Provenance of Cambrian conglomerates from New Zealand...

Journal of the Geological Society, London, Vol. 163, 2006, pp. 997–1010. Printed in Great Britain.

997

Provenance of Cambrian conglomerates from New Zealand: implications for the

tectonomagmatic evolution of the SE Gondwana margin

MARCUS GUTJAHR 1,2, JOHN D. BRADSHAW 1, STEVE WEAVER 1, CARSTEN MUNKER 3;4 &

TREVOR IRELAND 5

1Department of Geological Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealand2Present address: Institute for Isotope Geochemistry and Mineral Resources, Department of Earth Sciences, ETH-Zentrum,

NW81.2, Clausiusstrasse 25, 8092 Zurich, Switzerland (e-mail: [email protected])3Institut fur Mineralogie, Westfalische Wilhelms Universitat Munster, Corrensstrasse 24, 48149 Munster, Germany

4Present address, Universitat Bonn, Mineralogisch-Petrologisches Institut, 53115 Bonn, Germany5Australian National University, Canberra, ACT0200, Australia

Abstract: The oldest rocks in New Zealand are the Mid- to Late Cambrian intra-oceanic island arc rocks of

the Takaka terrane (Devil River arc). The provenance of Cambrian conglomerates stratigraphically above the

exposed arc succession was studied to constrain the late stages of arc evolution and its accretion to continental

crust. The Dead Goat Conglomerate contains two distinct groups of igneous clasts: (1) intermediate to felsic

volcanic clasts with moderately enriched light rare earth element (LREE) and high field strength element

(HFSE) contents and positive ENd500 (+2.1) that were derived from a medium-K calc-alkaline source, probably

the main sequence of the Devil River arc; (2) dioritic to metagranitic plutonic clasts strongly enriched in

LREE and HFSE and with ENd500 of +3.5 to +5.9 that were derived from a high-K arc source, probably the

uppermost units of the Devil River arc. This is consistent with a new U–Pb sensitive high-resolution ion

microprobe age of 496 � 6 Ma. The Lockett Conglomerate also contains two distinct groups of igneous clasts:

(1) ultramafic to intermediate igneous clasts identified as boninitic to transitional low-K calc-alkaline arc-

related rocks based on depleted REE and HFSE abundances; (2) ‘I’-type metagranitoid clasts derived from a

distinct Andean type continental margin, as indicated by ENd500 as low as �7.1. Both conglomerates contain

sandstone clasts derived from a common old, multi-cycle continental source with ENd500 of �14.2 to �15.7,

and no suitable source has been found in present-day New Zealand. The new provenance data from these

conglomerates constrain the time of accretion of the Devil River arc to the palaeo-Pacific Gondwana margin

and provide new information on the structural evolution of the accretionary event.

The oldest rocks of New Zealand are exposed in the Early to

Mid-Palaeozoic Takaka terrane in NW Nelson and SW Fiordland

(Fig. 1) (Cooper & Tulloch 1992), which are part of the Western

Province (Landis & Coombs 1967). These oldest units comprise

remnants of the late Mid- to Late Cambrian intra-oceanic Devil

River arc and the associated Haupiri Group sediments, forming

the western part of the Takaka terrane (Fig. 1). In the past two

decades many of the individual sedimentary and igneous suites

have been described in detail (e.g. Grindley 1980; Cooper 1989;

Pound 1993; Munker & Cooper 1999; Jongens et al. 2003). A

first reconstruction of the sequence was largely limited to the

igneous units (Munker & Cooper 1999). However, a fully

comprehensive reconstruction including the various sedimentary

units in the Haupiri Group has yet to be established. The aim of

the present study is to determine the provenance of igneous and

sedimentary clasts in two major conglomerate units within the

youngest sediments of the Devil River arc assemblage, to

constrain the later stages of arc evolution and, if possible, the

time of accretion.

Regional geology

The relics of the Cambrian Devil River arc are restricted to a

north–south-trending belt less than 15 km wide near the western

margin of the Takaka terrane (Fig. 1; Cooper 1989). The belt is

made up of 10 tectonic fault-bounded slices, each with a

coherent internal stratigraphy that differs somewhat from that of

adjacent slices, but that in some instances can be correlated

(Cooper & Tulloch 1992; Rattenbury et al. 1998; Munker &

Cooper 1999). Each slice contains igneous rocks of the Devil

River Volcanics Group and associated Haupiri Group sediments.

The Devil River Volcanics Group is divided into the older,

back-arc tholeiitic Mataki Volcanics, the younger back-arc

boninitic Cobb Igneous Complex, and the calc-alkaline arc-

related rocks of the Benson Volcanics, which evolved through

time from low-K to high-K type magmas (Munker & Cooper

1995, 1999). Basaltic to andesitic compositions are typical in the

Devil River arc (Munker & Cooper 1995, 1999) and felsic rocks

are rare, largely being restricted to the latest stages of volcanism

(McLean 1994; Funk 1996). Both the Benson and the Mataki

Volcanics formations are interbedded with Haupiri Group sedi-

ments. Furthermore, the Mataki Volcanics are intruded by dykes

of the Benson Volcanics type (Munker & Cooper 1999), support-

ing a close spatial relationship during the activity of the Devil

River arc. The back-arc boninitic Cobb Igneous Complex was

emplaced at shallow depths into Haupiri Group sediments and

the Mataki Volcanics.

Munker & Cooper (1999) reported the occurrence of boninitic

volcanic clasts in early Mid-Cambrian (Floran) conglomerates of

the Haupiri Group. Although similar in composition to rocks of

the Cobb Igneous Complex, these boninitic clasts stratigraphi-

cally underlie the Cobb Igneous Complex and Mataki Volcanics.

Fig.1.

(a)

Gen

eral

ized

geo

logy

of

Wes

tN

elso

n–

Nort

hW

estl

and,

South

Isla

nd,

New

Zea

land,

asgiv

enby

Cooper

&T

ull

och

(1992).

The

Anat

oki

Thru

stse

par

ates

the

Wes

tern

Bu

ller

terr

ane

fro

mth

e

Eas

tern

Tak

aka

terr

ane.

Cam

bri

anar

c-re

late

dro

cks

are

rest

rict

edto

the

Cen

tral

Bel

t.G

rey

box

indic

ates

posi

tion

of

the

fiel

dar

eash

ow

nin

(b).

Inse

t:in

ferr

edoff

shore

exte

nt

of

Ear

lyP

alae

ozo

icro

cks

of

the

Bu

ller

and

Tak

aka

terr

anes

.B

ox

inin

set

indic

ates

loca

tio

no

fm

ap(a

).M

TZ

,M

edia

nT

ecto

nic

Zo

ne.

(b)

Sch

emat

ico

utc

rop

of

the

Lo

cket

tan

dD

ead

Go

atco

ng

lom

erat

esN

ort

ho

fC

ob

bV

alle

y.M

apis

com

pil

edfr

om

Gri

ndle

y(1

98

0)

and

Po

un

d(1

99

3),

and

mod

ified

fro

mn

ewo

bse

rvat

ion

s.N

um

ber

sin

circ

les

refe

rto

com

po

nen

tco

un

tlo

cati

on

sw

her

eth

ed

ata

pre

sente

din

Tab

le1

wer

eco

llec

ted.

To

pog

rap

hic

con

tou

rin

terv

alis

40

m.

Th

esi

mp

lifi

edst

rati

gra

ph

icco

lum

nsh

ow

sth

em

ain

sed

imen

tary

form

atio

ns

of

the

Dev

ilR

iver

arc

wit

hth

eco

rres

pondin

gN

ewZ

eala

nd

Cam

bri

anst

ages

.A

ges

for

New

Zea

land

stag

esar

eta

ken

from

Cooper

(2004).

M . GUTJAHR ET AL.998

These clasts will be referred to hereafter as ‘pre-Devil River’

rocks.

The Haupiri Group sediments are interbedded with and overlie

the volcanic rocks and range in grain size from conglomerate to

mudstone, reflecting the wide variety of sedimentary environ-

ments in the Devil River arc. The provenance of the Haupiri

Group sediments is variable. Some units comprise predominantly

arc-derived material. The Junction Formation, on the other hand,

predominantly consists of continent-derived quartzo-feldspathic

sandstone, siltstone and channel-fill conglomerate (Pound 1993;

Roser et al. 1996; Wombacher & Munker 2000).

