Size and Depth of Ancient Magma Reservoirs Under Atolls and Islands of French Polynesia Using Gravity Data

Home , Huahine, Maiao, Manihi, Manuae (Society Islands), Maupiti, Moruroa

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 105, NO. B4, PAGES 8173{8191, APRIL 10, 2000

Size and depth of ancient magma reservoirs under atolls

and islands of FrenchPolynesia using gravity data

Val erie Clouard and Alain Bonneville

Lab oratoire de G eosciences Marines et T el ed etection, Universit edelaPolyn esie francaise,

Pap eete, Tahiti, FrenchPolynesia

Hans G. Barsczus,

GBE - UMR 5569 - CNRS and Universit edeMontp ellier I I, Montp ellier, France

Abstract. New insights into the structural and tectonic evolution of islands

and atolls in FrenchPolynesia are derived from the analysis of gravity data.

Free-air anomaly maps were constructed using gravity data from land surveys

of 25 islands combined with free-air anomaly data derived from satellite

altimetry and shipb orne gravimeters. Residual isostatic anomalies were

calculated using a three-dimensional 3-D, four-layer crustal mo del taking

into account the bathymetry of the sea o or, the top ography of the islands,

and the de ection of the lithosphere under the load of the volcano es. Twenty

of the 25 islands yield a p ositive residual anomaly, ranging b etween 9 and

60 mGal. Negative residual anomalies over four islands probably corresp ond

either to limestone dep osits or to fragmented material from aerial volcanism.

The o ccurrence of a p ositive anomaly provides some evidence that there

is a solidi ed magma chamb er at depth b eneath each of several islands in

FrenchPolynesia. A linear relation b etween the amplitude of the p ositive

gravity anomaly and island volume is also observed. Geological features which

generate these p ositive anomalies are describ ed by simple geometric mo dels

3-D ellipsoidal dense b o dies which lead to a rough estimate of the size and

depth of the magma chamb ers. We then prop ose a simple linear relation

between the volume of such magma chamb ers and the volume of islands that

can b e used for other extinct intraplate basaltic volcano es. The diameter of

the inferred magma chamb er and its pro jection on the surface of the islands

imply that calderas corresp ond mostly to collapses of the anks of the edi ces

rather than to caldera vertical subsidence.

1. Intro duction yr [Herron, 1972; Minster and Jordan, 1978]. The

Austral, So ciety, Pitcairn-Gambier, and Marquesas is-

FrenchPolynesia Figure 1 is lo cated to the south

land chains are comp osed mainly of islands, while the

of the central part of the Paci c Plate, b etween latit-

Tuamotu archip elago is comp osed entirely of atolls. All

udes 5 S and 30 S and longitudes 130 W and 160 W.

together, FrenchPolynesia consists of 34 islands and 84

The Austral, So ciety, Pitcairn-Gambier, and Tuamotu

atolls.

island chains trend N120 , the same direction of the

The islands rise 4500 m ab ove the sea o or, and their

presentPaci c Plate motion. The Marquesas Islands

maximum elevation ab ove sea level, at Tahiti, is 2200

have a N140 strike. The region is crossed by the Aus-

m. To a rst approximation, island ages and morpho-

tral and Marquesas Fracture Zones FZ with a N70

logies are consistent with a hotsp ot theory: the age of

strike. These FZs are the main tectonic features of the

each island within a chain increases from the southeast

sea o or and are related to the Farallon Ridge, which

to the northwest, based on radiometric dating of volcan-

b ecame extinct 9 Myr ago [Mammerickx et al., 1980].

ism from the present to ab out 16 Ma Table 1. The age

The average velo cityofPaci c Plate motion is 11 cm

of the o ceanic crust ranges b etween 45 and 85 Ma [Her-

ron, 1972; Mayes et al., 1990; Munschy et al., 1996].

Also at Centre IRD, Montp ellier, France

The Tuamotu Islands are separated from the other is-

land chains by basins exceeding 5000 m in depth. The

origin of the Tuamotu Islands, which rise ab ove a broad

plateau, is uncertain but has b een attributed to an ab or-

Pap er numb er 1999JB900393

ted ridge [Pautot, 1975], a hotsp ot [Okal and Cazenave,

0148-0227/00/1999JB900393$09.00 8173

8174 CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS

Marquesas Nuku Hiva Ua Pou Hiva Oa 10˚S

Marquesas FZ Manihi Mataiva Rangiroa Tuamotu Tupai Makatea Maupiti Bora Bora Tahaa Tetiaroa Manuae Huahine Moorea Maiao Tahiti Society 20˚S Pitcairn-Gambier Maria Moruroa Rimatara Rurutu Gambier Tubuai Austral FZ Austral Raivavae

Rapa 500 km 30˚S 150˚W 140˚W

-6000 -5000 -4000 -3000 -2000 -1000 0

meters

Figure 1. Bathymetric map [Sichoix and Bonnevil le, 1996] showing the volcanic chains of French

Polynesia, with archip elago names underlined. The names of the islands discussed in text are

shown. FZ stands for fracture zone.

1985], or an emerged submarine chain which originated 2. Data Presentation and Pro cessing

close to the East Paci c Rise [Ito et al., 1995].

2.1. Gravity Data

The age, p etrology, and geo chemistry of the French

Polynesian islands are relatively well known see sum-

From 1978 to 1981, 800 measurements were made by

maries by Brousse et al. [1990], Diraison [1991], and

Barsczus [1979, 1980, 1981] using a Worden 600 gravi-

Leroy [1994], but their subsurface structure, apart from

meter at Manuae, Maupiti, Tupai, Bora-Bora, Tahaa,

Moruroa and Fangataufa atolls [Guil le et al., 1992,

Huahine, Maiao, Tetiaroa, Mo orea, and Tahiti in the

1993], remains p o orly known. In this pap er, we address

So ciety Islands; Nuku Hiva, Hiva Oa, and Ua Pou in

the lateral density distribution using existing gravity

the Marquesas Islands; Maria, Rimatara, Rurutu, Tu-

data, which leads to a three-dimensional mo del of the is-

buai, Raivavae, and Rapa in the Austral Islands; Mor-

lands. Shipb oard and satellite-derived gravity data are

uroa and Gambier Islands in Gambier-Pitcairn align-

merged with previously unpublished land data. After

ment; and Mataiva, Manihi, Makatea, and Rangiroa in

adjusting these di erent data sets to a common ref-

the Tuamotu Islands. These data were merged with 137

erence system, a free-air anomaly FAA map is ob-

measurements made by Macheski [1965] at Tahiti and

tained. We then subtract a theoretical gravity eld

Mo orea using a Worden 300 gravimeter and with 320

computed using a three-dimensional, four-layer crustal

measurements made by Leroy [1994] at Tahiti using a

mo del based on a digital terrain mo del for bathy-

Lacoste and Romb erg mo del G gravimeter. All data are

metry and elevation. Residual isostatic anomaly maps

tied to PPT J7912, a reference station of the Interna-

[Blakely, 1995] are then pro duced for 17 islands and 8

tional Gravity Standardization Net established in 1971

atolls to obtain insights into the makeup of shield vol-

IGSN71 and lo cated at Pamatai in Tahiti. The PPT

cano es.

J7912 station where PPT stands for Pap eete and J7912

CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS 8175

Table 1. Radiometric Ages of Aerial Volcanism in FrenchPolynesian Studied Islands

Island Longitude W Latitude S Age, Ma Reference

Austral

0 0

Raivavae 147 39 23 52 5.6-7.6 Duncan and McDougal l [1976]

0 0

Rapa 144 20 27 36 5.9-5.2 Krummenacher and Noetzlin [1966]

0 0

Rimatara 152 48 22 39 4.8-28.6 Turner and Jarrard [1982]

0 0

Rurutu 151 21 22 29 0.6-12.3 Dalrymple et al. [1975], Duncan and

McDougal l [1976], and Turner and

Jarrard [1982]

0 0

Tubuai 149 30 23 23 8.5-10.6 Duncan and McDougal l [1976]

Marquesas

0 0

HivaOa 139 03 09 45 1.6-2.5 Duncan and McDougal l [1974]

0 0

Nuku Hiva 140 12 08 54 3.0-4.3 Duncan and McDougal l [1974]

So ciety

0 0

Bora Bora 151 44 16 30 3.1-3.4 Duncan and McDougal l [1976]

0 0

Huahine 151 00 16 45 2.0-2.6 Duncan and McDougal l [1976]

0 0

Maupiti 152 16 16 27 4.0-4.5 Duncan and McDougal l [1976]

0 0

Mo orea 149 50 17 33 1.5-1.7 Duncan and McDougal l [1976]

0 0

Tahaa 151 30 16 39 2.6-3.2 Duncan and McDougal l [1976]

0 0

Tahiti-iti 149 30 17 36 0.4-0.5 Leroy [1994]

0 0

Tahiti-nui 149 15 17 48 0.19-1.37 Leroy [1994]

Pitcairn-Gambier

0 0

Gambier Manureva 134 57 23 09 6.2 Guil lou et al. [1994]

0 0

Moruroa 138 54 21 51 11.0 Gil lot et al. [1992]

means that it was the twelth station made in 1979 of the Austral Islands A. Bonnevil le, personal commu-

was established by a Japanese team during a circum- nication, 1999. The entire o shore data set is referred

Paci c gravimetric connections campaign [Nakagawa et to here as the "marine data set." The error on FAA

al., 1983]. From this total "land data set", a FAA was satellite-derived data was estimated to b e 5 mGal by

then calculated using the GRS80 ellipsoid mo del. The Sandwel l [1992]. The shipb oard data error is 1 to 2

exp erimental error e related to this land data set can b e mGal. Consequently, a maximum error of 5 mGal for

expressed as the sum of the measurement error e es- the whole marine data set is assumed.

