Si

MI

a

ARRAA

KUCMSR

1

eC

R

0d

Precambrian Research 185 (2011) 55–64

Contents lists available at ScienceDirect

Precambrian Research

journa l homepage: www.e lsev ier .com/ locate /precamres

ignatures of continental collisions and magmatic activity in central Brazil asndicated by a magnetotelluric profile across distinct tectonic provinces

aurício S. Bologna1, Antonio L. Padilha ∗, Ícaro Vitorello, Marcelo B. Páduanstituto Nacional de Pesquisas Espaciais – INPE, C.P. 515, 12201-970 São José dos Campos, Brazil

r t i c l e i n f o

rticle history:eceived 22 June 2010eceived in revised form 1 December 2010ccepted 2 December 2010vailable online 13 December 2010

eywords:pper mantleentral Brazilagnetotellurics

eismic low velocity anomalyefertilization

a b s t r a c t

Broadband and long-period magnetotelluric soundings were collected along a 560 km E–W profile in aregion of central Brazil subjected to Neoproterozoic collision tectonics and Archean to Cretaceous mag-matic events. The profile crosses the northeastern portion of Phanerozoic sediments and volcanics of theParaná basin, the southern extension of the Neoproterozoic metasedimentary rocks of the Brasília belt,locally pervaded by Cretaceous alkaline magmas of the Alto Paranaíba igneous province (APIP), and Neo-proterozoic sedimentary cover and exposed Archean basement of the southern São Francisco craton. 2Dconductivity structures derived by joint inversions of the TE and TM polarization modes and a separateinversion of the tipper components show signatures of the past tectonomagmatic events that affected thearea. A gravity-defined suture zone beneath the Paraná basin is detected in the models as a subverticalconductor extending from crustal to upper mantle depths, which strengthen the interpretation of a Neo-proterozoic collision of the São Francisco craton and a continental block beneath the sedimentary basin.Deep underthrusting of organic graphite-bearing metasedimentary rocks composing a fossil suture zoneis proposed to explain such increase in electrical conductivity. A similar conductivity signature beneaththe sedimentary covered region of the São Francisco craton can be tentatively interpreted as anothersimilar cryptic suture zone. Other isolated high-conductivity anomalies at midcrustal depths below theParaná and APIP provinces are interpreted as fossil residues representing precipitated graphite and sul-fide derived from percolating volatiles during the emplacement of Cretaceous mafic-ultramafic volcanics.A very high conductivity wedge into the lower lithosphere is highlighted in the deep mantle beneath theAPIP volcanic complex, coincident with a zone of low velocity defined by seismic tomography. Geochemi-cal evidence indicates that the alkaline magmatism stemmed from a metasomatized upper mantle at onlyslightly raised temperatures. Thus, interconnected carbonatite melts of low melting point and graphitein the lithospheric mantle are the most likely candidates to explain this high conductivity. The seismicand conductivity anomalies are probably triggered by the same source mechanisms (incipient melting)

because they are plausibly related to the same magmatic process. Conductivity anomalies at crustal andmantle depths in the southern segment of the São Francisco craton suggest that its lithosphere wassignificantly affected by the several tectonomagmatic episodes it has experienced throughout its geo-logical history. Consequently, the enhanced conductivity in the lower crust can be genetically relatedto upwelling volatile-rich intrusions whereas upper mantle high conductivity can be related to refer-tilization by infiltrations of low-degree carbonatitic melts from deeper-sourced metasomatic processes.

. Introduction

In central Brazil, the surface geology points to a complex tectonicvolution, marked by Proterozoic continental collisions and Earlyretaceous to Eocene mantle magmatism. Because of the continen-

∗ Corresponding author. Tel.: +55 12 3208 7206; fax: +55 12 3941 1875.E-mail address: padilha@dge.inpe.br (A.L. Padilha).

1 Present address: Departamento de Geofísica, IAG, Universidade de São Paulo,ua do Matão 1226, 05508-090 São Paulo, Brazil.

301-9268© 2010 Elsevier B.V. oi:10.1016/j.precamres.2010.12.003

Open access under the Elsevier OA license.

© 2010 Elsevier B.V.

tal collisions, allochtonous terrains of all sizes are now comprisedinto large collage areas which display the typical imprint of theNeoproterozoic Brasiliano/Pan-African orogenic cycle (Brito-Nevesand Cordani, 1991). The magmatism includes the extensive Paranáflood basalt province and several alkaline mafic to ultramaficprovinces surrounding the large Paleozoic Paraná basin (Ulbrich

Open access under the Elsevier OA license.

and Gomes, 1981). Intruding the southern portion of the Neopro-terozoic Brasília fold belt and bounded by the Paraná basin to thesouthwest and the limits of the exposed São Francisco craton to thenortheast, the Cretaceous Alto Paranaíba igneous province (APIP)is the most important of these alkaline provinces. APIP magmas are

56 M.S. Bologna et al. / Precambrian Research 185 (2011) 55–64

Fig. 1. Generalized geological map in central-southern Brazil (modified from Valeriano et al., 2008). Symbols refer to: 1, Phanerozoic cover; 2, Cretaceous alkaline volcanicr of kimA asemu 8, Neog ain.

ckp

pss(ae2bgP

cmAuamAtttdd

laad

ocks and epiclastic sediments of the Mata da Corda group; 3, Cretaceous diatremesraxá group and associated tholeiitic mafic rocks, ophiolitic melange complexes, bnits (Canastra, Vazante and Ibiá groups); 7, granulitic nappe (Socorro-Guaxupé);roup); 9, Archean/Paleoproterozoic granite-greenstone and gneiss-migmatite terr

omposed of intrusions of kimberlites, lamproites, a large volume ofamafugites, and some carbonatite-bearing plutonic alkaline com-lexes (Gibson et al., 1995).

A Bouguer gravity map from this area reveals linear anomalies,robably related to sutures and juvenile material remnants fromeafloor spreading later deformed during the Neoproterozoic colli-ional events (Pereira and Fuck, 2005). In particular, a steep gradientlower values in the Brasília fold belt and the São Francisco craton,nd higher values in the Paraná basin) has been interpreted by sev-ral authors (e.g. Lesquer et al., 1981; Mantovani and Brito Neves,005) as a Neoproterozoic suture zone between a continental blockeneath the Paraná basin and the Brasília fold belt due to a conver-ence of the São Francisco plate and the block buried beneath thearaná basin.

Upper-mantle P- and S-wave traveltime tomography in SE andentral Brazil has shown that some regions of the alkaline mag-atism are characterized by lower seismic velocity anomalies.lthough low seismic velocities can be associated with a largelyndifferentiable combination of temperature and chemical vari-tions, raised temperatures in the upper mantle have been theirost common explanation in the region (Schimmel et al., 2003;ssumpcão et al., 2004). Unfortunately, heat flow data from cen-

ral Brazil are sparse and no reliable thermal model is available forhat region (Vitorello et al., 1980; Hamza and Munoz, 1996). Thus,he presence of anomalously high temperatures at upper mantleepths underneath the alkaline occurrences cannot be indepen-ently confirmed.

