The specific case of the Mid-Proterozoic rapakivi granites and associated suite within the context of the Columbia supercontinent

Precambrian Research 137 (2005) 1–34

The specific case of the Mid-Proterozoic rapakivi granites and associated suite within the context of the Columbia supercontinent J.L. Vigneresse ∗ CREGU, UMR CNRS 7566 G2R, BP 23, F-54501 Vandoeuvre Cedex, France Received 20 May 2003; accepted 17 January 2005

Abstract Rapakivi granites and the associated suite (anorthosite, mangerite and charnockite) that developed during the Mid-Proterozoic escape the global rules that control usual granitic magma formation. About 40 points that characterise the Proterozoic magmatism confirm the originality of the magmatism. Rapakivi granites developed without a global orogenic context within the supercontinent Columbia. The intrusions extend from the East European shield to the western US, through Fennoscandia, Greenland and Labrador. Other occurrences in the Amazonian shield, Australia or South China are less well documented. They show no definite trend in age or chemistry that would explain large-scale (mantle plume) effects. A mantle upwelling of material is contradictory with the re-assembly of a supercontinent because it occurred before this magmatic episode (1.9–1.8 Ga). At ∼1.7 Ga, the supercontinent Columbia amalgamated with the collision of the Yapavai and Mazatzal Provinces. A model is suggested that takes into account the supercontinent re-assembly, defining a downwelling flow in the mantle that anchors the continent above it. In contrast to a material flow, heat is still delivered to the base of the continental lithosphere, and is focused toward the juvenile suture zone. The base of the crust must reach 1200–1300 ◦ C before producing anorthosite magmas. Under such a high temperature gradient lasting over a long time, magmas are transferred toward the upper crust, giving the thin (5 km) and square-shape (100 km) that the intrusions presently have. Heat delivery is essentially conductive, leading to long time-spans for intrusions. The presence of the supercontinent, immobile over a descending cell (poloidal mode of convection) developed a tangential force (toroidal mode of convection) that partly split the continent through strike–slip deformation due to plate rotation. It developed progressively between 1.57 and 1.3 Ga, starting from Fennoscandia, and then passing to Amazonia, western US and Labrador in a clockwise sense. The associated rotation induced sinistral shear manifested by small-scale shear zones and the orientation of late magmatic facies (topaz-bearing granites) in each province. The Proterozoic magmatism appears to be unique because it requires a supercontinent with a zone of juvenile crust surrounded by older cratons. The present Moho still shows remnants of this process, having bumpy undulations that may reach 22 km in amplitude over a distance of 200 km. © 2005 Elsevier B.V. All rights reserved. Keywords: Columbia; Mid-Proterozoic; Rapakivi granite; AMCG suite; Mantle convection

∗

Tel.: +33 3 8368 4737; fax: +33 3 8368 4701. E-mail address: [email protected].

0301-9268/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.precamres.2005.01.001

2

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

1. Introduction Granites have been observed in many tectonic environments ranging from extensional to contractional, within plates or at plate margins (Pitcher, 1993). A broad classification distinguishes between I- and Stypes granites, depending on the source material they involve (White and Chappell, 1988). A third type, or A-type, has been added (Collins et al., 1982) that considers peralkaline granites. This classification remains in common use, although present models of granite generation are based on different premises related to the source (Thompson, 1982; Clemens, 2003).

Rapakivi granites constitute a specific case of Atype granites, ranging from quartz syenite to peralkaline granites (Collins et al., 1982). The rapakivi texture, as initially described (Sederholm, 1891), typically consists of ovoid megacrysts of alkali-feldspar mantled with a thin shell of oligoclase. Rapakivi granites are commonly part of a suite of variously evolved magmas. The suite is alkaline to peralkaline in character. Rock composition extends from anorthosite up to peraluminous granites, defining a broad range (Fig. 1). The range attests that such magmas require a large mantle component. Their chemical composition falls into the field of within-plate magmatism in diagrams relat-

Fig. 1. Map showing a reconstruction of the respective positions of Laurentia, Fennoscandia, Ukrainian and Amazonian shields between 1.8 and 1.3 Ga. The model is redrawn from a Columbia reconstruction (Zhao et al., 2002). Former orogenic belts that fit together are in dark grey. Archean cratons are indicated by dark grey. Sedimentary basins from that period (1.7–1.3 Ga) are with black and white and italic names aligned horizontally. Letters refer to shields: Siberia, Greenland, Fennoscandia, east European, Amazonia, North America, Australia, Antarctica, Tasmania, South Africa, South China, North China, India, Madagascar, Yavapai–Mazatzal and West Africa. Rapakivi granites are indicated with ages. Names of rapakivi granites are aligned vertically, with ages.

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

ing granite composition to a tectonic setting (Pearce et al., 1984). They are commonly referred to as “anorogenic” granites. Another definition suggests dealing with anorthosite, mangerite, charnockite and rapakivi granites (AMCG) magmas (Emslie et al., 1994; Duchesne et al., 1999). The AMCG suite appears specific to Mid-Proterozoic times (R¨am¨o and Haapala, 1995). They have ages (1.7–1.3 Ga) that cannot be related to any specific period of crustal collision or largescale orogeny. Hence, in the Fennoscandian shield, rapakivi granites were emplaced during a very specific period of time (1.65–1.5 Ga) about 200 Ma after the major thermal episode of the Svecofennian orogen (1.9–1.8 Ga). Nevertheless, rapakivi granites and their associated suite are not restricted to the MidProterozoic. Anorthosites are found around 900 Ma, after the amalgamation of Rodinia (Emslie, 1978), but they are generally restricted to the Proterozoic. Hence, the rapakivi texture is not restricted to a specific period of time, and it has been described in Phanerozoic granites, as young to Tertiary (R¨am¨o and Haapala, 1995). Rapakivi granites also fail to accommodate any standard model of heat delivery to the crust. Indeed, their time-scale does not correspond to any common timespan for other magmas. Their bulk emplacement takes place about 200–400 Ma after peak metamorphism, and a single pluton presents intrinsic ages ranging from 20 to 60 Ma for emplacement. The following points have been identified that characterise the rapakivi granites and their magmatic suite:

(8) (9) (10) (11) (12) (13) (14)

(15) (16) (17) (18) (19) (20) (21) (22)

(1) large intrusive areas (>7000 km × 1200 km), encompassing several ancient shields and younger crust; (2) the magmas present a wide range of composition, with a large mantle (∼50%) component; (3) the granites are commonly associated with anorthosite, mangerite and charnockite, defining an AMCG suite; (4) the intrusions remobilize mainly young juvenile crust (residence time about 200 Ma), and little Archean material; (5) all intrusions develop over a long time (from 200 to 400 Ma) after the peak of convergence; (6) no polarity in age is clearly reflected in the distribution of dates within a province; (7) the average intrusion ages are 1.57 Ga in Fennoscandia, 1.54 Ga in Amazonia, 1.48 Ga in

(23)

(24) (25) (26) (27)

(28)

3

middle US, 1.43 Ga in western US and 1.33 Ga in Labrador; the massifs are generally large, up to 100–200 km in diameter; no preferred shape, or grain, for individual massifs is evidenced in most provinces; finally, more evolved facies (topaz bearing granite) present a more or less elliptical shape; the massifs are very thin (3–8 km, as a maximum); a large spacing (100–150 km) exists between individual plutons; successive intrusions that build up a single massif last a long time (20–60 Ma); incoming magma batches are less and less evolved (gabbro-norite to biotite-granite) over time; the interval between magma pulses that form a single pluton is about 5–8 Ma; the ascent proceeds whilst the magma retains a large (up to 40%) crystal content; the pressure and temperature conditions reflect isothermal decompression; granitic magma intrusion takes place at 450–650 ◦ C; the surrounding crust is generally hot (550–700 ◦ C) in the amphibolitic facies; contact metamorphism is either lacking or poorly developed around plutons; the crust remains hot (>500 ◦ C for resetting Ar–Ar ages) during a long period (1.6–1.4 Ga); cooling finally develops over a very short time interval close to 1.4 Ga; the formation of anorthosite is experimentally constrained to 1.1–1.3 GPa, or 30–36 km, the base of the crust; the temperature of anorthosite formation is about 1300 ◦ C; rapakivi textures are generated by isothermal ascent from 600 MPa or 15 km; temperature estimates for granite formation are 780–720 ◦ C; whilst the AMCG suite is restricted to the Proterozoic, rapakivi textures occur at any time; bimodal magmatism with ilmenite-bearing magmas in the East (Fennoscandia, Europe, Labrador, Amazonia) whereas magnetite-bearing granites are common in the western US;

4

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

(29) large-scale tectonic context of anorogenic granites is poorly constrained when they are present; (30) when present, ductile shear zones are identified; (31) most identified shear zones are dextral in their geometry; (32) the obliquity between the elliptical last facies and local shear zones reflects a global sinistral strike–slip deformation, the local shear zones being only conjugate with the main shear; (33) the sinistral deformation progressively affects all provinces in a clockwise sense from Fennoscandia to Labrador; (34) sediments that would attest erosion or mantle swell generally lack; (35) the sedimentation, when existing, reflects internal basin, with few marine connections; (36) sedimentary series close to intrusions developed high-grade metamorphism (amphibolite to granulite facies); (37) sedimentary series far from the intrusions are cold (unmetamorphosed or greenschist facies); (38) both types of Proterozoic sedimentary series are commonly separated by a piece of Archean crust; (39) the convergence of continents lead to the formation of Columbia at about 1.9–1.8 Ga; (40) there is no apparent large motion in latitude of Fennoscandia and probably Laurentia between 1.7 and 1.2 Ga; (41) both Laurentia and Fennoscandia remain within about ±10◦ from the Equator during 500 Ma; (42) along both sides of the Grenville Front, Columbia never get really dismembered before Rodinia reassembly. The impressive number of specific points makes the rapakivi granites and their associated suite different from other magmatic events. They reinforce the idea of a specific episode of magmatism in the context of supercontinent re-assembly and fragmentation. The present paper considers the Mid-Proterozoic “anorogenic” magmatism, including the so-called rapakivi granites. A tentative interpretation is attempted for those magmas generated during the interval 1.6–1.3 Ga, corresponding to the time when plates were amalgamated into the supercontinent Columbia (Zhao et al., 2002). It necessarily requires examining the global context of a supercontinent and its effects on mantle convection. The paper is organised as follows.

A long introduction identifies the specific characters of the AMCG suite and reviews the former models for their heat source. Then, rapakivi granites are described more in detail, within their tectonic and sedimentary environments, in the framework of a supercontinent. Their generation obviously requires a mantle component, and their emplacement requires a long-lived heat source. Their setting within a supercontinent amalgamation also points to mantle convection. An antiplume model is finally presented, and discussed.

2. The heat source problem as a former model The above list that documents generation of rapakivi granites and their associated suite contains numerous contentious points. They gave rise to various interpretations. The major point of debates relates to the time scale of magma generation and emplacement, which indirectly poses the problem of the heat source. A very long-lasting cooking is obviously required before the magmatic suite is produced. Four major hypotheses are briefly reviewed, with their major points of contradiction. 2.1. The plume hypothesis The plume hypothesis has been formulated to account for the widespread magmatism in Fennoscandia and Laurentia. It is essentially based on a large mantle component reflected in the magma suite (Anderson and Bender, 1989). Those magmas have a high K and Fe content, associated with a reduced environment as displayed by low fO2 and fH2 O . Their slightly negative εNd values also attest to a small residence time within their protolith. Thus, just after the main episode of crustal formation, mantle-derived tholeiitic melts pond at the base of the crust, continuously brought in by asthenosphere upwelling (Frost and Frost, 1997). It induces a rifting episode in the crust, which produces extension. The continuous magma underplating induces partial melting of the lower crust that combines with magmas derived from fractionation of the underplated mafic magma. The resulting granitic melts are very hot with a low water activity. Further ascent, certainly by a diapiric mechanism, and hybridization between the granitic and mafic melts, leads to the formation of the large rapakivi batholiths (Frost and Frost, 1997). The

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

model is supported by the widespread occurrence of bodies with similar magma types and similar age ranges (Anderson and Bender, 1989). Nevertheless, the model of an underlying upwelling of the asthenosphere faces several points of contradiction. First, the region where the upwelling develops is quite large. In the suggested model, it ranges from the Southwestern US to Labrador. But, the same ages and types of magmatism are recognised in the Baltic shield, the NE European shield (Latvia, Poland, Ukraine), and also in the Amazonian craton. Similar structures are also observed in the South Australian craton (Karlstrom et al., 2001). Second, the existence of a plume would affect all terranes, and not specifically the just-accreted juvenile crust. Third, an ascending plume should be associated with the delivery of a voluminous amount of mantle material producing radial mafic dykes. For instance the Matachewan dike swarm, at 2.45 Ga, extends radially over more than 800 km in Canada (Halls, 1991). Fourth, the plume model supposes a deep origin for the plume, hence more or less fixed to a mantle position (Smith and Lewis, 1999). The continent moving above it should reflect a long trace, as for instance the hotspot tracks that presently end in Hawaii or in the Tuamotu Islands. The track is about 4000 km long, formed during about 200 Ma of plate drift over it. The age polarity of the intrusions along the track confirms the relative displacement. This is not observed in Laurentia–Fennoscandia during the Mesoproterozoic. Conversely, the paleo-positions of the supercontinent document a stationary position, or a slow drift along the Equator. In contrast, the plume hypothesis assumes a moving plume. The paleo-position data hardly fit any present observation of hotspot motion, which forms a stable reference frame for lithospheric motion (Morgan, 1971; Duncan and Richards, 1991). 2.2. Crustal overturn This model initiates from the observation that Phanerozoic and Proterozoic tectonic and magmatic episodes are not analogous to each other (Anderson and Bender, 1989). In particular, the magmatic episode during the Proterozoic involves a very young crust, recently formed (1.95–1.65 Ga). The less differentiated juvenile crust certainly presents lower melt productivity. It differs from a Phanerozoic orogen, in which the crust had time for internal differentiation, and thus

5

would more effectively produce magmas. In consequence, a thermal anomaly, resulting from mantle material upwelling or magma underplating would induce a profound disturbance in the lower crust, with formation and separation of potassic granitic magmas (Foley et al., 1987). The resulting disturbance later manifests itself by repeated intrusions of anorthosite and mafic related magmas, along with granitic magmas. The soformed magmas would explain the positive to slightly negative values of isotopic εNd . Because the model is restricted to a recent crust, it would have no analogue in most recent (Phanerozoic) crust, already strongly differentiated. However, anorthosite and rapakivi granitic massifs have been observed intruding at different times (1.7, 1.4 and 0.9 Ga) at least for this range of time. Effectively, the oldest and the most recent magmatic episodes are not associated with huge volumes of magma that span several hundred of million years. The widespread occurrence of such magmatism appears specific to the Mid-Proterozoic. In addition, the large volume of igneous intrusions contrasts with more recent typical anorogenic granites. Despite the fact that the Proterozoic magmatism seems unique, in both composition and duration, the crustal overturn model involves mantle superswells, in which the gravitational rise of solid undepleted mantle results in the production of high-Al basalt. The basalt later fractionates to produce anorthosite, and by interaction with lower crustal material produces granitic melts. The model sounds attractive on petrological grounds, but requires at least several hotspots to account for the various age episodes. 2.3. Thickened crust or crust in extension The formation of the rapakivi granites and the associated suite of A-type granite are generally considered as typical of anorogenic periods. They are not associated with major episodes of large-scale deformation that would mark any model of plate convergence. However, many models invoke a plume hypothesis, or some asthenosphere upwelling that would explain the magmatic character of the rocks such as rapakivi or fayalite-bearing granites. Such rocks are observed in the Yellowstone and Snake River series, attesting to the presence of an underlying hotspot. In consequence of that model, extension is often referred to

6

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

Fig. 2. Diagram similar to a Flinn diagram showing the ratio of the length/width (L/W) as a function of the ratio width/thickness (W/T) of various plutons, as deduced from gravity data inversion (Am´eglio et ., 1997). Circles reflect wedge-shaped plutons, and squares flatfloored massifs. Also plotted are curves for laccoliths and plutons deduced from 2D models (Petford et al., 2000). Rapakivi granites (Wiborg, Ahvenisto, both in Finland; Nordingr˚a, Sweden; Riga in Latvia; Korosten, Ukraine; Grahh Fjelde, Paatusoq and Qernertoq, southern Greenland) are shown by black boxes.

as being associated to the mantle upwelling. Another argument, often quoted for extension, is the occurrence of mafic dykes that would prove extensional environments. However, their restricted volume does not seem to reflect a large-scale extensional province (Anderson and Bender, 1989). Nevertheless, direct evidence for extension being associated with granite emplacement is generally lacking. The only evidence for a direct connection between emplacement and extensional shear faults is documented in Greenland (Hutton et al., 1990; Grocott et al., 1999). However, the extension model pertains more to thinsheet intrusions, with a thickness that rarely reaches 2 km, than regular plutons. Comparing their shape with other massifs (Fig. 2), the rapakivi granites are characterised by a very large width/thickness (W/T) ratio, whereas the length/thickness (L/W) ratio remains close to 1, indicating the anisotropic character of the crust at the time of emplacement (Hogan et al., 1998; Vigneresse et al., 1999). In contrast to those models showing a thin shape, the expected internal shape of the plutons emplaced during extension should be a vertical cylinder (Vigneresse et al., 1999). Situations have also been described that require a thickened crust to solve the heat problem (Thompson, 1999). The same as for the extension hypothesis, those models assume a thick crust, but fail in describing either how thick it was or how it formed. A thicker crust is commonly associated with a change in elevation, the

latter being soon restored by erosion. Thus, sedimentary basins that would record the material eroded from the mountains should form close to, or not too far from the region where the crust had been thickened. This supposes, when dealing with the huge region of southern Laurentia and Fennoscandia that some more or less linear mountain chain had been developed. No trace of such elevated regions presently exists, nor a mechanism that would have lead to mountain formation, especially during an “anorogenic” period. Thickening and subsequent lithospheric thinning have also been suggested (Loosveld, 1989), but again lacks the global context that would correspond to the observed tectonics in the whole region. The necessity to have feldspar involved in the reactions leading to anorthosites and rapakivi granites also precludes a thick crust, because feldspar is not stable below 40 km. Regional shortening has also been suggested for the Signal batholith, Arizona. The 1.42 Ga granite is emplaced within the Yavapai Province, Arizona, USA (Nyman and Karlstrom, 1997). The province is thrust against the southern Mazatzal province (1.8–1.6 Ga), as documented by seismic profiles (Tyson et al., 2002). It is not the only massif that appears to be emplaced during contraction at that time (Nyman et al., 1994; Duebendorfer and Christensen, 1995; Kirby et al., 1995). However, those massifs are quite late compared with the rapakivi granites sensu stricto in this region. Nevertheless, a period of global transpression seems to have existed at that time (Nyman et al., 1994). 2.4. Crustal tongue Post-collisional magmatic manifestations have also been invoked that would lead to a various types of magma production (Duchesne et al., 1999). In this model of crustal tongue melting, a piece of the lower mafic crust is pinched and underthrust at sufficiently high pressure (1.1–1.3 GPa) to reach the stability field of aluminous orthopyroxene. A temperature increase to 1200 ◦ C would lead to ferrodioritic magmas, whereas high-Al basalts would be produced at 1300 ◦ C. Restitic granulites from the generalised anatexis could result in classical A-type granitic melts (Clemens et al., 1986; Landenberger and Collins, 1996). Though elegant, this model ignores a long timedelay between the peak of convergence and the magma intrusion. Magma ascent through buoyancy and viscos-

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

ity contrast appears contradictory with a crust submitted to severe contraction. In that case, tectonic driven forces would lead to rapid magma ascent, of the order of some thousands of years (Petford et al., 2000). Any model of tectonically active crust intrinsically rules out models of diapiric ascent (Vigneresse and Clemens, 2000).