timated to b e 0.35 mGal standard deviation [Barsczus,

1980] and a function of the barometric leveling error dz

2.2. Free-air Anomaly Maps

estimated to b e 10 m, as follows:

When preparing comp osite FAA maps, the rst ad-

justment of land and marine gravity data is controlled

e = e +0:3086 dz 3:5 mGal : 1

visually. Close to lands, the values in the marine data

set, comp osed mainly of satellite data, app ear systemat-

Marine gravity data for the area compiled by Sichoix

ically lower by a few tens of milliGals than the ground

and Bonnevil le [1996] include 27 tracks 116,171 p oints

data set. This discrepancy may b e related to 1 the

from the National Geophysical Data Center NGDC

hanging of radar b eams on island anks [Barriot, 1987];

database whichwas obtained with go o d satellite navig-

2 a bias intro duced from the metho d to pro duce grav-

ation accuracy. After the removal of uncontrolled p eaks,

ity elds from the geoid signal due to crossover of un-

this set was tied to a Seasat and ERS-1 satellite-derived

evenly spaced ground tracks, esp ecially in the vicinityof

FAA grid [Smith and Sandwel l, 1995]. Wehave ad-

islands where data have b een removed [Filmer, 1991];

ded data from recent cruises in FrenchPolynesia to the

and 3 the size of the island itself when its diameter

Sichoix and Bonnevil le [1996] data set: EW9602 on the

ranges from 5 to 15 km, which is close to the resolution

R/V Maurice Ewing to the southeast of the Austral

of the satellite 20-30 km, [Sandwel l, 1992].

Islands [McNutt et al., 1997] and two cruises on the

R/V L'Atalante, ZEPOLYF1, south of the So ciety Is- To adjust these two data sets, we used only shipb oard

lands [Bonnevil le et al., 1997] and ZEPOLYF2, north data in the vicinity of islands up to 6 km o shore, ex-

8176 CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS

17ºS 17ºS

18ºS 18ºS

150ºW 149ºW 150ºW 149ºW

-50 -25 0 25 50 75 100 125 150 175 200

mGal

Figure 2. Example of the derivation of the FAA map for the Tetiaroa, Mo orea, and Tahiti

islands. left Gravity satellite data only [Smith and Sandwel l, 1995]. The thin black lines

represent the shipb oard pro les along which the measured gravityvalues were used to pro duce

an improved eld within the p olygons shown as white lines. right Final FAA map pro duced

by merging the satellite data outside the p olygons with the shipb orne data within the p olygons.

Note the improvement in anomaly resolution over and near the islands. Contour interval is 25

mGal.

1967. Data were carefully checked for quality and cept for Tahiti where this limit is extended to 60 km

sub jected to a crossover analysis, similar to the one b ecause these data have the same reference bases as

used by Smith [1993], but using Golbal Positioning Sys- the ground data and thus their datum are comparable

tem GPS navigated exp editions for reference. The Figure 2. Using a continuous curvature "splines-in-

most recent bathymetric pro les used, including multi- tension" algorithm GMT [Smith and Wessel, 1990],

b eam soundings from EW9602 and ZEPOLYF1 cruises, the land data set and shipb oard data in the vicinityof

are denser, often with a spacing of <500 m along the islands were merged with the complete marine data set

shiptrack. The older pro les haveatypical spacing of minus the data around the islands to obtain twotyp es

>2 km. of grids. The rst grid corresp onds to the complete

FAA map of FrenchPolynesia, which is a coarse grid

Elevation data on islands were obtained by digitiz-

0 0

2 x2 , such asisshown in the small region around the

ing top ographic maps of FrenchPolynesia [ Servicede

island of Tahiti in Figure 2. The second typ e corres-

l'Urbanisme, 1992] and from maps of the Atlas de la

p onds to a set of ner grids covering each island, which

Polyn esie Francaise [ORSTOM, 1993]. For the island

were used in calculating the prop erties and geometry of

of Tahiti we used a digital terrain mo del derived by

00 00

magma chamb ers. These are 30 x30 grids, constrained

G eoimage Sophia Antip olis, France from two stereo-

by go o d spatial coverage on land 1352 measurements

scopic SPOT images. The elevation on atolls was arbit-

for 2350km ,averaging ab out two stations within a

rarily set at 0 m over the lago on for depths b etween 1

00 00

60 x60 area. All the grids are available at URL

and 40 m and at 1 m ab ove sea level on the coral rim

http://www.ipgp.jussieu.fr/UFP/download.html.

for heights b etween 1 and 5 m. Top ographic land data

and bathymetric data were then combined and gridded

0 0

2.3. Top ographic Maps

to pro duce a 2 x2 grid for the whole area, and several

00 00

30 x30 grids covering each island.

We used the synthesized database of bathymetry

compiled by Sichoix and Bonnevil le [1996] that was

2.4. Flexural Mo del

based on 82 cruises plus additional soundings from the

Previous studies of the gravity and top ographyof French Service Hydrographique et Oc eanographique de

the Hawaiian archip elago [e.g., Walcott, 1970] show that la Marine, all acquired with satellite navigation since

CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS 8177

aerial ρ a topography sea water bathymetry ρ ρ w S

crust ρ f 6000m ρ flexure c

mantle ρ ρ c m

Figure 3. The theoretical FAA is obtained by computing the gravity e ect of volcanic construc-

tion ab ove a de ected crust. top The mo del involves six layers, with di erent densities . The

assumed densities for crust , mantle , and water are 2900, 3350 and 1030 kg m ,

c m w

resp ectively. The de ection is estimated using a 3-D mo del for a thin elastic plate loaded by

the top ography see text. The volcanic load is mo deled by three layers: the subaerial volcanic

construction , the submarine volcanic construction , the material lling the exure

a s f

with < < . b ottom The geometry of the system is mo deled by elementary prisms.

a s f

volcanic loads on the o ceanic lithosphere are supp orted of intraplate volcanism. Calmant and Cazenave [1986]

regionally rather than lo cally. In resp onse to applied and Young and Hil l [1986] invoke this concept to explain

loads the lithosphere acts as a thin elastic plate overly- rigidity anomalies obtained b elow the Co ok, Austral,

ing an inviscid uid [Watts et al., 1980]. The amplitude and So ciety Islands and b elow the Cap e Verde rise.

of the de exion dep ends on b oth the size of the volcanic

The numerical and three-dimensional approach de-

load and the e ective exural rigidity of the thin elastic

velop ed by Watts et al., [1975] is used here to solve the

plate [Watts et al., 1975]. The exural rigidity is linked

classical equation governing the deformation w p ositive

to the elastic thickness of the lithosphere that corres-

downward of an elastic layer under a volcanic load:

p onds to the upp er part of the lithosphere and dep ends

D r w + gw = q x; y 2

m l

mainly on the temp erature, and hence on the age of the

lithosphere at the time of loading [Watts, 1978].

where is the load density, is the mantle density,

l m

D the exural rigidity of the lithosphere, r is the bi-

Watts and Ribe [1984] obtained several values of the

harmonic op erator, q x; y is the vertical load derived

exural rigidityinvarious tectonic contexts. McNutt

from the top ographic grid obtained previously, and g is

[1984] intro duced the concept of the "e ective thermal

the acceleration due to gravityg =9:81 ms .

age," which corresp onds to the thermal age at the time

of loading. It can b e less than the lithospheric age if the

In our study the value of D has b een carefully chosen

plate has b een reheated, which is common for regions

for each island group using two metho ds. In the rst

8178 CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS

metho d, seismic refraction pro les provide the Moho

depth So ciety and west Tuamotu Islands [Danobeatia 16˚S

Tupai-3000

1995; Talandier and Okal, 1987] and Marque- et al., 40 km

Bora Bora

Caress et al., 1995]. After this depth is sas Islands [ -2000-3000 Tahaa -4000

compared with the Moho depth based on the exural Maupiti Huahine -2000

-1000

e select the D value that minimizes the di er- mo del w (a) Raiatea Tetiaroa

17˚S

ween observations and mo del. In the second

ence b et -3000

yed when seismic re ection data are un-

metho d, emplo -4000 Moorea

vailable, we used published D values. Typical values

a -4000 Tahiti -2000 22

21 Maiao

D vary from 1:5 10 N m to 8:0 10 N m for the of -3000 0

18˚S -1000

tral Paci c area [e.g. Calmant and Cazenave,

south cen A'

Filmer et al., 1993; Goodwil lie and Watts, 1993].

1987; 152˚W 151˚W 150˚W 149˚W

uchlower than the values exp ected for the These are m ρ

0 a

Watts et al., 1980]. range of crustal ages concerned [ ρ w

-4

hoice of exural rigidity

Complete discussions of the c ρ s

alues for our region are given by McNutt [1998] and v -8 f

(b)

[1997]. The following values for the exural ri- Sichoix -12

Depth (km) ρ

22 c 6000 m

giditywere selected: 4:0 10 N m for So ciety and

21 -16 ρ

:5 10 N m for Pitcairn-Gambier Marquesas Islands, 3 m

22 152˚W 151˚W 150˚W 149˚W

Islands, 1:2 10 N m for western Austral Islands,

3:5 10 N m for the eastern Austral Islands, and

21 300

8:0 10 N m for Tuamotu Islands. We assume that

200

the representative densities of the mantle and the load

3 3

(c)

= 3350 kg m and = 2700 kg m resp ect-

are 100

m l

ely [Caress et al., 1995; Sichoix, 1997]. The de ection

iv 0

0 0

is then computed for all the no des 2 x2 of the to- w -100

Free air anomaly (mGal) 152˚W 151˚W 150˚W 149˚W p ographic grid.