The Haupiri Group comprises mainly intra-arc sediments,

although the interlayering of back-arc tholeiitic Mataki Volcanics

at a number of localities (Siebert 1998; Munker & Cooper 1999)

suggests proximity to or gradation into a back-arc basin. The

distinctly continental character of the Junction Formation has led

to suggestions that the back-arc basin was bounded by the

Gondwana continent (Cooper & Tulloch 1992; Pound 1993;

Wombacher & Munker 2000).

Geological setting of the Dead Goat and the Lockettconglomerates

The Dead Goat Conglomerate (new name) is the thinner of the

two conglomerate units and lies stratigraphically near the top of

the Tasman Formation (see simplified stratigraphic column in

Fig. 1b). The Dead Goat Conglomerate is a granule- to cobble-

sized conglomerate deposited in a broad channel in a marine

environment. Pound (1993) included associated siltstones to-

gether with the conglomerate in the Dead Goat Member, the

youngest part of the Tasman Formation. Other sections of the

Dead Goat Conglomerate are likely to be exposed further to the

north of the study area shown in Figure 1b.

The stratigraphically overlying Lockett Conglomerate is more

diverse than the Dead Goat Conglomerate, ranging from granule-

to boulder-sized conglomerates, with single boulders reaching

1 m in diameter. At exposed contacts sandstone–siltstone lithol-

ogies of the Tasman Formation that are deposited above the Dead

Goat Conglomerate coarsen upward into pebble–cobble con-

glomerate of the Lockett Conglomerate. The basal Lockett

Conglomerate is marine, but the unit as a whole is regressive and

grades upward into braided fluvial deposits (Pound 1993). How-

ever, clear erosional discordant contacts between the Tasman

Formation and the Lockett Conglomerate are not exposed. In

most cases, later faulting disrupted the contacts between the two

formations. The matrix of the Lockett Conglomerate contains

detrital garnet and amphibole, which Coleman (1978) and Pound

(1993) interpreted to be derived from metamorphic continental

sources.

The uppermost part of the Tasman Formation contains trilo-

bites, including Dorypyge and Nepea cf. avara, that suggest a

Boomerangian age (Fig. 1b) (Munker & Cooper 1999). The

Lockett Conglomerate is regarded as Mindyallan. The conglom-

erate is over 1000 m thick (Grindley 1980) and the stratigraphic

top is unknown. The Lockett Conglomerate is cut by the Balloon

Melange, interpreted as a tectonic diapir by Jongens et al.

(2003). Both the Lockett Conglomerate and the Balloon Melange

are cut by dykes of Gendarme Dolerite with an 40Ar/39Ar age of

486 � 25 Ma (Munker & Cooper 1999).

Sampling and analytical methods

In our study, the relative abundance of the various components in the

conglomerates was assessed at one location in the Dead Goat Conglom-

erate and eight well-separated sites in the Lockett Conglomerate (Fig. 1b,

Table 1). Samples from the Dead Goat Conglomerate were taken from

the section SW of Mount Lockett and samples from the Lockett

Conglomerate were collected along strike between Cobb Valley and

Mount Lockett (Fig. 1b).

Major and trace element concentrations were determined by XRF using

a Philips PW2400 system at the University of Canterbury and following

the method of Norrish & Hutton (1969). All major element contents

discussed below are given loss on ignition (LOI) free. REE and high field

strength element (HFSE) analyses were made by inductively coupled

plasma-mass spectrometry (ICP-MS) using a VG Plasmaquad PQ1

system and following the procedure of Garbe-Schonberg (1993) at the

Universitat Kiel, Germany. Cross-calibration of the two methods was

ensured through a comparison of XRF data with ICP-MS data. Results

agree within better than 30%. ICP-MS results are presented in Table 2.

Detailed XRF results are available online at http://www.geolsoc.org.uk/

SUP18248. A hard copy can be obtained from the Society Library.

Felsic igneous clasts and sandstone clasts were analysed for whole-rock

Nd isotope composition and Sm–Nd concentrations by thermal ionization

mass spectrometry (TIMS) at the Institut fur Mineralogie, Universitat

Munster, Germany, using a VGSector54 multiple-collector system (Table

3). Samarium–neodymium concentrations were determined using a

Table 1. Summary of eight component counts for the Lockett Conglomerate (LC) and one count in the Dead Goat (DGC) Conglomerate

Count number

1 (LC) 2 (LC) 3 (LC) 4 (LC) 5 (LC) 6 (LC) 7 (LC) 8 (LC) LC average (%) 9 (DGC) (%)

Volcanic clasts 32 39 33 40 40 38 31 35 36 15Qz-diorites, granitoids 22 23 22 28 20 24 24 17 23 0Gabbros, diorites 14 6 12 7 7 10 15 14 11 0Ultramafic clasts 3 7 0 2 5 3 4 4 3 6Pink granitoids 0 0 0 0 0 0 0 0 0 5Purple volcaniclastic breccia 0 0 0 0 0 0 0 0 0 10White, grey, black chert 19 7 22 17 23 18 20 21 18 17Red chert 0 5 0 0 0 0 0 0 1 5Sandstone clasts 2 7 5 3 3 3 1 2 3 33Laminated sediment 0 1 2 0 0 0 0 0 ,1 5Unidentifiable 8 5 4 3 2 3 5 7 4 4Total 100 100

Locations of the individual counts are shown in Figure 1b. Fresh surfaces of at least 1 m2 were chosen. Count 6 was counted twice for the calculation of the averagecomposition of the Lockett Conglomerate (observed surface for count 6 was 2.7 m2). The gabbroic and dioritic clasts occur only in the Lockett Conglomerate, whereas the pinkgranitoid and volcaniclastic clasts occur only in the Dead Goat Conglomerate. Sandstone clasts are much more abundant in the Dead Goat Conglomerate. The igneous andsandstone clasts were analysed petrographically and geochemically.

LOCKETT AND DEAD GOAT CONGLOMERATE STUDY, NZ 999

http://www.geolsoc.org.uk/

mixed 149Sm–150Nd tracer. All isotope ratios for samples were calculated

back to an age of 500 Ma. Uncertainties concerning the age of the

granitic rocks (500 � 15 Ma) have a negligible effect on the initial

isotopic ratios and are within the analytical uncertainty. Neodymium

isotope results obtained by other workers and used here for correlation

were corrected to a La Jolla value of 0.51186.

A U–Pb zircon age was obtained for a metagranitic clast from the

Dead Goat Conglomerate by sensitive high-resolution ion microprobe

Table 2. Individual major and trace element contents of the mafic to intermediate samples

