February 2005 LIP of the Month

Corresponds to event #138 in LIP record database.

Late Paleoproterozoic Large Igneous Provinces from central and southeastern Brazil

Alexandre de Oliveira Chaves^* & Jose Marques Correia Neves
Comissao Nacional de Energia Nuclear (CNEN) - Centro de Desenvolvimento da Tecnologia Nuclear (CDTN). Rua Prof. Mario Werneck, s.No., Cidade Universitaria, Pampulha, Belo Horizonte, 30123-970, MG, Brazil

EXTRACTED FROM JOURNAL OF GEODYNAMICS, 2004 (in press, available online).
^*Corresponding author. Tel.: +55 31 34 99 31 40; Fax: +55 31 34 99 33 90.
aoc@cdtn.br (A.O. Chaves), nevesj@cdtn.br (J.M.C. Neves).

Introduction
The Brazilian Shield (Almeida and Hasui, 1984; inset map in Fig. 1) shows cratonic blocks (Amazonian, Sao Francisco and other smaller ones) surrounded by Proterozoic mobile belts. These belts belong to different structural provinces (Tocantins, Mantiqueira, Borborema), all with complex geological evolution. Large volcanic and sedimentary Phanerozoic basins are found in the Brazilian territory (Amazonas, Parnaiba and Parana) as well.

The Tocantins Proterozoic Structural Province and the Sao Francisco Craton (SFC) are situated in Central and Southeastern Brazil. Their lithotectonic domains are presented in Fig. 1. Meso to Neoproterozoic fold and thrust belts on the west side of the SFC show deformed low to medium grade metasediments and orthogneisses. These mobile belts were developed over the SFC with folds and faults striking NNE-SSW. The tectonic vergence is from west to east. The Neoproterozoic gravimetric anomaly of the suture zone is located farther west from these belts (Fig. 1, Strieder and Suita, 1999).

A lot of radiometric data between 1.8 and 1.7 Ga have been published for acid and basic igneous/metaigneous rocks of Central and Southeastern Brazil, which record widespread bimodal anorogenic magmatism at that time (e.g. Brito Neves et al., 1979; Pimentel et al., 1991; Dussin, 1994; Silva et al., 1995). Extensive intracontinental rift sedimentary sequences belonging to the basal formations of the Araí Group and southern Espinhaco Supergroup, which are located on the west and east borders of the SFC, respectively, are coeval with this magmatism. Could these magmatic rocks be Large Igneous Provinces (LIP) remnants?

The 1.8–1.7 Ga magmatic, tectonic and sedimentological events of Central and Southeastern Brazil
A giant tholeiitic dyke swarm, designated the Para de Minas swarm (1714+/-5 Ma, U-Pb by Silva et al., 1995), is found cutting migmatitic gneiss, granitoids and supracrustal rocks in the southernmost part of the SFC. It is also seen extending under the Neoproterozoic cover and Meso to Neoproterozoic thrust-belts west of this Craton in the aeromagnetic chart of Fig. 2. Even though this swarm extends along 800 km in this chart, it reaches ca. 1400 km up to its NW end in the Transbrasiliano Lineament at the Mato Grosso State (Borges and Drews, 2001).

Most of anisotropy of magnetic susceptibility (AMS) data from 1.72 Ga Para de Minas dykes reveals subhorizontal magma flow, since ca. 70% of the Kmax inclination values oscillate between 0 and 25o for normal and intermediate AMS fabrics (Raposo et al., 2004). Flow features found in dyke/wall rock contact match with the AMS results and show magma flow from NW to SE (Chaves, 2001). These facts seem to indicate a magma source centered at the NW end of the Para de Minas swarm. Furthermore, the central region of porphyritic dykes shows tabular plagioclase phenocrysts with their principal axes oriented having the same trend as those of dykes. This feature is in agreement with the subhorizontal magma flow.

In the neighborhood of the Brazilian capital, dioritic/gabbronoritic rocks of the Barro Alto mafic–ultramafic complex (Fig. 1) were dated at 1721+/-29 Ma using the U-Pb method (Suita et al., 1994). Although this complex has been affected by the Neoproterozoic Brasiliano Orogeny (770–795 Ma), the age based on Barro Alto zircon studies seems to be reliable (Takehara et al., 1999).