2.5. Isostatic Residual Anomaly

A three-dimensional, four-layer crustal mo del Fig-

(d) 0

ure 3 is used to compute a theoretical FAA. The in-

terface b etween the o cean and sea o or is given by the

Isostatic residual anomaly (mGal)

ymetric grid Figure 1. The interface b etween the bath -50

152˚W 151˚W 150˚W 149˚W

volcanic load, divided into three densitylayers, and the

o ceanic crust is given by the de ection grid, whichis

Figure 4. Residual isostatic anomaly over the So cie-

assumed also to de ne the interface b etween crust and

ty Islands obtained by subtracting the mo deled FAA

mantle. The crust is assumed to have a constant thick-

from the observed. a Bathymetric map of the So ci-

ness of 6000 m. This thickness has b een determined

ety Islands. Data along the pro le AA' solid dashed

from seismic refraction studies in the So ciety-Tuamotu

line are shown in Figures 4b, 4c and 4d; b the crustal

area [Danobeatia et al., 1995] and the Marquesas area

mo del used to compute the mo deled FAA. The elastic

[Caress et al., 1995]. The material lling the depressions

thickness is adjusted using the Moho depth given by

under the load is assumed to have a higher density

f seismic refraction pro les; c comparison b etween ob-

than the overlying volcanic material and the aer-

s served thick dashes and computed FAA thin line

ial part of the volcano . The sea o or normal depth

a along AA'; and d isostatic residual gravity anomaly

is given by the deep est p oint of the bathymetric grid

along AA'.

within the particular area b eing studied. The mo del

0 0

is divided into 2 x2 square prisms extending to Moho

its utilitytoprovide new insights on the makeup of is-

depth. The zero FAA is given by a at o ceanic crust

lands. According to Figure 3, only contrasts b etween

at the sea o or normal depth. The theoretical FAA is

mantle and crustal densities and b etween crust and in-

0 0

then computed for each no de of the 2 x2 grid by sum-

ll densities are relevant in the mo del, not their absolute

ming the gravity e ects of all the prisms [Plou , 1976] values. As the density contrasts are regional paramet-

pro duced by the density contrasts shown in Figure 3. A

ers rather than sp eci c to each island, we determined

similar computation was made for each island using the

these contrasts by minimizing the di erence b etween

00 00

ner grid 30 x30 .

computed and observed gravity anomaly for each ar-

chip elago. The b est estimate of the mantle-crust dens-

By subtracting the theoretical FAA grid from the

ity contrast is everywhere 450 kg m . Assuming a

observed anomaly, an isostatic residual gravity eld

value of 3350 kg m for the mantle density, the crust

[Blakely, 1995], similar to an isostatic Bouguer eld

density is then 2900 kg m .Thus the volcano submar-

Figure 4, is determined and investigated b elowasto

CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS 8179

Table 2. Prop erties of Islands Showing Positive Residual Isostatic Anomalies

Island Caldera Filling Submarine Amplitude Half

Island Volume, Diameter, Density , Density , g , Wavelength Lo cation

f s max

3 3 3

V ,km km kg m kg m mGal km

Austral

Raivavae 2159 3 2900 2900 16 7 south island cent.

Rapa 1500 3.5 and 5 2850 2500 17 7 caldera centered

Rurutu 1950 none 2900 2900 16 7.5 island centered

Tubuai 3273 2 2850 2750 26 11 west caldera cent.

Marquesas

HivaOa 7628 10 and 15 2800 2700 35 18 caldera centered

Nuku Hiva 5811 6 and 9 2800 2800 29 19 east caldera cent.

Ua Pou 2771 none 2800 2650 10 ? not enought pts

So ciety

Bora Bora 2338 3 2900 2900 16 6 caldera cent.

Huahine 2636 drowned 2850 2850 17 7 island centered

Maiao 1857 3 2850 2700 11 3 caldera centered

Manuae 2302 atoll 2800 2550 17 7 east lago on cent.

Maupiti 2132 drowned 2900 2850 18 5 caldera centered

Mo orea 3166 9 2850 2700 30 10 island centered

Tahaa 2784 4 2900 2700 21 12 caldera centered

Tahiti 12914 8 2900 2650 60 24 caldera centered

Tetiaroa 1300 atoll 2850 2650 11 4 lago on centered

Tupai 1442 atoll 2850 2800 111 5 lago on centered

Tuamotu-Gambier

Gambier 7287 drowned 2800 2600 34 15 caldera centered

Moruroa 4497 atoll 2600 2300 22 14 south lago on cent.

Rangiroa 14668 atoll 2750 2550 23 30 lago on centered

The half wavelength represents the anomaly width.

ine density , and, to a lesser extent, the in ll density no es. Machesky [1965] related p ositive Bouguer anom-

are the main factors controlling our mo dels. These alies to gabbros which outcrop in the caldera of Tahiti.

parameters are varied in the range 2300 to 2900 kg m More generally, Rymer and Brown [1986] prop osed a

and the residual anomaly is minimized by the technique classi cation of gravity anomalies in which p ositive an-

of least squares. omalies characterize mainly basaltic volcano es and are

caused by a relatively dense intrusive complex/magma

With this approach, we determined characteristic

b o dy which contrasts with its surroundings. Because

submarine and lling densities for each of the 25 islands

the FrenchPolynesian islands are basaltic shield volca-

considered Table 2. The average values of and

s f

no es, a pronounced p ositive anomaly mostly expresses

for all islands are 2700 and 2800 kg m , resp ectively

the dense cumulate core or solidi ed magmachamber

discussed later.

of the volcano. We will use known geological settings to

re ne the relation b etween a p ositive gravity anomaly

3. Results and Analysis

and a dense intrusive complex, in order to establish a

typ ology of gravity elds over the Polynesian volcano es.

Twenty of the 25 islands yield a gravity residual an-

Our results are sorted by archip elago and are shown in

omaly >5 mGal, the assumed uncertainty. The resid-

Table 2 and Table 3 according to the sign +/- of the

ual anomalies for these 20 islands are shown in Fig-

residual anomaly.

ures 5 to 8. To a rst approximation, the contours are

concentric, and their amplitude ranges from -20 to +60

3.1. So ciety Islands

mGal. Their wavelengths are similar to the diameter

of the volcano es, suggesting that the anomalies are re- The age of the o ceanic crust in the chain of the So-

lated to density contrasts at depth within the volca- ciety Islands ranges from 65 Ma in the south to 80 Ma

8180 CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS

a) b) Tahiti Nui 17˚30'S TAHITI -20

-10 10 10 MAIAO -15 40 20 -20 10 30 -30Tairapu 17˚39'S -10 0 5 10 0 6 km Tahiti Iti 2 km 149˚30'W 150˚36'W 16˚12'S TUPAI c) d) -20

TETIAROA 0 0 17˚00'S 10

2 km 2 km

149˚36'W 151˚48'W

-30 -20 -10 -5 0 5 10 15 20 25 30 40 50

mGal

Figure 5. Isostatic residual anomaly over individual islands of the So ciety Island chain where

a residual anomaly greater than 5 mGal is observed. a Tahiti, b Maiao, c Tetiaroa, d

Tupai, e Manuae, f Huahine, g Tahaa, h Mo orea, i Bora Bora, j Maupiti. Black dots

represent land gravity stations, and black dashed lines represent caldera b oundaries. Thick white

lines represent the coast, and thin white lines represent the marine gravity data.

750 km away in the north [Herron, 1972; Mayes et al., for the other islands studied here. Positive anomalies

1990]. This island chain is comp osed of ve atolls and are present at all of the islands and atolls Figure 5

nine islands of which Maupiti is the oldest ab out 4.3 Ma and range from 60 mGal over Tahiti Figure 5a, the

[Duncan and McDougal l, 1976] and Mehetia the young- largest island, to 11 mGal over Maiao Figure 5b, the

est 0 to 0.3 Ma [White and Duncan, 1996]. Because smallest island, and Tetiaroa Figure 5c, the smallest

Mehetia is asso ciated with the present hotsp ot, it has atoll Table 2. In the case of the islands the p ositive

b een intensely studied over the past 15 years [Talandier anomalies are lo cated over the inferred centers of the

and Okal, 1984a; Chemin ee et al., 1990; Binard et al., main feeding systems and havewavelengths similar to

1991]. The geo chronology, geo chemistry,volcanology, the diameter of the islands or the calderas. Positive

and geology of the So ciety Islands have b een reviewed anomalies are usually related to shield volcano es and

in some detail by Diraison et al. [1991] and more re- re ect a deep-seated, solidi ed magma chamb er or feed-

cently by White and Duncan [1996]. The aerial activity ing system [Rymer and Brown, 1986]. Thus the p ositive

of these volcano es can b e describ ed by three main se- anomalies over the atolls of Tetiaroa Figure 5c, Tupai

quential stages: the edi cation of a shield volcano, the Figure 5d, and Manuae Figure 5e may express un-

formation of a caldera, and the o ccurrence of p ostcal- derlying dense magma chamb ers.

dera volcanism.

