hauri, E. H., Wagner, T. P. & Grove, T. L. (1994). Experimental and natural partitioning of Th, U, Pb and other trace elements between garnet, clinopyroxene and basaltic melts. Chemical Geology 117, 149–166.[CrossRef][Web of Science]
Hirschmann, M. M. & Stolper, E. M. (1996). A possible role for garnet pyroxenite in the origin of the ‘garnet signature’ in MORB. Contributions to Mineralogy and Petrology 124, 185–208.[CrossRef][Web of Science]
Hofmann, A. W. (1988). Chemical differentiation of the Earth: the relationship between mantle, continental crust and oceanic crust. Earth and Planetary Science Letters 90, 297–314.[CrossRef][Web of Science]
Huang, Y.-M., Van Calsteren, P. & Hawkesworth, C. J. (1995). The evolution of the lithosphere in southern Africa: a perspective on the basic granulite xenoliths from kimberlites in South Africa. Geochimica et Cosmochimica Acta 59, 4905–4920.[CrossRef][Web of Science]
Irving, A. J. & Frey, F. A. (1978). Distribution of trace elements between garnet megacrysts and host volcanic liquids of kimberlitic to rhyolitic composition. Geochimica et Cosmochimica Acta 42, 771–787.[CrossRef][Web of Science]
Jacobs, J., Fanning, C. M., Henjes-Kunst, F., Olesch, M. & Paech, H.-J. (1998). Continuation of the Mozambique Belt into East Antarctica: Grenville-age metamorphism and polyphase Pan-African high-grade events in central Dronning Maud Land. Journal of Geology 106, 385–406.[Web of Science]
Jacobsen, S. B. (1999). The Geochemical Earth Reference Model: composition and evolution of the depleted mantle. EOS Transactions, American Geophysical Union 80, F1173.
James, D. E., Fouch, M. J., VanDecar, J. C. & van der Lee, S. (2001). Tectospheric structure beneath southern Africa. Geophysical Research Letters 28, 2485–2488.[CrossRef][Web of Science]
Johnson, C. M. & Beard, B. L. (1993). Evidence from hafnium isotopes for ancient sub-oceanic mantle beneath the Rio-Grande Rift. Nature 362, 441–444.[CrossRef]
Kamenetsky, V. S., Maas, R., Sushchevskaya, N. M., Norman, M. D., Cartwright, I. & Peyve, A. A. (2001). Remnants of Gondwanan continental lithosphere in oceanic upper mantle: evidence from the South Atlantic Ridge. Geology 29, 243–246.
Kempton, P. D. & Pearce, J. A. (2003). Indian MORB-source mantle: not just a case of plume contamination or sediment recycling. Geophysical Research Abstracts 5, abstract 04017 (CD-ROM).
Kempton, P. D., Pearce, J. A. & Tappin, D. (2000). Hf isotope evidence for mantle domain boundaries in the western Pacific. Journal of Conference Abstracts 5, 576.
Kempton, P. D., Pearce, J. A., Barry, T. L., Fitton, J. G., Langmuir, C. & Christie, D. M. (2002). Sr–Nd–Pb–Hf isotope results from ODP Leg 187: evidence for mantle dynamics of the Australian–Antarctic Discordance and origin of the Indian MORB source. Geochemistry, Geophysics, Geosystems 3, doi: 10.1029/2002GC000320.
Klein, E. M. & Langmuir, C. H. (1987). Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness. Journal of Geophysical Research 92, 8089–8115.
Kröner, A., Sacchi, R., Jaeckel, P. & Costa, M. (1997). Kibaran magmatism and Pan-African granulite metamorphism in northern Mozambique: single zircon ages and regional implications. Journal of African Earth Sciences 25, 467–484.[CrossRef]
Kröner, A., Willner, A. P., Hegner, E., Jaeckel, P. & Nemchin, A. (2001). Single zircon ages, PT evolution and Nd isotopic systematics of high-grade gneisses in southern Malawi and their bearing on the evolution of the Mozambique belt in southeastern Africa. Precambrian Research 109, 257–291.[CrossRef][Web of Science]
Lassiter, J. C., Hauri, E. H., Reiners, P. W. & Garcia, M. O. (2000). Generation of Hawaiian post-erosional lavas by melting of a mixed lherzolite/pyroxenite source. Earth and Planetary Science Letters 178, 269–284.[CrossRef][Web of Science]
Lee, D.-C., Halliday, A. N., Davies, G. R., Essene, E. J., Fitton, J. G. & Temdjim, R. (1996). Melt enrichment of shallow depleted mantle: a detailed petrological trace element and isotopic study of mantle-derived xenoliths and megacrysts from the Cameroon line. Journal of Petrology 37, 415–441.
