The Dharwar craton in South India, displays a natural cross section in the late-Archaean continental crust, from low grade greenschist facies in the North to high grade granulites to the South. It is intruded by a huge magmatic body emplaced at 2.52 Ga (the Closepet granite) which gives the unique opportunity to be continuously studied from its roots to the upper crustal levels. In the South, the root zone is heterogeneous and contains several plutonic facies ranging from poorly differentiated clinopyroxene-bearing monzonites to granites. Field and geochemical investigations recently performed in this area (Jayananda et al., 1995; Moyen et al., 1997) explain the genesis of the Closepet granite as follows: 1) a mantle-derived magma intruded the crust along a major crustal shear-zone, at depth it subsequently underwent limited fractional crystallisation (F > 0,9); 2) the hot magmatic intrusion induced partial melting of the surrounding 3.2 Ga-old gneissic crust, thus generating anatectic granites. There, both the "differentiated" mantle-derived magma and the crustal melt, mix together, thus giving rise to a wide compositional range of magmas. Our purpose, is to use a wide spectrum of independent chemical elements to build a synthetic model accounting for most of the observed geochemical variation in the Closepet granite. Our model can be summarized as follows: (1) The melting of the old gneissic crust can be adequately modelled in terms of equilibrium hydrous melting. (2) The source of the more mafic liquids is an enriched mantle whose mineralogy is recalculated. (3) The processes operating within the Closepet magmatic body (fractionnal crystallization and magmas mixing) are best described in term of multi-component mixing: felsic liquid, mafic liquid, and cumulate (incompletely removed from the remaining liquid). This allows for a simple mathematical formulation (linear equation), permitting a rapid calculation for every sample of the proportion of each single component involved in its formation. The resulting calculation not only have an internal coherence (low discrepancy between prediction and analytical data) but also are consistent with field and petrographic evidences.
Jayananda, M, Martin, H, Peucat, J-Jand Mahabaleswar, B, Contrib. Mineral. Petrol, 119, 314-329, (1995).
Moyen, J-F, Martin, Hand Jayananda, M, C. R. Acad. Sci. Paris, série II, 325, 659-664, (1997).
Numerical forward modeling of Lu-Hf systematics in the crust-mantle system through Earth history has been carried out using chondritic data (Blichert-Toft and Albarède, 1997) as a starting point, and a procedure identical to that used by Kramers and Tolstikhin (1997) for U-Th-Pb and Nägler and Kramers (1998) for Nd. The data for depleted mantle (176Hf/177Hf of c. 0.2832; Patchett, 1983) and continental crust (176Hf/177Hf ranging from 0.2810 to 0.2825; Vervoort and Patchett, 1996) can be reproduced in the framework of a two-reservoir mantle using reasonable partition coefficient values. Lu-Hf systematics of the depleted mantle are sensitive to crust-mantle recycling because the Lu/Hf ratio in the crust is less than 0.2x that of the mantle (compared to c. 0.5 for Sm/Nd) and Hf is strongly enriched in the upper crust (5-15 ppm). Lu/Hf contrasts between crust and mantle are further enhanced by erosion and fractionation of the Lu/Hf ratio in sediment transport (Patchett et al., 1984). The best fit crustal history scenario resulting from this modeling concurs with that obtained for U-Th-Pb and Sm-Nd. This entails rapid growth of the continental crust mass between c. 3.7 and 2 Ga (at which time about 60% of the present crust mass existed) followed by slower growth after 2 Ga. In particular, the modeling constrains the crust-to-mantle recycling flux in the Archean to at most 10% of the crustal growth rate. The slower net continental crust growth after 2 Ga is a result of increased recycling (presently about 50% of the crustal growth rate) rather than a slowing down of the continental crust formation process.
Two possible reasons why erosion of continental crust was less important in the Archean than at present could be, first, subdued topography as a result of more frequent lower crustal melting, and second, little or no land vegetation and hence slower rock weathering and soil formation. The likelihood of the first can be assessed in the light of evidence from Archean metamorphic provinces.
Blichert-Toft, J. & Albarède, F., Earth Planet. Sci. Lett., 148, 243-258, (1997).
Kramers, J.D. & Tolstikhin, I.N., Chem. Geol., 139, 75-110, (1997).
Naegler, FT & Kramers, JD, Prec. Res, 91, 233-252, (1998).
Patchett, PJ, Geochim. Cosmochim. Acta, 47, 81-91, (1983).
Patchett, PJ, White, WM, Feldmann, H, Kielinczuk, S & Hofmann, AW, Earth Planet. Sci. Lett, 69, 365-378, (1984).
Vervoort, J & Patchett, PJ, Geochim. Cosmochim. Acta, 60, 3717-3733, (1996).
Over the past decade the dominant model for the genesis of the magmas of the Archaean Tonalite-Trondhjemite-Granodiorite (TTG) suite has been that of the dehydration melting of a basaltic slab. Recently a number of observations have challenged this view. These include: the mismatch between the composition of experimentally derived slab melts and that of Archaean TTGs; recent geochemical studies of adakites, thought to be modern analogues of Archaean TTGs; arguments in favour of direct mantle melting to produce the diorites and high magnesian granodiorites also found in Archaean cratons.
Recent studies suggest that a multi-component melting model may better explain the origin of TTG melts. Berger & Rollinson (1997) showed that 2.7 Ga Archaean magmatic granulites in southern Zimbabwe have TTG affinities and are characterised by unusually high U/Pb ratios (model mu1=9.1), high Th/U (7.8), high initial 87Sr/86Sr (0.704) and old Nd-model ages (ca. 3.0 Ga). Their chemistry is interpreted as the mixing of an old crustal component with a younger mantle-derived component. Evidence from recent experimental studies by Rapp and Shimizu (1996) show that a process of hybridisation between a pristine melt of a basaltic parent and the mantle is also important in Archaean TTG genesis. Such a mixing process can explain the unusual composition of high-Mg REE-enriched diorites, frequently found as an early component of Archaean TTG suites. Taken together, these two studies show that the process of Archaean TTG genesis is more complex than has previously been recognised and that TTG compositions are better understood in terms of mixing processes in which slab melts mix with the mantle and/or older crust.
An alternative explanation of the multicomponent nature of Archaean TTGs comes from new studies of the Isua Greenstone belt in west Greenland. The Isua succession is a basalt-komatiite-chert association and is probably parental to the enclosing Amitsoq TTG gneisses. It is important to explore whether the partial melting of the integrated Isua succession would produce TTG melts with the same geochemical characteristics as those currently interpreted as the product of slab/mantle mixing.
In either case, the view that TTG melts are the product of mixing of several different components means that futures studies of TTGs will require their identification and an estimate made of their relative contribution to a particular TTG melt.
Berger M, Rollinson HR., Geochim. Cosmochim. Acta, 61, 4809-29, (1997).
Rapp RP, Shimizu N., J. Conf. Abs., 1, 497, (1996).
The Cordillera Blanca batholith, Peru is a metaluminous 'I'-type whose dominant high silica rocks (>70%) are Na-rich with many of the characteristics attributed to subducted oceanic slab melts. However, the position of the batholith and the age of the oceanic crust at the trench during the Miocene preclude slab melting, and the batholith rocks result instead from partial melting of newly underplated Miocene crust. In this dynamic model, basaltic material is melted to produce high Na, low HREE, high-Al trondhjemitic type melts with residues of garnet, clinopyroxene and amphibole. Such Na-rich magmas are characteristic of thick Andean crust (+ 60 km) and differ significantly from typical calc-alkaline, tonalite-granodiorite magmas formed in normal crust (ca. 30 km). 'Archean' signatures in Mesozoic (and younger) plutonic rocks from Antarctica, Chile, the Cascades and New Zealand, are also thought to reflect partial melting of basaltic lower crust and calls into question the traditional model of TTG formation in the Archean by slab melting. We propose a revised model of Archean TTG formation where rapid partial melting in magmatically thickened 'greenstone' crust produces partial melts of TTG composition similar to that proposed for the Mesozoic-recent TTG batholiths of Antarctica and the Andes. Transition from TTD to high-K tonalites and granodiorites does not require pre-existing evolved continental crust and can be achieved simply by melting mafic-intermediate K-rich rocks at source. Rapid melting of basaltic underplate by successive emplacement of numerous, thin (< 1 km thick) mafic sills overcomes the thermal 'problem' of melt generation at source. Emplacement of TTG magmas as tabular sheets (as opposed to diapiric blobs) at higher structural levels within greenstone basins ameliorates the space 'problem'.
