lithosphere. Using a regression equation obtained from a global distribution, data from TGA were integrated with those

1. Introduction

The African continent is made up of 4 main

Earth and Planetary Science Lettersobtained by other methods (gravitytopography coherence and thermo-mechanical analysis) providing a spatial coverage

sufficient to establish regional Te patterns for South America and Africa. A comparison and association between the Te

distributions for both continents indicates that for the African plate, the effective elastic thickness map clearly shows a

remarkable dichotomy of the Neoproterozoic rocks and reworked older rocks. But for the case of South American plate that

is moving faster than the African plate, lower Te values are observed only for areas where extensive tectonics with intense

volcanism has acted, suggesting that a colder mantle underlies this continental plate, while a hotter asthenosphere is observed

beneath the African plate. This is in part attributed to its relatively slow motion which prevented dissipating the earlier

developed high temperature.

D 2004 Elsevier B.V. All rights reserved.

Keywords: effective elastic thickness; tidal gravity anomaly; Gondwana; asthenosphere thermal stateLithosphere mechanical behavior inferred from tidal gravity

anomalies: a comparison of Africa and South America

Marta S.M. Mantovania,*, Wladimir Shukowskya,

Silvio R.C. de Freitasb, Benjamim B. Brito Nevesc

aIAG-USP, Rua do Matao, 1226-, 05508-090-Sao Paulo SP, BrazilbUFPr-CPGCG Centro Politecnico, 81531-990-Curitiba PR, Brazil

cIG-USP, Rua do Lago, 562, 05508-080 - Sao Paulo SP, Brazil

Received 19 May 2004; received in revised form 5 November 2004; accepted 7 December 2004

Available online 12 January 2005

Editor: S. King

Abstract

Earlier studies have shown that the amplitude difference of the M2 gravity tidal component (TGA) between the measured

and calculated response for a viscoelastic Earth is significantly correlated to the effective elastic thickness (Te) of the0012-821X/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.epsl.2004.12.007

* Correspon

30915034.

E-mail address: marta@iag.usp.br (M.S.M. Mantovani).230 (2005) 397412

www.elsevier.com/locate/epslest Africa, Congo,cratons of pre-Pan-African ageWding author. Tel.: +55 11 30914755; fax: +55 11Kaapvaal [1] and Tanzania [2]which have remained

stable for long periods of geological time. These

anetacratonic blocks are flanked by younger fold belts of

Neoproterozoic (Pan-African Cycle) and of Phaner-

ozoic development. Successive periods of rifting

occurred after these structures have stabilized [3].

This wide range of geological structures is ideal to

study the behavior of the rigidity parameter with

time.

Recent studies on the lithosphere evolution are

focusing their attention on the flexural rigidity

parameter (D), or alternatively on the effective elastic

thickness (Te), as indicative of the mechanism of

isostatic compensation beneath the plate. These

parameters are controlled by geological time scale

relaxation and, therefore, are strongly dependent on

the temperature and deformation rate [4]. This

implies that the lithosphere is stronger for slower

deformation rates and weaker when exposed to

higher temperature or, in other words, Te varies

laterally as a function of the thermal gradient and of

its derivative with time.

The stress pattern of a plate can be regarded as a

bpredictor componentQ of its evolutionary thermalstate, since brittle failure under tension occurs at lower

stress difference than under compression [5]. The

brittle to ductile transition of rocks is a function of

temperature, and so is dependent on the local geo-

therm [6]. It is also a function of the composition, as

most of the strength of the lithosphere lies in its upper

part, which can be broadly approximated by the

rheological behavior of dry olivine [7]. The occur-

rence of thermal insulation under large continental

masses [8] and the presence of mantle plumes [9]

increase the mantle temperature, leading to a change

in the evolutionary development of the stress pattern.

In time, this will affect the lateral Te distribution of the

plate. Therefore, Te can provide information on the

thermal state and on the mechanical properties of the

lithosphere through time.

The combination of Te estimates derived from tidal

gravity anomalies and from isostatic response model-

ing was successfully applied to construct a continental

scale Te map for the South American Plate [10,11].

This method is here applied to the continental part of

the African plate. The resultant Te patterns for both

plates are then compared, and the similarities related

to the elastic parameter of the main tectonic units of

M.S.M. Mantovani et al. / Earth and Pl398each continent are analyzed in the light of the Western

Gondwana breakup process.2. Methodology

2.1. Te from isostatic and thermo-mechanical analysis

In the isostatic analysis, Te is determined by

comparing the observed amplitude of the Bouguer

anomaly with its expected value for a locally

compensated topography. Forsyth [12] introduced

the coherence function taking into account the statisti-

cally independent subsurface loading, and making use

of the relationship between the Bouguer gravity and

topography as a function of the wavelength. The wave

number at the transition between coherent and

incoherent topography and gravity is a measure of

flexural rigidity, and hence of lithosphere effective

elastic thickness Te.

Admittance analysis [13], which considers com-

pensation only for surface loading, underestimates Te

when compared to the coherence analysis [12] which

places loads both on the surface and at the Moho.

Thermo-mechanical analysis takes into account addi-

tional parameters such as temperature, pressure,

mineralogy, xenoliths composition, age, and geometry

of geological structures, and may underestimate Te

where intraplate tectonic stresses exist. In addition,

where there is little power in the topography then any

spectral estimation technique will be biased to strong

values [14]. However, a recent study applied to

southern Africa shows that this bias can be taken into

account using a wavelet transform approach to the

analysis of gravity and topography data [15].

Assuming the lithosphere as an isotropic elastic

plate, Te data from coherence analyses are available,

for South America and Africa, over large areas (e.g.,

[1624]) or along extended profiles (e.g., [2530]).

Evaluations of Te that consider the lithosphere as a

broken elastic plate (e.g., [31]), and from thermo-

mechanical analysis (e.g., [32]) are also available.

In general, the resultant topography of an exten-

sional tectonic regime (e.g.: rifted areas, continental

sedimentary basins) and for large cratonic areas

allows one to perform 2D regional gravity surveys,

while in subduction areas and collisional tectonic

zones Te is preferentially obtained along profiles.

The spatial distribution of Te estimates determined

by isostatic response models or by thermo-mechanical

ry Science Letters 230 (2005) 397412analysis is not uniform over continental areas. This is

so because in many situations it is not possible to

The strong linear dependence between Te and TGA

creates an alternative for estimating the effective

elastic thickness where no large gravity surveys exis

or for small tectonic units for which the coherence

method is not applicable, as pointed out in [17]. When

Eq. (1) is used to estimate Te from existing TGA data

the uncertainty of the estimate is given by

rTe 7:39 7:72 TGA

cos2/ 11:22 TGA

2 228:3r2TGAcos4/

s

2The regression model (Eq. (1)) is shown by the solid

line in Fig. 1. The dashed lines delimit the F1standard deviation interval for predicted Te values

according to Eq. (2).

The regression analysis was based on worldwide

selected data, both from the point of view of tida

anetary Science Letters 230 (2005) 397412 399perform large gravity surveys or a set of long gravity

profiles such that the study area contains a single

tectonic unit, which is a premise for the method, as

pointed out in [12] and [17]. Alternatively, Te can be

estimated from tidal gravity anomaly (TGA), as

shown in [11].