Tahiti is comp osed of two di erentvolcano es, Tahiti

Figure 5 shows the residual gravity anomaly over 10 Nui and Tahiti Iti on the p eninsula. There are sucient

islands of the So ciety group. Owing to the relatively gravity data on Tahiti Nui to establish the existence of

dense coverage of bathymetry and gravity data, results a correlation b etween the p ositive anomaly and the cal-

for these islands are probably more reliable than those dera. On the other hand, the data coverage for Tahiti Iti

CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS 8181

e) i) MANUAE BORA BORA

-20 0 -10

16º 30'S 10

16º 36'S 3 km 3 km 154º 36'W 151º 48'W

f) HUAHINE g) Huahine Nui -10 16˚36'S

10 0

TAHAA Maroe -20 Bay 2 km 10 Vaihi 16˚48'S RAIATEA 16º 48'S 0 Huahine 3 km Iti

151º 00'W 151˚36'W

h) MOOREA

j) 16º 24'S MAUPITI

10 0 20 10 -30 3 km 0 17º 36'S 3 km

149º 48'W 152º 12'W

-30 -20 -10 -5 0 5 10 15 20 25 30 40 50

mGal

Figure 5. continued

is p o or, except for its western part, where a prominent tinct volcano es have b een p ostulated, suchasonTahiti

negative anomaly is observed. This negative anomaly is Iti. There is some debate ab out Huahine Figure 5f ,

not consistent with the existence of a third and separate which is comp osed of two separate islands lo cated in the

volcano as p ostulated by L eotot et al. [1990]. The neg- same lago on Huahine Nui and Huahine Iti, separated

ative anomaly of Tahiti Iti is discussed further b elow. by Maro e Bay and a large intrusion at MountVaihi.

The 10 mGal p ositive anomaly shown over the center of Deneufbourg [1965] considers that these two islands are

Tahiti Iti is not well constrained owing to the p o or data related to a single volcano separated by a large sub-

coverage on the eastern shore of the p eninsula and will merged caldera which formed Maro e Bay. In contrast,

not b e considered further. On Manuae Figure 5e the Brousse et al. [1983] p ostulate the presence of at least

anomaly is apparently lo cated close to the eastern side twovolcano es Huahine Nui and Huahine Iti. Figure

of this atoll. However, since there are no data for the 5f clearly indicates one p ositive anomaly centered over

western and southern sides, the lo cation of the anomaly Maro e Bay, which supp orts Deneufb ourg's interpreta-

is p o orly de ned. tion. A di erentinterpretation may b e appropriate in

Only one large p ositive anomaly is presentover each

the case of the islands of Tahaa Figure 5g and Rai-

island examined even though in some cases two dis-

atea, which are in the same lago on. Each island has

8182 CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS

b) RURUTU a) 23˚48'S 10 0

RAIVAVAE 0

10 22˚30'S

3 km 3 km 147˚36'W 151˚24'W

c) d) TUBUAI RAPA

Herani 10 27˚36'S

20 23˚24'S Hanareho 3 km

3 km -10

144˚24'W 149˚24'W

-30 -20 -10 -5 0 5 10 15 20 25 30 40 50

mGal

Figure 6. Isostatic residual anomaly over individual islands of the Austral Island chain where

an anomaly greater than 5 mGal is observed. a Rapa, b Rurutu, c Tubuai, d Raivavae.

Black dots represent land gravity stations, and black dashed lines represent caldera b oundaries.

Thick white lines represent the coast, and thin white lines represent the marine gravity data.

could also have pro duced the observed negative anom- a caldera. The two calderas are separated by25km.

aly on the coral reef of Bora Bora Figure 5i. Figure 5g shows a p ositive anomaly centered on the cal-

dera of Tahaa. Unfortunately, no gravity survey has

b een conducted on Raiatea. However the gravity resid-

3.2. Austral Islands

ual anomaly at Tahaa seems to show a separate feeding

The Austral Islands chain extends northwest for >

system.

1500 km from the active submarine volcano Macdon-

Mo orea Figure 5h has a 30 mGal anomaly, o set

ald [Norris and Johnson, 1969; Johnson and Malaho ,

by ab out 3 km from the center of a caldera, while Bora

1971; Talandier and Okal, 1984b] to the island of Rim-

Bora Figure 5i and Maupiti Figure 5j have 16 mGal

atara, which is ab out 16 Myr old [Turner and Jarrard,

and 18 mGal gravity anomalies, resp ectively, centered

1982], and the atoll of Maria Figure 1. The chain is

over asso ciated calderas.

comp osed of ve small islands and one atoll. The age

Thus each island in the So cietychain apparently has

progression of these islands is compatible with a hot-

only one deep-seated magma chamb er or feeding system.

sp ot origin, with the presentvolcanic activity related

Also, it is found that a linear correlation exists b etween

to the Macdonald submarine volcano [Duncan and Mc-

the amplitude of the p ositive anomalies and the island

Dougal l, 1976]. Young radiometric ages rep orted for

size for b oth islands and atolls.

Rurutu, lo cated toward the northern end of the chain,

Negative anomalies Table 3 are presentover Tahiti

can b e explained by another hotsp ot lo cated b etween

Iti at Taiarapu Figure 5a and Mo orea Figure 5g. On

Rurutu and Tubuai [Turner and Jarrard, 1982]. The

the basis of geological data [L eotot et al., 1990] those

existence of this hotsp ot was established by demonstrat-

anomalies can b e interpreted as the gravity e ects of

ing the di erences of chemical and isotopic signatures

secondary cones pro ducing di erentiated lavas of relat-

of the recentlavas from Rurutu with those pro duced at

ively low density. A large volume of limestone dep osits

the Macdonald hotsp ot [Barsczus et al., 1994]. Recent

CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS 8183

northernmost island, and Maria, a small atoll lo cated

west of Rimatara. a) to the north

UA POU

The lo cations of the p ositive anomalies in the Austral

e those in the So ciety Islands, are over the 5 km Islands, lik

9˚24'S

ter of the island in Rapa and Rurutu but over the

0 cen

western shore in Tubuai and over the southern shore

in Raivavae. The continuation of the anomaly to the

-40

avae is probably due to a complex mor- -10 south of Raiv

140˚00'W phology of the southern submarine ank. The lo cation

and uniqueness of the gravity anomaly show a single HIVA OA 0

Puamau magma source and imply that the so-called small cal-

dera see Figure 6d is likely related to the collapse of

10 Ootua a) 30 9˚48'S -20 GAMBIER 20Atuona -30 5 km -10 -10

0 139˚00'W c) 30 NUKU HIVA 10 -10 8˚48'S -30 20 23 12'S

0 5 km 10 20

Taiohae 135 00'W 5 km b) MORUROA 0 140˚00'W 21 48'S 5 km

-30 -20 -10 -5 0 5 10 15 20 25 30 40 50 -10 mGal 0

Figure 7. Isostatic residual anomaly over individual

islands of the Marquesas Island chain: a Ua Pou, b

a Oa, c Nuku Hiva. Black dots represent land

Hiv 138 48'W

vity stations, and black dashed lines represent cal-

gra c)

k white lines represent the coast,

dera b oundaries. Thic RANGIROA

and thin white lines represent the marine gravity data.

15 00'S

bathymetric and seismic data rep orted by McNutt et

[1997] reveal the complexityofoverlapping volcan-

al. 10 10 10

ism at the southeastern end of the Austral chain. The

0 0

hain b etween Tubuai and

Austral FZ crosses the island c 5 km

avae, forming a western and eastern segment of the

Raiv 10

Austral island chain, each showing a distinct isotopic

Barsczus et al., 1994; Hauri et al., 1997].

signature [ 148W

The ages of the o ceanic crust in the chain vary from

ab out 35 Ma to 80 Ma [Mayes et al., 1990].

-5 5 15 25

ws the residual anomaly over the Aus-

Figure 6 sho -30 -20 -10 0 10 20 30 40 50

Positive anomalies range from 16 mGal

tral Islands. mGal

at Raivavae Figure 6a and Rurutu Figure 6b to 26

Figure 8. Isostatic residual anomaly over a Gambier

mGal at Tubuai Figure 6c. Rapa Figure 6d presents

Islands, b Moruroa, c Rangiroa. Black dots repre-

a 17 mGal p ositive anomaly. These p ositive anomalies

sent land gravity stations. Thick white lines represent

are interpreted as expressing a single and deep solidi-

the coast, and thin white lines represent the marine

ed magma chamb er or feeding system. No signi cant

gravity data.

anomalies are observed at Rimatara, the smallest and

8184 CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS

Table 3. Prop erties of Islands Showing Negative Residual Isostatic Anomalies

Island Amplitude, mGal Half Wavelength, km Lo cation Interpretation

Bora Bora 20 4 north reef low density sup er cial b o dies

Mo orea 10 3 Maraarii secondary cone di erentiated lava

Nuku Hiva 30 10 Topuke stromb olian cone di erentiated lava

Tahiti Iti 10 6.5 NW secondary cones di erentiated lava

The half wavelength represents the anomaly width.

the anks of the volcano. On Tubuai the o set of the 3.4. Pitcairn-Gambier Islands

p ositive anomaly from the center of the island may in-

dicate subsidence of this part of the island due to an

The Pitcairn-Gambier island chain extends over 1650

underlying dense source Figure 6c, similar to that ob-

km from Hereheretue atoll to Pitcairn Island. Known

served in the Marquesas Islands [Chubb, 1930; Barsczus

K-Ar ages range from ab out 11 Ma at Moruroa [Guil le et

et al., 1992]; the anomaly pattern strongly suggests the

al., 1993] to zero in the active hotsp ot area, 70 to 100 km

presence of a single magma source. southeast of Pitcairn [Binard et al., 1992]. The Austral

FZ crosses this chain b etween Moruroa and Fangataufa.

The age of o ceanic crust ranges from ab out 25 Ma in the

southeast to 55 Ma to the northwest [Mayes et al., 1990;

3.3. Marquesas Islands

Munschy et al., 1996]. The chain includes the islands

Pitcairn and the Gambier and several atolls. Fangata-

The length of the Marquesas island chain is only 350

ufa and Moruroa have b een extensively drilled and thus

km, but it comprises a dozen islands and as many shal-

considerable data on the prop erties and geology of these

low seamounts. The age of the o ceanic crust ranges from

islands exist [see Guil le et al., 1993].