There is a convincing enough correlation between the composition of many Archean granitoids, and liquids produced in laboratory experiments on natural basalts at 1-4 GPa (e.g., Sen and Dunn, 1994; Rapp and Watson, 1995), to conclude that "wet melting" of garnet-amphibolite and/or eclogite was an important element of early continent formation (i.e., cratonization). Like many early Archean TTG (Tonalite-trondhjemite-granodiorite), wet melts of basalt are sodium-rich, high-Al2O3 granitoids which have high La/Yb, Sr/Nd, and Sr/Y ratios. Although virtually identical to Archean TTG in these and most other geochemical aspects, the experimental liquids have Mg-numbers (where Mg# = mole% Mg/(Mg+Fe)) that range from 25 to 45, whereas Mg-numbers of early-Archean TTG range from 25-53. And late-Archean, high-Mg tonalites from the Slave Province (Feng and Kerrich, 1992) and members of the "Sanukitoid Suite" from the Superior Province (Stern and Hanson, 1991) have Mg-numbers in the range 45-63, clearly inconsistent with a simple "basalt melting" origin. In order to reconcile these discrepancies between the experimental "TTG" liquids and Archean TTG, we have performed a series of experiments in which "TTG" melts are allowed to react with and assimilate a small amount of natural peridotite (Rapp et al., in press), simulating melt-rock reaction in a "proto-continental" setting. Peridotite assimilation by TTG melts produces hybridized, tonalitic liquids with elevated Mg-numbers (53-56), but trace element ratios that are largely unaffected by melt-rock reaction (i.e., high Mg-number tonalite with TTG trace element characteristics). As the melt:rock ratio in the experiments is reduced (i.e., more peridotite) even further, hybridized liquids evolve to high-Mg andesites compositions comparable to "parental" sanukitoid magmas, with Mg-numbers of 68-71.
A full continuum of processes can thus be envisioned, from the generation of "pristine" TTG magmas by wet melting of garnet amphibolite/eclogite, to hybridization of these melts by assimilation of peridotite, and the consequent metasomatism of the sub-continental mantle. The Mg-number of Archean granitoids is perhaps the best indicator of TTG lineage, and the extent to which the mantle was involved in early continent formation. The evidence thus far suggests an increasingly important role in the late-Archean, with crustal, basaltic (oceanic?) sources dominating in the early-Archean, although the exact tectonic setting remains debatable.
Sen, C & Dunn, T, Contrib. Mineral. Petrol, 117, 394-409, (1994).
Rapp, RP & Watson, EB, Jour. Petrology, 36, 891-931, (1995).
Feng, R & Kerrich, R, Chemical Geology, 98, 23-70, (1992).
Stern, RA & Hanson, GN, Jour. Petrology, 32, 201-238, (1991).
Rapp, RP, Shimizu, N, Norman, M & Applegate, GS, Chemical Geology, in press, (1999).
Monzogranitic and granodioritic granitoids and gneisses of the Wedza and Chilimanzi suites form large intrusive complexes in the south-eastern part of the Zimbabwe craton, between the Mutare-Odzi greenstone belt and the Northern Marginal Zone of the Limpopo Belt. While the older units of these complexes (Wedza suite) are of syn-kinematic origin, the younger Chilimanzi suite was emplaced late- to post-kinematically. Internal differentiation of the major elements indicates similar paths of magmatic evolution for both suites. Comparison with the major element distribution of older granodioritic/tonalitic intrusives within the Mutare greenstone belt suggests an overall magmatic evolution from primitive, tonalitic towards monzogranitic compositions. REE distributions of the older tonalitic units decrease from 100x enrichment (with respect to C1-Chondrite) for the light to only 10x for the heavy elements with a slight concave bend. REE of the majority of Wedza suite granitoids trend from 300-400x for the light to 10-25x C1-Chondrite for the heavy elements, a negative Eu-anomaly is well developed in most samples. The slope of the REE distribution of Chilimanzi granitoids is distinctly steeper (100-400x to 2-7x C1-Chondrite) and a negative Eu anomaly is only weakly developed. In comparison, REE distributions of the Chilimanzi-equivalent enderbitic intrusives of the Northern Marginal Zone (Limpopo Belt) gently slope from 40-105x to 2-10x C1-Chondrite with a positive Eu-anomaly. Whole rock Sm/Nd Nd-model ages (2.90-3.2 Ga) do not allow the distinction of the Wedza and Chilimanzi suite rocks. The model ages mark a discrete and well established Archean crust forming event in Africa. The small observed variation may indicate minor contribution of juvenile crust. U/Pb isotopic multigrain analyses of distinctly zoned zircons revealed highly discordant ages of 2507+111/-97 Ma and 2585+36/-33 Ma for the Wedza and 2402+12/-11 Ma and 2448+350/-167 Ma for the Chilimanzi suite. These ages confirm within the errors the intrusion age of about 2.6 Ga for the Wedza suite. However, the age of 2.4 Ga for the posttectonic Chilimanzi suite conflicts with the timing of the Great Dyke emplacement into an already consolidated crust. The geochemical and radiometric investigations suggest a dynamic crust forming process initialised at c. 3.2-2.9 Ga with the formation of a crustal protolith. The final stage of this process is marked by emplacement of large volumes of monzogranitic/granodioritc material at 2.6 to possibly as late as 2.4 Ga. The detailed reconstruction of this Archean process, however, is vitally dependent on high resolution chronometry for which single grain or SHRIMP determinations are more suitable than multigrain zircon U/Pb analysis.
Newly acquired U-Pb single zircon data are presented which provide the oldest and most precise dates so far obtained from the Zimbabwe craton and indicate a much more widespread outcrop of Early Archaean lithologies. This has significant repercussions when considering the tectonics and greenstone belt genesis of the craton.
Granitoids of the same and greater age as the 3.46 Ga Tokwe granitoid gneisses in the south central portion of the Zimbabwe craton are recognised in the central Midlands region, strongly suggesting synchronous formation of the two areas. An undeformed and relatively unmetamorphosed leucosome of the Tokwe gneisses constrains the timing of deformation and metamorphism in this region at c.3.35 Ga. This coincides with published detrital zircon and granitoid data which indicate the abundance of this phase in the south-central region. Other zircon results (Dougherty-Page 1994) indicate c.3.2 Ga granitoids in the northern, Shamva region of Zimbabwe, a phase also recognised around Tokwe.
We conclude that a single Early Archaean continental crustal segment existed in Zimbabwe, extending for over 450 km from the south-central Tokwe region, through the Midlands and up to the northern Shamva region of the present day craton. This c.3.4 Ga crustal block, which we propose to be called the Sebakwe Proto-craton, stabilised around 3.35 Ga with a later granitoid emplacement event around 3.2 Ga and provided the basement for the 3.0-2.6 Ga Late Archaean granite-greenstone magmatism. This second stage of evolution therefore resulted in the present predominance of these rocks, and occurred c.400 Ma after the initial stabilisation of the craton.
The synchroneity and extent of the Sebakwe Proto-craton is considered strong evidence supporting a predominantly intra-cratonic origin for the Late Archaean greenstone belts of Zimbabwe and refuting an arc accretion origin for the craton.
(Dougherty-Page, JS, Unpublished PhD thesis, Open University, Milton Keynes. ), (1994).
A model which explained the high grade Limpopo Mobile Belt in Southern Africa as resulting from a single continental collision in the late Archean (e.g. van Reenen et al., 1987 and Treloar et al., 1992) has been invalidated by recent structural, petrological and geochronological studies (e.g. Holzer et al., 1998). These have shown that the main tectonometamorphism in its most important province, the Central Zone (CZ), resulted from a transpressive orogeny at 2.0 Ga, in which medium to high pressures were reached and the p-T path had a clockwise sense. Nevertheless this zone also shows relics of late Archean high grade metamorphism. Further, in the Northern and Southern Marginal Zones (NMZ and SMZ) which adjoin the Zimbabwe and Kaapvaal cratons, respectively, the last high grade metamorphic episodes were indeed late Archean. The relics in the CZ are characterised by an anticlockwise p-T-evolution, at c. 2.55 Ga (Holzer et al., 1998). In the NMZ repeated crustal remelting (Berger and Rollinson, 1997) and intrusion of charnoenderbitic magmas continued to 2.58 Ga, producing counterclockwise p-T paths (Kamber and Biino, 1995). In contrast the SMZ consists of medium to high pressure granulites, which underwent a clockwise p-T-evolution at c. 2.69 Ga, followed by decompression and isobaric cooling (van Reenen et al., 1987). In the Northern Kaapvaal Craton (Renosterkoppies Greenstone Belt, Pietersburg area) tectonism took place under amphibolite facies conditions at c. 2.75 Ga and can therefore not be related to any events in the Limpopo belt. Thus the different tectonic units have different late Archean tectonometamorphic histories. Trace element geochemistry as well as Pb + Nd isotope characteristics of the SMZ are similar to those of the Kaapvaal Craton, with low Th and U concentrations around 2 and 0.7 ppm. Low U concentrations in the SMZ are not a consequence of high grade metamorphism. The NMZ, with high Th, U concentrations (10.8 and 2.5 ppm) and radiogenic Pb, resembles the adjoining Zimbabwe craton more. Using an approach similar to that of Ridley (1992) we consider the differences in late Archean tectonic styles between NMZ and SMZ+KC as a possible consequence of the differences in Th and U content of these provinces.
Berger M & Rollinson H, Geochimica Cosmochimica Acta, 61, 4809-4829, (1997).
Holzer L, Frei R, Barton JM & Kramers JD, Precambrian Research, 87, 87-115, (1998).
Kamber BS & Biino GG, Schweizer Mineralogisch Petrographische Mitteilungen, 75, 427-454, (1995).
Ridley J, Precambrian Research, 55, 407-427, (1992).