3. Rapakivi granites and anorogenic magmas Rapakivi granites are found in many countries, extending along 12,000 km, and intrusive in the same interval of time (1.6–1.3 Ga). This points to external conditions of formation that would certainly also be manifest through tectonic and sedimentary environments. All should be emplaced within the timeframe of a supercontinent. 3.1. Regions and ages of intrusion In Fennoscandia, the region where they were first described, rapakivi granites (Fig. 1) form an east–west trending belt about 600 km long, from Russia to Sweden. They extend from the Salmi massif, close to the Lake Ladoga, north of Saint Petersburg in Russia (Amelin et al., 1997), through Wiborg, about 100 km east of Helsinki, Finland (R¨am¨o and Haapala, 1995; ˚ Vaasjoki, 1996; Elliott, 2001), to the Aland archipelago in the Bothnian Sea (R¨am¨o and Haapala, 1995; Shebanov and Eklund, 1997) before cropping out in Ragunda and Nordingr˚a, central Sweden (Andersson, 1991; Persson, 1999; Lindh et al., 2001). However, they also occur within a north–south trending belt, when considering the massifs from Ukraine to Finland. The oldest intrusion is the Korosten massif, Ukraine (Amelin et al., 1997; Esipchuk and Skobelev, 1998), dated at 1.77–176 Ga (Sivoronov et al., 1998). The Mazury complex, Poland (Wiszniewska et al., 2002; D¨orr et al., 2002), is younger, dated at 1.55–1.50 Ga (D¨orr et al., 2002). It is close to the huge Riga batholith, Latvia (R¨am¨o et al., 1996) dated at 1.63–1.58 Ga. Many smaller satellite bodies (Undva, Taebla, Abja) located north–northeast of the main body, could imply connec˚ tion with the Wiborg and Aland batholiths in Finland (R¨am¨o et al., 1996; Andersson, 2001). The possible north–northeast and north–south trends demonstrates that granites are widespread over a large area, with-

7

out a clear zoning. Small granitic intrusions outcrop as satellites bodies of the larger intrusions. They are well developed in Fennoscandia with the massifs of Obbn¨as, Bodom (Kosunen et al., 2002), Onas, Ahvenisto and Suomenniemi, all associated with the large Wiborg massif. The massifs of Vehmaa, Laitila, Eurakoji and ˚ Peipohja are all associated with the Aland batholith (Eklund et al., 1998). Rapakivi granites develop in Sweden as smaller individual plutons, partly covered by subsurface sediments or sea (Andersson et al., 2002). The largest are the two massifs of Nordingr˚a and Ragunda, near the smaller massif of R¨od¨on (Andersson, 2001). Evidence of several ductile shear zones trending east–west over 200 km is documented by different lithologies on both sides (Torvela and Ehlers, 2002). Activation time is calibrated between 1.79 and 1.65 Ga, ˚ corresponding to the time of the Aland rapakivi granite emplacement. Intense dextral shearing is shown by the deflection of surrounding structures. A connection with granite emplacement has been suggested on the basis of the deflection in trend of several diabase dykes ˚ when approaching the major Aland massif (Ehlers and Ehlers, 1977). The same type of magmatism (Fig. 1) extends through southern Greenland (Dempster et al., 1991; Windley, 1991), Labrador (Ryan et al., 1998; Connelly and Ryan, 1999), and southern Laurentia (Emslie, 1978; Anderson et al., 1980; Anderson and Bender, 1989; Gower and Tucker, 1994; Karlstrom et al., 2001). Rapakivi granites are restricted to the Ketilidian orogen of South Greenland (Dempster et al., 1994; Garde et al., 2002). They develop at the southernmost area of Greenland, and are intrusive in either the psammite or pelite zones between 1755 and 1732 Ma (Brown et al., 2002). They form small massifs (Qernetoq, Graah Fjelde, Lindenow Fjord) that intrude an anisotropic pelitic or amphibolitic series (Grocott et al., 1999). They are marked by contrasting black and white facies, depending on the colour displayed by the alkali-feldspar ovoids (Brown et al., 1999). The connection with the nearby Makkovik Province, Labrador is immediate (Kerr, 1989; Ketchum et al., 2002). Small massifs dated at 1.75–1.71 Ga (Ketchum et al., 2001) are intensively sheared and lacerated between shear zones (Kaipokok, Kanairiktok), as in Greenland. Internal structures in all magmatic sequences indicate the intense shear deformation that developed just before rapakivi granite intrusion (Ketchum et al., 2001; Garde et al., 2002). A late

8

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

transpressional event characterised by ENE trending dextral shear zones is coupled with the emplacement of the rapakivi granites (McCaffrey et al., 2004). Farther north in the Nain Province, Labrador, two suites of anorthosites and granites have been identified, which present the same orientation and appearance (Ryan et al., 1998). The former, or Arnanunat Plutonic Suite (APS) dates from the Paleoproterozoic (2.1–2.0 Ga) and has a large magnetite content. The latter, or Nain Plutonic Suite (NPS) is much younger (1350–1290 Ma) and bears ilmenite (Connelly and Ryan, 1999; Royse et al., 1999). Both types, APS and NPS, can easily be identified on magnetic maps. Local NW–SE dextral shear during granite emplacement is compatible with the NE–SW trending sinistral Ketilidian orogen as a whole (Chadwick et al., 2000). On the southeastern margin of Laurentia, an intense magmatic activity develops between 1610 and 1420 Ma (Rivers, 1997), but lasts up to 1220 Ma (R¨am¨o et al., 2003). It is marked by high-grade metamorphism, up to granulite, within the Central Gneiss Domain. The low pressure, high temperature metamorphic terrane is characterised by intrusion of the AMCG suite at several places, north of the present Grenville Front. Intrusions include Michikamau Lake (1450 Ma), Mistastin (1420 Ma) and Harp Lake (1450 Ma). They are mainly anorthosites, with similar ages to the rapakivi granites of the Wakeham Supergroup (1497 Ma), just before the rapakivi of the Nain Province (Rivers and Corrigan, 2000). This part of the Grenvillian thrust belt could have been attached to the Amazonian craton (Sadowski and Bettencourt, 1996). Farther to the southwest, in the central and western US, anorthosites and related granites (Fig. 1) extend from Wisconsin to California (Anderson and Bender, 1989), intruding the Mazatzal and Yavapai provinces. The two provinces have been separated by their formation age, respectively, 2.1–2.6 and 1.7–1.8 Ga, as well as by their isotopic signature (Bennett and DePaolo, 1987). The oldest intrusives include the 1.76 Ga Horse Creek anorthosite complex, Wyoming (Scoates and Chamberlain, 1997; Frost et al., 2000). They record events prior to the amalgamation of the provinces to Laurentia, dated between 1720 and 1682 Ma in the Cerbat Mountains, Arizona (Duebendorfer et al., 2001). The collision is confirmed by seismic lines in northern Colorado and dated at the same time (Tyson et al., 2002). Intrusives form a suite aged 1.48–1.43 Ga

that dominates the northeast-trending Transcontinental Proterozoic provinces of the US (Anderson and Bender, 1989; Van Schmus et al., 1993). The older massifs are as the Wolf River batholith in Wisconsin, 1.5 Ma (Anderson et al., 1980) and the St. Francois Mountains in Missouri dated at 1.48 Ma (Menuge et al., 2002). The major episode of magmatism occurred at 1.44–1.36 Ga with numerous intrusions of anorthosite and charnockite. They define three geographic and petrographic subprovinces in Colorado and Wyoming (Anderson and Cullers, 1999). The Laramie Anorthosite Complex (Fuhrmam et al., 1988; Scoates and Frost, 1996) and Sherman ilmenite-bearing granite (about 1.43 Ga) comprise the northernmost belts of these intrusive rocks in the Colorado province (Smith et al., 1999). Magnetite-bearing granites, including some with rapakivi textures, form a southern belt that extends from the central mid-continent to California. The subprovince includes the Beer Bottle Pass pluton, Nevada (Duebendorfer and Christensen, 1995), the San Isabel and Eolus batholiths (Noel and Andronicos, 2002) in Colorado. The last subprovince extends in central Arizona. It contains biotite or two micas granites, like the Signal batholith, dated at 1.42 Ga (Nyman and Karlstrom, 1997) or the Sandia pluton, 1.4 Ga in age (Kirby et al., 1995) both in Arizona. In the Amazonian craton, two major provinces (Fig. 1) include rapakivi granites and anorthosite massifs that define two orogenic periodes. One is locally named Transamazonian, similar in age (1.7 Ga) to the Ketilidian in Greenland. The second, or Rondonian, develops farther south and is similar to the Elsonian orogen (1.5–1.3 Ga). Close to the Guyana Shield the Mucajai pluton is dated at 1544 Ma (Gaudette et al., 1996) and is associated with charnockites of the Kanuku complex. The Parguaza batholith, Venezuela (Gaudette et al., 1977; Miron Valdespino and Constanzo Alvarez, 1997), is a rapakivi granite (1545 Ma) that crops out at the extreme western part of the shield. Farther south, the Rondonia Tin Province also contains intrusions at 1.57–1.52 Ga that belong to a similar magmatic suite (Bettencourt et al., 1999; Payolla et al., 2002). This specific area, which can be considered as connected in a single continental landmass at that time (Karlstrom et al., 2001), covers about 7000 km × 1200 km. More than 70 massifs are considered, widespread over the whole area without any immediate chronological zoning. The spacing be-

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

9

Fig. 3. Map of the Salmi, Russia, intrusion, showing the various facies and age determinations (redrawn from Amelin et al., 1997). In the inset, ˚ various massifs from the Fennoscandian shield quoted in the text, with their inferred sub-surface bodies: Nordingr˚a, Ragunda, R¨od¨o, Aland, Laitila, Riga, Wiborg, B¨od¨om, Ahvenisto, Suomenniemi, Mazury and Korosten (see text).

tween massifs is quite irregular, but is of the order of 100–150 km. The total time-span of the intrusions ranges from 1.65 to 1.34 Ga (Fig. 1). In Australia and Antarctica, rapakivi granites and an anorthosite suite have been identified which have identical ages. Similarly, in the North China craton, rapakivi granites (Shachang massif) have an age of 1.7 Ga (R¨am¨o et al., 1995; Haapala et al., 2005). However, data on their age or chemical analyses are too sparse to include them in the present paper, as well as their respective position at that time. Major massifs, with their corresponding data on age, scale and thickness are listed (Table 1) and shown on the maps (Fig. 1 and partly Fig. 3). 3.2. Tectonic environments The tectonic environment that prevailed during the emplacement of rapakivi granites is still unclear. No global orogenic style can be attributed to this period, leading to the so-called “anorogenic” granites denomination. The period of intrusion postdates by 200–400 Ma the peak episode of plate convergence (1.9–1.8 Ga). Tectonic contexts as various as extension (Hutton et al., 1990) or contraction (Nyman and Karlstrom, 1997) have been advanced. At the time of

intrusion, the crust is assumed to be thick (Vorma, 1976), thin (Haapala and R¨am¨o, 1999), or thickened but actively thinning by extension (Loosveld, 1989) ˚ all et al., 2000). The large vaor distal orogenesis (Ah¨ riety in crustal or tectonic environments reflects often poorly constrained settings, or is based on general concepts, commonly unverified. For instance, the pressure constraints provided by experimental melting (Longhi et al., 1999) assume formation of crustal melts at about 1.1–1.3 GPa. The experimental data constrain the thickness of the crust to 30–36 km, reflecting a normal crust. The local tectonic context is also contentious. Apparently, no specific tectonic event of large-scale (orogenic) can be related to magmatic intrusions in Fennoscandia. Nevertheless, ductile dextral shear zones have been described in south Finland (Torvela and Ehlers, 2002) and east–west trending shear zones are documented in Latvia and Belarus (Taran and Bogdanova, 2001). They displace previous structures and are contemporaneous with the intrusion of the Riga batholith. In Greenland and Labrador, intense shear zones dismember all intrusions (John et al., 2002; McCaffrey et al., 2004). It remains difficult to assign a specific age to this intense shearing although it is certainly post-reworking. However, the dextral shear zones are

10

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

Table 1 Major batholiths in Laurentia, Fennoscandia, Amazonia and related provinces Country name

Size (km)

Russia Salmi

140 × 50

1.546–1.529

Finland Wiborg Ahvenisto Suomenniemi ˚ Aland Laitila Eurojaki Vehmaa

200 × 175 25 × 15 27 × 17 100 × 100 50 × 40 8×8 40 × 30

1.645–1.615 1.643–1.632 1.640–1.635 1.568–1.576 1.540–1.570

Sweden Nordingr˚a Ragunda R¨od¨o

150 × 100 50 × 20 37 × 20

5

Latvia Riga

250 × 225

1.630–1.580

Poland Mazury

155 × 150

Ukraine Korosten

125 × 100

0.5–3.0

Greenland Qernertoq Graah Fjelde Paatusoq

11 × 110.8 20 × 12 20 × 12

2 2

Labrador Makkovik Province Nain Province Mistatin Umiakovik Makhavinekh

3

1.720–1.715 1.35–1.29 1.318 1.322

10

References are quoted in the text.

1.77–1.76 1.755–1.732

100 × 75 110 × 45 40 × 10

North China craton Shachang

1.578 1.53–1.47 1.515–1.505

1.525–1.499

8–11

130 × 60 32 × 12

Age (Ga)

1.582–1.573

5×5

Western US WY WI MI NE NM AZ Amazonia Parguaza, Guyana Mucajai Rondonia Cachoerinha Serra de Providˆencia Pitinga

Thickness (km)

1.44–1.43 1.47 1.48 1.42–1.43 1.42 1.44–1.34 1.545 1.544 1.57–1.52 1.56–1.54 1.606–1.566 1.815–1.794 1.735–1.683

associated with granite emplacement (McCaffrey et al., 2004). There is growing evidence in the western US that tectonic events were coeval with intrusions. Several shear zones, the majority dextral, have been recognised that are associated with granite intrusions. They include the Homestake shear zone, Colorado (Shaw et al., 2001) dated at 1.45–1.38 Ga, the Cheyenne Belt, Wyoming (Karlstrom and Houston, 1984), the Beer Bottle Pass, South Nevada (Duebendorfer and Christensen, 1995) and the NE-trending shear zone associated with the Colorado mineral belt (Sims et al., 2002). All of these ductile structures trend roughly NE and are generally dextral. 3.3. Sedimentary constraints Another argument against a thick crust at this time concerns the poorly developed sedimentary basins that would have kept track of the erosion in the case of an elevated thickened crust. Proterozoic conditions of sedimentation have long been recognised as different from those acting during the Phanerozoic (Erikson et al., 2001; Aspler et al., 2001). Unequivocal aeolian deposits are rare prior to about 1.8 Ga, as well as foreshore deposits. The consequence of the latter observation is difficult to evaluate in terms of paleoenvironment, owing to the absence of pluricellular animal and plant fossils. Nevertheless, finding sedimentary sequences that would constrain the environment existing at 1.8–1.4 Ga is a hard task; only few have been identified (Fig. 1). Their restricted areas do not preclude a large erosion of a thickened crust at this time. The Belt-Purcell sedimentation Supergroup (Fig. 1) in the western USA covers about 130,000 km2 with a maximum thickness of 16 km (Hall and Veizer, 1996). The lower sediments are dated at 1470 Ma and the upper ones at about 1000 Ma (Frank et al., 1997; Lyons et al., 2000). The sediments include a basal sequence, with shales metamorphosed in the greenschist and amphibolite facies, carbonates and younger rocks, at about 920 Ma from dating of glauconite. The whole basin appears to have been unconnected with deep marine waters during most of the Mesoproterozoic. Four-kilometre thick sedimentation supergroup in Siberia is dated as middle to early Riphean (1727–860 Ma); the total area of the Uchur-Maya region, close to Yakutsk, west of the Aldan craton (1.9 Ga), comprises 100,000 km2 of sediment (Bartley

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

et al., 2001). The older sediments document shallow water, apparently with little connection to a deep ocean, but carbonates developed after 1.3 Ga in intermediate sequences. In central Canada (Fig. 1), several basins, as the Athabasca, Thelon and Hornby groups, may have been connected together (Thomas et al., 2000). The Athabasca Basin occupies nearly one-third of the surface area of the Saskatchewan Shield (Kyser and Kotzer, 1995). It comprises coarse fluvial to marine clastic sediments represented by quartz sandstones and conglomerates, and is dated at 1.7–1.6 Ga. The central part of the basin is about 1400 m thick, and mostly undeformed (Ross, 2000). The Vindhyan Supergroup outcrops over 110,000 km2 in Rajasthan, northern India, and is dated at 1600–710 Ma, with several hiatuses in the sedimentary sequence (Ray et al., 2003). It represents four major sequences that include sandstones, shales and carbonates with a few conglomeratic and volcanic beds. In Brazil (Fig. 1) the small Espinhac¸o basin, about 20,000 km2 , basin, contains about 4 km of sediment dated at 1710–900 Ma, if the lowermost sequence, the Archean Rio Parauna supergroup, is excluded (Martins-Neto, 2000). Sediments of the intermediate sequence, 1720–1500 Ma, are composed of sandstones and pelites of lacustrine and shallow marine environments. The Roraima Supergroup, also in Brazil, extends over 73,000 km2 and consists of sandstones deposited in deltaic and shallow marine environments (Santos et al., 2003). Other restricted basins are the Mount Isa and McArthur groups in Australia (Shen et al., 2002; Scott et al., 2002), the McNamara group and Solary Block (Bierlein, 1995), and the Bangemall Group, northwestern Australia (Buick et al., 1995). Subsequent geological developments mainly took place along the margins of the shield. During the later Proterozoic and throughout the Phanerozoic major sedimentary basins formed, notably in north and northeast Greenland (Independence Fjord Supergroup). In some places, accumulated sedimentary successions reach about 10–15 km in thickness (Henriksen et al., 2000). When these sedimentary basins are added together and shown on a global map (Fig. 1), they document a restricted volume. They consequently confirm that the erosion was not very drastic during the Mesoproterozoic, which rules out the presence of a high elevation terrane, hence of a thickened crust. A second point relates to the thermal state of the sedimentary formations. Sediments deposited prior to