Unit: LC LC LC LC LC LC LC LC DGC DGCRock type: boninite boninitic

andesiteboniniticandesite

hb-metapyroxenite

hb–epi-diorite

hb-metadiorite

hb-metadiorite

hb-metadiorite

diorite volcanicbreccia

Sample: LC94 LC52 LC49 LC85 LC130 LC71 LC123 LC63 DG102 DG150

Volatile-free(wt%)SiO2 55.15 56.46 58.05 50.40 52.86 53.83 55.97 56.22 56.49 53.30TiO2 0.17 0.30 0.27 0.35 0.23 0.19 0.70 0.32 0.83 1.16Al2O3 12.43 14.32 14.44 18.50 20.01 19.07 18.01 15.35 13.38 14.90Fe2O3T 10.65 10.16 7.72 11.46 7.14 7.99 10.12 8.95 14.67 11.95MnO 0.15 0.17 0.09 0.11 0.13 0.14 0.15 0.15 0.11 0.14MgO 10.60 6.56 7.22 6.70 6.63 8.62 5.48 7.97 4.81 2.99CaO 6.69 7.12 5.65 7.19 6.90 3.80 3.34 6.15 4.41 8.38Na2O 0.94 3.21 3.43 0.35 1.39 0.54 1.14 1.23 3.81 6.24K2O 3.19 1.67 3.09 4.92 4.64 5.78 5.02 3.61 1.25 0.59P2O5 0.02 0.03 0.04 0.02 0.06 0.03 0.07 0.04 0.26 0.36Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00LOI 2.39 2.06 2.09 2.63 3.16 3.36 2.07 2.41 7.41 3.98Original total 100.21 100.10 100.45 99.94 100.30 100.20 99.32 100.25 100.40 99.97ppmLi 11.1 9.7 6.4 7.5 9.2 13.4 10.9 6.2 12.3 16.4Sc 41 40 28 45 32 30 33 40 11 26Rb 108 49 99 197 119 101 136 134 14 35Sr 59 269 148 96 158 118 78 129 345 80Y 6.0 9.6 7.0 5.0 5.6 5.0 9.8 8.7 51 19Zr 9 15 27 8 20 15 35 30 337 91Nb 0.45 0.56 1.58 0.46 1.50 1.01 2.30 2.25 37 9.6Mo 0.10 0.12 0.15 0.10 0.08 0.13 0.24 0.75 1.32 0.31Cd 0.01 0.02 0.07 0.26 0.06 0.04 0.07 0.07 0.19 0.07Sb 0.18 0.18 0.15 0.16 0.12 0.23 0.24 0.14 1.24 1.35Cs 1.45 1.13 2.28 1.52 3.37 8.1 2.66 1.47 0.60 1.84Ba 693 751 537 444 451 896 745 505 2232 84La 1.03 2.59 3.92 1.99 4.6 2.90 5.6 5.6 31 15Ce 2.53 4.70 9.14 3.4 7.85 6.77 12 12 68 36Pr 0.34 0.61 1.00 0.49 0.92 0.77 1.40 1.31 8.3 4.84Nd 1.61 2.68 4.09 2.12 3.5 3.1 5.5 5.0 35 20Sm 0.51 0.82 1.02 0.54 0.74 0.73 1.33 1.14 8.6 4.4Eu 0.16 0.31 0.22 0.30 0.31 0.17 0.39 0.34 1.98 0.92Gd 0.69 1.14 1.13 0.65 0.84 0.81 1.49 1.29 8.8 3.9Tb 0.13 0.21 0.19 0.11 0.14 0.14 0.26 0.22 1.51 0.59Dy 0.97 1.56 1.25 0.80 0.94 0.93 1.69 1.42 9.7 3.5Ho 0.23 0.36 0.26 0.18 0.20 0.19 0.36 0.31 2.04 0.72Er 0.71 1.09 0.77 0.56 0.59 0.57 1.04 0.89 5.9 2.04Tm 0.11 0.17 0.11 0.09 0.09 0.08 0.15 0.13 0.87 0.29Yb 0.82 1.18 0.78 0.60 0.59 0.56 1.04 0.87 5.7 1.95Lu 0.13 0.18 0.12 0.10 0.09 0.08 0.16 0.14 0.85 0.29Hf 0.33 0.54 0.81 0.27 0.65 0.50 1.10 0.99 7.9 2.35Ta 0.05 0.06 0.16 0.05 0.13 0.10 0.21 0.19 2.17 0.51Tl 0.36 0.26 0.48 0.81 0.63 0.44 0.48 0.51 0.07 0.19Pb 0.67 1.55 0.93 1.69 2.23 1.03 3.12 2.62 9.8 14.9Th 0.32 0.56 1.43 0.25 1.13 0.73 2.09 1.97 5.4 4.3U 0.09 0.17 0.41 0.11 0.31 0.13 0.43 0.42 1.63 0.98Mg-no. 58.4 47.7 56.9 45.2 56.7 60.4 43.4 55.7 31.7 26.1CaO/Al2O3 0.54 0.50 0.39 0.39 0.34 0.20 0.19 0.40 0.33 0.56Th/Y 0.05 0.06 0.20 0.05 0.20 0.15 0.21 0.23 0.11 0.23Ti/Zr 110 121 60 269 70 74 119 63 15 76Zr/Y 1.5 1.6 3.8 1.6 3.5 3.0 3.6 3.5 6.6 4.9Th/Yb 0.39 0.48 1.84 0.42 1.91 1.30 2.02 2.27 0.95 2.20Ta/Yb 0.06 0.05 0.21 0.08 0.22 0.17 0.20 0.22 0.38 0.26LaN/SmN 1.27 2.00 2.43 2.32 3.87 2.49 2.64 3.08 2.30 2.11DyN/YbN 0.77 0.86 1.05 0.86 1.02 1.07 1.06 1.06 1.10 1.18Eu/Eu* 0.80 1.00 0.64 1.54 1.21 0.66 0.85 0.87 0.69 0.67LaN/YbN 0.85 1.49 3.40 2.23 5.17 3.48 3.62 4.33 3.67 5.14

Major element concentrations were determined by XRF analysis, trace element concentrations by ICP-MS. LC, Lockett Conglomerate; DGC, Dead Goat Conglomerate.

M. GUTJAHR ET AL.1000

(SHRIMP) using SHRIMP II at the Australian National University,

Canberra. Methods have been described in detail by Muir et al. (1996).

Results and a description of the detailed analytical procedure can be

found in the Supplementary Publication (see p. 999).

The Cambrian time scale used in this paper follows that of Davidek et

al. (1998) and Encarnacion et al. (1999). Absolute ages in the simplified

stratigraphic column in Figure 1b are taken from a recent compilation by

Cooper (2004) for the Cambrian stages in New Zealand. Detailed

geochemical data for the igneous clasts are given in Table 2 and the

Supplementary Publication. The classification of high-Mg rocks is after

Le Bas (2000). To avoid misinterpretations for the ultramafic igneous

clasts two samples in excess of 1000 ppm Cr and 200 ppm Ni were

dismissed as ultramafic cumulates and excluded from the geochemical

provenance analysis.

Results

Clast populations

Besides quartzose sandstone clasts, which contribute about a third

of the clast population in the Dead Goat Conglomerate, abundant

chert and clasts of diverse igneous origin are present (count 9 in

Table 1). The Dead Goat Conglomerate is dominantly poorly

sorted and only sandstone and chert clasts exceed cobble size.

Accessory amounts of clasts of laminated sandstone–siltstone

occur, which resemble sediments of the Tasman Formation.

The Lockett Conglomerate is dominated by igneous clasts,

which make up almost three-quarters of the clast population

(Table 1). In contrast to the Dead Goat Conglomerate, sandstone

clasts are only a minor contributor (3%). Sorting is very variable

in the Lockett Conglomerate, which Pound (1993) grouped into

eight depositional facies. Several blocks of mafic to intermediate

composition in the Lockett Conglomerate exceed 1 m in dia-

meter, whereas only few granitic clasts reach more than 50 cm.

The clast distribution in the Lockett Conglomerate is remarkably

uniform in different parts of the conglomerate (Table 1).

Petrology of the igneous and sandstone clasts

Igneous clasts in the Dead Goat Conglomerate. Of the various

types of igneous clasts that are present in the Dead Goat

Conglomerate, only two types listed in Table 1 were sufficiently

large to be analysed. The tuffaceous breccia clasts of intermedi-

ate to felsic compositions consist of dark red scoria and brown to

white volcaniclastic fragments with rare plagioclase phenocrysts.

Plutonic clasts are mainly pink metagranitic rocks with the

exception of one fresh dioritic sample. Most metagranitic

samples contain zones of granophyric intergrowth of quartz and

partly metasomatically replaced alkali feldspar, indicative of

either rapid cooling or chilling of the melts (Shelley 1993).

Almost all analysed samples contain secondary calcite, quartz

and chlorite. Comparison of the metamorphic mineral assem-

blage of the matrix and igneous clasts indicates post-depositional

metamorphism of the lower greenschist facies.

Igneous clasts in the Lockett Conglomerate. The majority of

volcanic clasts in the Lockett Conglomerate are effusive rocks of

mafic to intermediate composition. Cryptocrystalline hornblende

and feldspar needles dominate the matrix of the mafic to

intermediate samples, whereas the matrix of the rhyolitic samples

contains mainly cryptocrystalline quartz, an opaque mineral

(presumably magnetite), albitized feldspar and fibrous chlorite.