Treloar PJ, Coward MP & Harris NBW, Precambrian Research, 55, 571-587, (1992).
van Reenen DD, Barton JM, Roering C, Smit CA & van Schalkwyk JF, Geology, 15, 11-14, (1987).
The 250 km wide Limpopo belt of southern Africa is an east-northeast trending zone of granulite facies tectonites separating the granitoid-greenstone terranes of the Kaapvaal and Zimbabwe cratons. Large scale ductile shear zones are an integral part of Limpopo belt architecture. They define the boundaries between the belt and the adjacent cratons and separate internal zones within the belt. The shear zones forming the external (northern, southern and western) margins of the belt are interpreted as the structures along which uplift of the overthickened crust was accommodated. Although the kinematics of the belt are controversial, the largely contractional structural style and the syn-kinematic metamorphism are widely interpreted to be the products of collisional orogeny.
Two methods have been employed to provide temporal constraints on the age of the Limpopo Orogeny. The crystallization of granites interpreted to be the result of decompression melting during uplift of overthickened crust, provide a minimum age for the contractional stage of orogeny (McCourt et al., 1995). Pb-stepwise leaching of metamorphic silicates in high grade gneisses has been used to determine the age of syn-kinematic metamorphism, and therefore the age of orogenesis (Holzer et al., 1998).
This paper presents new U-Pb (SHRIMP) data for zircons from syn- and post-tectonic granites in the western part of the Central Zone, Limpopo belt, southern Africa. The Selebi-Pikwe and Makowe Hills granites are Late Archaean in age. Field and fabric relationships indicate that these are syntectonic granites which provide a minimum age constraint on the contractional stage of the Limpopo Orogeny. The Mahalapye granite in the extreme western part of the Central Zone is a post-tectonic intrusion that crystallized at 2023 ± 7 Ma This data suggests that mineral ages of ca 2.0 Ga from the eastern part of the Central Zone, date metamorphism during reworking of Archaean age shear zones rather than collision between the Kaapvaal and Zimbabwe cratons as has been suggested (Holzer et al., 1998).
Holzer, L, Frei R, Barton J M Jnr and Kramers J D, Precambrian Reseach, 87, 87-115, (1998).
McCourt S, van Reenen D D, Roering, C and Smit CA, GSSA Centennial Conference Abstracts, 1, 185-188, (1995).
Sm/Nd and 143Nd/144Nd ratios are used to calculate Nd model ages which are assumed to be representative of the average age of the respective crustal segment. The behaviour of the Sm-Nd system has been investigated on two examples of Proterozoic migmatites from the Central Zone of the Limpopo belt (South Africa) to assess whether Nd model ages on migmatites can be used to provide reliable constraints for the growth of the Archean continental crust.
Derivation of granitoid magma is chiefly associated with lower crustal anatexis. The behaviour of Sm-Nd system during partial melting of "fertile" lithologies (paragneisses) following the Bt dehydration reaction largely depends on the mineral composition of the protolith. Some pre-existing minerals such as garnet or monazite resisted isotopic exchange with the melt and/or co-existing minerals at the time of the metamorphic event, i.e. they represent petrographic, and therefore isotopic relics. Consequently, Nd isotope exchange between migmatitic components is often incomplete. Garnet is characterized by high Sm/Nd ratio and Nd radiogenic isotopic compositions. Its occurrence in the partial melt may dominate the HREE and raises the Nd isotopic compositions of leucosomes. In that case, their Nd model ages tend to be younger compared to those of the paleosomes. In contrast, monazite presents low Sm/Nd ratio and unradiogenic Nd isotopic composition and its entrainment or its dissolution in partial melt reduces the Nd isotopic compositon of the melt. Therefore, the resulting Nd model ages are older than those obtained on paleosomes. It appears that minerals may also fractionate the Sm/Nd ratio during intracrustal processes and may hamper Nd isotope exchanges between migmatitic components. For all these reasons, caution is needed in the use of the Nd model ages as indicator of crustal history.
Nevertheless, detailed isotopic studies performed on migmatitic rocks of the Limpopo Belt help to constrain the crustal history in southern Africa. The crustal evolution of the Limpopo Central Zone can be summarized into three main periods: (1) 3.2-2.9 Ga, (2) ~2.6 Ga, and (3) ~2.0 Ga. The two first periods are mainly characterized by magmatic activity leading to the formation of TTG such as the Sand River Gneisses or the Bulai Granite intrusion. The Early Proterozoic event took place under high-grade metamorphic conditions during which partial melting formed large amount of granitic melt.
The modern Earth loses heat primarily through thermal windows opened by separation of internally rigid lithosphere plates. Such plate tectonics began operating only after Archean time (e.g., Hamilton, 1998): so how, indeed, did the hotter Archean Earth lose heat? Archean cratons display no viable evidence for either divergent or convergent rigid-plate tectonics. Granite-and-greenstone terrains broadly expose rocks only minimally to moderately strained and metamorphosed, and plate boundaries should be obvious if present. There are no stratal wedges deposited on rifted margins, let alone thrust belts of such wedges recording subsequent collisions. There are no disjunct juxtapositions to suggest suturing, and no suture markers such as ophiolite slabs, polymict melange, and broken formation. There are no linear magmatic arcs. Collisional megathrust complexes, postulated from geochemistry where exposures, dating, or mapping are poor, have not been subtantiated. Instead, very thick sections of mostly submarine mafic, and subordinate ultramafic, volcanic rocks, and mostly younger subaerial and submarine felsic volcanic rocks and sediments, were oppressed into complex synforms between rising young domiform felsic batholiths mobilized by hydrous partial melting in the lower crust. Upper-crust granite-and-greenstone terrains underwent moderate regional shortening, decoupled from the lower crust, during compositional inversion accompanying doming, but cratonization soon followed. Tonalitic basement of unknown tectonic setting is preserved beneath some greenstone sections but supracrustal rocks commonly give way downward to correlative or younger plutonic rocks. Basal-crustal and mantle rocks are nowhere exposed. Ancient (3.8-4.3 Ga) zircons, as clastic grains in younger quartzites and xenocrysts in younger migmatites, are of unknown provenance in a pre-Archean Earth bombarded by megabolides. Magmatism in preserved Archean cratons was not continuous enough in time and space to have typified Earth heat loss, so cooling was accomplished mostly in the long-vanished parts of the Archean Earth's shell. Rationale.--Then as now, lateral temperature differences of hundreds of degrees characterized upper mantle, heat was lost primarily from hot regions, "plumes" probably did not exist, and developing continents were concentrated in cool regions. Hot-region upper mantle was partly molten, and voluminous magmas, mostly ultramafic, erupted through many ephemeral submarine vents and rifts focussed at the thinnest crust. Mantle was too hot and mobile to permit rigid lithosphere to form, and the dense volcanic crust was quickly recycled into the weak mantle with patterns more like those of lava lakes than of rigid plate tectonics. Sinking hydrothermally altered lavas gradually hydrated the mantle and made later plate tectonics possible. Surviving Archean crust is from regions of cooler, and more depleted, mantle, wherein greater stability permitted uncommonly thick volcanic accumulations from which voluminous partial-melt, low-density felsic rocks could be generated.
Hamilton WB, Precamb. Res, 91, 143-179, (1998).
With supracrustal successions and granite batholiths ranging in age from >3.52 Ga to 2.88 Ga the Pilbara Craton preserves one of the most complete records of middle Archean history. It evolved over two ~360 myr tectonic cycles and can be subdivided into six fault-bound tectonostratigraphic domains. Existing U-Pb in zircon geochronology indicates that the bulk of the intermediate to silicic igneous rocks in the Pilbara formed during seven periods of paired volcanic and plutonic activity. The distribution, relative abundance, petrogenesis and likely tectonic settings represented by each of these suites is critical to an understanding of the early growth and evolution of Pilbara continental crust. The extent of pre-3.5 Ga rocks is uncertain, but appears limited to one or two greenstone belts and minor phases in batholiths in the eastern Pilbara. Volcanic rocks are calc-alkaline dacites with intrusive anorthosites. The period 3.47 to 3.42 Ga represents the major episode of crustal growth in the eastern Pilbara domains with calc-alkaline basalts, andesites and dacites in most greenstone belts, and TTG suite granitoids in most batholiths. The compositions of calc-alkaline volcanic rocks resemble those from modern supra-subduction environments with TTG magmas derived via melting of underplated or subducted mafic crust. A second widespread magmatic episode at ~3.33 Ga in the eastern Pilbara involved rhyolites and I-type granitoids derived via extensive melting of older intermediate to silicic crust. After this time the locus of magmatism shifted to domains in the western and central Pilbara with TTG magmatism in the western Pilbara and calc-alkaline magmatism in the central Pilbara between 3.27 and 3.23 Ga. The bulk of west Pilbara greenstone belts and granite batholiths were generated in magmatic episodes at ~3.11 and 3.00 to 2.98 Ga with both episodes including calc-alkaline and TTG magmas. Late magmatism in the western Pilbara resulted from crustal melting by plume-derived mafic magmas at ~2.93 Ga. Western Pilbara domains were probably accreted to eastern Pilbara domains after this time (by 2.88 Ga) with localized crustal melting in the eastern Pilbara producing fractionated Sn- and Ta-bearing granites and pegmatites.
