Geochemical stratigraphy of Deccan flood basalts of the Bijasan Ghat

section, Satpura Range, India

H.C. Shetha, J.J. Mahoneyb, D. Chandrasekharama,*

aDepartment of Earth Sciences, Indian Institute of Technology, Powai, Bombay 400 076, IndiabSchool of Ocean and Earth Science and Technology, University of Hawaii, Honolulu, HI 96822, USA

Received 26 September 2002; revised 13 April 2003; accepted 15 April 2003

Abstract

The ,240-m-thick Bijasan Ghat section exposes Deccan basalt flows and dykes that have major and trace element compositions

suggestive of variable amounts of contamination by continental material. Most flows and a dyke are chemically similar to lavas of the

Poladpur Formation in the southwestern Deccan. One flow and one dyke are similar to lavas of the Bushe Formation in the southwest, and one

dyke is similar to lavas of the Mahabaleshwar Formation in the southwest. The type-sections of these formations are located ,350 km to the

south of Bijasan Ghat, but many Poladpur-type lavas, a few Mahabaleshwar-like lavas and dykes, and a few broadly Bushe-like lavas and

dykes have recently been identified in the Satpura region. Lavas resembling the thick Ambenali Fm. of the southwest are absent from Bijasan

Ghat and nearby areas. Our data further extend the known outcrop area of Poladpur-like flows (estimated at $3 £ 105 km2), one of the most

widespread Deccan magma types. Seven ‘giant plagioclase basalt’ flows encountered in seven different sections in this region cannot

unambiguously be correlated at present, either physically or chemically, and therefore are of little use in deciphering the complicated

stratigraphic and structural make-up of the regional lava pile.

q 2003 Elsevier Ltd. All rights reserved.

Keywords: Deccan; Flood basalt; Geochemistry; Petrogenesis; Satpura; Tapi; Stratigraphy

1. Introduction

Considerable advances have been made over the last two

decades in understanding the regional lava stratigraphy of

the Deccan flood basalt province of India (see Sen, 2001 for

a recent review). The province has a present-day area of

500,000 km2 (e.g. Wadia, 1975) on land and a large

additional area offshore of western India; an unknown, but

probably large amount of the province’s original extent has

been lost to erosion since the formation of the province

,65 million years ago (Ma). The Deccan lava pile is

particularly well exposed in the southwestern part of the

province (Fig. 1) in the Western Ghats range, and the

stratigraphic framework of this region is now known quite

well from extensive field, geochemical (including isotopic)

and palaeomagnetic work (e.g. Cox and Hawkesworth,

1985; Beane et al., 1986; Subbarao, 1988; Lightfoot et al.,

1990; Peng et al., 1994). On the basis of geochemical

characteristics and field markers, the Western Ghats

sequence, with a total stratigraphic thickness of

,3,000 m, has been divided into three subgroups and

eleven formations (Table 1). Many of the formations have

been subdivided into members and chemical types.

Geochemical stratigraphic work in the central part of

the Deccan province (Fig. 1) has identified the Thakur-

vadi, Khandala and Poladpur Formations (Fms.) to the

ENE of the Western Ghats (Subbarao et al., 1994; Peng,

1998). The Poladpur, Ambenali and Mahabaleshwar Fms.

also extend into the southeastern Deccan for hundreds of

kilometres from their type sections in the southwest (e.g.

Mitchell and Widdowson, 1991; Bilgrami, 1999). Lavas

isotopically and chemically resembling the Ambenali,

Poladpur, and Khandala Fms also have been documented

in the northeastern part of the province (Mhow,

Chikaldara, and Jabalpur areas) but, because most of the

Khandala- and Poladpur-like lavas have systematically

higher Pb isotopic ratios than their Western Ghats

counterparts, they appear to have been erupted from

different vents (Peng et al., 1998). Recent work on

1367-9120/03/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.

doi:10.1016/S1367-9120(03)00116-0

Journal of Asian Earth Sciences 23 (2004) 127–139

www.elsevier.com/locate/jseaes

* Corresponding author. Tel.: þ91-22-2576-7263; fax: þ91-22-2572-

3480/2576-7253.

E-mail address: dchandra@geos.iitb.ac.in (D. Chandrasekharam).

sections in the Tapi River valley and western Satpura

Range, north of the Western Ghats (Sheth et al., 1997;

Chandrasekharam et al., 1999, 2000; Mahoney et al.,

2000), has identified a considerable volume of Poladpur-

like lavas and a few Mahabaleshwar-like and broadly

Bushe-like lavas and dykes, as well as flows and dykes

that have no known counterparts in the southwestern

Deccan. Unlike the northeastern Deccan, no Ambenali-

type lavas have been encountered in this region. Also, the

broadly Bushe- and Mahabaleshwar-like lavas are not in

the same stratigraphic order as in the southwestern

Deccan.

The Satpura region is one of structural complexity;

field relationships are often complicated and stratigraphic

correlations unclear. As in the Western Ghats, geochem-

ical characteristics of the lavas are a powerful tool for

stratigraphic correlation, and for interpreting regional

structure and tectonics. Here, we present major and

trace element data for lava flows and dykes from the

Bijasan Ghat section (Fig. 1) and for a few ‘giant

plagioclase basalt’ (GPB) flows from nearby sections, and

discuss their implications for the regional stratigraphy and

structure.

2. Regional geology

The Satpura Range constitutes a horst between two

graben, the Narmada in the north and the Tapi in the south.

The Narmada and Tapi rivers flow westward along the

respective graben (Fig. 1). A roughly linear tract of post-

Deccan-Traps (Tertiary–Recent) alluvium, ,350 km long

and with an average width of 30 km, occupies the Tapi

Fig. 1. Map showing the location of Bijasan Ghat and other sections in the western Satpura–Tapi region. Inset shows the position of the area of study within the

Deccan province (gray).

Table 1

Southwestern Deccan formation stratigraphy (Subbarao and Hooper, 1988)

Group Subgroup Formation

Deccan Basalt Wai Panhala

Mahabaleshwar

Ambenali

Poladpur

Lonavala Bushe

Khandala

Kalsubai Bhimashankar

Thakurvadi

Neral

Igatpuri

Jawhar

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139128

Valley, and isolated inliers of the Deccan Traps are exposed

in it. The alluvium thickens from south to north, in general,

and extends to a depth of ,200 m below mean sea level in

places (as identified from borings), where it is .400 m thick

(e.g. Ravishanker, 1987). On the basis of seismic reflection

and refraction data, Kaila (1988) inferred the presence of

Mesozoic sedimentary basins below the Deccan Traps in the

Narmada and Tapi graben; the Narmada Mesozoic basin

averages 1000 m deep and the Tapi Mesozoic basin 1800 m

deep.

The Deccan lava pile in the Tapi Valley, seen in

discontinuous sections, has a complex structure and

shows variable, tectonically controlled dips in many

places. Guha (1995) concluded that physical correlation

of basalt flows across the Tapi Valley is not possible as

there is not sufficient continuity between even nearby

sections as a result of faulting. Moreover, he inferred the

presence of major faults, both basin-marginal and

sympathetic. Stratigraphic sequences must be deciphered

for three isolated sectors: the southern (flows south of the

alluvium), the central (flows of inliers exposed within

the alluvium), and the northern (flows exposed north of

the alluvium). Flow stratigraphic relationships remain

largely unknown.

