Imprints of a Proterozoic tectonothermal anomaly below the 1.1 Ga kimberlitic province of Southwest Cuddapah basin, Dharwar craton (Southern India

Geophys. J. Int. (2008) 172, 422–438 doi: 10.1111/j.1365-246X.2007.03623.xG

JITec

toni

csan

dge

ody

nam

ics

Imprints of a Proterozoic tectonothermal anomaly below the 1.1 Gakimberlitic province of Southwest Cuddapah basin, Dharwar craton(Southern India)

D. M. Mall, O. P. Pandey, K. Chandrakala and P. R. ReddyNational Geophysical Research Institute, Uppal Road, Hyderabad 500007, India. E-mail: om [email protected]

Accepted 2007 September 14. Received 2007 September 14; in original form 2006 July 24

S U M M A R YA detailed 2-D modelling of the seismic structure and other geological and geophysical sig-natures across the Cuddapah basin of the southern Indian shield suggests upwarping of high-velocity and high-density layers, which are observed close to the surface below the southwesternpart of the basin. This anomalous feature is constrained by (i) a strong gravity high anomalyof about 55 mGal, (ii) a 100 km wide high conductivity anomaly (resistivity < 100 �-m)extending from surface to a minimum depth of 50 km in the mantle lithosphere and (iii) largescale massive intrusive activity. These features are interpreted to be an expression of a thermalanomaly, which may have acted like a plume during the Proterozoic and could well correspondto a 1.1 Ga kimberlitic activity. Below this region, the thin granitic–gneissic crust is underlainby well-differentiated, high-velocity layers, possibly due to underplating and densification ofmuch of the crust by extruded magma.

Key words: Gravity anomalies and Earth structure; Magnetotelluric; Comtrolled source seis-mology; Dynamics: convection currents, and mantle plumes; Crustal structure; Physics ofmagma and magma bodies.

1 I N T RO D U C T I O N

The Indian subcontinent is a mosaic of several Archean-

Proterozoic cratons (viz. Dharwar, Bastar, Singhbhum, Aravalli and

Bundelkhand), separated by repeatedly rejuvenated grabens, sutures

and mega lineaments since at least 1.5 Ga (Naqvi & Rogers 1987;

Rogers & Callahan 1987). The largest of these, Dharwar craton

(∼3.5 Ga), is made up of three separate geotectonic segments: West-

ern Dharwar Craton (WDC), Eastern Dharwar Craton (EDC) and

Southern Granulite Terrain (SGT) (Fig. 1), which differ in compo-

sition, thickness and crustal seismic structure (Reddy et al. 2000;

Agrawal & Pandey 2004; Vijaya Rao et al. 2007). Among the three

segments, EDC and SGT have remained tectonothermally active

throughout the Proterozoic, whereas WDC has not. The northern

part of EDC is covered by 66 Ma Deccan volcanics (DVP in Fig. 1),

whereas, the southern part is dominated by the intensely deformed

crescent shaped Cuddapah basin. The origin and evolutionary his-

tory of this mid-Proterozoic basin has remained enigmatic in spite

of a large number of geological/geophysical investigations. It is

still being debated whether it was formed due to vertical tectonic

movements or extensional stretching or even a large meteoritic im-

pact (Drury & Holt 1980; Krishna Brahmam & Dutt 1985; Krishna

Brahmam et al. 1986; Anand et al. 2003; etc.). Although the

Cuddapah basin is contemporary to that of Eastern Ghat Mobile

Belt (EGMB) on its east (Fig. 1), their relationship is not yet un-

derstood. Our detailed 2-D geophysical modelling of the crustal

structure of this basin adds a new dimension to its origin and ad-

vocates the role played by a late Proterozoic deep tectonothermal

anomaly in the evolution of this basin.

2 G E O L O G Y A N D T E C T O N I C S

The Cuddapah basin sits in the southern part of the peninsular gneis-

sic complex of EDC. Broadly, this mid-Proterozoic basin can be

divided into three subbasins, Papaghni, Kurnool and Nallamalai

(Fig. 1), which are characterized by distinct differences in lithol-

ogy (Nagaraja Rao et al. 1987). On its east, it is bounded by two

major tectonic features, the Nellore Schist Belt (NSB) and EGMB

(Fig. 1). Its western part is pierced by numerous linear schist belts

(Tsundupalli, Veligallu and Kadiri), which abruptly terminate at its

southern boundary and reappear again in the form of the Jonnagiri

and Gadwal schist belts on its west and north. This basin is intruded

by large-scale volcanic material in the form of lava flows, sills, kim-

berlitic clan rocks and mid-Proterozoic dyke swarms. The initial

phase of extension and volcanism in the basin took place as early

as 1.9 Ga (Anand et al. 2003). A remarkable tectonic feature of this

basin is the presence of a belt of domal upwarp along an axis trend-

ing NW–SE running across middle of the basin (Narayana Swamy

1966). Detailed geotectonic history of this region has been discussed

in detail by Krishna Brahmam et al. (1986).

422 C© 2007 The Authors

Journal compilation C© 2007 RAS

Imprints of a proterozoic tectonothermal 423

Figure 1. Generalized geological map of Cuddapah basin along with location of DSS profile (solid line) and shot points (dotted circle). Numbers 1–5 represents

Kadiri, Veligallu, Tsundupalle, Jonnagiri and Gadwal schist belts, respectively. Location of Kadiri broad-band seismic station (KDR) is shown by star. Various

cratonic blocks are marked as Southern Granulite Terrain (SGT), Western Dharwar Craton (WDC), Eastern Dharwar Craton (EDC), Cuddapah basin (CB),

Nellore Schist Belt (NSB), Deccan Volcanic Province (DVP), Eastern Ghat Mobile Belt (EGMB) and Bastar Craton (BC). Geographic localities are marked as

Chennai (CHA), Hyderabad (HYD), Kadiri (KDR), Bangalore (BAN), Parnapalle (PP), Kavali (K) and Udipi (U).

