Geophys. J. Int. (2019) 219, 633–644 doi: 10.

1093/gji/ggz323
Advance Access publication 2019 July 19
GJI Geodynamics and Tectonics

Delimiting the Neoproterozoic São Francisco Paleocontinental Block

with P-wave traveltime tomography

Marcelo Peres Rocha ,1 Paulo Araújo de Azevedo,2 Marcelo Assumpção ,3 Antônio

Carlos Pedrosa-Soares,4 Reinhardt Fuck1 and Monica Giannoccaro Von Huelsen1
1 SeismologicalObservatory, Institute of Geosciences, University of Brası́lia, Brası́lia, DF, Brazil. E-mail: peres [email protected]
2 Universidade Federal do Oeste do Pará, Instituto de Engenharia e Geociências, Santarém, Pará, Brazil

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

3 Department of Geophysics, IAG, University of São Paulo, São Paulo, SP, Brazil
4 Universidade Federal de Minas Gerais, CPMTC-IGC, Campus Pampulha, Belo Horizonte, 31270–901 MG, Brazil

Accepted 2019 July 16. Received 2019 July 11; in original form 2017 October 1

SUMMARY
The São Francisco Paleocontinental Block (SFPB) represents part of the Congo-São Fran-
cisco Paleocontinent (CSFP), amalgamated around 2 Ga. In the Neoproterozoic, a branched
continental rift system evolved to ocean basins around most edges of the SFPB that remained
only partially linked to the Congo Paleocontinent by means of the Bahia-Gabon Continental
Bridge. After the Brasiliano—Pan-African orogeny, two relatively preserved CSFP sectors
formed the São Francisco and Congo cratons, surrounded by Neoproterozoic orogenic belts.
Recent results of upper mantle P-wave seismic tomography allowed us to suggest a delimi-
tation in lithospheric depths of the Neoproterozoic SFPB, which comprise the São Francisco
Craton, and that this would have been connected with the Congo Paleocontinent along the
Araçuaı́ Belt. It is characterized by high-velocity anomalies and its boundaries with other
blocks are marked by low-velocity anomalies at lithospheric depths. We tested the resolution
of the tomographic results through synthetic models obtained by a ray tracing scheme using the
observed ray configuration. We observe that the lateral resolution is adequate, but the method
used was not able to set the depth reached by the SFPB. Our results indicate that the SFPB area
in lithospheric depths is larger than the surface area ascribed to the São Francisco craton, and
thus, the SFPB basement deeply extends beneath neighboring orogenic regions, suggesting
that these Neoproterozoic mobile belts, such as Araçuaı́ Orogen and the Brasilia Fold Belt,
reworked the continental crust. We observe a low-velocity anomaly in the SFPB central region,
corresponding to the Pirapora aulacogen. Our results have a good spatial correspondence with
the low Bouguer anomalies used to define the SFPB in previous studies. The limits of the
SFPB are consistent with deviation of the mantle flow, as suggested by SKS fast polarization.
Key words: Seismic tomography.

to other cratons and tectonic history. Although it is almost com-

1 I N T RO D U C T I O N
pletely surrounded by Neoproterozoic orogens, it also records a
Cratons are considered as the most tectonic stable portions of con- long-lived connection with the African counterpart, the Congo Cra-
tinental plates, while the related mobile belts represent orogenic ton, prior to the Atlantic opening in the Cretaceous. This connec-
scars formed from plate convergence to collisional processes that tion is represented in the literature by the Bahia-Gabon Continental
resulted in the assembly of the paleocontinents. The Brasiliano–Pan Bridge (BCB, Fig. 1), formed around 2 Ga and preserved from
African orogenic belts represent a system of diachronic collisions the Brasiliano–Pan African orogeny (Torquato & Cordani 1981;
that amalgamated the Western Gondwana Paleocontinent at the end Porada 1989; Pedrosa-Soares et al. 2008). For some authors, the
of the Neoproterozoic (Fig. 1), including the region encompassed Congo-São Francisco Paleocontinent (CSFP) would be a dismem-
by the South American Platform (Brito-Neves et al. 1999; Alkmim bered fragment of the supercontinent Rodinia (Brito-Neves et al.
et al. 2001). 1999). Other authors suggest that the CSFP would be an isolated
Among the main cratons of the South America Platform, the continental piece, located far from Rodinia (Cordani et al. 2003a;
São Francisco Craton (SFC) stands out for its position in relation Kröner & Cordani 2003; Pisarevsky et al. 2003; Cordani et al.


C The Author(s) 2019. Published by Oxford University Press on behalf of The Royal Astronomical Society. 633
634 M.P. Rocha et al.

Platform, and to refine palaeogeographic reconstruction models.

In this way, we present results of traveltime seismic tomography
using P waves in the region of the SFPB to define its limits at
lithospheric depths and to assist the discussion about the tectonic
evolution of the South American Platform. The study region is in
eastern Brazil, covering the SFC and neighbouring orogenic belts,
as well as Phanerozoic basins, and parts of the Amazonian Craton
and Borborema Province (Fig. 2).
For this work we included new data in the database processed
in previous works (Rocha et al. 2011, 2016; Azevedo et al. 2015),
with a significant increase in the number of rays in the final database
reaching that difference to be greater than 44 per cent in relation to
the most recent work (Rocha et al. 2016).