The contact between the Satpura horst and the Tapi

graben is marked by the Satpura Foothill Fault, expressed by

steep fault scarps along its 300-km-length (Guha, 1995).

However, although the exposed part of the Satpura Foothill

Fault dips southward at a high angle, its geometry at depth is

unknown. Available seismic profiles across the Tapi basin

(Kaila, 1988) do not reveal any significant low-angle

reflectors, but Guha (1995) suggested a ‘domino-type’

fault-block structure, possibly with a listric geometry, for

the basin (Fig. 2).

Bijasan Ghat is located near the boundary between the

Tapi Valley and Satpura Range. The local Deccan lava pile

dips north or northwest, and the dips appear to decrease

northward. On National Highway 3, from Shirpur through

Sangvi to Bijasan Ghat (Fig. 1), several isolated outcrops of

the lava flows exhibit southerly dipping colonnades,

consistent with the northerly dips (,15–308) of the flows.

At the highest point of the Bijasan Ghat traverse, however,

the flows are horizontal and form small mesas. Similarly, in

the foothills below Mt. Toranmal (1,152 m), cuestas with

northerly dip-slopes have developed on the basalts, dipping

north at about 58, but the main Toranmal massif is made up

of nearly horizontal flows (Sheth, 1998; Mahoney et al.,

2000). Guha (1995) suggested that this variation in dip is a

manifestation of step-faulting toward the south, with the

dips of fault blocks decreasing toward the north (Fig. 2).

Volcanic rock types other than basalt are not found in the

region. The basalts are aphyric to highly phyric (mostly

plagioclase-phyric), and several GPB flows are present. The

GPBs have plagioclase phenocrysts .2 cm, and as much as

5 cm, in length. Also seen are numerous inter-flow boles,

which are most commonly red, but also green, brown, or

black. They are up to 2 m thick, and represent soils or

altered ash layers (e.g. Wilkins et al., 1994). Basaltic and

doleritic dykes are abundant and trend mostly ENE–WSW

and E–W, and rarely N–S. Obvious feeder dykes (dykes

passing into surface flows) were not observed.

3. Samples, petrography, and geochemical

analytical methods

Fig. 3 summarizes the stratigraphy of the Bijasan Ghat

section and the locations of our samples, along with the

stratigraphy of sections at Babakunvar Dongar and 3 km

south of Boradi (Fig. 1). Stratigraphic summaries of the

Toranmal and Shahada ridge–Nandarde ridge sections can

be found in Mahoney et al., (2000); Chandrasekharam et al.,

(1999), respectively. The Bijasan Ghat section exposes

mainly aphyric, ‘simple’ flows (thick flows with arge

colonnade tiers). In thin section, microphenocrysts (pheno-

crysts smaller than ,0.25 mm, and discernible only under

the microscope) of plagioclase, clinopyroxene and com-

pletely altered olivine are seen in small but variable amounts

in most samples. The sample SH28, however, has micro-

phenocrysts only of iron oxides. Plagioclase phenocrysts in

the GPB SH26 are as long as 5 cm. Six other GPBs come

from nearby sections, including the Babakunvar Dongar

section and the section south of Boradi. Geochemical data

for three of these GPBs have been published previously

(Chandrasekharam et al., 1999; Mahoney et al., 2000).

Major element concentrations were determined by X-ray

fluorescence spectrometry and trace element abundances by

inductively coupled plasma mass spectrometry on agate-

and alumina-ground powders at the University of Hawaii

(Table 2), following the methods of Norrish and Chapell

(1977) and Jain and Neal (1996), respectively. Sr abun-

dances were measured by inductively coupled plasma

atomic emission spectrometry at the Indian Institute of

Technology, Bombay.

Fig. 2. Schematic illustration of the ‘domino-type’ fault-block structure

proposed for the Tapi graben (Guha, 1995), with parallel, E–W-trending

faults produced by N–S extension. The dips of individual faults system-

atically decrease with depth and to the south, and the northward dips of

bedding become progressively steeper to the south.

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139 129

4. Chemical characteristics

All Bijasan Ghat rocks are tholeiitic basalts, with

normative hypersthene and quartz. They are relatively

evolved, with MgO contents ranging from 3.79 (in GPB

SH26) to 7.09 wt.% and Mg numbers (Mg#) ranging from

42.8 to 52.5, where Mg# ¼ [atomic Mg/(Mg þ

Fe2þ)] £ 100, assuming Fe3þ/Fe2þ ¼ 0.16. Weight loss on

ignition (LOI) values for the samples provide a rough idea

of the subaerial alteration suffered by the rocks. LOI values

are all .1.30 wt.% and reach much higher values in the

GPB SH26 (6.26 wt.%) and in flows SH35 and SH36 (both

Fig. 3. Stratigraphic columns for the Bijasan Ghat and Babakunvar Dongar sections, and the section 3 km south of Boradi. Sample numbers (without prefixes)

are shown in bold and elevations in italics. The plagioclase-phyric (pl-phyric) flows are divided into those with large (lg., $10 mm) and small (sm., ,10 mm)

plagioclase phenocrysts. Flow boundaries shown as continuous lines were observed in outcrop, whereas those shown as dashed lines are not exposed but were

inferred from the presence of features indicating proximity to a flow boundary (e.g. vesicular or amygdaloidal zones, breccias, significant petrographic

changes). Note that a flow boundary must also exist in the zone of no exposure between the locations of samples 32 and 33, because these samples have very

different geochemical characteristics.

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139130

Table 2

Major and trace element compositions of the Bijasan Ghat basalts and Satpura–Tapi GPBs

Elev.(m) 395 425 410 470 492 555 595 605 610 620 630 240 250 440

dyke dyke GPB flow flow flow flow flow flow dyke flow GPB GPB GPB

Sample SH20 SH25 SH26 SH28 SH29 SH32 SH33 SH35 SH36 SH37 SH39 SH7 SH55 SH63 BHVO-1 Recom. BHVO-1 Meas.

wt.%

SiO2 50.13 48.66 50.65 49.79 49.45 51.87 48.01 49.16 49.02 48.37 48.70 50.79 50.63 49.01 49.89 49.59

TiO2 2.54 3.33 2.30 2.21 2.41 1.39 2.32 2.16 2.21 2.37 2.37 2.86 2.94 2.97 2.74 2.69

Al2O3 13.99 13.09 18.96 13.47 14.22 15.11 14.44 13.99 14.33 13.61 13.76 12.93 12.89 15.29 13.71 13.70

Fe2O3* 15.08 16.30 10.84 15.12 13.74 12.78 15.31 14.96 14.86 15.29 15.19 15.58 15.80 14.80 12.44 12.39

MnO 0.18 0.21 0.14 0.21 0.16 0.17 0.21 0.18 0.18 0.20 0.20 0.20 0.17 0.16 0.17 0.17

MgO 4.85 5.61 3.79 5.93 5.66 5.64 6.88 7.09 6.20 6.43 6.72 4.82 4.59 4.55 7.26 7.22

CaO 9.32 10.19 10.24 10.43 10.82 10.41 11.03 10.69 10.79 10.90 10.94 9.21 8.37 10.17 11.42 11.32