Volcanic rocks belonging to the lower Warrawoona Group in the eastern sector of the Pilbara Craton in Western Australia vary in preservation from virtually undeformed lower greenschist grade to severely altered meta-amphibolites. U/Pb zircon dating of felsic formations indicates that emplacement of the lower Warrawoona group volcanics occurred before ca. 3.47 Ga (Barley, 1997). A 40Ar/39Ar age of 3.49 Ga of greenschist amphiboles from the North Star basalt in the Talga Talga section confirms the excellent preservation of these rocks and suggests that metamorphism occurred soon after extrusion. Similar lithologies have been recognised throughout the area in the Marble Bar Belt, the Kelly Belt, the Gorge Ranges and are remnants of a formerly wide spread upper plate. Granodiorites of 3.46 Ga (U/Pb zircon ages) have intruded into this upper plate sequence in the north Shaw and in the Mt. Edgar Batholith near Marble Bar. The upper plate sequence consists of an imbricated stack of thrust sheets with contrasting degrees of metamorphic overprinting, and is separated from lower plate gneisses by prominent mid crustal detachments. This configuration has been recognised in the northern and eastern Shaw Batholith, the southern Mt. Edgar Batholith, and in the northern margin of the Kurrana Batholith. The lower plate typically consists of banded grey gneisses that show evidence for a complex thermal history. The detachments have typically been the focus of late intrusion ranging in composition from gabbroic to muscovite bearing granitic sheets. Although similar in setting, a combination of kinematic and geochronological arguments suggests that the three identified detachments are not connected: the Split Rock Shear Zone in the Shaw Batholith is 3.46 Ga old although reactivation as young as 3.2 Ga cannot be ruled out. The South Edgar Marginal Shear is 3.31 Ga old although 3.47 Ga old gabbro sheets may point to an earlier component in this shear zone. The Kurrana Shear Zone predates 3.2 Ga as measured from the cooling age of metamorphic hornblende from the Middle Creek basement complex. The deposition age of the Mosquito Creek metasediments, which tectonically overly the Kurrana Shear Zone, is bracketed between 3.2 Ga, the age of high grade metamorphism in the Kurrana basement, and 2.9 Ga, the age of mafic sills in the eastern sector of the Mosquito Creek domain. We conclude that mid crustal detachments consistently yield kinematic data indicative of large scale horizontal motions at different periods in the Mid Archean tectonic evolution of the eastern Pilbara Craton. These we relate to cycles of extensional and compressional tectonics, which pre-date the final amalgamation of the East and West Pilbara Terranes at ca 2.9 Ga (Smith et al. 1998).
Barley ME, Greenstone Belts, Oxford University Press, 650-657, (1997).
Smith JB, Barley ME, Groves DI Krapez B, Mcnaughton NJ, Bickle MJ & Chapman HJ, Precambrian Research, 88, 143-171, (1998).
The late Archean (2.7 Ga) Schreiber-Hemlo greenstone belt is characterized by lithologically and geochemically diverse convergent margin igneous rocks. These include: (1) mafic to intermediate tholeiitic flows, (2) mafic to intermediate calc-alkaline flows, (3) felsic calc-alkaline flows, (4) syn-kinematic ultramafic (picrites) to felsic dykes, and (5) high-Al, TTG plutons. All these suites share positively fractionated REE patterns, and negative anomalies of Nb and Ti. Some ultramafic and intermediate dykes, and mafic to intermediate flows have negative Zr and Hf anomalies, whereas some intermediate and felsic calc-alkaline flows have positive Zr and Hf anomalies with respect to neighbouring REE.
Fractionated REE patterns, negative Nb and Ti anomalies, and transition metal contents are all consistent with a metasomatized mantle wedge source for the tholeiitic and calc-alkaline volcanic suites, and ultramafic to intermediate dykes. The geochemical characteristics of mafic to intermediate tholeiitic and calc-alkaline flows suggest a deeper and more primitive mantle source compositions for the calc-alkaline suite than the tholeiitic counterpart. Felsic flows and dykes, and TTG plutons are all defined by the enrichment of LREE, Th, and Zr, and by the depletion of Nb, Ti, Cr, Sc abundances. High Al2O3/TiO2, Zr/Y, Sr/Y, La/Ybn, and Gd/Ybn ratios for felsic rocks are consistent with slab melting and garnet + amphibole ± clinopyroxene residue in the source.
The existence of overlapping fields between mafic, intermediate, and felsic suites on trace element ratio diagrams suggest that processes controlling the production of these arc rocks were more complex than simple slab and/or wedge melting. This complexity may have resulted from a mixture of slab and wedge melts, second stage melting, magma mixing, fractional crystallization, partial equilibration with sub-arc wedge peridotite, crustal contamination, or some combination.
The Sumozero-Kenozero greenstone belt is ~400 km long and up to 50 km wide and comprises a 5-km thick oceanic plateau-like sequence of submarine komatiite-basalt lavas and volcaniclastic sediments. This unit is intruded and overlain by an island arc-like sequence of intermediate-felsic volcanic and subvolcanic rocks including andesitic basalts, andesites, dacites and rhyolites.
Komatiites were derived from a liquid containing ~30% MgO. This liquid was initiated at depths of 300-400 km in a mantle plume that was some 250°C hotter than the ambient mantle. Both komatiites and basalts of the lower sequence are strongly depleted in LREE, have high <epsilon>Nd(T) of +2.7±0.3, relatively unradiogenic Pb isotope compositions (µ m1 = 8.7±0.2) and show Nb-maxima (Nb/Nb* = 1.2±0.2, Nb/U = 43±6). These parameters are similar to those found in a number of uncontaminated early Precambrian greenstones and in recent Pacific OFB. They are regarded as plume source characteristics and provide further evidence for the existence of certain Nb-excess in the Archaean mantle due to the early extraction of large volumes of continental crust with low Nb/U ratios.
The intermediate-felsic volcanic and subvolcanic rocks of the upper unit are enriched in LREE, depleted in HFSE, but have positive <epsilon>Nd(T) values of +2.5±1.2. They represent both mantle wedge-derived basalt-andesite-dacite-rhyolite (BADR) and slab-derived (adakite) melts, erupted in the inner and frontal parts of an intraoceanic island arc, respectively.
U-Pb zircon age of 2875±2 Ma for the felsic volcanic rocks, and Pb-Pb and Sm-Nd ages of 2892±130 and 2916±117 Ma for the komatiites-basalts are in good agreement and correspond to the time of emplacement of the sequences.
The Sumozero-Kenozero greenstone belt reveals the coexistence of fragments of unsubductable oceanic crust represented by the lower mafic-ultramafic volcanic sequence, and the products of subduction-related magmatism. This observation is reconciled within a single model implying that the overthickened plume-derived oceanic crust reached an intraoceanic convergent plate boundary and was intruded and overlain by felsic melts coming from both subducting slab and overlying mantle wedge. Later on, the oceanic plateau together with the volcanic arc complex built on top of it were accreted to and obducted onto the continental crust of the Vodla block.
The late Archaean (~2.7 Ga) Ngezi Group greenstones in the Belingwe Greenstone Belt, south central Zimbabwe, show uplift and increased erosion during deposition of alluvial fans, local derivation of sedimentary material containing no arc derived component and eruption of komatiitic lava. We suggest sedimentation and volcanism occurred as a result of intracontinental extension associated with an active, late Archaean plume.
The Ngezi Group consists of a basal sedimentary sequence (Manjeri Formation), overlain by komatiitic and tholeiitic volcanic rocks (Reliance and Zeedebergs Formations), and a second sedimentary sequence (Cheshire Formation). Previous models invoked to explain the geologic history of the group, range from an allochthonous origin in a compressional setting to deposition in an extensional environment. Although the structure of the area and the volcanic rock geochemistry and petrography have been studied in detail it has still not been possible to distinguish between the two. This has led to implementation of an alternative approach. Studying the changes in the sedimentary facies and sediment geochemistry of the Manjeri Formation indicates the depositional environment prior to emplacement of the volcanic suite.
The Manjeri Formation was deposited in a fluviatile and shallow marine setting, with subsequent alluvial fans and fan-deltas during active tectonism. Changes in the degree of chemical weathering of the provenance area during deposition of the formation, as measured by CIA values, reflect uplift and increased rates of erosion. Detrital mineralogy and rare-earth element patterns, are consistent with derivation from very local sources. Palæogeographic variation in the measured Sm-Nd depleted-mantle model ages between 2.9 and 3.7 Ga are consistent with deposition over basement varying in age from ca. 2.9 Ga - 3.5 Ga, again suggesting local derivation. The facies and geochemical association imply sedimentation in an extensional continental setting.
The Ngezi Group is the type section of the Bulawayan Supergroup found throughout the Zimbabwean Province at ca. 2.7 Ga. Therefore, understanding the tectonic setting of the Ngezi Group is fundamental to a comprehension of the development of the Zimbabwean craton at this time.