2.2. Te from tidal gravity anomaly correlation

The periodic gravity variation with tides, driven by

mass redistribution, is a characteristic of the dynamic

response of the Earth to an external stimulation. The

attraction of the Moon and Sun causes most of the

Earths observed external tide potential, a super-

position of various frequency components of the

oscillation modes. These are known as long period,

diurnal, semi-diurnal, and ter-diurnal waves, accord-

ing to their oscillation period. The oscillations occur

in different environments, interacting with the atmos-

phere (atmosphere tides), with the oceans (ocean

tides), and with the different layers of the solid Earth:

lithosphere, mantle, and core (Earth tides). The

oceanic tides exert an additional gravitational effect

on the external solid layer of the Earth causing a

flexure of the lithosphere boundary, and a tilt of the

vertical component of gravity due both to the flexure

and to the modified mass distribution. These effects

are known as ocean loading. Among the various tidal

components, the semi-diurnal lunar wave (M2) can be

determined more precisely than other components

from observational data, and is thus preferred for the

present purpose over other components.

The tidal gravity anomaly (TGA) is a useful

measure of the discrepancy between the observed

and modeled tidal gravity, which carries information

on the Earths internal structure. It is defined as the in-

phase component of the vector difference between the

observed tidal gravity corrected for the ocean loading

and the tidal gravity model for a viscoelastic radially

symmetric Earth structure.

The high linear correlation (r=0.88) between thetidal gravity anomalies and Te estimates by the

isostatic response method available at 36 locations

worldwide allowed for the establishment of a linear

regression model [11]:

TGA

M.S.M. Mantovani et al. / Earth and PlTe 70:58F2:72 15:11F3:35 cos2/

1quality data and reliability of the corresponding

coherence Te estimates [11].

This new tool is applied to the 50 tidal gravity

stations presently available on the South American

plate, and to 34 available for the African plate. The

corresponding Te estimates are presented in Tables 1

and 2.

Data from isostatic response analysis and from

tidal gravity anomaly are merged to compose a larger

Te data set and to compare their distribution in the two

Fig. 1. Correlation of M2 tidal gravity anomaly and the lithosphere

effective elastic thickness. Solid line is the regression model (Eq.(1)). Dotted lines delimit F1 standard deviation for predicted Tevalues, according to Eq. (2) (after [11]).t

,

,

l

Table 2

Te from tidal gravity for South America

Id Lat Lon TGA Te Std

6911 14.73 61.15 0.44 74 66975 10.65 61.40 0.03 66 67118 9.94 84.05 0.20 69 67119 10.71 85.23 1.23 43 77201 10.51 66.93 0.69 78 67202 10.55 71.50 1.59 37 77203 8.81 70.86 1.08 46 77250 4.65 74.10 0.16 63 67251 10.39 75.54 0.89 81 67252 12.62 81.69 0.55 55 67253 7.90 72.49 0.72 78 67254 3.45 76.56 0.13 68 67281 5.17 52.68 0.03 65 67303 23.56 46.73 0.38 57 77305 25.45 49.24 0.32 58 77306 29.67 53.82 1.51 100 87307 20.46 54.62 0.08 67 67308 20.77 42.87 1.18 89 77309 15.61 56.13 0.22 61 67310 16.62 49.26 0.68 78 67311 6.53 37.14 0.37 72 67312 22.12 51.41 0.14 68 67313 5.06 42.77 1.15 45 77314 22.40 43.65 0.53 54 77315 3.17 59.83 0.47 74 67316 1.50 48.50 1.31 88 77317 12.97 38.48 0.56 76 67319 23.10 46.96 0.14 68 67408 16.43 71.57 0.41 73 67409 11.99 76.84 0.21 69 67410 6.76 79.84 0.68 77 67411 3.73 73.24 1.02 83 67412 14.82 74.94 0.47 74 67450 0.19 78.50 0.80 79 67451 2.18 79.87 0.21 69 67500 16.49 68.13 0.53 75 67505 17.77 63.19 0.38 73 67506 17.79 63.16 0.38 73 67601 20.29 70.04 0.90 83 77805 34.57 58.41 1.31 99 87810 37.32 59.08 0.20 71 77812 31.67 63.89 0.04 64 77813 27.47 58.78 0.84 47 77814 31.55 68.68 0.96 89 87815 24.73 65.49 0.34 73 77816 41.12 71.42 0.85 92 87817 54.82 68.33 0.52 38 117818 45.83 67.48 0.62 43 97819 36.40 64.20 0.66 83 87895 31.67 55.93 0.95 88 8Id is the station number; Lat and Lon are the coordinates of the

station; TGA is the M2 tidal gravity anomaly (micro Gal); Te is the

Effective Elastic Thickness (km) by Eq. (1); Std is the standard

deviation (km) by Eq. (2).

3801 26.0 28.0 0.58 78 73118 22.8 5.5 0.61 78 73030 2.6 36.9 0.81 51 63421 1.5 30.2 1.02 47 6

3031 3.3 38.6 0.21 62 63410 4.4 18.6 1.48 91 73399 18.8 7.3 1.07 86 73401 12.7 8.0 0.28 71 63010 15.6 32.5 0.13 68 63415 4.2 15.1 0.63 76 63505 15.1 39.3 1.94 102 83325 12.4 1.5 1.97 102 83807 33.9 18.9 0.50 78 73210 26.3 12.8 1.88 24 93005 24.1 32.6 0.51 55 7

3400 13.5 2.1 1.38 91 73312 12.6 12.2 0.37 72 63000 29.9 31.3 0.18 61 7

3321 6.2 5.0 0.14 68 63040 6.8 39.2 0.73 78 63020 2.0 45.4 1.13 46 7

3090 29.2 13.5 0.00 65 73112 36.8 3.0 0.60 82 83300 18.1 16.0 0.73 80 73311 14.4 17.0 0.56 76 63500 26.0 32.6 1.79 27 93601 18.9 47.6 0.18 62 6Id is the station number; Lat and Lon are the coordinates of the

station; TGA is the M2 tidal gravity anomaly (micro Gal); Te is the

Effective Elastic Thickness (km) by Eq. (1); Std is the standard

deviation (km) by Eq. (2).

continents, positioned for 10 My after break-up, at

~130 Ma (Fig. 2).

Contouring was accomplished using the Kriging

interpolation method [33], applying an exponential

variogram model with range of 500 km based on the

spatial autocorrelation analysis of Te [11]. The

Kriging method by itself provides the uncertainty of

the interpolated Te values, as presented in Fig. 3.

The complementary comparison between the two

continental lithosphere fractions, the South American

and African, is indeed effective to strengthen the

anetary Science Letters 230 (2005) 397412Table 1

Te from tidal gravity for Africa

Id Lat Lon TGA Te Std

3014 9.0 38.7 1.21 87 73451 2.6 29.8 0.58 76 63102 33.5 5.1 0.10 68 73806 32.4 20.8 0.07 64 73420 2.1 28.6 0.59 76 63901 22.3 17.1 1.57 98 83495 17.8 31.1 0.07 64 6

M.S.M. Mantovani et al. / Earth and Pl400

Fig. 2. Effective elastic thickness of South America and Africa obtained by the Krigging interpolation method. Data from isostatic response

analysis and from tidal gravity anomaly are merged to compose a larger Te data set and to compare their distribution in the two continents,

positioned for 10 Ma after break-up. Distribution of data and estimated uncertainty of the interpolated Te values are shown in Fig. 3.