Bimodal volcanic rocks near Brasília City (Fig. 1) are associated with the rift-related sedimentation of the Araí Group. This group is a 1500 m thick sedimentary sequence made up of continental to shallow marine sediments (Alvarenga et al., 2000). The volcanic rocks are represented by 1771+/-2Ma alkali-rich basalts and rhyolites near the base of this sedimentary pile. Furthermore, 1.77 Ga Soledade and Sucuri anorogenic undeformed granites are known in the surroundings of Brasilia City as well and are the plutonic equivalents of the Arai volcanics (Pimentel et al., 1991, U-Pb). Recent Nd isotopic data obtained by Pimentel and Botelho (2001) for these granites show a very large spread in _Nd(T) values. They interpreted this fact in two different ways: (i) the original magmas were the product of mixing between 1.8–1.7 Ga mafic magmas and felsic crust-derived melts, or (ii) the sialic source was very heterogeneous in terms of age and isotopic characteristics. They also considered a combination of these two explanations.

An expressive gravimetric low trending NW-SE in the Central-South SFC (Pirapora Low, Fig. 2), is observed on the Bouguer map from Alkmim and Martins-Neto (2001). It corresponds to a surface area of ca. 120,000 km2. Souza Filho (1995) has interpreted this feature as a giant graben that keeps at least 5000 m of sediments. The oldest sediments were generated during the southern Espinhaco Supergroup deposition, around 1.8–1.7 Ga, and they are coeval with the Arai Group sedimentation. Seismic profiles along this trough suggest the occurrence of Arai and southern Espinhaco sediments in its deep sections (Teixeira, 1993). The sedimentary pile of the Pirapora Graben also includes Neoproterozoic sedimentary rocks of the Sao Francisco Supergroup and Meso/Cenozoic sedimentary rocks (Brito Neves et al., 1995). Also with this NW-SE trend, a prominent normal fault near the axis of this graben is suggested in the aeromagnetic chart of Lasa (2001, area 1). This fault is probably a rift master fault and can be recognized by disrupted isogams on the aeromagnetic chart in Fig. 2.

The basal sequence of the Paleo/Mesoproterozoic southern Espinhaco Supergroup situated in the east side of the SFC (Fig. 1), whose thickest sediment pile is located in Pirapora Gravimetric Low (Dupont et al., 2004), is equivalent to that of the Arai Group. The paleocurrent studies of the Arai Group show sediments coming from north and northwest (Martins, 1999). The paleocurrents of the southern Espinhaco basal sequence, intruded by the Para de Minas mafic dykes, come from WNW to ESE (Martins-Neto, 1998). Thus, Arai and Espinhaco paleocurrents come from the northwest, approximately (Fig. 1). This fact suggests a maximum domal uplift in the area of Brasília City preceding the emplacement of the Para de Minas dykes. The present N-S Espinhaco trend seems to result from the Neoproterozoic Brasiliano Orogeny, whose tectonic vergence against the SFC was from the east.

Metarhyolites and Fe-rich metavolcanics dated at 1.77 by Brito Neves et al. (1979, U-Pb) and 1.71–1.72 Ga by Machado et al. (1989, U-Pb) and Dussin (1994, Pb-Pb) are found interbedded in the southern Espinhaco basal sequence. They are associated with greenschists, chemically classified as continental rift tholeiitic basalts (Chula, 1996). Furthermore, Almeida Abreu and Renger (2001) referred to maar craters containing diamond-bearing vent breccias intruding the basal sequence of the southern Espinhaco.

The 1729+/-14 Ma (Dossin et al., 1993; Pb-Pb) and 1770+/-30 Ma (Fernandes, 2001, U-Pb) anorogenic granites of the Borrachudos Suite are found near the southern Espinhaco rocks as well. They can be considered the plutonic equivalent of the rhyolites (Fig. 1). In the central SFC region, exactly in the Lagoa Real Uranium Province region, another anorogenic alkaline granite dated at 1725 +/-5Ma (Turpin et al., 1988, U-Pb) was found, the Sao Timoteo Granite.

Models for the origin of the 1.8–1.7 Ga geological activity in Central and Southeastern Brazil
Several geoscientists have suggested models to explain the origin of the 1.8–1.7 Ga tensional tectonics and the associated igneous activity in Central and Southeastern Brazilian Provinces.

Dussin (1994), Pimentel and Botelho (2001) and Fernandes (2001) proposed that the 1.8–1.7 Ga anorogenic volcanism and plutonism in Central and Southeastern Brazil, coeval with the Arai and Espinhaco rifting, represent a post-orogenic event following the Paleoproterozoic orogeny. They consider a model in which, after the Transamazonian Orogeny (2.1–1.9 Ga), a considerable time interval elapsed before the continental crust (amalgamated blocks) became gravitationally unstable and the crustal extension leading to upwelling/diapirism and decompression melting of the mantle took place. Mantle mafic magmas invaded the crust. Their heat triggered the crust melting, yielding felsic magmas. This was the model used by the last authors to explain the bimodal magmatism and rifting.