The magmatic supracrustal rocks of the Archean Sukumaland Greenstone Belt, NW-Tanzania, consist of thick successions of metabasalts, metagabbros as well as felsic and minor intermediate volcanic flows and pyroclastics. Igneous rocks of ultramafic composition are surprisingly scarce. Metasedimentary rocks comprise carbon-rich pelites, a prominent banded iron formation (BIF), and subordinately arenites and conglomerates. Previous field investigations suggested a magmatic evolution from an early mafic volcanic stage (Lower Nyanzian Supergroup) to a more mature felsic stage with contemporaneous chemical and minor clastic sedimentation (Upper Nyanzian Supergroup) (Borg, Shackleton, 1997). It had been speculated that the paleosome of the adjacent Kahama migmatitic gneisses might represent the significantly older basement to the greenstone belt (Dodson et al., 1973). However, new U/Pb single zircon isotopic age data partly contradict these previously established field relationships and suggest a more complex magmatic and structural evolution instead.
Two units of felsic metavolcanics define the age bracket for the deposition of the Nyanzian greenstone belt: Two older rhyolites yield ages of 2808±3 Ma and 2780±3 Ma, respectively, and one trachyandesite and one rhyolite provide ages of 2699±9 Ma and 2654 +15/-13 Ma, thus giving a minimum age for the greenstone belt. The paleosome of a migmatitic gneiss at Kahama has been dated at 2680±3 Ma thus being coeval with the late stage of the Nyanzian greenstone belt volcanism.
Previous authors suggested an event of widespread migmatisation of this region between 2530 Ma and 2570 Ma based on Rb/Sr whole rock ages of granitoids (Rammlmair et al., 1990). However, our new data suggest that this interpretation needs revision. It cannot be excluded that Rb/Sr whole rock systems have been disturbed by later hydrothermal redistribution of the alkalis coeval with granulitization of the lower crust as demonstrated elsewhere, e.g. for the Kapuskasing Structural Zone, Superior Province, Canada (Krogh, 1993). Furthermore, the wide range of metamorphic conditions of the Sukumaland Greenstone Belt rocks, from lower greenschist to upper amphibolite facies suggests considerable vertical movement after peak metamorphism.
Borg G, Shackleton RM, Greenstone Belts, Clarendon Press, Oxford, 608-619, (1997).
Dodson MH, Bell K, Shackleton RM, Fortschr. Miner, 50, 67-68, (1973).
Krogh TE, Earth Planetary Sci. Letters, 119, 1-18, (1993).
Rammlmair D, Höhndorf, A, Borg G, Hiza, GN, 15th Colloqu. Afr. Geol. Nancy, Abs.Vol, 43, (1990).
The Isua Greenstone Belt contains the oldest known, relatively well preserved, metavolcanic and metasedimentary rocks on Earth. The rocks are all deformed and many were substantially altered by one or more episodes of pervasive metasomatism (Rose et al., 1996; Rosing et al., 1996). Both the deformation and metasomatism were heterogeneous, and primary volcanic and sedimentary features are locally well preserved. New geological mapping has traced out gradations between the bestpreserved protoliths and their diverse deformed and metasomatised equivalents. The previously mapped stratigraphy (Nutman, 1986) was found to be of little value in understanding the geology. This stratigraphy is a muddled mixture of units defined by different and diverse criteria, and largely represents a misinterpretation of the primary nature of the rocks. The new work indicates that most of the Isua Greenstone Belt consists of fault-bounded rock packages, mainly derived from basaltic and high-Mg basaltic pillow lava and pillow lava breccia, and chert - BIF, with a minorcomponent of clastic sedimentary rocks derived from chert and basaltic volcanic rocks. An extensive unit previously mapped as felsic volcanic rocks was instead found to have been largely derived from metasomatised basaltic pillow lava and pillow breccia intruded by numerous sheets of tonalite.
Nutman AP, Gronlands Geologiske Undersogelse Bulletin, 154, 80 pp, (1986).
Rose NM, Rosing MT & Bridgwater DB, American Journal of Science, 296, 1004-1044, (1996).
Rosing MT, Rose NM, Bridgwater DB & Thomsen HS, Geology, 24, 43-46, (1996).
Kyanite has been found in muscovite-rich schists from the > 3700 Ma Isua supracrustal belt (ISB). At Isua kyanite is constrained to Al-rich metasomatites formed by transformation of felsic gneisses by fluids derived from ultramafic schists. Peridotite and dunite bodies were transformed into talc - anthophyllite - chlorite - magnesite schists during prograde reaction with ambient fluids in the ISB (Rosing and Rose, 1993). During amphibolite facies metamorphism and deformation the ultramafic schist released fluids, which interacted with neighboring lithologies. These fluids were buffered at silica activities an order of magnitude lower than that of quartz saturation by the talc - anthophyllite - magnesite mineral assemblage. At this low silica activity, chlorite in the ultramafic schists buffered (aAl+)/(aH+)3 activity ratios close to that of corundum saturation. These fluids caused silica depletion, and the high aluminum ion activities intersected kyanite saturation in the quartz saturated felsic rocks. The presence of kyanite is thus a consequence of the metasomatic reaction, and the timing of kyanite formation constrains a phase of fluid flow associated with deformation in high strain zones.
Progressive leaching in three steps of a pure separate of kyanite from a metasomatic kyanite - muscovite schist in the Southwestern part of the ISB defines a Pb-Pb isochron with an age of 2847 ± 26 Ma, MSWD = 0.70. This age is consistent with a Sm-Nd amphibole - plagioclase isochron of 2849 ± 116 Ma from mafic schists at Isua (Gruau et al., 1996) and a Pb step leaching age of 2840 ± 49 Ma (MSWD=1.43) on magnetite from Isua BIF (Frei et al., in press) and with regional granulite facies metamorphism in other parts of the Greenland Archaean block.
This study shows that the ISB was subjected to deformation and high grade metamorphism during the late Archaean, and that the presence of kyanite does not constrain early Archaean geotherms as suggested by Boak and Dymek (1980).
Boak, J. and Dymek, R.F., Geol. Soc. Amer. Abstr., 12, 389, (1980).
Gruau, G., Rosing, M.T., Bridgwater, D., Gill, R.C.O., Chem. Geol., 133, 225-240, (1996).
Frei, R., Bridgwater, D, Rosing, M.T. and Stecher, O., Geochim. Cosmochim Acta, In press
Rosing, MT & Rose, NM, Chem. Geol, 108, 187-200, (1993).
Archaean chemical sediments comprising cherts, oxide and sulphide facies iron formations contain information about the chemical state of the early oceans and, by implication, the tectonic processes which buffered the chemistry of the early oceans and atmosphere. We compare the chemistry and isotopic composition (13C, 34S, <epsilon>Nd and REE) of very well preserved oxide and sulphide facies ironstones sampled by the NERCMAR drill hole through the 2.7 Ga Manjeri Formation in the Belingwe Greenstone Belt with published and new data on the metamorphosed and deformed iron formations from the 3.7 Ga Isua Greenstone belt. 13C and 34S isotopic fractionations in the Belingwe samples may be interpreted in terms of a complex bacteria/archaea eclogical community. REE and Nd-isotopic variations may be modelled by contributions from a reduced hydrothermal component and a component surprisingly similar in REE pattern to modern seawater, given allowance for a small clastic component. Isua rocks are less well preserved but the overall similarity of the REE compositions implies deposition from a broadly similar ocean to that in the late Archaean.
Models for the evolution of the Earth's earliest continental crust are critically dependent upon reliable, correctly interpreted geochronology. Field relationships which might aid the interpretation of complex patterns of radiometric ages are often obscured or obliterated by later, polyphase tectonothermal processes, but developments in high-spatial resolution analysis (e.g. SIMS), together with imaging techniques such as CL or BSE, provide a possibility to date specific growth phases within zircons. Internal structure of zircon revealed by imaging has been used to infer an environment for crystallisation. For example, strongly oscillatory zoned zircon is commonly considered to have be crystallised from a silicic magma. Zircon crystallised under metamorphic conditions is sometimes accompanied by low Th/U ratios. There are, however, ambiguities in these interpretations, particularly when imaging methods reveal little or no internal structure.
SIMS analyses of REE, Y and Hf, performed in the same analytical sites as U-Th-Pb dating analyses, provides further information for elucidation of zircon growth histories. The SIMS technique has the advantage of minimal sample crater depth, thus permitting a high degree of confidence that the trace element analysis corresponds closely to the dated phase of the zircon. In this study, REE, Y and Hf data are presented for zircons from early Archaean Amîtsoq gneisses of Gothåbsfjord, West Greenland, for which there is an ongoing controversy about the interpretation of ion-probe ages in the range >3.8 - 3.6 Ga; either >3.8 Ga zircons are inherited in a ca. 3.65 Ga magma, or ca. 3.65 Ga metamorphism affects zircons in rocks crystallised at >3.8 Ga. In one of the studied samples (GGU 110999), CL-imaging reveals three phases of early Archaean zircon crystallisation. In order of decreasing age, (1) oscillatory zoned cores (>3.8 Ga) have steep HREE (YbN/GdN ~ 17 - 26) and LREE (PrN/GdN ~ 0.016 - 0.03), negative Eu anomalies (Eu/Eu* ~ 0.4), positive Ce anomalies (Ce/Ce* ~ 26 - 127), and near-chondritic Zr/Hf; (2) corrosive recrystallisation of a core (>3.74 Ga, low Th/U) has steep HREE (YbN/GdN ~ 55), flat LREE (PrN/GdN ~ 0.5), no Eu anomaly (Eu/Eu* ~ 1), small positive Ce anomaly (Ce/Ce* ~ 3), and Zr/Hf ~ 0.5x chondritic ; (3) zoned rims (ca. 3.65 Ga) have steep HREE (YbN/GdN ~ 54 - 31) and LREE (PrN/GdN ~ 0.018 - 0.06), negative Eu anomalies (Eu/Eu* ~ 0.16 - 0.32), positive Ce anomalies (Ce/Ce* ~ 72 - 423), and near-chondritic Zr/Hf.