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

2 D ATA A N D M E T H O D
For this study we used data acquired through a permanent net-
work recently installed in Brazil—Brazilian Seismographic Net-
work (RSBR, in Portuguese, Bianchi et al. 2018), with more than
80 stations (blue triangles in Fig. 2), providing data between 2011
Figure 1. Simplified sketch of the Neoproterozoic-Cambrian Brasiliano— and 2017, and through two temporary networks inside the study
Pan African orogenic system and related cratons (modified from Alkmim area: One with 15 stations (yellow triangles in Fig. 2) operated
et al. 2001). SFC, São Francisco Craton; BCB, Bahia-Gabon Cratonic Bridge
during about 4 yr, with events recorded between 2007 and 2013
(cf. Pedrosa-Soares et al. 2008); PC-RPC, Paranapanema and Rio de La Plata
(Azevedo et al. 2015), and another with 34 stations (red triangles in
cratons.
Fig. 2) operating between 2016 and 2017. We used stations from the
2003b; D’Agrella-Filho et al. 2004). However, regarding its recent permanent network of the Seismological Observatory of University
geological history, it can be stated that, with the rifting stages of the of Brası́lia (green triangles in Fig. 2), and some stations from global
Atlantic Ocean, the CSFP broke up into two parts. One of them is networks (black squares in Fig. 2). The data processed in this work
the SFC, of the South American Plate, while the other is the Congo were included in a database previously assembled by prior works
Craton, located in the African Plate. (VanDecar et al. 1995; Schimmel et al. 2003; Rocha et al. 2011,
The idea of the existence of an ancient São Francisco Paleocon- 2016; Azevedo et al. 2015) using old temporary networks deployed
tinental Block (SFPB), before the CSFP, arose from the work of mainly by the Seismological Center of the University of São Paulo
Almeida (1981). He postulated the existence of a larger and older (orange triangles in Fig. 2). A total of 339 stations were used to
continental block, the Paramirim Craton (PC), which would be the generate the tomographic images (all station shown in Fig. 2). We
ancestor of the SFC. According to Almeida (1981), the PC would used records of P and PKIKP phases for events with magnitudes
have reworked its limits during the Paleoproterozoic and Neopro- larger than 4.6 and 5.4 (mb), respectively. Events for phases P were
terozoic events, and the portion not deformed of this craton became chosen in the distance range of 30–95◦ , and for phases PKIKP in
the SFC in the Late Neoproterozoic. However, cratons started to a range of 150–180◦ (Fig. 3). We choose these ranges to avoid
be understood as differentiated lithospheric sectors (Jordan 1978, phase misidentification due to triplications, caused by the mantle
1981, 1988), characterized by having mantle roots that would reach transition zone or the inner core. The final database for the whole
depths of up to 400 km, indicating long-lasting stability. of Brazil consists of 55 704 time measurements for P and PKIKP
Based on the location of Neoproterozoic ophiolites and high- phases from 4989 events.
pressure metamorphic rocks, as well as on structural and gravi- The method used was the linearized inversion approach of Van-
metric data (Ussami 1993), Alkmim et al. (1993) presented a first Decar et al. (1995), based on the ACH inversion method (Aki-
attempt to outline the limits of the SFPB. Afterwards, Ussami (1999) Christoffersson-Husebye; Evans & Achauer 1993). This method
proposed a structure for the SFPB using thermomechanical criteria has been successfully used in several studies in different regions
from the mapping of lithospheric-scale features. More recently, a around the world (e.g. VanDecar et al. 1995; Wolfe 2002; Sol et al.
compilation of the seismological studies made by Assumpção et al. 2002; Schimmel et al. 2003; Bastow et al. 2005, 2008; Benoit et al.
(2004a), and a tomographic study carried out by Rocha et al. (2011), 2006; Lees 2007; Schmid et al. 2008; West et al. 2009; Rocha
indicated that the western limit of the SFC extends further west at et al. 2011, 2016; Azevedo et al. 2015). In this approach, travel-
lithospheric depths, supporting the hypothesis that the SFC is part of time relative residuals (see Evans & Achauer 1993) are inverted
a larger palaeocontinental block. Pereira & Fuck (2005) traced the simultaneously for 3-D velocity structure, earthquake relocations
limits of a former Paleoproterozoic continental plate, based on the and station terms. Observed times are obtained by phase picking of
integration of geological and geochemical data with regional gravi- the seismograms, and theoretical traveltimes have been calculated
metric and aeromagnetic results (Zang 1996; Steenkamp 1998). from a reference earth model (IASP91–Kennett & Engdahl 1991),
Finally, Assumpção et al. (2017) made a compilation of the avail- which was the same used as starting model for inversion. To ob-
able geophysical results to characterize the lithospheric features of tain more accurate picks, we used the Multi-Channel Phase Cross
the SFC. Correlation (MCPCC) developed by Schimmel et al. (2003). This
The knowledge of the geometry of the SFPB and its relation- method is an extension of the Multi-Channel Cross Correlation ap-
ship with other neighbouring cratonic blocks, is important to help proach (MCCC–VanDecar & Crosson 1990), including a scheme
the understanding of the tectonic evolution of the South American of amplitude unbiased Phase Cross-Correlation (PCC–Schimmel
 Delimiting São Francisco Block with Seismic Tomography 635

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

Figure 2. Location of the study region (dashed square) in the topographic map of South America. Solid black lines are the boundaries of the countries, and
solid black and white lines are the main geological provinces. AC, Amazonian Craton; AmB, Amazonian Basin; PnB, Parnaı́ba Basin; PB, Paraná Basin; BP,
Borborema Province; AB, Araçuaı́ Orogen; SFC, São Francisco Craton; and TP, Tocantins Province (Comprising: AFB—Araguaia Fold Belt, BFB—Brası́lia
Fold Belt and RFB—Ribeira Fold Belt). Triangles are the seismographic stations from different networks and institutions.

1999) rather than the classical cross-correlation. To perform the stations network to minimize the mapping of noise and inconsis-
pick, we used different combinations between the minimum and tencies as unrealistic structures into the central area. Further details
maximum frequencies of 0.01 and 4 Hz, where most of the events on the inversion and data processing procedures can be obtained in
were processed with frequencies between 0.5 and 1.5 Hz Schimmel et al. (2003), Rocha et al. (2011) and VanDecar et al.
The parameter space has been discretized in a dense grid of (1995).
1 167 080 knots: 40 knots in depth (depths between 0 and 1400 km), According to Sol et al. (2002), for the approach used in this
163 knots in latitude (latitudes between -43 and 10) and 179 knots work, the non-linearity can be adressed using travel time partial
in longitude (longitudes between –80 and –28) interpolated with derivatives with respect to the slowness parameter rather than ve-
splines under tension (Cline 1981; Neele et al. 1993). This interpo- locity. We used an iterative 3-D ray tracing, and a robust inversion
lation scheme provides a smooth slowness distribution and there- of the updated system of linearized equations, envolving a down-
fore permits an accurate ray tracing. The spacing between the knots weighting of the equations associated with high residuals, and also
at the central region of the model (0–30◦ S; 34–71◦ W; 0–700 km a conjugate-gradient algorithm (Paige & Saunders 1982) to con-
depth) was 0.25◦ and 25 km on horizontal and vertical directions, verge to the solution. We used for our results the 8th-interation of
respectively. This region is where we expect to obtain the best reso- downweighting of equations, where each one uses 2000 conjugate
lution and comprises our interpreted region. The spacing increases gradient steps. A regularization term (smoothness) was included
from the centre to the border of the grid to 0.5◦ and 50 km in an in the objective function to provide additional constraints on the
intermediate volume, and to 1◦ and 100 km at the edges of the model parameters, thereby reducing the non-uniqueness of the so-
model. The model parameterization extends outside the area of the lution. The correct level of regularization was determined somewhat
636 M.P. Rocha et al.

