Journal of Petrology Advance Access originally published online on July 2, 2004
Journal of Petrology 2004 45(8):1515-1537; doi:10.1093/petrology/egh014
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology 45(8) © Oxford University Press 2004; all rights reserved
Evidence for Early LREE-enriched Mantle Source Regions: Diverse Magmas from the c. 3·0 Ga Mallina Basin, Pilbara Craton, NW Australia
1 GEOLOGICAL SURVEY OF WESTERN AUSTRALIA, 100 PLAIN STREET, EAST PERTH, WA, 6004, AUSTRALIA
2 GEOSCIENCE AUSTRALIA, GPO BOX 378, CANBERRA, ACT, 2601, AUSTRALIA
RECEIVED JULY 10, 2002; ACCEPTED DECEMBER 15, 2003
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONThe c. 2·97–2·95 Ga magmatic history of the Mallina Basin, in the Pilbara Craton of NW Australia, includes what is perhaps the most lithologically diverse magmatism of any similar-sized Archaean terrain, and is unusual for similar-sized terrains of any age. The magmatism includes light rare earth element (LREE)-rich basaltic rocks, LREE-rich gabbros and rocks with boninite-like compositions (collectively the ‘Mallina mafic suite’), and high-Mg diorites (sanukitoids). The Mallina mafic suite is characterized by high primitive mantle normalized (La/Nb)PM (>3) and (La/Yb)PM (>2), and non-radiogenic Nd-isotopic compositions (
Nd(2·95 Ga) mostly <–1·0), suggesting that the magmas incorporated a crustal component. Despite having intruded through compositionally diverse continental crust, the magmatic rocks show a remarkably narrow range in La/Nb (
3·1), La/Sm (5·3) and La/Zr (0·15), and a small range of Nd(2·95 Ga) (–0·6 to –2·8) that is unlikely to be a result of assimilation of any single locally or regionally available crustal component. The Mallina mafic suite was probably derived from a mantle source that incorporated a homogeneous mix of ‘old Pilbara crust’ [i.e. >3·3 Ga, Nd(2·95 Ga) <–2·3, high La/Nb (Sm, Zr)] and crust that resembled the c. 3·12 Ga greenstones of the Whundo Group [Nd(2·95 Ga) >–0·4, low La/Nb (Sm, Zr)], which crop out to the NW of the basin. Compared with the Mallina mafic suite, the high-Mg diorites (sanukitoids) have higher Nd(2·95 Ga) (–0·4 to +1·2), suggesting a source that incorporated a greater proportion of the Whundo-like component. Evidence for enrichment of Archaean mantle source regions is typically extremely difficult to demonstrate and is primarily restricted to sequences that are c. 2·8 Ga or younger. The igneous rocks of the Mallina Basin, however, show that such sources existed by c. 3·0 Ga. Subduction of oceanic crust, including compositionally homogenized sediment, is the most obvious model for this mantle enrichment. KEY WORDS: Archaean; crustal evolution; enriched mantle; mafic magmas; boninite; sanukitoid; subduction
|
INTRODUCTION |
|---|
Although it is widely believed that convergent plate margin processes played a role in Archaean crustal evolution (e.g. Moorbath, 1977
; Martin, 1994, 1999; de Wit, 1998; Kusky & Polat, 1999), it is not clear when processes such as the enrichment of mantle source regions by subducted crust commenced. The typical geochemical signature of modern subduction environments reflects a mantle source enriched in large ion lithophile elements (LILE—Cs, K, Rb, Sr, Ba), Th, and often also in light rare earth elements (LREE; e.g. La, Ce), and relative depletion of high field strength elements (HFSE; e.g. Nb, Ta, Zr and Ti) (Pearce, 1982; Pearce & Cann, 1993). Such signatures, however, can also be acquired in non-subduction settings through crustal assimilation (e.g. Arndt & Jenner, 1986; Barley, 1986; Sun et al., 1989; Green et al., 2000). In both settings, the magmas have incorporated crust and might differ only in terms of where that process occurred (i.e. in the mantle or during melt migration). In oceanic settings, the opportunity for continent crustal assimilation is minimized, and so LILE, Th and LREE enrichments in many modern oceanic arc volcanics can be more confidently attributed to enrichment of the mantle source (e.g. Sun & Nesbitt, 1978; Cameron et al., 1979; Crawford et al., 1989). However, in settings that provide an opportunity for crustal assimilation, or in deformed terrains where an oceanic setting cannot be established, geochemistry alone commonly cannot uniquely identify the mechanism of enrichment. Opinion is strongly divided, for example, on the origin of crustal signatures in low-Ti continental flood basalts—both crustal assimilation (Mensing et al., 1984; Arndt et al., 1993, 2001) and enriched mantle source regions (Hooper & Hawkesworth, 1993; Hergt & Brauns, 2001; Riley et al., 2003) have been proposed. In many cases, only by establishing that the range of available and inferred crustal components are geochemically unsuitable contaminants (e.g. Tarney, 1992), and by demonstrating that subduction processes are capable of transporting material of appropriate composition to the site of melting, can such enrichments be attributed to a mantle source.
Geochemical evidence for subduction is not extensive in the Archaean igneous rock record, particularly before c. 3·0 Ga (Smithies, 2000; Smithies & Champion, 2000). The deformation and metamorphism typical of the Archaean record provides few cases where the tectonic environment can be unambiguously identified as oceanic and free of any possible influence of felsic crust. In general, the older the Archaean rock succession, the more difficult the geological and geochemical data are to interpret—but, potentially, the more important those interpretations become in terms of understanding the evolution of crustal growth processes. For example, it is clear that some c. 3·7–3·8 Ga basalts from West Greenland incorporated a crustal component (Polat et al., 2002), but whether this occurred in the mantle source or through later assimilation is extremely difficult to prove chemically.
One region that has been assigned to an oceanic environment is the Abitibi Subprovince of the Canadian Superior Province (e.g. Kerrich et al., 1998). Here, boninite-like rocks, high-Nb basalts, sanukitoids and calc-alkaline basalts and andesites provide evidence for modern-style subduction as far back as c. 2·7 Ga (Shirey & Hanson, 1984; Kerrich et al., 1998; Wyman et al., 2000; Polat & Kerrich, 2001, 2002). Shirey & Hanson (1984) and Polat & Kerrich (2002) showed that the mantle source for some of these magmas was enriched 200 Myr previously.
Here we report on the c. 2·97–2·95 Ga Mallina Basin in the NW Australian Pilbara Craton (Fig. 1). The geochemical diversity of igneous rocks in this small area is extremely unusual for terrains of any age. It includes three groups of LREE-rich basaltic rocks, LREE-rich gabbro, rocks with boninite-like compositions, and sanukitoids. More importantly, many of these rocks are characterized by high Th and LREE concentrations and high ratios of these elements against HFSE; features consistent with a subduction-modified source. Such a suggestion is, however, complicated by the fact that at least some of the magmatism occurred within a continental setting and with no contemporaneous subduction (Smithies et al., 2001). Barley (1986) and Arndt et al. (2001) concluded that the LREE enrichments observed in two of the basaltic suites probably resulted from crustal assimilation. However, we suggest that the balance of evidence presented here favours LREE enrichment of a mantle source.
|
Subduction in this part of the Pilbara Craton is speculated to have occurred at c. 3·12 Ga (Smith et al., 1998
) and at c. 3·01 Ga (Barley et al., 1984; Pike, 2001). Barley et al. (1984), however, recognized the difficulty in establishing the origin of incompatible trace-element enrichments in the c. 3·01 Ga felsic volcanics, as well as in older felsic volcanics within the craton, many of which are likely to be lower-crustal melts (Cullers et al., 1993) or low-pressure residues of extensive basalt fractionation. If our interpretation is correct, then the Mallina magmas provide some of the earliest direct evidence for mantle enrichment—possibly as far back as c. 3·12 Ga—and support subduction models for either or both the c. 3·01 and c. 3·12 Ga sequences. |
GEOLOGICAL BACKGROUND |
|---|
The ENE-trending Mallina Basin developed in the central Pilbara Craton, NW Australia, straddling the boundary between two distinct terranes (East and West Pilbara Granite–Greenstone Terranes; Fig. 1) that are characterized by greenstones aged c. 3·12 Ga and older (Van Kranendonk et al., 2002
). The youngest basement rocks are exposed to the NW and include c. 3·01 Ga volcaniclastic rocks of the Whim Creek Group, c. 3·015–3·02 Ga cherts and banded iron-formation of the Cleaverville Formation, c. 3·1 Ga and c. 2·99 Ga phases of the Caines Well Granitoid Complex, and rocks of the c. 3·12 Ga Whundo Group (Smithies et al., 2001). The last group is dominated by tholeiitic basalts, but also contains basalts with calc-alkaline affinities and up to 5% felsic volcanic rocks (R. H. Smithies & D. C. Champion, unpublished data; Hickman, 1997; Smith et al., 1998).