The similar trace-element compositions from phases 1 and 3 suggest zircon crystallisation in a similar, silicic magmatic environment, supporting a crustal evolution model with ca. 3.65 Ga TTG magmatism inheriting >3.8 Ga zircons. Further work is required to determine whether igneous and metamorphic growth zones in zircons can be distinguished by specific geochemical tracers.
Detailed chronological scales (4.4-3.4 Ga) of the Earth and Moon mantle and crustal events were constructed. There is a distinct correlation between of the duration (near 100 Ma) and the starting time of some cycles on the Earth and the Moon, accordingly (Ma): - and 4512±8.6; 4399±18 and 4418±10; 4301±66 and 4320±10; 4196±21 and 4214±4; 4115±5 and 4110±5; 4016±7 and 4010±4; 3913±19 and 3910±5; 3812±5 and 3811±4; 3715±10 and 3716±8; 3610±10 and 3614±10; 3512±10 and 3523±15; 3413±10 and 3413±17. This coherence of dynamics of the crust growth suggests an influence of, probably, an external - "outer-space" factor, common to both bodies. That factor was a "trigger" mechanism that initiated the mantle mafic-ultramafic and crustal acid magmatism and teconic activity. The origin of periodicity is likely to result from the change of gravity pull from the Galaxy Centre affecting the Solar System, as it follows its elliptical orbit, because the current half-duration (106-135 Ma) of the rotation period ("Galactic Year"- GY) is close to the identified periodicity of processes in the Early Archean and the accretion epoch. With the use of data on the length of separate cycles of the Moon and Earth, it is possible to calculate more precisely the length of the GY for the oldest history of Solar System (197.8±2 Ma). This study was supported by Russian Fund of Fundamtal Investigations (97-05-64863).
This contribution brings together my studies from opposite ends of the terrestrial age range and discusses their relevance to Archaean crustal genesis, building upon an earlier proposal (Osmaston 1992a).
At the far end, building iron cores in all the terrestrial planets by two-stage accretion or by percolation both face serious difficulties. An alternative is to react magmatic FeO with a reducing nebular atmosphere at the surface of a vigorously convecting protoplanet, to give Fe/FeS plus a hydrated crust, and then to 'subduct' them. This gives the early Earth a substantially wet mantle, with two consequences. Firstly, mantle viscosity is reduced by 1-2 orders of magnitude, giving ample ability to extract the early-Earth heat production. Secondly, komatiites were produced from a relatively wet mantle, evidenced by the constant association of felsics with them, by occurrences of komatiitic tuffs, by Nb anomalies and by S- and H2O-borne rich mineralizations (syn-eruptive segregation?). So Archaean greenstone belts may escape the subduction connotation and start their lives as MOR-equivalents.
At the near end of geological time, study of Phanerozoic subduction, including circum-Pacific histories and activity, shows the action of two major processes: subduction tectonic erosion (STE) and post-subduction magmatism (PSM), inaccurately called post-collision magmatism. In STE mechanical action at plate downbends removes material from the hanging wall and rapidly undercuts the margin by many hundred kilometres, rendering it liable to imbrication. The essential part played by pre-collision STE in construction of the Alps has been outlined (Osmaston, 1997). PSM (Osmaston, 1992a) is silicic-granitoid, lasts for up to 40 Ma, and is attributed to wholesale melting of subduction interface crustal material by heat that has soaked upward through the former slab. A diagnostic 'oceanward' migration ("sweepback") of magmatic onset is often seen.
Where data are available, both STE and PSM were confined to where the subducting plate was young (<70 Ma), which makes both processes highly likely in the Archaean. Central to my Alps synthesis, based on studies elsewhere (Osmaston, 1992b), was the recognition that subduction commonly begins within the oceanic domain, later reaching the continental margin by STE and imbrication. I propose that Archaean greenstone belts were originally passive margin oceanic crust that thus became part of an extensive subduction-undercut margin. TTG intrusion is seen as PSM when subduction was halted, perhaps by arrival of a micro-craton. These features will be discussed briefly by reference to the ~3.3 Ga Barberton and 2.7 Ga western Superior areas.
This interpretation implies that the late Archaean acceleration in TTG/greenstone belt addition to cratons represents an increasing frequency of interruption of subduction. Each such PSM event advected mantle heat to the surface that would otherwise have been returned to the mantle. The increasing greenstone/MOR-crest water depths in the late Archaean suggest a cooling mantle.
Osmaston, MF, 29th IGC, Kyoto, Abstr.1, 46, (1992a).
Osmaston, MF, 29th IGC, Kyoto, Abstr.1, 63, (1992b).
Osmaston, MF, EUG9, Strasbourg, Abstracts, 341, (1997).
1. There are two shields, the Ukrainian and Fennoscandian/Baltic shields (FS), in the East European Craton. The eastern part of the FS is extremely important for deciphering the Late Archean history because it is only there that rocks of this age are exposed over a large area. Three provinces: Karelian (KP), Belomorian (BP) and Kola, are distinguished in the Archean segment of the FS in terms of geological structure and evolution. KP is a granite - greenstone province slightly and incompletely reworked in Proterozoic time. BP is a mobile belt which evolved in polycyclic manner 2. A lateral sequence of Late Archean tectonic units was revealed in BP and in the eastern part of KP. From NE to SW it consists of the following structural - formational zones: 1) the Central Belomorian mafic zone (CBMZ) dominantly formed by mafic and ultramafic rocks, 2) the Chupa Paragneissic Belt (ChPB) composed of deep and repeatedly metamorphosed metagraywackes (mainly high-alumina gneisses), 3) the North Karelian system of greenstone belts (NKGB) dominated by volcanics of calc-alkaline series, 4) the North Karelian diorite-plagiogranitic batholith (NKB) and xenoliths of ultramafic to andesite - dacite composition that occur in it. 3. CBMZ is dominated by metabasalts (amphibolites) with widespread metaultrabasic rocks (metaperidotites, serpentinites and apoultramafic amphibolites) and extremely rare acid metavolcanics. The chemical composition of metabasalts in CBMZ is similar to that of MORB. The isotopic age of the rhyolite-dacites is 2 887 Ma (Bibikova et al., 1997). This association is interpreted as a fragment of the Late Archean ophiolitic complex (Slabunov & Stepanov, 1998).The supracrustal strata of NKGBS. consist of metabasalts, metakomatiites and felsic to intermediate metavolcanics. Metaandesites - metarhyolites make up a considerable portion of the sequence. The age of these volcanics was estimated at 2877- 2820 (Bibikova et al, 1995, 1997). Between NKGB and CBMZ there lies ChPB which consists of metagraywackes (garnet - biotite kyanite - bearing gneisses.) 5. This lateral series indicates Late Archean (3.0-2.8 Ga) subduction stages in the evolution of the eastern FS (Gaal & Gorbatchev, 1987; Bogdanova et. al., 1995). The oceanic lithosphere subducted from NE to SW under the subcontinental crust (1-st stage) and continental crust (2-nd stage) of the Karelian plate. 6. The collision events in the BP are represented by high pressure (6-12 kbar) and high temperature (500-700°C) kyanite - facies metamorphism (Volodichev, 1990), granitoid magmatism (Lobach-Zhuchenko et. al., 1995), the formation of folded nappe structure (Miller & Mil'kevich, 1995) and granite gneiss domes (Slabunov, 1993). Here,the time of collision is estimated at 2,7 -2,74 Ga (Bibikova, 1995). This stage of evolution in NE part of KP is accompanied by the generation of NKB. CBMZ marks a collision suture.
The Keivy A-granite complex consists mainly of sheet-like peralkaline granite bodies, granosyenite dykes and minor nepheline syenite fault-type intrusions with the total area of exposure ca. 2500 km2. The five fractions of zircons from aegirine-augite-lepidomelane-ferrohastingsite granosyenite of Western Keivy massif yield the U-Pb near concordant age of 2674±6 Ma. The emplacement ages for peralkaline granite magmatic phase according to U-Pb zircon data vary from 2610 Ma (White Tundra massif) to 2670 Ma (Western Keivy massif). The granites are spatially confined to voluminious gabbro-anorthosite magmatism of 2660-2680 Ma age (Bayanova et al., 1998). The predominantly "juvenile" Sm-Nd isotopic signatures from most suites of Keivy complex (epsNd(T) vary from +0.6 to +1.3; TDM =2780 Ma) suggest that they must be of mantle derivation, or have the short-lived crustal precursors (Zozulya et al., 1997). The granites are petrologically and geochemically similar to Phanerozoic A-type granitoids, presumably emplaced into noncompressive or extensional environments. The distinct tectonic regime of such type of granites indicates that the Keivy peralkaline granite magmatism can be regarded as a consequence of post-collisional events. Collision in the region has possibly taken place earlier than 2740 Ma (Mitrofanov et al., 1997). The granites studied were formed after the Late Archaean Keivy-Voronja greenstone belt evolution. The proposed geodynamic model of origin of Keivy A-granite complex suggests that the Archaean of NE Baltic shield was dominated by plume tectonics.