50 to 55 Ma [Kruse, 1988; Mayes et al., 1990; Munschy

Over the Gambier Islands Figure 8a, a large p osit-

et al., 1996]. K-Ar ages range from ab out 5.5 Ma at

ive anomaly 34 mGal o ccurs over the 10 islets in the

Eiao in the northwest to ab out 1.5 Ma at Fatu Ivain

center of a large lago on. These islets are the emerged

the southeast [Brousse et al., 1990]. A basalt sample

parts of an old volcano, and a caldera centered over the

dredged 50 km southeast of Fatu Iva yielded an age of

lago on has b een assumed [Guil le, 1974]. The at shap e

ab out 0.5 Ma [Desonie et al., 1993]. The Marquesas

of the gravity anomaly is probably related to the lack

FZ is lo cated southeast of the chain. To date, no active

of data in the lago on, and the anomaly magnitude may

hotsp ot generating the Marquesas volcanism has b een

actually b e a few milliGals more than that indicated.

identi ed.

Over the atoll of Moruroa Figure 8b, a p ositive anom-

Owing to the rugged top ography, gravity stations are

aly of 22 mGal is centered over the southern rim of the

sparse Figure 7, thus the resulting interpretations are

atoll. The p ositive gravity anomalies along the Pitcairn-

somewhat sp eculative. The shipb oard data coverage

Gambier island chain are interpreted as the remains of

around the islands is also p o or. The merging of these

ancient magmachamb ers, which coincide with caldera

two p o or data sets probably accounts for the non cir-

lo cations for Moruroa [Buigues et al., 1992], and for the

cular residual anomalies. The results for Ua Pou Fig-

Gambier Islands [Brousse et al., 1972].

ure 7a will not b e considered further b ecause of the

p o or distribution of data. Hiva Oa Figure 7b has a

prominent p ositive anomaly 35 mGal over the Atuona

3.5. Tuamotu Islands

caldera. It is p ossible that a separate p ositive anomaly

with a lower magnitude exists over the Puamau valley,

but the lack of o shore data prevents any clear descrip- Oriented N120 , the Tuamotu Archip elago consists

tion of the anomaly. The observations at Hiva Oa in- of ab out 60 atolls along two subparallel chains of length

dicate that at least one deep, solidi ed magma chamber 1200 km and width 400 km. Its northern and southern

is present at Hiva Oa. Obel lianne [1955] prop osed two termini are lo cated at the Marquesas Fracture Zone and

calderas, Atuona and Puamau. These two calderas are the Austral Fracture Zone, resp ectively. The age of the

separated by Mount Ootua, a trachytic diatreme, ex- sea o or ranges from ab out 30 Ma in the southeast to

pressed as 5 mGal negative anomaly. 65 Ma in the northwest [Mayes et al., 1990; Munschy

et al., 1996]. The atolls probably formed atop buried

The anomaly pattern obtained at Nuku Hiva Figure

volcano es.

7c may b e misleading b ecause of the p o or distribution

Among the four atolls for which gravity data are of data. However, the p ositive anomaly 29 mGal over

available, only Rangiroa Figure 8c shows p ositive an- the southern coast, where the data coverage is relatively

omalies. The eastern anomaly 23 mGal is only de ned good, may b e asso ciated with the caldera that forms

by one station and will not b e considered further. By Taiohae Bay.

CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS 8185

0 Sea c -5 a

-10 Crust Depth (km) Mantle

0 20 40 60 80 100 (km)

Figure 9. West-east cross section along latitude 19 06 N of Mauna Loa volcano, Hawai'i, extend-

ing from land to sea west-east [after Moore et al., 1995]. Horizontal layering indicates subaerial

lava; dashed layering indicates fragmental lava; ellipses indicate pillowlava; vertical lines indicate

sheeted dikes; dotted pattern indicates gabbro; unpatterned areas are giant blo cks of rst phase

of South Kona landslide; and solid black areas are magma, or plastic core of southwest rift zone

of Mauna Loa. The 3-D ellipsoidal chamb er used to mo del the anomaly see Figure 11 with a

p ositive density contrast is also drawn.

Society Islands 50 Gambier Islands Austral Islands Marquesas Islands 40

20 Rangiroa

10 residual isostatic anomaly (mGal)

0 0 2000 4000 6000 8000 10000 12000 14000

volcano volume (km 3 )

Figure 10. Plot of the maximum intensity of the p ositive residual anomaly mGal versus island

volume km . Shaded area represents the 5 mGal error asso ciated with the observed gravity

values. The central thick line represents the derived linear relation b etween volcano volume and

anomaly amplitude.

analogy with results obtained for other islands esp e- Hawaiian volcano [Eaton and Murata, 1960] prop osed

cially the Moruroa atoll weinterpret the western an- that the magma ascends due to buoyancy to a depth

omaly 17 mGal as the gravity signature of an ancient between 3 and 7 km, where it forms a shallow magma

magma chamb er and/or deep-seated feeding system. It reservoir. Intermittent eruptions o ccur within summit

is p ossible that the amplitude of this anomaly could calderas or laterally along rift zones. Ryan [1987] for-

have b een reduced due to the presence of a 2-km-thick mulated the neutral buoyancy hyp othesis in which the

coral limestone cap [Talandier and Okal, 1987] overly- magma ascends to depths where its density equals that

ing the prop osed magma chamb er. For the three other, of the surrounding material. This region of neutral

smaller atolls of Makatea, Manihi, and Mataiva not buoyancy controls the upp er limit of the levels of stable

shown, no appreciable anomaly is observed. magma storage. The summit reservoir elevation follows

the buildup of the volcano, leading to the lo cation of

the ro of at depths b etween 2 and 4 km, as observed on

4. Discussion

Kilauea and Mauna Loa volcano es Figure 9 [Moore

et al., 1995]. Another characteristic of hotsp ot volca-

Current understanding of volcanic edi ces asso ciated

no es is the existence of collapses. The Hawaiian Ridge

with hotsp ots and their tectonic b ehavior has b een de-

is characterized by giant submarine landslide debris,

rived from numerous studies, mainly of the Hawaiian

which playedakey role in the internal makeup of the

volcano es. The rst dynamic cross section of a typical

8186 CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS

volcano es [Moore et al., 1989]. These landslides can single relation b etween the size of the solidi ed magma

havevarious origins [e.g., Walker, 1988]: vertical sub- plumbing system and the present size of the volcano.

sidence by withdrawal of magma b eneath the summit Using a more realistic mo del, we examine usefulness of

area or lateral mass collapse. Using our gravity data this relation.

set, we will try to compare the Polynesian volcano es to

In most cases, the amplitudes of the gravity anom-

this classical Hawaiian mo del.

alies are based on a single pro le across the island. The

anomaly can b e mo deled, in a simple way, as the ef-

4.1. Gravity Anomalies

fect of a dense ellipsoid of revolution Figure 11. More

Positive residual gravity anomalies are calculated

sophisticated inverse or forward mo dels would b e unin-

over 20 islands, with amplitudes ranging from 9 to 60

formative in the absence of any other data constraints,

mGal Table 2. Generally, the anomalies are spatially

such as seismic or b orehole data. The selected ellips-

correlated with a known caldera e.g., Tahiti or the

oidal b o dy do es not re ect the complexity of magma

assumed main feeding conduit of volcano es of similar

reservoirs and feeding conduits. Nevertheless, it is a

widths. Thus solidi ed magma chamb ers and feeding

reasonable representation of the depth and volume of

conduits with higher densities than that of the sur-

the main system. Using this simple approachofan

rounding ro cks [Rymer and Brown, 1986; Rousset et

ellipsoidal magma chamb er, the radius, the attening,

al., 1989] pro duce the observed p ositive gravity anom-

and the depth of the chamb er of each island can b e de-

alies. Furthermore, for each island where the gravity

rived as a function of the residual anomaly.Atapoint

coverage is go o d the observed single p ositive anomaly

P lo cated at a distance r from the center of the ellipsoid

suggests that there is only one deep-seated, solidi ed

the vertical attraction g of an ellipsoid of revolution is

magma chamb er or main feeding system.

2 2 2 2 2

3a c 2z r 4a cGz

We also observe a go o d correlation Figure 10 b et-

[1 ] 4 g =

2 2 2

2 2 3=2

10r + z

3r + z

ween the amplitude of the p ositive anomaly and the

volume of the volcano. This volume is that b ounded

where z is the depth of the center of the ellipsoid, 2a is

by the volcano top ography and a reference plane cor-

the diameter of the circular section of the ellipsoid, c is

resp onding to the observed basis of the volcano on the

the length of the vertical axis of revolution see Figure

sea o or. Sp eci cally, the linear relation b etween the

9, is the density contrast, and G is the gravitational

volcano volume V and the amplitude of the p ositive

constant6:67 10 N m. The average density for

anomaly A is

ultrama c b o dies forming dense magma chamb ers or

feeding systems is roughly 3100-3300 kg m [Rymer

A =3:7 10 V +8:9; 3

p v

and Brown, 1986]: we used a value of 3100 kg m .

where V is in km and A is in mGal. This relation

The density contrast dep ends on the submarine load

v p

fails for the Rangiroa atoll, however, where the grav-

density assumed for each island, except for Rangiroa,

ity signal is lower than exp ected considering the vol-

where we set to 500 kg m .To constrain our res-

cano diameter. It has already b een p ointed out that

ults further, we also assume that the magma chamber

the amplitude of the gravity anomaly at Rangiroa may

depth lies b etween 2 and 4 km, the depth of neutral

have b een reduced due to a 2 km-thick coral limestone

buoyancy [Ryan, 1987]. Considering that the islands

cap. Thus the volcanic basement and the dense residue

have reached an elevation of 2000 m ab ove sea level,

of magmatic ro ck are lo cated at a greater depth than

the magma chamber would lie at depth of 0 m to 2

that for the other islands.

km b elow the sea level. Although the neutral buoyancy

region has b een de ned for activevolcano es, wehave

Negative anomalies Table 3 are observed in a few

assumed it applies here b ecause of the young ages of

cases, although their magnitude never reaches that of

the volcano es in our study area except Rangiroa. The

the p ositive anomalies. Moreover, in contrast to the

nal parameters of the ellipsoids, a, c and z , are those

p ositive anomalies, the negative anomalies are observed

that minimize the di erence b etween the isostatic resid-

on the island or atoll edge, p ossibly b ecause of lime-

ual anomalies and the attraction of the ellipsoidal b o dy

stone on reefs or fragmented material on the ank of

given by equation 4.

the edi ces.