Phenocryst phases, when present, are hornblende, augite, urali-

tized pyroxene (possibly opx), albitized and/or sericitized plagio-

clase in the mafic to intermediate samples, and quartz in the

rhyolitic samples. The compositions of the plutonic clasts range

from ultramafic to highly differentiated. Original ultramafic

mineral assemblages are fully replaced by sericite pseudomorphs

replacing pyroxene. Mafic to intermediate samples have a

mineral assemblage of hornblende, sericite pseudomorphs and

fibrous secondary actinolite. Intermediate samples contain in-

creasing amounts of quartz, decreasing amounts of hornblende

Table 3. Whole-rock Sm and Nd isotope data for igneous and sandstone clasts of the Lockett and the Dead Goat Conglomerate

Sample Rock type Nd (ppm) Sm (ppm) 143Nd/144Ndmeas147Sm/144Nd ENd500 TDM (Ga)

Dead Goat ConglomerateDG100 Metagranite 45.7 9.36 0.51270 0.124 5.9 0.78DG107 Metagranite 25.2 5.45 0.51260 0.131 3.5 1.02DG146 Metagranite 42.7 8.42 0.51264 0.119 4.9 0.84DG147 Tuffaceous

breccia15.7 3.24 0.51251 0.125 2.1 1.11

DG150 Tuffaceousbreccia

24.5 5.53 0.51255 0.136 2.1 1.20

DG97 Lithic arenite 33.5 6.17 0.51163 0.111 �14.2 2.25DG98 Lithic arenite 28.5 5.18 0.51161 0.110 �14.5 2.25DG148 Greywacke 28.5 5.18 0.51162 0.110 �14.3 2.24Lockett ConglomerateHL48 hb-

metagranodiorite6.95 1.41 0.51236 0.123 �0.7 1.33

HL50 Metagranite 11.8 2.50 0.51214 0.129 �5.4 1.82HL66 Metagranite 12.6 2.54 0.51203 0.122 �7.1 1.86HL84 Metagranite 17.9 4.12 0.51251 0.139 1.1 1.33HL138 Metagranite 14.6 3.73 0.51253 0.154 0.7 1.60HL58 Lithic arenite 23.8 4.36 0.51155 0.111 �15.7 2.36HL120 Lithic arenite 30.9 5.45 0.51155 0.107 �15.4 2.27HL140 Lithic arenite 20.6 3.66 0.51161 0.108 �14.3 2.21

ENd500 was calculated using ºSm ¼ 6:543 3 10�12a�1 with respect to a chondrite uniform reservoir with a present-day 143Nd/144Nd of 0.512638 and 147Sm/144Nd of 0.1966.143Nd/144Nd (LaJolla) ¼ 0:511859 � 23 (2�, n ¼ 28) Depleted mantle model ages TDM were calculated after Goldstein et al. (1984). Classification of the sandstone clasts afterPettijohn (1963). Over a period of 9 months, the La Jolla standard analysed at the University of Munster gave values of 0.511860 � 23 for 143Nd/144Nd (2�, n ¼ 28). Sm/Ndconcentrations were determined using a mixed 149Sm–150Nd tracer to a precision of 1% for the respective concentrations and 0.1% for the elemental ratio. One analysed blankfor Sm and Nd was 55 pg and 320 pg, respectively, and is negligible.


and actinolite, magnetite, calcite, chlorite and accessory epidote.

Virtually all volcanic and plutonic clasts show evidence of

lower greenschist facies metamorphism with the occurrence of

albitized, sericitized, saussuritized and calcitized feldspars, urali-

tized pyroxenes (if present), abundant chlorite and calcite in most

specimens, and epidote in a few samples. The common presence

of sericitized and/or albitized feldspar and uralite suggests a high

availability of water-rich fluids (Shelley 1993), contemporaneous

with only minor strain, as indicated by the lack of preferred

mineral orientations. The well-preserved state of alkali feldspars

in the matrix, compared with the highly altered condition of

alkali feldspars within igneous clasts, as well as the lack of

secondary actinolite in the matrix, suggests that the greenschist

metamorphism was pre-depositional.

Sandstone clast petrography. The sandstone clasts in the Dead

Goat Conglomerate and the Lockett Conglomerate are strikingly

similar in terms of petrography, geochemistry and Nd isotope

composition (see below). They are quartzose, medium-grained,

dark greyish green, well-indurated sandstone clasts that resemble

turbiditic sandstones. Biogenic material is absent. Monocrystal-

line, often undulose quartz grains and alkali felspar dominate the

sandstone clasts. Plagioclase is of minor importance and acces-

sory minerals include detrital muscovite, polycrystalline quartz,

rare micritic limestone fragments, and an opaque phase in

varying proportions. Secondary calcite and chlorite occur

throughout. In the classification scheme of Dickinson et al.

(1983), the sandstone clasts of the two conglomerates suggest a

common, purely continental origin (Fig. 2).

Geochemical compositions of clasts

The igneous clasts have undergone greenschist-facies meta-

morphism during which the large ion lithophile elements (LILE:

Na, K, Rb, Sr, Ba, Cs, U) were potentially mobilized. The

primary total alkali contents, for example, were severely dis-

turbed during the metamorphic event (Fig. 3b). Therefore the

geochemical discussion focuses on the REE, the HFSE (Zr, Nb,

Hf, Ta), and the elements Th, Sc and Y because these are

essentially immobile up to granulite-facies conditions (e.g. Wood

et al. 1976; Humphris & Thompson 1978).

Igneous clasts in the Dead Goat Conglomerate. The tuffaceous

volcanic clasts have an andesitic composition, covering a rela-

tively narrow range in SiO2 contents from 56 to 63 wt% at MgO

contents between 1.5 to 3.9 wt%. Relative to the Lockett

Conglomerate clasts they are enriched in HFSE and REE (Fig.

3). The analysed plutonic clasts have SiO2 contents above

65 wt% except for one diorite (53.3 wt% SiO2, 4.8 wt% MgO).

A prominent feature of the plutonic clasts in the Dead Goat

Conglomerate is the extreme enrichment in HFSE and REE (Figs

3 and 4a, b). Neodymium isotope ratios for the two volcanic and

three metagranitic clasts are presented in Table 3. The two

volcanic clasts have identical ENd500 of +2.1; the metagranitic

clasts have higher ENd500 values of +3.5, +4.9 and +5.9.

Uranium–lead SHRIMP age determinations were carried out on

a set of 15 zircons from one metagranitic clast of the Dead Goat

Conglomerate. The results obtained give a homogeneous age

population of 496 � 6 Ma (2� SE) (Fig. 5 and Supplementary

Publication).

Igneous clasts in the Lockett Conglomerate. Plutonic clasts of

the Lockett Conglomerate cover the entire spectrum from ultra-

mafic to highly felsic compositions, whereas most volcanic clasts

have an intermediate composition (Fig. 3). Incompatible trace

element concentrations for both volcanic and plutonic clasts in

the Lockett Conglomerate are much lower than those for the

igneous clasts in the Dead Goat Conglomerate (Fig. 3c and d).

The majority of the volcanic clasts have SiO2 contents between

53 and 63 wt% (Table 2 and Supplementary Publication).

Because of their high MgO contents (from 11.3 to 5.6 wt%,

respectively) most of the less evolved volcanic clasts can be

classified as boninites and boninitic andesites. Volcanic clasts

intermediate between dacitic and highly felsic compositions are

notably absent in the Lockett Conglomerate (Fig. 3). Three

pebble-sized rhyolitic samples have SiO2 contents between 73

and 75 wt%, but are depleted in incompatible trace elements to

the same extent as the mafic to intermediate volcanic clasts (Fig.

3). The plutonic clasts in the Lockett Conglomerate range in

SiO2 contents from 50 to 78 wt%.

For the ultramafic to intermediate samples, the volcanic and

plutonic clasts have similar MgO and trace element contents

(Fig. 3). Relative to normal mid-ocean ridge basalt (N-MORB)

both the volcanic and plutonic (ultra)mafic to intermediate clasts

are depleted in incompatible elements (Fig. 4c–e). In the

chondrite-normalized multi-element diagram the samples show

either a slight light REE (LREE)-enriched pattern characteristic

of subduction-related melts, or a V-shaped, middle REE

(MREE)-depleted pattern typical of boninitic melts (Fig. 4f–h)

(Crawford 1989). No clear covariation between REE concentra-

tions and LaN/SmN, LaN/YbN or Eu/Eu* with increasing SiO2

concentration can be observed (Table 2), therefore the geochem-

ical variation observed cannot simply reflect different degrees of

melt evolution.

Because of their relatively low Fe-number (i.e. Fe2O3T/

(Fe2O3T + MgO)) for a given SiO2 content the felsic clasts in

the Lockett Conglomerate can be classified as magnesian

metagranitoid rocks (Frost et al. 2001). Samarium–neodymium

results obtained for five metagranitic samples of the Lockett

Conglomerate gave ENd500 of �7.1 to +1.1 (Table 3). In recent

Fig. 2. QFL diagram for two sandstone clasts of the Lockett

Conglomerate (r) and the Dead Goat Conglomerate (d), respectively,

compared with the modal composition of the Benson–Tasman Melange

sandstones and the Junction Formation examined by Pound (1993).