M.S.M. Mantovani et al. / Earth and Planetary Science Letters 230 (2005) 397412 401

anetaM.S.M. Mantovani et al. / Earth and Pl402above assumptions and to emphasize the completely

different thermal history of these two segments,

partially, when were still connected (during the

Paleozoic), and even more after the continental drift.

The African portion underwent a more complex

history with more thermal events (slow-moving plate)

compared to the South American plate history.

Clearly, some areas have not remained stable since

break-up, and this reconstruction provides a means of

comparing post 130 Ma history.

3. Tectonic elements of the continental plates

To start any discussion related with the history of

any crustal segment one should begin from the

thermal history of that domain. To understand the

effective elastic thickness of South America and

Africa, one needs to understand the tectonic nature

of the different tectonic domains or units, their related

geophysical spatial pattern, the litho-structural com-

position, and its thermal history.

Fig. 3. Location of Te data and standard dery Science Letters 230 (2005) 397412Discussions of individual studies below intend to

show the available coherence analysis data, their

relative variation and consistency with each tectonic

province.

4. An outline of African tectonics

As it was done for South America [34], it is

possible to record and emphasize, primarily, the

dichotomy of Syn-Pan-African cratons (stable rem-

nants of Earths early continental lithosphere), and

their circumscribing mobile belts. To Northeast and

North (Mauritanides, Atlas) and SW (Cape Fold Belt)

of this general context, narrow Permo-Triassic mobile

belts are present (Fig. 4).

In fact, a critical analysis suggests that the Congo

craton is the site of the highest effective elastic

thickness (over 60 kmFig. 4) [15]. Furthermore, the

easternmost portion of the Damara Pan-African Belt

[35] and Gariep/Saldania are clearly incorporated to

these maximum domains. Therefore, it is not excluded

viation of the interpolated Te values.

anetaM.S.M. Mantovani et al. / Earth and Plthat the small thickness of the supra-crustal meta-

morphic covers (bschist beltsQ) of these mobile beltsand/or their allocthonous character (over imposing the

cratonic basement) provided for these high observed

Te signatures.

Another high effective elastic thickness (74 km)

portion is recorded in the southern part of the

West Africa craton, more precisely the LeoMan

Shield (mostly in Senegal) where Archean nuclei

and Paleoproterozoic rocks prevail. Additional little

Fig. 4. Schematic representation of the main tectonicry Science Letters 230 (2005) 397412 403spots of high Te (to NE and NW) should

correspond to fractions of the Paleoproterozoic

basement within the Pan-African mobile belts not

reworked (between 600 and 500 Ma), in central

Hoggar and South Nubia (KordofanKhartoum).

The last bluish set of high Te corresponds to the

southernmost portion of the Mozambique belt, the

so-called bMozambique ProvinceQ, a poorly knownarea, where older (pre-Neoproterozoic) rock units

occur.

units in Africa (modified after [3] and [36]).

The other African areas show the regional

distribution of the Pan-African belts (Neoprotero-

zoicCambrian), were superposed by several igneous

events (ring complexes and volcanic centers [36]

from the Upper Phanerozoic to the present. The

intraplate magmatism was less intense in the interior

of the cratonic nuclei and prevailingly present in

many extensional activities of the Pan-African belt

domains [3].

Over 600 ring complexes are known through

Africa and Arabia [36], of several ages, since

Neoproterozoic/Pan-African times (about 250 igneous

zones). During the Paleozoic (approximately 95

igneous zones) and especially during Mesozoic times

(over 230 igneous zones) these intraplate igneous

activities have continued, reaching the Cenozoic. It is

possible to say that over 80% of these activities

occurred within the Pan-African domains, many of

them taking advantage of previous tectonic disconti-

nuities formed during that cycle. The bpurpleQ zones(lower effective elastic thickness) clearly follow these

Phanerozoic magmatic zones (Fig. 5).

Additionally, continental rifting has been an

important tectonic process since Late Paleozoic times

eir rel

M.S.M. Mantovani et al. / Earth and Planetary Science Letters 230 (2005) 397412404Fig. 5. Distribution of ring complexes (Proterozoic to Tertiary) and thdomains are distinguished from the areas with tectono-thermal activities re

are the main igneous complexes (modified after [3639]).ationship to the Pan-African tectonic structure of Africa. The cratoniclated to the Pan-African belts (Based on the scheme of Fig. 4). Dots

(D=1.410 N m): much of this area is covered withextrusive volcanics from the rift and volcanic cones

aneta(including Mt. Kenya, Elgon, and Kilimanjaro) which

act as surface loads on the lithosphere. The intrusive

volcanic rock assemblages were classified as high-

density loads (e.g. dykes within the Gregory rift) and

low-density loads (e.g. dykes of the Kavirondo rift)

and the topography of the dome was attributed

primarily to surface volcanic loading, while a low

density (hot region or partial molten mantle) was

associated to deep subsurface loads. Ebinger et al.

[17] analyzed 4 tectonic features or provinces: the

Afar Plateau (Te=2149 km; 3 sub-regions), the East

African Plateau (Te=2764 km; 10 sub-regions),in Africa as a whole, and it has generated numerous

fault-bounded basins, according to a series of dis-

continuous and short-lived tectonic events. Seven

major events of tectonism have been identified from

the Permian up to Recent [3]. These extensional

events are reflecting (at the Earths surface) the

occurrence of deep-seated heat sources.

In this context it is important to consider the East

African rift system, where the modern (Cenozoic)

extensional structures (and associated volcanism) lie

within the Mozambique Belt (also called EAO=East

African Orogen) [37].

5. Analysis of results

5.1. Extensional tectonic regime structures

As emphasized above, the lower value of effective

elastic thickness (Teb30 km) in Fig. 2 correspond toPan-African basement zones where Permo-Triassic,

CretaceousEocene, and OligoceneRecent rifts

evolved (ex. East Africa rifts, Niger/Benue Through

Delta, etc.) and Phanerozoic volcanism (ex. Afar,

Cameroon Line, etc.) in different stages. Here, the

comparison of our map in Fig. 2 with Figs. 4 and 5 is

very interesting for the focusing of our discussions.

For the region of the Eastern African System of

Rifts, Te determinations are mainly given by many

authors focusing different objectives [9,15,17,24,

28,3941]. Inversion of the coherence for the Kenyan

rift yielded a best fitting elastic thickness of 25 km23

M.S.M. Mantovani et al. / Earth and Plstable cratons (Te=6480 km; 3 sub-regions), and

magmatic areas (without rifts Te=43 km; 1 sub-region). The apparent thinning of the elastic plate

beneath the uplifted East African System of Rifts and

Afar Plateau validate the presence of a heat source

beneath the uplifted regions. In this particular area of

overcompensation, a continental breakup is under

progress, with an incipient process of ocean formation

by an extensional regime of over 6000 km in length

[17,28]. This feature is typical of a so-called weak

lithosphere. The extensional process is a result of

convective forces in the underlying asthenosphere

because, due to the low heat conductivity of the

continental block and to its nearly stability (slow

motion), the African Plate favored the accumulation

of heat along time by insulation process inducing

convection currents [8], as well as from deep plumes

as for e.g. those rising from the coremantle boundary

(CMB) beneath southern Africa that may connect to

the hot zones in the upper mantle beneath the EAR

system (e.g. [4246]).

