Baltica – Amazonia – West Africa connection
Karlstrom et al. (2001) were the first to propose a link between Baltica, Amazonia and Africa. The SAMBA (South America – Baltica) model suggests that Baltica, Amazonia and West Africa were linked together, and remained consistent from 1800 Ma until 1300 Ma or even 800 Ma (Figure 1.30) (Johansson, 2009, 2014). Johansson (2009)’s model is mainly based on the distribution of orogenic and magmatic belts to provide a coherent evolution and continuity of these belts along the cratonic blocks, and the repartition of AMCG complexes. This model suggests that the Svecofennian orogenic belt (1900 – 1850 Ma) in Baltica is connected to the coeval Ventuari – Tapajós Province in the Amazonian craton. Moreover, the 1850 – 1650 Ma Transscandinavian Igneous Belt (TIB) and the 1640 – 1520 Ma Gothian belt have their continuation in the Rio Negro-Juruena Province (1780 – 1550 Ma) in the Amazonian craton.
According to this model, Baltica is well-linked with Laurentia but its orientation is different from the NENA “upside-down” configuration (Gower et al., 1990). The “right-way-up” orientation of Baltica relative to Laurentia is adopted based on geological grounds, and this supports a tight fit between SE Greenland and NW Fennoscandia as also suggested by Hoffman (1988) and Bridgwater et al. (1990). However, paleomagnetic data is consistent with a different connection between Baltica and Laurentia (Buchan, 2013).
New paleomagnetic and geochronological data for the Amazonian craton supports the SAMBA model with the 1790 Ma Avanavero pole (Bispo-Santos et al., 2014b). However, 1440 – 1420 Ma paleomagnetic poles – Indiavai pole (D’Agrella-Filho et al., 2012), Nova Guarita pole (Bispo-Santos et al., 2012), Salto do Céu sills pole (D’Agrella-Filho et al., 2012) are located at ca. 30° from the 1460 Ma mean pole for NENA (Laurentia and Baltica) and they don’t support the SAMBA model as viewed by (Bispo-Santos et al., 2014b). A possible explanation is that Internal block rotations within the Columbia supercontinent occurred between 1780 and 1400 Ma (See D’Agrella-Filho et al. (2016) for a discussion).
Among the latest models published in the literature, the SAMBA model is widely accepted (Eglington et al., 2013; Evans and Mitchell, 2011; Pehrsson et al., 2016; Salminen et al., 2015; Xu et al., 2014; Zhang et al., 2012).
Current models for Columbia
Current models are based on a larger paleomagnetic database, seeking to obtain a coherent kinematic evolution. Moreover, these data are combined with geochronological and stratigraphic compilations (Eglington et al., 2009; Pisarevsky et al., 2014). Metallogenic associations are also used (Pehrsson et al., 2016).
Pehrsson et al. (2016) propose a model that takes into account the evolution of cratons over time and which considers each province within the cratons (Figure 1.31). It’s not a palinspastic reconstruction because they consider rigid plates. This model is based on the updated paleomagnetic reconstructions of Evans and Mitchell (2011) and Zhang et al. (2012) and propose an evolution until the assembly of Rodinia of Li et al. (2013) and Li et al. (2008). The paleomagnetic “Upside-down” configuration between Baltica and Laurentia is accepted (Buchan, 2013). Siberia is linked to Laurentia in a tight-fit position (Buchan et al., 2016). The East Antarctica – Australia block is linked to Laurentia in a proto-SWEAT (“Southwest U.S – East Antarctic”) configuration (Zhang et al., 2012). Geological data support a tight connection of Australia with Laurentia (Betts et al., 2016; Thorkelson and Laughton, 2015). The SAMBA model is used to form a large landmass with Baltica – Amazonia and West Africa (Johansson, 2009, 2014). North China is near to the São Francisco-Congo craton and India, whereas Kalahari is drifting alone. In their model, Rio de la Plata, Amazonia, West Africa, and Congo/São Francisco formed a large united landmass. They consider India as divided in two parts – South and North India – but such separation is not supported by paleomagnetic data (Radhakrishna et al., 2013a; Radhakrishna et al., 2013b). Most of the Columbia supercontinent amalgamation occurred between 2200 – 1780 Ma, but it was finally assembled at ca. 1550 Ma. Finally, Kalahari and India did not took part of this Supercontinent, according to the model.