Na2O 2.17 2.03 2.96 1.90 1.84 1.97 1.58 1.48 1.59 1.77 1.75 2.21 2.00 1.84 2.24 2.24

K2O 0.98 0.25 0.56 0.43 0.46 0.44 0.11 0.37 0.11 0.21 0.17 0.75 1.22 0.38 0.51 0.52

P2O5 0.24 0.33 0.24 0.21 0.26 0.16 0.21 0.19 0.22 0.21 0.21 0.32 0.32 0.31 0.27 0.27

Total 99.48 100.00 100.68 99.70 99.02 99.94 100.10 100.27 99.51 99.36 100.01 99.67 98.93 99.48 100.65 100.11

LOI 2.82 1.62 6.26 1.34 2.74 2.82 4.24 5.80 5.80 1.80 3.06 1.36 2.36 2.78

Mg# 42.8 44.4 45.0 48.0 48.3 50.7 51.2 52.5 49.3 49.4 50.8 41.9 40.4 41.7

ppm

Ni 64 65 63 64 75 68 98 78 79 94 88 62 55 78 121 123

Cs nd nd 0.28 0.05 0.07 0.50 0.03 0.03 0.06 0.07 0.02 0.24 0.20 0.52 nd nd

Rb 22 4.4 14 7.7 8.8 11 3.0 3.8 2.6 2.6 1.6 15 31 14 9.7 8.6

Ba 278 105 186 107 114 171 44 64 56 64 62 265 285 124 133 137

Th 2.28 1.13 1.84 1.65 1.75 4.53 1.03 1.05 1.60 1.17 1.04 2.97 2.63 2.48 1.08 1.10

U 0.28 0.11 0.35 0.45 0.43 1.17 0.30 0.27 0.41 0.37 0.28 0.61 0.55 0.59 0.41 0.39

Nb 9.1 18 11 9.6 14 8.7 9.3 7.9 9.5 9.9 9.9 13 12 15 19 17

Ta 0.56 1.09 0.59 0.61 0.82 0.55 0.47 0.42 0.66 0.67 0.66 0.73 0.65 0.81 1.16 1.10

La 17.8 17.3 16.8 11.8 14.2 18.7 9.83 9.27 11.1 10.2 10.2 24.0 23.4 21.3 15.8 15.9

Ce 37.6 39.5 36.2 28.8 32.6 37.2 24.8 22.8 25.3 25.5 25.2 50.0 47.8 45.5 37.8 38.3

Pr 5.01 5.47 4.72 4.22 4.45 4.30 3.41 3.13 3.51 3.44 3.44 6.77 6.92 6.33 5.40 5.38

Pb 5.6 3.4 2.4 2.5 2.0 5.6 1.5 1.9 2.1 1.5 1.5 4.8 4.0 3.0 2.0 2.1

Sr 299 273 440 221 195 118 221 195 247 143 221 273 249 277 370

Nd 23.4 26.4 20.9 19.8 21.7 17.9 16.2 14.8 16.0 17.5 16.6 31.2 32.5 29.8 24.8 24.6

Zr 132 153 130 152 167 132 136 119 130 136 133 216 224 210 179 179

Sm 5.99 6.42 5.29 5.47 5.75 4.27 4.91 4.20 4.52 4.69 4.50 7.67 7.27 7.14 6.10 5.72

Eu 1.79 1.98 1.85 1.79 1.79 1.21 1.52 1.41 1.48 1.53 1.52 2.32 2.17 2.10 1.98 1.89

Gd 6.56 7.61 6.12 6.29 6.67 5.36 5.70 5.17 5.37 5.61 5.53 8.41 8.61 7.76 6.40 6.06

Tb 1.00 1.18 0.86 1.06 1.16 0.89 0.93 0.88 0.82 0.87 0.83 1.28 1.23 1.16 0.96 0.89

Dy 5.41 6.30 4.85 6.11 6.30 5.44 5.34 4.69 5.20 5.20 5.31 7.48 7.27 6.77 5.20 5.12

Y 29.7 34.7 26.8 34.2 35.3 33.0 28.7 25.6 30.0 29.1 28.5 38.5 40.0 37.8 27.0 24.0

Ho 1.13 1.27 0.94 1.25 1.30 1.12 1.02 0.95 1.04 1.10 0.98 1.42 1.41 1.34 0.99 0.91

Er 2.75 3.05 2.38 3.19 3.25 3.22 2.59 2.33 2.74 2.64 2.48 3.77 3.60 3.39 2.40 2.36

Tm 0.37 0.45 0.35 0.47 0.50 0.51 0.40 0.36 0.40 0.42 0.37 0.53 0.50 0.46 0.33 0.29

Yb 2.28 2.54 2.10 2.98 3.32 3.34 2.36 2.32 2.52 2.35 2.48 3.22 3.29 3.06 2.02 1.89

Lu 0.34 0.39 0.29 0.46 0.56 0.51 0.37 0.37 0.37 0.37 0.38 0.46 0.46 0.43 0.29 0.26

Relative uncertainties on major and minor elements are ,1%, and for SiO2, ,0.5%. For trace elements at .0.4 ppm in the rock, the relative uncertainty for most elements is 3.5% or better (range is 1–5%).

An indication of accuracy is given by measured and recommended (mainly from Govindaraju, 1989) values for standard rock BHVO-1. Fe2O3* is total iron measured as Fe2O3. The GPB SH7 was sampled along

the National Highway 3 between Shirpur and Sangvi at an elevation of 240 m.

H.C

.S

heth

eta

l./

Jou

rna

lo

fA

sian

Ea

rthS

ciences

23

(20

04

)1

27

–1

39

13

1

5.80 wt.%). Alteration has arguably caused some redis-

tribution of the mobile elements in some of the samples (see

below), as K, Rb and often U can be affected at rather

modest levels of alteration, and Ba and Na can be affected at

higher levels (e.g. Cox and Hawkesworth, 1985; Mahoney

et al., 1985, 2000; Beane et al., 1986; Lightfoot and

Hawkesworth, 1988). With advanced subaerial alteration,

many more elements, including Si, Ca, P, Ni, Sr, Y and the

lanthanide rare earths, can be affected, although elements

such as Al, Nb, Zr, Fe, Cr and Ti are relatively resistant (e.g.

Mahoney et al., 2000 and references therein).

The primitive-mantle-normalized multielement patterns

of the Bijasan Ghat flows and dykes are consistent with

variable contamination. Several flows have very similar

patterns. For example, flows SH33 and SH35 are chemically

very similar (Fig. 4a), with distinct Pb peaks, and Nb-Ta, Sr,

P and Eu troughs, characteristics consistent with contami-

nation by many types of continental crust and, for Sr, Eu,

and P, control by plagioclase and apatite fractionation.

Although SH35 has an even higher LOI value (5.80 wt.%)

than SH33 (4.24 wt.%), SH33 shows a pronounced K trough

and lower Rb and Ba than SH35, probably a result of loss

caused by alteration. For many elements, both patterns

broadly resemble that produced by 5% bulk contamination

of average Ambenali Fm. magma (the formation least

affected by continental material) by average Archaean felsic

crust (Fig. 4a). Note that the lavas have even lower Nb/La

and Ta/La than provided by this particular mixture,

however.

