Boomerang Seamount: the active expression of the Amsterdam^St. Paul hotspot, Southeast Indian Ridge

Boomerang Seamount: the active expression of theAmsterdam^St. Paul hotspot, Southeast Indian Ridge

K.T.M. Johnson a;*, D.W. Graham b, K.H. Rubin c, K. Nicolaysen d,D.S. Scheirer e, D.W. Forsyth e, E.T. Baker f , L.M. Douglas-Priebe b

a Department of Geology, Bishop Museum, Honolulu, HI 96817, USAb College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, OR 97331, USA

c Department of Geology and Geophysics, University of Hawaii, Honolulu, HI 96822, USAd Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA

e Department of Geological Sciences, Brown University, Providence, RI 02912, USAf Paci¢c Marine Environmental Laboratory, National Oceanic and Atmospheric Administration, Seattle, WA 98115, USA

Received 15 September 1999; received in revised form 18 August 2000; accepted 19 September 2000

Abstract

During a survey of the axis of the Southeast Indian Ridge (SEIR), we discovered a 1100 m tall, volcanically activesubmarine volcano, Boomerang Seamount, near the spreading center on the Amsterdam^St. Paul (ASP) Plateau. Thesummit of the volcano is 650 m below sea level and has a 200-m-deep, 2-km-wide circular caldera. Samples of very freshvolcanic glass, dated by the 210Po^210Pb technique at V5 months old, were collected from the floor of the caldera inMarch 1996. The volcano is 18 km northeast of Amsterdam Island near the intersection of a long spreading segmentand the Boomerang Transform Fault. It is built on the stationary Antarctic Plate, where widely scattered volcanicactivity thickens the crust, continuing to build the plateau. Water column profiles reveal a 1.7³C temperature anomalyand a 0.3 V stepped nephelometer (water column turbidity) anomaly within the caldera, nearly an order of magnitudelarger than other hydrothermal plume anomalies we measured. These anomalies suggest hydrothermal activity withinthe caldera. Volcanic glass compositions at two sample sites on the volcano summit are similar to each other and toAmsterdam and St. Paul Island basalts, but have some important differences as well. K2O/TiO2 ratios in BoomerangSeamount glasses are similar to St. Paul Island samples, but differ significantly from Amsterdam Island samples. Rareearth element patterns in lavas from Boomerang, Amsterdam, and St. Paul are similar. Sr, Nd, and Pb isotope ratios insamples from the Boomerang Caldera floor are similar to samples from Amsterdam Island. However, another samplefrom Boomerang Seamount deviates from a SEIR^St. PaulÂmsterdam mixing trend and shows evidence of mixingwith a Kerguelen-like source component. The geochemical complexity of these three closely spaced volcanic edifices onthe ASP Plateau suggests that the Boomerang Seamount source is heterogeneous on a very small scale. ß 2000Elsevier Science B.V. All rights reserved.

Keywords: mantle plumes; hot spots; Southeast Indian Ridge; hydrothermal conditions; Amsterdam Island; Saint Paul Island

0012-821X / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 0 0 ) 0 0 2 7 9 - X

* Corresponding author. Tel. : +1-808-848-4124; Fax: +1-808-847-8252; E-mail: [email protected]

EPSL 5633 3-11-00 Cyaan Magenta Geel Zwart

Earth and Planetary Science Letters 183 (2000) 245^259

www.elsevier.com/locate/epsl

1. Introduction

The Amsterdam^St. Paul (ASP) Plateau is anoceanic platform that rises 1000^1500 m abovethe adjacent sea£oor and sits atop the SoutheastIndian Ridge (SEIR) (Fig. 1). The SEIR itselfextends 6000 km from the Rodrigues Triple Junc-tion to the Macquarie Triple Junction and is oneof three major spreading centers in the IndianOcean. Volcanic construction on the ASP Plateauis widespread; it occurs at the SEIR spreadingcenter, in some di¡use o¡-axis areas, and at edi-¢ces such as Amsterdam and St. Paul Islands [1].A 53-day expedition to this region in FebruaryÂpril 1996 on the R/V Melville (Boomerang Ex-pedition, Leg 6) mapped some 1400 km of theSEIR with 75^100% Seabeam coverage andsampled along the spreading center at an averagedredge and core spacing of 14 km on either sideof the ASP Plateau, and 10 km on the ASP Pla-teau [2,3].

During the course of the survey, we discoveredan 1100-m-high seamount rising to a depth of 650m below sea level 18 km northeast of AmsterdamIsland and 120 km north of St. Paul Island (Fig.1b). The seamount is surmounted by a circular

summit caldera 2 km in diameter and 200 mdeep. The summit region is highly re£ective toside-scan sonar, indicating small-scale roughnessof the sea£oor that is characteristic of young, un-sedimented lava £ows [1]. Extremely fresh basaltglass was recovered from two sites on the sea-mount summit. The seamount appears to haveat least two 40-km-long rift zones extending northand southeast from its summit, giving it an arcu-ate shape in plan view. For this reason, and be-cause of the expedition's designation, the featurewas named Boomerang Seamount. In this paper,we describe the morphological, geochemical, andhydrothermal characteristics of Boomerang Sea-mount and discuss it in the context of mantleplume activity in the region.

2. Geologic setting

Australia and Antarctica began to separatearound 110 Myr, but sea£oor spreading alongwhat has become the SEIR began during the Eo-cene. This caused the Kerguelen Plateau and Bro-ken Ridge to separate at the site of the Kerguelenhotspot, now located V1400 km southwest of the

Fig. 1. (a) The Indian Ocean, its three main spreading centers (double lines), and selected hotspot features. Contour interval is1 km. Thick line is the Boomerang Leg 6 cruise track and the shaded box is coverage area of (b). 90³E ^ Ninetyeast Ridge; ASP^ Amsterdam^St. Paul Plateau; BR ^ Broken Ridge; KP ^ Kerguelen Plateau; SWIR ^ Southwest Indian Ridge; RTJ ^ 16 Ro-drigues Triple Junction; CIR ^ Central Indian Ridge; SEIR ^ Southeast Indian Ridge. (b) Blow-up of the ASP Plateau showing2000 m depth contour used to de¢ne the Plateau (dark gray), track line (dashed line), mapped ridge segments (double lines), frac-ture zones (heavy lines with italicized names), and the locations of Amsterdam and St. Paul Islands and Boomerang Seamount(star).