3 S E I S M I C DATA A N A LY S I S

3.1 DSS field procedure and the recording technique

Kavali-Udipi deep seismic sounding profile in the Indian Peninsu-

lar shield (Fig. 1) shot in 1972, was a collaborative effort between

NGRI (Hyderabad, India) and the USSR. Twenty six shot points,

approximately 30–40 km apart covering a distance of 600 km were

recorded at receiver spacing of 200 m. Shots were fired from drill

holes of 20 to 30 m deep, each loaded with 50 kg explosives of

open cast gelignite (O.C.G.). A minimum charge size of 50 kg and

a maximum of about 1500 kg were used as per varying distances

of recording along the profile. Therefore, a number of holes were

bunched within 10–15 m at a given shot point location for recording

at large distances. Vertical component geophones (10 Hz), in groups

of four, were planted in the ground at every 200 m and connected

to a 12-km-long seismic cable, which were in turn connected to

two 24 channel seismic recording units of type POISK-48 (Russian

made). The instruments provided a near uniform velocity response

within frequency range of 10–30 Hz. Rectangular coordinates of

receiver and shot point locations were determined by carrying out

traversing with reference to different great trigonometrical surveys

(GTS) stations, established by survey of India, along the seismic

line. Most shot points and receiver sites were located with an accu-

racy of 10 and 2 m horizontally and vertically, respectively. The two

seismic stations recorded the shots on photographic paper as moni-

tor records and also on magnetic tapes to be played back at different

filter and gain settings. The tapes were later played back to produce

additional photographic-paper records (termed playback records)

with 15 and 22 Hz high cut filters and appropriate gain settings.

The timing of each seismogram is accurate to within ∼10 ms. The

calibration of the instrument allows the data to be expressed in unit

of group velocity, allowing relative amplitude modelling at a partic-

ular recording position. The data are of good quality, as is evident

from clear first arrivals and strong wide-angle reflections at offsets

of several tens of kilometres, as shown in typical field seismograms

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

424 D. M. Mall et al.

Figure 2. Scan of paper field records: (a) shot point SP165, showing various phases at a distance range between ∼116 and 130 km. The seismogram was played

back at reduced gain to show later arrival reflected phases also and (b) monitor record of SP120 showing first arrivals in the distance range of 47–56 km.

(Fig. 2). We picked the arrival times for the identifiable phases from

the monitor and playback records. Various waves, starting from the

first arrivals from the shallow boundaries to the wide-angle reflec-

tions and the refracted waves from the Moho boundary were well

recorded. Out of 26 shot points along the 600 km long Kavali-Udipi

profile, we have selected six shot points SPs 360, 320, 280, 240, 165

and 120 for the present study, which pertains to the Cuddapah basin

as well as its adjoining region of EDC. Analogue play back records

of these shot points were digitized at a sampling interval of 4 ms

and converted into SEG-Y format for display by standard seismic

software. The composite observed seismograms of these shot points

were generated from the digitized data of individual spreads, which

were scaled to common maximum amplitude. The normalized seis-

mograms were plotted in reduced timescale. Some specimen records

from these shot points covering various distance ranges are shown

in Fig. 3.

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

Imprints of a proterozoic tectonothermal 425

Figure 3. Observed seismograms digitized and plotted in reduced traveltime (T – offset/6.0 km s−1), showing various refracted P1, P2, P3, P4, P6 and reflected

P2, P3, P4, P5 and P6 phases from shot points SP360 (upper panel), SP240 (middle panel) and SP165 (lower panel). All the distances marked on X -axis are

from ‘zero’ position of the model.

3.2 Modelling approach

Initial results of the Kavali-Udipi DSS transect were reported by

Kaila et al. (1979), using a single 1-D average velocity model

for the entire 600 km traverse. Part of this long profile covering

Cuddapah basin has been subsequently analysed by Tewari & Rao

(1987, 1990). Later, Reddy et al. (2004) re-interpreted the entire

EDC segment of the profile in order to compare structural and veloc-

ity variations within the Cuddapah basin and its adjacent peninsular

gneissic complex on its west.

In the present study, wide-angle data are modelled for 2-D

(isotropic) P-wave crustal structure similar to Fernandez Viejo &

Clowes (2003) using the traveltime inversion code described in de-

tail by Zelt & Smith (1992), Zelt & Forsyth (1994) and Zelt (1999).

In this study, an irregular grid of velocity and boundary nodes, be-

tween which linear interpolation is applied, parametrizes the model.

Layer boundaries are specified by an arbitrary number and spacing

of boundary nodes. A smooth layer boundary simulation is also used

in order to avoid scattering and focusing of ray paths, thus yield-

ing more stable inversion results. Within each layer, an arbitrary

number and spacing of upper and lower boundary velocity points

specify the P-wave velocity field. The traveltimes and their partial

derivatives with respect to velocity and depth of boundary nodes

are calculated using an efficient numerical solution (Zelt & Ellis

1988) of the ray tracing equations (Cerveny et al. 1977), coupled

with an automatic determination of ray take-off angles. The forward

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

426 D. M. Mall et al.

response of an initial model of the said description is compared with

the observed data, as usually is done in linearized inversion meth-

ods. The model parameters are then updated using the correction

vectors obtained from the damped least-squares inversion (Zelt &

Smith 1992) of traveltime residuals between the data being inverted

and the forward response. The process is repeated until a satisfac-

tory fit corresponding to a normalized χ2 value of ∼1 is achieved

between the data and the model response.

4 R E S U LT S

4.1 2-D seismic velocity structure

In deriving the subsurface seismic velocity structure, first arrival

refraction data and wide-angle reflection data are used to determine

the shallow and deep crustal horizons. The signal-to-noise ratio of

the first arrivals on most of the sections is good. Since the domi-

nant period of the data lies in the range between 60 and 70 ms, we

assigned uncertainties of ±50, ±65 and ±70 ms to the handpicked

first arrivals and reflections from the second, third and fourth lay-

ers, respectively. This uncertainty assignment is subjective and is

based on 60, 75 and 100 per cent of the dominant cycle of the data

to successively increasing segments of first arrivals. A calculation

was made of velocity required along a portion of ray path and if

the difference in velocity was greater than a reasonable level of lat-

eral heterogeneity (0.05 km s−1), the data were repicked. An initial

model for 2-D inversion is derived using the method of Sain & Kaila

(1996).

The data are modelled in terms of phases that are identifiable and

laterally coherent over many traces. A conventional nomenclature

is used for the first arrivals, marked as P1, P2 and P3 phases rep-

resenting the refraction from layer number 1, 2 and 3, respectively.