In this paper we describe the results of a deep magnetotel-uric (MT) study along a profile that crosses the gravity gradientnd one of the low velocity anomalies, positioned below the Araxálkaline–carbonatite complex in the southern part of the APIP. Theata were acquired, analyzed and modeled using the most cur-

berlitic affinity; 4, Cretaceous alkaline–carbonatite complexes; 5, Neoproterozoicent slivers and syn-collisional leucogranites; 6, Neoproterozoic quartzite-phylliteproterozoic autochthonous/para-autochthonous metasedimentary cover (Bambuí

rent equipment and techniques. As the MT method is particularlysensitive to the detection of any connected conductive medium, itcan provide unique geological information complementary to othergeophysical methods. The final 2-D conductivity model provides avertical image that shows significant lateral heterogeneity in thepresent-day lithospheric structure. It yields relevant new informa-tion on the current geophysical state of the crust and upper mantleunder the study area, from which inferences can be drawn on dif-ferent tectonomagmatic processes from the past.

2. Geological setting

The South American platform is characterized by a central coreformed by several Archean to Mesoproterozoic cratonic nuclei fromearlier supercontinents amalgamated during the NeoproterozoicBrasiliano/Pan African orogeny in the final assembly of West Gond-wana (Almeida et al., 2000). The central core is bordered by activeorogens to the west and northwest, and a Mesozoic–Cenozoic pas-sive continental margin to the northeast and east. In central Brazil,the São Francisco craton, which was continuous with the AfricanCongo craton prior to the opening of the South Atlantic, presentlyencompasses a substantial part of the southeastern Brazilian high-lands. Brasiliano fold-and-thrust belts fringe all sides of the cratonwith strong vergence towards its interior (Alkmim et al., 1993). Aschematic outline of the major geological provinces in this area isshown in Fig. 1.

The MT data used in this study in central Brazil were acquiredalong an E–W profile running from the Phanerozoic sediments andvolcanics of the Paraná basin to the west, across the Araxá synform(a regional fold within the southern APIP with a carbonatite com-plex at its southern border) over the Brasília fold belt in the center,

M.S. Bologna et al. / Precambrian Research 185 (2011) 55–64 57

Fig. 2. Outline of the main provinces in the study area and available geophysical information. MT sites are represented by open triangles for broadband-only soundings ando basinp dashet esents lcanick eriod

ac

SwtgueEtmebaeot“bb

iSPctptmspt(swvd

NnTttpc

pen squares for joint broadband and long period soundings. PB stands for Paranáosition of the gravity gradient (Lesquer et al., 1981) and the nearly circular shorthe low velocity seismic anomaly (Schimmel et al., 2003). Different grey tones repredimentary cover; 3, Neoproterozoic rocks of the Brasília belt; 4, Cretaceous voimberlites. Inset at bottom shows induction arrows at long-period MT sites for a p

nd the Neoproterozoic/Phanerozoic cover and exposed Archeanomplex of the São Francisco craton to the east (Fig. 2).

The Paraná basin is a large Phanerozoic intracratonic feature inouth America. Sediments were deposited on southwestern Gond-ana terrain from Late Ordovician to Late Cretaceous, in response

o a geological stabilization after the Brasiliano orogeny. Its strati-raphic succession outlines successive episodes of subsidence andplift, with thickness up to 7000 m in the central depocenter (Zalánt al., 1990). During breakup of the Gondwana continent in thearly Cretaceous, one of the most voluminous volcanic events inhe Earth’s history covered the basin with basaltic lava flows, while

any dykes and sills also intruded the sedimentary sequence (Melfit al., 1988). Although the origin of the lava flood was initiallyelieved to be related to the Tristan da Cunha plume (e.g., Whitend Mackenzie, 1989), a more recent geochemical and geophysicalvidence (Ernesto et al., 2002) appears to discard the hypothesisf a plume source for the generation of the lavas and restrictshe plume contribution to a source of heat only. A Proterozoiccratonic” nucleus was inferred beneath the basin (Paranapanemalock; Almeida et al., 2000), which is bounded by several mobileelts.

The Brasília belt is part of the large Neoproterozoic Brasil-ano/Pan African orogen that developed between the Amazonian,ão Francisco and Paraná-Rio de La Plata cratons in central Brazil.roterozoic metamorphic rock units constitute most of the belt,omprising passive margin sequences of the São Francisco con-inent, back-arc and fore-arc basin sequences, and a youngerost-inversion platform sequence deposited in a foreland basin ofhe São Francisco craton (Pimentel et al., 2001). In the southern seg-

ent of the belt, the Brasiliano deformation involved overthrustheets (nappes), transported eastward at least 150 km towards thelatform of the São Francisco craton, and subsequently stacked byhrust faults over Neoproterozoic pelitic and carbonatic sequenceFuck et al., 1994). In this area, the Brasília belt forms a complextructural pattern, locally characterized by foliations associatedith the thrust sheets and long lineaments corresponding to sub-

ertical lateral ramps or wrench faults originated from differentialisplacement of the thrust wedges.

The Late-Cretaceous alkaline magmas of the APIP occupy aW-elongated area between the São Francisco craton and theortheastern border of the Paraná basin (Gibson et al., 1995).

hey were emplaced dominantly into metasedimentary rocks ofhe southern Brasília belt and formed dykes, pipes, plugs, dia-remes, lava flows, pyroclastic deposits, and carbonatite-bearinglutonic complexes. The geological evolution of these carbonatiteomplexes was marked by multistage events of carbonatitic activ-

, BB for Brasília belt and SFC for the São Francisco craton. Long dashed line is thed contour between longitudes 46◦ and 47.5◦W indicates the surface projection of: 1, Archean/Paleoproterozoic rocks of the São Francisco craton; 2, Neoproterozoicrocks; 5, Phanerozoic cover; 6, Cretaceous carbonatite complexes; 7, Cretaceousof 106 s.

ity and related metasomatism. They are usually composed of botha silicate-rock sequence with dominant ultramafics and subordi-nate syenites, and a carbonatitic sequence. Ultramafic fine-graineddykes (phlogopite picrite and bebedourite) crosscut the coarse-grained plutonic rocks (Ribeiro et al., 2005). Chemical patternsdisplayed by some carbonatites and mica-rich rocks samples areindicative of their formation from the same parental magma, whichis thought to be derived by low-degree melting of a metasoma-tized lithospheric mantle, mainly characterized by the presence ofLREE-enriched phases (Traversa et al., 2001). This implies that a keycause of APIP alkaline magmatism is the combination of metaso-matized lithosphere underlain by mantle at only slightly elevatedtemperatures (Comin-Chiaramonti et al., 1997).