K.T.M. Johnson et al. / Earth and Planetary Science Letters 183 (2000) 245^259246

ASP 2 Plateau (Fig. 1) [4^7]. Since then, the SEIRhas been migrating to the northeast away fromthe hotspot and currently spreads at a full-rateof V65 mm/yr. Because the Antarctic Plate isessentially ¢xed in the hotspot reference frame[8], the Australian Plate migrates over the deepmantle at the total opening rate of the SEIRand the ridge migrates northeastward at approx-imately half that rate (32^33 mm/yr). The ASPhotspot was originally located beneath the Aus-tralian Plate, and it may have contributed to theformation of Ninetyeast Ridge until separation ofBroken Ridge and Ninetyeast Ridge from theKerguelen Plateau at about 40 Ma [9^11]. Satel-lite altimetry [12] shows a chain of seamountsconnecting the southern tip of the NinetyeastRidge to the northern edge of the ASP Plateau.This seamount chain appears to delineate the ASPhotspot track on the Australian Plate subsequentto the breakup of the Kerguelen Plateau and Bro-ken Ridge. Since the SEIR passed over the ASPhotspot 3^5 million years ago [13], the hotspotcurrently lies beneath the nearly stationary Ant-arctic Plate southwest of the SEIR. Its stationarylocation has allowed the formation of the shal-lowest parts of the ASP Plateau through contin-ued activity at island and seamount volcanoes.

3. Results

3.1. Seamount morphology

The currently most active surface expression ofthe ASP hotspot may be the Boomerang Sea-mount. The Boomerang Seamount lies nearly atthe intersection of a long, straight spreading seg-ment that bisects the ASP Plateau and a trans-form fault near the northwest edge of the plateau(Fig. 2a). This ridge segment (I2) shows high 3He/4He ratios (s 9 RA), a diagnostic characteristic ofthe ASP hotspot [3]. The southern section of seg-ment I2 (I2S) appears to terminate just to thesoutheast of the seamount, sidestepping the sum-mit through a short, 10-km-long spreading seg-ment on the northeast £ank of the seamount(I2N) that links to the Boomerang TransformFault (Fig. 2a). The southeast rift zone of Boo-

merang Seamount is parallel to the trend of seg-ment I2S and its orientation may be controlled byeither the pre-existing structure developed at thespreading axis or by the regional stress ¢eld con-trolling the orientation of the ridge axis.

Within the arc formed by the two main northand southeast rift zones is another shallow edi¢ce,smaller than the main edi¢ce, that possesses ashort NNE-trending ridge of its own. This smalleredi¢ce is located at the projected intersection ofsegment I2S and the Boomerang TransformFault. The base of Boomerang Seamount occursover a range of depths from 1600 m to 2000 m,thereby giving the seamount a range of base tosummit heights from 950 m to 1350 m; for sim-plicity in this paper, we have chosen 1100 m as anaverage height of the seamount. Using these pa-rameters, the volume of Boomerang Seamount iscalculated to be V300 km3, which when com-bined with the approximate age of the sea£ooron which it sits of 700 000 yr yields a minimumgrowth rate of V0.0004 km3/yr, or about 1.5 or-ders of magnitude smaller than the growth ratefor Mauna Loa volcano in Hawaii.

The surfaces of the volcanic edi¢ces are rela-tively smooth, with evidence for small satellitecones away from the summit area and lobe-shaped topographic plateaus and terraces down-slope of the ridge axes (Fig. 2a). The summit ofBoomerang Seamount contains a 2-km-wide cal-dera with a nearly constant 200 m depth and avolume of V0.63 km3. Within the SW section ofthe caldera, several smaller, nested collapse fea-tures can be seen. Small, discontinuous ring faultscarps can be clearly seen 1^2 km west of thecaldera (Fig. 2b). They may also be present inthe southeast sector of the summit where the mul-tibeam bathymetry data are less clear or incom-plete. The steep slopes of the caldera walls, gen-erally greater than 45³, suggest an origin byfaulting and subsequent mass-wasting ratherthan by explosive excavation or by build-up ofbasaltic lava £ows along ring fractures. Likewise,the near-circularity of the walls and their nearlyconstant height suggest formation of the calderaby down-drop of the center along a continuousring fault system, the upper reaches of whichnow form the caldera walls.


K.T.M. Johnson et al. / Earth and Planetary Science Letters 183 (2000) 245^259 247

3.2. Geochemistry

3.2.1. Basaltic glassVolcanic glass samples were collected from the

£oor of the summit caldera (wax core sampleWC44) and from just outside the caldera rim(WC45) (Fig. 2). Samples were collected using asingle-head wax coring device with a wax headdiameter of approximately 30 cm and a weightof approximately 315 kg. At these two sites, adouble hit technique was used to improve samplerecovery: after the initial impact at 90 m/min low-ering rate, the device was raised to 100 m above

the sea£oor, allowed to steady in the water col-umn, and lowered again at 90 m/min for a secondimpact. Although these cores were not located bytransponders, experience with transponder-navi-gated wax cores elsewhere suggests that the areaof sea£oor sampled using this double hit tech-nique is approximately 8^10 m in diameter. Sam-ple recovery in WC44 was V50 g of glass and inWC45 V250 g of glass and V150 g of red mud.

Rock compositions were determined on glasschips recovered from the wax core. The glass sam-ples are highly vesicular, very fresh and appear tobe quite young. Preliminary chemical groups were

Fig. 2. (a) Bathymetry of Boomerang Seamount region. Areas that are contoured at 200 m intervals are covered by Seabeam2000 data. Other areas are interpolated or controlled by other shipboard bathymetric measurements or satellite gravity data.Thick white lines represent ridge segments. (b) Top panel is Seabeam 2000 bathymetry (ping data) illuminated from the north,with wax core stations 44 and 45 plotted as blue stars. Bottom panels are bathymetric cross-sections of the caldera along sectionAÂP in the top panel, plotted at ¢ve times and one time vertical exaggerations.



determined on the basis of major element analysesof randomly selected glass chips from each core(WC44a^c and WC45a^h; Table 1) by electronmicroprobe. Each major element analysis in Table1 represents the average of at least three probepoints along a transect of a single chip of glass.Analyses with totals outside of 98^101% were ex-cluded. On this basis, we identi¢ed one chemicalgroup in WC44 and two groups in WC45. Theglasses are tholeiitic to transitional in major ele-ment composition and are similar to some St.Paul Island lavas. The glasses di¡er signi¢cantly,however, from the more proximal lavas erupted atAmsterdam Island, for example in K/Ti at a givenMg# (Fig. 3). Moreover, there are signi¢cantcompositional di¡erences between the two chem-ical groups recovered in WC45 (Table 1 and Fig.3), although these di¡erences may result from

shallow-level fractional crystallization of olivineand plagioclase þ clinopyroxene. The WC44glasses are similar in major element compositionto one of the groups in WC45, but are slightlymore evolved. New analyses of glasses from twodikes on St. Paul Island are also reported in Table1.

Rare earth element (REE) concentrations andpatterns in WC44 and WC45 glasses are similar tothose in basalts from St. Paul Island. The onlytwo rocks from Amsterdam Island with REEanalyses and two other rocks from St. Paul Islandhave positive Eu anomalies and appear to be pla-gioclase cumulates, so are not directly comparableto the glasses from Boomerang Seamount (Fig.4a).