Suita and Chemale (1997, 1998) believe that the bimodal magmatism observed in Tocantins and Sao Francisco Provinces during the Paleo/Mesoproterozoic transition could be assigned to the presence of picritic-tholeiitic magmatism affecting a hot Transamazonian lithosphere. The mafic magmatism was possibly caused by a giant plume event at that time. This plume-related event included underplating, thus enabling the heat conduction and the partial melting of the lower crust. According to these authors, this process caused widespread felsic and mafic magmatism in the aforementioned provinces. During the rift phase, sites for the deposition of thick and extensive intracontinental volcano-sedimentary sequences developed.

These two models agree in that the felsic magmatism was generated from the crust melting by heat transfer from deep mafic magmas. However, they are different in terms of the geodynamic mechanism that triggered the tensional tectonics and the huge magmatic activity. The first one is probably linked to the plate tectonics and the last one to a mantle plume.

Discussion
Earth history is punctuated by events during which large volumes of mafic magmas were generated and emplaced by processes unrelated to “normal” sea-floor spreading and subduction. These Large Igneous Provinces (LIPs) are best preserved in the Mesozoic and Cenozoic where they occur as continental flood basalts, volcanic rifted margins, oceanic plateaus, ocean basin flood basalts, submarine ridges, and seamount chains. Felsic rocks may also be represented. Many LIPs can be linked to regional-scale uplift, continental rifting and breakup, and climatic shifts. In the Paleozoic and Proterozoic, LIPs are typically deeply eroded. They are represented by deep-level plumbing systems consisting of giant dyke swarms, sill provinces and layered intrusions.

Based on geological and geochronological data, the 1.8–1.7 Ga events of Central and Southeastern Brazil presented in the previous item can be grouped in two well-defined periods. The first one occurred at 1.77 Ga and the last one around 1.72 Ga. Both are summarized in Table 1 and can be considered as LIP events. 

Table 1
Summary of the 1.77 and 1.72 Ga geological LIP events in Central and Southeastern Brazil and related strutuctural/sedimentological features

1.77 Ga event
Domal uplift of the lithosphere in the vicinity of Brasília City deduced from paleocurrent data.
Bimodal magmatism represented by basalts and anorogenic rhyolites (1771+/-2 Ma, 1770 Ma) and granites (1769+/-2 Ma, 1767+/-10 Ma, 1770+/-30 Ma) caused by deep hot mafic magmas/crust interaction.
Sedimentation of the southern Espinhaco Supergroup basal sequence accompanied by the formation of maar craters containing diamond-bearing vent breccias and deposition of the Arai Group within the Pirapora Graben.
Kimberlite intrusions?
Continental flood basalts? (Pirapora magnetic anomaly)

1.72 Ga event
Emplacement of the Para de Minas dyke swarm (1714+/-5 Ma), with the magma flowing from NW to SE.
Rift-related normal faulting with NW-SE strike.
Main phase of the NW-SE Pirapora Graben development and linked sedimentation.
Crustal melting yielding anorogenic rhyolites (1711+/-8 Ma, 1715+/-2 Ma) and granites (1729+/-14 Ma, 1725+/-5 Ma).
Emplacement of a mafic–ultramafic complex (1721+/-29 Ma).

The link between mantle plume activity and LIPs has been widely recognized. (e.g. Condie, 2001; Ernst and Buchan, 2003). The 1.77 Ga domal uplift in Central Brazil, deduced from the paleocurrent data of the Arai and Espinhaco basal sequences, represents a typical feature of mantle plume activity. The 1.77 Ga basalts and rhyolites interbedded in these sedimentary piles as well as the 1.77 Ga anorogenic undeformed granites found in Central and Southeastern Brazil are probably related to mantle plume mafic magmas that invaded and extensively melted the crust. During the plume-related initial rift phase, sites for the deposition of the Arai and Espinhaco sedimentary basal sequence were developed along the NW-SE Pirapora Graben.