Figure 1.31: Positions of cratons at 1.95 – 1.88 Ga (A) and Columbia supercontinent with its maximum packing at 1.60 – 1.40 Ma (B) (Pehrsson et al., 2016).
Also based on a large paleomagnetic database, Pisarevsky et al. (2014) suggested an alternative model for Columbia supercontinent between 1790 and 1270 Ma (Figure 1.32), the main features of which are summarized below. Like Pehrsson et al. (2016)’s model, Baltica and Laurentia are connected in an “upside-down” position as in NENA (Gower et al., 1990). Siberia was positioned close to the equator but ~2000 km separated from Laurentia between 1740 and 1720 Ma. New paleomagnetic data and LIP barcode strongly support the Laurentia and Siberia connection but some uncertainties on the position and orientation of Siberia persist (Buchan et al., 2016; Ernst et al., 2016a; Pisarevsky et al., 2008). Based on LIPs barcode correlation, Siberia is considered close to the Congo – São Francisco with a mantle plume center under the craton. Cederberg et al. (2016) updated this connection with inclusion of North China Craton (NCC).
In the Pisarevsky et al. (2014)’s model, Australia and East Antarctica are located close to north-western Laurentia in a proto-SWEAT configuration as in the Zhang et al. (2012) and Pehrsson et al. (2016)’s models with a certain distance (before final collision at ca. 1580 Ma). India was linked to Baltica during the Mesoproterozoic unlike the SAMBA model proposes. Evans (2013), however, pointed out that this position of India is unlikely because of the great distance it has to move towards Rodinia (Li et al., 2008). Another consequence of their model is the rejection to the SAMBA model. They consider Amazonia and West Africa are not part of the Columbia supercontinent and they drift as a single landmass. In this model we have the formation of proto-cratons between 2000 and 1800 Ma that drifted to form the “West-Nuna” or the “East-Nuna” since 1790 Ma with a final collision of these two large landmasses at ca. 1580 Ma. Break-up occurred between 1450 and 1380 Ma.
A significant amount of evidence corroborates the existence of the Columbia supercontinent, which was probably the first supercontinent in Earth’s history. Many different reconstructions exist for this supercontinent but some specificities are common to all models. The connection between Laurentia and Baltica appears to be strong although their orientations may vary between models. The same can be said for Amazonian craton and West Africa which apper linked together in practically all models. The proto-SWEAT connection is also generally well-accepted between proto-Australia – East Antarctica (“Mawsonland”) and Laurentia. Most proto-cratons amalgamated between 2000 and 1800 Ma and the maximum package seems to have occurred at ca. 1600 Ma. Break-up seems to have initiated at ca.1400 Ma but alternative models admit a later break-up with the formation of Rodinia. The lack of paleomagnetic data in this period is the main reason for these uncertainties.
Previous discussion shows that the geodynamic context (classified as “transitional”) in which this first supercontinent was assembled is different from the classical Phanerozoic plate tectonics. Most models do not take into account the geodynamical aspect. Recently, Meert (2014) observed similarities between the three supercontinents (Columbia, Rodinia, and Pangea). Laurentia, Baltica, and Siberia are always close together in the three supercontinents (in different configurations). With the modern-style plate tectonics, we tend to imagine a random drifting for the cratons through time and the probability to observe the same associations should be low. Meert (2014) called these landmasses as “strange attractors”. In opposition, cratons of “West-Gondwana” (South America, Africa) are referred as ‘spiritual interlopers” (similarities with large displacement). Some cratons are always isolated in different configurations and are referred to as “lonely wanderers”. This very interesting vision could suggest a dominant lid tectonic during the Proterozoic with episodes of true polar wander (TPW) (Meert, 2014).
Despite these new advances many uncertainties still persist. Thus, acquiring new paleomagnetic, geochronological, and structural data, and considering the prevailing geodynamics are essential to improve these models. We have especially seen that position of the Amazonian craton was not yet convincingly set in the different models. In the next section, we will focus on the importance of the Amazonian craton in the Columbia supercontinent.