Isotopic compositions of 87Sr/86Sr, 143Nd/144Nd,Fig. 3. K2O/TiO2 versus Mg# (Mg/[Mg+Fe*]U100) for Boo-merang Seamount samples and samples from Amsterdamand St. Paul Islands. Fields for basalts from the two islandsare drawn around samples with SiO2 9 51% for comparisonwith Boomerang Seamount data. The open triangles are datafor dikes from the northeast-facing scarp on the southernside of the breached caldera on St. Paul (this study and[17,32]). The three small squares are hawaiites from St. Paulwith SiO2 s 54% [32,33,41]. The two chemical groups inWC45 and their relationship with WC44 and some St. Paulsamples are shown. 2c error bars are shown for the Boomer-ang Seamount samples. All Fe is calculated as Fe2�. Datasources: [17,30^33,41]; F. Frey and D. Weis, unpublisheddata for Amsterdam Island; W. White, unpublished data forAmsterdam and St. Paul Islands.

Fig. 4. (a) Chondrite-normalized REE plot of samples fromBoomerang Seamount, Amsterdam and St. Paul Islands.Two St. Paul samples and the only two samples from Am-sterdam Island have lower REE abundances and positive Euanomalies suggestive of plagioclase accumulation. Data sour-ces: [17,18,21]; F. Frey and D. Weis, unpublished data forAmsterdam Island; W. White, unpublished data for Amster-dam and St. Paul Islands. Chondrite normalization values ofAnders and Grevesse [42]. (b) REE concentrations of thesediment recovered in WC45 normalized to the average ofthe two WC45 glass analyses. Note that the y-axis is linear.Propagated error bar is shown.



Tab

le1

Maj

oran

dtr

ace

elem

ent

com

posi

tion

ofgl

asse

sfr

omB

oom

eran

gSe

amou

ntan

dSt

.P

aul

Isla

nd

Boo

mer

ang

Seam

ount

St.

Pau

lIs

land

WC

44aa

WC

44b

WC

44c

WC

45a

WC

45ba

WC

45ca

WC

45d

WC

45e

WC

45f

WC

45g

WC

45h

WC

45se

dSP

-1aa

SP-5

ba

Lat

(³S)

37.7

2137

.721

37.7

2137

.716

37.7

1637

.716

37.7

1637

.716

37.7

1637

.716

37.7

1637

.716

38.7

238

.72

Lon

g(³

E)

77.8

2577

.825

77.8

2577

.832

77.8

3277

.832

77.8

3277

.832

77.8

3277

.832

77.8

3277

.832

77.5

577

.55

Dep

th(m

)87

587

587

564

764

764

764

764

764

764

764

764

7Si

O2

50.7

350

.98

51.0

750

.76

50.8

050

.99

50.9

850

.86

50.3

550

.55

50.3

950

.63

50.2

9T

iO2

2.63

2.69

2.60

2.12

2.11

2.53

2.14

2.55

2.13

2.54

2.16

3.29

3.23

Al 2

O3

13.5

413

.50

13.6

014

.26

14.0

913

.78

14.2

513

.66

14.0

713

.67

13.9

90.

4813

.37

13.3

3F

eO*

13.1

113

.30

13.0

711

.67

11.6

013

.09

11.9

813

.36

11.9

513

.23

11.8

449

.24

14.1

813

.92

MnO

0.22

0.20

0.24

0.21

0.21

0.22

0.18

0.18

0.19

0.22

0.20

0.04

90.

240.

22M

gO4.

664.

594.

715.

885.

814.

795.

964.

785.

804.

805.

780.

724.

494.

45C

aO9.

299.

229.

3610

.52

10.5

09.

2810

.66

9.43

10.5

69.

2810

.56

1.11

8.91

8.78

Na 2

O3.

093.

003.

062.

802.

823.

022.

933.

192.

833.

192.

903.

253.

14K

2O

0.75

0.78

0.76

0.61

0.62

0.74

0.62

0.76

0.62

0.75

0.63

0.99

0.97

P2O

50.

310.

310.

300.

250.

220.

300.

280.

300.

230.

300.

260.

420.

410.

45T

otal

98.3

398

.58

98.7

899

.10

98.7

998

.75

100.

0099

.09

98.7

698

.56

98.7

199

.76

98.7

7M

g/(M

g+F

e*)

0.38

790.

3809

0.39

110.

4730

0.47

180.

3949

0.47

030.

3897

0.46

390.

3929

0.46

550.

3606

0.36

31K

2O

/TiO

20.

287

0.28

90.

294

0.28

80.

292

0.29

10.

290

0.29

70.

292

0.29

70.

291

0.30

10.

301

Ti

23.4

V38

610

5C

r17

12C

o41

8.5

Ni

227.

6C

u57

136

Zn

118

481

As

1.8

44R

b15

.72.

9Sr

264

206

Y33

9.0

Zr

175

33N

b28

4.9

Mo

1.03

13C

s0.

198

0.06

7B

a20

571

La

20.4

19.8

16.6

17.6

5.7

21.2

20.4

Ce

44.4

41.5

36.0

38.7

10.4

48.6

46.6

Pr

5.2

1.3

Nd

21.3

23.4

18.0

20.2

6.0

24.7

24.0

Sm5.

15.

84.

85.

21.

46.

55.

6E

u1.

90.

463.

22.

9



Tab

le1

(con

tinu

ed)

Maj

oran

dtr

ace

elem

ent

com

posi

tion

ofgl

asse

sfr

omB

oom

eran

gSe

amou

ntan

dSt

.P

aul

Isla

nd

Boo

mer

ang

Seam

ount

St.

Pau

lIs

land

WC

44aa

WC

44b

WC

44c

WC

45a

WC

45ba

WC

45ca

WC

45d

WC

45e

WC

45f

WC

45g

WC

45h

WC

45se

dSP

-1aa

SP-5

ba

Gd

6.4

1.8

Tb

0.94

0.27

Dy

4.9

6.5

5.4

5.7

1.7

6.6

5.9

Ho

1.2

0.32

Er

3.5

3.6

3.6

4.1

0.94

4.2

4.0

Tm

0.49

0.14

Yb

2.4

3.5

2.7

3.4

1.0

4.2

4.0

Lu

0.44

0.20

Hf

4.9

1.1

Ta

1.9

0.36

Pb

1.53

70.

224

26.3

1.89

Th

2.37

2U

0.59

187

Sr/86

Sr0.

7038

10.

7045

60.

7036

014

3N

d/14

4N

d0.

5128

40.

5127

90.

5128

920

6P

b/20

4P

b19

.139

18.5

5418

.651

207P

b/20

4P

b15

.618

15.5

8615

.558

208P

b/20

4P

b39

.472

39.0

3638

.842

Maj

orel

emen

tsw

ere

anal

yzed

byel

ectr

onm

icro

prob

eat

Ore

gon

Stat

eU

nive

rsit

y(L

.D

ougl

as-P

rieb

e,an

alys

t).

Pre

cisi

onis

61%

for

maj

orel

emen

tsan

d5^

7%fo

rK

and

P.

