Time-space mapping of Easter Chain volcanism

EPSL ELSEVIER Earth and Planetary Science Letters 136 (1995) 197-212

Time-space mapping of Easter Chain volcanism

John M. O’Connor ay * , Peter Stoffers a, Michael 0. McWilliams b a Geology Institute, Christian-Albrechts University, Olshausenstrasse 40, D-241 18 Kiel. Germany

b Department of Geophysics, Stanford University, Stanford, CA 94305.2215. USA

Received 22 September 1994; revised 1 June 1995: accepted 12 September 1995

Abstract

New 4oAr/ 39Ar and published K-Ar ages show that the locus of volcanism along the Easter Volcanic Chain (EVC) has

shifted systematically from the Nazca Ridge, at about 26 m.y., to the recently active Sala y Gomez Island/Easter Island region. This indicates a plume rather than a hotline (i.e., mantle roll) origin for the EVC. The time-space distribution of

ages, combined with published ages for the Galapagos and Juan Femandez volcanic chains, is used to reconstruct Nazca plate velocities over the past 26 m.y. A plume now located in the region of Sala y Gomez Island is most compatible with these data. West of the plume, the EVC records neither Nazca nor Pacific plate motions. This section of the EVC may be related to westward channeling of plume material to the Pacific-Nazca spreading boundary region.

1. Introduction

The origin of the Easter Volcanic Chain (EVC),

southeast Pacific (Fig. 11, is a widely debated topic (e.g., [l-7]). One model holds that the EVC was formed by a mantle plume now located beneath Easter Island [ll. According to this model, the Pa- cific and Nazca sections of the EVC should have

formed simultaneously as the Pacific and Nazca plates drifted apart at a spreading axis located above the plume. This model successfully explains the

early stages of formation of the Walvis Ridge-Rio Grande Rise (e.g., [1,12,13]), an apparently similar volcanic feature in the South Atlantic. The time-

space distribution of volcanism along the Easter Island-Nazca Ridge (Fig. 1) section of the EVC

* Corresponding author. Fax: 0049 431 880 4376. E-mail:

[email protected]

should, therefore, represent a record of Nazca plate motion. The corresponding ‘mirror image’ of the

EVC on the Pacific plate should then document Pacific plate motion (Fig. 1, Table 11, but K-Ar ages do not become systematically younger between Sala y Gomez Island (1.94-1.3 Ma) [2,8,9] and Easter

Island (3.0-0.3 Ma) [8,9], as is predicted by such a model. It has been proposed that a mantle hotline

(attributed to mantle rolls [14- 163) might be upwelling along the EVC [2], leading to concurrent volcanism at multiple sites along the chain. In addi-

tion to the EVC, mantle rolls have been called on to explain other hotlines on the Pacific plate [ 17,181.

A modification to the Easter Island plume model

suggests that a plume is upwelling in an intraplate setting, east of Easter Island [3-S], following an earlier and more general ‘pipeline’ channeling model [ 191. Three locations have been proposed for an

intraplate plume: - 725 km east of the Easter microplate [5], Sala y Gomez Island [4], and - 360 km

0012-821X/95/$09.50 0 1995 Elsevier Science B.V. All rights reserved

SSDI 0012-821X(95)00376-X

198 J.M. O’Connor et al. / Earfh and Pianetary Science Letters 136 (199.5) 197-212

Fig. 1. Shown am SO80 (0) sample sites and “Ar/ s9Ar age data (Fig. 2 and Table 2) and GS7202 (0) sample site and associated K-Ar

age data [2]. Age ranges for Easter Island [8,9] (Fig. 2 and Table 2) and Sala y Gomez Island [2,8] are shown along with “‘Ar/ “9Ar age

ranges for submarine volcanism west of Easter Island and for Crough Seamount (Fig. 2 and Table 2). Seafloor age anomalies are shown as

long narrow lines, along with associated Chron numbers [lo]. Solid shading represents features that are equal or shallower than 3000 m, as

derived from Digital Bathymetric Database 5. Outline of East Pacific Rise and Easter and Juan Femandez microplates are from [7,11]. The

shaded disk at Sala y Gomez Island indicates our minimum estimate of the size of the zone over which the upwelling EVC plume is

impacting against the base of the lithosphere, see text for further discussion. Sample locations and descriptions are in Table 1.

east of Sala y Gomez Island at N 102” [3]. Geo-

chemical evidence points to sub-lithospheric chan-

neling of material from a plume, located at Sala y Gomez Island, to the East Rift of the Easter microplate [6]. More recently, side-scan mapping has revealed three large areas of young, low relief, vol-

canism consisting of fresh lava flows and young

volcanoes falling along, and to the north of, a line

through Easter and Sala y Gomez islands [20]. The three main areas mapped are west-northwest and northeast of Easter Island and at 25”5O’S, 104”2O’W, to the northeast of Sala y Gomez Island [20,21].