11

1.4 Ga generally have a high metamorphic grade (amphibolite to granulite facies). This is observed in the few basins that have been preserved in the immediate vicinity of the areas of magmatic activity. In basins far from this zone, often located in regions separated from the magmatic province by a piece of Archean crust, as in Athabasca, Siberia or India (Kyser and Kotzer, 1995; Bartley et al., 2001; Ray et al., 2003), sediments are unmetamorphosed as reflected by the presence of sandstone and limestone. The simultaneous presence of low-grade and high-grade basins indicates that the thermal perturbation associated with the magmatism was restricted to the inner portion of the supercontinent. It corresponds to the area where juvenile crust was surrounded by Archean cratons, and in which the associated magmatic suite developed. It thus demonstrates a thermal anomaly that focused on the most recently formed crust. 3.4. Paleo-reconstruction A paleo-reconstruction of the continents at about 1.6–1.5 Ga must first be considered, but this is far beyond the scope of this paper. It supposes that continents were amalgamated at 2.1–1.8 Ga into a supercontinent, called Columbia (Zhao et al., 2002, 2004). Paleomagnetic data that would document the position of Laurentia, Fennoscandia and South America are not sufficiently reliable for this period (Buchan et al., 2001; Pesonen et al., 2003). A proxy for the supercontinent has been constructed by fitting former structures together (Zhao et al., 2002). The next supercontinent reconstruction is for Rodinia at about 1.0 Ga, but it implies entering the debate between SWEAT (Moores, 1991), AUSWUS (Karlstrom et al., 1999), or modified models (Powell and Pisarevsky, 2002); these are the major plate reconstructions for the Rodinia supercontinent (see review in Piper, 2000; Kah and Bartley, 2001 and Meert and Powell, 2001). The models basically differ in the paleo-latitude of the former Australia–Antarctica block, on the eastern border of Laurentia. They globally agree on the relative position of Laurentia and Fennoscandia. South of the Laurentian margin, the Amazonian craton is supposed to fit a place that is adjacent or not to the western margin of the Ukrainian shield. In consequence and despite the uncertainty that underpins the reconstruction, a model of the Columbia

12

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

supercontinent (Zhao et al., 2002) is adopted (Fig. 1). The differences between Columbia and Rodinia appear very subtle (compare Zhao et al., 2002 and Karlstrom et al., 1999). The resemblance suggests that the time of 800 Ma spent after aggregation did not lead to a huge splitting and divergence amongst Laurentia and Fennoscandia. Probably, the most effective separation took place along the southern margin of Laurentia, leaving Amazonia and West Africa apart, and giving place to the Grenville Front during plate reassembly. The respective positions of smaller blocks (China, Siberia, India), as well as the position of the Australia–Antarctica block are not sufficiently constrained at the moment. By chance, they are of less importance in the magmatic history at those times. The paleomagnetic data-set for Laurentia and Fennoscandia offers not enough confidence for that period (Buchan et al., 2001). Many former pole determinations lack a precise age, owing to the poor reliability of Rb/Sr, if not K/Ar, ages (Piper and Stearn, 1977; Piper, 1980). There is only one reliable paleomagnetic pole at 1.5 Ga, for Laurentia between 1.75 and 1.3 Ga, which is not enough to constrain its position. More generally, there are no paleomagnetic poles between 1.5 and 1.25 Ga. Furthermore, poles from both plates must coincide with time. Finally, the amplitude of the true polar wander seems variable at these periods, which adds uncertainty, about 10◦ , to the dataset (Elston et al., 2002). Nevertheless, both Laurentia and Fennoscandia were very close to the paleo-equator (Elming et al., 2001; Powell and Pisarevsky, 2002). A large distance in longitude may separate two blocks at the same latitude because only latitude positions can be inferred from the paleomagnetic method. Block rotation around a vertical axis allows separation of the relative motion between blocks, but it requires having a close sampling in time, which is presently rarely the case. At 2.0 Ga, the Wyoming and Superior Provinces were at about 40◦ N, drifting to the south (Harlan et al., 2003). During the 1.93–1.78 Ga period, the Fennoscandian block remained at about 30◦ N in latitude (Mertanen and Pesonen, 1997). It drifted slowly southward to reach 10◦ S at 1.56 Ga (Pesonen and Mertanen, 2002). During the first period of time, the Ukrainian block was at about 5◦ N, just below Fennoscandia. It also slowly drifted to the south, but may be detached by about 5◦ from Fennoscandia at 1.74 Ga. During the time of the

rapakivi granite intrusion, the position of Fennoscandia was restricted to a fairly narrow latitudinal range between 0◦ and 27◦ N (Moakhar and Elming, 2000). At 1.5 Ga both blocks were either attached, or very close to each other (Buchan et al., 1993; Elming et al., 2001). At 1.5 Ga, Laurentia and Fennoscandia were probably attached together, because Laurentia was also at 5◦ N, close to the Equator (Elston et al., 2002). Later on, reliable poles for Fennoscandia indicate a position very close to the Equator, from 1.4 Ga up to about 1.2 Ga. At 1265 Ma, poles from both Fennoscandia and Laurentia allow a reconstruction. The match is consistent with Fennoscandia adjacent to present-day Greenland although their respective orientations are still a matter of debate (Pesonen and Mertanen, 2002). In this reconstruction, the 1.7–1.5 Ga Gothian and Labradorian belts are aligned. After that time Laurentia drifted northward. A very recent paper (Meert and Stuckey, 2002) traces back to 1.476 Ga the paleomagnetic pole for Laurentia. Poles from Laurentia, Fennoscandia, Siberia and Australia are similar, leading to a compatible reconstruction of a single continent, with Australia in the high latitude position (SWEAT model). In all reconstructions the regions where rapakivi granites were intrusive cluster in the interior of the supercontinent. Basins, with marine connections, sediments that show marine affinities as precursors of stromatolites as “Molar-tooth” structures (Frank and Lyons, 1998) are located in small areas that define the external border of the supercontinent. They constrain the geometry of the reconstruction.

4. Mantle component of the magma The specific character of the AMCG points to a mantle involvement on both geochemical and isotopic grounds. 4.1. Global composition Intrusions range from anorthosite to rapakivi granite with compositions ranging from ultramafic, calcalkaline, and apatite-rich lamprophyres to peraluminous granites. Roughly, they form a suite in which K2 O + Na2 O > 5.0, with K2 O/Na2 O > 0.5, and with SiO2 ranging from 32 to 78%. The ratio FeO/(FeO + MgO) is also commonly high, up to one

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

in Laurentian granites. Magmas are strongly enriched in LILE and HFSE (Eklund et al., 1998). They are also typically high in incompatible elements such as Zr, Nb, Ta, Th, including REE, but they are low in Co, Sc, Cr, Ni, Ba and Sr. Obviously rapakivi granites and their associated suite need a high mantle component to explain their chemical and isotopic composition. This, associated with low H2 O fugacity (fH2 O ) and oftenlow O2 fugacity (fO2 ), suggests that their minimum temperature of formation was elevated above 900 ◦ C (Clemens et al., 1986; Creaser et al., 1991). However, whereas the intrinsic origin of A-type granites is almost understood, their association with granitoids ranging from metaluminous to peraluminous to peralkalic is more contentious. Some peraluminous granites are very similar to S-type granites, even including late, evolved, topaz-bearing stocks such as Ahvenisto, Eurajoki and Kimy, all in Finland (Haapala, 1997; Alviola et al., 1999; Haapala and Lukkari, 2005) and Pitinga, Brazil (Lenharo et al., 2002) and tin mineralisation as Eurajoki at Finland (Haapala, 1997) and Sao Laurenco and Santa Clara, both in Brazil (Bettencourt et al., 1999). By contrast, many magnetite-bearing rapakivi granites bear much similarity with other I-type granites. 4.2. Anorthosite suite Problems dealing with the origin of rapakivi granites should also be examined in conjunction with those related to anorthosite magma production, and more generally within the AMCG suite (Duchesne et al., 1999; B´edard, 2001). The huge volume of such magmas appears specific to the Proterozoic (Ashwal, 1993). It poses the problem of source magmas that would account for the generation of extensive volumes of feldspar cumulates. The parental magmas could be mantle-derived, tholeiitic or high-Al basalts. Fractionation and cumulate flotation would lead to feldspar concentration. Another possibility favours a lower crustal source, already enriched in Al, with feldsparrich cumulates and after extensive anatexis leading to aluminium-rich cumulates. In both cases, problems still occur that link anorthosite magmas to cumulates, or to charnockites (B´edard, 2001). Recent melting experiments provide constraints on the region of production of such magmas, setting the stability field for aluminous orthopyroxene to

13

1.1–1.3 GPa (Longhi et al., 1999). Temperature conditions are assumed to range between 1300 and 1200 ◦ C. Experimental pressure and temperature conditions are constrained by field observations that document ubiquitous presence of orthopyroxene megacrysts in many anorthosite massifs. Orthopyroxene crystals, up to 1 m long, in anorthosite massifs, define magma flow during emplacement (see Fig. 2 in Barnichon et al., 1999). Therefore, they constrain the conditions of formation of anorthosites and associated suite to the lowermost crust (30–36 km). 4.3. Ilmenite-bearing versus magnetite-bearing granites The separation between the two types of magnetic carriers, hence in the amplitude of magnetic susceptibility, is regionally related. European granites, as well as those from the Nain province and Amazonian Shield contain ilmenite. Conversely, granites of the younger Labrador province and the western US contain magnetite and biotite-hornblende, with high magnetic susceptibility. Ilmenite-bearing rapakivi granites occur in Ferroscandian provinces, Poland, the Ukraine and in Labrador. They have high K2 O contents (up to 6%), higher than Na2 O, with high FeO/(FeO + MgO), commonly above 0.8. They are the most distinctive ironrich, but reduced granites, with low fO2 and fH2 O values (Anderson and Bender, 1989). Their high Fe content, along with the enrichment in incompatible elements, is likely to result from an extreme differentiation in a reducing environment (Frost and Frost, 1997). Widespread occurrence of fayalite within these granites suggests a high solidus (around 1000 ◦ C) temperature (Creaser et al., 1991). An origin through partial melting of tholeiitic basalts and later differentiation in an extensional environment has been suggested (Frost and Frost, 1997). This would explain the character of the magma, but also the occurrence of various suites from gabbro to ferrodiorite, including cumulate anorthosite and granites. In contrast, magnetite-bearing granites, as in western US, are less potassic, with lower content of LILE. Their fO2 values are accordingly higher, because of their oxidized character. This hydrous calc-alkaline composition also imposes higher water activity. Thus, reflects melting at lower temperature (above 900 ◦ C),

14

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

which also fits with the close occurrence of peraluminous granites. The later have very high oxygen isotopic ratios, with δ18 O commonly above 10‰ (Anderson and Morrison, 1998). 4.4. Isotopic data The isotopic signature of the AMCG suite generally agrees with a short residence time (R¨am¨o et al., 2001). An average εNd for the Finnish rapakivi granites is −1.39 ± 0.76 (R¨am¨o et al., 2001). These magmas result from a small degree of melting of deep crustal material having a 15% lower Sm/Nd value than the protolith that formed at a maximum of 2.0–2.2 Ga. The source material would be similar to the one that gave place to the 1.84–1.81 Ga post-kinematic granites in the same region. Lead isotopes confirm their derivation from young crust (Vaasjoki, 1996, 1997). Though a large mantle-derived component is obvious from the εNd values, this is more difficult to assess when dealing with Sr isotopes. Sr preferentially partitions into plagioclase. Nevertheless, the specific conditions of crystallisation leading to rapakivi textures produce Na-mantled K-feldspars. As a result, rapakivi granites are not the best-suited rocks for Rb/Sr analysis (Dempster et al., 1991; Patel et al., 1999). Rb/Sr age determinations using biotite yield ages systematically lower by about 200–250 Ma as compared to zircon ages. The example of the small R¨od¨o massif in Sweden is indicative of such discrepancy (Andersson et al., 2002). The massif is about 20 km in diameter at the present surface level of erosion. It should cool in a short time of the order of 20 ky (Vigneresse and Am´eglio, 1999). Even if it was doubled in surface area, the cooling time remains restricted to some tens of ky. A Sm/Nd age determination indicates an age of 1608 ± 45 Ma, whereas Rb/Sr ages are 1340 ± 39 Ma (Andersson et al., 2002). The closure temperature of biotite, at about 300 ◦ C, demonstrates that the cooling was very slow, occurring over about 300 Ma. This is about three orders of magnitude larger than the expected cooling time. In consequence, Rb/Sr ages in rapakivi granites should be treated with caution (Dempster et al., 1991). Hence, initial isotopic Sr values are reported in the literature that can be compared with initial εNd values (Jahn et al., 2000). When plotted in a diagram displaying the range of εNd versus Isr values, the respective fields and trends reflect the magma source conditions

Fig. 4. Diagram showing compiled analyses of granitoids from various regions (redrawn and adapted from Jahn et al., 2000). Her, Hercynian; Cal, Caledonian; Chi, S and N China; Aus, Lachlan Fold belt and New England; Cor, North America Cordillera. Small boxes indicate the field of rapakivi granites and their associated suite. R, Riga; F, Finland; P, Poland; L, Labrador. Also shown, in grey, data from A, Amazonia and G, Guyana. Data from North America (Anderson and Morrison, 1992) are separated into a western US (wA) heavy dotted and a light dotted box (eA) for eastern US (see text).

(Fig. 4). The diagram displays data for European Caledonian and Hercynian granites, i.e., granites having a similar source, and from Southeastern Australia, ranging the same period of time. It also displays data from Proterozoic granites of China. The diagram, already issued for determining the source conditions of granite generation (Jahn et al., 2000), may be used as a framework for the generation of rapakivi granites (Fig. 4). The total field for common granites extends from mantle conditions (positive εNd and low Isr ) to average crustal conditions (negative εNd and high Isr ). The residence time of crustal material is marked by decreasing the bulk εNd value by about −10 for a residence time of about 1 Ga. In this diagram, the rapakivi granites and associated suite plot close to the mantle field. Isr are generally low (0.703–0.709) demonstrating a moderate component of young crust in magma generation. A second conclusion is that the respective boxes of εNd − Isr for Latvia, Finland, Poland and Labrador are progressively shifted to more contamination (Fig. 4), thus to a larger crustal contribution. The shift also reflects the younger age (1.59–1.35 Ga) of intrusions in these regions. A similar shift is observed for rocks from the Amazonian shield. Data from northern US (Anderson and Morrison, 1992) have also been incorporated, with a separation into two blocks, for western

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

US (CO, NM, AZ, NE) and for eastern US (WI, MI, IL, NE, MS, OK). The two provinces also display an incomplete overlap, reflecting the age of the protolith (Fig. 4).

5. Conditions of emplacement The conditions of emplacement in each province point to a long-lived heat source that manifests itself by high temperature requirements, high-grade metamorphism and very specific conditions of pluton emplacement. 5.1. Pressure and temperature The pressure and temperature conditions leading to the formation of the rapakivi texture have been subject to a large debate (reviews in Emslie, 1991 and R¨am¨o and Haapala, 1995). A converging set of data indicates that isothermal decompression of the magma (Eklund and Shebanov, 1999) seems more likely than mixing, diffusion or exsolution processes. Barometric determinations of plagioclase-mantled alkali-feldspars define three groups of pressure. The core zones and mineral inclusions within the core of feldspar ovoids display the largest pressure (500–600 MPa), but they also include a second group of intermediate pressures (450–650 MPa). Cores of quartz megacrysts similarly present the two groups of values. In contrast, thermobarometric measurements on the matrix show low pressure (100–250 MPa). Temperature estimates in both groups are within the same range. The high-pressure group indicates 720 ◦ C in average, whereas the group of intermediate pressure presents temperatures ranging 720–780 ◦ C, cooling down 700 ◦ C for the low pressure group. Fluid interaction that resulted in the trapping of fluid inclusions developed at 550–600 ◦ C (Shebanov and Eklund, 1997). The range of pressure–temperature conditions suggests that ovoids and mantled plagioclases are the consequences of the isothermal decompression of the magma. Measurements of mineral proportions in granites that have a rapakivi texture (Eklund and Shebanov, 1999) estimate the following initial solid proportion; 20% alkali-feldspar, 12% plagioclase and 6% quartz. This means that the magma was 62% liquid and 38% solid. The large amount of the initial solid phase cor-

15

roborates other determinations (40% crystals) in southwest Finland (Shebanov and Eklund, 1997). The ascent caused partial dissolution of quartz and alkali-feldspar megacrysts while plagioclase precipitated. Due to the changing pressure and temperature conditions, the melt content of the magma increased by about 20%, thus increasing the ability of diffusion to develop. It would explain the rounded shape and deep embayments of quartz megacrysts, as well as the rounded shape of the ovoids with intergrowths in their intermediate zones. Late precipitation of plagioclase results in the mantling of the ovoids and separate euhedral crystals (Eklund and Shebanov, 1999). In this model, a specific type of fabric develops, because diffusion during crystal resorption takes advantage of crystal motion under shear flow. The above model (Eklund and Shebanov, 1999) explains the thermodynamical data observed in a detailed analysis at the crystal level. Moreover, it points to two major facts related to the conditions of magma ascent. First, the huge amount of the solid phase during the ascent is estimated at about 40% of solid phase at 100 MPa (Eklund and Shebanov, 1999). The solid phase content decreases during the ascent, making upwelling more likely. Second, the temperature loss is insignificant during the pressure loss of about 500 MPa. Though they do not correspond to a conductive geothermal gradient, these values indicate that the magma rose isothermally. This means that either magma rose very rapidly, or that the surroundings were at a temperature close to that of the magma, high enough to avoid heat loss. 5.2. Long-lived high-grade metamorphism The rocks that surround the granitic intrusions indicate high-grade metamorphism. Rather than reflecting peak metamorphic conditions, they reveal very long and stable high temperature (Shaw et al., 1999). Studies using 40 Ar/39 Ar overstepping assemblages on hornblende, muscovite, biotite and microcline in the Colorado Front Range indicate that the temperature remained above 525 ◦ C during the period from 1594 to 1390 Ma (Shaw et al., 1999). After that period cooling led to a temperature close to 300 ◦ C. The high temperature over a long time induced resetting of mica ages at 1.4 Ga. It also resulted in the production of staurotite, garnet and andalusite. The time-scale imposed

16

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

by the duration of the metamorphism cannot be explained by magmatic intrusions, dated at 1.7 and 1.5 Ga in the region. Intrusions cannot explain the long-lived thermal anomaly. Similar temperatures leading to reset micas and hornblende have been described in the East European Shield (Bogdanova et al., 2001), Poland (Marheine and Valverde-Vaquero, 2002), and Mount Isa Inlier, Australia (Spikins et al., 2002). In all these regions, the temperatures remained higher than 500 ◦ C for a long time, and sudden cooling occurred just after 1400 Ma. This age clearly postdates deformation of surrounding metamorphic rocks in the amphibolite facies. A long time after intrusion, but before cooling, is also indicated by paleomagnetic studies (Miron Valdespino and Constanzo Alvarez, 1997). Magnetite, titanomagnetite and hematite are the three different magnetic carriers in the Parguaza rapakivi granite, Venezuela. Because they have different blocking temperatures, each carries a different component of the bulk magnetisation. Two of them range between 400 and 600 ◦ C. The demagnetisation trajectories and the distance between the poles reflect the long cooling time of the granite.