Mg/

(Mg+

Fe*

)as

sum

esal

lF

eis

Fe2�

.M

ajor

elem

ents

inW

C45

sed

byIC

P-A

ES

atO

rego

nSt

ate

Uni

vers

ity

are

elem

enta

lw

t%.

Tra

ceel

emen

tsw

ere

ana-

lyze

dby

ICP

-MS

atU

nive

rsit

yof

Haw

aii,

exce

ptT

han

dU

conc

entr

atio

ns,

whi

chw

ere

anal

yzed

byT

IMS

onth

eU

HSe

ctor

54fo

llow

ing

anio

nex

chan

geex

trac

-ti

onfr

omdi

ssol

ved

glas

sch

ips

[14]

.P

reci

sion

,ba

sed

onre

plic

ate

anal

ysis

ofre

fere

nce

stan

dard

s,is

1^3%

for

RE

Ean

d2^

25%

for

othe

rIC

P-M

Str

ace

elem

ents

.T

han

dU

proc

edur

albl

anks

are6

2pg

and

exte

rnal

repr

oduc

ibili

ties

are

bett

erth

an0.

2%.

Dat

aar

ere

port

edre

lati

veto

ava

lue

of14

3N

d/14

4N

d=

0.51

1850

for

the

La

Jolla

Nd

stan

dard

,an

d87

Sr/86

Sr=

0.71

025

for

NB

S98

7Sr

.Is

otop

icfr

acti

onat

ion

corr

ecti

ons

are

148N

dO/14

4N

dO=

0.24

2436

(148N

d/14

4N

d=

0.24

1572

);P

bis

otop

era

-ti

osar

eco

rrec

ted

for

frac

tion

atio

nus

ing

the

NB

S98

1st

anda

rdva

lues

ofT

odt

etal

.[1

5].

Wit

hin-

run

erro

rson

the

isot

opic

data

abov

ear

ele

ssth

anor

equa

lto

the

exte

rnal

unce

rtai

ntie

son

thes

est

anda

rds:

the

tota

lra

nge

mea

sure

d(l

ast

2yr

)fo

rN

BS

987

Sris

aR

EE

inth

ese

sam

ples

wer

ean

alyz

edby

ion

mic

ropr

obe

atW

oods

Hol

eO

cean

ogra

phic

Inst

itut

ion.

Acc

urac

yþ

5^10

%;

prec

isio

nþ

1^8%

.



206Pb/204Pb, 207Pb/204Pb, and 208Pb/204Pb were de-termined on single chunks of glass from WC44and WC45. There are important di¡erences inisotopic compositions between Amsterdam andSt. Paul Islands and between Boomerang Sea-mount and the islands. Amsterdam Island hashigher 206Pb/204Pb, 208Pb/204Pb and 87Sr/86Srthan St. Paul Island (Fig. 5; [17^21]; F. Frey

and D. Weis, unpublished data; W. White, un-published data). WC44 plots close to AmsterdamIsland samples in all of the isotope plots. WC45 ishigher in 87Sr/86Sr than Amsterdam and St. Paulsamples, but lower in 206Pb/204Pb and 143Nd/144Nd(Fig. 5). Amsterdam, St. Paul, and WC44 plot inor close to the ¢eld for ODP site 756, drilled onthe southern terminus of the Ninetyeast Ridge

Fig. 5. Isotope compositions of Boomerang glasses and whole rock and glass samples from Amsterdam and St. Paul Islands([17^21]; F. Frey and D. Weis, unpublished data; W. White, unpublished data). Amsterdam and St. Paul Islands are distinct onall plots. WC44 plots close to Amsterdam samples, while WC45 has less radiogenic Pb and Nd but more radiogenic Sr and plotsaway from WC44 and the islands. WC45 also plots either in the Kerguelen ¢eld or between the other data and the Kerguelen¢eld. The `SEIRÂSP' ¢eld encloses SEIR data from the ASP Plateau; `SEIRô¡ ASP' ¢eld encloses all other available SEIRdata. Field labeled `Kerguelen Plume' described by Weis et al. [43]. Data sources for SEIR, Kerguelen and Ninetyeast Ridge:[17^22,43^49], K. Johnson, unpublished data. Error bars represent the total range of variation on analyses of isotope standardsmeasured in the University of Hawaii lab over the last 2 yr.



[22], in Nd^Sr space (Fig. 5a), but only St. Pauland WC45 plot in or close to this ¢eld in Sr^Pband Pb^Pb space (Fig. 5b,c). WC45 consistentlyplots between WC44, Amsterdam, and St. Pauland the ¢eld for Kerguelen Island (Fig. 5a^c).

3.2.2. 210Po^210Pb datingA single chip of glass from WC44 was dated

using the 210Po^210Pb technique [23] and wasfound to have an eruption window of 1 Novem-ber 1995 to 29 December 1995. Po degasses froma lava upon eruption; due to uncertainty in theextent of initial Po degassing, we report ages us-ing an èruption window' (the calendar dates cor-respond to maximum (100%) and minimum (75%)Po degassing based upon observed 100% ands 75% Po degassing during the 40-m-deep 1989eruption of Macdonald seamount [24] and the1991^1992 eruption of 9^10³N EPR [23], respec-tively). A major control on Po degassing is likelyhydrostatic pressure; because of the shallownessof this eruption, we feel the maximum age bestapproximates the actual age. Ages are calculatedby regressing a time series of 210Po measurementsto the radioactive ingrowth curve. Typically atleast three 210Po analyses are used but only twoanalyses were possible for WC44. Nevertheless, itis safe to conclude that the sample was on theorder of 5 þ 2 months old when collected on 30March 1996.

3.2.3. SedimentWC45 from the rim of the summit caldera re-

covered 250 g of glass and rock fragments withabundant red^brown sediment. The major ele-ment and chalcophile trace element compositionof this sediment (Table 1) is grossly similar tohigh-Fe, low-Mn ochres and ferruginous umbersfrom Cyprus and the TAG area of the Mid-At-lantic Ridge [25^27], although the WC45 sedimentis higher in Fe (49.24 elemental wt%) and lower inMn (491 ppm) than the ochres and umbers fromeither location. While the Fe/Mn ratio in WC45sediment is Vtwo times higher (V1000) than thatin Cyprus umbers and ochres (1^556), Ba/Cr, Sr/Cr, Cu/Zr, and Co/Ni in the WC45 sediment arewithin the ranges reported for Cyprus [26]. Pbconcentration in WC45 sediment (26 ppm) is

also similar to Pb values in Cyprus samples (12^219) [26].

Chondrite-normalized REE patterns from theWC45 sediment are parallel to WC44 and WC45glass patterns, but are lower in REE concentra-tion than the glasses (Fig. 4a). Normalizing WC45sediment to WC45 glass produces a relatively £atpattern at about 0.3 times the glass concentration(Fig. 4b).