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

Figure 3. Distribution of the events used in this work around the study area. The distances with respect to the centre (concentric circles) are 30◦ .

subjectively by examining the trade-off between model complexity is a robust feature. With the installation of the new RSBR stations
and data misfit (Fig. S1), where the point of greatest curvature on a along the coast, it was possible to improve the ray coverage and ob-
trade-off curve is the optimal value which allows to improve the fit tain more robust anomalies. Thus, it was possible to observe that the
of the data without increasing the complexity of the model. high-velocity anomaly under SFC also extends eastward underneath
the entire Araçuaı́ Orogen. A low-velocity anomaly can be observed
in the SFC centre. It is apparently robust since it is in a relatively
3 R E S U LT S well-covered area by a local network that has operated for at least
1 yr (Montes Claros stations), and by the RSBR stations. Beneath
3.1 Tomographic results the Parnaiba Basin (PnB in Fig. 1) we can observe a tendency of
high velocities that extend from the coast to the Araguaia Fold Belt
Our results are horizontal (Fig. 4) and vertical (Fig. 5) tomographic (AFB), and probably is related to another cratonic block under the
sections, where the velocity perturbations, relative to the regional PnB. Due to the low station coverage it is not possible to image it
average, are shown in an interval between cold (high velocities) and properly, especially in its limit with the SFPB (at its north-western
hot (low velocities) colours. Areas without data (ray density less limit).
than 20 rays/100 km3 ) are shown in black. The first 50 km in the Fig. 5 shows a vertical profile crossing the southern SFC portion
vertical sections (Fig. 5) were not shown due to the low ray crossing (X–X’; Fig. 5a), and three others, almost in NS direction, crossing
just below the stations (in the crust). the central part of SFPB (Y–Y’; Fig. 5b), the western portion (Z–Z’,
Fig. 4 shows horizontal sections at intervals of about 50 km from Fig. 5c) and its coastal portion (W–W’; Fig. 5d). A figure with the
the surface down to 300 km (the first horizontal section is at 75 km ray density for each profile shown in Fig. 5 can be found in the
since there is no resolution in the crust). A figure with the ray den- Supplementary Material (Fig. S3). In Fig. 5(a) we can observe a
sity (hit count) for each section shown in Fig. 4 can be found in the large high-velocity anomaly that practically extends from one ex-
Supplementary Material (Fig. S2). We can observe a predominance tremity to another of the profile, with exception of the westernmost
of high-velocity anomalies under the SFC and beyond its surface portion where a low-velocity anomaly appears. This low-velocity
boundaries with the orogenic belts (low-velocity anomalies), to the anomaly is spatially correlated with the Brası́lia fold belt (BFB) and
west and to the east. The central part of the figure has low ray cov- the high-velocity anomaly with the limits of the SFPB proposed by
erage mainly for the sections at the shallowest depths (regions in Alkmim et al. (1993). This anomaly reaches depths of about 450 km
black), however a continuity of the high-velocity anomaly can be ob- from west to the central part of the profile, and depths of 900 km in
served, which suggests that it is a single structure. The least sampled the eastern portion of the profile, mainly off-shore. In Figs 5(b)–(d)
part is the SFC northern boundary due to lack of stations, however, we can observe a high velocity anomaly crossing the entire SFC.
we can observe that, next to the coast where the ray crossing is This anomaly extends to the Borborema Province (BP), and sug-
larger, the high-velocity anomaly passes from the SFC surface limit gests that, at lithospheric depths, the boundary between these two
and enters under the Borborema Province. As formerly discussed units would be further north than the surface limits. However, in this
by Rocha et al. (2011, 2016), the high-velocity anomaly under the region this anomaly appears to be sectioned (see Fig. 5), suggesting
western portion of BFB (in central part of the Tocantins province) that it is another structure separate from the SFPB. Unfortunately,
remains, beyond the western SFC surface limit, suggesting that it
 Delimiting São Francisco Block with Seismic Tomography 637

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

Figure 4. Horizontal tomographic images for selected depths in the upper mantle. White squares are the stations and solid black lines the main geological
boundaries, highlighting the São Francisco Craton in the centre of the figures. The white lines are the profiles shown in Fig. 5.

despite being a region with relatively good coverage of stations and interpretation can be explored in the future, but not with the current
ray density (Fig. S2), we do not have resolution to fully interpret this station coverage.
structure due to the low ray crossing in depth (Fig. S3). Similar di- At the beginning of the profile Y–Y’ (Fig. 5b) we can observe a
visions appear in the central region of the SFPB (Figs 5c and d) and, strong low velocity anomaly crossing all the upper mantle. This
if we do not take into account the low resolution for small-scale anomaly was interpreted by VanDecar et al. (1995) as a fossil
structures, we might have to propose internal limits that indicate plume related with the Tristan da Cunha hotspot. The high-velocity
the compartmentalization of the SFPB into smaller blocks. This anomaly reaches depths up to 700 km. In the southern portion this
638 M.P. Rocha et al.

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

Figure 5. Vertical tomographic images of the profiles X-X’, Y-Y’, W-W’ and Z-Z’ shown in Fig. 4. The limits of the SFPB imaged (upper black bar)
and geological units crossed by the profiles (lower black bars) were indicated. TP—Tocantins Province; SFC—São Francisco Craton; AB—Araçuaı́ Belt;
PB—Paraná Basin; BP—Borborema Province. The dots in vertical sections are used to indicate the depths.