The São Francisco craton was the center of West Gondwana andits margins were initially formed during the Archean during sev-eral discrete events. Present geometry is linked intimately withPaleoproterozoic and Neoproterozoic orogenies. After the formerorogeny that brought together the different Archean crustal seg-ments of the craton, the latter, the Brasiliano orogeny, throughconstructive and destructive lithospheric processes, reworked thepaleocontinent margins, shaping them into their present forms(Almeida et al., 2000). The exact boundary of the craton at depthis still a matter of dispute, with the earlier outline defined throughstratigraphic associations and geochronological data (e.g., Almeida,1977) being challenged by regional geological and geophysicalstudies (e.g., Alkmim et al., 1993). The exposed basement in itssouthern part consists of Archean granite–greenstone terrains, inpart reworked during the Paleoproterozoic orogeny, which alsogenerated juvenile gneiss–migmatite–granitoid complexes andmetavolcano-sedimentary successions (Teixeira et al., 2000). Thisarea is also characterized by tholeiitic dyke swarms and sill-typeintrusions from Late Mesoproterozoic to Mesozoic ages. Vast areasof the craton are covered by pelitic and carbonatic rocks of theNeoproterozoic Bambuí Group, representing a post-glacial epicon-tinental marine environment.

3. Magnetotelluric data

MT data were recorded at 27 sites in the period range from0.0008 to 1024 s by commercial remote-referenced broadband MT

systems. These data image the geoelectric structures from the nearsurface to the deep crust. Time-domain electromagnetic (TDEM)data were also acquired at every site for static shift correction of theapparent resistivity data, related to local distortions in the electricfield arising from near-surface inhomogeneities (Sternberg et al.,

5 brian Research 185 (2011) 55–64

1rrtm

ttdfiml

aeWtoosttcsst

rat2gbttppssp

t(atotttawocaa

4

(tlfits

Fig. 3. 2D-resistivity model for the western profile derived from joint inversion of

Figs. 5 and 6 show the MT data, respectively, for western andeastern profiles as contoured pseudosections for the phases andapparent resistivities of both polarization modes. Visually the the-oretical response of the final models fits most of the observations

8 M.S. Bologna et al. / Precam

988; Meju, 1996). At 22 of these sites, other commercial remote-eferenced long period MT systems were operated in the periodange from 10 to 13,653 s, extending the depth range of investiga-ion obtained by broadband systems to great depths into the upper

antle.Modern techniques of robust data processing were applied

o estimate the MT tensor functions using the remote referenceechnique (Egbert, 1997). Reliable broadband MT responses wereerived at almost all stations, excepting for the vertical magneticeld with the broadband instruments. Thus, discussions about theagnetic transfer functions are based on data gathered from the

ong period stations.Fig. 2 displays rotated real induction arrows, at a period mainly

ffected by lower crust and upper mantle structure. Along the west-rn sites, over the Paraná basin, the arrows become aligned in aSW direction, pointing to an increase in conductivity towards

he center of the basin. Near null arrows at the center of the profile,ver the Brasília belt, suggest that the measurements were carriedut just above a major 1D conducting feature, coincident with theurface projection of the low velocity seismic anomaly. Two sites athe western border of this conducting feature point perpendicularo the profile, indicative of 3D effects probably due to anisotropicurrent flow at this region. At the eastern end of the profile, oppo-itely directed arrows suggest currents flowing in a 2D conductingtructure at the contact between Proterozoic deformed rocks andhe Archean terranes of the São Francisco craton.

Decomposition analysis was used to recover the undistortedegional impedance tensor (Groom and Bailey, 1989). The strikezimuths derived for four bandwidths from single site decomposi-ions were shown elsewhere (profile IBI in Fig. 5 of Padilha et al.,006). Over the Paraná basin, there is evidence of period-dependenteoelectric strike variation, especially for the westernmost sites,ut the overall strike is aligned in the NNW direction (or ENE due tohe 90◦ ambiguity in magnetotelluric strike analyses). Upper man-le structure in the Brasília belt appears to be nearly 1D, with littlehase difference between the two orthogonal directions at longeriods. Similarly, dimensionality of the long-period data at theites over the craton is weak, but the easternmost sites presenttrike angles mostly in the ESE (or NNE) direction at periods sam-ling lower crust and upper mantle structure.

The profile was divided into two overlapping segments overhe 1D Brasilia belt sites and a multi-site, multi-period analysisMcNeice and Jones, 2001) was applied to the data in order to derivecommon strike direction. Excluding the three long-period from

he westernmost sites over the Paraná basin and the five sites withnly broadband data, tensor decomposition gives acceptable fits tohe measured impedance data for a regional strike of N23◦W forhe westernmost sites and N18◦E for the easternmost sites overhe São Francisco craton, henceforth referred as western profilend eastern profile, respectively. These directions are consistentith the regional trend of exposed geological structures and with

rientation of induction arrows, roughly parallel to the profile, andonsequently perpendicular to the geologic strike for the sites notffected by the large conducting feature and neither by the localizednisotropic structure.

. Two-dimensional inversion

The 2D REBOCC inversion code of Siripunvaraporn and Egbert2000) was used to derive a model of the subsurface conductivity

hat reproduces the observed MT data. Due to the presence of aocalized midcrustal anisotropic layer near the center of the pro-le, the procedure described by Heise and Pous (2001) was usedo recover the geoelectric structure. Also, we opted for joint inver-ions of the two impedance polarizations (TE and TM modes) and

TE- and TM-mode data. GG and LSV are the locations of the gravity gradient and thelow seismic velocity anomaly, respectively. Crustal thickness is based on seismicdata (Berrocal et al., 2004). Labeled features A–E are discussed in the text.

a separate inversion of the magnetic transfer functions becausethere are no reliable vertical magnetic data in the 4 first decades(0.001–10 s) and theoretical studies have shown that arrows areseverely affected when anisotropic structures are present (Pek andVerner, 1997).

The western profile comprised 14 sites, including all long-period except the 3 westernmost stations of the Paraná basin andthe 7 nearly 1D stations over the Brasília belt. The eastern profilecomprised 12 sites, including all stations over the São Franciscocraton and the 7 stations over the Brasília belt. An error floor of5% in the complex impedance was used to prevent the inversionfrom being dominated by those data with unrealistically smallvariances. The starting models were either uniform half-spaces or alayered half-space determined from 1D inversion of the geometricmean of all sites.

The final 2D-conductivity models are presented in Figs. 3 and 4for the western and eastern profiles, respectively. They show a goodagreement in the overlapping zone of both profiles. The overall RMSmisfits are 1.9 and 2.0, respectively for the western and easternprofiles, which indicate a good fit between observations and modelpredictions. Some misfits are observed in a few noisy sites and inthe TE mode apparent resistivity data at long periods.

Fig. 4. The same as in Fig. 3 for the eastern profile. Additional features labeled F andG are discussed in the text.

M.S. Bologna et al. / Precambrian Research 185 (2011) 55–64 59

F odel st s, wh

wt

padcGztattttttatlt

rc

ig. 5. Comparison of field and model apparent resistivity and phase data for the mhe modeled data. The top two pseudosections are the apparent resistivity response

ell, with larger misfits being observed in the western profile athe long periods of the TE apparent resistivities.