3.3. Water column

Portable, wire-mounted instruments calledMiniature Autonomous Plume Recorders(MAPRs) designed to measure water columnphysical properties were used at all dredge andwax core sites [28]. The purpose of the MAPRis to provide rapid detection of hydrothermal ac-tivity during near-bottom operations where watercolumn information would otherwise not be ob-tained, thereby using valuable ship time more ef-¢ciently. The instrument contains a nephelometerto measure the backscatter intensity of the water(related to particulate matter in the water), and

Fig. 6. MAPR temperature and nephelometer pro¢les col-lected from the WC44 deployment. Dark lines are data col-lected on the downswing and lighter lines are from the up-swing. The dashed line marked àmbient pro¢le' is from thenearest deeper-water temperature pro¢le to Boomerang Sea-mount. The nephelometer pro¢les are a three-point movingmedian ¢lter that clearly shows the layered nature of theanomaly within the caldera (see text). The box around thelower portion of the temperature pro¢le draws attention tothe 1.6^1.7³C temperature anomaly discussed in the text.



temperature and pressure sensors. It was attachedto the hydro-wire 30 m above the wax core andrecords data during both lowering and raising ofthe wire [29].

The MAPR pro¢le collected from the center ofthe summit caldera is shown in Fig. 6. Within100 m of the caldera £oor, the nephelometeranomaly is 0.3 V, nearly an order of magnitudelarger than plume anomalies we measured else-where on the SEIR [28]. In£ections in the neph-elometer pro¢le suggest strati¢cation in severaldistinct water layers from the base of the calderato its rim (Fig. 6).

The temperature pro¢les of nearby MAPR low-erings are nearly identical in the 400^1000 mdepth range, and they parallel that of the deploy-ment into the summit caldera at depths shallowerthan 650 m. Below that depth, the caldera temper-atures are 1.6^1.7³C warmer than temperaturesboth extrapolated from the gradient above thesummit caldera and associated with similar waterdepths outside the caldera (Fig. 6).

4. Discussion

The discovery of an volcanically active sea-mount in a poorly studied region of the IndianOcean is a signi¢cant ¢nding. First, active volcan-ism on the ASP Plateau was poorly documentedprior to this survey, and the history of subaerialvolcanism at the two islands still remains poorlyunderstood [30^33]. Second, the current chemicalcomposition and geographic location of activeASP hotspot magmatism is important in under-standing plume^ridge interaction and in evaluat-ing the submarine eruptive history of the ASPhotspot. Third, the discovery of the volcanicallyactive Boomerang Seamount provides an oppor-tunity to study the e¡ects of volcanism at a shal-low submarine caldera upon the dispersal of hy-drothermal e¥uent, the formation of submarinemineral deposits, and the origin of biologicalniches in the Indian Ocean.

4.1. Active hotspot volcanism

O¡-axis volcanism on the ASP Plateau is widely

scattered rather than concentrated at a singleeruptive center [1,2]. Both Amsterdam Islandand St. Paul Island have had historical volcanicactivity [33]; the last Amsterdam eruption wasmore than 100 yr ago and the most recent vol-canic eruption on St. Paul was in 1793. We ob-served several active fumaroles along the innerwalls of the St. Paul caldera during a brief visitin 1996. Boomerang Seamount has clearly had themost recent eruption of these three volcanic edi-¢ces.

The basalt glasses from the two sites at Boo-merang Seamount are similar to, but have someimportant di¡erences from, volcanic rocks at bothAmsterdam and St. Paul Islands in major andtrace elements and isotopes. Moreover, there ap-pear to be signi¢cant major element and isotopicdi¡erences between WC44 and WC45, as well asmajor element di¡erences within WC45 itself.

The grouping of the data within Mg#^K2O/TiO2 space illustrates the di¡erences between Am-sterdam, St. Paul, and the Boomerang samples(Fig. 3). First, samples from Amsterdam and St.Paul Islands generally plot in distinct ¢elds,although one St. Paul sample plots within analyt-ical uncertainty of the Amsterdam ¢eld. Boomer-ang Seamount and Amsterdam Island, althoughseparated by only 18 km, are also di¡erent inK2O/TiO2 with Amsterdam Island being generallymore enriched at a given Mg#. St. Paul Islandlavas show much more compositional diversitythan Amsterdam lavas, but a cluster of samples,from chilled dike margins on the northeast-facingscarp of the southern side of the breached calderaon St. Paul (this study and [17,32]), are composi-tionally similar to some of the Boomerang Sea-mount samples (triangles in Fig. 3).

The major element di¡erences between WC44and WC45, and between the two compositionalgroups in WC45 may be largely explained bycrustal-level fractional crystallization of olivine,clinopyroxene, and plagioclase. However, the Sr,Nd, and Pb isotope compositions discussed belowrequire that di¡erent parental magmas are in-volved. These di¡erent parental magmas musthave been produced by partial melting of an iso-topically heterogeneous mantle source.

The close similarity between Boomerang Sea-



mount and St. Paul lavas shown in Fig. 3 doesnot extend to Sr^Nd^Pb isotope compositions.Plots of 87Sr/86Sr, 143Nd/144Nd, and 206Pb/204Pb(Fig. 5) show that, in general, lavas from Amster-dam and St. Paul Islands are isotopically distinct,although there is some overlap in 87Sr/86Sr vs.143Nd/144Nd (Fig. 5a). In plots of 206Pb/204Pb vs.87Sr/86Sr and 208Pb/204Pb, Boomerang samplesplot outside of the ¢eld for St. Paul lavas, butWC44 plots within the ¢eld for Amsterdam lavas(Fig. 5b,c). Although the Boomerang, Amster-dam, and St. Paul samples all plot in or nearthe ¢eld for other SEIR lavas from the ASP Pla-teau, only WC45 consistently plots near the ¢eldfor Kerguelen lavas, and it appears to deviatefrom any trajectory that might be de¢ned byWC44, Amsterdam, St. Paul, and o¡-plateauSEIR lavas.

We infer that the mantle underlying the Boo-merang Seamount area is isotopically heterogene-ous on a very small scale and that blobs of mantlewith a Kerguelen Island isotopic signature arecontained in the melting regime for BoomerangSeamount. The isotope data indicate that themantle source for the three volcanic centers prob-ably contains di¡erent proportions of at leastthree components. These components include aradiogenic Pb component, best characterized bythe WC44 and Amsterdam Island lavas; a de-pleted mantle component, exempli¢ed by MORBsampled along the SEIR away from the plateau;and an enriched component that is akin to theKerguelen hotspot, and best represented at Boo-merang Seamount by WC45. The scale of mantlesource heterogeneities must be very small, prob-ably on the order of 1 km, in order to account forthese measurable isotopic di¡erences at Boomer-ang Seamount.