The traveltimes of the first segments (P1 and P2 phases) for six shots

(SP360, SP320, SP280, SP240, SP165 and SP120) along the profile

were inverted simultaneously using the 2-D inversion code of Zelt

& Smith (1992), which determined the velocity variation in the first

and second layer, as both these layers are exposed on the profile.

The interface between the second and third layer and the velocity

in the third layer were then calculated by inverting traveltime data

of the second segment (P3 phase) for all shot points together, keeping

the velocity of the first and second layers fixed. Thus the modelling

approach was one of ‘layer stripping’, whereby successively deeper

layers were determined.

The reflection phases, which are convincingly strong and corre-

lated for long distances are taken for modelling. The reflections from

the top of the different crustal layers with velocities of 6.5, 6.6, 7.2,

7.35 and 8.2 km s−1 are designated as P2, P3, P4, P5 and P6. Fol-

lowing the procedures outlined above, a crustal velocity model was

derived for Cuddapah basin. Fig. 4 illustrates subsurface coverage

by refraction and reflection rays and Fig. 5 shows the fit between

observed and calculated traveltimes, respectively, for a number of

phases.

The P2, P2 and P3 phases provide good velocity control only to

a depth of about 10 km. The observed traveltimes show a good

match with those computed for the model. The traveltimes of the

P3 and P2 phases in the central part of the profile were inverted

jointly in order to constrain the interface location of the second

layer and the velocity variation of the third layer. In the eastern

part, the interface of the second layer was determined only from the

traveltime inversion of refraction data (P2 phase) of SPs 120 and 165.

The arrival of P3 phase for SP240 (Fig. 5) within an offset of 10 km

Figure 4. Ray diagrams showing the subsurface coverage by refracted rays

through the final velocity model from various shot points: (a) for shallow

layers including the basement, (b) from upper crustal layer with velocity of

6.5 km s−1, (c) from middle crustal layer with velocity of 6.6 km s−1, (d)

from lower crustal layer with velocity of 7.2 km s−1 and (e) from upper

mantle with velocity of 8.2 km s−1. Subsurface coverage by reflected rays

from various crustal layers from shot points SP360, SP320, SP280, SP240,

SP165 and SP120 are shown in figures f–k, respectively, where every 10th

ray is plotted for the sake of clarity. All the distances marked on X -axis are

from ‘zero’ position of the model.

indicates that a high-velocity layer lies at a shallow depth. The ver-

tical velocity variation within the layer and its base are determined

by simultaneous inversion of wide-angle reflection (P3 phase) data

from the third layer using the data of all shot points. Ray coverage

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

Imprints of a proterozoic tectonothermal 427

Figure 4. (Continued.)

for the reflection data is shown in Figs 4(f–k), and the respective

theoretical traveltimes (represented by lines) are superimposed on

the observed traveltimes in Fig. 5; differences range from 0.05 to

0.15 s. The intracrustal phases, P3 and P4 (Figs 4a–e) provide con-

straints on velocities in the upper and middle crust.

Phase P6 provided good coverage along the Moho. However, Pn

coverage along the profile was limited and provided velocities for

the mantle immediately below the Moho. Overall, the observed and

computed traveltimes for both phases have matched well with each

other. For phases from the mantle, interpretation was carried out

through forward modelling, keeping the geometry of the reflectors

as simple as possible (planar, dipping when needed). Rays have

been traced through the model to almost all observations. They

match the observed data with an overall rms traveltime residual of

65 ms, corresponding to a normalized χ2 value of almost one. The

number of rays traced through the model, rms traveltime residual

and normalized χ 2 value for various phases corresponding to all

the shot points taken together are shown in Table 1 to indicate the

goodness of fit. These values of χ2 assume that the arrival times

have been picked and identified correctly and assigned appropriate

uncertainties.

Fig. 6 shows the model parameter nodes used in the modelling. It

is to be noted that adding more nodes to a model typically reduces the

traveltime residual, but at the same time it concurrently decreases the

overall resolution of model parameters. Therefore, a velocity model

with fewer parameters defining a minimum structure is generally

sought, as it is most likely to be free from modelling artefacts, which

are not required by the data and do not accurately represent the true

structure (O’Leary et al. 1995). The final crustal velocity model is

presented in Fig. 7. The numbers given in the model indicate the

averaged interval velocity (km s−1) for various parts of the derived

model. The structure of the shallow part of the model to a depth

of 10 km is displayed in Fig. 4(a). The model was derived based

on its ability to trace rays through the model to all observation

points, and also on the trade-off between achieving a sufficiently

small traveltime residual of the order of data uncertainties and an

adequately high parameter resolution.

Using ray theory code (Zelt & Smith 1992) synthetic seismo-

grams were generated corresponding to the final velocity model. As

most of the traveltime modelling, this model has also been checked

by comparing the amplitudes of observed phases to those of the

theoretical ones for each shot point. This has been displayed for

several wide-angle phases from SP360, SP 240, SP 165 and SP 120

(Figs 8a–d). Phase P5 reflection corresponding to the velocity of

7.35 km s−1 observed in seismograms are very weak and are smeared

in background noise as the velocity contrast across this boundary is

low.

4.1.1 Modelling statistics and parameter uncertainties

The reliability of the velocity model (Fig. 7) was examined in terms

of resolutions and uncertainties of estimated model parameters from

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

428 D. M. Mall et al.

Figure 5. Comparison of observed and calculated traveltimes for refracted and reflected phases, plotted in reduced traveltime (T-D/6.0 km s−1) for all six shot

points SP360, SP320, SP280, SP240, SP165 and SP120 along the profile. The calculated traveltimes are indicated by continuous lines and the observed data

by vertical bars, the height of which represents the pick uncertainties. All the distances marked on X -axis are from ‘zero’ position of the model.

Table 1. Statistics of modelling results for each phase.