The paleomagnetic problem
This section is a brief summary of the published paper “Amazonian Craton paleomagnetism and paleocontinents” (D’Agrella-Filho et al., 2016) (see end of section). The evolution of the Amazonian craton has little resemblance to that recorded in other cratonic units of South America. It has more similarities with the evolution of West Africa craton, Laurentia and Baltica (Geraldes et al., 2001; Pesonen et al., 2003). The position of many of the units, especially the Amazonian craton, is yet poorly established due to the low quality of the world paleomagnetic database making reconstruction of the Proterozoic paleogeography highly speculative (Pesonen et al., 2003).
The Amazonian craton
The Amazonian craton is one of the main tectonic units of the South American Platform consisting of the Guiana and Central-Brazil (or Guaporé) shields, separated by the Amazon and Solimões basins. After the initial model of tectonic subdivision of the Amazonian craton, proposed by Amaral (1974), several other models of tectonic evolution have been proposed, which basically oppose two major theoretical schools: fixist against mobilistic school.
The fixist school considers the craton as a large Archean continental shield, affected by several episodes of crustal reworking through thermal events (Costa and Hasui, 1997; Hasui et al., 1984; Hasui, 1985). These authors defined the Amazonian craton as a mosaic due to the juxtaposition of twelve tectonic blocks (paleo-plates), which were assembled as a large landmass through diachronic collisions during the Archean and Paleoproterozoic. According to this model, at the end of Paleoproterozoic and at the beginning of the early Mesoproterozoic the newly assembled craton would be affected only by intraplate tectonic events, most likely extensional events. The fixist model was based primarily on geophysical data (gravimetric and magnetometric), available geological and structural interpretations at the time, and in a few radiometric data, especially those obtained by the K-Ar and Rb-Sr geochronological methods.
The mobilistic school proposes that the evolution of the Amazonian craton is the result of successive episodes of crustal accretion in Paleo and Mesoproterozoic, around an older core, stabilized at the end of the Archean (Cordani and Sato, 1999; Cordani et al., 1979; Cordani and Neves, 1982; Cordani and Teixeira, 2007; Tassinari and Macambira, 1999; Tassinari et al., 2000; Tassinari and Macambira, 2004; Teixeira et al., 1989).
The paleomagnetic problem
Among the models in the most recent literature, the tectonic divisions of Vasquez et al. (2008) or Santos et al. (2000) and Cordani and Teixeira (2007) are the most used (Figure 2.1). The models are similar, with some disagreements, especially regarding the boundaries of tectonic (geochronological) provinces.
The model of Vasquez et al. (2008), is a review of models of Santos et al. (2003a); Santos et al. (2000) and is based on the interpretations of new U-Pb and Sm-Nd data. Santos et al. (2003a); Santos et al. (2000) proposed a division of the craton in seven geochronological or tectonic provinces (Figure 2.1.A): the Transamazonic Province (2250 – 2000 Ma), Carajás Province (2530 – 3100 Ma), Central Amazon Province (1880 – 1700 Ma), Tapajós – Parima Province (2100 – 1870 Ma), Rio Negro Province (1860 – 1520 Ma), Rondônia – Juruena Province (1760 – 1470 Ma), and Sunsás Province (1330 – 990 Ma). Vasquez et al. (2008) updated this model based on the geological map of Pará state (Brazil). They proposed new domains: (1) the division of the Carajás Province in the Carajás and Rio Maria domains, (2) the division of Transamazonas Province in Carecuru, Paru, Amapá, Bacajá and Santana do Araguaia domains, (3) the division of the Central Amazonia Province in the Erepecuru – Trombetas (W and E) and Iriri – Xingu domains.
The model of Cordani and Teixeira (2007) is a review of previous models (Figure 2.1.B) (Cordani et al., 1979; Tassinari and Macambira, 1999; Tassinari and Macambira, 2004; Teixeira et al., 1989) with two Archean nuclei and five Proterozoic tectonic provinces. In this model, the core of the Amazonian craton consists in the Central Amazonian Province which is formed by two Archean nuclei, the Xingu – Iricoumé and Roraima blocks (3200 – 2600 Ma). The Maroni – Itacaiunas Province is constituted by mobile belts of ages between 2250 and 2050 Ma (Ledru et al., 1994) associated to the Siderian – Rhyacian orogenic events (“old-Transamazonian cycle”). The Archean basement of the Amazonian craton is covered by Proterozoic volcano – sedimentary units with little or no deformation. Accretionary belts occurred during the Paleo-Mesoproterozoic along the southwestern margin with the development of the Ventuari – Tapajós Province (2000 – 1800 Ma), the Rio Negro – Juruena Province (1780 – 1550 Ma), and the Rondonian – San Ignacio Province (1500 – 1300 Ma). The latter is characterized by the collision of the Paraguá terrane at ca. 1320 Ma (Bettencourt et al., 2010; Rizzotto and Hartmann, 2012). The final orogenic belt occurring to the west of the Amazonian craton is the Sunsas – Aguapeí (1250 – 1000 Ma) which highlights the Grenvillian collision between Amazonian and Laurentia cratons.