Collectively, Boomerang Seamount, Amster-dam, and St. Paul Island lavas generally overlapthe Sr^Nd^Pb isotope compositions seen forNinetyeast Ridge, in particular ODP site 756 onthe southern terminus of the Ninetyeast Ridge(Fig. 5). This is signi¢cant given the tectonic re-construction, wherein the Ninetyeast Ridge mayhave been built in part by the ASP hotspot[9,21,22]. A line of submarine volcanoes northeastof the ASP Plateau appears to be the trace of the

ASP hotspot. This trace would have formed onthe Australian Plate between the time that theKerguelen Plateau split from Broken Ridge andNinetyeast Ridge and the time that the ASPplume hotspot was `captured' by the approachingSEIR. It appears that the ASP hotspot stoppedforming this chain on the Australian Plate be-tween 5 and 10 Ma when hotspot volcanism wasintercepted by the SEIR, which had been migrat-ing towards it [1]. Once captured by the ridge, thehotspot was able to build up a shallow platformby adding to the volcanic accretion at the spread-ing center. Currently, the locus of ASP hotspotvolcanism appears to be wholly on the AntarcticPlate, where it has formed the islands and Boo-merang Seamount, among other constructive fea-tures. Thus, the observed isotopic similarities be-tween Boomerang Seamount, the islands, andNinetyeast Ridge are consistent with the genesisof these features from the same hotspot at di¡er-ent stages of its evolution. This is also consistentwith the assertion that the Ninetyeast Ridge wasformed by a combination of two plume sources,one of which is similar to that which formed St.Paul Island [21].

4.2. Sediment

The red^brown sediment recovered in WC45has both hydrothermal and non-hydrothermalcharacteristics. Its major and chalcophile trace el-ements are similar to hydrothermally derivedochres in both Cyprus [25,26] and the Mid-Atlan-tic Ridge at TAG [34], suggesting it was formed ina similar environment and by a similar mecha-nism to those deposits. In particular, Ba/Cr, Sr/Cr, Cu/Zr, and Co/Ni and Pb in the WC45 sedi-ment are within the ranges noted for Cyprusochres and low-Mn umbers [26]. These types ofdeposits are thought to form by oxidation of sul-¢de by a low temperature hydrothermal £uid,combined with precipitation of Fe oxide and silicafrom this £uid [35]. The non-hydrothermal char-acter of the sediment is suggested by Fig. 4. TheREE patterns for the sediment and the associatedglass are parallel and unfractionated relative toeach other, demonstrating a `rock-like' REE pro-¢le for the sediment. Thus, it is possible that the



sediment comprises rock particles, carrying theREE signature of the glass, and hydrothermallyderived Fe- and chalcophile-rich phases that di-lute the total REE concentrations but do notchange the REE signature.

It is also possible that the elevated chalcophileand metalloid concentrations we observe in sedi-ments of Boomerang Caldera are related to de-gassing during the recent eruption itself. Rubin[36] proposed a basalt degassing explanation (i.e.the same way Po is lost from basalts, which al-lows us to date them) for very similar enrichmentsof Cu, Zn, Pb, and As (as well even larger enrich-ments in Mo, Sb, Sn and Pb) in particles fromhydrocasts into the Loihi seamount pit craterthat formed following the 1996 eruption. TheseLoihi particles also had indistinguishable Pb andNd isotopic compositions from Loihi basalts, butslightly shifted Sr isotopic composition. The geo-chemical patterns of enrichment at Loihi could bepredicted from a combination of relative elementvolatility during magmatic processes on land andsolubility of these elements in seawater. The sameprocess may have occurred at Boomerang result-ing in the observed compositions of muds.

4.3. Water column

The MAPR pro¢le (Fig. 6) collected from thecenter of the Boomerang Caldera suggests hydro-thermal activity in the caldera. The 1.7³C temper-ature anomaly and the 0.3 V nephelometer anom-aly are nearly an order of magnitude larger thanthe other plume anomalies we measured. At deep-sea hydrothermal sites elsewhere, such large opti-cal backscatter anomalies have only been ob-served in the presence of volatile-rich £uids and/or relatively soon after an eruptive event whichinitiates or reinvigorates a hydrothermal system[37]. Although the precise source of the enhancedbackscatter has not yet been identi¢ed, Baker etal. [37] suggest that elemental sulfur or biologi-cally produced particles may e¤ciently scatterlight of the V1 Wm wavelength of the MAPRnephelometer.

The strati¢ed nephelometer pro¢le suggestslayering of distinct water masses from the baseof the caldera to its rim. A simple explanation

of this strati¢cation, seen on the up and downtransits for each of the two impacts of WC44, isthe in£uence of di¡erent chronic hydrothermalsources with di¡erent buoyancy £uxes. This hy-pothesis is supported by the presence of individualmaxima and minima in the nephelometer pro¢le(Fig. 6). An alternative hypothesis is strati¢cationresulting from double-di¡usive convection [38]. Ineither case, the stably strati¢ed layers are pre-served because the unbroken caldera rim createsan enclosed basin that prevents mixing with thesurrounding water column. Water within the cal-dera originates as spill-over around the calderarim, so its hydrographic properties re£ect thoseat a depth of V650 m, modi¢ed by hydrothermalactivity within the caldera. Staircase strati¢cationof the water column is a phenomenon observedunder hydrothermal conditions elsewhere. Layer-ing of CO2, density, temperature, conductivity,and pH in 100 and 200 m water column pro¢leswere observed at Lakes Nyos and Monoun inCameroon [39], and hot brine pools in deeps atthe bottom of the Red Sea are strati¢ed at depthscales of tens of meters [40].

5. Conclusions

A volcanically active submarine volcano, Boo-merang Seamount, was discovered 18 km north-east of Amsterdam Island on the ASP Plateaunear the SEIR. The seamount is in a region wherethe crust on the nearly stationary Antarctic Plateis being thickened by widespread volcanic activity.The 1100-m-high Boomerang Seamount is toppedby a 2 km circular caldera in which extremelyfresh, tholeiitic to transitional basalt glass wasrecovered. MAPR water column pro¢les withinthe caldera show evidence for hydrothermal activ-ity as measured by large, stepped nephelometeranomalies.

Volcanic glass was recovered from two sam-pling sites, one from the caldera £oor and theother along its rim. These glasses show chemicaland isotopic similarities to lavas from Amsterdamand St. Paul Island, but they di¡er in detail. REEpatterns in lavas from the three edi¢ces are sim-ilar. K2O/TiO2 ratios are similar to some basalts



from St. Paul Island, but di¡er signi¢cantly fromAmsterdam Island basalts. On the other hand, Sr,Nd, and Pb isotope ratios are similar in Amster-dam Island lavas and WC44 from the £oor ofBoomerang Seamount's summit caldera, but thesedi¡er signi¢cantly from St. Paul Island and WC45compositions. Although there are signi¢cant dif-ferences between basalts from Boomerang Sea-mount and from the islands of Amsterdam andSt. Paul, these basalts all display radiogenic 206Pb/204Pb ratios, and they are distinct from SEIRlavas erupted away from the ASP Plateau. Pres-ence of Kerguelen-type source mantle beneathBoomerang Seamount is supported by the similar-ity of WC45 and Kerguelen isotope ratios. Theconsistent similarity of isotope compositions ofBoomerang, Amsterdam, St. Paul, and youngestNinetyeast Ridge lavas lends support to the ideathat the ASP plume contributed to the formationof Ninetyeast, as suggested by others [9,21,22].