Table 1

Sample locations and descriptions

Sumple Deprh (m) Dercriptwn

Now Ridge

SO80 14DS-4 Salas y Gonrez Ridge

SO 8Oa 12DS-I

SOSOa I7DS-IA SO8Oa I8 DS-IA

23*16.92’S; 82’59.76’W 2324

25’38.54’s; 82’22.45’W 1253

24*55.90’s; 88’20.14’W I507

25’42.22’s; 93’14.5SW 1758

altered basalt, large plag phenocrysts

Fmegr. basalt, rare plag microphenos plag groundmass, olivine (5-10%)

vesic. alt.lava. large 01 and relative fresh plag phenos

Crough Seamounr SO 8Oa 67DS I 24-45.85’S; 121’40.4l’W 738 slightly alt, lava with large phenos of plag

SO 8Oa 7ADS-2 24’51.27’s: 21’53.63’W 1262 plate-like lava piece with large phenos of plag

Enrrer Island fRoiho IavafieId) EI9209 hqhly vesicular (30%) fine. grained lava. plag microphenos. some phcnos of 01

Moai Seomounr SO 8Oa 29DS-4 27’06.93’S; 109’39.62’W 1068 phorphontic lava, 25% fresh plag phenos. 10% unaltered 01 phenos.

Pukoa seamounr SO 80a 35DS-IA Umu Seamounr SO 8Oa 93DS.8

26’55.9O’S; 110’18.26’W

26’55.2OS. I10’58.27’W

907 very fresh lava, (30 to 40%) plag phenos. (5%) 01

1594 Vesic. lava with few very large plag phenoaysts

J.M. O’Connor et al./ Earth and Planetary Science Letters 136 (1995) 197-212 199

To establish which, if any, of the above models is

more applicable to the EVC, the R/V Sonne dredge sampled the Nazca Ridge-Crough Seamount section

of the Easter Volcanic Chain in 1992 122-241. Re- ported here are the results of a high precision “OAr/ 39Ar study (Fig. 2, Table 2) of plagioclase

separates from selected samples (Fig. 1, Table 1).

2. Analytical technique and results

Samples most suitable for “OAr/ 39Ar dating were

selected after careful inspection of thin sections.

Plagioclase separates were cleaned in 6% HF for

6-10 min, followed by ultrasonic washing. Between

13 and 75 mg of plagioclase was packed in 99.9%

pure aluminum foil packets, each of which was secured in one of a series of custom-drilled holes in

a 99.95% pure aluminum disk. Sample disks, inter-

spersed with a 3-dimensional array of 27.92 Ma TCR (856003) sanidine monitor [25], were secured to-

gether and then sealed in an aluminum can and irradiated for 1.5 h at a reactor power of 1 MW in

the Cd-shielded CLICIT facility [26] of the TRIGA

reactor at Oregon State University, USA. TCR standard was measured between 3 and 5

times for each monitor position in the stack. The

reproducibility on TCR was about 0.3%. J values range from 0.0003641 to 0.0003985. The appropriate J value and associated error was extrapolated for each sample position using a 3-dimensional least- squares cosine plane fit (C. Hall, unpublished, and Bogaard and Schimick in press). The uncertainty on each individual extrapolated J value is 0.12%. The maximum horizontal variation across the disk stack was < 2% and the maximum vertical variation was

9.5%. Aluminum packets containing plagioclase samples

were incrementally heated at Stanford University, in a UHV resistance furnace and gas released analyzed in a MAP216 rare gas mass spectrometer. The technique used is similar to that outlined in [27], with the

exception that electron multiplier current was mea-

sured directly using a Keithley 617 electrometer. The sensitivity was - 2E - 14 mol/nA of multiplier current. A mass discrimination of 292.6 (used in reducing all data) was estimated by repeated mea- surement of air samples. Blanks were measured at

regular intervals, normally once per day before start-

ing analyses at 8OO”C, lOOO”C, 1200°C and 1400°C. An estimate of the background is 4E - 16 mol 40Ar,

2E - 17 mol 39Ar and 1E - 17 mol 36Ar, at tempera- tures up to 1200°C. 40Ar and 36Ar increase exponen- tially in an atmospheric ratio to about 2E - 15 mol

40Ar by 1400°C. Isochrons have been calculated as

inverse isochrons using York’s least-squares fit,

which accommodates errors in both ratios and corre-

lation of errors [28]. A summary of the results is

shown in Fig. 2 and Table 2. Full data tables are

given in the Appendix.