The large spacing indicates mechanism of ascent driven by gravity instability (Vogt, 1974) rather than tectonic forces. All other arguments for a very hot source region at the base of the crust, a hot crust for a long time, light magmas and large spacing of intrusions, converge toward the same mechanism. Diapiric ascent was suggested for the Nain anorthosites (Royse and Park, 2000), Greenland granites (Bridgwater et al., 1974) and the Rogaland anorthosite massif, Norway (Barnichon et al., 1999), though the later is much younger. Evidence for the style of emplacement is provided by the modelling of flow that perfectly fits the orientation of large pyroxene megacrysts (Wiebe, 1986) that corroborate fabric measurements (Bolle et al., 2002). A diapiric ascent of the ACMG suite is consistent with most observed variables. The crust is hot within the ductile field; its temperature is bracketed between 780 and 550–600 ◦ C for the granites, from source to emplacement. The density of the magma, rich in light plagioclase, is much lower than the crust. Its viscosity (107 Pa s) is somehow higher than that of usual magmas, owing to the large amount of a solid phase. 5.4. 3D shape of the intrusions

5.3. Spacing of the intrusions The spacing between the rapakivi granites and associated suites is unusually large compared with the spacing between other granitic intrusions. A true spacing, i.e., a discrete periodicity is no longer interpreted as reflecting the initial instability (De Bremond d’Ars et al., 1995). Nevertheless, the spacing ranges between 50 and 70 km for volcanoes and plutons (Vogt, 1974; Rickard and Ward, 1981). The spacing associated with the AMCG suite is much larger. A rapid examination of maps (Anderson and Bender, 1989; R¨am¨o and Haapala, 1995) documents an average spacing between intrusions ranging from 100 to 150 km. The larger spacing would normally reflect a stronger crust for a magma with viscosity like that of a common granite. However, this effect is balanced by the increase in viscosity of rapakivi granites that incorporates a greater initial crystal content; this corresponds to a softened crust. In fact the large spacing reflects an increase of the elastic thickness of the plate that would range up to 60 km (Ten Brink, 1991). This therefore supposes that the elastic lithosphere is commonly thinner (or hotter) than normal lithosphere.

The global shape of rapakivi granites profoundly differs from that of other massifs. They have been surveyed through seismic reflection profiles and gravity inversion (Allen and Hinze, 1992; Jensen et al., 1997; Korsman et al., 1999; Louden and Fan, 1998; Funck and Louden, 1999; Korja et al., 2001; Wardle and Hall, ˚ 2002). For instance the Aland batholith crops out only ˚ on small islands in the Aland Archipelago, but it extends over about 110 km × 90 km according to the gravity map. From the seismic section and gravity interpretation the granite is about 8 km thick. Similarly, the Nordingr˚a granite in central Sweden is only 5 km thick, but has 50 km × 20 km lateral extent (Korja et al., 2001). A diagram similar to a Flinn diagram (Flinn, 1962) discriminates the three dimensional characteristics of the intrusions (Fig. 2). In this diagram the respective ratio of the length/width (L/W) is compared with the ratio of the width/thickness (W/T). The diagram has been used for granitic plutons that were investigated for their three-dimensional shape (Am´eglio et al., 1997). It clearly separates wedge-shaped plutons from flatfloored ones, reflecting the control of the emplacement

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

shape by regional tectonics. When rapakivi massifs are plotted on this diagram, they appear very different compared to other granitic massifs. Their L/W ratio is around one, reflecting a quasi-square shape at the surface. However, their very thin shape, compared with their lateral extent results in large to very large W/T, up to 40 in the case of Wiborg massif, provided a thickness of 5 km is assumed. The Ukrainian massif of Korosten extends over about 12,000 km2 and has a quasi-square shape (125 km × 100 km), but a very restricted thickness (0.5–3 km) estimated from seismic surveys (Sivoronov et al., 1998). In consequence, it also plots far on the right side of the diagram, in the field of thin plates (Fig. 2). The outcrop and relief in Greenland reveal the extreme thin-sheet character of the rapakivi granites (Hutton et al., 1990; Grocott et al., 1999). Thicknesses of 2 km (Graah Fjelde), but as low as 0.8 km (Qernertoq), are observed in the topography (Grocott et al., 1999). The undulating floor with depressions and deeply inward plunging features yielded models of crustal downfolding and floor depression (Bridgwater et al., 1974; Grocott et al., 1999). Most rapakivi granite plutons display an internal petrographic zonation. The last magmas, more evolved and toward topaz-bearing granites do not always present the general shape of the former intrusives. Those have an irregular, often quasi-square, bulk ˚ shape (Wiborg, Korosten, Aland). In contrast, the late facies have an elongate shape, more similar to other granitic intrusions. The late, not only the commonly topaz-bearing granites, but also smaller rapakivi granites have elliptical shapes. Their common dimension is about 15 km × 5 km, as documented in the Bodom and Obbn¨as massifs, south Finland (Kosunen, 1999; Kosunen et al., 2002); both trend N045. A similar trend is observed within the late topaz-bearing granites of Ahvenisto (Alviola et al., 1999) and Suomenniemi (R¨amo, 1999). Their orientation is commonly NE in present coordinates (Korosten, Mazury, Ahvenisto) in the Fennoscandian and European shield. Similar trends and elongated features are observed in the Amazonian shield. The orientations of Greenland and Labrador granites are more difficult to analyse, owing to the large shearing that partly dismembered the region. In central US granites from the Mazatzal Province has a bulk trend. In all regions, where it could be observed, the grain of the late facies is oblique to adjacent shear zones.

17

Internal ages within a single massif show a large variation in the emplacement dates. The rapakivi massif of Wiborg, Finland is a good example of such a variability (Elliott, 2001). The massif has an outcrop area of about 350 km × 200 km at the border between Finland and Russia, with no preferred orientation. Zircon ages for the emplacement of the whole massif range from 1615 to 1645 Ma for the whole massif. Individual values in each respective facies are constrained within ±2 to ±6 Ma (Alviola et al., 1999). Nevertheless, the time-span between separate intrusions spans 30 Ma. Similarly, the smaller massif of Salmi, close to Saint Petersburg, Russia (Fig. 3), has an age-range of more than 17 Ma (Amelin et al., 1997). At least six episodes of magmatism correspond to the intrusion of more evolved magmas. In consequence, any model of emplacement must explain how the pluton construction resulted from a rapid and isothermal ascent of magma, with a high solid fraction, and how the total time of pluton construction took place during about 30 Ma. A similar long-time has been ascribed for the formation of the Nain plutons (40 Ma), and the Korosten batholith in Ukraine (30 Ma) (Emslie et al., 1994; Hamilton et al., 1998).

6. Mantle convection and supercontinents Owing to the existence of the Columbia supercontinent and the ubiquitous role of the mantle in generating the magmas, the pattern of global mantle convection must be examined. 6.1. The global scheme Global mantle convection and motion of the lithospheric plates are intimately linked (Bercovici, 2003). In particular, downwelling zones of mantle convection act as local attraction points for continents (Gurnis, 1988; Zhong and Gurnis, 1993). However, the exact nature of mantle convection is still a matter of debate. Layered or non-layered convection through the entire mantle is not yet confirmed (Hofmann, 1997; Van der Hislt and Karason, 1999; Albar`ede and Van der Hilst, 1999). Certainly, the upper and lower parts of a hot mantle convect individually (Machetel and Weber, 1991). But small variations in the thermal conditions may lead to catastrophic mantle overturns (Brunet and

18

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

Fig. 5. Cartoon showing the assumed position of the supercontinent and development of strike–slip deformation. Poloidal flow in grey arrows and toroidal flow in black arrows. Late facies are represented with their orientation, as well as the local shear zones. They define a sinistral shear. The large ellipse in grey indicates the progressive advance of magmatism with time.

Machetel, 1998). They have been considered as the major causes of aggregation and dispersal of supercontinents (Hofmann, 1997; Condie, 1998). The total cycle of aggregation and dispersion takes about 700–800 Ma (Fig. 5). Modelling indicates that the total time subdivides into convergence for 325–350 Ma followed by divergence for 435–450 Ma (Tao and Jarvis, 2002). Assuming that oceanic plates cannot last longer than 200–250 Ma, before they start subducting, the time a supercontinent develops and stays aggregated is about 200–300 Ma, before it splits into smaller plates. A convective motion within a sphere commonly comprises a poloidal and a toroidal motion for all spherical harmonics. It reflects the usual decomposition of deformation into a divergent (poloidal) and a rotational (toroidal) mode (Jaeger, 1969). The l = 1 streamlines of the toroidal field are along circles parallel to the equator, with respect to the rotation pole. In contrast, those of the poloidal field are along meridian circles. For higher degrees, the toroidal streamlines are purely azimuthal, making close loops. In contrast, the poloidal streamlines have a radial component. The poloidal motion is associated with vertical flow, which diverges or converges at the Earth’s surface; it is expressed by the creation and destruction of plates (Hager and

O’Connell, 1978). Directly linked to advective heat transport by the vertical upwelling and sinks, it correlates with the release of gravitational potential energy. The toroidal plate motion at degree 1 represents the net rotation of the plates on the Earth’s surface. It is associated with strike–slip motion as transform faults, oblique subduction, or plate spin along a vertical axis (Bercovici et al., 2000). The exact significance of the toroidal component is not well understood, because it has no direct energy source due to its horizontal motion. It is often referred to as the non-convective component of the flow (Lithgow-Bertelloni et al., 1993). It minimises the dissipation of poloidal motion, enhancing the thermodynamic efficiency of the convection process (Bercovici, 1995; Tackley, 2000). Hence, the toroidal component is presently interpreted as being induced by the presence of the plates (Gable et al., 1991; Dubuffet et al., 2000). The spectral decomposition of the flow must be considered with respect to both the vertical (about 3000 km) and the horizontal dimensions of the cell. Both wavelengths should be compared with the respective dimension of the lithospheric plates. The spectral decomposition is marked by two low frequency peaks, notably at l = 2 and 6 (Anderson, 1998) that correspond

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

to wavelengths of 20,000 and 6000 km, respectively. The bulk surface of Columbia, with a maximum extent of 11,000 km × 7500 km (Fig. 1) requires an equivalent radius of 8000 km. The persistence of a supercontinent over a cell, much larger than twice the vertical dimension of the cell, has a blanketing effect on the released heat (Lowman and Jarvis, 1999; Grign´e and Labrosse, 2001). Two adjacent cells may therefore coalesce into a single one (Trubitsyn et al., 2003). It takes about 200 Ma for the focused heat to develop, inducing a thermal anomaly of 200 ◦ C. This is about double the average heat flow released from the mantle (Trubitsyn et al., 2003). Continental keels have apparently little effect on the long wavelength of the geoid, with degrees 2–5, corresponding to a wavelength of the order of 5000–10,000 km (Zhong, 2001). In contrast, they have a net influence on the degree 1 of the geoid, since they modify the viscosity distribution within the mantle (Tackley, 2000). They excite the degree 1 of the toroidal motion, i.e., the net rotation of the lithosphere. The latter is presently estimated to take place at a rate of about 2–4.5 mrad/Ma when compared with the hotspot reference frame (Gordon and Jurdy, 1986). However, continental keels, even as thick as 300 km, cannot achieve the necessary coupling with the deep mantle to produce the bulk rotation of the plate (Zhong, 2001). The toroidal component of the mantle flow decreases accordingly, as shown by the decrease of the toroidal–poloidal component with the decreasing number of plates (Gable et al., 1991; Monnereau and Qu´er´e, 2001). The formation of a supercontinent by successive aggregation of smaller continents globally corresponds to a converging flow toward a descending poloidal cell. In consequence, the formation of a supercontinent should result in the progressive slowing down of its absolute motion. Because the supercontinent is attracted toward the zone of downwelling flow of the mantle, it progressively aggregates (Fig. 5) when convergence leads to the disappearance of active subduction zones. Hence, slab pull is the dominant force of all forces that drive plate motions (Forsyhth and Uyeda, 1975). After amalgamating the supercontinent, the poloidal flow, or global convergent motion of the supercontinent, decreases, whereas the toroidal motion increases accordingly. Once formed, a supercontinent should consequently stay stable above the attraction zone of the de-

19

scending flow in the mantle (Mimouni and Rabinowicz, 1988). The interaction between old continental lithospheres and mantle currents would last about 500 Ma, which is actually the range of time between aggregation and disruption of a supercontinent.

6.2. An antiplume model While stationary above a downwelling cell of the mantle, a supercontinent receives heat from the mantle. Confusion often occurs between material flow as convective motion of the mantle, and heat flow that has a diffusive and an advective component. Above a descending convection cell, the material flow is downward, but heat continues to be evacuated by diffusion toward the surface. The continent, acting as an insulator, taps the heat released (Lowman and Jarvis, 1999). However, it manifests thermal effects only after some time, because heat diffuses through the lithosphere. This scheme is called antiplume, in contrast to the commonly conceived plume in which material and heat ascend together. In the antiplume model, the mantle material moves downward, but the heat released from the mantle still goes upward. Models of continental heating by a thermal plume indicate that about 185 Ma are required before the thermal effect is registered at a depth of 150 km, the base of the lithosphere (Monnereau et al., 1993). The time for the heat to diffuse through those 150 km is about 675 Ma, adopting a thermal diffusivity of 10−6 m2 s−1 ; this corresponds to the time before the full thermal perturbation occurs. Thus, a total of 860 Ma should be expected in the case of a common thermal plume. Reorganising the underlying cells and heating the continent takes a shorter time, about 250 Ma (Trubitsyn et al., 2003). This implies a total lifetime for convection of about 1.1 Ga to invert the sense of the mantle flux (D’Acremont et al., 2003). The long time before heating and rifting invalidates the plume hypothesis as a major means to break apart a supercontinent. Further possibilities for extensional rifting of a supercontinent are when the heat flow delivery is rapid, or when the plate is abnormally thin. In this case, the time delay for heat to reach the crust is shorter. In the case of a 50 km thick lithosphere, the time lag is 75 Ma. The time-lag between crustal thickening and melting is also too short, evaluated at 20–30 Ma (Thompson, 1999).

20

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

Rifting of a supercontinent does not necessarily require pure extension. It can also be expressed by strike–slip opening. Hence, a supercontinent is generally composed of Archean cratons surrounded by younger accreted belts. The Archean cratons have a deeper keel, owing to their older age. The keels act as a link with the underlying mantle motion. In contrast, the younger belts do not have such keels. The viscosity contrast induced by the difference in thickness of the old and juvenile lithosphere induces a toroidal component to the mantle flow (Tackley, 2000). It imposes a tangential motion to the supercontinent that leads to its break-up by strike–slip. If the half rate of present lithosphere rotation (Gordon and Jurdy, 1986) is applicable, a value of 3 mrad/Ma would induce a torque of only about 0.75 rad or 42◦ after 250 Ma. Shearinduced opening of the supercontinent does not avoid heat and magma continuously released by the mantle. Heat brought from the mantle is naturally focused toward this zone of splitting, simply by refraction of the keels (Nyblade and Pollack, 1993). Magma is also focused by the vorticity of the shear toward the zone of splitting.

7. An antiplume model for Columbia The model starts at 1.9–1.8 Ga (Fig. 5) after Laurentia and Fennoscandia have already been reunified. Other smaller continents have been incorporated into the supercontinent, but we lack data on the precise time of their collisions. The small Yavapai and Mazatzal Provinces joined the supercontinent later at 1.740–1.721 Ga (Duebendorfer et al., 2001). In this cartoon, the position of the several provinces forming Columbia is adopted from the reconstruction of Zhao et al. (2002). Certainly, the relative positions of Laurentia, joined with the Yavapai–Mazatzal Provinces and Greenland, Fennoscandia and possibly the East European Shield are correct. The respective positions of Antarctica and Australia are uncertain, as well as the position of Amazonia and West Africa. The positions of other blocks, such as India, South Africa and China, are approximate due to the lack of more precise data. The internal rotation of the supercontinent is not taken into account in this diagram, due to lack of longitudinal control of the several blocks. In the same way, initiation of large blocks fragmentation that would dis-

member the supercontinent is not considered due to the lack of more precise ages and location (Fig. 5). Four major points integrate into a logical simple scheme (Fig. 5). (1) The stationary situation of Columbia is documented by paleomagnetic reconstructions for the Laurentian and Fennoscandian shields (Buchan et al., 2001). (2) A bulk trend in ages of rapakivi granite formation extends from Fennoscandia (1.57 Ga), Amazonia (1.54 Ga) and western US (1.48 Ga), to centralnorth US (1.43 Ga) and Labrador (1.33 Ga). (3) Those regions approximately align on a small circle and each region represents magmatism in a clockwise sense. The clockwise development of magmatism with time documents an anticlockwise rotation of the plate. (4) In each region the position of shear zones and the orientation of possible late facies granites, also document opening during local sinistral deformation (Fig. 5). The long time that separates the peak metamorphism and onset of magmatism, the long period during which sediments remain hot, over 500 ◦ C, and the long time of a single massif emplacement, document a global process driven by heat conduction. 7.1. Bulk diffusional heat transport Heat transport was an important factor during the antiplume process. The magmatic process that develops after supercontinent re-assembly is extremely long. Rates are considerably slower compared with those of other intrusive mechanisms. Heating takes a long time, in the range of 250 Ma, consequently softening of the crust also takes time, leading to a very low deformation rate. Heat must be provided to the lithosphere, and through it, to the base of the crust in order to induce melting of the latter. The temperature at the base of the crust (1.1–1.3 GPa) must reach 1200–1300 ◦ C to reach the conditions of anorthosite production (Longhi et al., 1999). This roughly corresponds to a thermal gradient in the crust of 40 ◦ C/km. The time for the conductive heat to reach the base of the crust can be as long as 300 Ma if the lithosphere is 100 km thick. Heating the lithosphere also induces local partial melting, generat-

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

ing basaltic liquids that can later intrude the crust in the form of mafic dykes, prior to and synchronous with granite emplacement. However, their volume is not large enough to infer widespread extension (Anderson and Bender, 1989). Crustal thickness is also bracketed by several parameters. The depth to which the anorthosite should be produced (1.1–1.3 GPa) precludes large crustal thickness. The time after the orogenic peak also suggests a return to nearly isostatic conditions since erosion rates for the topography and viscosity values for the mantle preclude that abnormal crustal thickness could stand for 250 Ma. Assuming that the metamorphic peak does not require to be synchronous with the time of maximum thickening (Windley, 1993), the time to reach melting in a crust with double thickness remains about one order of magnitude less (Thompson, 1999) than the observed time-lag. The bulk trend observed in the ages of AMCG intrusions within the different provinces indicates that the crust slowly moved away from the major heat source in an ellipse with a size about 10,000 km × 4000 km (Fig. 5). Magmatism is delayed by about 30–50 Ma when passing from one region to another, from Fennoscandia, Amazonia and western US, to centralnorth US and Labrador. The time interval correlates with an average motion of about 2000 km, thus documenting a velocity of about 4 cm/year, which is a reasonable rate for plate motion. To accommodate a global rotation of 360◦ in 250 Ma, the average rotation rate should be 24 mrad/Ma, that is about eight times the average rotation rate under usual conditions (Gordon and Jurdy, 1986). In such a case, the pole of rotation is located at about 1000 km from the magmatic provinces, to fit the observations (Fig. 5). The estimated rotation rate roughly corresponds to an equivalent shear value (␥) of about 2, expressed by very low shear deformation. In this case, one cannot expect huge shear zones to develop, as in the Great Slave Shear Zone, Canada (Hanmer et al., 1992). The time lag after the peak metamorphism is about 230 Ma, considering the end of the major collisions at 1.8 Ga and the start of magmatism at 1.57 Ga. This corresponds to heat diffusion over about 90 km, a commonly observed thickness for young continental lithosphere. The rocks surrounding the intrusions indicate high metamorphic temperatures over 500 ◦ C, lasting up to 1.5–1.4 Ga, before they suddenly cool to less