Bayanova TB, Mitrofanov FP & Levkovich NV, Abstracts of ICOG-9, Chinese Science Bulletin, 43, 6, (1998).
Mitrofanov FP, Bayanova TB, Balabonin NL, Sorokhtin NO & Pozhilenko VI, S-Petersburg University Proceedings, 7, 5-18, (1997).
Zozulya DR, Balashov YA, Timmerman MJ & Vetrin VR, Abstracts EUG 9 Meeting, 135, (1997).
According to structural analyses, the Belomorian Belt is a tectonic pile of metasedimentary, metavolcanic and metaplutonic rocks which has been folded and metamorphosed several times. Our isotope dilution data indicate that the earliest metamorphic event took place c. 2850 Ma ago. Sm-Nd isotope studies of whole rock metasediment samples have constrained their mean protolith ages to between 3000 and 2860 Ma, indicating a short prehistory. A Neoarchaean origin of the Belt is supported by new secondary ion mass spectrometer (NORDSIM) zircon ages from discordant mafic dykes traversing the Belt.
To obtain a better insight into the spectrum of provenance ages in Belomorian metasediments, we have dated detrital zircons with the NORDSIM instrument in Stockholm. Zircons were separated from different localities and at various levels of the Belomorian tectonostratigraphical column. We have recognized three age groups of ancient cores at 3.2-3.1 Ga, 3.00-2.97 Ga and 2.93-2.90 Ga. The plus 3.1 Ga cores were obtained solely from localities in the northern part of the Belt. It has also been possible to distinguish three groups of metamorphic grains and overgrowths which are 2.84-2.80 Ga, 2.72-2.68 Ga and c.2.61 Ga old.
The NORDSIM data presented confirm the absence of detrital material older than 3.2 Ga in the Belomorian metasediments. This differs from the adjacent Karelian craton where crustal rocks of about 3.4 Ga have been recorded. If these ages are representative of the rocks discussed, our data suggest that the Belomorian Belt represents a Neoarchaean accretionary environment in the vicinity of the Karelian craton.
The Kolmozero-Voronja greenstone belt is located between two Upper Archaean terrains: Murmansk and Central Kola. Four suites are distinguished in the greenstone belt: Ljavozerskya (lower terrigenous formation), Polmostundrovskya (komatiite-tholeiite), Voronjatundrovskya (basalt-andesite-dacite) and Chervurtskya (an upper terrigenous formation). Small bodies (10-15 m in thickness) of ovoid plagioamphibolites are present among shistosed plagioamphibolites of the Polmostundrovsky suite. The ovoid plagioamphibolites are dark green rocks with large rounded aggregates of plagioclase (40-50 mm in diameter). Subconcordant bodies of quartz porphyry occur in the upper part of the Voronjatundrovskya suite. They are considered to be intrusive vein analogs of acid volcanites. Quartz porphyry is a fine-grained pale-grey rock with quartz and plagioclase impregnants (up to 5 mm). Zircon in the quartz porphyry consists of long prismatic and short prismatic crystals with an oscillatory zoning that is characteristic of magmatic crystallization. Eight fractions of these zircons yield a discordant U-Pb age of 2828±8 Ma, that we interpret as an age of intrusive emplacement of quartz porphyry that at the final stage of the belt development. Sphene in the ovoid plagioamphibolites consists of pale-yellow crystals of irregular crystallographic forms. Preliminary dating of three fractions of sphene yielded an U-Pb age of 2595±20 Ma, that probably is connected with the closure of the U-Pb isotopic system during the regional andalusite-sillimanite facies metamorphism.
Supracrustal assemblages of the late Archean (ca. 2.75-2.69 Ga) Schreiber-Hemlo greenstone belt of the Superior Province, Canada, are composed of tectonically juxtaposed fragments of oceanic plateaus, oceanic island arcs, and siliciclastic arc-derived trench turbidites. Following juxtaposition, these diverse lithologies were collectively intruded by syn-kinematic, ultramafic to felsic dykes, and TTG plutons, with subduction zone geochemical signatures. Overprinting relations between different sequences of structures suggest that this greenstone belt underwent three phases of deformation. During D1 oceanic plateau sequences, trench turbidites, and arc basalts were all tectonically juxtaposed as they were incorporated into an accretionary complex. D2 right-lateral transpression resulted in further fragmentation and mixing of oceanic plateau fragments, trench turbidites, and arc-derived mafic to felsic igneous rocks, with development of broken formations and a tectonic mélange.
The D2 strike-slip faults in the fore-arc region of the Schreiber-magmatic arc may have provided conduits for uprising melts from the descending slab, and induced decompressional partial melting in the sub-arc mantle wedge, to yield respectively syn-kinematic felsic and gabbroic intrusions. A similar close relationship between orogen-parallel strike-slip faulting and magmatism has recently been recognized in several Phanerozoic transpressional orogenic belts, including the North American Cordillera, Japanese island arcs, and British Caledonides, suggesting that as in Phanerozoic counterparts, orogen-parallel strike-slip faulting within the Schreiber-Hemlo greenstone belt played an important role in lateral crustal accretion and magma emplacement. It is suggested that subduction-accretion, with associated mélange formation and magmatism, was an important mechanism of continental growth in the late Archean southern Superior Province.
The origin of tonalites and trondhjemites (TT) is likely to be a result of partial melting of mafic sources, which is proved by experiments on melting of natural amphibolites. Using the published data (Rapp & Watson, 1995; Sen & Dunn, 1994; Rapp, 1994; Klein et al., 1994; Wyllie et al., 1997; etc.), we have calculated the concentrations of some trace elements in model tonalite - trondhjemite melts, generated at P from 3 to 22 kbar and T from 800 to 1040°C.
The whole rock and rare and REE- element composition was investigated along a profile about 500 km long - from Varanger and Kirkenes TT gneiss in Northern Norway via KSDB in Pechenga Proterozoic structure and towards the eastern part of the Murmansk block .
Plagiogneisses of the KSDB (2835±5 Ma) correspond to model tonalite- trondhjemite melts formed in equilibrium with different residues such as garnet amphibolite (biotite gneiss, A) and plagioamphibolite (epidote- biotite gneiss, B). The estimated composition of metabasic substratum is similar to tholeiite TH1, which is slightly enriched of the LREE (up to 15- 30 chondrite levels). Melting degree is 10- 20%. Plagiogneisses of A and B types could be formed under different P-T- conditions: the former at over 15 kbars and the latter at about 8 kbars.
Within the Kirkenes Gneiss (2804±9 Ma), the Varanger Gneiss (2813±6 Ma) and the Garsjo Complex (2700±150 Ma) we can distinguish three types of plagiogneisses of variable geochemistry. Some of them are similar to those from the KSDB, whereas others differ in trace element composition. On the other hand, trondhjemites extremely depleted in HREE occur solely in the Norvegian gneisses but they are not known in the KSDB. Protoliths of these rocks should be formed in equilibrium with eclogite residue at a pressure of more than 16 kbar.
Within the Murmansk block, from west to east, there is an increase in the REE content in the initial melts, a change in composition of protoliths from TH1 tholeiite with the highest content of REE to the subalkaline basalt, and in the eastern part of the Murmansk block the REE content is even higher than in TH2. The increase of alkalinity in the protoliths of TT- gneiss correlates with the abandance of the Late Archean peralkaline (2750 ±50 Ma) and alkaline (2760±60 Ma) granite massifs here.
What is the reason for the distinction in composition and origin of gneisses in the KSDB and gneisses on the surface? If the rocks from the KSDB section represent a deeper crustal level, they could have undergone partial melting and related enrichment of restites by HREE. But a verification of this assumption will require a more detailed research.
This work is a contribution to the Project IGCP-408.