Inferred tectonic settings are from Dickinson et al. (1983). Junction

Formation sandstones contain a significant proportion of lithic fragments,

indicating relative proximity to the Devil River arc during deposition.


intra-oceanic arcs, the range of ENd is generally confined to

values between +2 and +9 (White & Patchett 1984). Whereas the

ENd500 values of +1.1 to �0.7 for three of the measured clasts

could appear realistic for an intra-oceanic setting, values of �5.4

and �7.1 for the remaining two clasts clearly imply a significant

older crustal component.

Sandstone clasts in the Dead Goat Conglomerate and the Lockett

Conglomerate. Following the classification scheme of Bhatia &

Crook (1986), trace element ratios such as Th/Sc and Ti/Zr in the

quartzose sandstone clasts of both conglomerate units indicate an

active continental margin tectonic setting of deposition (Th/Sc

between 1.8 and 2.4; Ti/Zr between 11 and 15) (Bhatia & Crook

1986). ENd500 and depleted mantle model ages (TDM) for Lockett

Conglomerate sandstone clasts range from �15.7 to �14.3 and

from �14.5 to �14.2 for sandstone clasts of the Dead Goat

Conglomerate, corresponding to similar TDM of c. 2.2–2.3 Ga.

Discussion

Petrological, geochemical and isotopic lines of evidence point to

an intra-Devil River arc source for the igneous clasts in the Dead

Goat Conglomerate, as well as for the ultramafic to intermediate

igneous clasts in the Lockett Conglomerate. Conversely, the

metagranitic clasts in the Lockett Conglomerate and the sand-

stone clasts in the two conglomerates are exotic to the Devil

River arc and derived from a different, continental source.

A valuable feature for the identification of clast sources within

the Devil River arc is the previously reported progressive change

of igneous rock compositions through the evolution of the arc

from low-K to high-K compositions (Munker 2000).

Origin of the igneous clasts in the Dead GoatConglomerate

In Figure 6, the compositions of the igneous clasts in the Dead

Goat Conglomerate are compared with volcanic samples from

the Devil River arc (intermediate medium-K and high-K rocks

(Benson Volcanic Suite) and felsic high-K rocks (Snowden

Volcanic Suite) (Munker 2000). For the volcanic clasts, there is a

close compositional similarity to the intermediate medium-K and

high-K Devil River arc lavas, notably the enrichments in HFSE

and REE, particularly LREE (see also Fig. 4b). The Nd isotope

composition of the metavolcanic clasts (ENd500 of 2.1) matches

that reported from medium-K to high-K mineral separates from

within the Devil River arc (ENd500 of 2.1–4.1, Munker 2000).

For the metagranitic clasts in the Dead Goat Conglomerate the

agreement is not as good, as highly differentiated plutonic suites

have not yet been reported from the Devil River arc. The high

ENd500 values (3.5–5.9) of the metagranitic clasts in the Dead

Goat Conglomerate strongly suggest an intra-Devil River arc

origin, and the closest equivalents reported so far are felsic high-

K volcanic rocks emplaced in the latest stages of arc activity (i.e.

Snowden Volcanic Suite, Munker 2000). The extreme enrichment

in incompatible trace elements of the metagranitic clasts (Figs 4

and 6) points to a most feasible origin within the Devil River

arc, produced either contemporary with or subsequent to the

emplacement of the high-K Snowden Volcanic Suite. In compari-

son with the other plutonic clasts analysed from the Dead Goat

Conglomerate, the dioritic sample plotted in Figure 4 (sample

DG102) displays very high incompatible trace element contents.

This clast could possibly be derived from a different within-arc

source from the remaining more felsic plutonic clasts in the

Dead Goat Conglomerate.

Origin of the ultramafic to intermediate clasts in theLockett Conglomerate

The geochemical similarities of the volcanic and plutonic (ultra)-

mafic to intermediate clasts suggest a common petrogenetic

history. Because a direct assessment of the alkalinity of the

igneous clasts in the Lockett Conglomerate cannot be made, Th/

Yb v. Ta/Yb plots are employed because these elemental ratios

mimic the compositional differences observed between low-K

and high-K igneous suites (Fig. 7a; Pearce 1983). The composi-

tional spread of the ultramafic to intermediate clasts in the

Fig. 3. Selected Harker variation diagrams

for clasts of the Dead Goat Conglomerate

compared with samples from the Lockett

Conglomerate. Clasts in the Dead Goat

Conglomerate are significantly enriched in

HFSE and REE (as inferred from the Y

concentration) compared with igneous

clasts in the Lockett Conglomerate.

Samples DG102 and DG150 from the Dead

Goat Conglomerate, which were also

analysed for REE and HFSE, are

highlighted in (c). Most volcanic clasts in

the Lockett Conglomerate have

intermediate compositions, whereas

plutonic clasts range in composition from

ultramafic to highly felsic.


Fig. 4. N-MORB- and C1-normalized multi-element plots for two clasts of the Dead Goat Conglomerate (a, b) and the Lockett Conglomerate (c–h),

compared with several medium-K and high-K volcanic suites of the Devil River Volcanics (Peel, Cobb Flat, Circular Bush, Snowden Volcanic

Conglomerate and Snowden Volcanic Suite formations). (c–e) N-MORB-normalized and (f–h) C1-normalized multi-element plots of mafic to

intermediate clasts of the Lockett Conglomerate (LC) compared with samples of the ‘pre-Devil River’ source, the Cobb Igneous Complex and the low-K

Benson Volcanic Suite. Literature data are from Munker (2000). Average N-MORB composition is from Hofmann (1988); chondrite composition is taken

from Boynton (1984).


Lockett Conglomerate observed in Figure 7 corroborates the

variability illustrated in Figure 4c–h of a mixture of boninitic

and low-K subduction-related signatures. Within the Devil River

arc the low-K arc-related suites, the boninitic ‘pre-Devil River’

rocks, and suites of the boninitic Cobb Igneous Complex

presented by Munker (2000) are geochemically virtually identical

to the ultramafic to intermediate clasts in the Lockett Conglom-

erate (Figs 4 and 7). On the other hand, the geochemical

composition of the medium-K and high-K Devil River arc-

related rocks deviates substantially from that of the Lockett

Fig. 5. Tera–Wasserburg plot for SHRIMP U–Pb zircon data for the

cobble-sized metagranitic clast DG141 of the Dead Goat Conglomerate.

Fig. 6. Trace element variation plots for the plutonic and volcanic clasts

of the Dead Goat Conglomerate (DG), correlated with various medium-K

and high-K rocks of the Benson Volcanic Suite (BV), and five samples of

the high-K Snowden Volcanic Suite (SVS). Samples DG150 and DG102

were also analysed for REE and HFSE contents (see Fig. 4). Correlation

data are from McLean (1994), Funk (1996) and Munker (2000).

Fig. 7. Trace element variation plots for mafic to intermediate volcanic

and plutonic clasts of the Lockett Conglomerate, compared with possible

equivalents and sources. Clasts of the Lockett Conglomerate correlate

best with boninitic samples of the ‘pre-Devil River’ source and the Cobb

Igneous Complex (CIC), and moderately well with low-K Devil River

arc-related samples. Medium-K and high-K Devil River arc-related

volcanic rocks are generally more enriched in these elements.

Comparative data are from Munker (2000). r, Most boninitic sample

analysed by ICP-MS in the Lockett Conglomerate (LC94); m, sample of

the Lockett Conglomerate showing the most typical low-K arc-related TE

pattern (LC123); d, Lockett Conglomerate samples geochemically

intermediate between the two end-member compositions.


Conglomerate clasts. For this reason we propose a boninitic to

low-K Devil River arc-related source as the origin of these

clasts.

Origin of the metagranitic clasts in the LockettConglomerate

The metagranitic clasts in the Lockett Conglomerate are exotic

to the intra-oceanic Devil River arc and require a nearby

continental source area to supply these clasts. The palaeogeogra-

phical position of the Devil River island arc in the Mid- to Late

Cambrian was outboard from the palaeo-Pacific margin of SE

Gondwana. The Australian–Tasmanian–Antarctic section of the

Cambrian palaeo-Pacific margin of Gondwana is the most

proximal continent to the Devil River arc and the metagranitic

clasts are most likely to have been derived from this craton.