Previous studies have shown that the main period of deposition within the Mallina Basin occurred between 2·97 Ga and 2·95 Ga (Smithies et al., 2001) (Fig. 2). The main basin fill comprises medium- to coarse-grained wacke and conglomerate (Constantine Sandstone) and conformably overlying medium- to fine-grained siliciclastic turbidites (Mallina Formation). Along the northwestern margin of the basin, fine- to coarse-grained sediments of the Bookingarra Group filled local fault-bounded depocentres. Volcaniclastic turbidites sourced primarily from the unconformably underlying Whim Creek Group (Fig. 2) dominate the lower Bookingarra Group; siliciclastic turbidites become dominant up sequence (Pike, 2001; Pike & Cas, 2002).
|
Subaqueous flows of siliceous high-Mg basalt (Louden Volcanics) and of rocks that Arndt et al. (2001)
called ‘high-Si basalts’ (Mount Negri Volcanics) dominate the upper Bookingarra Group. Their sub-volcanic gabbroic equivalents (LREE-rich gabbro) are widespread throughout the southern and eastern part of the basin (Figs 1 and 2). They also occur up to 50 km NW of the basin, in the Cleaverville area, within a distinctly separate tectonostratigraphic terrane (West Pilbara Granite–Greenstone Terrane of Van Kranendonk et al., 2002) with a separate geological history prior to intrusion of the gabbros. Subaqueous flows and hyaloclastite deposits of high-Si basalt (south Mallina basalt) form a widespread discontinuous layer within siliciclastic turbidite units in the southern and central parts of the Mallina Basin and may be time correlatives of the Louden and Mount Negri Volcanics.
Rocks with compositions similar to boninites (referred to here as boninite-like rocks) form a thin (300 m) layer in the southern part of the basin (Fig. 1). They crop out over a strike length of more than 50 km, but their original extent has been obscured by later granite intrusion. Rare spinifex textures in the boninite-like rocks provide evidence of rapid cooling but no definitive evidence for an extrusive origin has been found. The extensive continuity of this thin layer in coarse-grained sedimentary rocks suggests it is a sill and so the textural evidence for rapid cooling probably indicates a high-level (sub-volcanic) intrusive origin (Smithies, 2002).
The boninite-like rocks were intruded by the alkaline Portree Granitoid Complex and by rocks of a high-Mg diorite (sanukitoid) suite, between 2·955 and 2·945 Ga (Smithies & Champion, 2000). Rocks of this suite form a NE-trending chain of high-level intrusions extending for over 150 km along the axis of the basin (Figs 1 and 2). In the east of the basin, the high-Mg diorite suite was emplaced into zones of active dilation related to extension along crustal-scale basin-parallel faults.
The available geochronology (Van Kranendonk et al., 2002, and references therein) indicates that the Louden Volcanics, high-Si basalts, LREE-rich gabbros and boninite-like rocks pre-date intrusion of the high-Mg diorite suite by up to 20 Myr. However, a close spatial relationship between the boninite-like rocks and the high-Mg diorite suite, and evidence that these rocks were emplaced at a high crustal level (Smithies, 2002), suggests that this magmatism occurred within the short interval between 2·955 Ga and 2·945 Ga.
Renewed basin extension resulted in additional sedimentation between 2·945 Ga and 2·935 Ga (Smithies et al., 2001). Voluminous high-K monzogranite swamped the region between c. 2·935 and 2·925 Ga, particularly adjacent to and south of the basin, but also within the Caines Well Granitoid Complex. High-K magmatism becomes systematically younger and less voluminous away from the Mallina Basin, indicating that the region beneath the basin was the focus of the 2·97–2·925 Ga tectonothermal event.
Massive outpourings of basaltic magma blanketed the region at c. 2·7 Ga (Nelson et al., 1992). These basalts of the Fortescue Group closely resemble Phanerozoic flood basalts in terms of composition, volume and environment of deposition (e.g. Nelson et al., 1992; Arndt et al., 2001) and share many compositional similarities with the Louden and Mount Negri Volcanics (Arndt et al., 2001).
Whereas the lower Bookingarra Group (Fig. 2) contains a minor juvenile felsic volcanic component (e.g. Pike, 2001; Pike & Cas, 2002), the main fill of the Mallina Basin contains no evidence of the felsic volcanic detritus that might be expected in a basin fringing a contemporaneous arc, and particularly a back-arc. Provenance studies of dated detrital zircon populations show that sources on both sides of the Mallina Basin made significant contributions to the sediment fill (Smithies et al., 2001).
The felsic volcanic and volcaniclastic rocks of the older Whim Creek Group, and felsic intrusions to the north of the Mallina Basin, provide evidence for a possible c. 3·01–3·015 Ga arc (Barley et al., 1984; Pike & Cas, 2002), but the Mallina Basin and associated magmatism evolved at least 40 Myr later. Nevertheless, some modern convergent margins record a punctuated volcanic history that spans much longer periods (e.g. >70 Myr for the Greater Antilles; Peate et al., 1997). Hence, the exact tectonic setting of the Mallina Basin remains unclear, and although Smithies et al. (2001) preferred a continental rift setting, we cannot discount a convergent margin setting as proposed by Smith et al. (1998). An important point, however, is that there is no regional evidence for subduction contemporaneous with the main (2·97– 2·955 Ga) or later (2·945–2·935 Ga) depositional phases of the Mallina Basin, and hence with the magmatism discussed here.
|
PETROGRAPHY OF THE LREE-RICHROCKS OF THE MALLINA BASIN |
|---|
All rocks of the Mallina Basin have recrystallized at lower to middle greenschist facies. Recrystallization has been most intense for the mafic rocks (Louden Volcanics, high-Si basalts, LREE-rich gabbros, boninite-like rocks), which only locally preserve a primary mineralogy and are now typically dominated by actinolite, chlorite and epidote, and also serpentine and talc in the more magnesian rocks (boninite-like rocks and some Louden Volcanics). Rocks of the high-Mg diorite suite typically show only incipient replacement of primary mafic phases by actinolite and chlorite and of plagioclase by epidote and sericite.
Vesicular flows are common within the Mount Negri Volcanics, whereas the south Mallina basalt locally forms hyaloclastite deposits. Relict primary mineralogy is dominated by locally acicular pyroxene and plagioclase. Dykes and sills of LREE-rich gabbro comprise late interstitial graphic intergrowths of quartz and plagioclase and a pyroxene-rich primary mineralogy that is now completely recrystallized to actinolite.
The Louden Volcanics locally show well-developed pyroxene spinifex textures. Aphyric rocks are locally common and may form extensive units many metres thick. Non-cumulate rocks are dominated by acicular skeletal crystals up to 5 cm in length with clinopyroxene rims that enclose a core dominated by chlorite and minor serpentine (after olivine), and rare relicts of orthopyroxene. Some rocks contain very rare and late-crystallizing plagioclase. Cumulate-textured rocks occur both in dykes and in thicker flow units and typically contain phenocrysts of olivine enclosed in orthopyroxene and rimmed by clinopyroxene.
The boninite-like rocks are typically a schistose assemblage of actinolite, chlorite and serpentine with accessory epidote, talc, plagioclase and quartz. Rarely preserved primary textures include oriented pyroxene (now actinolite) spinifex textures.
The rocks of the Mallina high-Mg diorite suite range from diorite and monzodiorite to tonalite and granodiorite (Smithies & Champion, 2000). Mesocratic hornblende–biotite granodiorite is the most common. The most mafic rocks are fine-grained melanodiorite forming chilled margins to some intrusions. These contain phenocrysts of plagioclase, clinopyroxene, hornblende and locally, orthopyroxene.
|
GEOCHEMISTRY OF THE LREE-RICH ROCKS OF THE MALLINA BASIN |
|---|
Table 1 contains new analyses of the Louden Volcanics (nine), Mount Negri Volcanics (10), south Mallina basalts (10) and LREE-rich gabbros (eight) as well as representative analyses covering the compositional range of the boninite-like rocks and high-Mg diorite suite. Despite variable degrees of recrystallization, the mafic igneous rocks of the Mallina Basin show moderately well-constrained compositional variations for some elements that are typically highly mobile under low-grade metamorphism (e.g. Smithies & Champion, 2000
; Smithies, 2002). The boninite-like rocks, in particular, were collected from four separate sites and show contact metamorphic assemblages equivalent to a range from greenschist to lower amphibolite facies, yet they preserve near-chondritic Th/U ratios of 3–4. Mobile elements such as Sr show rather wide variations within individual units (Fig. 3) and interpretations based on such elements (e.g. Sr, Ba, Na, K, Ca) are avoided.
|
|
Analytical procedures
Major and trace elements
Rock samples were crushed in a jaw crusher and then ground in a tungsten carbide ring mill. Major elements (Si, Ti, Al, Fe, Mn, Mg, Ca, Na, K, P and S) were determined at Geoscience Australia, in Canberra, by wavelength-dispersive X-ray fluorescence (XRF) on fused discs using methods similar to those of Norrish & Hutton (1969)
. Precision for these elements is better than ±1% of the reported values. Loss on ignition (LOI) was determined by gravimetry after combustion at 1100°C. FeO abundances were determined at Geoscience Australia by digestion and electrochemical titration using a modified technique based on Shapiro & Brannock (1962). Trace elements were determined at Geoscience Australia and at the Department of Geology, Queensland University. Concentrations measured at the two laboratories agree to within 10% for all elements. At Geoscience Australia, Ba, Cr, Cu, Ni, Sc, V, Zn and Zr were determined by wavelength-dispersive XRF on pressed pellets using methods similar to those described by Norrish & Chappell (1977), whereas Cs, Ga, Nb, Pb, Rb, Sr, Ta, Th, U, Y and the REE were analysed by inductively coupled plasma mass spectrometry (ICP-MS; Perkin Elmer ELAN 6000) using methods similar to those of Eggins et al. (1997). At the University of Queensland, 100 mg of sample was weighed into a polytetrafluoroethylene (PTFE) high-pressure bomb and dissolved in a mixture of HF and HCl at 180°C for 24 h. Following drying, the sample was re-fluxed with 1 ml of HNO3 three times before final dissolution in 2 ml of HNO3 and 4 ml of Milli-q® water. The solution was checked for insoluble fluorides before dilution to 1000 times. ICP-MS analyses were carried out using methods similar to those of Eggins et al. (1997).