high-velocity anomaly extends below the Araçuaı́ Belt, showing that that there is a good lateral recovery of the input feature for both
the basement of this orogenic belt has a ‘cratonic’ seismic response. models. The differences are mainly related to the intensity of the
anomalies, since the same inversion parameters of the real models
were considered (weight of regularization, number of iterations,
3.2 Resolution tests etc.). The anomalies obtained for Model 2 are slightly more intense
than those of Model 1. This can be explained because the thickness
To assess the lateral and vertical resolution of our results, we per- of the simulated structure for Model 2 is 50 km greater than in
formed resolution tests using synthetic models (Figs 6 and 7). We Model 1, causing the rays to travel a greater distance within the high-
inverted travel times calculated using a 3-D ray tracer for geometric velocity anomaly. It is also observed that both models are smoother
models based in the real/interpreted structure, and with the same than the input model, since strong spatial velocity gradients are
inversion parameters. The observed event/station configuration was reduced through the regularization of the inversion (e.g. Schimmel
used. A Gaussian residual time error component, with a standard et al. 2003).
deviation of 25 per cent of the RMS, was added to the travel times Some low-velocity anomalies appear around the synthetic SFPB
to simulate the noise of the data. The procedure of the synthetic test simulated in the input model. These are artefacts of the inversion
is the same used by Rocha et al. (2011), Azevedo et al. (2015) and since there are no low-velocity structures in any of the input models.
Schimmel et al. (2003). Rocha et al. (2011) observed a similar effect in synthetic tests made
We used two different models simulating the SFPB observed to test the resolution of the anomaly attributed to the slab of the
in the real results. The lateral dimensions were the same for both Nazca Plate observed in their results. In that work, the artefacts
models (Fig. 6a), but the depth reached for each model was different, were attributed to low resolution in the lower mantle. However,
where the SFPB extends down to 200 and 250 km depth for Model 1 a better explanation would be that, as the regularization of the
and Model 2, respectively. The difference in the vertical dimensions inversion algorithm seeks to make the overall mean of the model
was designed to estimate the vertical resolution. The maximum close to zero, it tries to compensate the presence of large structures
depths that were chosen to be modelled were based on the synthetic including artefacts with opposite anomalies. Also, the low-velocity
models used by Rocha et al. (2011), which suggest that the depth of anomalies might be due to the random noise added to the synthetic
the root of the SFC is 200 km, but considering a smaller block with input travel-times.
dimensions limited by the eastern border of the SFC. Profile X–X’ Fig. 7 shows the vertical profiles with the input models (Figs 7a
in the Fig. 7 are in same position of the profile X–X’ in Fig. 5. and b) and the results of the inversion of the synthetic travel-times
Fig. 6 shows horizontal cross sections at 150 km depth, resulting (Figs 7c and d). As for the horizontal figures, we can also observe
from the inversion of the synthetic travel times for Model 1and that the anomalies recovered for Model 2 are more intense than for
Model 2 (Figs 6b and c, respectively) and cross sections at the Model 1. For both models the recovered anomalies are smeared in
bottom of the Model 1 (200 km, Fig. 6d) and Model 2 (250 km, depth when compared to the input model, which might be due to
Fig. 6e), as well as the input model (Fig. 6a). We can observe
 Delimiting São Francisco Block with Seismic Tomography 639

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

Figure 6. Geometric synthetic models (horizontal slices) built to analyse the resolution of the tomographic results. (a) Slice at 150 km depth for the input
models. (b) and (d) are, respectively, the reconstructions for depths of 150 and 200 km for Model 1 (maximum depth of 200 km), and (c) and (e) are, respectively,
the reconstructions for depths of 150 and 250 km for Model 2 (maximum depth of 250 km).

a high density of ray-paths (see Fig. S3) coming from a particular anomalies reach greater depths (up to 700 km) than in the western
direction (e.g. events from the Andes). This effect appears mainly portion (up to about 400 km). This is similar to what happens in real
in regions with lower ray coverage, as in the central and eastern data, except for the central portion (Fig. 5a), where the anomaly does
portion of the profile X–X’ (Figs 6c and d and Fig. S3a), where the not exceed 400 km. In fact, as mentioned in previous works (e.g.
640 M.P. Rocha et al.

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

Figure 7. Geometric synthetic models (vertical slices) built to analyse the resolution of the tomographic results. (a) and (b) are the X–X’ profile (same of the
real results) for Model 1 (maximum depth of 200 km) and Model 2 (maximum depth of 250 km), respectively. (c) and (d) are, respectively, the reconstructions
of the profile X–X’ for (a) and (b). As explained in the main text, this test highlights the limited depth resolution of the method used in this study.

Rocha et al. 2011; VanDecar et al. 1995; Schimmel et al. 2003), basement at lithosphere depths, and that the SFPB extends far from
the method used here does not have a good vertical resolution. the northern limit of the SFC. This interpretation agrees with the
Finally, in the central part of the structure there are no low-velocity limits proposed by Pereira & Fuck (2005) based on gravimetric
anomalies, which suggests that the low-velocity anomaly observed anomalies. The northern boundary of the SFC, which borders the
in the central part of the SFPB in the real inversion results is not an PB, is defined by a complex system of orogenic belts comprising
artefact. the Rio Preto Belt, in its western portion, the Riacho do Pontal Belt
in the central portion and the Sergipano Belt in the eastern portion,
closer to the continental margin. This orogenic system, together
4 DISCUSSION with the Pernambuco-Alagoas Domain, forms the Southern Sub-
From the results obtained in this work, the limits of the SFPB in province of the BP (Brito-Neves et al. 2001) and was formed by
the lithospheric mantle were interpreted as the limits of the high- the convergence between the CSFP and the West Africa/São Luis
velocity anomaly under the SFC, which extends beyond the surface Craton during the formation of Gondwana in the Brasiliano-Pan
limits of this craton both to the west and to the east (cf. Alkmim et al. African orogeny (Cruz et al. 2014). According to our interpreted
1993), but also to the north. The high-velocity signature is common limit, this orogenic system could be formed over the northern SFPB
for regions with cratonic characteristics once they are more stable edge, which could represent its basement, however, as mentioned
and older, and thus colder and stiffer. Fig. 8 shows a proposal for the above, there is a tendency of splitting the high-velocity anomaly,
SFPB lithospheric limits based on our results. For this interpretation, which could indicate that there would be a smaller block under the
we consider the limits of the high-velocity anomaly and the regional Sergipana Belt separate from the SFPB. Unfortunately, there is not
geological information. enough resolution to state this, and to be consistent we use the limits
The northern boundary of the SFPB includes a region with a of the high-velocity anomaly for our interpretation.
lack of stations; however, part of this limit could be imaged in the The high-velocities observed under the Araçuaı́ orogen suggest
region closest to the coast. The high-velocity observed in our re- extensions southward of both the São Francisco and Congo palao-
sults exceeds the surface limits of the SFC and practically reaches continental blocks. In fact, the Archean and Paleoproterozoic ter-
the Pernambuco Lineament (orange solid line in Fig. 8a). This ranes of these blocks amalgamated around 2 Ga, as a result of a
lineament represents an important tectonic boundary within the complex collage of magmatic arcs formed in continental margins
Borborema Province (PB), limiting its transversal and southern do- and intra-oceanic settings (Noce et al. 2007). After this Paleopro-
mains (Neves et al. 2008). The high-velocity anomalies under the terozoic event of continental amalgamation, the region remained
Borborema Southern subprovince would imply that it has a cratonic preserved from oceanic opening until ca. 650–600 Ma, when an
 Delimiting São Francisco Block with Seismic Tomography 641

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

Figure 8. Comparison between tomographic (a) and Bouguer (b) anomalies in the SFC region. Solid/dashed blue lines are the SFPB limits proposed in this
work based on the tomographic results. Dashed red line is the SFPB limit based on the Bouguer anomaly. Dashed white lines indicate the Paramirim (PaAl)
and Pirapora (PiAl) Aulacogens. Orange and Purple solid lines indicate the Pernambuco and Patos lineaments, respectively, which separate the main tectonic
domains of the Borborema Province: SD, Southern Subprovince, TD, Transversal Subprovince; and ND, Northern Subprovince. The Goiás Magmatic Arc and
the Goiás Massif are the green and brown solid lines, respectively. SKS fast orientations are resented in white bars (Melo & Assumpção 2018), green bars
(Prado et al. 2012), and brown bars (Bastow et al. 2015). The red arrow is the absolute plate motion for the study area based on the model HS3-NUVEL1A
(Gripp & Gordon 2002; http://tectonics.rice.edu/calculators/hs3.html).