Average (Chi square)

No of uncertainty rms normalized

Phase observations (ms) misfit χ2

P1 130 50 0.059 0.825

P2 2364 50 0.053 1.139

P3 878 65 0.072 1.215

P4 483 65 0.068 1.093

P5 581 65 0.069 1.138

P6 136 65 0.084 1.696

P2 95 100 0.098 0.977

P3 240 100 0.106 1.120

P4 307 100 0.107 1.226

P5 276 100 0.112 1.459

P6 191 100 0.128 1.638

the diagonal elements of resolution and covariance matrices, respec-

tively (Zelt & Smith, 1992). Generally resolution values range be-

tween 0 and 1, and depend on relative number of rays sampling in

each model parameter and azimuth. The model is finalized having

resolution values greater than 0.5 (Fig. 6) except at a few boundary

nodes at the edge and in the deeper part of the models due to the

insufficient ray coverage. Table 1 present statistics associated with

the inversion of traveltimes. A qualitative evaluation of Table 1 and

ray paths for the complete shot-receiver distribution (Figs 4 and 5),

suggests that the resolution is best in the upper crust, and poorer

in the middle and lower crust. This indicates that the model is well

resolved and reliable as the resolution values of associated model

parameters are generally greater than 0.5 (Zelt & Smith 1992).

The uncertainties in model parameters under linear assumptions

are very small (<0.1 km) and are considered as statistical un-

certainty. To account for the non-linearity of traveltime inversion

and to provide significant insight into the model constraints, a uni-

parameter uncertainty test (Zelt & Smith 1992; Zelt, 1999) is per-

formed. Applying this type of error analysis to all velocity and

boundary nodes could take longer than the time required in obtain-

ing the final model. Therefore, we have performed both positive

and negative parameter perturbation tests on one representative ve-

locity and boundary node for each layer. To obtain the absolute

uncertainty, we have perturbed its value from that in the final model

and hold it fixed while inverting the observed data involving all

other parameters that were determined at the same time as the per-

turbed parameter during the inversion for the final model. Then, we

have increased the perturbation until the final model so obtained

is unable to fit the observed data. The maximum perturbation of

the parameter that allows a comparable fit to the observed data be-

comes an estimate of its absolute uncertainty as described by Zelt &

Smith (1992) and Zelt (1999). Tests on absolute velocity uncertain-

ties for crustal layers show that the values lie between +0.15 and

+0.25 km s−1 along the DSS profile. Further, the absolute uncer-

tainties of boundary depths for various crustal layers lie in the range

between +0.6 and +1.0 km. The same damping factor of 1.0 is used

for all of the above tests.

Although, for the final model (Fig. 7), amplitude information

contributed to the interpretation, but these constraints cannot be

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

Imprints of a proterozoic tectonothermal 429

Figure 6. Velocity (circles) and boundary (squares) nodes used in the present study. The relative size of the symbols indicates the resolution of the particular

model parameters which range between 0.1 and 0.99.

Figure 7. Final velocity model derived from the present study. Numbers in the model are average P-wave velocity (in km s−1) at various depths. Inverted

triangles at top of the model show the location of shot points. Projected location of Kadiri (broad-band seismic site) together with Parnapalle is also shown

along the profile.

factored into the statistics. Further, the principal sources of error

not included in the statistics are the possibilities of misidentification

of phases and the effects of 3-D structures when applying a 2-D

analytical approach.

4.2 Broad-band seismic velocity structure

Besides DSS studies, EDC is also covered by a dense broad-band

seismic network. Using this network, Gupta & Rai (2005) derived

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

430 D. M. Mall et al.

Figure 8. Comparison of observed and synthetic seismograms of shot points SP360 (a), SP240 recorded towards west (b), SP240 recorded towards east (c),

SP165 (d) and SP120 (e) showing the degree of match. The observed seismogram of shot points showing the respective wide-angle phases P2, P3, P4 and P6

from the tops of the 6.45, 6.60, 7.20 and 8.2 km s−1 velocity layers, respectively (marked by solid arrow) are shown in upper panels. Seismograms are plotted

with a reduction velocity of 6.0 km s−1. 2-D synthetic seismogram generated corresponding to the final velocity model (Fig. 7) for above mentioned shot points

showing wide-angle reflected phases are plotted in the lower panels. Corresponding ray diagram can be seen in Fig. 4.

high-resolution S velocity structure and thickness of the crust below

EDC using a receiver function technique. One of the network stations

Kadiri (Fig. 1) is located at about 40 km southeast of Parnapalle.

They too found a high shear velocity of 3.75 km s−1 (equivalent to

Vp ∼ 6.5 km s−1) at a depth of about 4 km only. It is intriguing to note

that below 4 km depth, seismic velocities obtained near Parnapalle

as well as KDR (Kadiri) are higher than those obtained below mafic

granulitic terrain (Reddy et al. 2003), where recorded metamorphic

pressures in some exposed sections reach as high as 10–11 kb. A

comparison of crustal velocity (Vp) as obtained in various depth

intervals with the corresponding Vs and electrical conductivity is

shown in Table 2. This table also includes the estimated density and

expected rock types in these depth intervals.

4.3 Gravity and magnetic studies

The gravity map of the SW Cuddapah basin, bounded by the latitudes

14◦–16◦N and longitudes 77.5◦–79◦E, is shown in Fig. 9. Within

this region, the Bouguer gravity anomalies are predominantly neg-

ative, varying between −50 and −110 mGal. In general, the gravity

anomalies decrease towards east but that cannot be explained by

the sediments. The observed undulations in the deep crustal lay-

ers are well reflected in the broad regional feature observed on the

gravity profile. In order to understand the nature and thickness of

the crust, the observed gravity field has been interpreted by many

workers (Glennie 1951; Qureshy et al. 1968; Grant 1983; Krishna

Brahmam & Dutt 1985; Rambabu 1993), who put forward various

explanations for the observed features, including meteoritic impact.

Since the seismically derived velocity structure obtained below

southwestern part of Cuddapah basin (Fig. 7) corresponds to a

large positive gravity anomaly of about 55 mGal centred around

Gandikota, Tadipatri and Parnapalle (Fig. 9), we make an attempt

to quantitatively model the gravity field along the seismic profile

which is shown in Fig. 9. For this purpose, the Bouguer anomaly

map has been digitized at an equal interval of 5 km along the line

by linearly interpolating between the contours, and taking care to

ensure that essential parts of the anomaly are included. To avoid the

edge effect of the 2-D density model, the profiles were extended to

about 1000 km on both sides.