Neodymium isotopes
Sm–Nd isotopic analyses (Table 2) were determined by isotope dilution at both the Research School of Earth Sciences, Australian National University (ANU), Canberra, and at VIEPS Radiogenic Isotope Laboratory, Department of Earth Sciences, La Trobe University, Victoria. Analytical techniques at Canberra follow the approach outlined by Sun et al. (1995) and Champion & Sheraton (1997), whereas those for La Trobe follow that reported by Waight et al. (2000). Reported 143Nd/144Nd ratios have been normalized to 146Nd/144Nd = 0·7219 for mass fractionation correction. Average 143Nd/144Nd values of the La Jolla and BCR standards measured at ANU are 0·511872 ± 2 (2
, n = 85), and 0·512653 ± 5 (2, n = 8), respectively. Errors are typically ±0·5 epsilon units. La Trobe data, originally reported normalized to 0·511860 for La Jolla, have been renormalized to the ANU La Jolla value (0·511872). Differences between samples analysed at both laboratories are less than half an epsilon unit.
|
Louden Volcanics, high-Si basalts and intrusive equivalents
Rocks of the Louden Volcanics have high SiO2 (51–56 wt %), Th and LREE concentrations (Table 1) and have been included within the class of Precambrian mafic rocks known as siliceous high-Mg basalt (Sun et al., 1989
).
Sampling of fine-grained and typical spinifex-textured Louden Volcanics for the current study revealed no rocks with MgO >9·0 wt %, although dyke rocks with well-developed olivine-cumulate textures have MgO up to 18 wt % (R. H. Smithies & D. C. Champion, unpublished data). Approximately one-third of the samples studied by Arndt et al. (2001) have MgO >10·0 wt %, but many of these appear to be from cumulate layers (and dykes?) including all samples with MgO >13 wt %.
The Mount Negri Volcanics and the south Mallina basalts have lower MgO contents and Mg-number than the Louden Volcanics and higher Th, Zr and LREE concentrations (Table 1). These rocks share moderate to low TiO2 (0·39–0·66 wt %) and high Al2O3/TiO2 (22·5–48·6), and have silica values between 53 and 59 wt %, comparable with modern continental flood basalts (Arndt et al., 2001).
On primitive mantle normalized trace-element diagrams (Fig. 4), the Louden Volcanics, LREE-rich gabbro and the south Mallina basalts show virtually identical patterns that include significant enrichments in Th, Zr and LREE, with (Th/Gd)PM 9–11 and flat normalized heavy REE (HREE) patterns (Fig. 4). The Mount Negri Volcanics have more fractionated HREE concentrations [(Gd/Lu)PM 1·8 compared with 1·1 for the Louden Volcanics] but have virtually identical normalized trace-element patterns to the other basaltic rocks for elements between Th and Gd (Fig. 4). On variation diagrams involving LREE and Zr, the Louden Volcanics, south Mallina basalts, LREE-rich gabbros and the Mount Negri Volcanics show a single linear trend with a constant slope that projects back to the origin (Fig. 5). Slightly poorer correlations are produced when these trace elements are plotted against Nb and Yb.
|
|
Despite the similar trace-element patterns, absolute abundances of Th to Lu in the south Mallina basalts are 2–3 times larger than those in both the Louden Volcanics and LREE-rich gabbros (Fig. 4). There are also very subtle but persistent differences in such ratios as Nb/Th, Th/U, Th/Nd, and Eu/Eu* (Table 1). The south Mallina basalts show lower Ti/Zr and Al2O3/TiO2 (Table 1), and along with some of the LREE-rich gabbros, have lower Cr than the Louden Volcanics.
Boninite-like rocks
Rocks with compositional similarities to Phanerozoic boninites are extremely rare in Archaean successions, but have also been reported from the c. 2·7 Ga Abitibi Subprovince in Canada (Kerrich et al., 1998), and the c. 3·7 Ga Isua greenstone belt of Greenland (Polat et al., 2002). The examples from the Mallina Basin (Smithies, 2002) have SiO2 between 52 and 54 wt %, Mg-number between 65 and 69, TiO2 <0·3 wt %, and high Al2O3/TiO2 (62–74) and CaO/TiO2 (29–31) (Table 1). The concentrations of Th, Zr and LREE are lower than in the Louden Volcanics, high-Si basalts and LREE-rich gabbro, but are still strongly enriched (Fig. 4). Strongly fractionated LREE [(La/Gd)PM 4·5] and HREE [(Gd/Yb)PM 0·58] result in prominent ‘U-shaped’ normalized patterns, which are also a characteristic of Phanerozoic boninites (Fig. 6). Because of these close compositional similarities to Phanerozoic boninites, we refer to these rocks as boninite-like rocks. Notably, on variation diagrams for LREE and Zr (Fig. 5), the boninite-like rocks fall on the same linear trends defined by rocks of the Louden Volcanics, south Mallina basalts, LREE-rich gabbros and the Mount Negri Volcanics. In the following discussion these rocks, including the boninite-like rocks, will be collectively referred to as the ‘Mallina mafic suite’.
|
High-Mg diorite suite
Archaean granitoids with compositions similar to those of the high-Mg diorite suite have elsewhere been grouped into the ‘Archaean sanukitoid suite’. This term was first used by Shirey & Hanson (1984)
to describe a suite of dioritic to granodioritic rocks, the more mafic of which (60% SiO2 or less) have MgO >6 wt %, Mg-number >60, Ni and Cr >100 ppm, coupled with high LILE and Th, low HFSE (Fig. 4) and high Na2O, K2O, LREE and LREE/HREE [e.g. (La/Yb)PM 15–55] (Table 1). Archaean sanukitoids are now recognized as a widespread, but minor (probably <<5% of Archaean granitic rocks) component of many Archaean terrains younger than 2·8 Ga (Shirey & Hanson, 1984; Stern & Hanson, 1991; Evans & Hanson, 1997).
The only older documented examples are those from the Mallina Basin (Smithies & Champion, 2000). These rocks show a wide range of La/Nb ratios (3·1–7·5) compared with the Mallina mafic suite and also have variably higher La/Zr and La/Sm ratios (Fig. 5).
Nd-isotopic compositions
Neodymium isotopic data for all major igneous rock types within the Mallina Basin are presented in Table 2. The isotopic compositions of the Louden Volcanics and Mount Negri Volcanics are very similar, with initial Nd (at 2·95 Ga) between –1·5 and –2·8. The data for two samples of the boninite-like rocks overlap this range but extend to more primitive values (–0·9 and –1·7), as do the data for the south Mallina basalts (–0·6 to –1·0) and the LREE-rich gabbros (–0·5 and –1·5). These data, together with TDM model ages older than 3·2 Ga, clearly indicate the presence of an old (LREE-enriched) ‘crustal’ component within the rocks. The high-Mg diorite suite is consistently and significantly more radiogenic than the Mallina mafic suite, with Nd values between –0·1 and +0·8. One sample of the felsic alkali granites of the Portree Granitoid Complex has an Nd value of +1·1.