Flows SH36 and SH39 and dyke SH37 are likewise very

similar in most elements (Fig. 4b), with distinct Pb peaks

and pronounced lows at Rb and K, the latter probably

largely a result of alteration. SH36 shows a peak at Th and

U. All three patterns also have small but distinct troughs at P

and Eu, and the SH37 pattern has a pronounced Sr trough. In

contrast to SH33 and SH35, these three samples do not show

a depletion in Nb or Ta relative to La.

Patterns for the dyke SH25 and flows SH28 and SH29

(Fig. 4c) also exhibit distinct Pb peaks and Sr and P troughs,

and small Eu troughs. (Nb/La)N and (Ta/La)N vary from ,1

Fig. 4. Primitive-mantle-normalized multielement patterns for Bijasan Ghat flows and dykes and Satpura–Tapi GPBs. Similar patterns are grouped together in

each panel. Patterns of hypothetical bulk mixtures of Ambenali or average T-MORB magmas and average Archaean felsic crust (Rudnick and Fountain, 1995)

are also shown for comparison in panels a and d. T-MORB represents a 50:50 mixture of average N-MORB (normal-MORB) and average E-MORB (enriched-

MORB). MORB compositions and the primitive-mantle normalizing values are from Sun and McDonough (1989). In panel a, the Ambenali average pattern

(see Mahoney et al., 2000 for data sources) was adjusted to the same Lu-value as that of SH33 and SH35 (by multiplying the Ambenali abundances by 0.88) in

order to minimize differences related to different amounts of crystal fractionation. In panel d, the T-MORB values were multiplied by 1.26 to obtain the same

Lu value as that of SH32, and the Ambenali values were multiplied by 0.81 to adjust the Ambenali pattern to the same Lu value as that of SH20.

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139132

(SH28) to ,1, where N is the estimated primitive mantle

ratio (Sun and McDonough, 1989). The dyke SH25 appears

to have lost Rb, K, and U and has a marked Ti peak.

The patterns for dyke SH20 and flow SH32 (Fig. 4d) have

more pronounced Pb peaks and larger troughs at Nb and Ta.

The lows at P, Sr, and Eu are variable. SH20 has a small Ti

peak, whereas SH32 shows a Th-U peak. The dyke SH20

may have lost some U, whereas SH32 may have lost Rb and

K. The Ambenali magma type of the Deccan resembles

average T-MORB (transitional MORB) in isotopic charac-

teristics and many element ratios (e.g. Mahoney, 1988;

Lightfoot and Hawkesworth, 1988), and the SH32 pattern is

rather similar, from Nb to Lu, to a hypothetical 75:25 mixture

of T-MORB and average Archaean felsic crust. The pattern

of the dyke SH20 is only roughly similar to it, because SH20

has significantly higher concentrations of the middle rare

earths (Sm–Tb) and Sr, P, and Ti.

Patterns for GPBs of the region (Fig. 4e and f) all have

distinct Pb peaks, and all except SH105 have P troughs. All

GPBs except SH26 (from Bijasan Ghat) also have Sr

troughs, and many have small troughs at Eu. Several

(Fig. 4e) have pronounced Nb–Ta troughs and two have K

peaks and RbN .ThN. The GPBs SH47 and SH105 lack

Nb–Ta troughs (Fig. 4f). As a group, the GPBs are more

enriched in the light rare earths than the other flows and

dykes, with (La/Sm)N ¼ 2.69 2 3.18 and (La/

Yb)N ¼ 4.87 2 8.51. In comparison, the Bijasan Ghat

basalts have (La/Sm)N ¼ 1.29 2 2.82 and (La/

Yb)N ¼ 2.85 2 5.75.

5. Regional geochemical comparisons and stratigraphic

correlation

To compare our samples to the geochemically well-

characterized southwestern Deccan formations, we used

three approaches: (i) binary plots of incompatible element

abundances and element-element ratios, (ii) comparison of

the primitive-mantle-normalized multielement patterns of

the Bijasan Ghat samples with patterns of the southwestern

lavas, and (iii) discriminant function analysis employing

major elements and several commonly analyzed trace

elements.

5.1. Binary chemical diagrams

The Nb/Zr ratio changes little even during extreme

alteration and is therefore one of the more useful in

petrogenetic interpretation of altered basalt (e.g. Mitchell

and Widdowson, 1991; Widdowson et al., 2000). Fig. 5a is

Fig. 5. Variation diagrams of (a) Nb/Zr vs. Ba/Y, (b) Sr vs. Nb/Zr, (c) Ba/Zr vs. Nb/Zr, and (d) Nb/Y vs. Rb/Y for the Bijasan Ghat samples and the Satpura–

Tapi GPBs. Fields for several southwestern Deccan formations are from the data of Beane (1988).

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139 133

a plot of Nb/Zr vs. Ba/Y. Y is also resistant to alteration, and

Ba is less mobile than K or Rb. Both ratios are insensitive to

fractional crystallization of minerals common in basalt.

Data for the Bijasan Ghat flow samples and dykes SH20 and

SH37 fall within the Poladpur Fm. field, mostly in the area

of overlap among it and the Khandala, Ambenali and Bushe

Fm. fields. The data point for dyke SH25 falls within the

Mahabaleshwar Fm field. In plots of Sr vs. Nb/Zr (Fig. 5b)

and of Ba/Zr vs. Nb/Zr (Fig. 5c), data for most Bijasan Ghat

flow samples are again seen to plot in the area of overlap

between the Poladpur, Khandala, Ambenali and Bushe Fm.

fields. The data point for dyke SH25 again lies well within

the Mahabaleshwar Fm. field in Fig. 5a and b. In a plot of

Nb/Y vs. Rb/Y (Fig. 5d), most Bijasan Ghat flow samples

fall again within the area of overlap between the Poladpur,

Ambenali and Bushe Fm. fields, and the data point for dyke

SH25 falls within the Ambenali Fm. field, just outside the

edge of the Mahabaleshwar Fm. field.

Thus, based on the element-ratio plots, most of the

Bijasan Ghat rocks could be grouped with either the

Poladpur or Ambenali Fms. The difficulty of distinguishing

some Ambenali and Poladpur compositions on the basis of

the most commonly analyzed trace elements is well

recognized (e.g. Cox and Hawkesworth, 1985; Devey and

Lightfoot, 1986). Lightfoot et al. (1990) proposed that Ba

contents can be used to distinguish between Ambenali and

Poladpur lavas, with the Ambenali generally having

,100 ppm Ba and the Poladpur generally .100 ppm Ba.

We disagree; Ba is certainly useful in discriminating the

Mahabaleshwar Fm. (Ba .100 ppm) from the Ambenali

Fm., but all Ambenali Fm. flows and the great majority of

Poladpur Fm. flows contain Ba ,100 ppm (Beane, 1988;

P.R. Hooper, pers. comm., 2003), the exception being a

group of coarsely plagioclase-phyric Poladpur Fm. flows.

Five of the Bijasan Ghat flows have Ba p100 ppm; another

flow (SH28) and a dyke (SH25) have values of ,100 ppm

within analytical error. Because Ba may have been some-

what mobile during alteration, the Ba evidence is somewhat

ambiguous. The best evidence, in the absence of isotopic

data, for the Poladpur-Fm. affinity of the Bijasan Ghat flows

in question is the large Pb spikes in their primitive-mantle-

normalized multielement patterns (Fig. 4). These flows are

evidently contaminated by Pb-rich continental material, and

therefore none of them can be linked with the Ambenali Fm.