To obtain 2-D crustal density structure, the concept of interac-

tive forward gravity modelling, with a-priori seismic information

has been used. The geometry of the initial density model was con-

strained by the crustal seismic section (Fig. 7). The velocity–density

relationship of Christensen & Mooney (1995) was used to obtain

the average density within the individual polygons. Finally, we used

the GMSYS (2002) software to model the observed gravity field.

The computed gravity response using the derived density values

could match with the observed gravity with a few perturbations to

the initial density model (Fig. 9). The perturbed densities (ρ) and

depths to the various interfaces in individual polygons fall within

±0.05 g cm−3 and ±0.5 km, respectively, which are well within

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

Imprints of a proterozoic tectonothermal 431

Figure 8. (Continued.)

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

432 D. M. Mall et al.

Figure 8. (Continued.)

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

Imprints of a proterozoic tectonothermal 433

Table 2. Comparison of Vp as obtained at the projected location of Kadiri on the seismic profile (Fig. 7) with that of (i) shear velocity

(Vs) derived from broad-band seismics at Kadiri and (ii) electrical conductivity. Estimated corresponding density and expected rock

types at various depths are also included.

Layer Deptha Vpa Vs

b Densityc Resistivityd Rock

no. (km) (km s−1) (km s−1) (g cm−3) (�-m) type

2 0–2 5.90–6.00 3.10 2.68–2.69 180–320 Granites/volcanics

3 2–12 6.50– 6.60 3.50–3.95 2.88–2.91 <10 Basalt/volcanics/intermediate granulites

4 12–26 6.60–6.8 3.63–3.95 2.92–2.98 <10–56 Volcanics/intermediate granulites

5 26–37 7.2–7.3 3.63–4.10 3.12–3.15 32–100 Underplated volcanic magma/mafic granulites

6 37–43 7.35 4.10–4.75 3.15 56–100 Underplated volcanic magma/mafic granulites

7 43–50 8.2 4.75 3.40 56–100 Ultramafics

aAt the projected location of Kadiri on the seismic profile (Fig. 7).bBroad-band seismics (Gupta & Rai 2005).cEstimated density based on Christensen & Mooney (1995).dNaganjaneyulu & Harinarayana (2004).

the intrinsic uncertainty of the density estimates as well as derived

crustal seismic model. The observed gravity field along the seis-

mic profile is well accounted for by the derived crustal structure

with a thick anomalous density layer (3.07–3.16 g cm−3) above the

Moho, which has been attributed to extrusion/underplating by a ba-

sic magma. The presence of high-density mafic material at shallow

depths is manifested in the form of an intense magnetic anomaly with

amplitude of 580 nT (Babu Rao et al. 1987). Earlier, a long linear

greenstone schist belt was suspected to continue below the anoma-

lous gravity and magnetic region (Rama Rao et al. 1993).

4.4 Magnetotelluric (MT) studies

The depicted high gravity/high velocity anomalous feature is also

reflected in the wide-band magnetotelluric study which was carried

out along the same DSS transect (Naganjaneyulu & Harinarayana

2004). In their study, MT data were acquired in the frequency range

of 8192–0.001 Hz using a GMS-05 unit of M/s METRONIX with a

station interval of 12–18 km. The acquired MT data were subjected

to robust processing, decomposition and static shift correction be-

fore deriving the 2-D model, using the ‘Rapid Relaxation Inver-

sion (RRI)’ approach (Smith & Booker 1991). The modelling result

(Fig. 10) suggests the presence of an anomalous feature, associated

with very high conductivity (resistivity less than 100 �-m), below

the southwestern part of Cuddapah basin. This structure is about

100 km wide and extends to a depth of at least 50 km the source

of which could be of basic/ultra basic in nature (Naganjaneyulu &

Harinarayana, 2004).

Although no simple relationship exists between rock types and

their resistivities, mafic and ultramfic rocks/magmas are invari-

ably associated with high conductivity anomalies, based on labora-

tory and field measurements, compared to acidic rocks like granite

(Beblo 1982; Simpson & Bahr 2005). As expected, at subsurface

depth, this feature is bounded on both sides by very high resistiv-

ities, corresponding to granitic–gneissic upper crust. For example,

towards the western side of the anomalous conductive body, the re-

sistivity of the crust reaches as high as 10 000–30 000 �-m from

the surface to a depth of 8–10 km.

5 D I S C U S S I O N S

The large positive oval shaped gravity anomaly with a magnitude of

about 55 mGal, surrounded by a trough of gravity lows on all sides

(Fig. 9) remains one of the most striking features of the Cuddapah

basin. The underlying feature associated with this gravity anomaly

appears to have played an important role in the evolution of this

basin, which previously has been attributed to several factors such

as the thickening of a basaltic layer, upwarping of the crustal column,

intrusion of feeder dykes/ sills and existence of a high density–high

velocity mafic body (Krishna Brahmam et al. 1986 and references

therein), or even a meteoritic impact (Krishna Brahmam & Dutt

1985). We feel, our integrated multidisciplinary geophysical studies

have provided some definitive clues to the structure, geometry and

petrophysical properties of the crust underneath Cuddapah basin,

which has been debated for so long.

From near the surface to a depth of about 40 km, observed veloci-

ties are very high (Vp ∼ 6.50–7.35 km s−1) beneath the southwestern

part of the basin. Near Parnapalle (Fig. 7), this high velocity layer is

found at the shallowest level of 0.4 km and at much deeper depths

on either side of it along the profile. At Kadiri (KDR), situated at

about 40 km southeast of Parnapalle (Fig. 1), such velocities are

found at the depth of 4 km (Gupta & Rai 2005). Similar disposi-

tion is seen in other velocity and density layers too (Figs 7 and 9).

The delineated high seismic velocity at Parnapalle coincides almost

one to one with the high-conductivity and high-gravity anomalies

(Figs 9 and 10). The cause of such high conductivity anomaly

has been attributed to basic/ultra basic nature of the source

(Naganjaneyulu & Harinarayana 2004). Therefore, in all likeli-

hood, these anomalies correspond to the same feature. Considering

the shape, size, depth extent and structural disposition of under-

lying structures (Fig. 7), this appears to be a case of mantle up-

welling related to a long sustained deep tectonothermal anomaly,

which may have acted like a plume. In view of the above, con-

tinuation of long linear greenstone schist belt below the anoma-

lous gravity and magnetic region (Rama Rao et al. 1993) appears

untenable.