Poudjom Djomani et al. [19] investigated the

relationship between different tectonic structures in

the West African region, as Cretaceous rifts (Benue),

Tertiary domal uplift (Adamawa volcanic uplift),

TertiaryRecent volcanoes (Cameroon Volcanic Line

or CVL), Tertiary sedimentary basins (Chad basins),

and cratonic region (Archean reworked Congolese

craton). The mentioned structures may be seen in

Fig. 4.

By the use of the coherence function analysis,

these authors obtained the minima lithosphere

strength (Te=1420 km) beneath the active rifted

and volcanic areas (Benue, CVL, and Adamawa) and

the maxima (Te ~40 km) corresponding to the

reworked unit in Congo. In the mentioned work of

Poudjom Djomani et al. (1995), only the northernmost

part of the Congo area has been analyzed.

5.2. Sedimentary basins

5.2.1. Africa

Although kinematic models of lithosphere exten-

sion account satisfactorily for the structure and

evolution of many sedimentary basins, there is little

agreement about the main aspects of the dynamical

problem [7]. Large, long-lived, and extensive con-

tinental sedimentary basins are generally associated

ry Science Letters 230 (2005) 397412 405to large parts of cratons and three of these basins

(Chad, Iullmedden, and Congo) in Africa are

anetadescribed by [18]. The three basins contain mega-

sequences initiated in Late Jurassic to Early Creta-

ceous. The Chad basin, in particular, is located above

Early Cretaceous rifts which are connected at depth

to the Atlantic margin via the Benue Trough [3].

Coherence studies indicate that the lithosphere

underlying the Congo basin has a Te value in excess

of 100 km, whereas the Chad and Iullmedden overlie

substantially weaker lithosphere (Te=2030 km). The

contrasting value compared with that given in [19] is

related to the size of the analyzed area and a relative

low topography relief, for which spectral analysis

cannot be applied. A Te=84 km value for the Chad

basin reported for a coherence analysis of super-

imposed adjacent areas is rejected by these authors

due to the inclusion of mirrored wavebands which

may have caused an artificial value. Newman and

White [7] concluded that the lithosphere underlying

the Congo basin is strong, whereas the Chad and

Iullmedden basins as well as the respectively

adjacent Hoggar and Darfur Domes overlie substan-

tially weaker lithosphere, suggesting that this vast

area has been previously weakened [47]. These areas

belong to an ancient basement that was reworked

during the Pan-African orogenic cycle; many rifts

and plutonicvolcanic centers are present in these

areas (Fig. 5).

Analyzing the isostatic anomalies, Newman and

White [7] suggested that uplifts may be related to

convective upwelling in the asthenosphere involving

lateral density contrasts; this implies the existence of

thermal anomalies in the mantle underneath.

5.2.2. South America

The larger intracratonic sedimentary basins of

South America were analyzed by [22], Te=1266

km, and by [23], Te=2458 km. The Parnaiba basin

covers an area of approximately 600000 km2 of the

western part of Northeast Brazil; its maximum thick-

ness comprises about 3500 m of Silurian to Creta-

ceous sediments, intruded by magmatic rocks of

Permo-Triassic to Juro-Cretaceous age. The lower

Te values for the Parnaiba basin correspond to the

smaller areas used for coherence inversion; looking to

the shape of coherence plots, the Te=58+4/6 kmvalue is preferred [23].

M.S.M. Mantovani et al. / Earth and Pl406The Parana basin, located in central-south-eastern

South America covers an area of about 1,700,000km2 and is filled by Ordovician to Cretaceous

sediments and Cretaceous volcanic rocks. Vidotti

[22] calculated the effective elastic thickness of this

lithosphere sector using the coherence analysis

technique [12]. As result of her analysis, [22]

concluded that this area is underlain by a bstrongQ(rather rigid) lithosphere with bweakerQ (less rigid)areas within. However, from Figs. 5.3 and 5.6 in

[22], the clear correlation observed between the size

of each analyzed sub-area and Te, supports [17]

argument for the underestimation of Te (when

analyzed areas are too small); therefore her max-

imum Te value obtained for the largest analyzed area

(Te=66+6/4 km) is the best estimate for the Paranabasin. The evolution of the geometry of the Chaco

foreland basin (Bolivia) using seismic reflection,

gravity, and well log data was examined by [30]; the

best fit between their computational results and

experimental data was obtained for an elastic thick-

ness value ranging from 29 to 31 km.

A large volume of continental flood basalts (CFB)

erupted prior to the Gondwana break-up in the

ParanaChaco basin (308S to 108S). Eruptionsoccurred for about 10 My, with a peak of intense

activity between 133 and 130 Ma. The large amount

of erupted lava is attributed to the presence of a plume

[4850], and therefore associated with a hot astheno-

sphere that partially melted the lithosphere to produce

the CFB. TGA (Te=4768 km) values are in agree-

ment with those calculated by other techniques (such

as the coherence analysis).

The Amazonas basin is an elongated NE

structure that splits the Amazonian craton into two

large pieces. Although this basin was gravity

surveyed for petroleum exploration, its shape favors

a coherence analysis technique only along profiles.

Nunn and Aires [51] modeled 4 profiles crossing

the Medio Amazonas Basin obtaining a Te=1520

km up to a maximum value of 40 km; although

their flexure model considers the basin to be

completely filled by sediments of density 2.55 g/

cm3, gravity records indicate intrusions or partial

replacement of lower crust by mantle material.

Taking into account the inadequate assumption

(which ignores the intruded density material), they

concluded that Te was lower than expected and

ry Science Letters 230 (2005) 397412suggested that more complex rheological models

should be applied.

aneta5.3. Large cratonic areas

5.3.1. Africa

Flexural rigidity in cratonic areas was investigated

by [17,21,24]. To estimate the effective elastic thick-

ness of the continental lithosphere these authors used

the coherence technique [12].

A). For the major tectonic provinces of South

Africa (Kaapvaal, to the south, Limpopo belt at the

center, and Rhodesia/Zimbabwe to the north, which

together form the Kalahari craton, Fig. 4), values of

Te=72 km for the Archean Kaapvaal craton and

Te=38 (East) km to 48 (West) km for the Mesoproter-

ozoic NamaquaNatal mobile belt were obtained by

[21]. Stark et al. [15] obtained similar values of Te

using the wavelet transform mapping method.

Doucore et al. [21] considered each tectonic

province as an independent coherent domain on the

basis of topographic features and isostatic response.

From geological and geophysical considerations, they

suggested that the contrast in flexural rigidity of the

Kaapvaal and Namaqua-Natal provinces can be

attributed to combined effects of compositional and

thickness differences of the lithosphere and to the

present asthenosphere heat flow variation. In [24] a

value of Te=64 km for the Kalahari craton was

obtained: a number between the two independent

domains of [21].