The regions marked by the letters from A to G in Figs. 3 and 4resent the major electrical features of the models. They are: (A)subvertical conductor extending from crustal to upper mantle

epths (total conductance up to 1000 S) beneath the Paraná basin,oincident with the location of the gravimetric gradient, shown asG (see Fig. 2), that has been interpreted as a signature of a sutureone (Lesquer et al., 1981); (B1–B4) a horizontal sequence of sec-ioned alternating conductors and nonconductors at crustal depthslso beneath the Paraná basin and extending up to its border withhe Brasília belt; (C1–C3) some isolated horizontal crustal conduc-ors mostly concentrated beneath the Brasília belt in the region ofhe Araxá complex; (D) a large conductor at lower crust-upper man-le depths, located under the Neoproterozoic sedimentary cover ofhe São Francisco craton; (E) a huge upper mantle highly conduc-ive zone, coincident with the position of the seismic low-velocitynomaly (LSV) under the Araxá complex; (F) a lower crust conduc-or beneath the Archean rocks of the São Francisco craton, probably

inked to (G) another deep upper mantle conductor under the cra-on.

Because the usual 2D inversion of data affected by anisotropyecover the anisotropy as a sequence of alternating resistive andonductive dyke-like media, features B1–B4 can be replaced by an

hown in Fig. 3. The left column gives the field data, whereas the right column givesereas the bottom two are the phases.

anisotropic mid- to low-crust structure. Using the formulae of Eiseland Haak (1999), an intrinsic anisotropy ratio of 12 was estimatedfrom the average width and resistivities observed in the model. Thisratio is in close agreement with that of a crustal anisotropic layerdetected by another MT profile, positioned about 100 km north ofthe limits of the seismic anomaly and crossing the cluster of dia-tremes in the central part of the APIP (Bologna et al., 2005). Also,the behavior of the induction arrows at the two sites to the westof the deep conducting structure can be explained by the presenceof crustal anisotropy. As shown by Pek and Verner (1997), a largelateral contrast in conductivity overlain by an anisotropic layerdeflects the arrows, which point towards the direction of preferredconductivity within the anisotropic layer instead of perpendicularto the predominant induced currents.

Separate tipper inversion using only the long period data wasperformed to get additional information on lithospheric conductivestructures. It helps to locate the horizontal position of the mostprominent conductors, but does not determine their depths. Thethree westernmost sites over the Paraná basin and the easternmost

site over the São Francisco craton were excluded from the inversionof the tipper data that used the remaining 18 long-period MT sites.Based on data quality, the period band was limited to the intervalof 20–800 s and an error floor of 0.025 was assumed for the realand imaginary tipper components. The REBOCC program was used

60 M.S. Bologna et al. / Precambrian Research 185 (2011) 55–64

ulated

ih

rsaT

Fp

Fig. 6. Comparison of observed and calc

n the inversion procedure and the initial model was a 100 �malf-space.

The final RMS misfit of the tipper inversion was 2.1 and the

esistivity model is shown in Fig. 7. The pseudosections in Fig. 8how that the final model response reproduces the observed datadequately, with larger misfits in the imaginary tipper component.hey are concentrated at short periods in the anisotropic region

ig. 7. Resistivity depth section from 2D-inversion of tipper data along the entirerofile.

pseudosections for the model of Fig. 4.

beneath the Paraná basin and at the whole period band in the borderbetween the Brasília belt and the São Francisco craton.

Due to the overall high resistivity of crustal rocks in the region,the signal penetration depth exceeds several tens of kilometers atthe shortest period available for the tipper responses. Hence, crustalstructure is not resolved in the resistivity-depth distribution modelof Fig. 7. Also, by defining the structural boundaries of regions ofenhanced conductivity, the tipper data are more suited to map geo-logical structures marked by large lateral conductivity contrasts.Both aspects are reflected in the model of Fig. 7 where only thelargest conductivity anomaly (feature E) and the two main verticalstructures (features A and D) are resolved.

5. Discussion

All the conductivity features highlighted by the final 2D modelsappear to be robust and independent of the initial model.

5.1. Feature A – a Neoproterozoic suture related to continental

collision

Feature A is spatially associated with the gravity-defined suturezone beneath the Paraná basin. It must be noted that a magnetome-ter array study carried out over the same region of the Brazilian

M.S. Bologna et al. / Precambrian Research 185 (2011) 55–64 61

F e fieldi

scsbo

smJfbtoectd

bovadcdlN

5a

tcaM2toefsm

ad

ig. 8. Pseudo-sections of the magnetic transfer functions. The left column gives thmaginary and real parts of the functions, respectively.

hield showed negligible vertical response at periods samplingrustal and upper mantle depths (Padilha et al., 2006), that woulduggest the presence of a deep conductor underneath, but spacingetween the stations (around 100 km) would restrain the detectionf such a localized conductor.

Regarding the source of the anomaly, several studies havehown that a subvertical increase of electrical conductivity is com-only observed in fossil suture zones (e.g., Korja and Hjelt, 1993;

ones et al., 2005; Selway et al., 2009). Wherever it is possible toollow them to the surface, these conductors appear to be causedy deep underthrusting of organic graphite-bearing metasedimen-ary rocks down the suture zones. The formation of graphite fromrganic carbon would require high temperatures unless strainnergy derived from the transpressive regime is present. In thisase, graphitization can occur at greenschist facies pressure andemperature conditions normally found at typical mid-crustalepths (Nover et al., 2005).

Following this interpretation, feature A could be related toiogenic carbon, in the form of graphite, derived from a formerrganic-rich basin deeply underthrust during Neoproterozoic con-ergent tectonic processes that involved the São Francisco platend a cratonic nucleous beneath the Paraná basin. Geological evi-ence for these basins underlining the São Francisco plate borderonsists of the continental sedimentary deposits and acid volcanicsated at ca. 1750 Ma in the surrounding fold belts and in the fore-

and (Alkmim et al., 1993). These are associated with records ofeoproterozoic ensialic rifts, which later evolved into ocean basins.

.2. Features B and C – residual remains of Cretaceous magmaticctivity at midcrustal depths

Absence of any obvious crustal sedimentary sources in the cen-ral part of the profile eliminates the hypothesis of a biogenicomponent to explain the B and C anomalies. Features B1–B4re correlated to the midcrustal anisotropic layer detected by anT profile across the central part of the APIP (Bologna et al.,

005). A complex tectonic evolutionary model was proposed forhat region, with Proterozoic transpressional deformations alongbliquely convergent blocks reactivated during distinct magmaticpisodes. The source of the conduction would probably comerom interconnected grain-boundary graphite, sulfides and hyper-

aline fluids, exsolved and precipitated from upwelling Cretaceousagma.Similarly, features C1–C3 are related to the Araxá magmatic

ctivity and can be ascribed to residual remains of magmaifferentiation at midcrustal depths or to unexposed intrusive

data and the right column the modeled data, with Im(T) and Re(T) indicating the

carbonatite-bearing complexes, also detected in aeromagnetic data(Bosum, 1973). The enhanced conductivity can be explained byelectronic conduction through interconnected graphite in carbon-rich rocks precipitated from magmatic fluids moving along faultzones, and associated mineralization (see Nover, 2005).