21

than 300 ◦ C. However, the Ar reset ages also represent a temporal trend because cooling occurred first in Sweden (1490 Ma), as documented by hornblende and white mica data (Beunk and Page, 2001). A similar age occurs in the East European craton (Bogdanova et al., 2001). Cooling only took place at 1390 Ma in Colorado (Shaw et al., 1999) and at 1392 Ma in Wisconsin (Holm and Lux, 1998). The polarity observed in ages of the intrusions is also present in the cooling ages. A younger age, around 1350 Ma, is observed in NE Australia (Spikins et al., 2002). The large spacing (100–150 km) between intrusions, associated with their specific shape (up to 100 km of lateral extent, but less than 5 km thick), as well as the time span of intrusions with a high crystal content (up to 40%), a lighter magma density (anorthosite and granite), combined with the surrounding hot character of the crust, all lead to the conclusion that magma ascent developed owing to gravitational instabilities. The mode of intrusion has been confirmed by numerical modelling of the intrusion mechanism of an anorthosite massif in Norway (Barnichon et al., 1999), though the intrusion is much younger (930 Ma). Structures indicated by huge pyroxene crystals and flow directions estimated by the anisotropy of magnetic susceptibility (Barnichon et al., 1999; Diot et al., 2003) indicate that a similar process developed leading to the emplacement of ilmenite-rich deposits linked to the anorthosite massif in Rogaland, Norway. The distance between plutons is about four times the depth of the source region, indicating a ratio of 1/2 between vertical and horizontal dimensions, which is a reasonable size for Rayleigh instabilities. The light density of anorthosites and granites justifies a diapiric upwelling of the magma (Royse and Park, 2000). Both types concern the lighter felsic rocks. The surrounding crust was at a high temperature, leading to plastic deformation, which made possible a low inertia flow of magma. The thin-plate shape of the intrusions also indicates magma spreading close to the surface (a few tens of MPa in many cases), and suggests magma upwelling and lateral extension, when magma pressure could not counteract the tectonic load (Hogan et al., 1998). The large dimension of the massifs, up to 100 km, compared with the thickness of less than 5 km, certainly indicates gravity-induced upwelling, which turned into horizontal flow. The large time-span (20–50 Ma) required to build a pluton exceeds the time-scale generally calcu-

22

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

lated (about 10 ka) to build granitic plutons that result from tectonic forces (Harris et al., 2000; Petford et al., 2000). This confirms that deformation is not very active during magma emplacement. It is consistent with the velocity of such instabilities that have a rate of vertical motion about 1 mm/year in an ascent of about 30 km. The corresponding rate would be of the order of 10−16 s−1 , which is not very fast, if a 5-km thick magma is transported upwards. Upward motion should result in various degrees of melting and mixing between mantle magma and crustal-derived material, leading to the various types and compositions of intrusive magmas (Duchesne et al., 1999; Longhi et al., 1999). The whole magmatic episode is calibrated in pressure and temperature using experimental petrology (Creaser et al., 1991; Longhi et al., 1999; Wiszniewska et al., 2002), P–T measurements on granitic melts (Frost and Frost, 1997; Duchesne et al., 1999; Eklund and Shebanov, 1999) and dated by U–Pb, Sm–Nd and Re–Os methods (Amelin et al., 1997; Morgan et al., 2000; D¨orr et al., 2002). In addition, Ar–Ar ages provide the cooling age, which is obviously younger (Bogdanova et al., 2001). A specific example is provided by the small massif of Salmi, Russia (Fig. 3). Six successive episodes of magmatism have been reported (Amelin et al., 1997) that developed over 17 Ma. Anorthosites were the first intrusives with apatite ages of 1569 ± 7 and 1567 ± 4 Ma that formed at the base of the crust at 1.3 GPa at 1300 ◦ C (Longhi et al., 1999). Gabbro-norites and monzonites follow. The first magma gave rise to a two pyroxene- and plagioclasebearing rock, which cannot be found in the mantle (Wiszniewska et al., 2002). The rocks have an average age of 1546.7 Ma and precede a syenogranite dated at 1543.4 Ma. The following magma includes the most important part of the granites with rapakivi textures, and constitutes the early Salmi wiborgite and pyterlite at 1540.6–1537.9 Ma, before a main biotite granite at 1538.4–1535.0 Ma and the late Salmi pyterlite at 1535 Ma. The Salmi olivine gabbro (1530.6 Ma) and the Uljalegi amphibolite-bearing granite (1529.9 Ma) are the late intrusions (Amelin et al., 1997). The late granite may be estimated to form at 500 MPa and about 850 ◦ C, common temperatures and pressure for biotite breakdown (Pati˜no Douce, 1996). Each magmatic episode had a probable duration of 2.7–3.4 Ma with mutual intervals of 3.5–5.0 Ma. The total magmatic episode can be calibrated in tempera-

ture and pressure, extending over 400 ◦ C for a pressure difference of 800 MPa; the latter corresponds to about 23-km in depth. The measured time interval of 17 Ma is similar to the computed time, about 16 Ma, for heat to diffuse over 23 km. The time interval between magma pulses corresponds to the time required for heat to conduct about 2 km upwards, which adds to the time required for melting. The progressive cooling of the temperature front can be attributed to the upward migration of magma that progressively looses energy through latent heat release induced by the progressive melting of the rocks at shallower levels. This implies a model of a stationary plate heated from below for a long time interval that also controls the reaction kinetics leading to rapakivi texture (Eklund and Shebanov, 1999). Finally, because heat is continuously provided, this explains the formation of a crustal mush with a large solid phase still able to move. The individual shapes of the large rapakivi granites do not indicate of any preferential orientation. In contrast, the last facies suggests an elliptical shape. When plotted in their supposed geographical position (Fig. 5), the late granitic magmas (topaz-bearing granites) are oblique to the orientation of adjacent shear zones in each region (Fig. 5). Considering the discrete time interval for most granitic pluton production, the late facies orientation in each region is oblique to the trend of the dextral shear zones. Obliquity defines a bulk sinistral shear deformation, with corresponding extensional joints and conjugated shear zones (Fig. 5). The global organisation does not appear as fortuitous, but reflects the progressive shearing of the supercontinent. The shear deformation, combined with the opening, jumps progressively from one province to the next, from Fennoscandia to Labrador with time (Fig. 5). This defines a global anticlockwise rotation along a vertical axis, the cause of which is attributed to the rotational motion of the supercontinent, after the converging poloidal motion has attracted all plates.

8. Discussion 8.1. The very specific conditions for the AMCG suite development Anorthosites are largely restricted to the Proterozoic (Emslie, 1978), as are the huge rapakivi granite plutons.

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

However, younger rapakivi granites also occur locally in limited areas, up to the Tertiary (R¨am¨o and Haapala, 1995; Haapala et al., 2005). The global tectonic context that led to the formation of the AMCG suite during the Mesoproterozoic is very specific and requires several ingredients: (1) a supercontinent that formed from a descending mantle convective cell (poloidal component); (2) the supercontinent must include both young juvenile crust and old Archean cratons that are both marked by lithosphere with contrasting age and thickness; (3) fragments of lithosphere with contrasting ages induce a tangential flow and concentrate the heat flow (toroidal component); (4) an high temperature (1300 ◦ C) must be reached at the base of the crust; (5) hence, the plate must remain stationary for a long time before it is warmed up; (6) the large-scale tectonic environment of the supercontinent remains stable, without large internal deformation; (7) global warming of the crust develops for a long time (>200 Ma), progressively induced by heat diffusion and melting. These points are certainly not always realised at the end of each supercontinent cycle. In the case of Columbia, the lithosphere pinned to the respective Archean cratons contrast with the lithosphere associated with young juvenile crust. Some Archean cratons were much older than 700 Ma, whereas the youngest were only 200 Ma old. The contrast emphasises the crucial role of the rate of heating in the development of magmas. At present, few studies consider heating rates in experimental petrology. Rates control the time to reach equilibrium during melting, but they can also bring equilibrium to unexpected situations, as reflected in the development of rapakivi textures by displacing phase boundaries (Dempster et al., 1994). In contrast to the Columbia supercontinent, magmatism after the assembly of Rodinia at about 1000 Ma gave rise to anorthosites at 950–900 Ma (Emslie, 1989). In Norway, anorthosites are dated at 900–930 Ma (Bingen and van Breemen, 1998) and the time-lag after peak metamorphism is only 100 Ma, but the age difference of the various segments of lithosphere is smaller. The anorthosites crop out over modest surface areas

23

compared with those in Mesoproterozoic times. Close to the Grenville Front, the situation seems different because many anorthosites (Adirondacks, Morin, . . .) and AMCG are dated at 1.2–1.1 Ga (Rivers, 1997; Hanmer et al., 2000). They predated the re-assembly of Rodinia that did not take place before 1.0 Ga. This situation may be re-examined in the light of the whole scheme of events. If the southeastern province of Laurentia had been connected to Amazonia that is with the peak of rapakivi intrusions at 1.54 Ga, then some fragments of the supercontinent may have rifted, leading to convergence at 1.3 Ga and closure of successive magmatic arcs with an actual global trend N030 (Hanmer et al., 2000; Gower and Krogh, 2002). This led to AMCG magmatism, due to a still hot crust. Conditions required for the above points were obviously not realised during the Phanerozoic, while dismembering Pangea. A supercontinent with less contrasted lithosphere is certainly not able to induce a large enough toroidal component. Hence, heat release would preferentially result in rifting, leading to new ocean formation that splits apart the former supercontinent. Rapakivi granites formed during the Brasiliano orogen, or Panafrican orogen, at 600–550 Ma, and massifs developed from fayalite- or magnetite-bearing magmas (Galindo et al., 1995). Several such as Umarizal (Archanjo et al., 1998) or Pombal (Archanjo et al., 1994) have been studied for their emplacement. They all record a close association with large ductile shear zones that extend over thousands of kilometres through NE Brazil and W Africa. 8.2. The significance of the ilmenite-bearing versus magnetite-bearing granites The bimodal ilmenite-bearing and magnetitebearing suite is a puzzling feature of Meso-Proterozoic granites. Ilmenite-bearing rocks developed mostly in Fennoscandia, East European craton, Greenland and Amazonia between 1.6 and 1.5 Ga. Magnetite-bearing granites were late (<1.48 Ga) and occur in the western US and Nain Province, Labrador. The change in magma composition can be attributed to either a change in the conditions of formation, or in the conditions of emplacement. Ilmenite-bearing magmas are traditionally generated within a reducing environment (Frost and Frost, 1997). They also require a high temperature and restricted fluid environment with low oxygen

24

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

Fig. 6. Zircon age distribution (redrawn and adapted from Condie, 1998) for intrusions on Earth. The scale, on the left, is in frequency of ages. Above is a similar diagram for rapakivi granites and igneous associated suite for the period 1.7–1.4 Ga. The scale is on the right, with ordinate displaced with respect to the former frequency diagram. The ilmenite-bearing (black and grey chessboard) and the magnetitebearing (black) granites are identified. BIF occurrences (in white) are also represented (Isley and Abbott, 1999). Duration estimates of supercontinents are plotted, reflecting the correlation between the supercontinent cycle and granite formation, revealing the indirect role of the mantle.

and water fugacity. In contrast, magnetite-bearing granites require less heat and more fluids. Nevertheless, the change in oxidation state of the magma after a time is important to consider because similar events also affected redox conditions and are recorded in sediments. These changes are consistent with a global change in atmosphere-ocean interactions (Fig. 6). They all occurred during a limited period of the Mid-Proterozoic, between 2.0 and 1.6 Ga, and include banded iron formations (BIF) that ceased at 1.8 Ga (Isley and Abbott, 1999), the change in ocean oxygen productivity (Bjerrum and Canfield, 2002) at about 1.9 Ga, and the shift in δ13 N to positive values between 2.0 and 1.6 Ga (Beaumont and Robert, 1999). The significant oxidation of the Earth’s surface between 2.0 and 1.8 Ga (Canfield, 1998) was marked by a change in sulfur isotopic concentration and its separation in sulphate and sulfide. Because the oxidation states of the ocean and atmosphere were intimately linked, especially through a global CO2 balance, the oxidation state of the granites could reflect the change that affected meteoritic water at those times. It is commonly accepted that the oceans of the primitive Earth were anoxic, i.e., without oxygen, in contrast with the oxic oceans today (Canfield, 1998).

This is partly connected with the appearance of life. Massive banded iron formations could only form in anoxic oceans because iron is removed from water when it reacts with oxygen. The BIF deposition ceased at about 1.8 Ga (Isley and Abbott, 1999). However, an intermediate ocean stage has been recently suggested, in which the deep sea did not become rich in oxygen, but rather rich in hydrogen sulfide (Canfield, 1998). In this model, anoxic waters probably persisted for some time after the end of BIF deposits. Sulphur isotopes support this evolution, which started at about 1.8 Ga and ended at about 1.0 Ga (Canfield, 1998). However, the precise time of change has not yet been determined. Fluctuations in isotopic Sr (87 Sr/86 Sr) in well-preserved marine carbonate rocks reflect changes in the relative contributions of continental versus mantle chemical reservoirs to oceanic composition (Veizer, 1989; Lyons et al., 2000). Geochemical evaluations of the Proterozoic oceanic conditions involve the sulfur cycle, black shale geochemistry (metals, including Fe/Al and Mo/Al approaches, C–S relationships, and sulfur isotopes), and carbonateassociated sulphate tracing in dolostones and limestones. These measurements constrain sulphate and oxygen availability in the Mesoproterozoic oceans. They have implications for oxygenation of the atmosphere and deep ocean. They have been applied to the 1.4 Ga Belt Supergroup of the northwestern United States, the 1.2 Ga Apache Group of central Arizona, and the 1.7 Ga McNamara Group of Queensland, Australia (Luepke and Lyons, 2001; Lyons et al., 2003). The major argument that the deep ocean became oxidised at 1.8 Ga relates to the disappearance of BIF at this time. Their formation requires anoxic deep water to deliver Fe2+ by hydrothermalism to the site of deposition (Holland, 1973). Oxygenation of the oceans would produce Fe3+ that gives place to insoluble ferrooxyhydroxides, removing Fe and precluding BIF formation. Independent geochemical evidence documents a rise of O2 partial pressure between 2.4 and 2.0 Ga (Karhu and Holland, 1996). Simple modelling suggests such a scheme was unavoidable if atmospheric O2 was about one-third of the present level and if biological productivity was comparable to that of modern oceans (Anbar and Knoll, 2002). Hence, sulphur isotope data from sulfides and sulphates in individual basins suggest a low-sulphate ocean between 1.8 and 1.0 Ga (Shen et

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

al., 2002). The difference in isotopic sulphur (δ34 S) between sulphate and sulfide rarely exceeded 45‰ until 1.0 Ma which indicates that ocean oxygenation was less extensive than today (Anbar and Knoll, 2002). This suggests that only the shallow depths of the oceans were oxygenated during this time, and that more extensive oxygenation did not occur until 1.0 Ga. The determination of the global-scale event that triggered the onset of oxygenation of oceans is still poorly constrained. Inspection of sediments is not sufficiently conclusive, owing to the paucity of deposits, and the lack of fauna and vegetation that could provide information on the oxygen conditions. Nevertheless, it appears that sediments displayed changing isotopic conditions at about 1700 Ma and a very flat δ13 C profile close to 0 ± 2‰ between 1850 and 1250 Ma (Bartley et al., 2001; Anbar and Knoll, 2002; Ray et al., 2003), documenting unique long-term stability of the carbon cycle before the onset of large-amplitude δ13 C variations at about 800 Ma leading to the snowball Earth (Hoffmann and Schrag, 2002). Nevertheless, the profound change in the atmosphere/ocean O2 balance should have implications for the redox state of percolating fluids and associated magmatism. Granitic magmas reflect a progressive melting of hydrous minerals of the crust and should reflect the change within a time-lag. The puzzling change of oxidation stage of the intrusive magmas, separated in time and by province, cannot be understood without the interaction between the atmosphere and the ocean. The change from ilmenite- to magnetite-bearing granites may also be thought of in terms of temperature. The ilmenite-bearing granites are restricted to the first emplaced magmas that generally document a higher temperature of formation and emplacement, represented by a close association with anorthosites, and a long time of emplacement. Conversely, magnetitebearing granites are often associated with biotite and biotite-muscovite granites, reflecting a lower temperature of genesis. The dichotomy can thus be an effect of heat supply. As the supercontinent rotates over the heat source, the temperature of formation decreases, leading first to high temperature magmas, and the anorthosites and associated norites at the base of the crust. This leads to the formation of ilmenite-bearing granites. With time, the temperature of the source region decreases, and leads to the formation of magnetite-bearing granites.

25

8.3. Consequences for the Moho topography Development of Rayleigh instabilities followed by subsequent magma ascent to the upper crust should profoundly mark the base of the crust. Hence, maps in Finland, Eastern Europe, or Labrador show a bumpy interface of the Moho (Korsman et al., 1999; Jensen and Thybo, 2002; Louden and Fan, 1998; Funck and Louden, 1999). The depth to the Moho ranges from 41 to 63 km, displaying about 50% variation in thickness, which is considerable (Fig. 7). Furthermore, the wavelength associated with such bumps is extremely short, around 200 km. The corresponding slope is about 13◦ , and seems unrelated to any structure. A shallow Moho is often associated with a rapakivi body, as reflected by Bouguer anomaly maps. Negative gravity anomalies combine both the effect of lighter granitic intrusion and Moho undulations (Korja et al., 2001). Nevertheless, combined interpretations from seismic reflection lines can easily define both signals. The large gradient in the Moho topography had later consequences in the formation of large sandstone basins (Korja and Heikkinen, 1995; Puura and Floden, 1999) at 1.5 Ga

Fig. 7. Map of Moho topography, redrawn from R¨am¨o and Haapala (2003). Data were compiled from various sources. Rapakivi granites ˚ are in grey, delineating the massifs of Riga, Wiborg, Salmi, Aland and Nordingr˚a. Depth values to the Moho are in km. A profile (stippled line) is taken from the map and shown with a horizontal/vertical true scale. The place of the Wiborg massif is indicated with a thickness of 5 km.