In this work, we will follow the evolutionary model of Cordani and Teixeira (2007) which is adopted by several other authors (Bettencourt et al., 2010; Cordani et al., 2009; Schobbenhaus et al., 2004).
The paleomagnetic problem
Figure 2.1: Tectonic models for the Amazonian craton. A: Model of Santos et al. (2003a) adapted from Roverato et al. (2016). B: Model of Cordani and Teixeira (2007) with localization of different paleomagnetic studies for the Amazonian craton (D’Agrella-Filho et al., 2016; Teixeira et al., 2016). Inset: Sketch of the southwestern part of the Amazonian Craton showing the Paraguá Terrain and Alto Guaporé, Aguapeí and Nova Brasilândia belts (modified after D’Agrella-Filho et al. (2012)).
We saw in the previous chapter that there are many models for the position of the Amazonian craton in the Columbia supercontinent constrained by paleomagnetic data. All Amazonian paleomagnetic poles from Paleo-Mesoproterozoic times are described in Table 1 of the attached paper (D’Agrella-Filho et al., 2016), which synthesizes their tectonic implications for paleocontinents. The progress in the Amazonian paleomagnetism can be attributed mainly to three research groups (Figure 2.1.B): (i) one carried out by the Princeton group (led by Tullis C. Onstot) in the 1980s. This group worked mainly on Paleoproterozoic rocks from Venezuela (green stars in Figure 2.1.B). (ii) A second group, with paleomagnetic work developed on Paleoproterozoic rocks from the French Guiana (blue stars in Figure 2.1.B), whose paleomagnetic results were published in the 2000s. This group was led by French researchers: Sébastian Nomade (at that time in Berkeley Geochronology Center, USA) and Hervé Théveniaut from BRGM (Bureau de Recherches Géologiques et Minières, France). (iii) The third influential group (and currently in activity) in the history of paleomagnetism of the Amazonian craton is from IAG-USP (Brazil). This group has worked with geological units with ages varying since Paleoproterozoic up to Cambrian (yellow stars in Figure 2.1.B). Adding to these three groups, independent studies (purple stars in Figure 2.1.B) have also achieved significant results (Castillo and Costanzo-Alvarez, 1993; Valdespino and Alvarez, 1997; Veldkamp et al., 1971). Below, we describe a synthesis of the main tectonic implications of the Paleoproterozoic paleomagnetic data on the formation of the Columbia supercontinent. The first paleomagnetic data obtained by the Princeton Group led to the proposition of a possible connection between Amazonian and West African cratons along the Guri (in Amazonia) and Sassandra (West Africa) shear zones (Onstott, 1981a). This proposal was later on corroborated by new paleomagnetic data from Paleoproterozoic igneous and metamorphic rocks from French Guiana (Nomade et al., 2003).
In the 2000s, new paleomagnetic expeditions were carried out in the Guiana shield by the BRGM (Bureau de Recherches Géologiques et Minières, France). They produced a large amount of paleomagnetic data and new poles for the Guiana shield (Costal Late granite, Approuague River granite, Mataroni River granite, Tampok granite, Tumuc granite, Armontabo River granite) (Théveniaut et al., 2006). We can highlight the very good paleomagnetic OYA pole determined by Nomade et al. (2001), and dated by Ar-Ar at ca. 2036 ± 14 Ma (cooling age of the tonalite). Théveniaut and Delor (2003) were the first authors to propose a review of the paleomagnetic results and quantify the reliability of data for the Amazonian craton. Based on new paleomagnetic data on well-calibrated in age plutonic and metamorphic rocks, (Théveniaut et al., 2006) proposed the first apparent polar wander path for the Amazonian craton between 2155 and 1970 Ma.