5.2. Comparison of multielement patterns

The primitive-mantle-normalized multielement patterns

of flows SH28 and SH29 (Fig. 6a) are rather similar to that

of the Visapur member of the Poladpur Fm. in most

elements, whereas they are distinct from the Ambenali Fm.

pattern in their large Pb peaks, relative enrichments in Rb,

Ba, Th and U, and in Nb–Ta–La relationships. Like

Poladpur Fm. lavas, the Bijasan Ghat samples have much

flatter patterns than Khandala Fm. lavas. Patterns for

Bijasan Ghat flows SH33 and SH35 and, to a lesser extent,

flows SH36 and SH39 and dyke SH37, also are Poladpur-

like (Fig. 6b and c).

Bushe Fm. patterns are flatter than those of the

Poladpur from Sr to Lu, and have much larger Pb peaks

and more pronounced Nb–Ta and Ti troughs, in particular

(e.g. Lightfoot et al., 1990; Mahoney et al., 2000). Bijasan

Ghat flow SH32 has Bushe-type characteristics, the closest

Bushe Fm. member to flow SH32 being the Pingalvadi

(Fig. 6d). SH32, however, has a lower Sr abundance than

Bushe Fm. basalts and its pattern also shows a sizeable

Eu–Ti trough. Broadly Bushe-like patterns have been

reported for dyke SH49 from Boradi (Chandrasekharam

et al., 1999) and several flows (e.g. SH114) from

Toranmal (Mahoney et al., 2000). The dyke SH20 also

shows a few similarities with the Bushe Fm., but is much

more enriched in the elements from Sr to Tb (Fig. 6e).

The pattern of the dyke SH25, however, is similar to the

average Mahabaleshwar Fm. pattern in many elements

except, in particular, for the higher Nb/La and Ta/La and

a much larger Ti peak in the SH25 pattern (Fig. 6f). Also

shown for comparison in Fig. 6f are patterns of

Mahabaleshwar-like dyke SH41 from the Shahada ridge

(Chandrasekharam et al., 1999) and Mahabaleshwar-like

flow SH115 from Toranmal (Mahoney et al., 2000).

In summary, none of the Bijasan Ghat patterns resembles

the Ambenali Fm. pattern. Ambenali-like lavas are also

absent in the nearby Toranmal section (Mahoney et al.,

2000). Patterns of most Bijasan Ghat flows and dyke SH37

are similar to those of the Poladpur Fm. in the southwestern

Deccan, whereas flow SH32 and dyke SH20 display general

similarities to Bushe Fm. lavas. However, the Bushe Fm.

underlies the Poladpur Fm. (Table 1). Therefore, the Bushe-

like flow SH32, sandwiched between a group of Poladpur-

like flows, cannot be correlated with the Bushe Fm. if the

Poladpur-like Bijasan Ghat flows in fact belong to the

Poladpur Fm. Peng et al. (1998) and Mahoney et al. (2000)

also found several broadly Bushe-like flows interspersed

within thick Wai-Subgroup-type sequences in the north-

eastern Deccan and at Toranmal, respectively; these flows

have combined Nd–Sr–Pb isotopic characteristics outside

the range measured for Bushe Fm. lavas. On the other hand,

the presence of the rather Mahabaleshwar-Fm.-like dyke

SH25 within a Poladpur-Fm.-like lava sequence is consist-

ent with the southwestern Deccan stratigraphic order, and

with relationships in sections near Shahada and Nandarde,

which expose Poladpur-Fm.-type lavas intruded by Maha-

baleshwar-Fm.-like dykes (Fig. 1; Sheth et al., 1997;

Chandrasekharam et al., 1999, 2000).

5.3. Discriminant function analysis

To further evaluate the Bijasan Ghat basalts, we

performed a discriminant function analysis using the

major elements and several of the most commonly analyzed

trace elements, a southwestern Deccan data set consisting of

623 samples from all the southwestern formations except

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139134

the Panhala, and the SPSS 7.5 for Windows (Student

Version) software. Data for the rare earths, Pb, Th, U, etc.

are available for a relatively small number of southwestern

Deccan basalts; hence, these elements were not used.

Following Peng et al. (1998) and Mahoney et al. (2000), no

derived variables (Zr/Y, Mg#, etc.) were used, and Cr was

not used because of contamination problems with some of

the southwestern Deccan Cr analyses. Among the major

elements, MnO was excluded because some of the south-

western Deccan samples lack Mn data and many measure-

ments are of low precision, yet the range of values is very

limited (,0.18 2 0.22 wt.%). Na2O was not used because

of its limited range of variation, variable analytical quality

in the southwestern data set, and alteration effects in some

samples. The trace elements used were Ba, Ni, Sr, Zr, Y, and

Nb. We performed the analysis both with and without K2O

as a discriminating variable and obtained virtually identical

results. We also performed the analysis both with and

without Fe2O3*, MgO and Ni (elements whose concen-

trations can show significant intra-flow variation as a result

of local fractional crystallization or phenocryst accumu-

lation). The best separation between the southwestern

formation fields was achieved when K2O, Fe2O3*, MgO

and Ni were not used as discriminating variables, and the

results of this analysis are reported here.

For the analysis, both the southwestern Deccan and

Bijasan Ghat data were transformed to standardized values

(Z-scores). The program calculated the F-statistic (essen-

tially the ratio of between-group variability to within-group

variability) for each variable, and the discriminant func-

tions, group centroids, and Mahalanobis distance of each

sample from the closest formation centroid. For the

southwestern Deccan data set, eight canonical discriminant

functions were obtained. The first two functions together

account for 76.2% of the total variance in the southwestern

Deccan data set, and 80.3% of the southwestern Deccan

samples, when run as unknowns, were classified correctly;

i.e., matched with their correct formations (cf. Peng et al.,

1998).

Table 3 summarizes the results of the analysis for the

Bijasan Ghat samples and GPBs, listing the closest south-

western-formation matches (if any), the corresponding

Mahalanobis distances, and the scores of the first two

discriminant functions. These scores are plotted in Fig. 7,

which also shows the fields defined by the relevant

southwestern Deccan formations. Five Bijasan Ghat

Fig. 6. Comparison of primitive-mantle-normalized multielement patterns of the Bijasan Ghat flows and dykes with those of selected southwestern Deccan

lavas, members, or formation averages (main data sources are Beane et al., 1986; Beane, 1988; P. R. Hooper, manuscript in preparation). The patterns are

compared at the same Lu value.

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139 135

samples were matched with the Ambenali and two with the

Poladpur at relatively small Mahalanobis distances. Flows

SH28 and SH29 were grouped with the Poladpur Fm.,

consistent with the shapes of their normalized multielement

patterns and their positions in Fig. 5. Flows SH33, SH35,

SH36 and SH39, and dyke SH37, were grouped with the

Ambenali Fm., although these basalts are most similar to

the Poladpur Fm. in their normalized multielement patterns.