B). For the Congo craton a value of Te=101 km

was calculated [24] in agreement with the conclusions

for the Congo basin [7]. For Tanzania craton which is

underlain by hot asthenosphere [4246,52] Te=64F5km. Unlike the other cratons, there is considerable

power in topography here.

As early mentioned in [17], it was shown that Te

estimates using the coherence technique must fulfill

the assumptions imposed by this method, one of

which is to analyze a large enough area covering the

structure.

5.3.2. South America

A). For South America, applying the coherence

method to a large regional gravity survey that covers

Uruguay and the southern portion of Rio Grande do

Sul State in Brazil, values for the Rio de La Plata

craton (RLPC) of Te=95 Km are reported [20]. For

M.S.M. Mantovani et al. / Earth and Pllatitudes between 358 and 258 and longitudes(658, 508), the RLPC is clearly depicted by anintense high (Te=88100 km). RbSr geochronology

from the RLPC of Uruguay was described by [53,54].

Ages measured in granitoids from western Uruguay

(Piedra Alta terrane) range from 1900 to 2200 Ma.

Low Sr initial ratios (N0.7022) are a commoncharacteristic of these rocks. These results confirm

the Paleoproterozoic ages obtained for different

lithologies [53]. This Early Precambrian age of this

lithospheric sector suggests that it has not been

significantly reworked (during late Proterozoic

cycles), and therefore it is cold and rigid in agreement

with the measured bhighQ TGA.B). A study using the coherence technique along

profiles [25], evaluated TeN85 km for the westernGuyana shield and for the southwestern Central

Brazilian shield, that are parts of exposures of the

Amazonian craton. Ussami and Molina [55] obtained

TeN85 km for the eastern margin of the Amazoniancraton using the model of a lithosphere bbroken plateQand assuming as load for their model, the Araguaia

belt.

Only two tidal stations are available for the

Amazonian craton (AC), which limits the resolution

of its boundary outline. Estimated values (Te=7488

km) are similar to those of San Francisco craton

(SFC). The oldest ages of the Amazon craton are

reported within the Carajas area, ranging from 3.1 to

2.5 Ga [56]. Its geochronological pattern decreases in

age from NE to SW, and at least five provinces that

behaved as stable platforms at the end of Meso-

Proterozoic are identified [56]. From the two available

tidal stations it is possible to devise the mentioned

NESW trend, although additional TGA stations in

the area are needed to clarify this feature. Due to the

lack of TGA stations, the AC tectonic boundary is not

clearly imaged.

C). Coherence determinations of Te for the San

Francisco craton (SFC) are not available in the

literature, up to now. Age provinces of the SFC

(3.452.0 Ga) are consistent with an Archean and

Early Proterozoic evolution for the continental crust.

The major tectonomagmatic events of SFC occurred

between 2.1 and 2.0 Ga, at the late stage of its

evolution and consolidation. Teixeira [57] identified

some episodes south of SFC that occurred at 2.82.7

Ga ago.

ry Science Letters 230 (2005) 397412 407This is in agreement with an old and cold

lithosphere, which justifies the observed Te (76 to

aneta89 km). Regardless of its older age, relative to the Rio

de La Plata Craton, the SFC has a slightly lower

rigidity. This could be explained by accepted average

lithosphere thickness: thinner for Archean provinces

compared with Proterozoic provinces [58]. Clearly

associated to SFC, the Congo craton is distinguished

in the west side of the African continent.

5.4. Collisional tectonic structures

5.4.1. Africa

For the Cape Fold Belt, a Permo-Triassic colli-

sional orogen, the southernmost structure of South

Africa, [24] obtained a value of Te=18 km. No Te

records were reported for the Mauritanides structure

of NW Africa.

For the African plate, the effective elastic thickness

map clearly shows the dichotomy of the pre-Pan-

African (older than Late Mesoproterozoic) and the

Pan-African (younger than 900 Ma) regions/mobile

belts (Neoproterozic rocks and older rocks reworked

in that period). But in the case of South America, less

than 5% is in red color while at least 95% are from

green to blue (70NTe/kmN100).

5.5. Subduction tectonic structures

5.5.1. South America

In the South American continent, subduction

structures are associated to the Andean Cordillera,

and Te determinations were obtained along several

profiles. Whitman [26] used the seismically con-

strained shape of the Moho in NW Argentina and

compared it to the gravity data to obtain the flexural

rigidity of the foreland lithosphere (10211022 N m);

he concluded that the corresponding Te (612 km)

was a factor 2 to 4 less than that estimated for the

Bolivian Altiplano [59]. Fan et al. [31] constrained

their study to the Peruvian Andes, and obtained a Te

varying from 25 to 55 km. Their model was based on

the flexural analysis of [59,60]. Watts et al. [27]

presented eight profiles between latitudes 10 and28, in correspondence to the Nazca Plate subduc-tion. In this segment, Te contour lines increase from

25 km near the shore to 100 km where the Brazilian

Shield outcrops; between Central Andes and the fold

M.S.M. Mantovani et al. / Earth and Pl408thrust belt, Te varies from 50 to 75 km. Stewart and

Watts [25] analyzed 58 profiles between latitudes108N and 358S. They used the bbroken plate modelQ,and divided the area into northern and southern

Andes. Results are presented along each profile as

well as a Te contour map of western South America:

Te ranges from 25 km near the shore to N85 km in thecentral Bolivian Range. Along the Andes, the highest

values are located between latitudes 08 and 108.Contours cover part of the western portion of

continental shields where Te extrapolates 85 km.

Along most of the Andean cordillera (for latitudes

08 to 458), estimates of Te (6989 km) belong to thehigher group. This is not in agreement with estimates

of radiometric age, but does fit with the lithospheric

thickness. In fact, a high Te, which is associated with

the rigidity parameter D, is consistent with the

existence of deep seismicity.

Hypocenters of deep earthquakes in the South

American Cordillera indicate that the down going slab

may be divided into discrete segments. The segments

beneath northern and southern Peru and beneath

central Chile have shallow dips (about 108; [61].Although slightly displaced, probably due to gridding

effects, these shallow segments are perceptible in Fig.

2 (a relatively lower Te).

The westward elongation of a high (7593 km),

confirmed by two tidal stations, reaches the southern

Andean Cordillera at latitude 428 to 338. Part ofthis westward extent covers the Chilenia terrane

described by [62]. In the basement of the Central

Andean Chain, Chilenia, Cuyania, and Pie de Palo are

small Proterozoic exotic terranes that have probably

been sutured to Gondwana during the Paleozoic [63].

The Chilenian basement shows a PbPb age of

1069F36 Ma, and records a complex Precambrianhistory [63]. This may explain the higher intensity of

the blue color within the Cordillera structure.

A different interpretation is given to the volcanic

province of the Patagonia microplate (Southern

South-America; centered at 708W, 458S). Theobserved geometry for the Patagonia low (Te=3843

km) coincides with that estimated by seismic aniso-

tropy of surface and body waves [64]. The evolution

of the Andes in this area began in the Middle to

Upper Jurassic with extrusion of voluminous acid

tuffs and lavas [65]. The origin of these volcanic

rocks seems to be related to crustal extension and

ry Science Letters 230 (2005) 397412anatexis predating the opening of the Atlantic and the

Magellan marginal basin.

underlain plate is supported by the seismic records,

which delineate the subduction path, and the

anetavariation in the dip (declivity of the plate) is

reflected in the color nuances (the color intensity

in correspondence of the Nazca plate differs from the

northern and southern segments).