5.3. Feature D – a cryptic suture or residual remains fromCretaceous magmatic activity beneath the Neoproterozoicsedimentary cover of the São Francisco craton

Feature D is positioned within the cratonic domain and presentsa resistivity two orders of magnitude less than that of the resistivecrust in this region. Except for the particular cases of exotic carbon-and sulfide-rich lithologies, high conductivity at crustal depths isfrequently attributed to proportionally minute amounts of highconductors in interconnected grain-boundary arrangements. Thecommonly mentioned conductors are saline fluids, graphite andmetallic oxides (Jones, 1992). Fluids are an unlikely source for thefeature here because it is not probable for fluids to remain trappedat middle-lower-crust depths of a stable area for long periods oftime. Metallic oxides can also be discarded because they normallyoccur in discrete ore bodies and there is no evidence of large-scalemineral resources in aeromagnetic data from this part of the craton.

Conductivity increase at great depths below long-stable geo-logical terrains can be explained by the presence of a solid-phaseconductor, principally graphite. As previously discussed in relationto feature A, graphitic horizons have been interpreted as tectonicmarkers possibly related to old orogenies. These ancient tectonicprocesses often leave components that are good electrical con-ductors as traces of deep deformations of different crustal units,revealing thus the location of palaeosuture zones or even morespecific tectonic assemblages (Boerner et al., 1996).

Feature D could thus be an expression of an older suture whichhas somehow been preserved within the São Francisco cratonicrocks. Alternatively, the feature could be related to a major accre-tionary event to the southwestern margin of the Archean craton.It has been proposed that the present day surface limit of the SãoFrancisco craton is only the stable foreland block of a larger Pale-oproterozoic plate, which included the entire foreland domain ofthe Brasilia belt (Sanfranciscana continental plate – Alkmim et al.,1993; Pereira and Fuck, 2005). A seismic high-velocity anomaly

west of the surface limit of the São Francisco craton supports thishypothesis (Assumpcão et al., 2004). Also, the geographic posi-tion and the strike of this proposed suture are in close agreementwith another local gravity gradient which would be the geophysi-cal divider that separates the São Francisco craton from the Brasília

6 brian R

bgadhaakpstz

ttcwht

5u

tocrwtkir(

abioetstaanpcwnCt

1aatcdcab(tee

2 M.S. Bologna et al. / Precam

elt domains. The enhanced conductivity implies the presence ofraphitic or sulfide-rich sedimentary rocks, probably formed inrestricted oceanic environment, that have undergone complex

eformation and underthrusting in the deep crust. On the otherand, it must be observed that suture zones resulting from theccretion of terrains after subduction of an intervening ocean usu-lly also produce a series of ophiolites and volcanism, which are notnown to occur in this area. Subsequent weathering and metamor-hism would explain the absence of such characteristics in someutures. Nonetheless, additional studies are necessary in the areao detect other geological and geophysical features typical of theseones.

Another possibility for the source of this geoelectric signature ishat the widespread Cretaceous magmatic activity responsible forhe B and C crustal anomalies beneath Paraná basin and Brasília beltan also be responsible for this lower crust/upper mantle anomalyithin the São Francisco cratonic domain. In this case, instead oforizontal sill-like material the magmatic feature would appear inhe form of a large, subvertical, dyke-like intrusion.

.4. Feature E – signature of Cretaceous magmatic activity atpper mantle depths

The enhanced conductor E at the topmost upper mantle beneathhe Araxá carbonatite complex represents one of the rare casesf extraordinary conductivity anomaly in the upper mantle thatannot be associated with a thinned lithosphere. The anomalyeaches a conductivity value close to 0.1 S/m at about 100 km depthhere it is at least three orders of magnitude more conductive than

he adjacent lithosphere underlying the NW Paraná basin. Othernown examples of enhanced conductivity in cratonic lithospherenclude the central Slave craton and the western part of the Supe-ior Province, both in Canada (Jones and Craven, 2004), South AfricaJones et al., 2009) and northern Australia (Selway et al., 2009).

In our study region, the spatially coincident seismic low-velocitynomaly has been interpreted as due to higher temperatureseneath the alkaline provinces (contrast of almost 150 ◦C, accord-

ng to Assumpcão et al., 2004), associated with a remnant effectf Cretaceous hot mantle upwellings. However, it must be consid-red that most of such temperature pulse would have subsidedhrough heat dissipation since the Cretaceous events and any pos-ible residual temperature anomaly that could have persisted sincehe passage over a hotspot would now occupy a broader bulget asthenospheric depths (see Gallagher and Brown, 1997; Eatonnd Frederiksen, 2007). Thus, large temperature differences areot expected to have persisted from the Cretaceous up to theresent-day at shallow upper mantle depths in such a tectoni-ally inactive area. Furthermore, a temperature gain of about 150 ◦Could explain a conductivity increase of only one order of mag-itude, at the most, of the observed anomaly for mantle rocks.onsequently, high temperature is not expected to be the source ofhe upper mantle conductivity anomaly.

Alternatively, the small seismic velocity anomaly (less than%, following Schimmel et al., 2003) and the large conductivitynomaly could be caused by chemical variations, including volatilend melt content. Thus, proposed effects of dissolved water on elec-rical conductivity and seismological characteristics should also beonsidered (Karato, 2006). For instance, the conductivity of anhy-rous mantle minerals would be raised by one order of magnitudeonsidering undersaturated hydrous minerals with 0.1 wt.% H2O,ccording to Yoshino et al. (2009). On the other hand, the car-

onatite complexes of Brazil are rich in hydrous mantle mineralssuch as phlogopite) that can be stable under upper mantle condi-ions within the depth range of the observed seismic anomaly (Satot al., 1997) and can lower the seismic velocities significantly (Glahnt al., 1992). Phlogopite, or more properly tetra-ferriphlogopite, has

esearch 185 (2011) 55–64

also been suggested to enhance mantle conductivity in at leastone order of magnitude at about 150 km (Boerner et al., 1999).This possibility must be regarded with caution because the con-ductivity of such rocks has not been measured at appropriateconditions in laboratory (Ledo et al., 2004) and it is questionedwhether these minerals can compose broad interconnected net-work at lithospheric mantle. Also, contrary to the hydrous mineralsconductivity hypothesis, an occurring volcanic activity would per-sist until the low-melting point metasomatized layer is depleted,consuming most of the hydrous minerals. Moreover, regional MTresponses at long periods indicate that the conductive structurein central-southeastern Brazil is only weakly anisotropic at uppermantle depths (anisotropy factor around 3; Padilha et al., 2006),a typical result of a dry lithospheric mantle (e.g., Du Frane et al.,2005).