26

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

that largely overlie rapakivi plutons. They also controlled the emplacement of numerous diabase dykes that are marked by strong reflectors, using listric faults in the intermediate crust down to 15 km (Korja et al., 2001). The bumpy character of the ancient Moho, and the present Moho jumps at the suture zones of the 1.9–1.8 Ga orogen in the Baltic shield (BABEL Working Group, 1993) indicate that the bumps reflect events that took place at that time. Similarly, the Trans-Hudson Orogen, Canada, occurs in a restricted area, about 50 km in width, where the Moho suddenly deepens to 58 km (Lucas et al., 1993). This severely contrasts with the present flat Moho observed under recent convergent orogens, such as the Hercynian (Meissner and Tanner, 1993) or the Sierra Nevada (Savage et al., 1994; Jones et al., 1994). The undulating Proterozoic Moho contrasts with the more recent Moho that returned to a flat topography after orogenic processes. The rapid transformation of pyroxene granulites at the base of a thickened crust into eclogite may explain the rapid return to a flat Moho in convergent orogens. Conversely, crust which has not been thickened during supercontinent splitting cannot return to a flat Moho because basal rocks do not achieve pressures high enough to reach eclogite transformation. Eclogite xenoliths are divisible into two groups, depending on whether they represent samples from the lower crust or deeper in the mantle (Rudnick et al., 1998). The former show an average pressure of equilibration close to 900 MPa, but which may vary from 800 to 1400 MPa (Rudnick, 1992). The second group that may represent former continental roots, has transformed into eclogites, and is also divisible into two groups (Barth et al., 2001). Many eclogites with a high Mg number record conditions of equilibration between 1080 and 1133 ◦ C with pressures in the range of 4.0–4.5 GPa. Conversely eclogites with a low Mg number span a low temperature range from 880 to 930 ◦ C and a pressure between 3.3 and 3.6 GPa (Barth et al., 2001). Such high pressures, corresponding to depths from 92 to 126 km, indicate that rocks were in the deepest region of a thickened crust. In the Proterozoic crustal rocks never reached that pressure, which prevented them from being transformed into eclogite. Thus, interactions between the crust and the mantle were different in the Proterozoic. The Moho undulations, as observed in Fennoscandia (Fig. 7) have been interpreted as crustal thinning

due to listric fault zones in connection with the rapakivi granites (Korja et al., 2001; R¨am¨o and Haapala, 2003). However, the correlation with Moho uplift, as evidenced on the Babel 1 line, is not observed at a map scale (Fig. 7). Though an uplift may be associated with the massifs of Wiborg and Nordingr˚a, no such feature ˚ is present in the case of the massifs of Salmi or Aland (Fig. 7). The association of a granitic massif with a thinned Moho does not present any coherent orientation on a map, reflecting the poor correlation with a global tectonic pattern. The widespread occurrence of such undulations, not always spatially associated with a granitic massif, is better explained as residual Moho topography remaining after magma removal.

9. Conclusions The specific case of “anorogenic” granites that lead to rapakivi granites and associate suites, including anorthosite and peraluminous rocks, reveals a strong mantle control on their mode of generation. A model is suggested in which they are indirectly a consequence of the supercontinent amalgamation. The plume hypothesis, which implies material upwelling, cannot be invoked because the supercontinent is located above a point of downwelling mantle flow (converging flow). However, even in a converging flow, heat is still delivered toward the surface. Within a supercontinent, the zone of juvenile crust focuses heat flow, leading to a hot and ductile crust. The toroidal part of the convective motion induces strike–slip deformation within the zone, inducing partial melting and diapiric magma ascent. No global age polarity is required in this system. Finally, the supercontinent splits and the fragments start aggregating toward a new supercontinent cycle. Considering the magmatic period lasting from 1.6 to 1.3 Ga that led to rapakivi granite formation, the above model takes into account the time delay between the major episode of accretion (1.8 Ga) and the widespread granite formation time. It corresponds to a diffusion time necessary for heat to warm the lithosphere. This explains the long time during which newly formed crust remains hot and ductile. The large timespan (17–50 Ma) during which the rapakivi granites and associated suites intruded the crust is explained by a progressive warming of the crust that successively

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

induces melting as the heat wave progresses upward. This is consistent with a large component of mantlederived material incorporated in granite and its associated suite, including anorthosites. The polarity of ages within the respective provinces documents a global rotation of the supercontinent. Magma ascent was controlled by diapirism, because it developed within a hot and plastic crust. Strike–slip certainly helped in segregating magma. This also explains how highly viscous magma, owing to crystal fraction, can ascend buoyantly. However, this model is restricted to supercontinent dismembering and cannot be directly transposed to other situations.

Acknowledgements It is a pleasure to acknowledge Joffi Eklund who introduced me to rapakivi granites while staying in Finland for a sabbatical period in Turku University, Finland. During that period, I acknowledge a travel and lodging grant (mid-January to mid-March 2000) from Turku University Foundation. O.T. R¨am¨o and I. Haapala are thanked for organising the special session on “Granitic systems—state of the art and future avenues” honouring I. Haapala’s retirement in Helsinki. I realised at this occasion the many problems that definitively make the “anorogenic” granites specific, and thus intellectually very attractive. Many thanks to K.C. Condie and O. Eklund who all alone encouraged me with this paper and who provided valuable comments and criticisms. D.H. Abbott and P.J. Tackley pointed out the blanketing effect of a supercontinent during their positive comments and reviews. The important editorial work done by D.H. Abbott and B. Windley is also warmly acknowledged. I also acknowledge CREGU and UMR 7566 for funding studies that were not explicitly planned.

References ˚ all, K.I., Connelly, J., Brewer, T.S., 2000. Episodic rapakivi magAh¨ matism due to distal orogenesis? Correlation of 1.69–1.50 Ga orogenic and inboard “anorogenic” events in the Baltic shield. Geology 28, 823–826. Albar`ede, F., Van der Hilst, R.D., 1999. New mantle convection model may reconcile conflicting evidence. EOS 80, 535–539.

27

Allen, D.J., Hinze, W.J., 1992. Wisconsin gravity minimum solution to a geologic and geophysical puzzle and implications for cratonic evolution. Geology 20, 515–518. Alviola, R., Johanson, B.S., R¨am¨o, O.T., Vaasjoki, M., 1999. The Proterozoic Ahvenisto rapakivi-massif-type anorthosite complex, southeastern Finland; petrography and U–Pb chronology. Precambrian Res. 95, 89–107. Am´eglio, L., Vigneresse, J.L., Bouchez, J.L., 1997. Granite pluton geometry and emplacement mode inferred from combined fabric and gravity data. In: Bouchez, J.L., Hutton, D.H.W., Stephens, W.E. (Eds.), Granite: from Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publications, Dordrecht, pp. 199–214. Amelin, Y.V., Larin, A.M., Tucker, R.D., 1997. Chronology of multiphase emplacement of the Salmi rapakivi granite–anorthosite complex, Baltic shield: implications for magmatic evolution. Contrib. Mineral. Petrol. 127, 353–368. Anbar, A.D., Knoll, A.H., 2002. Proterozoic ocean chemistry and evolution: a bioinorganic bridge? Science 297, 1137–1142. Anderson, D.L., 1998. The scales of mantle convection. Tectonophysics 284, 1–17. Anderson, J.L., Bender, E.E., 1989. Nature and origin of Proterozoic A-type granitic magmatism in the southwestern United States of America. Lithos 23, 19–52. Anderson, J.L., Cullers, R.L., 1999. Paleo- and Mesoproterozoic granite plutonism of Colorado and Wyoming, in Proterozoic magmatism of the Rocky Mountains and environs (Part I). Rocky Mountain Geol. 34, 149–164. Anderson, J.L., Morrison, J., 1992. The role of anorogenic granites in the Proterozoic crustal development of North America. In: Condie, K.C. (Ed.), Proterozoic Crustal Evolution. Elsevier, New York, pp. 263–299. Anderson, J.L., Morrison, J., 1998. Oxygen isotope systematics of 1.3 to 1.6 Ga granites of Laurentia. Geol. Soc. Am. Abstr. Programs 30, A89. Anderson, J.L., Cullers, R.L., Van Schmus, W.R., 1980. Anorogenic metaluminous and peraluminous granite plutonism in the MidProterozoic of Wisconsin, USA. Contrib. Mineral. Petrol. 74, 311–328. Andersson, U.B., 1991. Granitoid episodes and mafic-felsic magma interaction in the Svecofennian of the Fennoscandian Shield, with main emphasis on the ∼1.8 Ga plutonics. Precambrian Res. 51, 127–149. Andersson, U.B., 2001. An overview of the geochemical evolution in the Mesoproterozoic (1.58–1.50 Ga) anorogenic complexes of central Sweden. Z. Geol. Wiss. 29, 455–470. Andersson, U.B., Neymark, L.A., Billstr¨om, K., 2002. Petrogenesis of Mesoproterozoic (Subjotnian) rapakivi complexes of central Sweden: implications from U–Pb zircon ages, Nd, Sr and Pb isotopes. Trans. R. Soc. Edinb. Earth Sci. 92, 201– 228. Archanjo, C.J., Bouchez, J.L., Corsini, M., Vauchez, A., 1994. The Pombal granite pluton: magnetic fabric, emplacement and relationships with the Brasiliano strike–slip setting of NE Brazil (Paraiba State). J. Struct. Geol. 16, 323–335. Archanjo, C.J., Macedo, J.W.P., Galindo, A.C., Ara´ujo, M.G.S., 1998. Brasiliano crustal extension and emplacement fabrics of

28

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

the mangerite-charnockite pluton of Umarizal, north-east Brazil. Precambrian Res. 87, 19–32. Ashwal, L.D., 1993. Anorthosites. Springer, Heidelberg, 422 p. Aspler, L.B., Wisotzek, I.E., Chiarenzelli, J.R., Losonczy, M.F., Cousens, B.L., McNicoll, V.J., Davis, W.J., 2001. Paleoproterozoic intracratonic basin processes, from breakup of Kenorland to assembly of Laurentia: Hurwitz Basin, Nunavut, Canada. Sediment. Geol. 141/142, 287–318. BABEL Working Group, 1993. Integrated seismic studies of the Baltic shield using data in the Gulf of Bothnia region. Geophys. J. Int. 112, 305–324. Barnichon, J.D., Hanevith, H.B., Hoffer, B., Charlier, R., Jongmans, D., Duchesne, J.C., 1999. The deformation of the Egersund–Ogna anorthosite massif, south Norway: finiteelement modelling of diapirism. Tectonophysics 303, 109– 130. Barth, M.G., Rudnick, R.L., Horn, I., McDonough, W.F., Spicuzza, M.J., Valley, J.V., Haggerty, S.E., 2001. Geochemistry of xenolithic eclogites from West Africa. Part I: A link between low MgO eclogites and Archean crust formation. Geochim. Cosmochim. Acta 65, 1499–1527. Bartley, J.K., Semikhatov, M.A., Kaufman, A.J., Knoll, A.H., Pope, M.C., Jacobsen, S.B., 2001. Global events across the Mesoproterozoic–Neoproterozoic boundary: C and Sr evidence from Siberia. Precambrian Res. 111, 165–202. B´edard, J.H., 2001. Parental magmas of the Nain Plutonic Suite anorthosites and mafic cumulates: a trace element modelling approach. Contrib. Mineral. Petrol. 141, 747–771. Bennett, V.C., DePaolo, D.J., 1987. Proterozoic crustal history of the western United States as determined by neodymium isotopic mapping. Geol. Soc. Am. Bull. 99, 674–685. Bercovici, D., 1995. On the purpose of toroidal flow in a convecting mantle. Geophys. Res. Lett. 22, 3107–3110. Bercovici, D., 2003. The generation of plate tectonics from mantle convection. Earth Planet. Sci. Lett. 205, 107– 121. Bercovici, D., Ricard, Y., Richards, M.A., 2000. The relation between mantle dynamics and plate tectonics: a primer. In: Richards, M.A., Gordon, R., Van der Hilst, R. (Eds.), The History and Dynamics of Global Plate Motion. Geophysical Monograph, vol. 121. American Geophysical Union, pp. 5–46. Bettencourt, J.S., Tosdal, R.M., Leite Jr., W.B., Payolla, B.L., 1999. Mesoproterozoic rapakivi granites of the Rondˆonia Tin Province, southwestern border of the Amazonian craton, Brazil. I. Rreconnaissance U–Pb geochronology and regional implications. Precambrian Res. 95, 41–67. Beunk, F.F., Page, L.M., 2001. Structural evolution of the accretional continental margin of the Paleoproterozoic Svecofennian orogen in southern Sweden. Tectonophysics 339, 67–92. Bierlein, F.P., 1995. Rare-earth elements geochemistry of clastic and chemical metasedimentary rocks associated with hydrothermal sulphide mineralisation in the Olary Block, South Australia. Chem. Geol. 122, 77–98. Bingen, B., van Breemen, 1998. U–Pb monazite ages in amphiboliteto granulite-facies orthogneiss reflect hydrous mineral breakdown reactions: Sveconorwegian Province of SW Norway. Contrib. Mineral. Petrol. 132, 336–353.

Bjerrum, C.J., Canfield, D.E., 2002. Ocean productivity before about 1.9 Gyr ago limited by phosphorus adsorption onto iron oxides. Nature 417, 159–162. Bogdanova, S., Page, L.M., Skridlaite, G., Taran, L.M., 2001. Proterozoic tectonothermal history in the western part of the East European craton: 40 Ar/39 Ar geochronological constraints. Tectonophysics 339, 39–66. Bolle, O., Trindade, R.I.F., Bouchez, J.L., Duchesne, J.C., 2002. Imaging downward granitic magma in the Rogaland igneous complex, SW Norway. Terra Nova 14, 87–92. Bridgwater, D., Sutton, J., Watterson, J., 1974. Crustal downfolding associated with tectonic activity. Tectonophysics 21, 57–77. Brown, P.E., Dempster, T.J., Hutton, D.H.W., Becker, S.M., 2002. Extensional tectonics and mafic plutons in the Ketilidian rapakivi granite suite of South Greenland. Lithos 96, 1–13. Brown, P.E., Fallick, A.E., Becker, S.M., Dempster, T.J., Hutton, D.H.W., 1999. The rapakivi granites of South Greenland-stable isotope characteristics of their black and white facies and the nature of their protolith. Lithos 46, 485–504. Brunet, D., Machetel, P., 1998. Large-scale tectonic features induced by mantle avalanches with phase, temperature and pressure lateral variations of viscosity. J. Geophys. Res. B103, 4929–4945. Buchan, K.L., Mortensen, J.K., Card, K.D., 1993. Northeast-trending early Proterozoic dykes of southern Superior Province: multiple episodes of emplacement recognized from integrated paleomagnetism and U–Pb geochronology. Can. J. Earth Sci. 30, 1286–1296. Buchan, K.L., Ernst, R.E., Hamilton, M.A., Mertanen, S., Pesonen, L.J., Elming, S.A., 2001. Rodinia: the evidence from integrated paleomagnetism and U–Pb geochronology. Precambrian Res. 110, 9–32. Buick, R., Des Marais, D.J., Knoll, A.H., 1995. Stable isotope compositions of carbonates from the Mesoproterozoic Bangemall Group, northwestern Australia. Chem. Geol. 123, 153–171. Canfield, D.E., 1998. A new model for Proterozoic ocean chemistry. Nature 396, 450–453. Chadwick, B., Garde, A.A., Grocott, J., McCaffrey, K.J.W., Hamilton, M.A., 2000. Ketilidian structure and the rapakivi suite between Lindenow Fjord and Kap Farvel, South-East Greenland. Geol. Greenland Surv. Bull. 286, 50–59. Clemens, J.D., 2003. S-type granitic magmas—petrogenetic issues, models and evidence. Earth Sci. Rev. 61, 1–18. Clemens, J.D., Holloway, J.R., White, A.J.R., 1986. Origin of an A-type granite: experimental constraints. Am. Mineral. 71, 317–324. Collins, W.J., Beams, S.D., White, A.J.R., Chappell, B.W., 1982. Nature and origin of A-type granites with particular reference to southeast Australia. Contrib. Mineral. Petrol. 80, 189–200. Condie, K.C., 1998. Episodic continental growth and supercontinents: a mantle avalanche connection? Earth Planet. Sci. Lett. 163, 97–108. Connelly, J.N., Ryan, A.B., 1999. Age and tectonic implications of Paleoproterozoic granitoid intrusions within the Nain Province near Nain, Labrador. Can. J. Earth Sci. 36, 833–853. Creaser, R.A., Price, R.C., Wormald, R.J., 1991. A-type granite revisited: assessment of a residual-source model. Geology 19, 163–166.

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34 D’Acremont, E., Leroy, S., Burov, E.B., 2003. Numerical modelling of a mantle plume: the plume head-lithosphere interaction in the formation of an oceanic large igneous province. Earth Planet. Sci. Lett. 206, 379–396. De Bremond d’Ars, J., Jaupart, C., Sparks, R.S.J., 1995. Distribution of volcanoes in active margins. J. Geophys. Res. B100, 20421–20432. Dempster, T.J., Hutton, D.H.W., Harrison, T.N., Brown, P.E., Jenkin, G.R.T., 1991. Textural evolution of the rapakivi granites, south Greenland—Sr, O and H isotopic investigations. Contrib. Mineral. Petrol. 107, 459–471. Dempster, T.J., Jenkin, G.R.T., Rogers, G., 1994. The origin of rapakivi texture. J. Petrol. 35, 963–981. Diot, H., Bolle, O., Lambert, J.M., Launeau, P., Duchesne, J.C., 2003. The Tellnes ilmenite deposit (Rogaland, South Norway): magnetic and petrofabric evidence for emplacement of a Tienriched noritic crystal mush in a fracture zone. J. Struct. Geol. 25, 481–501. D¨orr, W., Belka, Z., Dirk, M., Scatstok, J., Valverde-Vaquero, P., Wiszniewska, J., 2002. U–Pb and Ar–Ar geochronology of anorogenic granite magmatism of the Mazury complex, NE Poland. Precambrian Res. 119, 101–120. Dubuffet, F., Rabinowicz, M., Monnereau, M., 2000. Multiple scales in mantle convection. Earth Planet. Sci. Lett. 178, 351–366. Duchesne, J.C., Li´egeois, J.P., Vander Auwera, J., Longhi, J., 1999. The crustal tongue melting model and the origin of massive anorthosites. Terra Nova 11, 100–105. Duebendorfer, E.M., Christensen, C., 1995. Synkinematic (?) intrusion of the “anorogenic” 1425 Ma Beer Bottle Pass pluton, southern Nevada. Tectonics 14, 168–184. Duebendorfer, E.M., Chamberlain, K.R., Jones, C.S., 2001. Paleoproterozoic tectonic history of the Mojave–Yavapai boundary zone: perspective from the Cerbat Mountains, northwestern Arizona. Geol. Soc. Am. Bull. 113, 575–590. Duncan, R.A., Richards, M.A., 1991. Hotspots, mantle plumes, flood basalts and true polar wander. Rev. Geophys. 29, 31–50. Ehlers, C., Ehlers, M., 1977. Shearing and multiple intrusion in the ˚ diabases of Aland Archipelago, SW Finland. Geol. Surv. Finland Bull. 289, 38. Eklund, O., Shebanov, A.D., 1999. The origin of rapakivi texture by sub-isothermal decompression. Precambrian Res. 95, 129– 146. Eklund, O., Konopelko, D., Rutanen, H., Fr¨ojd¨o, S., Shebanov, A.D., 1998. 1.8 Ga Svecofennian post-collisional shoshonitic magmatism in the Fennoscandian shield. Lithos 45, 87–108. Elliott, B.A., 2001. Crystallization conditions of the Wiborg rapakivi batholith, SE Finland: an evaluation of amphibole and biotite mineral chemistry. Mineral. Petrol. 72, 305–324. Elming, S.A., Mikhilova, N.P., Kravchenko, S., 2001. Paleomagnetism of Proterozoic rocks from the Ukrainian Shield: new tectonic reconstruction of the Ukrainian and Fennoscandian shields. Tectonophysics 339, 19–38. Elston, D.P., Enkin, R.J., Baker, J., Kisilevsky, D.K., 2002. Tightening the belt: paleomagnetic–stratigraphic constraints on deposition, correlation and deformation of the Middle Proterozoic (ca 1.4 Ga) Belt-Purcell Supergroup, United States and Canada. Geol. Soc. Am. Bull. 114, 619–638.