ephemeral and narrow oceanic branch formed within the southern and one would expect a low-velocity signature in the tomographic
segment of the Araçuaı́-West Congo basin system. Even during results. However, this structure is displaced to the south in relation
this stage, the Bahia-Gabon Continental Bridge remained unbro- to the observed low-velocity anomaly (Fig. 8a). In this case, if both
ken, to the north, linking the São Francisco Paleopeninsula and are related, we could interpret that the Pirapora Aulacogen has a
the Congo Paleocontinent (Fig. 1). The resultant basin system deep expression, however, this region does not have a density of
was similar to a large gulf (an inland-sea basin) only floored by a stations enough to resolve properly a structure with the dimensions
narrow oceanic lithosphere along its southern axial zone (Pedrosa- of this aulacogen.
Soares et al. 2001; Alkmim et al. 2006, 2008, 2017). Therefore, The high-velocity anomalies considered to be related to a cratonic
the Araçuaı́-West Congo Orogen evolved in a highly ensialic sector, signature, especially in relation to the SFC, have a good spatial cor-
with the outstanding and widespread influence of the Archean- respondence with the low Bouguer anomalies, which have been used
Paleoproterozoic lithosphere of the CSFP. This important ensialic for the delimitation of the SFPB (Alkmim et al. 1993; Ussami 1993;
influence is clearly shown by the chemical and isotopic signatures Pereira & Fuck 2005). An exception is in the coastal region where
of ophiolites and orogenic granites found in the Araçuaı́ Orogen Bouguer anomalies appear significantly larger when compared to
(Pedrosa-Soares et al. 1998; Gonçalves et al. 2016; Tedeschi et al. its innermost part, which can be explained by the increase of den-
2016). This scenario can explain the ‘cratonic’ signature shown by sity of the oceanic lithosphere. Some minor differences between the
our studies within the Araçuaı́ Orogen, that is this signature reflects seismic and gravity anomalies can be observed in the southeast and
the ancient lithosphere of the CSFP (today SFPB), preserved in the west portions of the SFPB. In the southeast part this discrepancy
basement of the orogen. This interpretation would have implica- must be related to the presence of the low-velocity anomaly, inter-
tions in the models that explain the opening of the Atlantic Ocean. preted by VanDecar et al. (1995) as the result of the impact of a
One implication could be that the SFPB would be connected to plume on the edge of the SFPB. Although this effect is not evident
the Congo Craton along its length (along the coast following the in the gravimetric maps, we can observe a tendency of segmentation
Araçuaı́ Belt) during the opening of the Atlantic Ocean and not only of the Bouguer anomaly in the same place where the low velocity
by the Bahia-Gabon Bridge, as is generally indicated in the littoral, anomaly is. In the western part, the low velocity anomaly in BFB,
serving as a barrier to this process. as observed by Azevedo et al. (2015), would be segmented into
The low-velocity anomaly at the SFC central region is well- two parallel structures, one related to the Goiás Magmatic Arc and
correlated spatially with the northern limit of the Pirapora Aulaco- the other to the Goiás Massif. The crust under these two structures
gen (Alkmim et al. 2017). Low-velocity anomalies normally have is thinner, suggesting that there is also a lithospheric thinning (see
been interpreted as weakness zones or lithospheric thinning (e.g. Rocha et al. 2016). In this case, with the rise of the asthenosphere,
Assumpção et al. 2004b; Azevedo et al. 2015; Rocha et al. 2016), the resulting effect would be to increase the Bouguer anomaly and
and since the Pirapora Aulacogen is an aborted rift channel (Alkmim decrease the velocity anomaly. This is the pattern observed in the
et al. 2007), it is an area of weakness due to lithospheric thinning, region of the Goiás Magmatic Arc, but not in the region of the
642 M.P. Rocha et al.

Goiás Massif. The Bouguer anomalies in the Goiás Massif are basin only floored by oceanic crust in a relatively narrow and short
smaller compared to the Goiás Magmatic Arc, suggesting that this zone (see references in previous sections). This result suggests that
would have a similar gravimetric behaviour to that of the SFC, but the SFPB would be connected to the Congo Craton along the coast
due to its smaller lithospheric thickness, the Bouguer anomaly is following the Araçuaı́ Belt during the opening of the Atlantic Ocean
slightly larger than the rest of the SFC. With respect to tomographic and not only by the Bahia-Gabon Bridge, as is generally indicated
anomalies, the low-velocity anomalies would be in agreement with in the Paleogeographic reconstructions of the Gondwana, serving
the region of lithospheric thinning. This interpretation supports the as a barrier to this process.
possibility that this massif could be part of the SFC, as suggested The southern limit of the SFPB defined with seismic tomography
by Alkmim et al. (2001), but we prefer to follow the limits of the has, generally, a good spatial correlation with the limits of the
tomographic anomaly to define the SFPB limits, given that the study Bouguer anomaly, which has been used to define its limits. However,
presented here has no resolution at shallow depths. one of the exceptions is an intense low-velocity anomaly at the limits
The vertical similarity between the results of real data and Model between SFC and BFB.
2 suggest that the roots of SFPB reach depths of 250 km, mainly The western limit of the SFPB defined with seismic tomogra-