Plumes do not necessarily have to be very large; in compara-

tively smaller volcanic regions, they are only 100–200 km wide,

containing a column of hot rocks rising from great depths in the

mantle (Ritter et al. 2001). In normal cases, such plumes with tem-

peratures 100–200 ◦C in excess of the background temperature, can

cause regional uplift coupled with considerable extension (White &

McKenzie 1989), thereby leading to massive injection of volcanic

material in the lower crust. Under the influence of excess temper-

ature, magmatic material associated with a rising plume diverges

laterally when it encounters a strong viscosity gradient (Meissner

1986). Cooling of such magma results in the formation of a homo-

geneous layer containing mafic residues and stratified melt products

at the base of crust. Such layers could be as thick as 10–15 km,

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

434 D. M. Mall et al.

Figure 9. Bouguer anomaly map of the Cuddapah basin (modified after Vasanthi & Mallick 2005) along with the location of DSS profile shown as thick

line (upper panel). Locations of Parnapalle (PP), Tadpatri (TD) and Gandikota (GT) are also shown. Comparison of observed Bouguer gravity (crosses) and

computed gravity response (solid lines), using the seismically derived density model, is shown in the middle panel. The density values (g cm−3) within each

trapezoid are shown in the lower panel.

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

Imprints of a proterozoic tectonothermal 435

Figure 10. 2-D Geoelectric depth section along DSS profile below Cuddapah basin and adjoining areas (Naganjaneyulu & Harinarayana 2004).

typically characterized by a density around 3.10 ± 0.10 g cm−3 and

a P-wave velocity of ∼7.30 ± 0.20 km s−1.

Inferences drawn from the present study, would indicate that simi-

lar phenomena may have been active below Cuddapah basin through

the entire mid-Proterozoic period, perhaps initially by lithospheric

stretching and later on by deep seated tectonothermally induced

plume activity. In this region, magmatic activity seems to have be-

gun as early as 1.9–1.8 Ga. and continued episodically to about

1.0 Ga. (Mallikarjuna Rao et al. 1995). This activity was intense

during the later half of this period, with a strong magmatic pulse at

1.1 Ga. The occurrence of (i) volcanic lava flows and sills as old as

1.9 Ga (Anand et al. 2003), (ii) 1.6 to 1.0 Ga massive dyke swarms

and (iii) widespread 1.1 Ga diamondiferous kimberlitic intrusions

(Anil Kumar et al. 1993) in an area of about 500 km in radius (Fig.

11), bear the testimony to such activity. Due to episodic magmatic

activity, the entire crustal column below Cuddapah basin has be-

come more mafic and dense on account of forceful injection and

underplating/accretion of magma at the base of the crust, thereby

resulting in significantly higher crustal velocities, even higher than

those encountered in the exposed mafic granulitic terrain of southern

India, as mentioned earlier. 40Ar–39Ar ages would confirm periodic

crustal dilation and thermal resetting at different periods beneath

this region (Mallikarjuna Rao et al. 1995).

The underplated layer below the southwestern region appears to

be as thick as 15–20 km within which the velocities are in the range

of 7.10–7.35 km s−1, density 3.07 to 3.16 g cm−3 (layers 5 and 6;

Figs 7 and 9). This underplated layer has made the crust thicker

(40–44 km) compared to 33–36 km in the surrounding parts of

EDC (Gupta & Rai 2005; Rai et al. 2003) upon which Cuddapah

basin sits. Such extensive underplating and crustal thickening would

indicate the possibility of direct asthenospheric interaction with the

base of the crust.

Apparently, such accretion is quite common below intra plate hot

spots, active and passive volcanic areas, as well as rift zones all

over the world. Examples include: western margin and Rajmahal

traps, India (Singh & Mall 1998; Mall et al. 1999; Singh

et al. 2004), Costa Rican Isthmus, Caribbean Plateau (Sallares

et al. 2001), basin and Range province of northwestern Nevada,

USA (Catchings & Mooney 1991), Lofoten volcanic margin, north-

ern Norway (Mjelde et al. 1996), Kenya Rift (Green et al. 1991)

and many passive volcanic margins (White & McKenzie 1989).

Since the kimberlitic intrusive activity is widespread and well dated

(Anil Kumar et al. 2003), the proposed thermal plume could be

a 1.1 Ga kimberlitic one, which is also close to the accretion

period of Rodinia (Condie 2001). The postulated mechanism is

compatible with the composition of the crust and its gravity field

response.

Further, in view of absence of either shocked quartz or an Iridium

anomaly, it would not be possible to justify a meteoritic impact ex-

planation (Krishna Brahmam & Dutt 1985) for the observations seen

here. More importantly, the feature fails several criteria necessary

to define impact structures. For example, impact structures are con-

spicuously associated with circular gravity and magnetic lows (as

opposed to the highs observed here) caused by buried post impact

sediments and ejecta and the presence of brecciated crystalline rocks

over fractured basement. Impact structures are also characterized by

reduced seismic velocities (Pilkington & Grieve 1992) as opposed

to the high velocities seen here. Obviously, none of these conditions

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

436 D. M. Mall et al.

Figure 11. (a) Relationship between kimberlite, lamproites and diamond occurrences (Vasanthi & Mallick 2005) in and around Cuddapah basin and location

of tectonothermal anomaly. 1: kimberlites, 2: lamproite, 3: diamond finds, 4: Eastern Ghat Mobile Belt (EGMB), 5: schist belt, 6: location of proposed thermal

anomaly, 7: DSS profile, 8: basic intrusives, lavas. (b) Distribution of mafic dyke swarms surrounding the SW of Cuddapah basin (Murthy 1987) within a radius

of about 500 km.

are met, if we take a close look at the observed surface/subsurface

structural and geophysical features (Figs 7, 9 and 10). We, therefore

conclude that our preferred interpretation for the feature below the

southwestern part of the Cuddapah basin is associated with a deep

seated late Proterozoic mantle thermal anomaly possibly resulting

from a plume.