Noting the overlap in abundances of the commonly

analyzed elements between the Poladpur and Ambenali

Fms., Mahoney et al. (2000) also observed that several

isotopically Poladpur-like basalts with Poladpur-type multi-

element patterns in the Toranmal section were matched with

the Ambenali Fm. in discriminant function analysis. We

interpret the combined major and trace element data and

discriminant function analysis results to indicate an overall

Poladpur-type affinity for these Bijasan Ghat basalts.

Flow SH32 was grouped with the Bushe Fm., which is

consistent with its Bushe-like multielement pattern. Dykes

SH20 and SH25 were not matched with any formation,

although the normalized multielement pattern of SH25

rather resembles the average Mahabaleshwar Fm. pattern

and in Fig. 7 the SH25 data point is closest to the

Mahabaleshwar Fm. field. The GPBs SH26 and SH55

could not be matched with any formation but GPB SH7

from the Shirpur–Sangvi traverse was matched with the

Khandala Fm., and GPB SH63 from the Babakunvar Dongar

section with the Igatpuri–Jawhar Fms., both at rather large

Mahalanobis distances. The significance of the matches for

the GPBs is not clear, as the samples are rich in phenocrysts

and may contain excess phenocrysts.

5.4. The GPBs, and structural complications

The usefulness of GPB flows as field and chemical

markers in the stratigraphy of the lower part (Kalsubai

Table 3

Summary of discriminant function analysis

Sample 1st Fm.

match

Mahalanobis

distance

Function 1 Function 2

SH39 (flow) Ambenali 5.12 2.032 20.329

SH36 (flow) Ambenali 4.36 1.686 20.327

SH35 (flow) Ambenali 8.67 1.659 0.069

SH33 (flow) Ambenali 7.37 1.743 20.036

SH32 (flow) Bushe 7.61 24.698 1.922

SH29 (flow) Poladpur 3.79 0.313 0.495

SH28 (flow) Poladpur 7.15 20.711 1.015

SH26 (GPB) ? 4.009 0.277

SH20 (dyke) ? 4.840 2.971

SH25 (dyke) ? 7.191 21.839

SH37 (dyke) Ambenali 9.55 1.400 0.313

SH7 (GPB) Khandala 20.2 1.102 4.664

SH55 (GPB) ? 0.563 5.899

SH63 (GPB) Igatpuri-Jawhar 12.1 1.051 1.711

Bijasan Ghat flows are listed in order of decreasing elevation. Samples

with Mahalanobis distance .24 (corresponding to a probability ,0.002)

are shown with question marks.

Fig. 7. Values of the first two canonical discriminant functions for the Bijasan Ghat rocks, with fields and centroids for several southwestern Deccan

formations. Function 1 ¼ 20.446SiO2 2 0.129Al2O3 þ 1.312TiO2 þ 0.124CaO þ 1.398P2O5 þ 0.503Ba þ 0.302Sr 2 1.699Zr 2 0.727Y 2 0.022Nb.

Function 2 ¼ 20.203SiO2 þ 0.078Al2O3 2 0.257TiO2 2 0.125CaO þ 0.453P2O5 þ 0.866Ba 2 0.251Sr þ 0.941Zr 2 0.050Y 2 1.221Nb. Note that the

oxide and elemental abundances in these equations are Z-score-standardized values (see Peng et al., 1998 and references therein). Formation centroids (shown

by black circles) have the following abbreviations and function scores: Igatpuri–Jawhar (IJ) 1.058, 1.066; Khandala (K) 0.677, 2.313; Bushe (B) 23.884,

0.099; Poladpur (P) 20.537, 20.662; Ambenali (A) 1.460, 21.122; Mahabaleshwar (M) 2.453, 21.581.

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139136

Subgroup) of the Western Ghats sequence has been well

documented (Hooper et al., 1988). The Kalsubai Subgroup

has six GPBs. They are not present together in any single

vertical section and are all similar in appearance. Therefore,

the chances of miscorrelation in the field are high, and

chemical ‘fingerprinting’ is required to check stratigraphic

correlations. Despite the wide compositional range of each

GPB, all can be distinguished straightforwardly on chemical

variation diagrams and/or with isotopic ratios (Hooper et al.,

1988; Peng et al., 1994).

In the southwestern Deccan, GPBs are found only within

the lower formations (Kalsubai Subgroup). The presence of

several GPBs in chemically Wai-Subgroup-like sequences

in the Satpura–Tapi region suggests that the region’s

eruptive history is at least partly independent of the

southwestern Deccan’s. Most highly plagioclase-phyric

flows and GPBs may not have flowed great distances,

because of their expected high viscosity (e.g. Mahoney,

1988). If so, the GPBs of the Satpura–Tapi region are

unlikely to be stratigraphically equivalent to any of the

GPBs of the Kalsubai Subgroup.

In the Satpura–Tapi region, no section we have studied

has more than one GPB flow. Within this region, field-based

correlation of the GPBs is rendered impossible by faulting,

erosion, and discontinuous exposures. Therefore, geochem-

ical data are essential for correlation. Consider the following

ratios of incompatible, alteration-resistant elements in three

of the Satpura–Tapi GPBs: Shahada ridge GPB SH47:

Nb/Zr ¼ 0.12 (^0.008), Nb/Y ¼ 0.60 (^0.042); Nandarde

ridge GPB SH52: Nb/Zr ¼ 0.08 (^0.006), Nb/Y ¼ 0.35

(^0.024); Toranmal GPB SH105: Nb/Zr ¼ 0.12 (^0.008),

Nb/Y ¼ 0.45 (^0.031). (The 2s errors on these ratios are

7%, calculated using a value of 5% (the maximum) for

errors on individual element concentrations.) SH47 appears

very similar to SH105 in many alteration-resistant incom-

patible elements (Table 4; Fig. 4f), but there are also

significant differences between the two; particularly in the

heavy rare earths. SH47 has phenocrysts only of plagio-

clase, whereas SH105 also contains microphenocrysts of

plagioclase, olivine, and clinopyroxene. For these two

flows, ratios of the light rare earths to the heavy ones (e.g.

La/Yb) are also substantially different. Also, based on

available chemical and isotopic data (Chandrasekharam

et al., 1999; Mahoney et al., 2000), no lava flow overlying

the Shahada ridge GPB can be correlated with any flow

overlying the Toranmal GPB.

The Nandarde ridge GPB (SH52) may be the same as the

GPB sampled 3 km south of Boradi (SH55) because the

normalized multielement patterns and key inter-element

ratios for both are similar (Fig. 4e, Table 4). However, some

ratios are substantially different (e.g. Nb/Th, Zr/Y). Note

that the Nandarde ridge and the section 3 km south of

Boradi are only a few kilometres apart, although there is no

physical continuity between them.