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

at its western portion, being greater than that suggested by Rocha phy does not have a good spatial correlation with the limits of the
et al. (2011), whose results of the resolution tests indicated a depth Bouguer anomaly. Probably the gravimetric measurements are more
of 200 km. Unfortunately, it is not possible to estimate the depth in influenced by the presence of the Goiás Massif at the crust while
its eastern portion due to the low density of rays coming from the the seismic tomography image the asthenosphere below it.
east, that reduce the ray crossing in this region, and consequently The Pirapora aulacogen is spatially close to a low-velocity
the resolution. anomaly isolated in the centre of the SFC, indicating that it can
Results of SKS fast orientations measured for stations of the have a deep expression, and can be a weakness zone or a thin litho-
central and southern Brazil (Melo & Assumpção 2018), for stations spheric region.
in BFB (Prado et al. 2012), and for stations at the BP (Bastow et al. Comparison between real data and synthetic results, and the sim-
2015) were compared with the tomographic results presented here ilarities observed in vertical profiles, suggest that the São Francisco
(Fig. 8). The fast orientations can be interpreted as being due to a cratonic root reaches 250 km in depth.
diversion of the mantle flow around the cratonic lithospheric keels The SKS fast anisotropy orientations show an alignment along
(McKenzie 1979; Heintz et al. 2003; Zhao 2005; Bokelmann & the low-velocity channel (thin lithosphere) at the edges of the SFPB,
Wüstefeld 2009; Assumpção et al. 2011; Tian & Santosh 2015). and can be interpreted as mantle flow deviating around the thicker
In Fig. 8 we can observe that the orientation of the fast component lithosphere of the SFPB.
of the SKS wave is generally parallel to the edges of the SFPB,
especially at its boundaries with the Paranapanema Block and the
Amazon Craton (along BFB) to the west, and with the BP to the AC K N OW L E D G E M E N T S
north. The mantle material seems to flow along the low-velocity
channel at the edges of the SFPB, due to the absolute motion of the Authors thank Coordenação de Aperfeiçoamento de Pessoal de
South America Plate (red arrow in Fig. 8) relative to the hotspots Nı́vel Superior (CAPES) and the Instituto Nacional de Ciência e
(HS-NUVEL1A Model, Gripp & Gordon 2002) towards the west, Tecnologia—Estudos Tectônicos (INCT-ET) for the MSc and PhD
along areas of shallower asthenosphere, that is, along the region of scholarships granted to P. A. Azevedo. To CNPq for Research Pro-
the fold belts. ductivity Grant provided to M. P. Rocha, A.C. Pedrosa-Soares,
R. Fuck and M. Assumpção. To Petrobras, for providing fund-
ing through the projects Brazilian Seismographic Network (RSBR)
and Lineamento Tranbrasiliano. We used data from stations of the
5 C O N C LU S I O N S Seismological Center of University of São Paulo, Institute of Tech-
The SFPB is larger than the homonymous craton (SFC) at litho- nological Research of São Paulo (IPT), Seismological Observa-
spheric depths, with its boundaries extending beyond the SFC sur- tory and Laboratory of Lithospheric Studies of the University of
face limits, both to the east and west, and in part to the northern Brasilia, GTSN and GEOSCOPE Networks and ETH-Zürich. John
limit. These limits are in agreement with the Bouguer anomalies VanDecar kindly provided the Regional Seismic Tomography code.
with exception in the coastal region due to the influence of the Tomographic models and interpreted limits of the PSF are avail-
oceanic lithosphere in the gravimetric anomalies, and also with a able at https://doi.org/10.5281/zenodo.3252163. We would like to
small difference in its southern and western portions. thank Martin Schimmel, Carlos Alberto Moreno Chaves, Giuliano
The high velocity observed below the Southern Subprovince of Sant’Anna Marotta and Ian Bastow for the discussions about the
the Borborema Province could be an extension of the SFPB under trade-off curves and the criteria for choosing the regularization pa-
that province, suggesting that the lithospheric basement of the SD is rameters. We also thank Professor Fernando Flecha de Alkmim for
part of the SFPB, and that the lithospheric limit between SFC and BP giving his time to discuss the results of this work. We thank all staff
is more to the north than the surface limits. However, a tendency of and students that helped in the station installation. We thank the re-
division can be observed in the high-velocity anomaly in that region viewers by the valuable suggestions and corrections, and especially
suggesting the existence of another small block separated from the the Editor Dra. Ana Ferreira for her important comments that have
SFPB. It is not possible to confirm any of these interpretations due improved this paper.
to the low resolution in the region.
The Araçuaı́ orogen shows a high P-wave velocity that can be
related to extensions of both the basements of the São Francisco REFERENCES
and Congo paleocontinental blocks. Its craton-like seismic signature Alkmim, F.F., Brito-Neves, B.B.de & Alves, J.A.A., 1993. Arcabouço
could reflect the confined nature of the Araçuaı́-West Congo orogen, tectônico do cráton São Francisco – Uma Revisão, in Simpósio Sobre
a singular case of orogen formed from the closing of an inland–sea o Cráton Do São Francisco. pp. 45–62, SGB/SGM/CNPq.
 Delimiting São Francisco Block with Seismic Tomography 643

Alkmim, F.F., Kuchenbecker, M., Reis, H.L.S. & Pedrosa-Soares, A.C., Cordani, U.G., Brito-Neves, B.B.de & D’Agrella-Filho, M.S., 2003a. From
2017. The Araçuaı́ Belt, in São Francisco Craton, Eastern Brazil: Tectonic rodinia to Gondwana: a review of the available evidence from South
Genealogy of a Miniature Continent, pp. 255–276, eds Heilbron, M., America, Gondwana Res., 6, 275–283.
Cordani, U.G. & Alkmim, F.F., Springer International Publishing. Cordani, U.G., D’Agrella-Filho, M.S., Brito-Neves, B.B.de & Trindade,
Alkmim, F.F., Marshak, S. & Fonseca, M.A., 2001. Assembling West Gond- R.I.F., 2003b. Tearing up Rodinia: the Neoproterozoic palaeogeography
wana in the Neoproterozoic: clues from the São Francisco craton region, of South American cratonic fragments, Terra Nov., 15, 350–359.
Brazil, Geology, 29, 319. Cruz, R.F. da, Pimentel, M.M., Accioly, A.C.de A. & Rodrigues, J.B.,
Alkmim, F.F., Marshak, S., Pedrosa-Soares, A.C., Peres, G.G., Cruz, S.C.P. 2014. Geological and isotopic characteristics of granites from the Western
& Whittington, A., 2006. Kinematic evolution of the Araçuaı́-West Congo Pernambuco-Alagoas Domain: implications for the crustal evolution of
orogen in Brazil and Africa: nutcracker tectonics during the Neoprotero- the Neoproterozoic Borborema Province, Brazilian J. Geol., 44, 627–652.
zoic assembly of Gondwana, Precambrian Res., 149, 43–64. D’Agrella-Filho, M.S., Pacca, I.I.G., Trindade, R.I.F., Teixeira, W., Raposo,
Alkmim, F.F., Pedrosa-Soares, A.C., Noce, C.M. & Cruz, S.C.P., 2007. Sobre M.I.B. & Onstott, T.C., 2004. Paleomagnetism and 40Ar/39Ar ages of
a evolução tectônica do orógeno araçuaı́-congo ocidental, Geonomos, 15, mafic dikes from Salvador (Brazil): new constraints on the São Francisco
25–43. craton APW path between 1080 and 1010 Ma, Precambrian Res., 132,