Recently, Bispo-Santos et al. (2014a) published a robust pole for the well-dated 1980 – 1960 Ma (U-Pb on zircons) Surumu volcanics from northern Roraima State (Brazil). This pole helped to improve the Théveniaut et al. (2006)’s APWP. Bispo-Santos et al. (2014a) argue that the present 2100 – 1960 Ma paleomagnetic data from Amazonian and West African cratons support the connection between these cratons along the Guri and Sassandra lineaments as previously proposed by Onstott et al. (1981) and Nomade et al. (2003). These results imply that a large landmass was probably formed at 2.0 Ga by proto-Amazonia, West Africa, and another cratonic block (probably Sarmatia/Volgo-Uralia from Baltica) that collided during the 2200 – 2000 Ma Maroni – Itacaiunas mobile belt (D’Agrella-Filho et al., 2016). According to these authors, this continental block collided with Fennoscandia to form Columbia at about 1.79 – 1.78 Ga ago.
Bispo-Santos et al. (2008) calculated a paleomagnetic pole for the well-dated 1790 Ma Colíder group from the Central – Brazil shield. They proposed a configuration for the Columbia supercontinent with a connection between Baltica, North China craton and Amazonian craton as we saw previously. It is interesting to note that results from the 1440 – 1420 Ma Nova Guarita mafic dike swarm (Bispo-Santos et al., 2012), Indiavaí Suite (D’Agrella-Filho et al., 2012) and Salto do Céu sills (D’Agrella-Filho et al., 2016) support a similar connection between Baltica – NCC – Amazonia. These geological units are also from the Central – Brazil shield located at the southern part of the Amazonian craton.
Recently, however, Bispo-Santos et al. (2014b) performed a paleomagnetic study on the Avanavero sills from the Guiana shield. These rocks are well-dated by U-Pb on baddeleyite at ca. 1790 Ma (Reis et al., 2013). The Avanavero pole passes a baked contact test and supports the SAMBA model (Johansson, 2009) where Baltica was directly linked to Amazonian craton and West Africa. The inconsistence of the 1790 Ma and the 1440 – 1420 Ma poles from the Central – Brazil shield with those from Guiana shield, Baltica and Laurentia (in the Columbia reconstruction of Bispo-Santos et al. (2014b)) may be interpreted as either: (i) dextral strike-slip movements occurred between Central – Brazil shield and Guiana shield, after 1420 Ma (Bispo-Santos et al., 2014b); (ii) counterclockwise rotation of Amazonia/West Africa occurred at some time between 1780 and 1440 Ma inside Columbia (D’Agrella-Filho et al., 2016a, b); or (iii) Amazonia/West Africa did not take part of Columbia (Pisarevsky et al. (2014). The last two alternatives assume that the Colíder pole did not represent a primary magnetization.
Paleomagnetic problem and birth of this study
Despite the accumulation of new paleomagnetic data obtained for the Amazonian craton during the Paleo – Mesoproterozoic, it is yet difficult to define an apparent polar wander path for this craton for more recent periods than 1960 Ma (Surumu pole). Figure 2.2 shows possible scenarios for the Amazonian craton’s APWP according to the use of the Avanavero pole (trajectory 1), Colíder pole (trajectory 2), or yet using their anti-poles which define, respectively, the trajectories 3 and 4. The great age difference between the 1960 Ma Surumu pole and the 1790 Ma Avanavero and Colíder poles shows clearly the need of new poles for this interval. In addition, there is yet a large uncertainty in the position of the Amazonian craton at ca. 1790 Ma. Indeed, as discussed above, there were two different poles of the same age, the Colíder pole (Bispo-Santos et al., 2008) and the Avanavero pole (Bispo-Santos et al., 2014b) that involve distinct configurations within the Columbia supercontinent (Figure 5 of the attached paper). Therefore, we need new key poles for the Amazonian craton, particularly between 1960 (Surumu) and 1790 Ma (Avanavero, Colíder) to constrain the position of the craton and establish its APWP. It may be noted that there is no paleomagnetic pole at 1880 Ma for the North China Craton (NCC) to test a possible connection with the Amazonian craton as suggested by Bispo-Santos et al. (2008).