|
REGIONAL AND LOCAL Nd-ISOTOPIC RELATIONSHIPS |
|---|
The isotopic composition of the Mallina mafic suite varies within a range of
2 epsilon units (Nd(2·95 Ga) between –0·6 and –2·8; Fig. 7) over a wide geographical region and over a wide range in trace-element compositions that includes variations in Nd concentration from 3 to 27 ppm. Volcanic rocks of the c. 3·12 Ga Whundo Group, and associated granites, which locally form the western basement to the Mallina Basin, have more radiogenic isotopic compositions, with Nd(2·95 Ga) ranging from –0·4 to +1·8. The oldest (c. 3·1 Ga) phase of the Caines Well Granitoid Complex (Fig. 1) is a tonalite with an Nd(2·95 Ga) of +1·0 and a TDM model age of c. 3·1 Ga (Table 2), and is isotopically indistinguishable from both the c. 2·945 Ga alkaline Portree Granite and the Whundo Group. Similarly, 2·93 Ga to 2·85 Ga granites within and immediately east of the Mallina Basin also show an isotopic range (Nd(2·95 Ga) = –0·2 to +3·4) similar to that of the Whundo Group. The Whundo Group has the highest Nd values of any supracrustal sequence known from the Pilbara Craton. Champion & Smithies (1998) have argued that rocks of the Whundo Group, or isotopically similar but as yet unsampled crust, underlie and extend to the east of the Mallina Basin to form the source of these c. 3·1–2·93 Ga granites. Also clear from these data is that the non-radiogenic signatures of the Mallina mafic suite could not have been derived by incorporation of ‘Whundo-like’ lower crust alone—an older and less radiogenic crustal input is required (Fig. 7).
|
The TDM model ages of the Mallina mafic suite suggest a source component older than c. 3·3 Ga. Although no crust older than c. 3·27 Ga is known from the central or western Pilbara Craton, granites of that age have TDM model ages of c. 3·4 Ga and older, and
Nd(2·95 Ga) values between –2·3 and –4·0 (Fig. 7). Another source of non-radiogenic Nd is from the rocks of the older (c. 2·85–3·52 Ga; Van Kranendonk et al., 2002) eastern Pilbara Craton. Rocks from this region that both are older than 2·95 Ga and have TDM model ages >3·3 Ga, have Nd(2·95 Ga) values between –2·3 and –8·5 (median of 26 analyses = –5·2; Table 2) coupled with Nd contents largely between 15 and 50 ppm. Thus the most radiogenic samples from both east Pilbara crust and older parts of the west Pilbara crust (here collectively referred to as ‘old Pilbara crust’) slightly overlap, or are within error of, the low Nd end of the range for the Mallina mafic suite. |
CRUSTAL OR SUB-CRUSTAL CONTAMINATION? |
|---|
The Nd-isotopic evidence for a significant crustal component within all 2·97–2·95 Ga igneous rocks of the Mallina Basin is consistent with the Th-, Zr- and LREE-rich nature of these rocks. For the high-Mg diorite suite and the boninite-like rocks, these compositional features have been attributed to a metasomatized mantle source (Smithies & Champion, 2000
; Smithies, 2002). In contrast, Arndt et al. (2001) have suggested that the petrogenesis of the Louden and Mount Negri Volcanics is similar to that of Phanerozoic continental flood basalts and that the LREE enrichments are the result of crustal assimilation. In this section, we investigate the importance of crustal contamination and mantle source enrichment to the compositional evolution of the Mallina magmas, and, in particular, in producing the very narrow range of incompatible trace-element ratios and Nd-isotopic compositions shown by the Mallina mafic suite.
Assimilation of crust—AFC?
If the Mallina magmas have been contaminated during ascent it was either by basin sediment or basement. The Cleaverville LREE-rich gabbros, however, have not intruded the Mallina Basin. They were emplaced into a distinctly separate tectonostratigraphic terrain yet have the same trace-element and Nd-isotopic composition as the Louden Volcanics. Additionally, the south Mallina basalts have significantly higher Th, HFSE and REE concentrations than sampled sedimentary rocks of the basin (Table 3), as do the Mount Negri Volcanics for some trace elements. This renders significant contamination by sedimentary material from the Mallina Basin as an unlikely mechanism for producing the enriched signature of the Mallina magmas. The possibility that basement rocks have been assimilated can be examined using the known compositional range for old Pilbara crust and for rocks of the Whundo Group. These two components effectively bracket the known compositional range of pre-3·0 Ga Pilbara crust, and probably also of pre-3·0 Ga Archaean crust in general.
|
On a plot of
Nd vs 1/Nd (Fig. 7), most rocks of the Mallina mafic suite lie on a broadly linear trend that lies close to the more radiogenic end of the field for old Pilbara crust. Simple mixing models show that contamination by basement comprising such crust might be a plausible explanation for this trend, although >30% to 45% input of old Pilbara-like crust containing 50 ppm Nd is required for the LREE-enriched Mount Negri Volcanics and south Mallina basalts, respectively, and similar values are obtained for other elements (e.g. La, Zr). If more realistic average crustal Nd concentrations of 25–30 ppm are considered, then more than 50% crustal input is required for the Mount Negri Volcanics.
A wide range of possible parental magmas (komatiite to basalt) and the heterogeneous composition of old Pilbara crust (Figs 7 and 8) ensure that assimilation–fractional crystallization (AFC) models can provide some plausible numeric solutions. For example, taking a parental melt with a trace-element composition between enriched mid-ocean ridge basalt (E-MORB; Sun & McDonough, 1989) and c. 3·4 Ga basalts from the eastern Pilbara Craton (Arndt et al., 2001), the AFC models of Aitcheson & Forrest (1994) can account for the incompatible trace-element compositions of the Mount Negri Volcanics by assimilation of 20–25% old Pilbara crust. However, we consider it unlikely that these models can explain the very narrow range of incompatible trace-element ratios observed within the regionally widespread Mallina mafic suite. Plots of La vs Zr and Sm for the Mallina mafic suite (Fig. 8) illustrate this problem. The rocks of the Mallina mafic suite plot within a very narrow array with a constant slope yielding a high La/Zr ratio of 0·15 and high La/Sm ratio of 5·3, very close to average Archaean crustal values (0·16 and 5·0, respectively; Taylor & McLennan, 1985). However, Taylor & McLennan (1985) argued that average Archaean crust largely represents a bimodal mixture of high La/Sm felsic rocks [tonalite–trondhjemite–granodiorite (TTGs), with an average La/Sm of 10; Martin, 1994] and mafic (–ultramafic) rocks. Thus, average Archaean felsic igneous rocks alone are an unlikely contaminant within the Mallina mafic suite. The plot of La vs Sm, in particular, shows that the linear trend for the Mallina mafic suite lies at the lower limit of the very wide range of La/Sm values from outcropping old Pilbara crust (>1000 analyses of largely felsic rocks). If individual members of the Mallina mafic suite are, indeed, related to each other by AFC-type processes then any viable crustal contaminant must not only have had a La/Sm ratio atypical of most of the exposed felsic rocks in the region but also a ratio that varied within a very narrow range (5·3) over the regional geographical extent of the suite. Further, if AFC processes involved partial melts of crust, rather than the less likely case of bulk crustal assimilation, then that crustal component (before melting) must have had a lower La/Sm (<5·3), well out of the field of outcropping old Pilbara crust, or of felsic Archaean crust in general. Thus, there is no indication that compositional evolution within or between the individual suites was controlled by varying degrees of fractional crystallization accompanied by assimilation of old Pilbara felsic crust.
|
Assimilation of crust—decoupled assimilation and fractional crystallization?
The possibility that the Mallina mafic suite results from early assimilation in a lower-crustal chamber followed by homogenization, prior to fractional crystallization without further assimilation, also needs to be evaluated.
The Louden Volcanics show petrographic evidence for early crystallization of olivine followed by orthopyroxene. As such, the LREE and Zr, which are highly incompatible during crystallization of both minerals, may have become strongly enriched over significant crystallization intervals, with little change in La/Sm and La/Zr ratios (Fig. 8). However, there are several problems with a suggestion that melts parental to the boninite-like rocks—the most primitive of the mafic suite—assimilated crust in a lower-crustal chamber prior to various degrees of fractionation dominated by olivine and orthopyroxene (perhaps in a series of higher-level chambers), to produce the other rocks of the mafic suite.
First, the narrow range of La/Sm and La/Zr ratios in the Mallina mafic suite requires that the parental magma, after contamination, was extremely homogeneous. However, Kerr et al. (1995) and Cadman et al. (2001) have made the point that such homogeneity is unlikely unless magmas are able to flow turbulently. This is certainly likely to be an issue where the crust that is sampled by a chamber is itself notably inhomogeneous. We can only infer that the lowest Pilbara crust is similarly heterogeneous to the exposed Pilbara crust; our Nd-isotopic data certainly support this inference.
Second, some members of the Louden Volcanics have twice the concentration of Ni and three times the Th and LREE concentrations of boninite-like rocks with similar Mg-number. Similarly, on plots of Mg-number vs Ce, Y and Zr (Fig. 9), for example, the south Mallina basalts and the Louden Volcanics overlap extensively in terms of Mg-number and show similar trends but at significantly different absolute trace-element concentrations. Third, La/Yb ratios of the Mount Negri Volcanics (12·5) and south Mallina basalts (8·3) are significantly higher than those of the Louden Volcanics (6·5) or the boninite-like rocks (3·5). Fractionation of olivine and orthopyroxene alone cannot account for this. Addition of clinopyroxene to the fractionating assemblage will produce magmas with higher La/Yb, but will increase La/Zr and significantly lower La/Sm (Fig. 8). Extensive removal of plagioclase may have the opposite affect but similar Eu/Eu* values (0·9) in the Louden and Mount Negri Volcanics show that this is not a factor here. Finally, the Mallina mafic suite is preserved (i.e. minimum original extent) over an area of 18 000 km2, possibly requiring a single chamber of unreasonably large dimensions, or that contamination of different parental magma pulses in separate chambers coincidentally produced uniform La/Sm and La/Zr ratios.
|
We cannot totally rule out the possibility that compositional trends in the Mallina mafic suite are a result of crustal assimilation; however, we do not favour such a model.