Lenardic, A., Moresi, L. & Muhlhaus, H. (2003). Longevity and stability of cratonic lithosphere: insights from numerical simulations of coupled mantle convection and continental tectonics. Journal of Geophysical Research 108, article number 2303.
le Roex, A. P. (1985). Geochemistry, mineralogy and magmatic evolution of the basaltic and trachytic lavas from Gough Island, South Atlantic. Journal of Petrology 26, 149–186.
le Roex, A. P. & Erlank, A. J. (1982). Quantitative evaluation of fractional crystallization in Bouvet Island lavas. Journal of Volcanology and Geothermal Research 13, 309–338.[CrossRef][Web of Science]
le Roex, A. P., Dick, H. J. B., Erlank, A. J., Reid, A. M., Frey, F. A. & Hart, S. R. (1983). Geochemistry, mineralogy and petrogenesis of lavas erupted along the Southwest Indian Ridge between the Bouvet triple junction and 11 degrees east. Journal of Petrology 24, 267–318.[Web of Science]
le Roex, A. P., Dick, H. J. B., Reid, A. M., Frey, F. A., Erlank, A. J. & Hart, S. R. (1985). Petrology and geochemistry of basalts from the American–Antarctic Ridge, Southern Ocean: implications for the westward influence of the Bouvet mantle plume. Contributions to Mineralogy and Petrology 90, 367–380.[CrossRef][Web of Science]
le Roex, A. P., Dick, H. J. B. & Fisher, R. L. (1989). Petrology and geochemistry of MORB from 25°E to 46°E along the Southwest Indian Ridge: evidence for contrasting styles of mantle enrichment. Journal of Petrology 30, 947–986.
le Roex, A. P., Dick, H. J. B. & Watkins, R. T. (1992). Petrogenesis of anomalous K-enriched MORB from the Southwest Indian Ridge: 11°53'E to 14°38'E. Contributions to Mineralogy and Petrology 110, 253–268.[CrossRef][Web of Science]
Le Roux, P. J., le Roex, A. P., Schilling, J. G., Shimizu, N., Perkins, W. W. & Pearce, N. J. G. (2002a). Mantle heterogeneity beneath the southern Mid-Atlantic Ridge: trace element evidence for contamination of ambient asthenospheric mantle. Earth and Planetary Science Letters 203, 479–498.[CrossRef][Web of Science]
Le Roux, P. J., le Roex, A. P. & Schilling, J.-G. (2002b). MORB melting processes beneath the southern Mid-Atlantic Ridge (40–55°S): a role for mantle plume-derived pyroxenite. Contributions to Mineralogy and Petrology 144, 206–229.[Web of Science]
Lin, P. N., Stern, R. J., Morris, J. & Bloomer, S. H. (1990). Nd and Sr-isotopic compositions of lavas from the northern Mariana and southern Volcano arcs: implications for the origin of island arc melts. Contributions to Mineralogy and Petrology 105, 381–392.[CrossRef][Web of Science]
Mahoney, J. J., Natland, J. H., White, W. M., Poreda, R., Bloomer, S. M., Fisher, R. L. & Baxter, A. N. (1989). Isotopic and geochemical provinces of the western Indian Ocean spreading centers. Journal of Geophysical Research 94, 283–296.
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., Fisher, R. L. & Natland, J. H. (1992). Southwestern limits of Indian Ocean Ridge mantle and the origin of low 206Pb/204Pb mid-ocean ridge basalts: isotope systematics of the Southwest Indian Ridge (17°–50°E). Journal of Geophysical Research 97, 19771–19790.[CrossRef]
Mahoney, J. J., Sinton, J. M., Kurz, M. D., Macdougall, J. D., Spencer, K. J. & Lugmair, G. W. (1994). Isotope and trace element characteristics of a super-fast spreading ridge: East Pacific Rise, 13–23°S. Earth and Planetary Science Letters 121, 173–193.[CrossRef][Web of Science]
Mahoney, J. J., Graham, D. W., Christie, D. M., Johnson, K. T. M., Hall, L. S. & Vonderhaar, D. L. (2002). Between a hotspot and a cold spot: isotopic variation in the Southeast Indian Ridge asthenosphere, 86°E–118°E. Journal of Petrology 43, 1155–1176.