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

Almeida, F.F.M.de, 1981. O Cráton do Paramirim e suas relações com o do 55–77.
São Francisco, in Simpósio Do Cráton Do Sçao Francisco e Suas Faixas Evans, J.R. & Achauer, U., 1993. Teleseismic velocity tomography using
Marginais. pp. 1–10, SBG. the ACH method: theory and application to continental-scale studies, in
Assumpção, M., An, M., Bianchi, M., França, G.S.L., Rocha, M., Barbosa, Seismic Tomography: Theory and Practice, pp. 319–360, eds Iyer, H.M.
J.R. & Berrocal, J., 2004a. Seismic studies of the Brası́lia fold belt at & Hirahara, K., Chapman and Hall.
the western border of the São Francisco Craton, Central Brazil, using Gonçalves, L., Alkmim, F.F., Pedrosa-Soares, A.C., Dussin, I.A., Valeriano,
receiver function, surface-wave dispersion and teleseismic tomography, C.de M., Lana, C. & Tedeschi, M., 2016. Granites of the intracontinental
Tectonophysics, 388, 173–185. termination of a magmatic arc: an example from the Ediacaran Araçuaı́
Assumpção, M., Azevedo, P.A., Rocha, M.P. & Bianchi, M.B., 2017. Litho- orogen, southeastern Brazil, Gondwana Res., 36, 439–458.
spheric Features of the São Francisco Craton, in São Francisco Craton, Gripp, A.E. & Gordon, R.G., 2002. Young tracks of hotspots and current
Eastern Brazil: Tectonic Genealogy of a Miniature Continent, pp. 15–25, plate velocities, Geophys. J. Int., 150, 321–361.
eds Heilbron, M., Cordani, U.G. & Alkmim, F.F., Springer International Heintz, M., Vauchez, A., Assumpção, M., Barruol, G. & Egydio-Silva, M.,
Publishing. 2003. Shear wave splitting in SE Brazil: an effect of active or fossil upper
Assumpção, M., Guarido, M., van der Lee, S. & Dourado, J.C., 2011. Upper- mantle flow, or both? Earth planet. Sci. Lett., 211, 79–95.
mantle seismic anisotropy from SKS splitting in the South American sta- Jordan, T.H., 1988. Structure and formation of the continental tectosphere,
ble platform: a test of asthenospheric flow models beneath the lithosphere, J. Petrol., Special Vo, 11–37.
Lithosphere, 3, 173–180. Jordan, T.H., 1981. Continents as a chemical boundary layer, Philos. Trans.
Assumpção, M., Schimmel, M., Escalante, C., Roberto Barbosa, J., Rocha, R. Soc. A Math. Phys. Eng. Sci., 301, 359–373.
M. & Barros, L.V., 2004b. Intraplate seismicity in SE Brazil: stress con- Jordan, T.H., 1978. Composition and development of the continental tecto-
centration in lithospheric thin spots, Geophys. J. Int., 159, 390–399. sphere, Nature, 274, 544–548.
Azevedo, P.A., Rocha, M.P., Soares, J.E.P. & Fuck, R.A., 2015. Thin litho- Kennett, B.L.N. & Engdahl, E.R., 1991. Traveltimes for global earthquake
sphere between the Amazonian and São Francisco cratons, in central location and phase identification, Geophys. J. Int., 105, 429–465.
Brazil, revealed by seismic P-wave tomography, Geophys. J. Int., 201, Kröner, A. & Cordani, U., 2003. African, southern Indian and South Amer-
61–69. ican cratons were not part of the Rodinia supercontinent: evidence from
B Brito-Neves, B., Santos, E. & Van Schmus, W., 2001. Tectonic history field relationships and geochronology, Tectonophysics, 375, 325–352.
of the Borborema Province, northeastern Brazil, in Tectonic Evolution of Lees, J.M., 2007. Seismic tomography of magmatic systems, J. Volc.
South America, pp. 151–182, eds Cordani, U.G., Milani, E.J., Thomaz Geotherm. Res., 167, 37–56.
Filho, A. & Campos, D.A., SBG: 31st International Geological Congress. McKenzie, D., 1979. Finite deformation during fluid flow, Geophys. J. Int.,
Bastow, I.D., Julià, J., do Nascimento, A.F., Fuck, R.A., Buckthorp, T.L. 58, 689–715.
& McClellan, J.J., 2015. Upper mantle anisotropy of the Borborema Melo, B.C. & Assumpção, M., 2018. Mantle anisotropy and asthenospheric
Province, NE Brazil: implications for intra-plate deformation and sub- flow around cratons in southeastern South America, Geophys. J. Int., 215,
cratonic asthenospheric flow, Tectonophysics, 657, 81–93. 494–506.
Bastow, I.D., Nyblade, A.A., Stuart, G.W., Rooney, T.O. & Benoit, M.H., Neele, F., VanDecar, J. & Snieder, R., 1993. The use of P wave amplitude data
2008. Upper mantle seismic structure beneath the Ethiopian hot spot: Rift- in a joint inversion with travel times for upper mantle velocity structure,
ing at the edge of the African low-velocity anomaly, Geochem., Geophys. J. geophys. Res.: Solid Earth, 98, 12033–12054.
Geosyst., 9. Neves, S.P., Bruguier, O., Bosch, D., Silva, J.M.R. & Mariano, G., 2008.
Bastow, I.D., Stuart, G.W., Kendall, J.-M. & Ebinger, C.J., 2005. Upper- U-Pb ages of plutonic and metaplutonic rocks in southern Borborema
mantle seismic structure in a region of incipient continental breakup: Province (NE Brazil): timing of Brasiliano deformation and magmatism,
northern Ethiopian rift, Geophys. J. Int., 162, 479–493. J. South Am. Earth Sci., 25, 285–297.
Benoit, M.H., Nyblade, A.A. & VanDecar, J.C., 2006. Upper mantle P-wave Noce, C.M., Pedrosa-Soares, a.C., Silva, L.C.S. & Alkmim, F.F., 2007.
speed variations beneath Ethiopia and the origin of the Afar hotspot, O embasamento arqueano e paleoproterozóico do Orógeno Araçuaı́,
Geology, 34, 329. Geonomos, 15, 17–23.
Bianchi, M.B. et al., 2018. The Brazilian Seismographic Network (RSBR): Paige, C.C. & Saunders, M.A., 1982. LSQR: an algorithm for sparse linear
improving seismic monitoring in Brazil, Seismol. Res. Lett., 89, 452–457. equations and sparse least squares, ACM Trans. Math. Softw., 8, 43–71.
Bokelmann, G.H.R. & Wüstefeld, A., 2009. Comparing crustal and mantle Pedrosa-Soares, A.C., Alkmim, F.F., Tack, L., Noce, C.M., Babinski, M.,
fabric from the North American craton using magnetics and seismic Silva, L.C. & Martins-Neto, M.A., 2008. Similarities and Differences
anisotropy, Earth planet. Sci. Lett., 277, 355–364. Between The Brazilian and African Counterparts of The Neoproterozoic
Brito-Neves, B.B.de, Campos-Neto, M. da C. & Fuck, R.A., 1999. From Araçuaı́-West Congo Orogen, Geol. Soc. London, Spec. Publ. 294, 153
Rodinia to western Gondwana; an approach to the Brasiliano-Pan African LP – 172.
cycle and orogenic collage, Episodes, 22, 155–166. Pedrosa-Soares, A.C., Noce, C.M., Wiedemann, C.M. & Pinto, C.P., 2001.
Cline, A.K., 1981. FITPACK: A Software Package for Curve and Surface The Araçuaı´-West-Congo Orogen in Brazil: an overview of a confined
Fitting Employing Splines Under Tension. Dep. of Computer Sciences - orogen formed during Gondwanaland assembly, Precambrian Res., 110,
University of Texas, Austin. 307–323.
644 M.P. Rocha et al.