3. Time-space distribution of EVC volcanism

@Ar/ 39Ar and published K-Ar ages become sys-

tematically younger between the N 26 Ma South- west Nazca Ridge and the Sala y Gomez Island/Easter Island region (Fig. l), indicating a

plume origin for the EVC. The time-space distribu-

tion of volcanism along the EVC should, therefore,

record the speed of the Nazca plate for the past N 26 Ma (Fig. 3). However, in order to calculate velocities

and total reconstruction poles for the complete 0- N

26 m.y. time period, the EVC plume needs to be located; that is, where the vertically rising mantle

plume is impacting against the base of the oceanic lithosphere. As discussed earlier, three locations have been proposed for the EVC plume: Easter Island [ 11, Sala y Gomez Island [4,6] and some 360 km east of Sala y Gomez Island, at N 102”W [3]. We have evaluated our data using these three proposed loca-

tions of the EVC plume-hotspot (Fig. 3). A 30 Ma K-Ar age [2] for the eastern end of the Sala y Gomez Ridge is significantly older than the 22.0 +

0.5 @Ar/ 39Ar age reported here from the same

region. We have not included this age in our study because a submarine whole-rock sample > 20 Ma will probably be too altered to be reliably dated using the K-Ar technique. However, as seawater alteration usually results in measured ages being too young (due to a loss of radiogenic argon from alter-

ation products and the addition of potassium via seawater weathering), this may not fully explain an older K-Ar age. Nonetheless, we use the much more reliable 40Ar/ 39Ar age measured on plagioclase. A younger whole-rock K-Ar age of _ 8 Ma has,

however, been used in this study, both because it is

200 J.M. O’Connor et al./Earth and Planetary Science Letters 136 (1995) 197-212

a 100 -

g 80-- SOS0 14DS-4 (72.6 mg) A A

Nazca Ridge 1400~

A 1200” 5 60-- 25.8 f 0.6 Ma

1150”

z % 4o-- 11000

& 950” 1000” 10.500

@z 20--

o- 0.0 0.2 0.4 0.6 0.8 1.0

60 - 248 -- SO80 12DS-l(l3.3 mg)

E Sala y Gomez Ridge

936 -- 22.0 f 0.5 Ma B 800” 11000 1200~

1300” _ -

$24-1 gooo 9500 JOOOO 10500

$12 --

Oj I I I I- I I I

0.0 0.2 0.4 0.6 0.8 1 .O

40

s32--

SO80 17 DS-1A (74.0 mg)

E Sala y Gomez Ridge

14.9 f. 0.2 Ma

ot I I I I I I I I I I 0.0 0.2 0.4 0.6 0.8 1.0

$40 _t

SOS0 lgDS-1A (74.6 mg) SaIa Y Gomez Ridge I

Ol I I I I I I I I I

0.0 0.2 0.4 0.6 0.8 1.0

1.0 , I

go.8 --

&I 0 0.6 --

E19209 (63.2 mg) Easter Island

131f19ka

0.004

0.003

0.002

0.001

SO80 14DS-4 (72.6 mg) Nazca Ridge

MSWD =O.Ol

0.000 J 1 0.000 0.006 0.012 0.018 0.024 0.031

“.““-r

304f22

0.003.

0.002-

SOS0 l2DS-l(l3.3 mg) Sala y Gomez Ridge

0.001. 21.6 f 1.1 Me MSWD =0.07

o.oool TrII\ 0.000 0.008 0.016 0.024 0.040 O.(

“.V_

29.5 f 0.3

0.003.

0.002*

SO80 17 DS-IA (74.0 mg) Sala J Gomez Ridge

0.001 14.9 f 0.3 Ma

* MSWD =0.14

0

“.uv-t

29552

o,ool . 11.75~ 0.3 Ma MSWD =0.59

SO80 18DS-1A (74.6 mg) Sala y Gomez Ridge

34

2

0.000 0.013

0.004 ‘I 299 f 15

0.003 .