Many examples worldwide show a very close relationship of carbonatitic and kimberlitic rocks to mantle plume magmatism (Bell, 2001; Heaman and Kjarsgaard, 2000; Haggerty, 1999). The diamonds found in metaconglomerates of the southern Espinhaco basal pile are probably related to kimberlitic/lamproitic volcanism affecting this basin. This basal pile also hosts maar craters containing diamond-bearing vent breccias generated generally by phreatomagmatic activity, which increased its explosive action (Nixon, 1995). Thus, although Paleoproterozoic primary diamondiferous rocks in southern Espinhaco have not been located precisely, it is probable that diamonds were brought to the sedimentary basin surface by kimberlitic (or lamproitic) magmatism related to the 1.77 Ga mantle plume.

Although there is no confirmation from drill core, the enigmatic large Pirapora magnetic anomaly clearly shown inside the Pirapora Gravimetric Low (Fig. 2) can be interpreted as plume-related continental flood basalts interbedded in the basal sediments of the NW-SE Pirapora Graben.

The aforementioned features point to 1.77 Ga mantle plume activity in Central and Southeastern Brazil. Since the mantle plume/lithosphere interaction is usually shorter than 20 Myr (Ernst and Buchan, 2002), another plume should be responsible for the activity at around 1.72 Ga. This fact would mean the difference between our two-plume model and the single-plume model proposed by Suita and Chemale (1997, 1998).

When a mantle plume reaches the base of the lithosphere with its head and conduit (tail) structure (e.g. Richards et al. 1989), its head flattens against the lithosphere. Far away from the plume center, the mafic magma spreads laterally in the lithosphere as dykes (e.g. Ernst et al. 2001). The anisotropy of magnetic susceptibility data and flow features in dyke/wall rock contact regarding the giant Para de Minas swarm reveal the lateral spreading of the mafic magma 1400 km away from the magmatic source. This can be explained by the activity of a 1.72 Ga mantle plume in the Central and Southeastern Brazil. The slightly high initial 87 Sr/86 Sr (0.70562) found by Chaves and Correia Neves (2004) for Para de Minas dykes suggests that they were contaminated during their crustal emplacement.

Ernst and Buchan (1997) proposed that mafic–ultramafic complexes can be fed from the system of a plume-related giant radiating dyke swarm and that they may be located either near or far from the plume center. Probably there is a correlation between the Para de Minas swarm and dioritic/gabbronoritic rocks of the Barro Alto mafic–ultramafic complex, taking into account their ages and trace element chemistry (chemical data from Oliveira, 1993 and Chaves, 2001).

Rubin (1992) refers to close relationship between graben subsidence and dyke-induced faulting in Hawaiian volcanic rift zones. Mege et al. (2000) and Ernst et al. (2001) also describe graben development related to dyke emplacement on Mars and Venus. The parallelism between the plume-related Para de Minas dykes and the NW-SE Pirapora Graben (Fig. 2) suggests that the Graben resulted from normal faulting induced by Para de Minas dykes (assuming that the interpretations from Hawaiian and extra-terrestrial dykes can be applied to our intracontinental dyke swarm). The prominent normal fault shown in Fig. 2 near the axis of the Pirapora Graben is probably a rift master fault developed at this time. The real density of the Para de Minas dykes is not shown in Fig. 2 due to the aeromagnetic resolution used. On average, they are 5 km apart from each other in southern SFC and keep the same pattern under the cratonic cover in the Pirapora Graben (Chaves, 2001).

The Pirapora Graben subsidence certainly was accompanied by an increasing sedimentation rate both in Arai and southern Espinhaco. In the deep crustal sections of this Graben the heat transfer from the plume-related rising mafic magmas and probably the regional decompression (deflation) under the Graben yielded extensive crustal melting and generated the acid anorogenic volcanism and plutonism around 1.72 Ga, which were sometimes accompanied by plume-related basic volcanics.

Conclusions
The 1.77 and 1.72 Ga plume events in Central and Southeastern Brazil are certainly responsible for LIPs which can be identified by extensive anorogenic bimodal magmatism and the intrusion of a giant mafic dyke swarm. These events can also be related to the emplacement of dioritic/gabbronoritic rocks of the Barro Alto mafic–ultramafic complex and diamond-bearing rocks. The possibility of the occurrence of Late Paleoproterozoic continental flood basalts in the studied area can be presumed taking into account the large magnetic anomaly inside the coeval Pirapora Graben. The rift sedimentary sequences of the southern Espinhaco Supergroup and Arai Group accumulated in this graben during the Late Paleoproterozoic/Mesoproterozoic transition. Therefore, the widespread magmatism, rifting and sedimentation yielded by these two endogenic events insert them among the most prominent geodynamic processes of the Brazilian Precambrian Geology.

References