For Patagonia, there is a quite active tectono-

thermal history since the Neoproterozoic III and

especially during all the Paleozoic and Mesozoic

somehow associated to the evolution of SouthernNeogene volcanic activity, mainly in the western

and central Patagonian plateau, is attributed to the

interaction of upwelling sub-slab asthenosphere flow-

ing around the trailing edge of the descending Nazca

Plate [66]. Compositional distribution of these lavas

and the modeled anomalous mantle potential temper-

ature are explained by these authors as a bweakplumeQ beneath the slab window [66].

6. Discussions and conclusions

For the African plate, the effective elastic thickness

map clearly shows the dichotomy of the pre-Pan-

African and the Pan-African mobile belts (Neoproter-

ozic rocks and older rocks reworked in that period).

But in the case of South America, less than 5%

corresponds to Teb70, while at least 95% are withinthe interval 70NTe/km N100.

Although Watts and Burov [14] do not relate the

seismogenic Ts layer to Te, they ponder that brittle

and ductile deformation fields are roughly equally

involved in the support of loads; and because in the

continents there may be more than one brittle-ductile

transition (BDT), the elastic portion of the lithosphere

is more complex than for the oceanic plates. Accord-

ing to Kuznir and Karner [67] Te correspond to the

strong portion of the lithosphere which can be thicker

than 100 km for complex continental sectors and up to

50 km for the oceanic lithosphere with more than one

BDT.

The high values of Te observed for the Andean

range can be clearly related to a BDT effect, the

juxtaposition of a continental plate underlain by a

cold and rigid oceanic plate of the subduction

process. In spite of the depth, the rigidity of the

M.S.M. Mantovani et al. / Earth and PlAndes and to the formation of the Austral Atlantic

Ocean. This intense volcano-plutonic activity corre-sponds to the lower values (reddish) of the effective

elastic thickness. The low Te value and its closest

relationship with Ts relate this sector to its thermal age

[65].

In correspondence to the Amazonian craton, to

the San Francisco craton (the nucleus and its

extension under the marginal belts) and to the

assemblage of the Luis Alves, Rio de La Plata,

and partially Pampia cratons (since their limits are

not distinguishable in this analysis scale and method-

ology) the highest effective elastic thickness

(between 70 and 100 km) with some local attenu-

ation is observed.

Intermediate values are observed for the Guyana

Shield (north of the Amazon craton), probably due to

the partial connection of this area with the tectono-

thermal activity due to its interaction with the

Caribbean plate.

Comparing the color intensity of the South

America cratonic blocks with those in the African

plate, we may attribute a shallower root to the western

cratons of Gondwana compared to its central and

eastern cratons.

The Mesozoic magmatism is present and exposed

(Takutu graben and associated volcano plutonism),

but apparently is not considerable to explain the Te

observed values (between 60 and 70 km).

The Brasiliano age domains (Neo-Proterozoic), as

Tocantins, Borborema (including the portion covered

by the Parnaiba basin) and Mantiqueira, show

distinctive and coherent values (50 to 70 km; from

yellow to light red). To the south (Dom Feliciano

belt) and to the north (Espirito Santo granulitic belt)

are exceptional among the Brasiliano areas because

their Te values are not low (the observed colors

match those from the adjacent cratons). The explan-

ation of scarce presence of supra-crustals (predom-

inance of Neo-Proterozoic high-grade rocks) is yet

speculative, due to the inappropriate distribution of

measured Te.

It is convenient here to remark that the Phanerozoic

history of the Brasiliano dated domains were sub-

mitted only to a few, shallow and of little significance

events of magmatism. This fact distinguishes the

Brasiliano Domains from the above-discussed Neo-

Proterozoic domains in Africa (Pan-African ages), and

ry Science Letters 230 (2005) 397412 409that were benefited by a rigorous magmatic and

Phanerozoic extensive history.

the Phanerozoic. While the Pan-African domains are

persistently red-colored (Teb60 km and even b20

anetakm), the correlated Brasiliano domains show Te

values of 50 to 70 km. This confirms a history of

plate domains, plate, and asthenospheric conditions

completely different between the two continents.

Since the Paleozoic (when the two continents were

assembled and formed Gondwana), and especially

during the Pangea history (Permo-Triassic) and during

the continental break-up (Upper Triassic to the

present), these differences were already present,

supporting the existence of a contrasting astheno-

sphere beneath the two blocks. In other words, a

colder asthenosphere beneath the bfast-movingQ SouthAmerican plate compared to the hotter correspondent,

developed under the African plate due to its slow

motion which preserved the earlier developed high

temperature. This is in agreement with Nyblade and

Robinson [69] that attributed the mantle beneath the

African superswell to heat insulation by the super-

continent Pangea in the Late Paleozoic and Early

Mesozoic, providing possible explanation for why

deep mantle beneath the African superswell may have

elevated temperatures.

Acknowledgments

The authors are grateful to Dr. Scott King, Cindy

Ebinger, and another unknown reviewer for com-

ments and corrections that highly improved the

original manuscript. ORB and ICET provided access

to the tidal gravity database. Figures, cited as

extracted from other authors were redrawn by A.

Rugenski. CNPq, FAPESP, and CAPES financially

supported this research, through exchange coopera-The low Te value observed for the Chaco Plain is

connected to the ParanaChaco sedimentary basin

structure [68] and is interpreted as being the site of the

thermal anomaly that produced the intense volcanism

just prior to the continental split [50].

Thus, the pre-drift between Africa and South

America is in general good (Fig. 2) showing excellent

location for the cratonic domains and the Pan-African/

Brasiliano mobile belts. For the last, it is also

observed the significant difference of behavior during

M.S.M. Mantovani et al. / Earth and Pl410tion among institutions, travel expenses, and gradu-

ate scholarship.References

[1] D.E. James, M.J. Fouch, S. VanDecar, S. van der Lee,

Kaapvaal seismic group, Tectospheric structure beneath south-

ern Africa, Geophys. Res. Lett. 28 (2001) 24852488.

[2] C.J. Ebinger, Y. Poudjom Djomani, E. Mbede, A. Foster, J.B.

Dawson, Rifting Archean lithosphere: the EyasiManyara

Natron rifts, East Africa, J. Geol. Soc. (Lond.) 154 (1997)

947960.

[3] J.J. Lambiase, The framework of African rifting during the

Phanerozoic, J. Afr. Earth Sci. 8 (2/3/4) (1989) 183190.

[4] G. Ranalli, Rheology of the Earth: Deformation and Flow

Processes in Geophysics and Geodynamics, Allen & Unwin,

London, 1987, p. 365.

[5] M.K. McNutt, Implications of regional gravity state of stress in

the Earths crust and upper mantle, J. Geophys. Res. 85 (B11)

(1980) 63776396.

[6] C.J. Hawkesworth, S. Kelley, S.P. Turner, A. LeRoex, B.

Storey, Mantle processes during Gondwana break-up and

dispersal, J. Afr. Earth Sci. 28 (1) (1999) 239261.