Despite the good Nd isotopic agreement between metagranitic

clasts and igneous suites in different sections along this con-

tinental margin (Fig. 8) the observed wide range of Nd isotope

compositions alone does not allow an unambiguous identification

of the source area of these clasts. Moreover, only poor agreement

is observed when using Zr and Y as additional proxies to ENd500

(Fig. 9). On the other hand, data for comparison are scarce and

the source might be either not exposed or no longer preserved.

Origin of the sandstone clasts in the conglomerates

Combining the petrological, trace element and Nd isotope data,

the sandstone clasts in the conglomerates can be described as

being derived from a turbiditic quartzo-feldspathic sandstone

suite deposited adjacent to an old cratonic province in which

early Proterozoic rocks were extensively eroded. The very nega-

tive ENd500 values preclude a significant input of juvenile arc-

related material.

In an earlier study of the Lockett Conglomerate, Pound (1993)

inferred the sandstone clasts to be derived from the Junction

Formation, a petrologically very similar quartzo-feldspathic

sandstone suite exposed in a separate tectonic slice in the Takaka

terrane (Munker & Cooper 1999). The results of the present

study argue against this conclusion. The sandstone clasts

analysed from the Dead Goat Conglomerate and the Lockett

Conglomerate are petrologically more mature than those ana-

lysed by Pound (1993) (Fig. 2). Additionally, the Nd isotope

composition of the sandstone clasts in the Dead Goat Conglom-

erate and the Lockett Conglomerate is less radiogenic (ENd500 in

Fig. 8. Range in ENd500 of felsic clasts examined from the Lockett and Dead Goat conglomerates compared with published data for felsic igneous suites

along the Australian–Tasmanian–Antarctic segment of the Cambrian palaeo-Pacific margin of Gondwana. d, individual results. All correlation samples

depicted here are derived from igneous suites that were emplaced into continental crust. Data from Antarctica are geographically grouped. Literature

sources: (1) Turner et al. (1993); (2) Foden et al. (2002); (3) Whitford et al. (1990); (4) Borg et al. (1987); (5) Armienti et al. (1990); (6) Cox et al.

(2000); (7) Borg et al. (1990); (8) Wareham et al. (2001).


the range of �14.2 to �15.7) than that of the Junction Formation

(ENd500 in the range of �9.7 to �11.2) (Fig. 10; Wombacher &

Munker 2000). No other sandstone suite of appropriate age and

similar source composition has been reported in New Zealand

(see Cooper (1989) and Roser et al. (1996) for an overview).

Hence, the source is either not exposed in present-day New

Zealand, located in the Antarctic–Tasmanian section of the

Cambrian palaeo-Pacific margin of Gondwana, or has been

eroded and is no longer preserved.

In Figure 10, a plot of ENd500 v. Th/Sc is used to discriminate

Devil River arc-related sedimentary suites fed with volcaniclastic

detritus from continental sandstones along the Cambrian palaeo-

Pacific margin of Gondwana, following the approach of Womba-

cher & Munker (2000). Compared with a variety of sandstone

suites from different sections along the palaeo-Pacific continental

margin of Gondwana the sandstone clasts in the Dead Goat and

Lockett conglomerates are the least radiogenic in Nd isotope

composition, reflected in lowest ENd500 (Fig. 10). Although

sandstone sequences of the Kanmantoo Group and the Adelaide

Fold Belt in Australia geochemically and in Nd isotope composi-

tion correlate with sandstones analysed from the Junction Forma-

tion presented by Wombacher & Munker (2000), this agreement

cannot be observed for sandstone clasts of the Dead Goat and

Lockett conglomerates. The sandstone suites of the Wilson and

Bowers terranes in Antarctica (Henjes-Kunst & Schussler 2003)

represent the closest equivalent. However, the sedimentary

sequences in the Bowers terrane are unlikely to represent a

possible source but rather are a probable lateral continuation of

the Devil River arc sequence in present-day New Zealand (see

Weaver et al. 1984; Bradshaw et al. 1985).

Quartzo-feldspathic sandstone suites of Cambrian or older age

are widely exposed along the Antarctic Cambrian palaeo-Pacific

margin (Rowell et al. 2001; Goodge et al. 2002), but published

Nd isotope data are scarce (Borg et al. 1990; Turner et al. 1993;

Cox et al. 2000; Henjes-Kunst & Schussler 2003). The sandstone

clasts in the Dead Goat and Lockett conglomerates were fed

from the Cambrian Gondwana palaeo-Pacific continental margin

but an unequivocal source area cannot be nominated.

Tectonic implications

From consideration of tectonic models proposed in earlier studies

and incorporating the data presented above, our results allow for

a tighter reconstruction of the structural evolution of the

Cambrian palaeo-Pacific margin of SE Gondwana. This region

underwent profound deformation in the course of the Cambrian

Ross–Delamerian orogenic event. Although still under debate,

the favoured structural setting along the Cambrian palaeo-Pacific

margin of Gondwana is that of eastward-dipping subduction in

the SE Australian and Tasmanian segment (e.g. Flottmann et al.

1998; Meffre et al. 2000; Foster et al. 2005) and westward-

dipping subduction in the Antarctic segment of this continental

margin (Kleinschmidt & Tessensohn 1987; Encarnacion &

Grunow 1996; Munker & Crawford 2000). Only a small fraction

of the original island-arc sequence is preserved in New Zealand,

but sufficient information is available to conclude that the Devil

River intra-oceanic island arc was formed outboard of the

Antarctic palaeo-Pacific margin.

The Devil River arc-related sediments of the Haupiri Group

are intruded by dykes of the back-arc tholeiitic Mataki Volcanics

Formation (Fig. 11a) (Munker & Cooper 1999), providing strong

support for the back-arc character of the Haupiri Group sedi-

ments. The Tasman Formation, which hosts the Dead Goat

Conglomerate, is part of the Haupiri Group. Furthermore, at a

few locations the Lockett Conglomerate rests conformably on

sediments of the Tasman Formation (Fig. 1b), indicating that the

sediments of the Haupiri Group and the Lockett Conglomerate

were deposited in the same sedimentary basin. Both the Dead

Goat Conglomerate and the Lockett Conglomerate contain clasts

that are in part derived from the Devil River arc and the

Gondwana continental margin (Fig. 11b and c), therefore the

back-arc basin of the Devil River arc was situated between the

Gondwana continental margin and the Devil River arc. For this

reason subduction underneath the Devil River arc must have been

westward-dipping. To accrete the Devil River intra-oceanic arc to

Fig. 9. Nd isotope ratios of the clasts examined from the Dead Goat and

Lockett conglomerates plotted against Zr and Y concentrations,

compared with several igneous suites from the Cambrian Australian–

Tasmanian–Antarctic palaeo-Pacific margin of Gondwana. No

satisfactory correlation between felsic clasts of the Lockett Conglomerate

and other igneous suites can be observed. The intra-oceanic signature of

igneous clasts deposited in the Dead Goat Conglomerate is reflected in

the offset of this field from those of the other igneous suites. KGSA,

Granites, Kanmantoo Group, South Australia (Turner et al. 1993); NVL,

Granite Harbour Intrusives, Northern Victoria Land (Borg et al. 1987;

Armienti et al. 1990); SVL, Dv1a and Dv1b suites, Southern Victoria

Land (Cox et al. 2000); BGA, granitoids of the Beardmore Glacier Area,

Central Transantarctic Mountains (Borg et al. 1990); LVG, Liv Volcanic

Group, Central Transantarctic Mountains (Wareham et al. 2001).


Fig. 10. ENd500 v. Th/Sc for the sandstone clasts of the Lockett Conglomerate and the Dead Goat Conglomerate compared with the compositions of

sandstones and shales from various sources within the Devil River arc and the Junction Formation (Wombacher & Munker 2000), Sandstone and shale

sequences of the Kanmantoo Group and the Adelaide Fold Belt in SE Australia are shown for comparison (Turner et al. 1993), as well as compositions of

various sandstone sequences in the Bowers terrane (BT), Robertson Bay terrane (RBT) and Wilson terrane in Antarctica (Henjes-Kunst & Schussler

2003). The composition of the sandstone clasts of the Lockett Conglomerate (r) and the Dead Goat Conglomerate (d) are distinct from all other

sedimentary suites shown here. The closest equivalent can be found in a few locations in the Bowers and Wilson terranes (see Henjes-Kunst & Schussler

(2003) for details).