29

Emslie, R.F., 1978. Anorthosite massifs; rapakivi granites and Late Proterozoic rifting of North America. Precambrian Res. 7, 61–98. Emslie, R.F., 1991. Granitoids of rapakivi granite-anorthosite and related associations. Precambrian Res. 51, 173–191. Emslie, R.F., Hamilton, M.A., Thi´erault, R.J., 1994. Petrogenesis of a Mid-Proterozoic Anorthosite–Mangerite–Charnockite–Granite (AMCG) complex: isotopic and chemical evidence from the Nain Plutonic Suite. J. Geol. 102, 539–558. Erikson, P.G., Martins-Neto, M.A., Nelson, D.R., Asper, L.B., Chiarenzelli, J.R., Catuneau, O., Sarkar, S., Alterman, W., Rautenbach, C.J. de W., 2001. An introduction to Precambrian basins: their characteristics and genesis. Sediment. Geol. 141–142, 1–35. Esipchuk, K., Skobelev, V., 1998. Mineralogy of the Korosten rapakivi granites (Ukrainian shield). In: Proceeding of the Institute of Fundamental Studies, Kyev, Znannya, pp. 45–56. Flinn, D., 1962. On folding during three-dimensional progressive deformation. Geol. Soc. Lond. Q. J. 118, 385–433. Foley, S.F., Venturelli, G., Green, D.H., Toscani, L., 1987. The ultrapotassic rocks: characteristics, classification and constraints for petrogenetic models. Earth Sci. Rev. 24, 81–134. Forsyhth, D., Uyeda, S., 1975. On the relative importance of the driving forces of plate motion. Geophys. J. Royal Astronom. Soc. 43, 163–200. Frank, T.D., Lyons, T.W., 1998. “Molar-tooth” structures: a geochemical perspective on a Proterozoic enigma. Geology 26, 683–686. Frank, T.D., Lyons, T.W., Lohmann, K.C., 1997. Isotopic evidence for the paleoenvironmental evolution of the Mesoproterozoic Helena Formation, Belt Supergroup, Montana, USA. Geochim. Cosmochim. Acta 61, 5027–5041. Frost, C.D., Frost, B.D., 1997. Reduced rapakivi-type granites: the tholeiite connection. Geology 25, 647–650. Frost, C.D., Chamberlain, K.R., Frost, B.R., Scoates, J.S., 2000. The 1.76-Ga Horse Creek anorthosite complex, Wyoming: a massif anorthosite emplaced late in the Medicine Bow orogeny. Rocky Mountain Geol. 35, 71–90. Fuhrmam, M.L., Frost, B.R., Lindsley, D.H., 1988. Crystallization conditions of the Sybille monzosyenite Laramie anorthosite complex, Wyoming. J. Petrol. 29, 699–729. Funck, T., Louden, K.E., 1999. Wide-angle seismic transect across the Torngat Orogen, northern Labrador: evidence for a Proterozoic crustal root. J. Geophys. Res. B104, 7463–7480. Gable, C.W., O’Connell, R.J., Travis, B.J., 1991. Convection in 3 dimensions with surface plates. Generation of toroidal flow. J. Geophys. Res. B96, 8391–8405. Galindo, A.C., Dall’Agnol, R., McReath, I., Lafon, J.M., Texeira, N., 1995. Evolution of Brasiliano-age granitoid types in a shear environment. Umarizal Caraubas region, Rio Grande do Norte, north-east Brazil. J. S. Am. Earth Sci. 8, 79–95. Garde, A.A., Hamilton, M.A., Chadwick, B., Grocott, J., McCaffrey, K.J.W., 2002. The Ketilidian orogen of South Greenland: geochronology, tectonics, magmatism, and fore-arc accretion during Palaeoproterozoic oblique convergence. Can. J. Earth Sci. 39, 765–793. Gaudette, H.E., Mendoza, V.E., Hurley, P.M., Fairbain, H.W., 1977. Geology and age of the Parguaza rapakivi granite, Venezuela. Geol. Soc. Am. Bull. 89, 1335–1340.

30

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

Gaudette, H.E., Olszewski, W.J., Santos, J.O.S., 1996. Geochronology of Precambrian rocks from the northern part of the Guiana Shield, State of Roraima, Brazil. J. S. Am. Earth Sci. 9, 183– 195. Gordon, R.G., Jurdy, D.M., 1986. Cenozoic global plate motions. J. Geophys. Res. B91, 12389–12406. Gower, C.F., Krogh, T.E., 2002. A U–Pb geochronological review of the Proterozoic history of the eastern Grenville Province. Can. J. Earth Sci. 39, 795–829. Gower, C.F., Tucker, R.D., 1994. Distributions of pre-1400 Ma crust in the Grenville Province: implications for rifting in the Laurentia–Fennoscandia during geon 14. Geology 22, 827–830. Grign´e, C., Labrosse, S., 2001. Effects of continents on Earth cooling: thermal blanketing and depletion in radioactive elements. Geophys. Res. Lett. 28, 2707–2710. Grocott, J., Garde, A.A., Chadwick, B., Cruden, A.R., Swager, C., 1999. Emplacement of rapakivi granites by floor depression and roof uplift in the Palaeoproterozoic Ketilidian orogen, South Greenland. J. Geol. Soc. Lond. 156, 15–24. Gurnis, M., 1988. Large scale mantle convection and the aggregation and dispersal of supercontinents. Nature 332, 695–699. Haapala, I., Lukkari, S., 2005. Petrological and geochemical evolution of the Kymi stock, a topaz granite cupola within the Wiborg rapakivi batholith, Finland. Lithos 80, 347–362. Haapala, I., R¨am¨o, O.T., 1999. Rapakivi granites and related rocks. Precambrian Res. 95, 1–7. Haapala, I., 1997. Magmatic and postmagmatic processes in tinmineralized granites: topaz-bearing leucogranite in the topazbearing leucogranite of the Eurajoki rapakivi stock, Finland. J. Petrol. 38, 1645–1659. Haapala, I., R¨am¨o, O.T., Frindt, S., 2005. Comparison of Proterozoic and Phanerozoic rift-related basaltic-granitic magmatism. Lithos 80, 1–32. Hager, B., O’Connell, R., 1978. Subduction zone dip angles and flow driven by plate motion. Tectonophysics 50, 111–133. Hall, S.M., Veizer, J., 1996. Geochemistry of Precambrian carbonates. VII. Belt Supergroup, Montana and Idaho, USA. Geochim. Cosmochim. Acta 60, 667–677. Halls, H.C., 1991. The Matachewan dyke swarm, Canada: an early Proterozoic magnetic field reversal. Earth Planet. Sci. Lett. 105, 279–292. Hamilton, M.A., Ryan, A.B., Emslie, R.F., Ermanovics, I.F., 1998. Identification of Palaeoproterozoic anorthisitic and monzonitic rocks in the vicinity of the Mesoroterozoic Nain Plutonic Suite, Labrador. U–Pb evidence, radiogenic age isotopic studies. Geol. Surv. Can. Curr. Res. 11, 23–40. Hanmer, S., Bowring, S., van Breemen, O., Parrish, R., 1992. Great Slave Lake shear zone, NW Canada: mylonitic record of Early Proterozoic continental convergence, collision and indentation. J. Struct. Geol. 14, 757–773. Hanmer, S., Corrigan, D., Pehrsson, S., Nadeau, L., 2000. SW Grenville Province, Canada: the case against post-1.4 Ga accretionary tectonics. Tectonophysics 319, 33–51. Harlan, S.S., Geissman, J.W., Premo, W.R., 2003. Paleomagnetism and geochronology of an Early Proterozoic quartz diorite in the southern Wind River Range, Wyoming, USA. Tectonophysics 362, 105–122.

Harris, N., Vance, D., Ayres, M., 2000. From sediment to granites: timescales of anatexis in the upper crust. Chem. Geol. 162, 155–167. Henriksen, N., Higgins, A.K., Kalsbeek, F., Pulvertaft, T.C.R., 2000. Greenland from archaean to quaternary. Descriptive text to the geological map of Greenland. Geol. Greenland Surv. Bull. 185, 93. Hoffmann, P., Schrag, D.P., 2002. The snowball Earth hypothesis: testing the limits of global change. Terra Nova 14, 129–155. Hofmann, A.W., 1997. Mantle geochemistry: the message from oceanic volcanism. Nature 385, 219–229. Hogan, J.P., Price, J.P., Gilbert, M.C., 1998. Magma traps and driving pressure: consequences for pluton shape and emplacement in an extensional regime. J. Struct. Geol. 20, 1155–1168. Holland, D.H., 1973. The oceans, a possible source of iron in ironformations. Econ. Geol. 68, 1169–1172. Holm, D., Lux, D., 1998. Depth of emplacement and tilting of the Middle Proterozoic (1470 Ma) Wolf River batholith, Wisconsin: 40 Ar/39 Ar thermochronologic constraints. Can. J. Earth Sci. 35, 1143–1151. Hutton, D.H.W., Dempster, T.J., Brown, P.E., Becker, S.M., 1990. A new mechanism of granite emplacement: rapakivi intrusions in active extensional shear zones. Nature 343, 452–454. Isley, A.E., Abbott, D.H., 1999. Plume-related mafic volcanism and the deposition of banded iron formation. J. Geophys. Res. B104, 15461–15477. Jaeger, J.C., 1969. Elasticity, Fracture and Flow with Engineering and Geological Applications. Methuen, London, 268 pp. Jahn, B.M., Wu, F., Chen, B., 2000. Granitoids of the Central Asian Orogenic Belt and continental growth in the Phanerozoic. Trans. R. Soc. Edinb. Earth Sci. 91, 181–193. Jensen, S.L., Janik, T., Thybo, H., POLONAISE Profile P1 Working Group, 1997. Seismic structure of the Palaeozoic Platform along POLONAISE’97 profile P1 in northwestern Poland. Tectonophysics 314, 123–143. Jensen, S.L., Thybo, H., The POLONAISE’97 Working Group, 2002. Moho topography and lower crustal wide-angle reflectivity around the TESZ in southern Scandinavia and northeastern Europe. Tectonophysics 360, 187–213. John, W.F., Ketchum, J.W.F., Culshaw, N.G., Barr, S.M., 2002. Anatomy and orogenic history of a Paleoproterozoic accretionary belt: the Makkovik Province, Labrador, Canada. Can. J. Earth Sci. 39, 711–730. Jones, C.H., Kanamori, H., Roecker, S.W., 1994. Missing roots and mantle “drips”: regional Pn and teleseismic arrival times in the southern Sierra Nevada and vicinity, California. J. Geophys. Res. B99, 4567–4601. Kah, L.C., Bartley, J.K., 2001. Rodinia and the Mesoproterozoic earth–ocean system. Precambrian Res. 111, 1–3. Karhu, J.A., Holland, H.D., 1996. Carbon isotopes and the rise of atmospheric oxygen. Geology 24, 867–870. Karlstrom, K.E., Houston, R.S., 1984. The Cheyenne belt—analysis of a Proterozoic suture in southern Wyoming. Precambrian Res. 25, 415–446. Karlstrom, K.E., Ah¨all, K.I., Harlan, S.S., Williams, M.L., McLelland, J., Geissman, J.W., 2001. Long-lived (1.8–1.0 Ga) convergent orogen in southern Laurentia, its extensions to Australia

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34 and Fennoscandia, and implications for refining Rodinia. Precambrian Res. 111, 5–30. Karlstrom, K.E., Harlan, S.S., Williams, M.L., McLelland, J., Geisslan, J.W., Ah¨all, K.I., 1999. Refining Rodinia: geologic evidence for the Australian–western US connection in the Proterozoic. GSA Today 9, 1–7. Kerr, A., 1989. Geochemistry of the Trans-Labrador granitoid belt, Canada. A quantitative comparative study of a Proterozoic batholith and possible Phanerozoic counterparts. Precambrian Res. 45, 1–17. Ketchum, J.K.F., Barr, S.M., Culshaw, N.G., White, C.E., 2001. U–Pb ages of granitoid rocks in the northwestern Makkovik Province, Labrador: evidence for 175 million years of episodic synorogenic and postorogenic plutonism. Can. J. Earth Sci. 38, 359–372. Ketchum, J.K.F., Culshaw, N.G., Barr, S.M., 2002. Anatomy and orogenic history of a Paleoproterooic accretionary belt: the Makkovik Province, Labrador, Canada. Can. J. Earth Sci. 39, 711–730. Kirby, E., Karlstrom, K.E., Andronicos, C., Dallmeyer, R.D., 1995. Tectonic setting of the Sandia pluton: an orogenic 1.4 Ga pluton in New Mexico. Tectonics 14, 185–201. Korja, A., Heikkinen, P.J., 1995. Proterozoic extensional tectonics of the central Fennoscandian Shield: results from the Baltic and Bothnian echoes from the Lithosphere experiments. Tectonics 14, 504–517. Korja, A., Heikkinen, P., Aaro, S., 2001. Crustal structure of the northern Baltic Sea paleorift. Tectonophysics 331, 341–358. Korsman, K., Korja, T., Pajunen, M., Virransalo, P., and GGT/SVEKA Working Group, 1999. The GGT/SVEKA Transect: structure and evolution of the continental crust in the Paleoproterozoic Svecofennian orogen in Finland. Int. Geol. Rev. 41, 287–333. Kosunen, P., 1999. The rapakivi granite plutons af Bodom and Obbn¨as, southern Finland: petrography and geochemistry. Bull. Geol. Soc. Finland 71, 275–304. Kosunen, P., R¨am¨o, O.T., Vaasjoki, M., 2002. The Bodom and Obbn¨as rapakivi granites, southern Finland: distinct composition implies a Paleoproterozoic terrane boundary. In: Lahtinen, R., Korja, A., Arhe, K., Eklund, O., Hjelt, S.E., Pesonen, L.J. (Eds.), Lithosphere 2002, Proceedings of the Second Symposium on the Structure, Composition and Evolution of the Lithosphere in Finland. Institute of Seismology, University of Helsinki, Report S-42, Espoo, Finland, pp. 49–53. Kyser, T.K., Kotzer, T.G., 1995. Petrogenesis of the Proterozoic Athabasca Basin, northern Saskatchewan, Canada, and its relation to diagenesis, hydrothermal uranium mineralization and paleohydrogeology. Chem. Geol. 120, 45–89. Landenberger, B., Collins, W.J., 1996. Derivation of A-type granites from a dehydrated charnockitic lower crust: evidence from the Chaelundi complex, eastern Australia. J. Petrol. 37, 145–170. Lenharo, S.R., Moura, M.A., Botelho, N.F., 2002. Petrogenetic and mineralization processes in Paleo- to Mesoproterozoic rapakivi granites: examples from Pitinga and Goias, Brazil. Precambrian Res. 119, 277–299. Lindh, A., Andersson, U.B., Lundqvist, T., Claesson, S., 2001. Evidence of crustal contamination of mafic rocks associated with

31

Rapakivi rocks: an example from the Nordingr˚a complex, Central Sweden. Geol. Mag. 138, 371–386. Lithgow-Bertelloni, M., Richards, M., Ricard, Y., O’Connell, R., Engebreston, D., 1993. Toroidal–poloidal partitioning of plate motions since 120 Ma. Geophys. Res. Lett. 20, 375–378. Longhi, J., Vander Auwera, J., Fram, M.S., Duchesne, J.C., 1999. Some phase equilibrium constraints on the origin of Proterozoic (massif) anorthosites and related rocks. J. Petrol. 40, 339–362. Loosveld, R.J.H., 1989. The synchronism of crustal thickening and low pressure facies metamorphism in the Mount Isa Inlier, Australia. 2. Fast convective thinning of mantle lithosphere during crustal thickening. Tectonophysics 165, 191–218. Louden, K.E., Fan, J., 1998. Crustal structures of Grenville, Makkovik, and southern Nain provinces along the Lithoprobe ECSOOT Transect: regional seismic refraction and gravity models and their tectonic implications. Can. J. Earth Sci. 35, 711–730. Lowman, J.P., Jarvis, G.T., 1999. Effects of mantle heat source distribution on supercontinent stability. J. Geophys. Res. B104, 12733–12746. Lucas, S.B., Green, A., Hajnal, Z., White, D., Lewry, J., Ashton, K., Weber, W., Clowes, R., 1993. Deep seismic profile across a Proterozoic collision zone: surprises at depth. Nature 363, 339–342. Luepke, J.J., Lyons, T.W., 2001. Pre-Rodinian (Mesoproterozoic) supercontinental rifting along the western margin of Laurentia: geochemical evidence from the Belt-Purcell Supergroup. Precambrian Res. 111, 79–90. Lyons, T.W., Kah, L.C., Gellatly, A.M., 2003. The Precambrian sulfur isotope record of evolving atmospheric oxygen. In: Erikson, P.G., Altermann, W., Nelson, D.R., Mueller, W.U., Catuneanu, O. (Eds.), The Precambrian Earth: Tempos and Events. Elsevier, pp. 421–439. Lyons, T.W., Luepke, J.J., Schreiber, M.E., Zieg, G.A., 2000. Sulfur geochemical constraints on Mesoproterozoic restricted marine deposition: lower Belt Supergroup, northwestern United States. Geochim. Cosmochim. Acta 64, 427–437. Machetel, P., Weber, P., 1991. Intermittent layered convection in a model mantle with an endothermic phase change at 670 km depth. Nature 350, 55–57. Marheine, D., Valverde-Vaquero, P., 2002. Recognition of preSveconorwegian cooling ages in the eastern European craton, Central Poland: new 40 Ar–39 Ar dating in the 1.8 Ga Kampinos complex. Precambrian Res. 118, 169–177. Martins-Neto, M.A., 2000. Tectonics and sedimentation in a Paleo/Mesoproterozoic rift-sag basin (Espinhac¸o basin, southeastern Brazil). Precambrian Res. 103, 147–173. McCaffrey, K.J.W., Grocott, J., Garde, A.A., Hamilton, M.A., 2004. Attachment formation during partitioning of oblique convergence in the Ketilidian orogen, south Greenland. In: Grocott, J., McCaffrey, K.J.W., Taylor, G., Tikoff, B. (Eds.), Vertical Coupling and Decoupling in the Lithosphere, vol. 227. Geological Society of London, pp. 231–248 (special publication). Meert, J.G., Powell, C. McA., 2001. Assembly and break-up of Rodinia: introduction to the special volume. Precambrian Res. 110, 1–8. Meert, J.G., Stuckey, W., 2002. Revisiting the paleomagnetism of the 1.476 Ga St. Francois Mountains Igneous Province, Missouri, Tectonics 21, 10.1029/2000TC001265.