Mantle source enrichment?
Independent lines of evidence support a contaminated-mantle origin for the igneous rocks in the Mallina Basin.
The boninite-like rocks of the Mallina Basin have high Al2O3/TiO2 and CaO/TiO2 ratios and low (Gd/Yb)PM ratios, and were derived from a strongly refractory mantle source. Their LREE abundance patterns would have been significantly influenced by only minor contamination. These rocks have Yb/Ti ratios that are higher than those of most common magmas. The ratios would have had to have been even higher if the boninite-like magmas were contaminated during ascent, given the significantly lower Yb/Ti ratios of most local and regional Archaean felsic rocks. The only local compositionally plausible contaminant (in terms of Yb/Ti ratios) is HREE-rich TTG that crops out to the south of the Mallina Basin. However, this is very rare at present levels of exposure, and is invariably interleaved with more voluminous TTG that is a compositionally unsuitable contaminant (Smithies, 2002) (e.g. very low La/Th). Accordingly, it is most likely that LREE enrichments in the boninite-like rocks also reflect derivation from a LREE-enriched mantle source.
A mantle source component is required for the high-Mg diorite suite to explain the high Mg-number and high Cr and Ni concentrations, but even the most primitive diorites also show extreme enrichment in Th and LREE (Fig. 4) with no apparent correlation with Mg-number, Nd, Cr, Ni or SiO2 (Fig. 10). Plutons with higher Mg-number (>55) and Cr (>150 ppm) often have amongst the highest SiO2 (>66 wt %), K2O (up to 4 wt %) and Th (up to 33 ppm), suggesting that the enrichment within the suite was not simply the result of crystal fractionation or crustal contamination (Smithies & Champion, 2000). A significantly LILE-enriched mantle source is the most reasonable explanation for the composition of these rocks (Shirey & Hanson, 1984; Smithies & Champion, 2000).
|
, Jallagoonina; 
, Peawah West; 
, Peawah East; *, Mallindra; 
, Jones Well; x, Wallareenya; open star, Stock 1. Modified from Smithies & Champion (2000)The Nd-isotopic composition of the Mallina mafic suite appears to be largely controlled by a mixture of non-radiogenic crust, similar to old Pilbara crust (itself a mixed source), and variably depleted (non-chondritic) mantle (
Nd(2·95 Ga) > +1·0). Within these broad parameters, it is difficult to rule out a small crustal component similar to Whundo-like depleted crust (Nd(2·95 Ga) of –0·4 to +1·8), as suggested by relationships on plots of Nd vs 1/Nd, La vs Nb and La vs Sm (Figs 7 and 8). The involvement of both Whundo-type and old Pilbara-like crust increases the potential compositional heterogeneity of any proposed bulk crustal contaminant. The very narrow range of trace-element ratios shown by the Mallina mafic suite would indicate that whatever the required mix of crustal components, the bulk crustal contaminant was compositionally homogenized before it was incorporated into the magmas. If we assume that the Mallina mafic suite was derived from a contaminated-mantle source, then the required crustal components were probably incorporated into the mantle as compositionally homogenized subducted sediment. To assess such a model we must evaluate whether the required homogeneity is realistic in terms of the observed compositional range of Archaean sediments and modern subducting sedimentary piles, and whether such homogeneity can survive subduction and be imparted on the magmas of the Mallina mafic suite.
A subducted sediment model
According to Taylor & McLennan (1985) and Vroon et al. (1995), turbidites form good average samples of old upper-crustal material. Vroon et al. (1995) analysed sediments representative of those subducting beneath the Banda Arc, in east Indonesia. Although they identified four discrete provenance areas based on isotopic variations, trace-element ratios such as La/Sm and La/Zr showed very little variation. Thus, the modern record shows that compositional homogeneity in subducting sedimentary piles, in terms of some trace-element ratios, is possible at least at the scale of a single arc. For some trace-element ratios (e.g. La/Sm) this is a general feature of sediments entering modern subduction zones (Fig. 8), although La/Zr ratios appear to fall along two discrete trends. Archaean shales, including five samples from the Mallina Basin, show even less variation in La/Sm and La/Zr (Fig. 8). One possible reason for this might be that early Archaean crust showed less compositional diversity than does modern crust. It might be expected, for example, that early Archaean sea-floor or marginal accumulations lacked carbonate and biological additions, as suggested by the dearth of such rocks in early Archaean successions. In addition, chemical sediments such as Archaean chert typically have very low trace-element abundances (Kato et al., 1998), and at least some of these rocks are simply silicified shales (e.g. Van Kranendonk et al., 2002). Thus, the suggestion that sediments subducted into the Archaean mantle beneath the NW Pilbara Craton were homogeneous in terms of critical trace-element ratios appears reasonable. Vroon et al. (1995) attributed the isotopic variation found within the Banda Arc sediments to large variations in the age of the four hinterlands, from Archaean to Palaeozoic. By contrast, sediment subducted at c. 3·0 Ga must have been overwhelmingly dominated by material with a much smaller age range (400 Myr), and so greater isotopic homogeneity should be expected in an Archaean subducting sediment pile.
Even given a compositionally homogeneous subducting sediment pile, subduction offers numerous mechanisms through which heterogeneity can be introduced into the mantle (e.g. variation in degrees of subduction dehydration and/or melting, and in subsequent interaction with peridotite). Thus, compositionally homogeneous subducted components do not necessarily lead to homogeneous mantle metasomatism. Nevertheless, Vroon et al. (1995) found that many trace-element characteristics of the Banda Arc volcanics are remarkably similar to those of the subducting sediments. Cousens et al. (2001) found that Proterozoic mantle-derived magmas emplaced across more than 240000 km2 of the NW Canadian Shield had enriched incompatible element concentrations and constant trace-element and isotopic compositions, with no systematic variation between Nd and SiO2 that could be linked to crustal contamination. Those workers inferred an extensive mantle reservoir that was homogeneously metasomatized during an episode of Archaean subduction. They speculated that flat-subduction, a style of subduction favoured by elevated geotherms in the Archaean (e.g. Abbott et al., 1994), provides a mechanism allowing metasomatic components to infiltrate a broad area of mantle. These cases suggest that in some circumstances, the subduction-modified source for magmas might be compositionally homogeneous in terms of trace-element and isotopic ratios, and over areas far greater than those inferred for the Mallina mafic suite. We propose that the enriched trace-element features of the Mallina mafic suite reflect a mantle source that was homogeneously contaminated by a sediment-dominated subduction component.
|
COMPOSITIONAL VARIATION IN THE MALLINA MAFIC SUITE |
|---|
An interesting feature of the Mallina mafic suite is that whereas La/Sm and La/Zr ratios show very little variation, concentrations of LREE, Zr, Th, etc. vary significantly (5–30 ppm for La). If crustal assimilation had little effect on the composition of these rocks then these variations must be explained in terms of variations in source enrichment, partial melting and subsequent crystal fractionation.
It is widely believed that the incompatible element budget in a subduction-modified mantle is controlled by the subducted component, via mantle metasomatism (e.g. Kay, 1980; Plank & Langmuir, 1993). This is particularly the case if the source regions had undergone prior melt extraction. Low TiO2 concentrations, high Al2O3/TiO2, [Gd/Lu]PM 1 and [Dy/Lu]PM <1 provide evidence for a variably refractory source for the south Mallina basalt and the Louden Volcanics. This may also be suggested by low [Ti/Yb]PM in the south Mallina basalts (0·45, compared with 0·3–0·4 for the boninite-like rocks) and Louden Volcanics (0·7–0·8). Significantly lower [Gd/Lu]PM and higher Al2O3/TiO2 at similar Mg-number for the boninite-like rocks (Table 1) suggests a strongly refractory source. Thus, it is likely that the concentrations of most incompatible trace elements in the source for the Mallina mafic suite, before metasomatism, were negligible compared with any additions made during subduction-related metasomatism. Exceptions here might be trace elements such as Nb and Yb that show poorer correlations with La (Fig. 5) and were probably in very low concentration in the metasomatizing agent. Provided the metasomatic component was homogeneous over the scale required to produce the source for the Mallina mafic suite, then reasonably small variations in the amount of metasomatic enrichment will be reflected in large changes in concentrations at constant ratios.