A C K N O W L E D G M E N T S

We are extremely thankful to Drs Valenti Sallares, Andrew Gorman

and Jun Korenaga for their positive comments, which have im-

mensely helped to improve the manuscript. We are also thankful to

Drs D. Sarkar, U. Raval and H.C. Tewari for useful discussions and

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

Imprints of a proterozoic tectonothermal 437

Mr O. Prasad Rao for drafting figures. The permission accorded by

the Director, National Geophysical Research Institute, Hyderabad

to publish this work is gratefully acknowledged. Dr P.R. Reddy and

K. Chandrakala thank CSIR, New Delhi for Emeritus scientist and

SRF positions, respectively.

R E F E R E N C E S

Agrawal, P.K. & Pandey, O.P., 2004. Unusual lithospheric structure and

evolutionary pattern of the cratonic segment of the south Indian shield,

Earth Planets Space, 56, 139–150.

Anand, M., Gibson, S.A., Subba Rao, K.V., Kelley, S.P. & Dickin, A.P., 2003.

Early Proterozoic Melt Generation Processes beneath the Intra-cratonic

Cuddapah basin, Southern India, J. Petrology, 44, 2139–2171.

Babu Rao, V., Atchuta Rao, D., Rama Rao, Ch., Sharma, B.S.P., Bhaskar

Rao, D.S., Veeraswamy, K. & Sharma, M.R.L., 1987. Some salient re-

sults of interpretation of aeromagnetic data over Cuddapah basin and ad-

joining terrain, South India, in Purana Basins of Peninsular India, Vol.6,

pp. 295–312, Geol, Soc. India Mem.

Beblo, M., 1982. Electrical conductivity (resistivity) of minerals and rocks

at ordinary temperature and pressures, in Landolt-Bornstein, Vol. 1,

pp. 239–253, ed. Angenheister, G., Springer-Verlag, Berlin-Heidelberg.

Catchings, R.D. & Mooney, W.D., 1991. Basin and range crustal and upper

mantle structure, northwest to central Nevada, J. geophys. Res, 96(B4),

6247–6267.

Cerveny, V., Molotkov, I.A. & Psencik, I., 1977. Ray Method in Seismology,p. 214, University of Karlova, Prague.

Christensen, N.I. & Mooney, W.D., 1995. Seismic velocity structure and

composition of the continental crust: A global view, J. geophys. Res.,100, 9761–9788.

Condie, K.C. 2001. Continental growth during formation of Rodinia at 1.35–

0.90 Ga, Gond. Res., 4, 5–16.

Drury, S.A. & Holt, R.W., 1980. The tectonic framework of the South Indian

craton: a reconnaissance involving LANDSAT imagery, Tectonophysics,65, T1–T5.

Fernandez Viejo, G. & Clowes, R.M., 2003. Lithospheric structure beneath

the Archaean Slave Province and Proterozoic Wopmay orogen, north-

western Canada, a LITHOPROBE refraction/wide-angle refraction sur-

vey, Geophys. J. Int., 153, 1–19.

Glennie, E.A., 1951. The Cuddapah basin in India and crustal warping, Mon.Nat. R. Astron. Soc. Geophys. Suppl., 6, 168–176.

GMSYS, 2002. Two and half D gravity magnetic modeling system for OASIS

MontajTM Geosoft Canada

Grant, F.S., 1983. Results of preliminary interpretation studies, Report

on ‘Visit to NGRI, Hyderabad, India’ Project IND/79/047. Sept-Oct.,

pp. 5–17 (Restricted).

Green, W.V., Ulrich, Achauer, & Meyer, R.P. 1991. A three-dimensional

seismic image of the crust and upper mantle beneath the Kenya rift, Nature,354, 199–203.

Gupta, S. & Rai, S.S., 2005. Structure and evolution of south Indian crust

using teleseismic waveform modeling, Him. Geol., 26, 109–123.

Kaila K.L. et al., 1979. Crustal structure along Kavali-Udipi profle in the

Indian peninsular shield from deep seismic sounding, J. Geol. Soc. India,20, 307–333.

Krishna Brahmam, N. & Dutt, N.V.B.S., 1985. Possible relationship be-

tween initiation of Papaghni basin and meteoritic impact, in Workshopon “Purana Basins (middle to later Proterozoic) of Peninsular India”,Hyderabad, Dec. 29–31, pp. 48–49.

Krishna Brahmam, N., Sharma, J.R.K., Aravamadhu, P.S. & Subba

Rao, D.V., 1986. Explanatory Broucher on Bouguer Gravity AnomalyMap (NGRI/GPH-6) of Cuddapah Basin (India), NGRI, Hyderabad,

India.

Kumar, A., Padma Kumari, V.M., Dayal, A.M., Murthy, D.S.N. &

Gopalan, K., 1993. Rb-Sr ages of Proterozoic kimberlites of India, ev-

idence for Contemporaneous emplacement, Precamb. Res., 62, 227–

237.

Mall, D.M., Rao, V.K. & Reddy, P.R. 1999. Deep sub-crustal features in the

Bengal basin: Seismic signature for plume activity, Geophys. Res. Lett.,26, 2545–2548.

Mallikarjuna Rao, J., Bhattacharji, S., Rao, M.N. & Hermes, O.D., 1995.40Ar- 39Ar ages and geochemical characteristics of the dolerite dykes

around Proterozoic Cuddapah basin, South India, Mem. Geol. Soc. Ind.,33, 307–328.

Meissner, R., 1986. The continental crust: A geophysical approach, p. 426,

Academic Press, New York.

Mjelde R. et al., 1996. The continent/ocean transition of the Lofoten volcanic

margin N. Norway, J. Geodyn., 22, 189–206.

Murthy, N.G.K., 1987. Mafic dyke swarms of the Indian shield, in Maficdyke swarms, Vol.34, pp. 393–400, eds Halls, H.C. & Fahrig, W.F., Geol.

Assn. Canada, Special paper.

Naganjaneyulu, K. & Harinarayana, T., 2004. Deep crustal electrical signa-

tures of Eastern Dharwar Craton, India, Gond. Res., 7, 951–960.

Nagaraja Rao, B.K., Rakurkar, S.T., Ramalingaswamy, G. & Ravindra Babu,

B., 1987. Purana Basins of Peninsular India, Vol.6, pp. 33–86, ed. Rad-

hakrishna, B.P., Geol. Soc. Ind., Mem.Naqvi, S.M., & Rogers, J.J.W., 1987. Precambrian Geology of India, p. 223,

Oxford University Press, New York.