Mattielli, N., Weis, D., Blichert-Toft, J. & Albarède, F. (2002). Hf isotope evidence for a Miocene change in the Kerguelen mantle plume composition. Journal of Petrology 43, 1327–1339.
McDonough, W. F. (1990). Constraints on the composition of the continental lithospheric mantle. Earth and Planetary Science Letters 101, 1–18.[CrossRef][Web of Science]
McDonough, W. F. & McCulloch, M. T. (1987). The southeast Australian lithospheric mantle: isotopic and geochemical constraints on its growth and evolution. Earth and Planetary Science Letters 86, 327–340.[CrossRef][Web of Science]
McKenzie, D. & O'Nions, R. K. (1991). Partial melt distributions from inversion of rare-earth element concentrations. Journal of Petrology 32, 1021–1091.
Meert, J. G. (2003). A synopsis of events related to the assembly of East Gondwana. Tectonophysics 362, 1–40.[CrossRef][Web of Science]
Meisel, T., Walker, R. J. & Morgan, J. W. (1996). The osmium isotopic composition of the Earth's primitive upper mantle. Nature 383, 517–520.[CrossRef]
Menzies, M. A. & Murthy, V. R. (1980). Enriched mantle: Nd and Sr isotopes in diopsides from kimberlite nodules. Nature 283, 634–636.[CrossRef]
Michard, A., Montigny, R. & Schlich, R. (1986). Geochemistry of the mantle beneath the Rodriguez triple junction and the South-East Indian Ridge. Earth and Planetary Science Letters 78, 104–114.[CrossRef][Web of Science]
Milner, S. C. & le Roex, A. P. (1996). Isotope characteristics of the Okenyenya igneous complex, northwestern Namibia: constraints on the composition of the early Tristan plume and the origin of the EM 1 mantle component. Earth and Planetary Science Letters 141, 277–291.[CrossRef][Web of Science]
Mühe, R., Devey, C. W. & Bohrmann, H. (1993). Isotope and trace element geochemistry of MORB from the Nansen–Gakkel Ridge at 86°N. Earth and Planetary Science Letters 120, 103–109.[CrossRef][Web of Science]
Mühe, R., Bohrmann, H., Garbe-Schönberg, D. & Kassens, H. (1997). E-MORB glasses from the Gakkel Ridge (Arctic Ocean) at 87°N: evidence for the Earth's most northerly volcanic activity. Earth and Planetary Science Letters 152, 1–9.[CrossRef][Web of Science]
Müller, M. R., Robinson, C. J., Minshull, T. A., White, R. S. & Bickle, M. J. (1997). Thin crust beneath Ocean Drilling Program borehole 735B at the Southwest Indian Ridge? Earth and Planetary Science Letters 148, 93–107.[CrossRef][Web of Science]
Nelson, D. R., Chivas, A. R., Chappell, B. W. & McCulloch, M. T. (1988). Geochemical and isotopic systematics in carbonatites and implications for the evolution of ocean island sources. Geochimica et Cosmochimica Acta 52, 1–17.[CrossRef][Web of Science]
Niu, Y. & Batiza, R. (1991). An empirical method for calculating melt compositions produced beneath mid-ocean ridges: applications for axis and off-axis (seamounts) melting. Journal of Geophysical Research 96, 21753–21777.[CrossRef]
Nowell, G. M., Kempton, P. D., Noble, S. R., Fitton, J. G., Saunders, A. D., Mahoney, J. J. & Taylor, R. N. (1998). High precision Hf isotope measurements of MORB and OIB by thermal ionisation mass spectrometry: insights into the depleted mantle. Chemical Geology 149, 211–233.[CrossRef][Web of Science]
O'Neill, H. S. C. (1981). The transition between spinel lherzolite and garnet lherzolite, and its use as a geobarometer. Contributions to Mineralogy and Petrology 77, 185–194.[CrossRef][Web of Science]
Patchett, P. J. & Tatsumoto, M. (1980a). Hafnium isotope variations in oceanic basalts. Geophysical Research Letters 7, 1077–1080.[Web of Science]
Patchett, P. J. & Tatsumoto, M. (1980b). A routine high-precision method for Lu–Hf isotope geochemistry and chronology. Contributions to Mineralogy and Petrology 75, 263–267.[CrossRef][Web of Science]
Patchett, P. J., White, W. M., Feldmann, H., Kielinczuk, S. & Hofmann, A. W. (1984). Hafnium/rare earth element fractionation in the sedimentary system and crustal recycling into the Earth's mantle. Earth and Planetary Science Letters 69, 365–378.[CrossRef][Web of Science]
Paulsson, O. & Austrheim, H. (2003). A geochronological and geochemical study of rocks from Gjelsvikfjella, Dronning Maud Land, Antarctica—implications for Mesoproterozoic correlations and assembly of Gondwana. Precambrian Research 125, 113–138.[CrossRef][Web of Science]
Pearce, J. A., Baker, P. E., Harvey, P. K. & Luff, I. W. (1995). Geochemical evidence for subduction fluxes, mantle melting and fractional crystallization beneath the South Sandwich Island Arc. Journal of Petrology 36, 1073–1109.