Because the LREE and Zr are highly incompatible in olivine and orthopyroxene (e.g. Suhr et al., 1998), progressive partial melting will form melts that vary in concentrations of these trace elements but not in La/Sm and La/Zr ratios, provided these minerals greatly dominate the residual source assemblage. Early crystallizing olivine and orthopyroxene and typically late crystallizing clinopyroxene, in the Louden Volcanics, is possibly consistent with a clinopyroxene-poor harzburgitic source. Low Gd/Yb in the Mallina mafic suite suggests that garnet was not a residual phase, and the anhydrous primary mineralogy of the rocks suggests that if the metasomatized mantle source contained amphibole, then like clinopyroxene (and possibly garnet), it was consumed during melting.
If we assume that the boninite-like rocks resulted from a refractory source, subsequently weakly metasomatized prior to 20% partial (batch) melting, leaving a harzburgite residue, then doubling the extent of metasomatic source enrichment and reducing the degree of partial melting to 10% (to produce the more evolved high-Si basaltic rocks) will alone account for 65% of variation in La concentration while maintaining constant La/Sm ratio. Given that distribution coefficients for La and Sm in both olivine and orthopyroxene are very small (<0·03; e.g. Suhr et al., 1998), these calculations are insensitive to variations in the proportions of residual olivine and orthopyroxene.
Nevertheless, it is likely that crystal fractionation, dominated by olivine and orthopyroxene (Fig. 8), resulted in at least some variations in absolute concentrations of all incompatible trace elements (including Nb and HREE) within the various magmas of the mafic suite.
|
DISCUSSION |
|---|
In Archaean terrains where geological relationships are commonly obscured, a temptation is to infer a modern-style subduction setting based largely, or wholly, upon the compositions of igneous rocks and on the recognition that compositionally similar rocks have formed at modern convergent margins. Barley (1986)
showed some of the dangers of this approach, which emphasizes the many superficial similarities between Archaean and modern plate tectonics, but pays little attention to the equally numerous differences [see review of Archaean tectonics by Bleeker (2002)]. High LREE/HFSE and LILE/HFSE ratios in igneous rocks can be interpreted in terms of a mantle source that was contaminated (or enriched) by crust, or of a mantle-derived melt that assimilated crust during emplacement. The distinction between the two processes is vital in interpreting the tectonic setting of magmatism, but in the absence of additional evidence, a relationship to modern-style subduction process cannot simply be assumed. Of particular interest to Archaean geology is the identification of the stage when enriched mantle source regions began contributing to crustal evolution—such a point might be when modern-style subduction became a viable process. There are very few examples where such a source has been clearly identified for rocks older than c. 2·8 Ga. The c. 2·97–2·95 Ga magmatic history of the Mallina Basin includes three groups of LREE-rich basaltic rocks, LREE-rich gabbro, rocks with boninite-like compositions and sanukitoids. This magmatism was a result of melting mantle source regions, but it also shows crustal signatures including high Th, Zr and LREE concentrations and high Th/HFSE and LREE/HFSE ratios.
Arndt et al. (2001) pointed out the compositional similarities between rocks of the Mallina mafic suite and continental flood basalt suites, including rocks of the c. 2·7 Ga Fortescue Group, which overlie the mafic suite. The crustal signature in many continental flood basalts has been attributed to crustal assimilation (Mensing et al., 1984; Arndt et al., 1993, 2001). However, enriched mantle source regions are commonly also invoked (e.g. Hooper & Hawkesworth, 1993), and Hergt & Brauns (2001) and Riley et al. (2003) have provided convincing evidence that the voluminous Mesozoic Tasmanian dolerites (Ferrar large igneous province) and the Karoo large igneous province come from such source regions. Studies of the c. 2·7 Ga Fortescue Group provide no consensus on the origin of their crustal signature (Nelson et al., 1992; compare Arndt et al., 2001). Interestingly, however, Nelson et al. (1992) suggested that if crustal assimilation was not a factor, then a 3078 ± 107 Ma isochron may date a Sm/Nd fractionation event in the subcontinental lithosphere beneath the Pilbara Craton. This would be consistent with a subduction event at either c. 3·01 Ga (Pike, 2001) or c. 3·12 Ga (Smith et al., 1998).
We cannot totally rule out a role for assimilation in developing crustal signatures in the igneous rocks within the Mallina Basin and our preferred model, an enriched mantle source, itself relies on limited evidence that metasomatism of the mantle may result in a compositionally homogeneous source. However, the balance of evidence presented here favours a metasomatized mantle source for these signatures. A future test of this interpretation might come from a study of mobile LILE in melt inclusions in early formed olivine, or through an estimation of the initial 187Os/188Os isotopic composition in chromite separates (e.g. Riley et al., 2003).
If our interpretation is correct, then by c. 2·95 Ga two distinct enriched mantle sources had developed beneath the central part of the Pilbara Craton. The combined geographical extent of the Mallina mafic suite is at least 180 km in an east–west direction and the range for the high-Mg diorite suite is at least 150 km, and the two suites show complete geographical overlap. The entire magmatic period probably occurred within a narrow interval between c. 2·955 Ga and c. 2·945 Ga. Consequently, the respective mantle source regions almost certainly coexisted and were not widely separated spatially.
We suggest that c. 3·12 Ga Whundo-like mafic crust and homogeneous sediment derived from old (>3·3 Ga) Archaean terrains was subducted to the SE and that partial melts derived from the subducted sediments infiltrated the mantle wedge. This probably occurred at c. 3·01 Ga, the depositional age of the Whim Creek Group, but might be as old as c. 3·12 Ga. Mantle in that region was refractory harzburgite, having previously yielded mafic and ultramafic magmas including those that formed the Whundo Group.
Resulting metasomatism was homogeneous with regard to ratios involving LREE and Zr, at least over an area large enough to provide the source that later produced the Mallina mafic suite. Because relative Nb depletions in the Mallina mafic suite are not matched by relative depletions in other fluid-immobile elements such as Th and Zr (Fig. 4), the metasomatic medium was probably a partial melt rather than a fluid.
Compared with the source for the Mallina mafic suite, contamination of the source for the high-Mg diorite suite was by a component with a considerably lower ratio of old Pilbara-like crust to basaltic Whundo-like crust. Smithies & Champion (2000) suggested that the mantle source for the high-Mg diorite suite was metasomatized by addition of up to 40% of slab-melt prior to magma genesis. This would explain the more radiogenic Nd-isotopic compositions and the higher LREE concentrations and LREE/HREE ratios of the high-Mg diorite suite, particularly if melting of the slab occurred at pressures high enough to stabilize garnet. Selected trace-element variations also show clear compositional differences between the respective source regions of the high-Mg diorite suite and the Mallina mafic suite (Fig. 5). We suggest that at greater depths, within the stability field of garnet, mafic Whundo-like crust itself partially melted to produce small volumes of adakitic magmas. These continuously interacted with mantle material (e.g. Kelemen, 1995; Rapp et al., 1999) as they ascended, but because of low melt/wall-rock ratios, froze (e.g. Rapp et al., 1999) close to the base of the lithosphere, where they provided the source for subsequent high-Mg diorite suite magmatism.
The variably metasomatized sub-Mallina mantle remained essentially inert for between 50 and 160 Myr when, at c. 2·955 Ga, it partially melted in a non-subduction setting. Smithies & Champion (2000) have attributed the late tectonothermal anomaly that caused this magmatism to either a plume or active rifting of the Mallina Basin. The observation that slightly later (c. 2·94–2·93 Ga) monzogranitic magmatism becomes less voluminous and younger away from the Mallina Basin indicates that this tectonothermal anomaly was centred on the Mallina Basin. The involvement of a mantle plume is, to some extent, supported by the work of Oversby (1976), who indicated that peak regional metamorphism occurred throughout the Pilbara Craton at c. 2·95 Ga, and also by a c. 2·9 Ga Re–Os isochron age for komatiitic rocks in the western part of the Pilbara Craton (Meisel et al., 2001). Alternatively, late removal, or break-off, of the fossil subducted slab at c. 2·95 Ga may have caused substantial asthenospheric upwelling, leading to partial melting of the previously metasomatized mantle.
|
SUPPLEMENTARY DATA |
|---|
Supplementary data for this paper are available on Journal of Petrology online.
|
ACKNOWLEDGEMENTS |
|---|
This study has benefited greatly through discussions with, and suggestions from, Arthur Hickman, Paul Morris, Steve Sheppard, Ian Tyler, and Martin Van Kranendonk. Journal reviews by Nick Arndt, Mark Barley, Ali Polat and an anonymous reviewer helped to greatly improve the manuscript. Lisa Cosgrove is thanked for drafting the figures. Published with the permission of the Director, Geological Survey of Western Australia and the Chief Executive Officer, Geoscience Australia.
|
FOOTNOTES |
|---|
* Corresponding author. E-mail: hugh.smithies{at}doir.wa.gov.au
|
REFERENCES |
|---|
Abbott, D. H., Drury, R. & Smith, W. H. F. (1994). A flat to steep transition in subduction style. Geology 22, 937–940.
Aitcheson, S. J. & Forrest, A. H. (1994). Quantification of crustal contamination in open magmatic systems. Journal of Petrology 35, 461–488.