Like the Toranmal section, the Bijasan Ghat section is

made up mostly of simple flows with large colonnades. Yet

the GPBs (Toranmal: SH105, Bijasan Ghat: SH26) are

different (Fig. 4e and f; Table 4); for example, SH105 Nb/

Zr ¼ 0.12 (^0.008); SH26 Nb/Zr ¼ 0.08 (^0.006). Also,

the Shirpur–Sangvi GPB SH7 appears distinct from the

Bijasan Ghat GPB SH26, SH7 having Nb/Zr ¼ 0.06

(^0.004). The Bijasan Ghat and Babakunvar Dongar

sections are both situated along the same segment of the

Satpura range (Fig. 1), and are both made up of simple

flows. The Bijasan Ghat GPB SH26 is very similar to the

Babakunvar Dongar GPB SH63 in ratios of the alteration-

resistant incompatible elements (e.g. Nb/Zr, Nb/Th, Nb/Pb

and Nb/Y; Table 4). The primitive-mantle-normalized

patterns of both (Fig. 4e and f) are also similar in many

elements except, in particular, for the negative Sr and Eu

anomalies in the SH63 pattern, which the SH26 pattern

lacks. We conclude that the seven GPBs in our area of study

are not correlateable based on the existing geochemical

data, with the possible exception of SH52 and SH55, and

SH26 and SH63. Isotopic data are required for the

unambiguous correlation of these GPBs. Presently, how-

ever, the GPBs of this region, unlike those of the Western

Ghats, appear not to be very useful as either field or

chemical stratigraphic marker horizons.

6. Discussion and concluding remarks

The existence of Poladpur-like basalts along the

Bijasan Ghat section is not too surprising, because

chemically and isotopically Poladpur-like basalts consti-

tute a large part of the thick Toranmal section only

,70 km away. A few broadly Mahabaleshwar-like and

Bushe-like flows are also present at Toranmal. Noting the

presence of Bushe-like dykes at Toranmal and in the

Tapi Valley (Sheth, 1998; Chandrasekharam et al., 1999),

and that the Bushe-like Toranmal flows are not in the

southwestern Deccan stratigraphic order, Mahoney et al.

(2000) suggested that the Bushe-like flows may have had

relatively local feeder vents. The Bushe-like Bijasan Ghat

flow SH32, sandwiched between Poladpur-like flows, also

may have had a relatively local source.

Subbarao et al. (1994) noted that the southerly regional

formational dips of the southwestern Deccan flatten out and

become slightly northerly to the north of Igatpuri and

slightly easterly to the east of Igatpuri. Also, the lower

formations (Kalsubai Subgroup) extend north to Kondaibar,

near the Tapi River, and appear to dip below the surface

south of the latitude of Toranmal. These observations are

consistent with the presence of Poladpur-like flows at

Toranmal and Bijasan Ghat. Wai-Subgroup-like lavas are

also known from the area between Buldana and Igatpuri

(Subbarao et al., 1994; Peng, 1998), and from the

northeastern Deccan (Mahoney, 1988; Peng et al., 1998).

Flows with Poladpur- and Khandala-like chemical compo-

sitions dominate at Mhow, with a few Bushe-like flows,

and Ambenali-like flows are absent. The Jabalpur and

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139 137

Chikaldara areas have Ambenali-type basalts above Polad-

pur-like lavas, which are in turn underlain by lavas resem-

bling some members of the Khandala Fm (Peng et al., 1998).

However, most of the northeastern Poladpur- and Khandala-

like lavas are different from the chemically similar south-

western basalts in having consistently higher 206Pb/204Pb

ratios, and Peng et al. (1998) argued that they must have

erupted from different vents than the southwestern basalts.

Importantly, Khandala- and Poladpur-like dykes remain to

be found in the northeast or elsewhere in the Deccan.

Our work, along with other recent studies (e.g. Peng et al.,

1998; Chandrasekharam et al., 1999; Mahoney et al., 2000;

Widdowson et al., 2000; Sano et al., 2001), confirms that

several major magma types or formations defined in the

southwestern Deccan have a substantial geographic distri-

bution; in particular, Poladpur-type basalts are some of the

most widespread, found in locations as far apart as 800 km,

and with a possible original extent of $3 £ 105 km2; i.e.

about 60% of the present-day area of the Deccan province.

The Ambenali Fm. is also widespread; the reason for its

absence in sections such as Bijasan Ghat (present work),

Toranmal (Mahoney et al., 2000), Shahada and Nandarde

ridges (Chandrasekharam et al., 1999), and Mhow (Peng

et al., 1998) is as yet unknown. Continued sampling and

chemical and isotopic characterization of sections in the

Satpura–Tapi–Narmada region will provide valuable

insights into the stratigraphy and structure of this vast

volcanic province.

Acknowledgements

This work was supported by NSF Grant EAR-9418168 to

J.M. and a Dept. of Science and Technology (Govt. of India)

grant to D.C. We thank P.R. Hooper for sharing his

unpublished southwestern Deccan data, D. VonderHaar for

assistance in the lab, Z.X. Peng for help with discriminant

function analysis, and N. Hulbirt for advice on illustration.

Critical reviews by P.R. Hooper and G. Sen improved the

manuscript.

References

Beane, J.E., 1988. Flow stratigraphy, chemical variation and petrogenesis

of Deccan flood basalts from the Western Ghats, India. PhD

Dissertation, Washington State University, Pullman.

Beane, J.E., Turner, C.A., Hooper, P.R., Subbarao, K.V., Walsh, J.N., 1986.

Stratigraphy, composition and form of the Deccan basalts, Western

Ghats, India. Bulletin of Volcanology 48, 61–83.

Bilgrami, S. Z., 1999. A reconnaissance geological map of the eastern part

of the Deccan Traps (Bidar–Nagpur). Geological Society of India

Memoir 43(1), 219–232.

Chandrasekharam, D., Mahoney, J.J., Sheth, H.C., Duncan, R.A., 1999.

Elemental and Nd–Sr–Pb isotope geochemistry of flows and dikes

from the Tapi rift, Deccan flood basalt province, India. Journal of

Volcanology and Geothermal Research 93, 111–123.

Chandrasekharam, D., Vaselli, O., Sheth, H.C., Keshav, S., 2000.

Petrogenetic significance of ferro-enstatite orthopyroxene in basaltic

dikes from the Tapi rift. Deccan flood basalt province, India. Earth and

Planetary Science Letters 179, 469–476.

Table 4

Comparison of some key chemical parameters of the Satpura–Tapi GPBs

Sample SH7 SH26 SH47 SH52 SH55 SH63 SH105

Section Shirpur–Sangvi Bijasan Ghat Shahada ridge Nandarde ridge South of Boradi Babakunvar Dongar Toranmal

Elev.(m) 240 410 195 255 250 440 700

Mg# 41.9 45.0 42.0 37.0 40.4 41.7 44.7

TiO2 2.86 2.30 3.02 3.35 2.94 2.97 2.95

Nb 13 11 21 15 12 15 18

Ba 265 186 158 285 285 124 141

Sr 273 440 251 320 249 277 182

Zr 216 130 178 176 224 210 155

Y 38.5 26.8 34.1 42.0 40.0 37.8 40.2

Nb/Th 4.4 6.0 11 7.0 4.6 6.1 10

Nb/Zr 0.06 0.08 0.12 0.08 0.05 0.07 0.12

Nb/La 0.54 0.65 0.98 0.69 0.51 0.71 1.06

Nb/Pb 0.37 0.22 0.13 0.22 0.33 0.20 0.12

Nb/Y 0.34 0.41 0.61 0.36 0.30 0.40 0.45

Zr/Y 5.61 4.85 5.22 4.19 5.60 5.56 3.86

Nd/Sm 4.07 3.95 4.12 3.93 4.47 4.17 4.02

La/Sm 3.13 3.18 3.08 2.82 2.97 2.69 2.90

La/Yb 7.45 8.00 8.51 6.60 7.11 6.96 4.87

Pb/Nd 0.154 0.115 0.092 0.107 0.122 0.101 0.093

Pb/Zr 0.022 0.018 0.015 0.018 0.018 0.014 0.014

Ba/Y 6.88 6.94 4.63 6.79 7.13 3.28 3.51

Ba/Ti 0.016 0.014 0.009 0.015 0.017 0.007 0.008

Ba/Zr 1.23 1.43 0.89 1.62 2.11 0.69 0.91

Ba/Nb 20 17 7.7 19 24 8.2 7.8

Ba/Sr 0.97 0.42 0.63 0.89 1.14 0.45 0.77

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139138

Cox, K.G., Hawkesworth, C.J., 1985. Geochemical stratigraphy of the

Deccan Traps at Mahabaleshwar, Western Ghats, India, with impli-

cations for open system magmatic processes. Journal of Petrology 26,

355–377.