Based on similarities between isotopic signatures it has beenproposed that the Late Cretaceous alkaline magmatism aroundthe Paraná basin was developed from the cooler margins of thesame anomalous thermal regime responsible for the Paraná basintholeiites (Comin-Chiaramonti et al., 1997). These authors sug-gested that a hydrous mantle source at normal mantle potentialtemperature (1280 ◦C) would be responsible for producing the alka-line melt from relatively low degrees of melting of lithosphericmantle rocks. For the deep areas surrounding the anomaly, thatshow much smaller conductivities, laboratory based conductiv-ity of anhydrous mantle minerals, at the above temperature level,gives conductivities of about 0.01 S/m (Yoshino et al., 2009), whichare compatible with the values found at depths of 150 km out-side the anomaly. Such temperature–depth relationship suggeststhe occurrence of a regional geotherm under steady state ther-mal conduction much less than 50 mW/m2. Such geotherm wouldstill allow incipient melting of carbonated peridotites below depthsof 110 km (Dasgupta and Hirschmann, 2006). Therefore, previousmetasomatic processes in the southern São Francisco cratonic litho-spheric mantle are likely responsible in fostering the conditionsfor low-degree carbonatite melting which would produce smallamounts of interconnected melts with variable CO2 and Fe con-tent. There is petrologic evidence for traces of carbonatite meltsin Cretaceous mafic dikes occurring to the northeast of the studyarea (Rosset et al., 2007). The effects of carbonatite melts on electri-cal conductivity are well discussed in the literature (Gaillard et al.,2008). Velocity perturbations due to a low amount of carbonatemelt or iron content increase are within the RMS misfit of thevelocity models derived for this area (Feng et al., 2007).

Widespread carbonatite-bearing intrusions in the APIP indicatethat carbon is the most likely candidate to explain the enhancedconductivity in the upper mantle, above the depth of graphite-diamond stability field. For geotherms of 50–40 mW/m2 that areexpected for the study area, graphite would occur only above100–150 km. The Araxá complex was the most affected by carbon-atitic activity, which was recurrent until the last evolution stages ofthe magmatic events (Ribeiro et al., 2005). Interconnected graphiteon grain boundaries form is known to reduce the electrical resistiv-ity in the topmost upper mantle to the levels detected in the studyarea (see, Jones, 1999).

5.5. Features F and G – signatures of Archean to Cretaceousmagmatic activity in the São Francisco cratonic lithosphere

High conductivity layers at the base of the crust and in a deeperregion of the upper mantle beneath the southeastern portion of

the São Francisco craton were already reported in a recent study byPinto et al. (2010). They integrated the results of two other radial MTprofiles across the exposed Archean rocks of the craton with avail-able data of Bouguer gravity and geoid anomalies that occur in theeastern portion of the study area. A high density/conductivity layer

brian R

aoetrdtimlpose

6

BsemtPtbmgboscrTtt

atrcdaptoditdratimo

A

(((Ci

M.S. Bologna et al. / Precam

t the base of the crust was interpreted, respectively, as evidencef underplating associated with Mesoarchean and Proterozoicpisodes of tholeiitic dike intrusion, causing crustal deformationshat might have facilitated later infiltration of carbon- and metal-ich volatiles. The intrusive basaltic magmas were assumed to beenser, as required by the gravity signal, with an additional conduc-ive solid phase (graphite, sulfide) to generate the high conductivityn the lower crust. In the southern São Francisco cratonic upper

antle, metasomatic processes would foster the conditions for aow-degree carbonatite melting because of its relatively low tem-erature melting point. Forward geoid modeling, using the scopef the low electrical resistivity region within the mantle as a con-traint, entails a density increase (40–50 kg/m3) possibly due to Fenrichment of mantle minerals related to refertilization.

. Summary and conclusions

An MT survey across an old continental lithosphere in centralrazil reveals a richness of conductivity anomalies on differentcales. The region has experienced a broad range of tectomagmaticvents including periods of compressional deformation and mag-atic activity, which have contributed to the total expression of

he electrical signatures. A collisional plate boundary beneath thearaná basin inferred by gravity data appears as a major subver-ical linear conductor in the MT model. In this case, MT proves toe an efficient technique in imaging old sutures where metasedi-entary rocks containing interconnected conducting phases, like

raphite, produce a high-conductivity contrast and make it possi-le to image the present geometry of the structures. The capabilityf the method to find markers related to continent-continent colli-ions is still more crucial in the southern part of the São Franciscoraton, characterized by limited outcrops, deep weathering, thickegolith and sedimentary cover and lack of systematic mapping.he possibility of the presence of a suture zone beneath the Neopro-erozoic sedimentary cover of the craton opens new questions tohe debate on the West Gondwana continental accretion processes.

Several nearly horizontal high-conductivity crustal anomaliesre imaged below the Paraná basin and APIP domains. They areentatively associated with fluid-deposited graphite and sulfideelated to the emplacement of Cretaceous mafic-ultramafic vol-anics. A well-resolved high conductivity anomaly at upper mantleepths is spatially correlated with a zone of low seismic velocitynd the surface expression of the Araxá alkaline–carbonatite com-lex within the southern APIP province. Geochemical studies showhat the APIP alkaline magmatism was a result of small degreesf melting of a metasomatically enriched source at upper mantleepths. Therefore, the spatially coincidental seismic and conductiv-

ty anomalies are probably related to the same magmatic source ofhe Araxá complex. The high conductivity is likely related to fluid-eposited interconnected carbon and incipient carbonatite meltsespectively at the topmost and lowermost upper mantle, whichlso explains the low seismic velocity zone. The MT soundings overhe southern segment of the São Francisco craton provide a physicalmage of a cratonic lithosphere affected by a complex tectonomag-

atic evolution involving continental collisions and rejuvenationf cratonic lithosphere.

cknowledgements

This study was supported by research grants from FAPESP

95/0687-4 and 00/00806-5) and fellowships from CNPq142617/97-0, 350683/94-8 and 351398/94-5) and FAPESP01/02848-0). Thanks to the dedicated field and lab crew at INPE.areful reviews from Kate Selway and one anonymous reviewer

mproved the original manuscript.

esearch 185 (2011) 55–64 63

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.precamres.2010.12.003.

References

Alkmim, F.F., Brito-Neves, B.B., Alves, J.A.C., 1993. Arcabouco tectônico do Cráton doSão Francisco: uma revisão. In: Dominguez, J.M.L., Misi, A. (Eds.), O Cráton doSão Francisco. SBG-SGM-CNPq, Salvador, Brazil, pp. 45–62.

Almeida, F.F.M., 1977. The São Francisco Craton. Rev. Bras. Geociênc. 7, 349–364.Almeida, F.F.M., Brito-Neves, B.B., Carneiro, C.D.R., 2000. The origin and evolution of

the South American Platform. Earth Sci. Rev. 50, 77–111.Assumpcão, M., Schimmel, M., Escalante, C., Barbosa, J.R., Rocha, M., Barros, L.V.,

2004. Intraplate seismicity in SE Brazil: stress concentration in lithospheric thinspots. Geophys. J. Int. 159, 390–399.

Berrocal, J., Marangoni, Y., Sá, N.C., Fuck, R., Soares, J.E.P., Dantas, E., Perosi, F., Fer-nandes, C., 2004. Deep seismic refraction and gravity crustal model and tectonicdeformation in Tocantins Province, Central Brazil. Tectonophysics 388, 187–199.