Acknowledgements

We thank Captain Eric Buck and the crew ofthe R/V Melville for their high level of profession-alism during this long and sometimes di¤cultvoyage. We are particularly grateful to Gene Pil-lard, Terry Naumann, and David Preston fortheir tireless handling of deck operations. BillWhite, Fred Frey, and Dominique Weis kindlyprovided previously unpublished analyses of sev-eral Amsterdam and St. Paul Island samples. Wethank Terres Australes et Antarctiques Franc°aises(TAAF) for their clearance to work in territorialwaters around Amsterdam and St. Paul Islands.K.N. thanks J.J. Mahoney and K. Spencer forassistance with the Sr, Nd and Pb isotopic analy-ses. Constructive reviews by Rodey Batiza andHubert Staudigel improved the manuscript. Thiswork was supported by NSF-9415948, NSF-9505667, and the NOAA VENTS Program. Thisis PMEL contribution 2073.[FA]

References

[1] J.A. Conder, D.S. Scheirer, D.W. Forsyth, Sea£oor

spreading on the Amsterdam^St. Paul Plateau, J. Geo-phys. Res. 105 (B4) (2000) 8263^8277.

[2] D.S. Scheirer, D.W. Forsyth, J.A. Conder, M.A. Eberle,S.-H. Hung, K.T.M. Johnson, D.W. Graham, Anomaloussea£oor spreading of the Southeast Indian Ridge near theAmsterdam^St. Paul Plateau, J. Geophys. Res. 105 (B4)(2000) 8243^8262.

[3] D.W. Graham, K.T.M. Johnson, L. Douglas-Priebe, J.E.Lupton, Hotspot-ridge interaction along the SoutheastIndian Ridge near Amsterdam and St. Paul Islands: He-lium isotope evidence, Earth Planet. Sci. Lett. 167 (1999)297^310.

[4] J.C. Mutter, S.C. Cande, The early opening between Bro-ken Ridge and Kerguelen Plateau, Earth Planet. Sci. Lett.65 (1983) 369^376.

[5] A.A. Tikku, S.C. Cande, Early Tertiary reconstructions ofthe Australian and Antarctic plates, EOS Trans. Am.Geophys. Union 77 (1996) 689.

[6] S.C. Cande, J.C. Mutter, A revised identi¢cation of theoldest sea-£oor spreading anomalies between Australiaand Antarctica, Earth Planet. Sci. Lett. 58 (1982) 151^160.

[7] J.C. Mutter, K.A. Hegarty, S.C. Cande, J.K. Weissel,Breakup between Australia and Antarctica; a brief reviewin the light of new data, Tectonophysics 114 (1985) 255^279.

[8] A.E. Gripp, R.G. Gordon, Current plate velocities rela-tive to the hotspots incorporating the NUVEL-1 globalplate motion model, Geophys. Res. Lett. 17 (1990) 1109^1112.

[9] B.P. Luyendyk, W. Rennick, Tectonic history of aseismicridges in the eastern Indian Ocean, Geol. Soc. Am. Bull.88 (1977) 1347^1356.

[10] R.A. Duncan, M. Storey, The life cycle of Indian Oceanhotspots, in: R.A. Duncan, D.K. Rea, R.B. Kidd, U. vonRad, J.K. Weissel (Eds.), Synthesis of Results from Sci-enti¢c Drilling in the Indian Ocean, Geophys. Monogr.70, Am. Geophys. Union, Washington, DC, 1992, pp. 91^103.

[11] J.-Y. Royer, M.F. Co¤n, Jurassic to Eocene plate recon-structions in the Kerguelen Plateau region, in: S.W. Wise,Jr., R. Schlich (Eds.), Proc. ODP, Sci. Results, 120, OceanDrilling Program, College Station, TX, 1992, pp. 917^928.

[12] W.H. Smith, D.T. Sandwell, Measured and EstimatedSea£oor Topography (version 4.2), World Data Center-A for Marine Geology and Geophysics, 1997.

[13] J.-Y. Royer, R. Schlich, Southeast Indian Ridge betweenthe Rodriguez Triple Junction and the Amsterdam andSaint-Paul Islands: detailed kinematics for the past 20m.y., J. Geophys. Res. 93 (1988) 13524^13550.

[14] K.H. Rubin, M.C. Smith, M.R. Per¢t, L.F. Sacks, D.Christie, Geochronology and petrology of lavas fromthe 1996 North Gorda Ridge eruption, Deep-Sea Res. II45 (1998) 2751^2799.

[15] W. Todt, R.A. Cli¡, A. Hanser, A.W. Hofmann, Evalua-tion of a 202Pb^205Pb double spike for high-precision lead



isotopic analyses, in: A. Basu, S. Hart (Eds.), Earth Pro-cesses, Reading the Isotopic Code, 1996, pp. 429^437.

[16] J. Mahoney, C. Nicollet, C. Dupuy, Madagascar basalts:tracking oceanic and continental sources, Earth Planet.Sci. Lett. 104 (1991) 350^363.

[17] L. Dosso, H. Bougault, P. Beuzart, J.-Y. Calvez, J.-L.Joron, The geochemical structure of the SouthÊast In-dian Ridge, Earth Planet. Sci. Lett. 88 (1988) 47^59.

[18] A. Michard, R. Montigny, R. Schlich, Geochemistry ofthe mantle beneath the Rodriguez triple junction and theSouthÊast Indian Ridge, Earth Planet. Sci. Lett. 78(1986) 104^114.

[19] B. Hamelin, B. Dupre, C.-J. Alle©gre, Pb^Sr^Nd isotopicdata of Indian Ocean ridges: new evidence of large-scalemapping of mantle heterogeneities, Earth Planet. Sci.Lett. 76 (1986) 288^298.

[20] B. Dupre, C.J. Alle©gre, Pb^Sr isotope variation in IndianOcean basalts and mixing phenomena, Nature 303 (1983)142^146.

[21] F.A. Frey, D. Weis, Temporal evolution of the Kerguelenplume: Geochemical evidence from V38 to 82 Ma lavasforming the Ninetyeast Ridge, Contrib. Mineral. Petrol.121 (1995) 12^28.

[22] D. Weis, F.A. Frey, Isotope geochemistry of NinetyeastRidge basement basalts: Sr, Nd, and Pb evidence for in-volvement of the Kerguelen Hotspot, in: J. Weissel, J.Peirce, E. Taylor, J. Alt et al. (Eds.), Proc. ODP, Sci.Results, 121, Ocean Drilling Program, College Station,TX, 1991, pp. 591^610.