[7] R. Newman, N. White, Rheology of continental litho-

sphere inferred from sedimentary basins, Nature 385

(1997) 621624.

[8] D.L. Anderson, Hotspots, polar wander, Mesozoic convection

and the geoid, Nature 297 (1982) 391393.

[9] W.J. Morgan, Hotspot tracks and the opening of the Atlantic

and Indian Oceans, in: C. Emiliani (Ed.), The Sea, vol. 7,

Wiley, New York, 1981, pp. 443487.

[10] M.S.M. Mantovani, W. Shukowsky, S.R.C. de Freitas,

Tectonic pattern of South America inferred from tidal

gravity anomalies, Phys. Earth Planet. Inter. 114 (1999)

9198.

[11] W. Shukowsky, M.S.M. Mantovani, Spatial variability of tidal

gravity anomalies and its correlation with the effective elastic

thickness of the lithosphere, Phys. Earth Planet. Inter. 114

(1999) 8190.

[12] D.W. Forsyth, Subsurface loading and estimates of the flexural

rigidity of continental lithosphere, J. Geophys. Res. 90 (1985)

1262312632.

[13] M.K. McNutt, Compensation of oceanic topography: an

application of the response function technique to the surveyor

area, J. Geophys. Res. 84 (1979) 75897598.

[14] A.B. Watts, E.B. Burov, Lithospheric strength and its relation-

ship to the elastic and seismogenic layer thickness, Earth

Planet. Sci. Lett. 213 (2003) 113131.

[15] C.P. Stark, J. Stewart, C.J. Ebinger, Wavelet transform

mapping o effective elastic thickness and plate loading:

validation, using synthetic data and application to the study

of Southern African tectonics, J. Geophys. Res. 108 (B12)

(2003) 25582577.

[16] T.D. Bechtel, D.W. Forsyth, V.L. Sharpton, R.A.F. Grieve,

Variations in effective elastic thickness of the North American

lithosphere, Nature 343 (1990) 636638.

[17] C.J. Ebinger, T.D. Bechtel, D.W. Forsyth, C.O. Bowin,

Effective elastic plate thickness beneath the East African

ry Science Letters 230 (2005) 397412and Afar plateaus and dynamic compensation of the uplifts,

J. Geophys. Res. 94 (B3) (1989) 28832901.

M.S.M. Mantovani et al. / Earth and Planetary Science Letters 230 (2005) 397412 411[18] R.W. Hartley, P.A. Allen, Interior cratonic basin of Africa:

relation to continental break-up and role of mantle convection,

Basin Res. 6 (1994) 95113.

[19] Y.H. Poudjom Djomani, J.M. Nnange, M. Diament, C.J.

Ebinger, J.D. Fairhead, Effective elastic thickness and

crustal thickness variations in west central Africa inferred

from gravity data, J. Geophys. Res. 100 (B11) (1995)

22.04722.070.

[20] M.S.M. Mantovani, W. Shukowsky, S.E. Hallinan, Analise da

espessura elastica efetiva no segmento litosferico Rio de

LaPlata-Dom Feliciano, An. Acad. Bras. Cienc. 67 (2) (1995)

200220.

[21] C.M. Doucoure, M.J. de Wit, M.F. Mushyandebvu, Effective

elastic thickness of the continental lithosphere in South Africa,

J. Geophys. Res. 101 (B5) (1996) 11.29111.303.

[22] R.M. Vidotti, Lithospheric structure beneath the Parana and

Parnaiba basins, Brazil, from regional gravity analyses, PhD,

The University of Leeds, School of Earth Sciences, UK (1998)

96 pp.

[23] M.A. de Souza, Regional Gravity Modeling and Geo-history

of the Parnaba Basin (NE Brazil), PhD thesis, Univ. of

Newcastle upon Thyne, (1996) 126 pp.

[24] R. Hartley, A.B. Watts, J.D. Fairhead, Isotasy of Africa, Earth

Planet. Sci. Lett. 137 (1996) 118.

[25] J. Stewart, A.B. Watts, Gravity anomalies and spatial

variations of flexural rigidity at mountain ranges, J. Geophys.

Res. 102 (B3) (1997) 53275352.

[26] D. Whitman, Moho geometry beneath the eastern margin of

the Andes, northwest Argentina and its implications to the

effective elastic thickness of the Andean foreland, J. Geophys.

Res. 99 (B8) (1994) 15.27715.287.

[27] A.B. Watts, S.H. Lamb, J.D. Fairhead, J.F. Dewey, Litho-

spheric flexure and bending of the Central Andes, Earth

Planet. Sci. Lett. 134 (1995) 921.

[28] C.J. Ebinger, N.J. Hayward, Soft plates and hot spots: views

from Afar, J. Geophys. Res. 101 (B10) (1996) 21.85921.876.

[29] N.M. Upcott, R.K. Mukasa, C.J. Ebinger, Along-axis segmen-

tation and isostasy in the Western rift, East Africa, J. Geophys.

Res. 101 (B2) (1996) 32473268.

[30] L. Coudert, M. Frappa, C. Viguier, P. Arias, Tectonic

subsidence and crustal flexure in the Neogene Chaco basin

of Bolivia, Tectonophysics 243 (1995) 277292.

[31] G. Fan, T.C. Wallace, S.L. Beck, C.G. Chase, Gravity anomaly

and flexural model: constraints on the structure beneath the

Peruvian Andes, Tectonophysics 255 (1996) 99109.

[32] L.L. Lavier, M.S. Steckler, The effect of sedimentary cover on

the flexural strength of the continental lithosphere, Nature 476

(1997) 476479.

[33] G. Davis, Statistics and Data Analysis in Geology, 2nd edition,

John Wiley & Sons, NY, 1986, p. 646.

[34] M.S.M. Mantovani, W. Shukowsky, S.R.C. de Freitas, Tidal

gravity anomalies as a tool o measure rheological properties of

the continental lithosphere: application to the South American

Plate, J. South Am. Earth Sci. 14 (2001) 114.

[35] C. Hartnady, P. Joubert, C. Stowe, Proterozoic crustal

evolution in Southwestern Africa, Episodes 8 (4) (1985)236240.[36] J.R. Vail, Ring complexes and related rocks in Africa, J. Afr.

Earth Sci. 8 (1) (1989) 1940.

[37] A.M. Goodwin, Precambrian Geology. The Dynamic Evolu-

tion of the Continental Crust, Academic Press, London,

1991, p. 666.

[38] P.G. Eriksson, The 2.72.0 Ga volcano-sedimentary record of

Africa, India and Australia: evidence for global and local

changes in sea level and continental freeboard, Precambrian

Res. 97 (1999) 269302.

[39] C. Petit, C.J. Ebinger, Flexure and mechanical behavior of

cratonic lithosphere: gravity models of East African and

Baikal rifts, J. Geophys. Res. 105 (B8) (2000) 19.15119.162.

[40] C.J. Ebinger, G.D. Karner, J.K. Weissel, Mechanical

strength of extended continental lithosphere: constraints

from the western rift system, East Africa, Tectonics 10

(1991) 12391256.

[41] N.M. Upcott, Structural segmentation of continental rifts as

seen from ship and land gravity data: examples from East

Africa, PhD thesis, (1994) 248 pp., University of Leeds,

England.