Arndt, N. T. & Jenner, G. A. (1986). Crustally contaminated komatiites and basalts from Kambalda, Western Australia. Chemical Geology 56, 229–255.[CrossRef][Web of Science]
Arndt, N. T., Czamanske, G. K., Wooden, J. L. & Fedorenko, V. A. (1993). Mantle and crustal contributions to continental flood volcanism. Tectonophysics 223, 39–52.[CrossRef][Web of Science]
Arndt, N., Bruzak, G. & Reischmann, T. (2001). The oldest continental and oceanic plateaus: geochemistry of basalts and komatiites of the Pilbara Craton, Australia. Geological Society of America, Special Paper 352, 359–387.
Barley, M. E. (1986). Incompatible-element enrichment in Archaean basalts: a consequence of contamination by older sialic crust rather than mantle heterogeneity. Geology 14, 947–950.
Barley, M. E., Sylvester, G. C. & Groves, D. I. (1984). Archaean calc-alkaline volcanism in the Pilbara Block, Western Australia. Precambrian Research 24, 285–319.[CrossRef][Web of Science]
Bickle, M. J., Bettenay, L. F., Chapman, H. J., Groves, D. I., McNaughton, N. J., Campbell, I. H. & deLaeter, J. R. (1989). The age and origin of the younger plutons of the Shaw Batholith in the Archaean Pilbara Block, western Australia. Contributions to Mineralogy and Petrology 101, 361–376.[CrossRef][Web of Science]
Bickle, M. J., Bettenay, L. F., Chapman, H. J., Groves, D. I., McNaughton, N. J., Campbell, I. H. & deLaeter, J. R. (1993). Origin of the 3500–3300 Ma calc-alkaline rocks in the Pilbara Archaean: isotopic and geochemical constraints from the Shaw batholith. Precambrian Research 60, 117–149.[CrossRef][Web of Science]
Bleeker, W. (2002). Archaean tectonics: a review, with illustrations from the Slave craton. In: Fowler, C. M. R., Ebinger, C. J. & Hawkesworth, C. J. (eds) The Early Earth: Physical, Chemical and Biological Development. Geological Society, London, Special Publications 199, 151–181.
Brauhart, C. W & Morant, P. (2000). Discussion on ‘Geochemistry and geodynamic setting of volcanic and plutonic rocks associated with early Archean volcanogenic massive sulfide mineralisation, Pilbara Craton’ by Vearncombe, S. E., Kerrich, R., 1999. Precambrian Research 98, 243–270. Precambrian Research 104, 95–99.[CrossRef][Web of Science]
Cadman, A. C., Tarney, J., Bridgwater, D., Mengel, F., Whitehouse, M. J. & Windley, B. F. (2001). The petrogenesis of the Kangamiut dyke swarm, W. Greenland. Precambrian Research 105, 183–203.
Cameron, W. E., Nisbet, E. G. & Dietrich, V. J. (1979). Boninites, komatiites and ophiolitic basalts. Nature 280, 550–553.[CrossRef]
Champion, D. C. & Sheraton, J. (1997). Geochemistry and Nd isotope systematics of Archaean granites of the Eastern Goldfields, Yilgarn Craton, Australia: implications for crustal growth processes. Precambrian Research 83, 109–132.[CrossRef][Web of Science]
Champion, D. C. & Smithies, H. (1998). Archaean granites of the Yilgarn and Pilbara Cratons. The Bruce Chappell Symposium: Granites, Island Arcs, the Mantle and Ore Deposits, Abstract Volume. AGSO, Record 1998/33, 24–25.
Cousens, B. L., Aspler, L. B., Chiarenzelli, J. R., Donaldson, J. A., Sandeman, H., Peterson, T. D. & LeCheminant, A. N. (2001). Enriched Archean lithospheric mantle beneath western Churchill Province tapped during Paleoproterozoic orogenesis. Geology 29(9), 827–830.
Crawford, A. J., Falloon, T. J. & Green, D. H. (1989). Classification, petrogenesis and tectonic setting of boninites. In: Crawford, A. J. (ed.) Boninites. London: Unwin–Hyman, pp. 1–49.
Cullers, R. L., DiMarco, M. J., Lowe, D. R. & Stone, J. (1993). Geochemistry of a silicified, felsic volcaniclastic suite from the early Archaean Panorama Formation, Pilbara Block, Western Australia: an evaluation of depositional and post-depositional processes with a special emphasis on the rare-earth elements. Precambrian Research 60, 99–116.[CrossRef][Web of Science]
de Wit, M. J. (1998). On Archean granites, greenstones, cratons and tectonics: does the evidence demand a verdict? Precambrian Research 91, 181–226.[CrossRef][Web of Science]
Eggins, S. M., Woodhead, J. D., Kinsley, L. P. J., Mortimer, G. E., Sylvester, P., McCulloch, M. T., Hergt, J. M. & Handler, M. R. (1997). A simple method for the precise determination of >40 trace elements in geological samples by ICPMS using enriched isotope internal standardisation. Chemical Geology 134, 311–326.[CrossRef][Web of Science]
Evans, O. C. & Hanson, G. N. (1997). Late- to post-kinematic Archean granitoids of the S.W. Superior Province: derivation through direct mantle melting. In: de Wit, M. J. & Ashwal, L. D. (eds) Greenstone Belts. Oxford: Oxford University Press, pp. 280–295.
Green, M. G., Sylvester, P. J. & Buick, R. (2000). Growth and recycling of early Archaean continental crust: geochemical evidence from the Coonterunah and Warrawoona Groups, Pilbara Craton, Australia. Tectonophysics 322, 69–88.[CrossRef][Web of Science]
Hergt, J. M. & Brauns, C. M. (2001). On the origin of Tasmanian dolerites. Australian Journal of Earth Sciences 48, 543–549.[CrossRef][Web of Science]
Hickman, A. H. (1997). A revision of the stratigraphy of Archaean greenstone successions in the Roebourne–Whundo area—west Pilbara. Western Australia Geological Survey, Annual Review 1996–97, 76–81.
Hooper, P. R. & Hawkesworth, C. J. (1993). Isotopic and geochemical constraints on the origin and evolution of the Columbia River Basalts. Journal of Petrology 34, 1203–1246.
Kato, Y., Ohta, I., Tsunematsu, T., Watanabe, Y., Isozaki, Y., Maruyama, S. & Imai, N. (1998). Rare earth element variations in mid-Archean banded iron formations: implications for the chemistry of ocean and continent and plate tectonics. Geochimica et Cosmochimica Acta 21/22, 3475–3497.
Kay, R. W. (1980). Volcanic arc magmas: implications of a melting–mixing model for element recycling in the crust–upper mantle system. Journal of Geology 88, 497–522.[Web of Science]
Kelemen, P. B. (1995). Genesis of high Mg andesites and the continental crust. Contributions to Minerology and Petrology 120, 1–19.[Web of Science]
Kerr, A. C., Kempton, P. D. & Thompson, R. N. (1995). Crustal assimilation during turbulent magma ascent (ATA); new isotopic evidence from the Mull Tertiary lava succession, N.W. Scotland. Contributions to Mineralogy and Petrology 119, 142–154.[Web of Science]
Kerrich, R., Wyman, D., Fan, J. & Bleeker, W. (1998). Boninite series: low-Ti tholeiite associations from the 2·7 Ga Abitibi greenstone belt. Earth and Planetary Science Letters 164, 303–316.[CrossRef][Web of Science]
Kusky, T. M. & Polat, A. (1999). Growth of granite–greenstone terranes at convergent margins, and stabilization of Archaean cratons. Tectonophysics 305, 45–73.
Martin, H. (1994). The Archean grey gneisses and the genesis of continental crust. In: Condie, K. C. (ed.) Archean Crustal Evolution. Amsterdam: Elsevier, pp. 205–259.
Martin, H. (1999). Adakitic magmas: modern analogues of Archaean granitoids. Lithos 46, 411–429.[CrossRef][Web of Science]
McCulloch, M. T. (1987). Sm–Nd isotopic constraints on the evolution of Precambrian crust in the Australian continent. In: Kroner, A. (ed.) Proterozoic Lithospheric Evolution. American Geophysical Union, Geodynamic Series 17, 115–130.
Meisel, T., Moser, J. & Wegscheider, W. (2001). Recognizing heterogeneous distribution of platinum group elements (PGE) in geological materials by means of the Re–Os isotope system. Analytical Chemistry 370, 566–572.
Mensing, T. M., Faure, G., Jones, L. M., Bowman, J. R. & Hoefs, J. (1984). Petrogenesis of the Kirkpatrick Basalt, Solo Nunatak, Northern Victoria Land, Antarctica, based on isotopic compositions of strontium, oxygen and sulfur. Contributions to Mineralogy and Petrology 87, 101–108.[CrossRef][Web of Science]
Moorbath, S. (1977). Ages, isotopes and evolution of Precambrian continental crust. Chemical Geology 20, 151–187.[CrossRef][Web of Science]
Nelson, D. R., Trendall, A. F., de Laeter, J. R., Grobler, N. J. & Fletcher, I. R. (1992). A comparative study of the geochemical and isotopic systematics of late Archaean flood basalts from the Pilbara and Kaapvaal Cratons. Precambrian Research 54, 231–256.[CrossRef][Web of Science]
Norrish, K. & Chappell, B. W. (1977). X-ray fluorescence spectrometry. In: Zussman, J. (ed.) Physical Methods in Determinative Mineralogy, 2nd edn. London: Academic Press, pp. 201–272.