Pearce, J. A., Kempton, P. D., Nowell, G. M. & Noble, S. R. (1999). Hf–Nd element and isotope perspective on the nature and provenance of mantle and subduction components in western Pacific arc–basin systems. Journal of Petrology 40, 1579–1611.[CrossRef][Web of Science]
Pertermann, M. & Hirschmann, M. M. (2003). Partial melting experiments on a MORB-like pyroxenite between 2 and 3 GPa: constraints on the presence of pyroxenite in basalt source regions from solidus location and melting rate. Journal of Geophysical Research 108, paper number 2000JB000118.
Plank, T. & Langmuir, C. H. (1998). The chemical composition of subducting sediment and its consequences for the crust and mantle. Chemical Geology 145, 325–394.[CrossRef][Web of Science]
Price, R. C., Kennedy, A. K., Riggs-Sneeringer, M. & Frey, F. A. (1986). Geochemistry of basalts from the Indian Ocean triple junction: implications for the generation and development of Indian Ocean ridge basalts. Earth and Planetary Science Letters 78, 379–396.[CrossRef][Web of Science]
Pyle, D. G., Christie, D. M. & Mahoney, J. J. (1992). Resolving an isotopic boundary within the Australian–Antarctic discordance. Earth and Planetary Science Letters 112, 161–178.[CrossRef][Web of Science]
Pyle, D. G., Christie, D. M., Mahoney, J. J. & Duncan, R. A. (1995). Geochemistry and geochronology of ancient southeast Indian and southwest Pacific seafloor. Journal of Geophysical Research 100, 22261–22282.[CrossRef]
Rehkämper, M. & Hofmann, A. W. (1997). Recycled ocean crust and sediment in Indian Ocean MORB. Earth and Planetary Science Letters 147, 93–106.[CrossRef][Web of Science]
Rehkämper, M. & Halliday, A. N. (1998). Accuracy and long-term reproducibility of lead isotopic measurements by MC-ICP-MS using an external method for correction of mass discrimination. International Journal of Mass Spectrometry and Ion Processes 58, 123–133.
Robinson, C. J., White, R. S., Bickle, M. J. & Minshull, T. A. (1996). Restricted melting under the very slow-spreading Southwest Indian Ridge. In: MacLeod, C. J., Tyler, P. A. & Walker, C. L. (eds) Tectonic, Magmatic, Hydrothermal and Biological Segmentation of Mid-Ocean Ridges. Geological Society, London, Special Publications 118, 131–141.[CrossRef]
Rudnick, R. L. (1990). Nd and Sr isotopic compositions of lower crustal xenoliths from North Queensland, Australia: implications for Nd model ages and crustal growth processes. Chemical Geology 83, 195–208.[CrossRef][Web of Science]
Rudnick, R. L. & Fountain, D. M. (1995). Nature and composition of the continental crust: a lower crustal perspective. Reviews of Geophysics 33, 267–309.[CrossRef][Web of Science]
Rudnick, R. L., McDonough, W. F., McCulloch, M. T. & Taylor, S. R. (1986). Lower crustal xenoliths from Queensland, Australia: evidence for deep crustal assimilation and fractionation of continental basalts. Geochimica et Cosmochimica Acta 50, 1099–1115.[CrossRef][Web of Science]
Saal, A. E., Rudnick, R. L., Ravizza, G. E. & Hart, S. R. (1998). Re–Os isotope evidence for the composition, formation and age of the lower crust. Nature 393, 58–61.[CrossRef]
Salters, V. J. M. (1996). The generation of mid-ocean ridge basalts from the Hf and Nd isotope perspective. Earth and Planetary Science Letters 141, 109–123.[CrossRef][Web of Science]
Salters, V. J. M. & Dick, H. J. B. (2002). Mineralogy of the mid-ocean-ridge basalt source from neodymium isotopic composition of abyssal peridotites. Nature 418, 68–72.[CrossRef]
Salters, V. J. M. & Hart, S. R. (1989). The hafnium paradox and the role of garnet in the source of mid-ocean-ridge basalts. Nature 342, 420–422.[CrossRef]
Salters, V. J. M. & Hart, S. R. (1991). The mantle sources of ocean ridges, islands and arcs: the Hf-isotope connection. Earth and Planetary Science Letters 104, 364–380.[CrossRef][Web of Science]
Salters, V. J. M. & Longhi, J. (1999). Trace element partitioning during the initial stages of melting beneath mid-ocean ridges. Earth and Planetary Science Letters 166, 15–30.[CrossRef][Web of Science]
Salters, V. J. M. & Stracke, A. (2004). Composition of the depleted mantle. Geochemistry, Geophysics, Geosystems 5, doi: 10.1029/2003GC000597.