[42] J. Ritsema, H.J. vanHeijst, J.H. Woodhouse, Global transition

zone tomography, J. Geophys. Res. 109 (B2) (2004)

101029101047.

[43] J. Ritsema, H.J. vanHeijst, New seismic model of the upper

mantle beneath Africa, Geology 25 (1) (2000) 6366.

[44] R. Montelli, G. Nolet, F.A. Dahlen, G. Masters, E.R. Engdahl,

S.H. Hung, Finite frequency tomography reveals a variety of

plumes in the mantle, Science 303 (2004) 338343.

[45] E. Debayle, B.L.N. Kenneth, Anisotropy in the Australasian

upper mantle from Love and Rayleigh waveform inversion,

Earth Planet. Sci. Lett. 184 (2000) 339351.

[46] J. Ritsema, J. van Heijst, J.H. Woodhouse, Complex shear

wave velocity structure imaged beneath Africa and Iceland,

Science 286 (1999) 19251928.

[47] A.E. Ibrahim, C.J. Ebinger, J.D. Fairhead, Lithospheric

extension northwest of the central African shear zone in

Sudan from potential field studies, Tectonophysics 255 (1996)

7997.

[48] C.J. Hawkesworth, M.S.M. Mantovani, P.N. Taylor, Z. Palacz,

Coupled crustmantle systems: evidence from the Parana of

South Brazil, Nature 322 (1986) 356359.

[49] D.W. Peate, M.S.M. Mantovani, C.J. Hawkesworth, Geo-

chemical stratigraphy of the Parana CFB: borehole evidence,

Rev. Bras. Geocienc. 18 (2) (1988) 212221.

[50] S. Turner, M. Regelous, S. Kelley, C.J. Hawkesworth, M.S.M.

Mantovani, Magmatism and continental break-up in the South

Atlantic: high precision 40Ar39Ar geochronology, Earth

Planet. Sci. Lett. 1221 (1994) 333348.

[51] A.J. Nunn, J.R. Aires, Gravity anomalies an lexure of the

lithosphere at the Middle Amazon Basin, Brazil, J. Geophys.

Res. 93 (B1) (1988) 415428.

[52] D.S. Weeraratne, D.W. Forsyth, K.M. Fisher, Evidence for an

upper mantle plume beneath the Tanzanian craton from

Rayleigh wave tomography, J. Geophys. Res. 108 (B9)

(2003) 24272446.

[53] C. Cingolani, R. Varela, L. Dalla Salda, J. Bossi, N. Campal,L. Ferrando, D. Pineyro, A. Schipilov, RbSr geochronology

from the Rio de La Plata Craton of Uruguay, South American

Symposium on Isotope Geolog, June 1518, Campos do

Jordao, SP, 1997, pp. 775 (Extended abstract).

[54] L.A. Hartmann, J.A.D. Leite, N.J. McNaughton, J.O.S. Santos,

Deepest exposed crust of BrazilSHRIMP establishes three

events, Geology 27 (1999) 947950.

[55] N. Ussami, E.C. Molina, Flexural modeling of the Neo-

proterozoic Araguaia belt, Central Brazil, J. South Am. Earth

Sci. 12 (1999) 8798.

[56] C.C.G. Tassinari, K.M. Mellito, L.V. Rodrigues, The geo-

chronological map of the Amazonian craton in Brazil, South

American Symposium on Isotope Geolog, June 1518,

Campos do Jordao, SP, 1997, pp. 326329 (Extended

Abstracts).

[57] W. Teixeira, Evolucao tectonotermal proterozoica do craton

de Sao Francisco, com base em interpretacoes geocronolog-

icas, KAr, Simposio sobre o Craton de Sao Francisco

Evolucao Tectonica e Metalogenetica, Salvador, BA, vol. 2,

1993, pp. 1820.

[58] K.D. Nelson, A unified view of craton evolution motivated by

recent deep seismic reflection and refraction result, Geophys.

J. Int. 105 (1991) 2535.

[62] V.A. Ramos, G.I. Vujovich, Alternativas de la evolucion del

borde occidental de Ameica del Sur durante el Proterozoico,

Rev. Bras. Geocienc. 23 (3) (1993) 194200.

[63] V.A. Ramos, M.A.S. Basei, Gondwana, Perigondwanan, and

exotic terranes, South American Symposium on Isotope

Geology, 1518 June, 1997, pp. 250252.

[64] V. Babuska, J. Plomerova, Seismic anisotropy and large scale

fabric of the continental mantle lithosphere, IUGG99,

Birmingham, 2630 July 1999 (Abstracts B144, JSS44/

13-JSS44/B4).

[65] F. Herve, E. Godou, M. Parada, V. Ramos, C. Rapela, C.

Mpdozis, J. Davidson, A general review on the Chilean

Argentine Andes, with emphasis on their early history, in:

J.W.H. Monger, J. Francheteau (Eds.), Circum-pacific Oro-

genic Belts and Evolution of the Pacific Ocean Basin,

Gedynamic Series, vol. 18, 1987, pp. 97113 ILP contribution

n. 0132.

[66] C. Lomnitz, A statistical argument for the existence of a

discontinuity in some subduction zones, J. Geophys. Res. 78

(1973) 25152612.

[67] N. Kuznir, G. Karner, Dependence on flexural rigidity of

continental lithosphere on rheology and temperature, Naure

316 (1985) 138142.

[68] P.V. Zalan, S. Wolff, M.A.M. Astolfi, I.S. Vieira, J.C.J.

Conceicao, V.T. Appi, E.V.S. Neto, J.R. Cerqueira, A.

M.S.M. Mantovani et al. / Earth and Planetary Science Letters 230 (2005) 397412412flexure of the Brazilian Shield beneath the Bolivian Andes,

Earth Planet. Sci. Lett. 75 (1985) 8192.

[60] H. Lyon-Cahen, P. Molnar, Constraints on the structure of the

Himalaya from an analysis of gravity anomalies and a

flexural model of the lithosphere, J. Geophys. Res. 88 (1983)

81718191.

[61] M. Barazangi, B.L. Isaacks, Spatial distribution of earthquakes

and subduction of he Nazca Plate beneath South America,

Geology 4 (1976) 686692.Marques, The Parana basin, Brazil, in: M.W. Leighton, D.R.

Kolata, D.F. Olts, J.J. Eidel (Eds.), Interior Cratonic Basins,

AAPG Mem., vol. 51, 1990, pp. 681708.

[69] A.A. Nyblade, S.W. Robinson, The African superswell,

Geophys. Res. Lett. 21 (9) (1994) 765768.[59] H. Lyon-Cahen, P. Molnar, G. Suarez, Gravity anomaly and

Lithosphere mechanical behavior inferred from tidal gravity anomalies: a comparison of Africa and South AmericaIntroductionMethodologyTe from isostatic and thermo-mechanical analysisTe from tidal gravity anomaly correlation

Tectonic elements of the continental platesAn outline of African tectonicsAnalysis of resultsExtensional tectonic regime structuresSedimentary basinsAfricaSouth America

Large cratonic areasAfricaSouth America

Collisional tectonic structuresAfrica

Subduction tectonic structuresSouth America

Discussions and conclusionsAcknowledgmentsReferences