Table of contents :
Chapter.1: Paleoproterozoic Era and the Columbia supercontinent
1.1 Paleoproterozoic geodynamics
1.1.1 Earth’s Atmosphere, Hydrosphere, and Biosphere
1.1.2 Cooling of the mantle and crustal evolution
1.1.3 Stabilization of cratons
1.2 Definition and evolution of supercontinents
1.3 Evidence for a Paleoproterozoic supercontinent
1.4. Models for the Columbia supercontinent
Chapter 2: Position of the Amazonian craton in Columbia: The paleomagnetic problem
2.1 The Amazonian craton
2.2 Paleomagnetic database for the Amazonian craton – implications to the paleocontinent Columbia
2.3 Paleomagnetic problem and birth of this study
2.4 Paper “Amazonian Craton paleomagnetism and paleocontinents” (co-author)
Chapter. 3: The Carajás Province, Sampling
3.1 Target of the study: The Uatumã LIP, a Paleoproterozoic SLIP
3.2 The Carajás Province
3.3 Sampling and geological setting
3.3.1 Tucumã area
3.3.2 São Felix do Xingu area
Chapter. 4: Methodology
4.1.1 Paleomagnetic sampling
4.1.2 Anisotropy of magnetic susceptibility (AMS)
4.1.3 The remanent magnetization
4.1.4 Demagnetization techniques
188.8.131.52 Alternating Field (AF) demagnetization
184.108.40.206 Thermal demagnetization Summary
220.127.116.11 LTD demagnetization
4.1.5 Magnetic mineralogy
4.1.6 Analysis of components
4.1.7 Field tests and paleomagnetic stability
18.104.22.168 Reversals test
22.214.171.124 Baked contact test
126.96.36.199 Regional consistency
4.1.8 Paleomagnetic pole
4.1.9 Paleogeographic reconstruction in the Precambrian
188.8.131.52 GAD through Precambrian?
184.108.40.206 Paleolatitude reconstruction
220.127.116.11 Comparison between two cratons
18.104.22.168 True Polar Wander (TPW) reconstruction
4.2.1 U-Th-Pb system
4.2.2 SHRIMP analysis
4.2.3 LA-ICPMS analysis
4.3 Geochronological and paleomagnetic systems
Chapter. 5: Petrology and magnetic mineralogy of the Tucumã dike swarms; overview of the dike swarm of the Uatumã event
5.1.1 Field observations
5.1.2 Microgranitic dikes
22.214.171.124 Sequence of crystallization
5.2 Mineral chemistry of microgranites
5.4 Magnetic properties
5.4.1 Magnetic Mineralogy
126.96.36.199 Hysteresis curves
188.8.131.52 Isothermal remanent magnetization (IRM) curves
184.108.40.206 Kruiver’s analysis
220.127.116.11 Day plot
18.104.22.168 Thermomagnetic curves
5.4.2 Summary for the magnetic mineralogy
5.5 Whole rock geochemistry
5.5.1 Major and trace elements geochemistry
5.5.2 Relation between petrology and magnetism
5.6 Paper of da Silva et al. (2016) (co-author)
Chapter. 6: AMS and paleomagnetic data for the Tucumã dike swarms
6.1 Magnetic Mineralogy
6.2 Anisotropy of magnetic susceptibility (AMS)
6.3 Paleomagnetic results
6.3.1 Magnetic components
6.3.2 Mean directions and paleomagnetic poles
6.4 Baked contact tests
6.5 Reliability of Tucumã poles
Chapter. 7: Turmoil before the boring billion: Paleomagnetism of the 1880 – 1860 Ma Uatumã event in the Amazonian craton
7.2 Geological setting and lithology
7.3 Sampling and analytical methods
7.4 U-Pb Geochronology
7.5 Paleomagnetic results
7.6 Baked contact test
7.7 Magnetic mineralogy
7.8 Oxide textural analysis
7.9.1 U-Pb geochronology
7.9.2 Confidence of the paleomagnetic poles
7.9.3 Paleomagnetic discrepancies between 1.9 – 1.8 Ga
7.9.4 True polar wander and paleogeography at 1880 – 1860 Ma
7.9.5 Geological turmoil during the amalgamation of the Supercontinent Columbia