Pedrosa-Soares, A.C., Vidal, P., Leonardos, O.H. & Brito-Neves, B.B.de, Torquato, J.R. & Cordani, U.G., 1981. Brazil-Africa geological links, Earth-
1998. Neoproterozoic oceanic remnants in eastern Brazil: Further evi- Sci. Rev., 17, 155–176.
dence and refutation of an exclusively ensialic evolution for the Araçuai´– Ussami, N., 1999. Estrutura e Limites da Placa Litosférica São Franciscana:
West Congo orogen, Geology, 26, 519–522. Contribuição Gravimétrica. IAG-USP. Tese de Livre Docência, Instituto
Pereira, R.S. & Fuck, R.A., 2005. Archean nucleii and the distribution of de Astronomia, Geofı́sica e Ciencias Atmosfericas, Universidade de São
kimberlite and related rocks in the São Francisco Craton, Brazil, Rev. Paulo.
Bras. Geociências, 35, 93–104. Ussami, N., 1993. Estudos geofı́sicos no Cráton do São Francisco: estágio
Pisarevsky, S.A., Wingate, M.T.D., Powell, C.M.C.A., Johnson, S. & Evans, atual e perspectivas, in Simpósio Sobre o Cráton Do São Francisco, pp.
D.A.D., 2003. Models of Rodinia assembly and fragmentation, in Protero- 35–44, SGB/SGM/CNPq, Salvado.
zoic East Gondwana: Supercontinent Assembly and Breakup, pp. 35–55, VanDecar, J.C. & Crosson, R.S., 1990. Determination of teleseismic rela-
eds Yoshida, M., Windley, B.F. & Dasgupta, S., The Geological Society tive phase arrival times using multi-channel cross-correlation and least
of London. squares, Bull. seism. Soc. Am., 80, 150–169.
Porada, H., 1989. Pan-African rifting and orogenesis in southern to equato- VanDecar, J.C., James, D.E. & Assumpção, M., 1995. Seismic evidence for
rial Africa and eastern Brazil. Precambrian Res., 44, 103–136. a fossil mantle plume beneath South America and implications for plate

Downloaded from https://academic.oup.com/gji/article-abstract/219/1/633/5536249 by UFOPA user on 11 March 2020

Prado, E.M.G., Soares, J.E.P. & Fuck, R.A., 2012. Anisotropia do manto driving forces, Nature, 378, 25–31.
litosferico por re-polarização de ondas SKS: implicações na evolução West, J.D., Fouch, M.J., Roth, J.B. & Elkins-Tanton, L.T., 2009. Vertical
tectônica do Brasil central, in V Simpósio Brasileiro de Geofı́sica, SBGf. mantle flow associated with a lithospheric drip beneath the Great Basin,
Rocha, M.P., Azevedo, P.A.de, Marotta, G.S., Schimmel, M. & Fuck, R., Nat. Geosci., 2, 439–444.
2016. Causes of intraplate seismicity in central Brazil from travel time Wolfe, C.J., 2002. Assessing the depth resolution of tomographic models of
seismic tomography. Tectonophysics, 680, 1–7. upper mantle structure beneath Iceland, Geophys. Res. Lett., 29, 1015.
Rocha, M.P., Schimmel, M. & Assumpção, M., 2011. Upper-mantle seismic Zang, M.W., 1996. Magnetic and gravity study of Eastern Bahia. Intern.
structure beneath SE and Central Brazil from P- and S-wave regional Report, DeBeers Bras.
traveltime tomography, Geophys. J. Int., 184, 268–286. Zhao, L., 2005. Using shear wave splitting measurements to investigate
Schimmel, M., 1999. Phase cross-correlations: design, comparisons, and the upper mantle anisotropy beneath the North China Craton: dis-
applications, Bull. seism. Soc. Am., 89, 1366–1378. tinct variation from east to west, Geophys. Res. Lett., 32, L10309,
Schimmel, M., Assumpção, M. & VanDecar, J.C., 2003. Seismic velocity doi:10.1029/2005GL022585.
anomalies beneath SE Brazil from P and S wave travel time inversions, J.
geophys. Res.: Solid Earth, 108, 2191.
Schmid, C., van der Lee, S., VanDecar, J.C., Engdahl, E.R. & Giardini, D., S U P P O RT I N G I N F O R M AT I O N
2008. Three-dimensional S velocity of the mantle in the Africa-Eurasia
Supplementary data are available at GJ I online.
plate boundary region from phase arrival times and regional waveforms,
J. geophys. Res., 113, B03306, doi:10.1029/2005JB004193. Figure S1. Trade-off curve between the RMS Travel-time Residual
Sol, S., Thomson, C.J., Kendall, J.-M., White, D., VanDecar, J.C. & Asudeh, Reduction Percentage and the RMS Model Roughness for different
I., 2002. Seismic tomographic images of the cratonic upper mantle be- smoothness parameters used in the regularization of the inversion.
neath the Western Superior Province of the Canadian Shield—a remnant The value used for the final figures is indicated by the red arrow.
Archean slab? Phys. Earth planet. Inter., 134, 53–69. Figure S2. Ray density (hit count) for each horizontal image in
Steenkamp, B., 1998. Some questions and answers about the Mesozoic
Fig. 4 of the main text.
diamond mineralization potential of the Paramirim Shield. Intern. Report,
Figure S3. Ray density (hit count) for each vertical image in Fig. 5
DeBeers Bras.
Tedeschi, M. et al., 2016. The Ediacaran Rio Doce magmatic arc revisited of the main text.
(Araçuaı́-Ribeira orogenic system, SE Brazil), J. South Am. Earth Sci., Please note: Oxford University Press is not responsible for the con-
68, 167–186. tent or functionality of any supporting materials supplied by the
Tian, X. & Santosh, M., 2015. Fossilized lithospheric deformation revealed authors. Any queries (other than missing material) should be di-
by teleseismic shear wave splitting in eastern China, GSA Today, 25, 4–10.
rected to the corresponding author for the paper.