For each island, Table 4 gives the diameter and the

4.2. Size and Depth of Solidi ed Magma

depth of the magmachamb er ro of. For the So ciety Is-

Chamb ers

lands the magma chamb er ro of depth varies from 100

Residual gravity anomalies are asso ciated with shal- m for Maiao, Maupiti, Mo orea and Tupai to 1.3 km for

low mass inhomogeneities. Robertson [1967] tried to Tahiti. The diameter of the chamb er ranges from 2 km

mo del the gravity anomalies observed on six of the Co ok for Maiao to 14 km for Tahiti. For the Austral Islands

Islands as caused byvertical cylindrical dense cores with the sizes of chamb er range from 3.6 to 7 km. The Gam-

a radius equal to the radius of the island at sea level. bier Islands, Rangiroa, and Marquesas islands generally

Despite the simplicity of the mo del, which uses a uni- have larger magma chamb ers from 8.4 to 11.6 km in

form density for the whole island, Rob ertson found a diameter lying at greater depths up to 5.9 km. Plot-

CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS 8187

80 Observed residual anomaly Gravity induced by a buried ellipsoidal body 60 17 30'S 0

40 -20 20

Gravity anomaly (mGal) 20 149W

0 0

-4 c a yy

Depth (km) -8

0 yy4 8 12 16 20 24 28 32 36 40

Length of the profile across Tahiti (km)

Figure 11. Gravity anomaly pro duced by a 3-D ellipsoidal chamb er with a p ositive density

contrast see text with the surrounding ro cks on Tahiti. The isostatic residual anomaly in

milliGals is represented as a dotted line, and the computed anomaly is represented as a solid

line. The inset shows the lo cation of the pro le on Tahiti Island.

for all the islands Table 2. On this lo cal scale the ting the magma chamber volume V km versus the

density of the masses ab ove sea level has a ma jor ef- volcano volume V km Figure 12 gives the following

fect on the terrain correction. An average value of 2300 linear relation:

kg m for , the density of the aerial part of the is-

V =7:14 10 V 91: 5

m v

lands, provides suitable results for all the islands. The

value of the lling density was determined jointly by

This relation gives a metho d to infer the rough size of

the same RMS minimization metho d and varies from

the solidi ed magmachamb er using only the volcano

volume. To compare with other intraplate volcano es, we

plotted data from the Kilauea [Ryan, 1988] and Piton

Reunion Island

Lesquer, 1990] volcano es. Although the size

des Neiges [ Society Islands olcano es is dicult to estimate since they can of these v 1000 Gambier Islands

3 Austral Islands

yvolcano es, not b e easily separated from other nearb Marquesas Islands

Rangiroa

the Piton des Neiges p oint on the graph ts well with

Kilauea

the linear curve determined for the islands of French

100

Polynesia. The prop osed relation in equation 5 do es

not apply for Kilauea, probably b ecause the volume of

this volcano is very dicult to estimate. Therefore our

metho d probably applies only to extinct volcano es and

thus entirely solidi ed magmachamb ers. magma chamber volume (km )

1 y 4.3. Load Densit 1000 10000

volcano volume (km 3 )

On a regional scale a submarine volcanic load density

Figure 12. Computed solidi ed magma chamber

of 2700 kg m provides reasonable results for the cal-

volume versus the island volume using a logarithmic

culated de ection surface and for the theoretical FAA

representation. The trend of the plot in gray corres-

Figure 4. This value maythus suitably represent the

p onds to a linear relation. FrenchPolynesian volcano es

actual average regional load density for volcanic island

are represented by solid geometrical symb ols see le-

chains. For individual islands, however, the RMS error

gend, and volcano es from Hawai'i and Reunion Islands

between theoretical and observed FAA leads to a range

are plotted for comparison using op en circles. of submarine load densities, from 2300 to 2900 kg m

8188 CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS

Table 4. Characteristics of Magma Chamb ers and Comparison With the Size of the Caldera

Caldera Solidi ed Magma Chamber = 3100 kg m

Island Diameter, Center Center Diameter Flattening Volume, Ro of Depth,

km Longitude W Latitude S a,km c=a km km

Austral

0 0

Raivavae 3 147 39 23 53 5.4 0.85 70 0.2

0 0

Rapa 3.5 and 5 144 21 27 36 3.6 0.83 20 0.2

0 0

Rurutu 152 21 22 29 3.8 1.0 29 0.1

0 0

Tubuai 2 149 31 23 23 7.0 0.77 138 0.2

Marquesas

0 0

HivaOa 10 and 15 139 03 09 48 8.4 1.52 473 0.3

0 0

Nuku Hiva 6 and 9 140 06 08 54 8.4 1.43 443 0.2

So ciety

0 0

Bora Bora 4 151 45 16 31 6.4 0.78 107 0.2

0 0

Huahine 151 00 16 46 5.2 0.85 62 0.3

0 0

Maiao 150 37 17 40 2.0 0.9 4 0.1

0 0

Manuae 154 39 16 32 4.0 0.75 25 0.3

0 0

Maupiti 3 152 16 16 27 4.0 1.0 33 0.1

0 0

Mo orea 7 149 49 17 32 6.2 0.97 121 0.1

0 0

Tahaa 4 151 28 16 38 6.2 0.81 101 0.8

0 0

Tahiti 7 149 26 17 38 14.0 0.77 1108 1.3

0 0

Tetiaroa 149 34 17 01 2.8 1.0 12 0.3

0 0

Tupai 151 49 16 16 3.6 0.72 18 0.1

Tuamotu-Gambier

0 0

Gambier 12 134 58 23 09 9.6 1.04 482 1.6

0 0

Moruroa 138 54 21 52 7.2 0.83 163 1.8

0 0

Rangiroa 147 48 15 05 11.6 0.86 705 5.9

Po orly constrained.

2800 to 2900 kg m , except for Moruroa and Rangiroa, Tubuai, Maupiti, Tahiti Nui, and Bora Bora are prob-

where it is lower. ably characterized by classic calderas centered ab ove the

magma chamb er. Even for these edi ces, however, the

caldera ank is always op en to the sea, and this could

4.4. Calderas or Collapses?

b e related to ank instabilities.

Are the computed lo cations and diameters of the an-

cient magmachamb ers consistent with those of known

calderas? It app ears that the collapse diameter is some-

5. Conclusion

times larger than the computed diameter of the asso ci-

A gravity study of 25 volcanic islands in FrenchPoly-

ated magma chamb er. In addition, the collapses are

nesia leads us to prop ose a single linear relation b etween

not always lo cated directly ab ove the chamb ers them-

the island size and the solidi ed magma chamb er size,

selves e.g., north of Huahine. Considering the inferred

the latter b eing derived from the p ositive isostatic re-

magma chamb er lo cation, the collapse of Tahaa is shif-

sidual gravity anomaly. This linear relation app ears ap-

ted west, that of Mo orea is shifted north, and those of

propriate for other hotsp ot volcano es, such as Piton des

Huahine and Raivavae are clearly indep endent of the

Neiges in Reunion Island. This relation may b e used as

lo cation of the magma chamb er. Hiva Oa, Nuku Hiva,

a simple and viable to ol to infer the rough size of the

and Rapa have outer collapses larger than the prop osed

solidi ed magmatic feeding system for extinct o ceanic

solidi ed reservoirs. These collapses are probably not

intraplate volcano.

linked to any caldera subsidence but rather to ank in-

stabilities e.g., landslides such as describ ed on other The results also illustrate the wide range of submar-

intraplate volcano es suchasHawaii or Reunion islands ine volcanic load densities, from 2300 to 2900 kg m ,

[Moore et al., 1989; L enat and Labazuy, 1990]. How- and con rm the imp ortance of the increase of density

ever, inner collapses of Rapa and Hiva Oa, the eastern with depth, from 2300 kg m for subaerial lava to 2900

collapse of Tahaa, and the collapses of Gambier Islands, kg m for the deep est part of the volcano which lls

CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS 8189

the exure of the o ceanic crust. On a regional scale the Bonnevill e, A., V. Clouard, P. Beuzart, I. Klaucke, R. Le

Suav e, B. Loubrieu, P. Saget, and Y. Thomas, Possible

b est tting load densityis2700 kg m .

control of So ciety Islands volcanism by a preexisting vol-

At the island scale the residual isostatic anomaly

canic chain abstract, Eos Trans. AGU, 78,45, Fall

helps to de ne the actual lo cation of volcano es. For

Meet. Suppl., F725, 1997.

example, previous studies assume multiple volcano es

Brousse, R., J.C. Philipp et, G. Guille, and H. Bellon, G eo-

on the same island e.g., Huahine, Tubuai, Tahiti Iti.

chronom etrie des ^ les Gambier Oc ean Paci que, C. R.