[23] K.H. Rubin, J.D. Macdougall, M.R. Per¢t, 210Po^210Pbdating of recent volcanic eruptions on the sea £oor, Na-ture 368 (1994) 841^844.

[24] K.H. Rubin, J.D. Macdougall, Submarine magma degass-ing and explosive magmatism at Macdonald (Tamarii)seamount, Nature 341 (1989) 50^52.

[25] A.H.F. Robertson, Cyprus umbers: Basalt^sediment rela-tionships on a Mesozoic oceanic ridge, J. Geol. Soc.Lond. 131 (1975) 511^531.

[26] A.H.F. Robertson, Origins of ochres and umbers: Evi-dence from Skouriotissa, Troodos Massif, Cyprus, Trans.Inst. Min. Metall. Sect. B 85 (1976) B245^B251.

[27] R.A. Mills, T. Clayton, J.C. Alt, Low-temperature £uid£ow through sul¢dic sediments from TAG: Modi¢cationof £uid chemistry and alteration of mineral deposits, Geo-phys. Res. Lett. 23 (1996) 3495^3498.

[28] D.S. Scheirer, E.T. Baker, K.T.M. Johnson, Detection ofhydrothermal plumes along the Southeast Indian Ridgenear the Amsterdam^St. Paul Plateau, Geophys. Res.Lett. 25 (1) (1998) 97^100.

[29] E.T. Baker, H.T. Milburn, MAPR: a new instrument forhydrothermal plume mapping, RIDGE Events 8 (1997)23^25.

[30] B.M. Gunn, S.E. Abranson, J. Nougier, N.D. Watkins,A. Hajash, Amsterdam Island, an isolated volcano in thesouthern Indian Ocean, Contrib. Mineral. Petrol. 32(1971) 79^92.

[31] B.M. Gunn, N.D. Watkins, W.E. Trzcienski, J. Nougier,

The Amsterdam^St. Paul volcanic province, and the for-mation of low Al tholeitic andesites, Lithos 8 (1975) 137^149.

[32] R. Hebert, Petrology of a SouthÊast Indian Ridge seg-ment between 28³ and 41³S and the origin of St. PaulIsland, Indian Ocean: I. Mineralogy and major and minorelements chemistry, O¢oliti 12 (1987) 357^391.

[33] J. Nougier, J.W. Thomson, Amsterdam and Saint PaulIslands, in: W.E. LeMasurier, J.W. Thomson, P.E. Baker,P.R. Kyle, P.D. Rowley, J.L. Smellie, W.J. Verwoerd(Eds.), Volcanoes of the Antarctic Plate and SouthernOceans, Antarctic Research Series 48, American Geophys-ical Union, Washington, DC, 1990, pp. 442^445.

[34] R.A. Mills, H. Elder¢eld, Rare earth element geochemis-try of hydrothermal deposits from the active TAGMound, 26³N Mid-Atlantic Ridge, Geochim. Cosmo-chim. Acta 59 (1995) 3511^3524.

[35] M.K. Tivey, S.E. Humphris, G. Thompson, M.D. Han-nington, P.A. Rona, Deducing patterns of £uid £ow andmixing within the TAG active hydrothermal mound usingmineralogical and geochemical data, J. Geophys. Res. 100(1995) 12527^12555.

[36] K.H. Rubin, Degassing of metals and metalloids fromerupting seamount and mid-ocean ridge volcanoes: Ob-servations and predictions, Geochim. Cosmochim. Acta61 (1997) 3525^3542.

[37] E.T. Baker, A. Tennant, R.A. Feely, G.T. Lebon, S.L.Walker, Field and laboratory studies on the e¡ect of par-ticle size and composition on optical backscattering mea-surements in hydrothermal plumes, Deep-Sea Res. (2000),in press.

[38] J.S. Turner, Buoyancy E¡ects in Fluids, Chapter 8, Cam-bridge Univ. Press, Cambridge, 1973.

[39] M. Kusakabe, G.Z. Tanyileke, S.A. McCord, Recent pHand CO2 pro¢les at Lakes Nyos and Monoun, Came-roon: Implications for the degassing strategy and its nu-merical simulation, J. Volcanol. Geotherm. Res. (2000), inpress.

[40] D.A. Ross, Temperature structure of Red Sea brines, in:E.T. Degens, D.A. Ross (Eds.), Hot Brines and RecentHeavy Metal Deposits in the Red Sea, Springer-Verlag,New York, 1969, pp. 148^152.

[41] M. Girod, G. Camus, Y. Vialette, Sur la presence detholeiites a© l'|ªle Saint-Paul (Ocean Indien), Contrib. Min-eral. Petrol. 33 (1971) 108^117.

[42] E. Anders, N. Grevesse, Abundances of the elements:meteoritic and solar, Geochim. Cosmochim. Acta 53(1989) 197^214.

[43] D. Weis, F.A. Frey, H. Leyrit, I. Gautier, Kerguelen Ar-chipelago revisited: geochemical and isotopic study of thesoutheast province lavas, Earth Planet. Sci. Lett. 118(1993) 101^119.

[44] W.M. White, A.W. Hofmann, Sr and Nd isotope geo-chemistry of oceanic basalts and mantle evolution, Nature296 (1982) 821^825.

[45] D. Weis, J.-F. Beaux, I. Gautier, A. Giret, P. Vidal, Ker-guelen archipelago: geochemical evidence for recycled ma-



terial, in: S.R.H.a.L. Gu«len (Ed.), Crust/Mantle Recyclingat Convergence Zones, NATO ASI Series, Kluwer Aca-demic Press, Amsterdam, 1989, pp. 59^63.

[46] D. Weis, Y. Bassias, I. Gautier, J.-P. Mennessier, Dupalanomaly in existence 115 Ma ago: evidence from isotopicstudy of the Kerguelen Plateau (South Indian Ocean),Geochim. Cosmochim. Acta 53 (1989) 2125^2131.

[47] H.-J. Yang, F.A. Frey, D. Weis, A. Giret, D. Pyle, G.Michon, Petrogenesis of the £ood basalts forming thenorthern Kerguelen Archipelago: Implications for theKerguelen Plume, J. Petrol. 39 (1998) 711^748.

[48] D. Weis, F.A. Frey, A. Giret, J.-M. Cantagrel, Geochem-ical characteristics of the youngest volcano (Mount Ross)in the Kerguelen Archipelago: Inferences for magma £ux,lithosphere assimilation and composition of the KerguelenPlume, J. Petrol. 39 (1998) 973^994.

[49] D. Weis, F.A. Frey, Role of Kerguelen Plume in generat-ing the eastern Indian Ocean sea£oor, J. Geophys. Res.101 (1996) 13831^13849.



Date post:	04-Mar-2023
Category:	Documents
Upload:	independent
View:	0 times
Download:	0 times

Boomerang Seamount: the active expression of the Amsterdam^St. Paul hotspot, Southeast Indian Ridge

Documents