Salters, V. J. M. & White, W. M. (1998). Hf isotopic constraints on mantle evolution. Chemical Geology 145, 447–460.[CrossRef][Web of Science]
Salters, V. J. M., Longhi, J. E. & Bizimis, M. (2002). Near mantle solidus trace element partitioning at pressures up to 3·4 GPa. Geochemistry, Geophysics, Geosystems 3, doi: 10.1029/2001GC000148.
Scherer, E. E., Cameron, K. L., Johnson, C. M., Beard, B. L., Barovich, K. M. & Collerson, K. D. (1997). Lu–Hf geochronology applied to dating Cenozoic events affecting lower crustal xenoliths from Kilbourne Hole, New Mexico. Chemical Geology 142, 63–78.[CrossRef][Web of Science]
Schmitz, M. D., Vervoort, J. D., Bowring, S. A. & Patchett, J. P. (2004). Decoupling of the Lu–Hf and Sm–Nd systems during the evolution of granulitic lower crust beneath southern Africa. Geology 32, 405–408.
Scotese, C. R. (2004). A continental drift flipbook. Journal of Geology 112, 729–741.[CrossRef][Web of Science]
Shirey, S. B., Bender, J. F. & Langmuir, C. H. (1987). Three-component isotopic heterogeneity near the Oceanographer Transform, Mid-Atlantic Ridge. Nature 325, 217–223.[CrossRef]
Simon, N. S. C., Carlson, R. W., Pearson, D. G. & Davies, G. R. (2002). The Lu–Hf isotope composition of cratonic lithosphere: disequilibrium between garnet and clinopyroxene in kimberlite xenoliths. Geochimica et Cosmochimica Acta 66(Supplement), A717.
Simon, N. S. C., Carlson, R. W., Davies, G. R., Nowell, G. M. & Pearson, D. G. (2003). Os–Sr–Nd–Hf isotope evidence for the ancient depletion and subsequent multi-stage enrichment history of the Kaapvaal cratonic lithosphere. In: Extended Abstracts of the Eighth International Kimberlite Conference, abstract number 117 (CD-ROM).
Staudigel, H., Plank, T., White, W. M. & Schminke, H. U. (1996). Geochemical fluxes during seafloor alteration of the basaltic upper crust: DSDP Sites 417 and 418 (Overview). In: Bebout, G. E., Scholl, D. W., Kirby, S. H. & Platt, J. P. (eds) Subduction: Top to Bottom. Geophysical Monograph, American Geophysical Union 96, 19–38.
Stern, R. J. (1994). Arc-assembly and continental collision in the Neoproterozoic East African orogen: implications for the consolidation of Gondwanaland. Annual Review of Earth and Planetary Sciences 22, 319–352.[Web of Science]
Storey, B. C. (1995). The role of mantle plumes in continental breakup: case histories from Gondwanaland. Nature 377, 301–308.[CrossRef]
Storey, M., Saunders, A. D., Tarney, J., Gibson, I. L., Norry, M. J., Thirlwall, M. F., Leat, P., Thompson, R. N. & Menzies, M. A. (1989). Contamination of Indian Ocean asthenosphere by the Kerguelen–Heard mantle plume. Nature 338, 574–576.[CrossRef]
Storey, M., Mahoney, J. J. & Saunders, A. D. (1998). 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. Geophysical Monograph, Washington American Geophysical Union 100, 95–122.