In South India, the Dharwar craton presents a natural cross-section through the late-Archaean continental crust. It is cut by several granitic bodies, among which the Closepet granite, which is a huge body (400 x 30 km) that can be observed from the lower crustal levels (granulite facies: P=7-8 kb) to the upper crust (P=2-3 kb). Based on field work three main structural zones are distinguished: (1) A root zone, highly heterogeneous both in petrology (from monzonite to granite) and in texture (phenocryst accumulation, enclaves, etc.). There, a mantle-derived magma underwent fractional crystallisation which was followed by mingling between the residual liquids and melts generated by anatexis of the surrounding gneissic basement. (2) A "channel zone", where evidences of large scale magma ascent can be observed. (3) A zone of superficial intrusions, consisting in independent homogeneous intrusive bodies. There enclaves or phenocryst are very scarce. Zones (1) and (2) are in physical continuity whereas zones (2) and (3) are separated by a "magmatic gap", consisting in a network of dykes providing a magmatic connection between both parts. On both sides of the gap, rocks have exactly the same chemical composition, although only differentiated facies are found in the superficial intrusions. In order to account for all these data, the following model is proposed: Magmas generated in deep crustal levels, progressively rise towards the surface. They present a great petrological heterogeneity as well as contrasted enclaves and phenocryst load, which result in highly variable viscosity and density. At a peculiar crustal level, corresponding to the gap, the ascent of the magmatic mush is stopped. There, only less viscous lighter, enclave- and phenocryst-free differentiated liquids can continue their ascent through the network of narrow dykes. These liquids go up into the upper crust and fill rather small pockets giving rise to the superficial intrusions. On the contrary, the highly heterogeneous and viscous magmas remain at lower crustal levels where they constitute the main mass of the Closepet granite. Consequently, the magmatic gap, which could correspond to a major rheological interface, operates as a "filter" discriminating magmas as function of both their density and viscosity. In conclusion, the Closepet granite, provides an unique opportunity to reconstruct the whole anatomy of a granite batholith, from its root to the sub-surface and to observe and discuss all the physical processes and mechanisms of its emplacement.
Lewisian grey gneisses are a tonalite-trondhjemite-granodiorite (TTG) suite similar to rocks found in Archæan terranes worldwide. The high-grade metamorphism has been dated at 2.7 Ga. The Lewisian is divided into three, the northern, central and southern regions. In the northern region, particularly, there is a suite of trondhjemitic to granitic, so-called 'late' igneous sheets.
H2O-saturated partial melting experiments were carried out on prograde, amphibolite-facies biotite and hornblende tonalites from the northern region, at 0.6 GPa. Melt compositions are unlike those of the late sheets. This suggests that the sheets are not the simple products of anatexis of the TTG gneisses. They must either be residual melts after fractional crystallisation or products of partial melting of rocks other than Lewisian grey gneisses. The second origin is favoured because of the absence of exposed more mafic intrusive rocks (of this age) that could represent the mafic differentiates or parent magmas. Unless the Lewisian terrane has been transported a great distance with respect to the (necessarily) underlying protoliths of the late sheet magmas, these rocks must still lie in the deep crust of NW Scotland.
LISPB seismic data (Bamford et al., 1978) indicates the presence of a crustal layer beneath the Lewisian, separated by a discontinuity. Velocity data indicate that this layer is not mafic. The best model would be granulites of slightly higher density than the Lewisian. This unexposed crust may be the protolith for the late sheets. Unless the Lewisian has been thrust over this material, this suggests that the Lewisian is not the true juvenile felsic basement of the region.
Furthermore, a SHRIMP age of 3.4 Ga, from a zircon core in the northern region (Kinny and Friend, 1994), is much older than the inferred protolith ages for the TTG gneisses, ~ 2.8 Ga in the northern and 3.0 Ga in the central region (Friend and Kinny, 1995). Thus, there are two possibilities. There may have been mafic crust present that underwent metamorphism long before the intrusion of the TTG protolith. The TTG rocks might then have been derived through partial melting of this old mafic crust. Alternatively, there may have been felsic crust already present, and the 3.4 Ga age could represent the crystallisation of this magmatic addition. In either case, the TTG gneisses cannot be the oldest crust in Britain. If the second possibility is correct, the Lewisian is not even the oldest continental crust here. The LISBP data would seem to favour the existence of pre-Lewisian intermediate to felsic crust.
If these suggestions are valid, there are important implications for the timing of crust formation in this part of Europe and for the geological evolution of the British Isles.
Bamford D, Nunn K, Prodehl C & Jacob B, Geophys. J. R. Astronom. Soc, 54, 43-60, (1978).
Friend CRL & Kinny PD, Geology, 23, 1027-1030, (1995).
Kinny PD & Friend CRL, Min Mag, 58A, 481-482, (1994).
The Pilbara granite-greenstone terrain in NW Australia is one of the few places that offer the opportunity to study the generation and evolution of the crust in early to mid-Archaean times. This study focuses on refining the structural and magmatic history of the eastern Pilbara, with the Mt Edgar Batholith and surrounding greenstone belts as a key area.
The Mt Edgar Batholith is a granitoid complex with a sequence of pre-, syn- and post-tectonic intrusive phases, containing an originally subhorizontal mid-crustal detachment zone. This zone is found in the internal part of the Batholith, where it displays a causal relation between uni-directional NE-SW extension and magma genesis. Along the south(west)ern margin, the geometry of structures indicates this zone was tilted partly actively and partly passively during deformation to form the 70 km long, now steeply dipping, 2-3 km wide, Southern Edgar Marginal Shear Zone (SEMSZ). Early movement on this zone juxtaposed migmatitic gneisses adjacent to greenschist and lower greenschist facies supracrustals. Kinematic analyses consistently give a greenstone belt up movement although an opposite sense may be found in the supracrustals.
An integrated thermo-tectonic study of the different intrusive phases associated with the SEMSZ provides geochronological constraints on its activity. Zircon SHRIMP U-Pb crystallization ages for granitoid sheets range between 3312 and 3465 Ma. Hornblende 40Ar/39Ar cooling ages for this area indicate cooling at 3315-3325 Ma.
Evidence for an early deformation phase in the SEMSZ comes from a gabbro/diorite complex (U/Pb age 3465 Ma) with syn-tectonic dolerite sills. A related swarm of dolerite dykes (Ar age >3400 Ma) exploited a conjugate set of NE-SW extensional faults in a felsic extrusive unit. The dykes are feeders for the overlying basaltic units, which are now, as are the felsics, part of a thrust sheet.
Part of the SEMSZ footwall is formed by ~3315 Ma TTG sheets and plutons. Less deformed plutons of similar age have intruded into the hanging wall of the SEMSZ, providing an upper age for its main activity. Later deformation phases overprinting the SESMZ include periods of NNE-SSW, E(NE)-W(SW), and NW-SE compression.
This study indicates that a mid-crustal detachment played a major role in the emplacement of the circa 3315 Ma Mt Edgar granitoid suites and that this occurred during a uni-directional tectonic transport to the NE. Structures within the migmatitic gneisses and the thermal gradients across the detachment at this time are consistent with an extensional tectonic regime. That is, a regime similar to that proposed for the earlier phase of granitoid emplacement at circa 3460 Ma in the Eastern Pilbara (Zegers, 1995).
Zegers TE, Geologica Utraiectina, 146, 208, (1996).
The only remaining areas of pristine 3.6- 2.7 Ga crust on Earth are parts of the Kaapvaal and Pilbara cratons. General similarities of their rock records, especially of the overlying late Archean sequences, suggest that they were once part of a larger Vaalbara supercontinent. Here we show that the present geochronological, structural and paleomagnetic data support such a Vaalbara model at least as far back as 3.1 Ga, and possibly further back to 3.6 Ga.
Garnet plagiogranitoids occur in the northeastern part of the Laplandian Granulite Belt. This rock was crystallised from a melt. This is confirmed by findines of magmatic breccias with xenoliths of basic rocks, enderbites, GRT-BT gneisses, OPX-dioritogneisses. The plagiogranitoids have magmatic hypidiomorphic-granular structures. Granulation takes place only on the rim of large grains of plagioclase. GRT and BT are metamorphic minerals. TPF program (IEM, Tchernogolovka) was used for the termobarometry.
T and P estimated for metamorphic GRT-BT paragenesis in garnet plagiogranitoids, acid granulites and rocks of xenolithes are similar (661°-729° for plagiogranitoids, 617°-686° for GRT-BT gneisses of xenoliths and 706°-712° for acid granulites). Magmatic structures, in the one hand, and the granulitic T and P of the formation of GRT-BT paragenesis, in the other, testify to the synmetamorphic (syntectonic) crystallization of garnet plagiogranitoids under condition of the granulite facies metamorphism.
Petrochemical data do not contradict the possibility for the plagiogranitoids to originate by diatexis of acid granulites. The geochemical similarity of plagiogranitoids and acid granulites is obvious. All these rocks distinctly differ from each other by FeO, MnO, MgO, Na2O, P2O5, Y contents and index of total femicity F. These distinguishing features are typical for processes of palingenesis. This allows a conclusion to be made that garnet plagiogranitoids crystallised as aresult of melting of host rocks - acid granulites.
The absence of stratification in the north part of the Lapland Granulite Belt was confirmed by the data obtained by SC EGGI from CDP vibroseis survey along two extensive profiles near the Russian Finnish border, and it is connected with palingenesis processes.
This data are consistent with the models implying that these rocks were derived from acid granulites in the upper part of the Lapland Granulite Belt as a result of gravity tectonics and the related to E-W extension at the final period of thrusting. This deformation stage was characterised by persistently high temperatures and increasing water activity. The regressive zoning in garnets in the plagiogranitoids shows that T decreased from 672° (in the center of garnet grains) to 592° (on rims garnet grains) in the process of the crystallization. Such zoning is absent in garnets from xenoliths.
Index of EUG 10 Volume
Further EUG 10 Information
Index of the Journal of Conference Abstracts
Cambridge Publications Home Page
Last Updated on Wednesday, March 17, 1999.
© 1997 Cambridge Publications