However, our results clearly demonstrate the presence

Acad. Sci., Ser. D, 274, 1995-1998, 1972.

of a single volcanic system for these islands and for the

Brousse, R., C. Macherey, E. Berger, and G. Boutault,

atolls, except for the Marquesas Islands. Finally,we

L'ile de Huahine: trois volcans successifs Archip el de

la So ci et e, Polyn esie, C. R. Acad. Sci., Ser. II, 296,

show that the ma jority of the so-called calderas are re-

1559-1562, 1983.

lated to ank instabilities which induce huge landslides,

Brousse, R., H.G. Barsczus, H. Bellon, J.M. Cantagrel,

rather than to caldera subsidence. Thus the caldera

C. Diraison, H. Guillou, and C. L eotot, Les Marquises

center may often b e a p o or indicator of the lo cations

Polyn esie Francaise: Volcanologie, g eochronologie, dis-

and other attributes of now consolidated and inactive

cussion d'un mo d ele de p ointchaud, Bul l. Soc. G eol.

ancient magma reservoirs.

Fr., 6, 933-949, 1990.

Buigues, D., A. Gachon, and G. Guille, L'atoll de Moruroa

Acknowledgments. We are grateful for ORSTOM in-

Polyn esie Francaise: Structure et evolution g eologique,

stitute, now IRD, that funded all the gravity surveys con-

Bul l. Soc. G eol. Fr., 5, 645-657, 1992.

ducted by one of us H.G.B. in FrenchPolynesia b etween

Caress, D.W., M.K. McNutt, R.C. Detrick, and J.C. Mutter,

1978 and 1981 and to I. Leroy who made the complete grav-

Seismic imaging of hotsp ot-related crustal underplating

ity survey of Tahiti in 1994. V.C. is supp orted by a grantof

b eneath the Marquesas Islands, Nature, 373, 600, 1995.

the ZEPOLYF program. We thank particularly the asso ci-

Calmant, S., and A. Cazenave, The e ective elastic litho-

ate editor, T. Hildenbrand, whose remarks and suggestions

sphere under the Co ok Austral and So ciety Islands, Earth

have greatly improved this pap er. This work also b ene ted

Planet. Sci. Lett., 77,, 187-202, 1986.

of the constructive reviews made by J.F. L enat, M.P.Ryan,

Calmant, S., and A. Cazenave, Anomalous elastic thickness

and H. Rymer.

of the o ceanic lithosphere in the south central Paci c,

Nature, 328, 236-238, 1987.

Chemin ee, J.L., R. H ekinian, J. Talandier, F. Albar ede,

References

C.W. Devey,J.Francheteau, and Y. Lancelot, Campagne

Barriot, J.P., La d etermination du g eoide par altim etrie

Teahitia-1, G eologie d'un p ointchaud actif: Teahitia-

oc eanique et gravim etrie : Quelques asp ects du traite-

Mehetia, ^ les de la So ci et ePaci que Central Sud, Ocea-

mentetinterpr etation g eologique sur l'Oc ean Indien partie

nol. Acta, 10, 225-239, 1990.

NO et M editerrann ee o ccidentale. Ph.D. thesis, 240 pp,

Chubb, L.J., The geology of the Marquesas Islands, Bul l.

Univ. des Sciences et Techniques du Languedo c, Mont-

68, B.P. Bishop Mus., Honolulu, Hawaii, 1930.

p ellier, France, 1987.

Dalrymple, G.B., R.D. Jarrard, and D.A. Clague, Potas-

Barsczus, H.G., R esultats provisoires de quelques mesures et

sium-argon ages of some volcanic ro cks from the Co ok

liaisons gravim etriques e ectu ees en Polyn esie Francaise,

and Austral Islands, Geol. Soc. Am. Bul l., 86, 1463-1467,

Intern. Rep., Cent. ORSTOM de Pap eete, Tahiti, 1979.

1975.

Barsczus, H.G., Reconnaissance gravim etrique de Tubuai

Danob eitia, J.J., et al., Deep seismic reconnaissance b eneath

Iles Australes, Intern. Rep., Cent. ORSTOM de Pa-

So ciety Archip elago South Paci c, pap er presented at

p eete, Tahiti, 1980.

EGS Assembly, Eur. Geophys. So c., Hamburg, Germany,

Barsczus, H.G., Reconnaissance gravim etrique de Maiao

1995.

Tupai-Manu, Intern. Rep., Cent. ORSTOM de Pa-

Deneufb ourg, G., Carte g eologique, Notice explicative sur la

p eete, Tahiti, 1981.

feuille Huahine, Bur. des Rec. G eol. et Mini er., Paris,

Barsczus, H.G., P.E. Filmer, and D.L. Desonie, Cataclys-

1965.

mic collapses and mass wasting pro cesses in the Marque-

Desonie, D. L., R.A. Duncan, and J.H. Natland, Temp oral

sas abstract, Eos Trans. AGU, 73 14, Spring Meet.

and geo chemical variabilityofvolcanic pro ducts of the

Suppl., 313, 1992.

Marquesas hotsp ot, J. Geophys. Res., 98, 17,649-17,665,

Barsczus, H.G., G. Guille, R.C. Maury, C. Chauvel, and H.

1993.

Guillou, Two magmatic sources at Rurutu Island Aus-

Diraison, C., Le volcanisme a erien des archip els p olyn esiens

tral Islands, FrenchPolynesia and the Austral "Hotline"

de la So ci et e, des Marquises et des Australes-Co ok, Ph.D.

abstract, Eos Trans. AGU, 7544, Fall Meet. Suppl.,

thesis, 413 pp, Univ. de Bretagne Occidentale, Brest,

323, 1994.

France, 1991.

Binard, N., R. H ekinian, J.L. Chemin ee, R.C. Searle, and

Diraison, C., H. Bellon, C. Leotot, R. Brousse, and H.

P. Sto ers, Morphologica l and structural studies of the

Barsczus, L'alignementdelaSoci et ePolyn esie fran-

So ciety and Austral hot sp ot regions in the South Paci c,

caise: Volcanologie, g eo chronologie, prop osition d'un

Tectonophysics, 186, 293-312, 1991.

mo d ele de p ointchaud, Bul l. Soc. G eol. Fr., 162, 479-

Binard, N., R. H ekinian, and P. Sto ers, Morphostructural

496, 1991.

study and typ e of volcanism of submarine volcano es over

Duncan, R.A., and I. McDougall, Migration of volcanism

the Pitcairn hotsp ot in the South Paci c, Tectonophysics,

with time in the Marquesas Islands, FrenchPolynesia,

206, 245-264, 1992.

Earth Planet. Sci. Lett., 21, 414-420, 1974.

Blakely, R.J., Potential Theory in Gravity and Magnetic Ap-

Duncan, R.A., and I. McDougall, Linear volcanism in French

plications, 441 pp., Cambridge Univ. Press, New York,

Polynesia, J. Volcanol. Geotherm. Res., 13, 197-227, 1995.

8190 CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS

Lesquer, A., Deep structure of Reunion Island from grav- 1976.

ity data, in Le volcanisme de la R eunion-Monographie, Duncan, R.A., M.R. Fisk, W.M. White, and R.L. Nielsen,

edited by J.F. L enat, pp 19-27, Cent. Rech. Volcanol., Tahiti: Geo chemical evolution of a FrenchPolynesian vol-

Clermont-Ferrand, France, 1990. cano, J. Geophys. Res., 99, 24,341-24,357, 1994.

Macheski, L.F., Gravity relations in American Samoa and Eaton, J.P., and K.J. Murata, Howvolcano es grow, Science,

the So ciety Islands, Paci c Sci., 19, 367-373, 1965. 132, 925-938, 1960.

Mammerickx, J., E. Herron, and L. Dorman, Evidence for Filmer, P.E., Flexure of the o ceanic lithosphere in the vicin-

two fossil spreading ridges in the southeast Paci c, Geol. ity of the Marquesas Islands, Ph.D. thesis, 234 pp, Mass.

Soc. Am. Bul l., 91, 263-271, 1980. Inst. of Technol., Cambridge, 1991.

Mayes, C.L., L.A. Lawver, and D.T. Sandwell, Tectonic his- Filmer, P.E., M.K. McNutt, and C.J. Wolfe, Elastic thick-

tory and new iso chron chart of the south Paci c, J. Geo- ness of the lithosphere in the Marquesas and So ciety Is-

phys. Res., 95 8543-8567, 1990. lands, J. Geophys. Res., 98, 19,565-19,577, 1993.

McNutt, M.K., Lithospheric exure and thermal anomalies, Gillot, P.Y., Y. Cornette, and G. Guille, Age K-Ar et con-

J. Geophys. Res., 89, 11,180-11,194, 1984. ditions d' edi cation du soubassementvolcanique de l'atoll

McNutt, M.K., Sup erswells, Rev. Geophys., 362, 211-244, de Moruroa Paci que Sud, C. R. Acad. Sci., Ser. II,

1998. 314, 393-399, 1992.

McNutt, M.K., D.W. Caress, J. Reynolds, K.A. Jordahl, and Go o dwillie, A.M., and A.B. Watts, An altimetric and ba-

R.A.Duncan, Failure of plume theory to explain midplate thymetric study of elastic thickness in the central Paci c

volcanism in the southern Austral islands, Nature, 389, Ocean, Earth Planet. Sci. Lett., 118, 311-326, 1993.

479-482, 1997. Guille, G., Contributional' etude g eologique des ^ les de

Minster, J.B., and T.G. Jordan, Present-day plate motions, Polyn esie francaise: Les Gambiers, Ph.D. thesis, 215 pp,

J. Geophys. Res., 83, 5331-5354, 1978. Univ. Paris XI, Orsay,France, 1974.

Mo ore, J.G., W.B. Bryan, M.H. Besson, and W.R. Normark, Guille, G., R.C. Maury, D. Buigues, H. Bellon, A. Gachon,

Giant blo cks in the South Kona landslide, Hawaii, Geo- and M. Caro , L'Atoll de Moruroa Polyn esie francaise;

logy, 23, 125-128, 1995. I I I, Conclusions g en erales, Bul l. Soc. Geol. Fr., 163,

Mo ore, J.G ., D.A. Clague, R.T. Holcomb, P.W. Lipman, 681-685, 1992.