Norrish, K. & Hutton, J. T. (1969). An accurate X-ray spectrographic method for the analysis of a wide range of geological samples. Geochimica et Cosmochimica Acta 33, 431–453.[CrossRef][Web of Science]
Oversby, V. M. (1976). Isotopic ages and geochemistry of Archaean acid igneous rocks from the Pilbara, Western Australia. Geochimica et Cosmochimica Acta 40, 817–829.[CrossRef][Web of Science]
Pearce, J. A. (1982). Trace element characteristics of lavas from destructive plate boundaries. In: Thorpe, R. S. (ed.) Andesites: Orogenic Andesites and Related Rocks. Chichester: Wiley, pp. 525–548.
Pearce, J. A. & Cann, J. R. (1993). Tectonic setting of basic volcanic rocks determined using trace element analyses. Earth and Planetary Science Letters 19, 290–300.
Peate, D. W., Pearce, J. A., Hawkesworth, C. J., Colley, H., Edwards, C. M. H. & Hirose, K. (1997). Geochemical variations in Vanuatu Arc lavas: the role of subducted material and a variable mantle wedge composition. Journal of Petrology 38, 1331–1358.[CrossRef][Web of Science]
Pike, G. (2001). The facies architecture of two contrasting volcano-sedimentary basin successions from the Archaean Whim Creek Belt, North Pilbara Terrain, Western Australia: the intra-continental arc-related Whim Creek Group and plume-related, continental rift-hosted Bookingarra Group. Ph.D thesis, Monash University, Clayton, Vic., 465 pp.
Pike, G. & Cas, R. A. F. (2002). Stratigraphic evolution of Archaean volcanic rock-dominated rift basins from the Whim Creek Belt, west Pilbara Craton, Western Australia. International Association of Sedimentologists, Special Publications 33, 213–234.
Plank, T. & Langmuir, C. H. (1993). Tracing trace elements from sediment input to volcanic output at subduction zones. Nature 362, 739–743.[CrossRef]
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]
Polat, A. & Kerrich, R. (2001). Magnesian andesites, Nb-enriched basalt-andesites, and adakites from late-Archean 2·7 Ga Wawa greenstone belts, Superior Province, Canada: implication for late Archean subduction zone petrogenetic processes. Contributions to Mineralogy and Petrology 141, 36–52.
Polat, A. & Kerrich, R. (2002). Nd-isotope systematics of 2·7 Ga adakites, magnesian andesites, and arc basalts, Superior Province: evidence for shallow crustal recycling at Archean subduction zones. Earth and Planetary Science Letters 202, 345–360.[CrossRef][Web of Science]
Polat, A., Hofmann, A. W. & Rosing, M. T. (2002). Boninite-like volcanic rocks in the 3·7–3·8 Ga Isua greenstone belt, West Greenland: geochemical evidence for intra-oceanic subduction zone processes in the early Earth. Chemical Geology 184, 231–254.[CrossRef][Web of Science]
Rapp, P. R., Shimizu, N., Norman, M. D. & Applegate, G. S. (1999). Reaction between slab-derived melts and peridotite in the mantle wedge: experimental constraints at 3·8 GPa. Chemical Geology 160, 335–356.[CrossRef][Web of Science]
Riley, T. R., Leat, P. T., Storey, B. C., Parkinson, I. J. & Millar, I. L. (2003). Ultramafic lamprophyres of the Ferrar large igneous province: evidence for a HIMU mantle component. Lithos 66, 63–76.[CrossRef][Web of Science]
Shapiro, L. & Brannock, W. W. (1962). Rapid analysis of silicate, carbonate and phosphate rocks. US Geological Survey Bulletin 1144-A.
Shirey, S. B. & Hanson, G. N. (1984). Mantle-derived Archaean monzodiorites and trachyandesites. Nature 310, 222–224.[CrossRef]
Smith, J. B., Barley, M. E., Groves, D. I., Krapez, B., McNaughton, N. J., Bickle, M. J. & Chapman, H. J. (1998). The Sholl Shear Zone, West Pilbara: evidence for a domain boundary structure from integrated tectonic analyses, SHRIMP U–Pb dating and isotopic and geochemical data of granitoids. Precambrian Research 88, 143–171.[CrossRef][Web of Science]
Smithies, R. H. (2000). The Archaean tonalite–trondhjemite– granodiorite (TTG) series is not an analogue of Cenozoic adakite. Earth and Planetary Science Letters 182, 115–125.[CrossRef][Web of Science]
Smithies, R. H. (2002). Archaean boninite-like rocks in an intracratonic setting. Earth and Planetary Science Letters 197, 19–34.[CrossRef][Web of Science]
Smithies, R. H. & Champion, D. C. (2000). The Archaean high-Mg diorite suite: links to tonalite–trondhjemite–granodiorite magmatism and implications for early Archaean crustal growth. Journal of Petrology 41, 1653–1671.
Smithies, R. H., Nelson, D. R. & Pike, G. (2001). Development of the Archaean Mallina Basin, Pilbara Craton, northwestern Australia; a study of detrital and inherited zircon ages. Sedimentary Geology 141–142, 79–94.
Stern, R. A. & Hanson, G. N. (1991). Archean high-Mg granodiorite: a derivative of light rare earth element-enriched monzodiorite of mantle origin. Journal of Petrology 32, 201–238.
Suhr, G., Seck, H. A., Shimizu, N., Gunther, D. & Jenner, G. (1998). Infiltration of refractory melts into the lowermost oceanic crust: evidence from dunite- and gabbro-hosted clinopyroxenes in the Bay of Islands ophiolites. Contributions to Mineralogy and Petrology 131, 136–154.[CrossRef][Web of Science]
Sun, S.-S. & McDonough, W. F. (1989). Chemical and isotopic systematics of oceanic basalts: implications for mantle compositions and processes. In: Saunders, A. D. & Norry, M. J. (eds) Magmatism in the Ocean Basins. Geological Society, London, Special Publications 42, 313–345.
Sun, S.-s. & Nesbitt, R. W. (1978). Geochemical regularities and genetic significance of ophiolitic basalts. Geology 6, 689–693.
Sun, S-.S., Nesbitt, R. W. & McCulloch, M. T. (1989). Geochemistry and petrogenesis of Archaean and early Proterozoic siliceous high-magnesian basalts. In: Crawford, A. J. (ed.) Boninites. London: Unwin–Hyman, pp. 149–173.
Sun, S-.s, Warren, R. G. & Shaw, R. D. (1995). Nd isotope study of granites from the Arunta Inlier, central Australia: constraints on geological models and limitation of the method. Precambrian Research 71, 301–314.[CrossRef][Web of Science]
Tarney, J. (1992). Geochemistry and significance of mafic dyke swarms in the Proterozoic. In: Kondie, K. C. (ed.) Proterozoic Crustal Evolution. Amsterdam: Elsevier, pp. 151–179.
Taylor, S. R. & McLennan, S. M. (1985). The Continental Crust: its Composition and Evolution. Oxford: Blackwell.
Van Kranendonk, M. J., Hickman, A. H., Smithies, R. H., Nelson, D. R. & Pike, G. (2002). Geology and tectonic evolution of the Archaean North Pilbara Terrain, Pilbara Craton, Western Australia. Economic Geology 97, 695–732.
Vroon, P. Z., van Bergen, M. J., Klaver, G. J. & White, W. M. (1995). Strontium, neodymium, and lead isotopic and trace-element signatures of the East Indonesian sediments: provenance and implications for Banda Arc magma genesis. Geochimica et Cosmochimica Acta 59, 2573–2598.[CrossRef][Web of Science]
Waight, T. E., Maas, R. & Nicholls, I. A. (2000). Fingerprinting feldspar phenocrysts using crystal isotopic composition stratigraphy: implications for crystal transfer and magma mingling in S-type granites. Contributions to Mineralogy and Petrology 139, 227–239.[CrossRef][Web of Science]
Wyman, D. A., Ayer, J. A. & Devaney, J. R. (2000). Niobium-enriched basalts from the Wabigoon subprovince, Canada: evidence for adakitic metasomatism above an Archean subduction zone. Earth and Planetary Science Letters 179, 21–30.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. SHIMIZU, E. NAKAMURA, and S. MARUYAMA The Geochemistry of Ultramafic to Mafic Volcanics from the Belingwe Greenstone Belt, Zimbabwe: Magmatism in an Archean Continental Large Igneous Province J. Petrology, November 1, 2005; 46(11): 2367 - 2394. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||