Fig. 11. Simplified schematic sketch map

illustrating the late-stage tectonic evolution

of the Devil River Arc complex and the

final accretion to the Gondwana continental

margin. Munker & Crawford (2000) have

provided a detailed schematic model of the

early stages in the evolution of the Devil

River Arc. (See text for discussion.)


the Gondwana continental margin the back-arc basin between the

arc and the continent had to be closed, either by strike-slip

movement or by initiation of a second subduction zone between

the arc and the continent. Subduction-related magmatism has

been reported for the entire Cambrian Antarctic palaeo-Pacific

margin (see references in Fig. 8 caption), hence the initiation of a

second subduction zone is more feasible (Fig. 11b and c).

The U–Pb SHRIMP age of 496 � 6 Ma for the metagranitic

clast of the Dead Goat Conglomerate (Fig. 5) requires that high-

K igneous suites of the Devil River arc were uplifted and eroded

immediately after emplacement. In addition, the Lockett Con-

glomerate contains boulder-sized ultramafic to intermediate

clasts derived from older sequences of the Devil River arc (see

above). To supply these clasts to the Lockett Conglomerate

significant uplift and erosion was necessary to expose the

boninitic to low-K igneous sequences of the Devil River arc to

erosion (Fig. 11c), whereas at the same time the Antarctic

continental margin was close enough to supply boulder-sized

Andean-type metagranitic clasts. Accretion ended with the

emplacement of the diapiric Balloon Melange in the Latest

Cambrian (Fig. 11d; see Jongens et al. 2003).

We cannot pinpoint an exact location for the accretion of the

Devil River arc to the Antarctic continental margin. However, the

polarity of subduction of the different segments along the SE

Gondwana Cambrian palaeo-Pacific margin (see above), the

timing of events recorded by the Granite Harbour Intrusives in

Antarctica (e.g. Encarnacion & Grunow 1996), and the spatial

and temporal distribution of deformational phases along the

Transantarctic Mountains (e.g. Rowell et al. 2001) make the area

of Southern Victoria Land or the northern Central Transantarctic

Mountains the most likely location of accretion of the Devil

River arc to the Cambrian palaeo-Pacific Gondwana margin.

Conclusions

Conglomerate clasts from Cambrian conglomerates in New

Zealand provide important information relevant to tectonic events

in SE Gondwana close to the Mid-Cambrian–Late Cambrian

boundary. A back-arc basin (documented by the tholeiitic Mataki

Volcanics Formation) separated the intra-oceanic Devil River arc

from the Gondwana margin until the Mid-Cambrian. This back-

arc basin was subsequently closed by the onset of a second

subduction system along the Gondwana margin, allowing Gond-

wana-derived material to be deposited in the intra-arc basins by

the end of the Mid-Cambrian. The overlying conglomerates of

late Mid-Cambrian age contain granitoids that were probably

derived from the Ross–Delamerian continental margin and were

deposited in a marginal marine to fluvial setting. The coarse

sizes of this debris (.50 cm) imply a continuous subaerial slope

of moderate gradient from continental margin to arc basin. The

most likely position of accretion of the Devil River intra-oceanic

arc to the Cambrian palaeo-Pacific margin was in the area of

Southern Victoria Land or the northern Central Transantarctic

Mountains.

The Brian Mason Trust of the University of Canterbury is thanked for

providing financial support for the fieldwork, the geochemical analysis and

the SHRIMP dating. K. Mezger (ZLG Munster) generously provided access

to the ZLG research facilities for the Sm–Nd isotope analysis. D. Garbe-

Schonberg in Kiel (Germany) facilitated REE and HFSE analyses. SHRIMP

measurements were conducted at the Australian National University in

Canberra. The research of M.G. was supported by a University of Canter-

bury’s Master’s Scholarship. The Department of Conservation, New

Zealand, provided free accommodation facilities during the fieldwork in

Cobb Valley, NW Nelson. We are grateful to T. Waight and J. Foden for their

constructive reviews of the manuscript, and D. Peate is thanked for editorial

handling and additional constructive reviews.

References

Armienti, P., Ghezzo, C., Innocenti, F., Manetti, P., Rocchi, S. & Tonarini,

S. 1990. Isotope geochemistry and petrology of granitoid suites from Granite

Harbour Intrusives of the Wilson Terrane, Northern Victoria Land, Antarctica.

European Journal of Mineralogy, 2, 103–123.

Bhatia, M.R. & Crook, K.A.W. 1986. Trace element characteristics of

graywackes and tectonic setting discrimination of sedimentary basins.

Contributions to Mineralogy and Petrology, 92, 181–193.

Borg, S.G., Stump, E., Chappell, B.W., McCulloch, M.T., Wyborn, D.,

Armstrong, R.J. & Holloway, J.R. 1987. Granitoids of Northern Victoria

Land, Antarctica: implications of chemical and isotopic variations to regional

crustal structure and tectonics. American Journal of Science, 287, 129–169.

Borg, S.D., DePaolo, D.J. & Smith, B.M. 1990. Isotope structure and tectonics of

the Central Transantarctic Mountains. Journal of Geophysical Research,

95(B5), 6647–6667.

Boynton, W.V. 1984. Cosmochemistry of the rare earth elements: meteorite

studies. In: Henderson, P. (ed.) Rare Earth Element Geochemistry. Elsevier,

Amsterdam, 89–94.

Bradshaw, J.D., Weaver, S.D. & Laird, M.C. 1985. Suspect terranes and

Cambrian tectonics in Northern Victoria Land, Antarctica. In: Howell, D.C.

(ed.) Tectonostratigraphic Terranes of the Circum-Pacific Region. Circum-

Pacific Council for Energy and Mineral Resources, Houston, TX, 467–479.

Coleman, A.C. 1978. Composition of the Lockett Conglomerate Lake Sylvester

Area, northwest Nelson. New Zealand Journal of Geology and Geophysics,

21, 549–552.

Cooper, R.A. 1989. Early Paleozoic terranes in New Zealand. Journal of the Royal

Society of New Zealand, 19, 73–112.

Cooper, R.A. 2004. New Zealand Geological Timescale 2004/2 Wallchart. Institute

of Geological & Nuclear Sciences Information Series, 64.

Cooper, R.A. & Tulloch, A.J. 1992. Early Paleozoic terranes in New Zealand and

their relationship to the Lachlan Fold Belt. Tectonophysics, 214, 129–144.

Cox, S.C., Parkinson, D.L., Allibone, A.H. & Cooper, A.F. 2000. Isotopic

character of Cambro-Ordovician plutonism, Southern Victoria Land, Antarc-

tica. New Zealand Journal of Geology and Geophysics, 43, 501–520.

Crawford, A.J. (ed.) 1989. Boninites. Unwin Hyman, London.

Davidek, K., Landing, E., Bowring, S.A., Westrop, S.R., Rushton, A.W.A.,

Fortrey, R.A. & Adrain, J.M. 1998. New uppermost Cambrian U–Pb date

from Avalonian Wales and age of the Cambro-Ordovician boundary.

Geological Magazine, 135, 305–309.

Dickinson, W.R., Beard, L.S., Brakenridge, G.R., Erjavec, J.L. & Fergusen,

R.C. 1983. Provenance of North American Phanerozoic sandstones in relation

to tectonic setting. Geological Society of America Bulletin, 94(2), 222–235.

Encarnacion, J. & Grunow, A. 1996. Changing magmatic and tectonic styles

along the paleo-Pacific margin of Gondwana and the onset of Early Paleozoic

magmatism in Antarctica. Tectonics, 15(6), 1325–1341.

Encarnacion, J., Rowell, A.J. & Grunow, A. 1999. A U–Pb age for the

Cambrian Taylor Formation, Antarctica: implications for the Late Cambrian

time scale. Journal of Geology, 107, 497–504.

Flottmann, T., Haines, P., Jago, J., James, P., Belperio, A. & Gum, J. 1998.

Formation and reactivation of the Cambrian Kanmantoo Trough, SE

Australia: implications for early Palaeozoic tectonics at eastern Gondwana’s

plate margin. Journal of the Geological Society, London, 155, 525–539.

Foden, J.D., Elburg, M.A., Turner, S.P., Sandiford, M., O’Callaghan, J.O.

& Mitchell, S. 2002. Granite production in the Delamerian orogen, South

Australia. Journal of the Geological Society, London, 159, 557–575.

Foster, A.F., Gray, D.R. & Spaggiari, C. 2005. Timing of subduction and

exhumation along the Cambrian East Gondwana margin, and the formation of

Paleozoic backarc basins. Geological Society of America Bulletin, 117, 105–

116.