Narayana Swamy, S., 1966. Tectonics of the Cuddapah basin, J. Geol. Soc.Ind., 7, 33–50.

O’Leary, D.M., Ellis, R.M., Stephenson, R.A., Lane, L.S. & Zelt, C.A.,

1995. Crustal structure of the northern Yukon and Mackenzie delta, north-

western Canada, J. geophys. Res., 100(B7), 9905–9920.

Pilkington, M. & Grieve, R.A.F., 1992. The geophysical signature of terres-

trial impact craters, Rev. Geophys., 30, 161–181.

Qureshy, M.N., Krishna Brahmam, N., Aravamadhu, P.S. & Naqvi, S.M.,

1968. Role of granitic intrusions in reducing the density of the crust and

other related problems as illustrated from a gravity study of the Cuddapah

basin, India, Proc. Roy. Soc. Lond., A304, 449–464.

Rai, S.S., Priestly, K., Suryaprakasam, K., Srinagesh, D., Gaur, V.K. & Du,

Z., 2003. Crustal shear velocity structure of the South India shield, J.geophys. Res., 108(B2), 2088, doi:10.1029/2002JB001776.

Rama Rao, Ch., Bhaskara Rao, D.S. & Atchuta Rao, D., 1993. Delineation

of possible extension of schist belt complex in the southwestern part of the

Cuddapah basin from regional aeromagnetic anomaly map—a directional

spectral approach, Proceedings of the 28th Annual convention and theSeminar on “Geophysics for rural development”, Ind. Geophy. Un., 99–

103.

Ram Babu, H.V., 1993. Basement structure of the Cuddapah basin from

gravity anomalies, Tectonophysics, 223, 411–422.

Reddy, P.R., Chandrakala, K. & Sridhar, A.R. 2000. Crustal velocity

structure of the Dharwar craton, India, J. Geol. Soc. Ind., 55, 381–

386.

Reddy, P.R., Rajendraprasad, B.R., Vijaya Rao, V., Sain, K., Prasad Rao,

P., Prakash, Khare & Reddy, M.S., 2003. Deep seismic reflection

and refraction/wide-angle reflection studies along Kuppam-Palani tran-

sect in the southern Granulite Terrain, Mem. Geol. Soc. Ind., 50, 79–

106.

Reddy, P.R., Chandrakala, K., Prasad, A.S.S.S.R.S. & Rama Rao, Ch., 2004.

Lateral and vertical crustal velocity and density variations in the South-

western Cuddapah basin and adjoining Eastern Dharwar Craton, Curr.Sci., 87, 1607–1614.

Ritter, J.R.R., Jordan, M., Christensen, U.R. & Achauer, U., 2001. A mantle

plume below the Eifel volcanic fields, Germany, Earth Planet, Sci. Lett.,186, 7–14.

Rogers, J.J.W. & Callahan, E.J., 1987. Radioactivity, heat flow and rifting of

the Indian continental crust, J. Geol., 95, 829–836.

Sain, K. & Kaila, K.L., 1996. Ambiguity in the solution of the velocity

inversion problem and a solution by joint inversion of seismic refraction

and wide angle reflection times, Geophys. J. Int., 124, 215–227.

Sallares, V., Danobeitia, J.J. & Flueh, E., 2001. Lithospheric structure of

the Costa Rican Isthmus: effects of subduction zone magmatism on an

oceanic plateau, J. geophys. Res., 106, 621–643.

Simpson, F. & Bahr, K., 2005. Practical Magnetotellurics, p. 254, Cambridge

University Press, Cambridge, UK.

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS

438 D. M. Mall et al.

Singh, A.P. & Mall, D.M., 1998. Crustal accretion beneath the Koyna coastal

region (India) and late Cretaceous geodynamics, Tectonophysics, 290,285–297.

Singh, A.P., Kumar, Niraj & Singh, Bijendra, 2004. Magmatic underplating

beneath the Rajmahal Traps: gravity signature and derived 3-D configu-

ration, Proc. Ind. Acad. Soc. (Earth Planet. Sci.), 113, 759–769.

Smith, J.T. & Booker, J.R., 1991. Rapid inversion of two and three-

dimensional magnetotelluric data, J. geophys. Res., 96, 3905–3922.

Tewari, H.C. & Rao, V.K., 1987. A high velocity intrusive body in the upper

crust in the south-western Cuddapah basin as delineated by deep seismic

soundings and gravity modeling, in Purana Basins of Peninsular India,Vol.6. pp. 349–356, ed. Radhakrishna, B.P., Geol. Soc. Ind. Mem.

Tewari, H.C. & Rao, V.K., 1990. Crustal velocity model in the eastern part

of the Indian Peninsular shield, J. Geol. Soc. Ind., 36, 475–483.

Vasanthi, A. & Mallick, K., 2005. Bouguer gravity modeling of central

Cuddapah basin, J. Geol. Soc. Ind., 66, 171–184.

Vijaya Rao, V., Sain, K. & Prasad, B.R., 2007. Dipping Moho in the southern

part of eastern Dharwar craton, India, as revealed by the coincident seismic

reflection and refraction study, Curr. Sci., 93, 330–336.

White, R.S. & McKenzie, D.P. 1989. Magmatism at rift zones: the generation

of volcanic continental margins and flood basalts, J. geophys. Res., 94,7685–7729.

Zelt, C.A., 1999. Modeling strategies and model assessment for wide-angle

seismic travel time data, Geophys. J. Int., 139, 183–204.

Zelt, C.A. & Ellis, R.M., 1988. Practical and efficient ray tracing in two-

dimensional media for rapid travel time and amplitude forward modeling,

Can. J. Explor. Geophys., 24, 16–31.

Zelt, C.A. & Smith, R.B., 1992. Seismic travel time inversion for 2-D crustal

velocity structure, Geophys. J. Int., 108, 16–34,

Zelt, C.A. & Forsyth, D.A., 1994. Modeling wide-angle seismic data for

crustal structure: Southern Grenville province, J. geophys. Res., 99(B6),

11687–11704.

C© 2007 The Authors, GJI, 172, 422–438

Journal compilation C© 2007 RAS