Devey, C.W., Lightfoot, P.C., 1986. Volcanological and tectonic control of

stratigraphy and structure in the western Deccan Traps. Bulletin of

Volcanology 48, 195–207.

Govindaraju, K., 1989. 1989 Compilation of working values and sample

descriptions for 272 geostandards. Geostandards Newsletter 13, 1–113.

Guha, S., 1995. Tectonic framework and evolution of the Tapi basin - an

intracratonic half-graben in west-central India. Indian Minerals 49,

61–78.

Hooper, P.R., Subbarao, K.V., Beane, J.E., 1988. The giant plagioclase

basalts (GPBs) of the Western Ghats, Deccan Traps. Geological Society

of India Memoir 10, 135–144.

Jain, J.C., Neal, C.R., 1996. Report of the ICP-MS facility, 1993–1996.

Notre Dame Univ. Open File Report, p. 30

Kaila, K.L., 1988. Mapping the thickness of Deccan Trap flows in India

from DSS studies and inferences about a hidden Mesozoic basin in the

Narmada–Tapti region. Geological Society of India Memoir 10,

91–116.

Lightfoot, P., Hawkesworth, C., 1988. Origin of Deccan Trap lavas:

evidence from combined trace element and Sr-, Nd- and Pb-isotopic

studies. Earth and Planetary Science Letters 91, 89–104.

Lightfoot, P.C., Hawkesworth, C.J., Devey, C.W., Rogers, N.W., van

Calsteren, P.W.C., 1990. Source and differentiation of Deccan Trap

lavas: implications of geochemical and mineral chemical variations.

Journal of Petrology 31, 1165–1200.

Mahoney, J.J., 1988. Deccan Traps. In: Macdougall, J.D., (Ed.),

Continental Flood Basalts, Kluwer, Dordrecht, pp. 151–194.

Mahoney, J.J., Macdougall, J.D., Lugmair, G.W., Gopalan, K., Krishna-

murthy, P., 1985. Origin of contemporaneous tholeiitic and K-rich

alkalic lavas: a case study from the northern Deccan plateau, India.

Earth and Planetary Science Letters 73, 39–53.

Mahoney, J.J., Sheth, H.C., Chandrasekharam, D., Peng, Z.X., 2000.

Geochemistry of flood basalts of the Toranmal section, northern Deccan

Traps, India: implications for regional Deccan stratigraphy. Journal of

Petrology 41, 1099–1120.

Mitchell, C.H., Widdowson, M., 1991. A geological map of the southern

Deccan Traps, India and its structural implications. Journal of

Geological Society of London 148, 495–505.

Norrish, K., Chapell, B.W., 1977. X-ray fluorescence spectrometry. In:

Zussman, J., (Ed.), Physical Methods in Determinative Mineralogy,

second ed., Academic Press, New York, pp. 201–272.

Peng, Z.X., 1998. Geochemical studies of the northwestern, central and

northeastern Deccan Traps, India: sources, petrogenesis, and relations

to southwestern Deccan stratigraphy, PhD Dissertation, University of

Hawaii, Honolulu.

Peng, Z.X., Mahoney, J.J., Hooper, P.R., Harris, C., Beane, J.E., 1994. A

role for lower continental crust in flood basalt genesis? Isotopic and

incompatible element study of the lower six formations of the

western Deccan Traps. Geochimica et Cosmochimica Acta 58,

267–288.

Peng, Z.X., Mahoney, J.J., Hooper, P.R., Macdougall, J.D., Krishnamurthy,

P., 1998. Basalts of the northeastern Deccan Traps India: isotopic and

elemental geochemistry and relation to southwestern Deccan strati-

graphy. Journal of Geophysical Research 103, 29843–29865.

Ravishanker, 1987. Neotectonic activity along the Tapi–Satpura lineament

in central India. Indian Minerals 41, 19–30.

Rudnick, R., Fountain, D.M., 1995. Nature and composition of the

continental crust: a lower crustal perspective. Reviews of Geophysics

33, 267–309.

Sano, T., Fujii, T., Deshmukh, S.S., Fukuoka, T., Aramaki, S., 2001.

Differentiation processes of Deccan Trap basalts: contribution from

geochemistry and experimental petrology. Journal of Petrology 42,

2175–2195.

Sen, G., 2001. Generation of Deccan Trap magmas. Proceedings of the

Indian Academy of Sciences (Earth and Planetary Sciences) 110,

409–431.

Sheth, H.C., 1998. Geochemistry, petrogenesis, stratigraphy and structure

of Deccan flood basalts of the western Satpura–Tapi region, India. PhD

Dissertation, Indian Institute of Technology, Bombay.

Sheth, H.C., Duncan, R.A., Chandrasekharam, D., Mahoney, J.J., 1997.

Deccan Trap dioritic gabbros from the western Satpura–Tapi region.

Current Science 72, 755–757.

Subbarao, K.V., (ed.), 1988. Deccan Flood Basalts. Geological Society of

India Memoir 10, 393.

Subbarao, K.V., Hooper, P.R., 1988. Reconnaissance map of the Deccan

Basalt Group in the Western Ghats, India. Geological Society of India

Memoir 10 (enclosure).

Subbarao, K.V., Chandrasekharam, D., Navaneethakrishnan, P., Hooper,

P.R., 1994. Stratigraphy and structure of parts of the central Deccan

basalt province: eruptive models. In: Subbarao, K.V., (Ed.), Volcanism,

Wiley, New Delhi, pp. 321–332.

Sun, S.-s., McDonough, W.F., 1989. Chemical and isotopic systematics of

oceanic basalts: implications for mantle composition and processes.

Geological Society of London Special Publication 42, 313–345.

Wadia, D.N., 1975. Geology of India, fourth ed., Tata McGraw-Hill, New

Delhi, p. 508.

Widdowson, M., Pringle, M.S., Fernandez, O.A., 2000. A post K–T

Boundary (Early Palaeocene) age for Deccan-type feeder dykes, Goa,

India. Journal of Petrology 41, 1177–1194.

Wilkins, A., Subbarao, K.V., Ingram, G., Walsh, J.N., 1994. Weathering

regimes within the Deccan basalts. In: Subbarao, K.V., (Ed.),

Volcanism, Wiley, New Delhi, pp. 217–231.

H.C. Sheth et al. / Journal of Asian Earth Sciences 23 (2004) 127–139 139