Boerner, D.E., Kurtz, R.D., Craven, J.A., 1996. Electrical conductivity and Paleo-Proterozoic foredeeps. J. Geophys. Res. 101 (13), 775–813, 791.

Boerner, D.E., Kurtz, R.D., Craven, J.A., Ross, G.M., Jones, F.W., Davis, W.J., 1999. Elec-trical conductivity in the Precambrian lithosphere of Western Canada. Science283, 668–670.

Bologna, M.S., Padilha, A.L., Vitorello, Í., 2005. Geoelectric crustal structures off theSW border of the São Francisco craton, central Brazil, as inferred from a magne-totelluric survey. Geophys. J. Int. 162, 357–370.

Bosum, W., 1973. O levantamento aeromagnético de Minas Gerais e Espírito Santoe sua seqüência quanto à estrutura geológica. Rev. Bras. Geociênc. 3, 149–159.

Brito-Neves, B.B., Cordani, U.G., 1991. Tectonic evolution of South America duringthe Late Proterozoic. Precambrian Res. 53, 23–40.

Comin-Chiaramonti, P., Cundari, A., Piccirillo, E.M., Gomes, C.B., Castorina, F., Censi,P., De Min, A., Marzoli, A., Speziale, S., Velásquez, V.F., 1997. Potassic and sodicigneous rocks from Eastern Paraguay: their origin from the lithospheric mantleand genetic relationships with the associated Paraná flood tholeiites. J. Petrol.38, 495–528.

Dasgupta, R., Hirschmann, M.M., 2006. Melting in the Earth’s deep upper mantlecaused by carbon dioxide. Nature 440, 659–662.

Du Frane, W.L., Roberts, J.J., Toffelmier, D.A., Tyburczy, J.A., 2005. Anisotropyof electrical conductivity in dry olivine. Geophys. Res. Lett. 32,doi:10.1029/2005GL023879.

Eaton, D.W., Frederiksen, A., 2007. Seismic evidence for convection-driven motionof the North American plate. Nature 446, 428–431.

Egbert, G.D., 1997. Robust multiple-station magnetotelluric data processing. Geo-phys. J. Int. 130, 475–496.

Eisel, M., Haak, V., 1999. Macro-anisotropy of the electrical conductivity of the crust:a magnetotelluric study of the German Continental Deep Drilling site (KTB).Geophys. J. Int. 136, 109–122.

Ernesto, M., Marques, L.S., Piccirillo, E.M., Molina, E.C., Ussami, N., Comin-Chiaramonti, P., Bellieni, G., 2002. Paraná Magmatic Province–Tristan da Cunhaplume system: fixed versus mobile plume, petrogenetic considerations andalternative heat sources. J. Volcanol. Geoth. Res. 118, 15–36.

Feng, M., Van der Lee, S.V., Assumpcão, M., 2007. Upper mantle structure of SouthAmerica from joint inversion of waveforms and fundamental mode group veloc-ities of Rayleigh waves. J. Geophys. Res. 112, doi:10.1029/2006JB004449.

Fuck, R.A., Jardim de Sá, E.F., Pimentel, M.M., Dardenne, M.A., Soares, A.C.P., 1994.As faixas de dobramentos marginais do Cráton do São Francisco: síntese dosconhecimentos. In: Dominguez, J.M.L., Misi, A. (Eds.), O Cráton do São Francisco.SBG-SGM-CNPq, Salvador, Brazil, pp. 161–185.

Gaillard, F., Malki, M., Marziano, G.I., Pichavant, M., Scaillet, B., 2008. Carbonatitemelts and electrical conductivity in the asthenosphere. Science 322, 1363–1364.

Gallagher, K., Brown, R., 1997. The onshore record of passive margin evolution. J.Geol. Soc. Lond. 154, 451–457.

Gibson, S.A., Thompson, R.N., Leonardos, O.H., Dickin, A.P., Mitchell, J.G., 1995. Thelate Cretaceous impact of the Trindade mantle plume – evidence from large-volume, mafic, potassic magmatism in SE Brazil. J. Petrol. 36, 189–229.

Glahn, A., Sachs, P.M., Achauer, U., 1992. A teleseismic and petrological study of thecrust and upper mantle beneath the geothermal anomaly Urach/SW-Germany.Phys. Earth Planet. Inter. 69, 176–206.

Groom, R.W., Bailey, R.C., 1989. Decomposition of magnetotelluric impedance tensorin the presence of local three-dimensional galvanic distortion. J. Geophys. Res.94, 1913–1925.

Hamza, V.M., Munoz, M., 1996. Heat flow map of South America. Geothermics 6,599–646.

Heise, W., Pous, J., 2001. Effects of the anisotropy on two dimensional inversionprocedure. Geophys. J. Int. 147, 610–622.

Jones, A.G., 1992. Electrical conductivity of the continental lower crust. In: Fountain,

D.M., Arculus, R.J., Kay, R.W. (Eds.), Continental Lower Crust. Elsevier, Amster-dam, pp. 81–143.

Jones, A.G., 1999. Imaging the continental upper mantle using electromagneticmethods. Lithos 48, 57–80.

Jones, A.G., Craven, J.A., 2004. Area selection for diamond exploration using deep-probing electromagnetic surveying. Lithos 77, 765–782.

6 brian R

J

J

K

K

L

L

M

M

M

M

N

N

P

P

P

P

P

4 M.S. Bologna et al. / Precam

ones, A.G., Ledo, J., Ferguson, I.J., 2005. Electromagnetic images of the Trans-Hudsonorogen: the North American Central Plains anomaly revealed. Can. J. Earth Sci.42, 457–478.

ones, A.G., Evans, R.L., Muller, M.R., Hamilton, M.P., Miensopust, M.P., Garcia, X.,Cole, P., Ngwisanyi, T., Hutchins, D., Fourie, C.J.S., Jelsma, H., Aravanis, T., Pettit,W., Webb, S., Wasborg, J., The SAMTEX Team, 2009. Area selection for diamondsusing magnetotellurics: examples from southern Africa. Lithos 112S, 83–92.

arato, S., 2006. Influence of hydrogen-related defects on the electrical conductivityand plastic deformation of mantle minerals: a critical review. In: Jacobsen, S.D.,van der Lee, S. (Eds.), Earth’s Deep Water Cycle. American Geophysical Union,Washington, DC, pp. 113–129.

orja, T., Hjelt, S.E., 1993. Electromagnetic studies in the Fennoscandian Shield:electrical conductivity of Precambrian crust. Phys. Earth Planet. Int. 81, 107–138.

edo, J., Jones, A.G., Ferguson, I.J., Wolynec, L., 2004. Lithospheric structure of theYukon, northern Canadian Cordillera, obtained from magnetotelluric data. J.Geophys. Res. 109, B04410, doi:10. 1029/2003JB002516.

esquer, A., Almeida, F.F.M., Davino, A., Lachaud, J.C., Maillard, P., 1981. Significa-tion structurale des anomalies gravimétriques de la partie sud du Craton de SãoFrancisco (Brèsil). Tectonophysics 76, 273–293.