Frost, B.R., Barnes, C.G., Collins, W.J., Arculus, R.J., Ellis, D.J. & Frost,

C.D. 2001. A geochemical classification for granitic rocks. Journal of

Petrology, 42, 2033–2048.

Funk, T. 1996. Geologische und geochemische Untersuchung an kambrischen

Vulkaniten Neuseelands. Diploma thesis, University of Gottingen.

Garbe-Schonberg, C.-D. 1993. Simultaneous determination of thirty-seven trace

elements in twenty-eight international rock standards by ICP-MS. Geostan-

dards Newsletter, 17, 81–97.

Goldstein, S.L., O’Nions, R.K. & Hamilton, P.J. 1984. A Sm–Nd study of

atmospheric dusts and particulates from major river systems. Earth and

Planetary Science Letters, 70, 221–236.

Goodge, J.W., Myrow, P., Williams, I.S. & Bowring, S.A. 2002. Age and


provenance of the Beardmore Group, Antarctica: constraints on Rodinia

supercontinent breakup. Journal of Geology, 110, 393–406.

Grindley, G.W. 1980. Sheet S13 Cobb, 1st. Geological Map of New Zealand

1:63 360, Department of Scientific and Industrial Research, Wellington.

Henjes-Kunst, F. & Schussler, U. 2003. Metasedimentary units of the Cambro-

Ordovician Ross orogen in Northern Victoria Land and Oates Land:

implications for the provenance and geotectonic setting from geochemical

and Nd–Sr isotope data. Terra Antartica, 10(3), 105–128.

Hofmann, A.W. 1988. Chemical differentiation of the Earth: the relationship

between mantle, continental crust and oceanic crust. Earth and Planetary

Science Letters, 90, 297–314.

Humphris, S.E. & Thompson, G. 1978. Hydrothermal alteration of oceanic basalts

by seawater. Geochimica et Cosmochimica Acta, 42, 107–125.

Jongens, R., Bradshaw, J.D. & Fowler, A. 2003. The Balloon Melange,

northwest Nelson: origin, structure and emplacement. New Zealand Journal

of Geology and Geophysics, 46, 437–448.

Kleinschmidt, G. & Tessensohn, F. 1987. Early Paleozoic westward directed

subduction at the Pacific margin of Antarctica. In: McKenzie, G.D. (ed.)

Gondwana Six: Structure, Tectonics, and Geophysics. American Geophysical

Union, Washington D.C. 89–105.

Landis, C.A. & Coombs, D.S. 1967. Metamorphic belts and orogenesis in Southern

New Zealand. Tectonophysics, 4, 501.

Le Bas, M.J. 2000. IUGS reclassification of the high-Mg and picritic volcanic

rocks. Journal of Petrology, 41, 1467–1470.

McLean, D.R. 1994. The geology and geochemistry of the Cambrian Devil River

Volcanics, Anatoki Range, northwest Nelson. Master’s thesis, University of

Canterbury, Christchurch.

Meffre, S., Berry, R.F. & Hall, M. 2000. Cambrian metamorphic complexes in

Tasmania: tectonic implications. Australian Journal of Earth Sciences, 47,

971–985.

Muir, R.J., Ireland, T.R., Weaver, S.D. & Bradshaw, J.D. 1996. Ion microp-

robe dating of Paleozoic granitoids: Devonian magmatism in New Zealand

and correlations with Australia and Antarctica. Chemical Geology, 127,

191–210.

Munker, C. 2000. The isotope and trace element budget of the Cambrian Devil

River Arc system, New Zealand: identification of four source components.

Journal of Petrology, 41, 759–788.

Munker, C. & Cooper, R.A. 1995. The island arc setting of a New Zealand

Cambrian volcano-sedimentary sequence: implications for the evolution of

the southwest Pacific Gondwana fragments. Journal of Geology, 103, 687–

700.

Munker, C. & Cooper, R.A. 1999. The Cambrian arc complex of the Takaka

Terrane, New Zealand: an integrated stratigraphical, paleontological and

geochemical approach. New Zealand Journal of Geology and Geophysics, 42,

415–445.

Munker, C. & Crawford, A.J. 2000. Cambrian arc evolution along the SE

Gondwana active margin: a synthesis from Tasmania–New Zealand–Austra-

lia–Antarctica correlations. Tectonics, 19, 415–432.

Norrish, K. & Hutton, J.T. 1969. An accurate X-ray spectrographic method for

the analysis of a wide range of geologic samples. Geochimica et Cosmochi-

mica Acta, 33, 431–454.

Pearce, J.A. 1983. The role of sub-continental lithosphere in magma genesis at

destructive plate margins. In: Hawkesworth, C.J. & Norris, M.J. (eds)

Continental Basalts and Mantle Xenoliths. Shiva, Nantwich, 230–249.

Pettijohn, F.J. 1963. Chemical Composition of Sandstones. US Geological Survey,

Professional Papers, N-40S.

Pound, K.S. 1993. Geology of the Lower Paleozoic Haupiri Group rocks, Cobb

Valley area, northwest Nelson, New Zealand. PhD thesis, University of Otago,

Dunedin.

Rattenbury, M.S., Cooper, R.A. & Johnston, M.R. 1998. Geology of the Nelson

area. Geological map 9 (1:250 000). Institute of Geological and Nuclear

Sciences, Lower Hutt.

Roser, B.P., Cooper, R.A., Nathan, S. & Tulloch, A.J. 1996. Reconnaissance

sandstone geochemistry, provenance, and tectonic setting of the lower

Paleozoic terranes of the Weast Coast and Nelson, New Zealand. New

Zealand Journal of Geology and Geophysics, 39, 1–16.

Rowell, A.J., Van Schmus, W.R., Storey, B.C., Fetter, A.H. & Evans, K.R.

2001. Latest Neoproterozoic to mid-Cambrian age for the main deformational

phases of the Transantarctic Mountains: new stratigraphic and isotopic

constraints from the Pensacola Mountains, Antarctica. Journal of the

Geological Society, London, 158, 295–308.

Shelley, D. 1993. Igneous and Metamorphic Rocks under the Microscope:

Classification, Textures, Microstructures, and Mineral-preferred Orientations.

Chapman & Hall, London.

Siebert, C. 1998. Geology and geochemistry of metavolcanic rocks in the Baton

Valley area, northwest Nelson. Diploma thesis, Georg August University,

Gottingen.

Turner, S., Foden, J., Sandford, M. & Bruce, D. 1993. Sm–Nd evidence for

the provenance of sediments from the Adelaide Fold Belt and southeastern

Australia with implications for episodic crustal addition. Geochimica et

Cosmochimica Acta, 57, 1837–1856.

Wareham, C.D., Stump, E., Storey, B.C., Millar, I. & Riley, T.R. 2001.

Petrogenesis of the Cambrian Liv Group, a bimodal volcanic rock suite from

the Ross Orogen, Transantarctic Mountains. Geological Society of America

Bulletin, 113(3), 360–372.

Weaver, S.D., Bradshaw, J.D. & Laird, M.G. 1984. Geochemistry of Cambrian

volcanics of the Bowers Supergroup and implications for the early Paleozoic

tectonic evolution of northern Victoria Land, Antarctica. Earth and Planetary

Science Letters, 68, 128–140.

White, W.M. & Patchett, P.J. 1984. Hf–Nd–Sr isotopes and incompatible trace

element abundances in island arcs: implications for magma origin and crust–

mantle evolution. Earth and Planetary Science Letters, 67, 167–185.

Whitford, D., Crawford, T., Korsch, M. & Craven, S. 1990. Mt. Read

Volcanics: strontium and neodymium isotope geochemistry.In: IAVCEI

Abstracts: Ancient Volcanism and Modern Analogues. 1, 122.

Wombacher, F. & Munker, C. 2000. Pb, Nd, and Sr isotopes and REE

systematics of Cambrian sediments from New Zealand: implications for the

reconstruction of the Early Paleozoic Gondwana margin along Australia and

Antarctica. Journal of Geology, 108, 663–686.

Wood, D.A., Gibson, I. & Thompson, R.N. 1976. Element mobility during zeolite

facies metamorphism of the Tertiary basalts of eastern Iceland. Contributions

to Mineralogy and Petrology, 55, 241–254.

Received 27 January 2005; revised typescript accepted 23 March 2006.

Scientific editing by David Peate


Provenance of Cambrian conglomerates from New Zealand...

Documents

Transcript of Provenance of Cambrian conglomerates from New Zealand...