Subbarao, K. V. & Hedge, C. E. (1973). K, Rb, Sr and 87Sr/86Sr in rocks from the Mid-Indian Ocean Ridge. Earth and Planetary Science Letters 18, 223–228.[CrossRef][Web of Science]
Sun, S.-S. (1980). Lead isotopic study of young volcanic rocks from mid-ocean ridges, ocean islands and island arcs. Philosophical Transactions of the Royal Society of London, Series A 297, 409–445.[CrossRef]
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.
Taylor, S. R. & McClennan, S. M. (1985). The Continental Crust: its Composition and Evolution. Oxford: Blackwell Scientific.
Todt, W., Cliff, R. A., Hanser, A. & Hofmann, A. W. (1996). Evaluation of a 202Pb–205Pb double spike for high-precision lead isotope analysis. In: Basu, A. & Hart, S. R. (eds) Earth Processes: Reading the Isotopic Code. Geophysical Monograph, American Geophysical Union 95, 429–437.
Tucker, R. D., Ashwal, L. D., Handke, M. J., Hamilton, M. A., Le Grange, M. & Rambeloson, R. A. (1999). U–Pb geochronology of the Archean and Proterozoic rocks of north–central Madagascar. Journal of Geology 107, 135–153.[CrossRef][Web of Science]
Vervoort, J. D., Patchett, P. J., Blichert-Toft, J. & Albarède, F. (1999). Relationships between Lu–Hf and Sm–Nd isotopic systems in the global sedimentary system. Earth and Planetary Science Letters 168, 79–99.[CrossRef][Web of Science]
Vervoort, J. D., Patchett, J. P., Albarède, F., Blichert-Toft, J., Rudnick, R. & Downes, H. (2000). Hf–Nd isotopic evolution of the lower crust. Earth and Planetary Science Letters 181, 115–129.[CrossRef][Web of Science]
Walker, R. J., Carlson, R. W., Shirey, S. B. & Boyd, F. R. (1989). Os, Sr, Nd and Pb isotope systematics of southern African peridotite xenoliths: implications for the chemical evolution of subcontinental mantle. Geochimica et Cosmochimica Acta 53, 1583–1595.[CrossRef][Web of Science]
Weaver, B. L., Wood, D. A., Tarney, J. & Joron, J. L. (1986). Role of subducted sediment in the genesis of ocean island basalts: geochemical evidence from South Atlantic Ocean islands. Geology 14, 275–278.
Weis, D. & Frey, F. A. (1996). Role of the Kerguelen Plume in generating the eastern Indian Ocean seafloor. Journal of Geophysical Research 101, 13831–13849.[CrossRef]
White, W. M., Hofmann, A. W. & Puchelt, H. (1987). Isotope geochemistry of Pacific mid-ocean ridge basalt. Journal of Geophysical Research 92, 4881–4893.
Williams, G. A. & Turekian, K. K. (2004). The glacial–interglacial variation of seawater osmium isotopes as recorded in Santa Barbara Basin. Earth and Planetary Science Letters 228, 379–389.[CrossRef][Web of Science]
Woodhead, J. D., Hergt, J. M., Davidson, J. P. & Eggins, S. M. (2001). Hafnium isotope evidence for ‘conservative’ element mobility during subduction zone processes. Earth and Planetary Science Letters 192, 331–346.[CrossRef][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]
Yang, H.-J., Frey, F. A., Weis, D., Giret, A., Pyle, D. G. & Michon, G. (1998). Petrogenesis of the flood basalts forming the northern Kerguelen Archipelago: implications for the Kerguelen Plume. Journal of Petrology 39, 711–748.[CrossRef][Web of Science]
Zhang, S.-Q., Mahoney, J. J., Mo, X.-X., Ghazi, A. M., Milani, L., Crawford, A. J., Guo, T.-Y. & Zhao, Z.-D. (2005). Evidence for a widespread Tethyan upper mantle with Indian Ocean-type isotopic characteristics. Journal of Petrology 46, 829–858.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  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]
|
|
|
|
 |
|
 J. G. Fitton The OIB paradox Geological Society of America Special Papers, January 1, 2007; 430(0): 387 - 412. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||