32

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

Meissner, R., Tanner, B., 1993. From collision to collapse: phases of lithospheric evolution as monitored by seismic records. Phys. Earth Planet. Interiors 79, 75–86. Menuge, J.F., Brewer, T.S., Seeger, C.M., 2002. Petrogenesis of metaluminous A-type rhyolites from the St. Francois Mountains, Missouri and the Mesoproterozoic evolution of the southern Laurentian margin. Precambrian Res. 113, 269–291. Mimouni, A., Rabinowicz, M., 1988. The old continental shield stability related to mantle convection. Geophys. Res. Lett. 15, 68–71. Miron Valdespino, O.E., Constanzo Alvarez, V., 1997. Paleomagnetic and rock magnetic evidence for an inverse zoning in the Parguaza batholith (southwestern Venezuela) and its implications about tectonics of the Guyana shield. Precambrian Res. 85, 1–25. Moakhar, M.O., Elming, S.A., 2000. A paleomagnetic analysis of rapakivi intrusions and related dykes in the Fennoscandian Shield. Phys. Chem. Earth A25, 489–494. Monnereau, M., Qu´er´e, S., 2001. Spherical shell models of mantle convection with tectonic plates. Earth Planet. Sci. Lett. 184, 575–587. Monnereau, M., Rabinowicz, M., Arquis, E., 1993. Mechanical erosion and reheating of the lithosphere: a numerical model for hotspot swells. J. Geophys. Res. B98, 809–823. Moores, E.M., 1991. Southwest US-East Antarctic (SWEAT) connection: a hypothesis. Geology 19, 425–428. Morgan, J.W., Stein, H.J., Hannah, J.L., Markey, R.J., Wiszniewska, J., 2000. Re–Os study of Fe–Ti–V oxide and Fe–Cu–Ni sulfide deposits, Suwalki anorthosite massif, northeast Poland. Mineral. Deposita 35, 391–401. Morgan, W.J., 1971. Convection plumes in the lower mantle. Nature 230, 42–43. Noel, M.E., Andronicos, C.L., 2002. Structure and metamorphism of the Eolus granite, Needle Mountains, Colorado: implications for regional metamorphism and Proterozoic tectonics in southern Colorado, GSA Meeting Paper, 107-3. Nyblade, A.A., Pollack, H.N., 1993. A global analysis of heat flow from Precambrian terrains: implications for the thermal structure of Archean and Proterozoic lithosphere. J. Geophys. Res. B98, 12207–12218. Nyman, M.W., Karlstrom, K.E., 1997. Pluton emplacement processes and tectonic setting of the 1.42 Ga Signal batholith, SW USA: important role of crustal anisotropy during regional shortening. Precambrian Res. 82, 237–263. Nyman, M.W., Karlstrom, K.E., Kirby, E., Graubard, C.M., 1994. Mesoproterozoic contractional orogeny in western North America: evidence from 1.4 Ga plutons. Geology 22, 901–904. Patel, S.C., Frost, C.D., Frost, B.R., 1999. Contrasting responses of Rb–Sr systematics to regional and contact metamorphism, Laramie Mountains, Wyoming, USA. J. Metamorph. Geol. 17, 259–269. Pati˜no Douce, A.E., 1996. Effects of pressure and H2 O content on the composition of primary crustal melts. Trans. R. Soc. Edinb. Earth Sci. 87, 11–21. Payolla, B.L., Bettencourt, J.S., Kozuch, M., Leite Jr., W.B., Fetter, A.H., Van Schmus, W.R., 2002. Geological evolution of the basement rocks in the east-central part of the Rondˆonia Tin Province,

SW Amazonia craton, Brazil: U–Pb and Sm–Nd isotopic constraints. Precambrian Res. 119, 141–169. Pearce, J.A., Harris, N.B.W., Tindle, A.G., 1984. Trace element discrimination diagrams for the tectonic interpretation of granitic rocks. J. Petrol. 25, 956–983. Persson, A.I., 1999. Absolute (U–Pb) and relative age determinations of intrusive rocks in the Ragunda rapakivi complex, central Sweden. Precambrian Res. 95, 109–127. Pesonen, L.J., Mertanen, S., 2002. Paleomagnetic configuration of continents during the Proterozoic. In: Lahtinen, R., Korja, A., Arhe, K., Eklund, O., Hjelt, S.E., Pesonen, L.J. (Eds.), Lithosphere 2002, Proceedings of the Second Symposium on the Structure, Composition and Evolution of the Lithosphere in Finland. Institute of Seismology, University of Helsinki, Report S-42, Espoo, Finland, pp. 103–109. ˚ Mertanen, S., Pisarevsky, S., D’AgrellaPesonen, L.J., Elming, S.A., Filho, M.S., Meert, J.G., Schmidt, P.W., Abrahamsen, N., Bylund, G., 2003. Palaeomagnetic configuration of continents during the Proterozoic. Tectonophysics 375, 289– 324. Petford, N., Cruden, A.R., McCaffrey, K.J.W., Vigneresse, J.L., 2000. Granite magma formation, transport and emplacement in the Earth’s crust. Nature 408, 669–673. Piper, J.D.A., Stearn, J.E.F., 1977. Palaeomagnetism of the dyke swarms of the Gardar Igneous Province, south Greenland. Phys. Earth Planet. Interiors 14, 345–358. Piper, J.D.A., 1980. Palaeomagnetic study of the Swedish Rapakivi suite: Proterozoic tectonics of the Baltic shield. Earth Planet. Sci. Lett. 46, 443–461. Piper, J.D.A., 2000. The Neoproterozoic supercontinent: Rodinia or Palaeopangaea? Earth Planet. Sci. Lett. 176, 131–146. Pitcher, W.S., 1993. The Nature and Origin of Granite. Chapman and Hall, London, 322 p. Powell, C.McA., Pisarevsky, S.A., 2002. Late Neoproterozoic assembly of east Gondwanaland. Geology 30, 3–6. Puura, V., Floden, T., 1999. Rapakivi-granite-anorthosite magmatism—a way of thinning and stabilisation of the Svecofennian crust, Baltic Sea Basin. Tectonophysics 305, 75–92. R¨am¨o, O.T., Haapala, I., 1995. One hundred years of rapakivi granite. Mineral. Petrol. 52, 129–185. R¨am¨o, O.T., Haapala, I., 2003. Petrogenesis of the rapakivi granites: what do the classic Fennoscandian suites imply. In: R¨am¨o, O.T., Kosunen, P.J., Lauri, L.S., Karhu, J.H. (Eds.), Granitic systems: State of the Art and Future. Helsinki University Press, Helsinki, pp. 84–88. R¨amo, O.T., 1999. Sr isotopic composition of Finnish rapakivi granites: the Suomenniemi batholith. Bull. Geol. Soc. Finland 71, 339–345. R¨am¨o, O.T., Haapala, I., Vaasjoki, M., Yu, J.H., Fu, H.Q., 1995. 1700 Ma Shachang complex, northeast China: Proterozoic rapakivi granite not associated with Paleoproterozoic orogenic crust. Geology 23, 815–818. R¨am¨o, O.T., Huhma, H., Kirs, J., 1996. Radiogenic isotopes of the Estonian and Latvian rapakivi granite suites: new data from the concealed Precambrian of the East European craton. Precambrian Res. 79, 209–226.

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34 R¨am¨o, O.T., McLemore, V.T., Hamilton, M.A., Kosunen, P.J., Heizler, M., Haapala, I., 2003. Intermittent 1630–1220 Ma magmatism in central Mazatzal province: new geochronologic piercing points and some tectonic implications. Geology 31, 335–338. R¨am¨o, O.T., Vaasjoki, M., M¨antt¨ari, I., Elliott, B.A., Nironen, M., 2001. Petrogenesis of the post-kinematic magmatism of the central Finland granitoid complex. I. Radiogenic isotope constraints and implications for crustal evolution. J. Petrol. 42, 1971–1993. Ray, J.S., Veizer, J., Davis, W.J., 2003. C, O, Sr and Pb isotope systematics of carbonate sequences of the Vindhyan Supergroup, India: age, diagenesis, correlations and implications for global events. Precambrian Res. 121, 103–140. Rickard, M.J., Ward, P., 1981. Paleozoic crustal thickness in the southern part of the Lachlan orogen deduced from volcano and pluton-spacing geometry. J. Geol. Soc. Aust. 28, 19–32. Rivers, T., Corrigan, D., 2000. Convergent margin on southeastern Laurentia during the Mesoproterozoic: tectonic implications. Can. J. Earth Sci. 37, 359–383. Rivers, T., 1997. Lithotectonic elements of the Grenville Province: review and tectonic implications. Precambrian Res. 86, 117–154. Ross, G.M., 2000. Proterozoic stratigraphy of western Canada: a short review. In: Kyser, K. (Ed.), Tracing Fluid Histories of Sedimentary Basins, vol. 28. Mineralogical Association of Canada Short Course, pp. 211–224. Royse, K.R., Park, R.G., 2000. Emplacement of the Nain anorthosite: diapiric versus conduit ascent. Can. J. Earth Sci. 37, 1195– 1207. Royse, K.R., Noble, S.R., Tarney, J., Cadman, A.C., 1999. Countryrock contamination of marginal mafic granulites bordering the Nain Plutonic Suite: implication for mobilization of Sr during high grade contact metamorphism. Can. J. Earth Sci. 36, 985–997. Rudnick, R.L., 1992. Xenoliths—Samples of the lower continental crust. In: Fountain, D.M., Arculus, R., Kay, R. (Eds.), The Continental Lower Crust. Elsevier, Amsterdam, pp. 269–316. Rudnick, R.L., McDonough, W.F., O’Connell, R.J., 1998. Thermal structure, thickness and composition of continental lithosphere. Chem. Geol. 145, 399–416. Ryan, B., Phillips, E., Swetz, J., Machado, G., 1998. A tale of more than ten plutons. Geological survey, Newfoundland Department of Mines and Energy, Current Research Report 98-1, 143–171. Sadowski, G.R., Bettencourt, J.S., 1996. Mesoproterozoic tectonic correlations between eastern Laurentia and the western border of the Amazon craton. Precambrian Res. 76, 213–227. Santos, J.O.S., Potter, P.E., Reis, N.J., Hartmann, L.A., Fitcher, I.R., McNaughton, N.J., 2003. Age, source and regional stratigraphy of the Roraima Supergroup and Roraima-like outliers in northern South America based on U–Pb geochronology. Geol. Soc. Am. Bull. 115, 331–348. Savage, M.K., Li, L., Eaton, J.P., Jones, C.H., Brune, J.N., 1994. Earthquake refraction profiles of the root of the Sierra Nevada. Tectonics 13, 803–817. Scoates, J.S., Chamberlain, K.R., 1997. Orogenic to post-orogenic origin for the 1.76 Ga Horse Creek anorthosite complex, Wyoming, USA. J. Geol. 105, 331–343. Scoates, J.S., Frost, C.D., 1996. A strontium and neodynium isotopic investigation of the Laramie anorthosites, Wyoming, USA:

33

implications for magma chamber processes and the evolution of magma conduits in Proterozoic anorthosites. Geochim. Cosmochim. Acta 60, 95–107. Scott, D.L., Rawlings, D.J., Page, R.W., Tarlowski, C.Z., Idnurm, M., Jackson, M.J., Southgate, P.N., 2002. Basement framework and geodynamic evolution of the Palaeoproterozoic superbasins of north-central Australia: an integrated review of geochemical, geochronological and geophysical data. Aust. J. Earth Sci. 47, 341–380. Sederholm, J.J., 1891. Ueber die finnl¨andischen Rapakiwigesteine. Tschermaks Mineralogische und Petrographische Mitteilungen 12, 1–31. Shaw, C.A., Karlstrom, K.E., Williams, M.L., Jercinovic, M.L., McCoy, A.M., 2001. Electron-microprobe monazite dating of ca 1.71–1.63 Ga and ca 1.45–1.38 Ga deformation in the Homestake shear zone, Colorado: origin and early evolution of a persistent intracontinental tectonic zone. Geology 29, 739– 742. Shaw, C.A., Snee, L.W., Selverstone, J., Reed Jr., J.C., 1999. 40 Ar/39 Ar thermochronology of Mesoproterozoic metamorphism in the Colorado Front Range. J. Geol. 107, 49–67. Shebanov, A.D., Eklund, O., 1997. Interaction between acid and basic magmas in the Hammarudda Complex, SW Finland: P–T conditions obtained by mineralogical thermometry and barometry. Petrology 5, 141–166. Shen, Y., Canfield, D.E., Knoll, A.H., 2002. Middle Proterozoic ocean chemistry: evidence from the McArthur Basin, northern Australia. Am. J. Sci. 302, 81–109. Sims, P.K., Stein, H.J., Finn, C.A., 2002. New Mexico structural zone—an analogue to the Colorado mineral belt. Ore Geol. Rev. 21, 211–225. Sivoronov, A.A., Bobrov, A.B., Lysak, A.M., Malyuk, B.I., Koliy, V.D., Sveshnikov, K.I., 1998. Principal geological units and their tectonic evolution in the north-western part of the Ukrainian shield. Eurobridge Report, 18 p. Smith, A.D., Lewis, C., 1999. The planet beyond the plume hypothesis. Earth Sci. Rev. 48, 135–182. Smith, D.R., Noblett, J., Wobus, R.A., Unruh, D., Douglass, J., Beane, R., Davis, C., Goldman, S., Kay, G., Gustavson, B., Saltoun, B., Stewart, J., 1999. Petrology and geochemistry of late-stage intrusions of the A-type, Mid-Proterozoic Pikes Peak batholith (Central Colorado, USA): implications for petrogenetic models. Precambrian Res. 98, 271–305. Spikins, R.A., Foster, D.A., Kohn, B.P., Lister, G.S., 2002. Post-orogenic (<1500 Ma) thermal history of the PalaeoMesoproterozoic, Mt. Isa province, Australia. Tectonophysics 349, 327–365. Tackley, P.J., 2000. Self-consistent generation of tectonic plates in time-dependent, three-dimensional mantle convection simulations. 1. Pseudoplastic yielding. Geochem. Geophys. Geosyst. 1, 2000GC000036. Tao, W., Jarvis, G.T., 2002. The influence of continental surface area on the assembly time for supercontinents. Geophys. Res. Lett. 29, 10.1029/2001GL013712. Taran, L.N., Bogdanova, S.V., 2001. The Fennoscandia–Samartia junction in Belarus: new inferences from a PT study. Tectonophysics 339, 193–214.

34

J.L. Vigneresse / Precambrian Research 137 (2005) 1–34

Ten Brink, U., 1991. Volcano spacing and plate rigidity. Geology 19, 397–400. Thomas, D.J., Matthews, R.B., Sopuck, V., 2000. Athabasca basin (Canada) unconformity-type uranium deposits: exploration model, current mine developments and exploration directions. In: Cluer, J.K., Price, J.G., Struhsacker, E.M., Hardyman, R.F., Morris, C.L. (Eds.), Geology and Ore Deposits 2000, The Great Basin and Beyond, Geological Society of Nevada Symposium Proceedings, vol. 1, pp. 1–23. Thompson, A.B., 1982. Dehydration melting of pelitic rocks and the generation of H2 O-undersaturated granitic liquids. Am. J. Sci. 282, 1567–1595. Thompson, A.B., 1999. Some time-space relationships for crustal melting and granitic intrusion at various depths. In: Castro, A., Fernandez, C., Vigneresse, J.L. (Eds.), Understanding Granites: Integrating New and Classical Techniques, vol. 168. Geological Society of London, pp. 7–25 (special publication). Torvela, T., Ehlers, C., 2002. The south Finland shear zone—ductile shearing of Paleoproterozoic crust in SW Finland. In: Lahtinen, R., Korja, A., Arhe, K., Eklund, O., Hjelt, S.E., Pesonen, L.J. (Eds.), Lithosphere 2002, Proceedings of the Second Symposium on the Structure, Composition and Evolution of the Lithosphere in Finland. Institute of Seismology, University of Helsinki, Report S-42, Espoo, Finland, pp. 137–138. Trubitsyn, V.P., Mooney, W.D., Abbott, D.H., 2003. Cold cratonic roots and thermal blankets: how continents affect mantle convection. Int. Geol. Rev. 45, 479–496. Tyson, A.R., Morozova, E.A., Karlstrom, K.E., Chamberlain, K.R., Smithson, S.B., Dueker, K.G., Foster, C.T., 2002. Proterozoic Farwell Mountains-Lester Mountain suture zone, northern Colorado: subduction flip and progressive assembly of arcs. Geology 30, 943–946. Vaasjoki, M., 1996. Explanation to the geochronological map of southern Finland. The development of the continental crust with special reference to the Svecofennian orogeny, Geological Survey of Finland, Report of Investigation 35, 30 p. Vaasjoki, M., 1997. Rapakivi granites and other postorogenic rocks in Finland: their age and the lead isotopic composition of certain associated galena mineralizations. Geol. Surv. Finland Bull. 294, 64. Van der Hislt, R.D., Karason, H., 1999. Compositional heterogeneity in the bottom 1000 kilometers of Earth’s mantle: toward a hybrid convection model. Science 283, 1885–1888. Van Schmus, W.R., Bickford, M.E., et al. (Eds.), 1993. Transcontinental Proterozoic Provinces, Precambrian—Conterminous U.S., The Geology of North America C-2, Geological Society of America, 171–334.

Veizer, J., 1989. Strontium isotopes in seawater trough times. Annu. Rev. Earth Planet. Sci. 17, 141–167. Vigneresse, J.L., Clemens, J.D., 2000. Granitic magma ascent and emplacement: neither diapirism nor neutral buoyancy. In: Vendeville, B., Mart, Y., Vigneresse, J.L. (Eds.), Salt, Shale and Igneous Diapirs in and around Europe, vol. 174. Geological Society of London, pp. 1–19 (special publication). Vigneresse, J.L., Tikoff, B., Am´eglio, L., 1999. Modification of the regional stress field by magma intrusion and formation of tabular granitic plutons. Tectonophysics 302, 203–224. Vogt, P.R., 1974. Volcano spacing, fractures and the thickness of the lithosphere. Earth Planet. Sci. Lett. 21, 235–252. Vorma, A., 1976. On the petrochemistry of rapakivi granites with special reference to the Laitila massif, southwestern Finland. Geol. Surv. Finland Bull. 285, 98. Wardle, R.J., Hall, J., 2002. Proterozoic evolution of the northeastern Canadian Shield: Lithoprobe Eastern Canadian Shield OnshoreOffshore Transect (ECSOOT), introduction and summary. Can. J. Earth Sci. 39, 563–567. White, A.J.R., Chappell, B.W., 1988. Some supracrustal (S-type) granites of the Lachlan Fold belt. Trans. R. Soc. Edinb. Earth Sci. 79, 169–181. Wiebe, R.A., 1986. Lower cumulates nodules in Proterozoic dykes of the Nain complex, evidence for the origin of the anorthosites (Labrador, Canada). J. Petrol. 27, 1253–1275. Windley, B.F., 1991. Early Proterozoic collision tectonics and rapakivi granites as intrusions in an extensional thrust-thickened crust: the Ketilidian orogen, South Greenland. Tectonophysics 195, 1–10. Windley, B.F., 1993. Proterozoic anorogenic magmatism and its orogenic connections. J. Geol. Soc. Lond. 150, 39–50. Wiszniewska, J., Claesson, S., Stein, H., Vander Auwera, J., Duchene, J.C., 2002. The north-eastern Polish anorthosite massifs: petrological, geochemical and isotopic evidence for a crustal derivation. Terra Nova 14, 451–460. Zhao, G., Cawood, P.A., Wilde, S.A., Sun, M., 2002. Review of global 2.1–1.8 Ga orogens: implications for a pre-Rodinia supercontinent. Earth Sci. Rev. 59, 125–162. Zhao, G., Sun, M., Wilde, S.A., Li, S., 2004. A PaleoMesoproterozoic supercontinent: assembly, growth and breakup. Earth Sci. Rev. 67, 91–123. Zhong, S., Gurnis, M., 1993. Dynamic feedback between a continent-like raft and thermal convection. J. Geophys. Res. B98, 12219–12232. Zhong, S., 2001. Role of ocean-continent contrast and continental keels on plate motion, net rotation of lithosphere and the geoid. J. Geophys. Res. B106, 703–712.