W.R. Normark, and M.E. Torresan, Pro digious submar- Guille, G., G. Gouti ere, and J.F. Sornein, Les atol ls de

ine landslide s on the Hawaiian ridge, J. Geophys. Res., MururoaetdeFangataufa Polyn esie Francaise, vol I.

94, 17,465-17,484, 1989. G eologie, P etrologie, Hydrog eologie, 168 pp., Masson, Pa-

Munschy, M., C. Antoine, and A. Gachon, Evolution tecto- ris, 1993.

nique de la r egion des Tuamotu, o c ean Paci que central, Guillou, H., P.Y. Gillot, and G. Guille, Age K-Ar et p os-

C. R. Acad. Sci., Ser. II, 323, 941-948, 1996. ition des ^ les Gambier dans l'alignement du p ointchaud

Nakagawa, I., et al., Precise calibration of scale values of de Pitcairn Paci que Sud, C. R. Acad. Sci., Ser. I I,

Lacoste & Romb erg gravimeters and international gravi- 318, 635-641, 1994.

metric connections along the circum-Paci c zone, Tech. Hauri, E.H., J.C. Lassiter, and H.G. Barsczus, Lithospheric

Rep. Kyoto Univ., Kyoto, Japan, 1983. controls on the comp ositions of Austral Islands lavas: Iso-

Norris, A., and R.H. Johnson, Submarine volcanic eruptions topic evidence from Raivavae and the southern Australs

recently lo cated in the Paci c by SOFAR hydrophones, abstract, Eos Trans. AGU, 7846, Fall Meet. Suppl.,

J. Geophys. Res., 74, 650-664, 1969. F828, 1997.

Ob elliann e, J.M., Contribution al' etude g eologique des ^ les Herron, E.M., Sea- o or spreading and the cenozoic history

des Etablissements Francais de l'Oc eanie. Sci. Terre, 3, of the east-central Paci c, Geol. Soc. Am. Bul l., 83, 1671-

1-146, 1955. 1692, 1972.

Okal, E.A., and A. Cazenave, A mo del for the plate tec- Ito, G., M. McNutt, and R.L. Gibson, Crustal structure of

tonic evolution of the east-central Paci c based on Seasat the Tuamotu Plateau, 15 S, implicati ons for its origin, J.

investigations, Earth Planet. Sci. Lett., 72, 99-116, 1985. Geophys. Res., 100, 8097-8114, 1995.

ORSTOM, Atlas de la Polyn esie Francaise, Paris, 1993. Johnson, R.H., and A. Mahalo , Relation of Macdonald vol-

Pautot, G., Analyse structurale de l'archip el des Tuamotu: cano to migration of volcanism along the Austral chain,

Origine volcano-tectonique, pap er presented at 3 eme Col- J. Geophys. Res., 76, 3282-3290, 1971.

lo que des Sciences de la Terre, Univ. Montp ellier, Mont- Krummenacher, D., and J. No etzlin, Datations de coul ees

p ellier, France, 1975. eruptives dans di erents archip els de la Polyn esie Fran-

Plou , D., Gravity and magnetic elds of p olygonal prisms caise, Bull. So c. Geol. Fr., 3, 131, 1966.

and application to magnetic terrain corrections, Geophys- Kruse, S.E., Magnetic lineations on the anks of the Mar-

ics, 41 4, 727-741, 1976. quesas swell: Implications for the age of the sea o or, Geo-

Rob ertson, E.I., Gravity e ects of volcanic islands, N. Z. J. phys. Res. Lett., 15, 573-576, 1988.

Geol. Geophys., 106, 1466-1483, 1967. L enat, J.F., and P. Labazuy, Submarine morphologies and

Rousset, D., A. Lesquer, A. Bonnevill e, and J.F. L enat, structures of Reunion, in LeVolcanisme de la R eunion-

Complete gravity study of Piton de la Fournaise volcano, Monographie, edited by J.F. L enat, pp. 43-74, Cent.

Reunion Island, J. Volcanol. Geotherm. Res., 36, 37-52, Rech. Volcanol., Clermont-Ferrand, France, 1990.

1989. L eotot, C., P.Y Gillot, F. Guichard, and R. Brousse, Le vol-

Ryan, M.P., Neutral buoyancy and the mechanical evolution can de Taravao Tahiti: Un exemple de volcanisme p oly-

of magmatic systems, in Magmatic Processes: Physico- phasique asso ci e a une structure d'e ondrement, Bul l.

chemical Principles, edited by B.O. Mysen, Geo ch. So c. Soc. Geol. Fr., 6, 951-961, 1990.

Sp ec. Pub. 1, 259-287, 1987. Leroy, I., Evolution des volcans en syst eme de p ointchaud:

Ryan, M.P., The mechanics and three-dimentional internal Ile de Tahiti, archip el de la So ci et ePolyn esie Francaise,

structure of active magmatic systems : Kilauea Volcano, Ph.D. thesis, 271 pp, Univ. Paris XI, Orsay,France, 1994.

CLOUARD ET AL.: GRAVITY SIGNATURE OF MAGMA RESERVOIRS 8191

Walker, G.P.L., Three Hawaiian calderas: An origin through Hawaii, J. Geophys. Res., 93, 4213-4248, 1988.

loading by shallowintrusions?, J. Geophys. Res., 93, Rymer, H. and G.C. Brown, Gravity elds and the inter-

14,773-14,784, 1988. pretation of volcanic structures: Geological discrimina-

Watts, A.B., An analysis of isostasy in the world's o ceans, tion and temp oral evolution, J. Volcano. Geotherm. Res.,

1, Hawaiian-Emp eror seamountchain, J. Geophys. Res., 27, 229-254, 1986.

83, 5989-6004, 1978. Sandwell, D.T., Antartic marine gravity eld from high-

Watts, A.B., and N.M. Rib e, On geoid heights and exure density satellite altimetry, Geophys. J. Int., 109, 437-448,

of the lithosphere at seamounts, J. Geophys. Res., 89, 1992.

11,152-11,170, 1984. Service de l'Urbanisme, Carte top ographique 1:20,000, Pa-

Watts, A.B., J.R. Co chran, and G. Selzer, Gravity anom- p eete, Tahiti, FrenchPolynesia, 1992.

alies and exure of the lithosphere: A three-dimensional Sichoix, L., Le volcanisme de Polyn esie Francaise : Ca-

study of the Great Meteor Seamount, northeast Atlantic, ract erisation des p oints chauds et du sup erb omb ement

J. Geophys. Res., 80, 1391-1399, 1975. a partir d'une nouvelle synth ese bathym etrique, Ph.D.

Watts, A.B., J.H. Bo dine, and N.M. Rib e, Observations of thesis, 297 pp, Univ. Fr. du Pac., Pap eete, Tahiti, 1997.

exure and the geological evolution of the Paci c Ocean Sichoix, L., and A. Bonnevill e, Prediction of bathymetry in

basin, Nature, 283, 532-537, 1980 FrenchPolynesia constrained by shipb oard data. Geo-

White, W.M., and R.A. Duncan, Geo chemistry and geophys. Res. Lett., 23, 2469-2472, 1996.

chronology of the So ciety Islands: New evidence for deep Smith, W.H.P., On the accuracy of digital bathymetric data,

mantle reclycling, in Earth Processes: Reading the Iso- J. Geophys. Res., 98, 9591-9603, 1993.

topic Code, Geophys. Monogr. Ser. vol 95, edited byA. Smith, W.H.P., and D.T. Sandwell, Marine gravity eld from

Basu and S. Hart, pp. 183-206, AGU, Washington, D.C., declassi ed Geosat and ERS-1 altimetry abstract, Eos

1996. Trans. AGU, 76 46, Fall Meet. Suppl., F156, 1995.

Young, R.Y., and I.A. Hill, An estimate of the e ective Smith, W.H.P., and P.Wessel, Gridding with continuous

elastic thickness of the Cap e Verde Rise, J. Geophys. Res., curvature splines in tension, Geophysics, 55, 293-305,

91, 4854-4866, 1986. 1990.

Talandier, J., and E.A. Okal, The volcanoseismi c swarms of

1981-1983 in the Tahiti-Mehetia area, FrenchPolynesia,

J. Geophys. Res., 89, 11,216-11,234, 1984a.

V. Clouard and A. Bonneville, Lab oratoire de G eo-

Talandier, J., and E.A. Okal, New surveys of Macdon-

sciences Marines et T el ed etection, Universit edelaPoly-

ald seamount, southcentral Paci c, following volcanoseis-

n esie Francaise, BP 6570 Faaa A erop ort, Tahiti, French

mic activity, 1977-1983, Geophys. Res. Lett., 1, 813-816,

Polynesia. e-mail:[email protected]; b [email protected]

1984b.

H.G. Barsczus, GBE - UMR 5569 - CNRS, and Uni-

Talandier, J., and E.A. Okal, Crustal structure in the So ciety

versit edeMontp ellier I I, 34095 Montp ellier Cedex 5.

and Tuamotu Islands, FrenchPolynesia, Geophys. J. R.

Astron. Soc., 88, 499-528, 1987.

e-mail:[email protected]

Turner, D.L., and R.D. Jarrard, K-Ar dating of the Co ok-

Austral island chain: A test of the hot sp ot hyp othesis,

J. Volcano. Geotherm. Res., 12, 187-220, 1982.

Walcott, R.I., Flexure of the lithosphere at Hawaii, Tectono-

Received January 29, 1998; revised Octob er 8, 1999; accepted

Novemb er 4, 1999.

physics, 9, 435-446, 1970.

This preprintwas prepared with AGU's L T X macros v5.01.

E File jgr99_pp formatted June 22, 2000.