antovani, M.S.M., Brito Neves, B.B., 2005. The Paranapanema lithospheric block:its importance for Proterozoic (Rodinia, Gondwana) supercontinent theories.Gondwana Res. 8, 303–315.

cNeice, G.W., Jones, A.G., 2001. Multisite, multifrequency tensor decomposition ofmagnetotelluric data. Geophysics 66, 158–173.

eju, M.A., 1996. Joint inversion of TEM and distorted MT soundings: some effectivepractical considerations. Geophysics 61, 56–65.

elfi, A.J., Piccirillo, E.M., Nardy, A.J.R., 1988. Geological and magmatic aspects of theParaná basin: an introduction. In: Piccirillo, E.M., Melfi, A.J. (Eds.), The MesozoicFlood Volcanism from the Paraná Basin (Brazil): Petrogenetic and GeophysicalAspects. IAG/USP, São Paulo, Brazil, pp. 1–13.

over, G., 2005. Electrical properties of crustal and mantle rocks - a review of labo-ratory measurements and their explanation. Surv. Geophys. 26, 593–651.

over, G., Stoll, J.B., von der Gonna, J., 2005. Promotion of graphite formation bytectonic stress: a laboratory experiment. Geophys. J. Int. 160, 1059–1067.

adilha, A.L., Vitorello, I., Pádua, M.B., Bologna, M.S., 2006. Lithospheric and sub-lithospheric anisotropy beneath central-southeastern Brazil constrained by longperiod magnetotelluric data. Phys. Earth Planet. Inter. 158, 190–209.

ek, J., Verner, T., 1997. Finite-difference modelling of magnetotelluric fields in two-dimensional anisotropic media. Geophys. J. Int. 128, 505–521.

ereira, R.S., Fuck, R.A., 2005. Archean nucleii and the distribution of kimberlite andrelated rocks in the São Francisco craton, Brazil. Rev. Bras. Geociênc. 35, 93–104.

imentel, M.M., Dardenne, M.A., Fuck, R.A., Viana, M.G., Junges, S.L., Fischel, D.P., Seer,

H.J., Dantas, E.L., 2001. Nd isotopes and the provenance of detrital sediments ofthe Neoproterozoic Brasília Belt, central Brazil. J. S. Am. Earth Sci. 14, 571–585.

into, L.G.R., Pádua, M.B., Ussami, N., Vitorello, I., Padilha, A.L., Braitenberg, C., 2010.Magnetotelluric deep soundings, gravity and geoid in the South São Franciscocraton: geophysical indicators of cratonic lithosphere rejuvenation and crustalunderplating. Earth Planet. Sci. Lett. 297, 423–434.

esearch 185 (2011) 55–64

Ribeiro, C.C., Brod, J.A., Junqueira-Brod, T.C., Gaspar, J.C., Petrinovic, I.A., 2005.Mineralogical and field aspects of magma fragmentation deposits in acarbonate–phosphate magma chamber: evidence from the Catalão I complex,Brazil. J. S. Am. Earth Sci. 18, 355–369.

Rosset, A., De Min, A., Marques, L.S., Macambira, M.J.B., Ernesto, M., Renne, P.R., Pic-cirillo, E.M., 2007. Genesis and geodynamic significance of Mesoproterozoic andEarly Cretaceous tholeiitic dyke swarms from the São Francisco craton (Brazil).J. S. Am. Earth Sci. 24, 69–92.

Sato, K., Katsura, T., Ito, E., 1997. Phase relations of natural phlogopite with and with-out enstatite up to 8 GPa: implication for mantle metasomatism. Earth Planet.Sci. Lett. 146, 511–526.

Schimmel, M., Assumpcão, M., VanDecar, J.C., 2003. Seismic velocity anomaliesbeneath SE Brazil from P and S wave travel time inversions. J. Geophys. Res.108 (10), JB0001871029/2001.

Selway, K., Hand, M., Heinson, G.S., Payne, J.L., 2009. Magnetotelluric constraints onsubduction polarity: reversing reconstruction models for Proterozoic Australia.Geology 37, 799–802.

Siripunvaraporn, W., Egbert, G., 2000. An efficient data-subspace inversion for two-dimensional magnetotelluric data. Geophysics 65, 791–803.

Sternberg, B.K., Washburne, J.C., Pellerin, L., 1988. Correction for the static shiftin magnetotellurics using transient electromagnetic soundings. Geophysics 53,1459–1468.

Teixeira, W., Sabaté, P., Barbosa, J., Noce, C.M., Carneiro, M.A., 2000. Archean andPaleoproterozoic tectonic evolution of the São Francisco Craton, Brazil. In: Cor-dani, U.G., Milani, E.J., Thomaz Filho, A., Campos, D.A. (Eds.), Tectonic Evolutionof South America. SBG, Rio de Janeiro, Brazil, pp. 101–137.

Traversa, G., Gomes, C.B., Brotzu, P., Buraglini, N., Morbidelli, L., Principato, M.S.,Ronca, S., Ruberti, E., 2001. Petrography and mineral chemistry of carbonatitesand mica-rich rocks from the Araxá complex (Alto Paranaíba Province, Brazil).An. Acad. Bras. Ciênc. 73, 71–98.

Ulbrich, H.H.G.J., Gomes, C.B., 1981. Alkaline rocks from continental Brazil. Earth Sci.Rev. 17, 135–154.

Valeriano, C.M., Pimentel, M.M., Heilbron, M., Almeida, J.C.H., Trouw, R.A.J., 2008.Tectonic evolution of the Brasília Belt, Central Brazil, and early assembly ofGondwana. In: Pankhurst, R.J., Trown, R.A.J., Brito Neves, B.B., Wit, M.J. (Eds.),West Gondwana: Pre-Cenozoic Correlations across the South Atlantic Region.Geological Society, London, Special Publication 294, pp. 197–210.

Vitorello, Í., Hamza, V.M., Pollack, H.N., 1980. Terrestrial heat flow in the Brazilianhighlands. J. Geophys. Res. 85, 3778–3788.

White, R., Mackenzie, D., 1989. Magmatism at rift zones: the generation of volcaniccontinental margins and flood basalts. J. Geophys. Res. 94, 7685–7729.

Yoshino, T., Matsuzaki, T., Shatsky, A., Katsura, T., 2009. The effect of water on the

electrical conductivity of olivine aggregates and its implications for the electricalstructure of the upper mantle. Earth Planet. Sci. Lett. 288, 291–300.

Zalán, P.V., Wolf, S., Astolfi, M.A.M., Vieira, I.S., Conceicão, J.C.J., Appi, V.T., Neto, E.V.S.,Cerqueira, J.R., Marques, A., 1990. The Paraná Basin Brazil. In: Leighton, M.W.,Kolata, D.R., Oltz, D.F., Eidel, J.J. (Eds.), Interior Cratonic Basins, 51. AAPG Memoir,Tulsa, OK, pp. 681–708.