0.002 - 0.001. 139 f 20 ka

MSWD =0.03

0.000. 0.000 1.100

0.026 0.039 0.052 0.065

1

ET9209 (63.2 me) Easter Island

59

$

2.200 3.300 4.400 5.500

39ArP!34r

Fig. 2. 40Ar/‘9Ar incremental heating and inverse isochron data for SO80 samples. Heating steps included in plateau and isochron

calculations are indicated by shading. Weighted mean plateau ages are shown for all plateaus. The f 1 c error bars in the plateau plots are

without etror in J. All samples analyzed were plagioclase separates. A summary of age data is given in Table 2.

J.M. O’Connor et al./Earrh and Planetary Science Letters 136 (1995) 197-212 20

b

3

g

$2 SO80 29DS-4 (72.2 mg)

;i Moai Seamount

g! 230f7Ska

b l--

9 800”

0-i 9000

1VW I , I I I

0.0 0.2 0.4 0.6 0.8 1.0

51 I %4 E. t

SO80 35 DS-IA (71.6 mg) Pukao Seamount I

9 3 -- 630 It 180 ka 1400”

0.0 0.2 0.4 0.6 0.8 1.0

30

924

SO80 93DS-8 (70.7 mg) Umu Seamount I

c

&I 2.4 It 0.54 Ma

ol 18 $ 0 12 a I I I

I I I I

0.0 0.2 0.4 0.6 0.8 1.0

SO80 71DS-2 (75.5 mg) Crough Seamount

7.6f 0.2Ma

0t I I I I

I I 0.0 0.2 0.4 0.6 0.8 1.0

60 - I

g48 -- SO80 67DS-l(72.2 mg)

Crougb Seamount (west slope) 12% 14fjoo

236 -” 8.4 f 0.1 Ma

z % 24 -- &

1150”

( 12 -300’ 875’ 925” 975” 1025’ 1075”

o! I I I I I I I I

0.0 0.2 0.4 0.6 0.8 1.0

Cumulative3gAr

0.000 1 I 0.000 0.640 1.280 1.920 2.560 3.200

0.004

0.003

0.002

0.001

0.000 0.1 DC

SO80 35 DS-lA (71.6 mg) Pukao Seamount

\ lo 0.280 0.560 0.840 1.120 1.3

0.004 - 282f 24 SO80 93DS-8 (70.7 mg)

Umu Seamount

c$

2

0.001 - 3.0 f 1.1 Ma MSW=0.19

0.000 0.071 0.141 0.212 0.282 0.353

0.004 - 293 f 15 SO80 71DS-2 (75.5 mg)

0.003 - Crougb Seamount

0.002 a4 -

7.7 i 0.5 Ma & o.ool. MSWD =O.Ol

0.000 J \I 0.000 0.019 0.038 0.057 0.076 0.095

0.000 0.000 0.017 0.034 0.05 1 0.068 0.085

3gAr/41r Fig. 2 (continued).

202 J.M. O’Connor et al./Earth and Planetary Science Letters 136 (199.5) 197-212

significantly less altered and because the analysis has been duplicated.

The time-space (i.e., distance from Easter Island) distribution of EVC volcanism is linear to a first approximation (Fig. 3a). However, the offsets of the Nazca Ridge and 22 Ma Sala y Gomez Ridge sample sites from this overall trend (Fig. 3a) suggest a much slower velocity for the Nazca plate during earlier times. For the purposes of our study we assume that this time of slower plate speed ended at - 15 Ma (Fig. 3a), as there are insufficient data to document this slowdown in more detail. This time corresponds with a major change in plate motion in the South Pacific that occurred between C5 and C7 [lo]. As a result, the Farallon plate split to form the Cocos and Nazca plates and a change in spreading direction occurred, from a northeasterly to a northwesterly direction [lo]. We consider it reasonable to assume that, during this time of plate reorganization and the associated transition from Nazca Ridge to Sala y Gomez Ridge volcanism (Fig. l), a broad zone of volcanism existed at the eastern end of the Sala y Gomez Ridge, at the intersection of the Nazca Ridge and the Sala y Gomez Ridge (Fig. 1). As the 22 Ma Sala y Gomez Ridge sample site would have been located on the eastern edge of such a broad transition

Table 2 Summary of age calculations from argon isotopic data

zone (Fig. 11, this could explain why it is offset from the overall time-space distribution of 2 15 Ma EVC volcanism (Fig. 3). We therefore evaluate also the possibility that the N 26 Ma Nazca Ridge age is a more reliable constraint on plate motion than the 22 Ma Sala y Gomez Ridge sample site (Fig. 3a).

A major problem with an Easter Island plume is that it is not possible to fit a convincing regression line to EVC ages 4 22 Ma or I 14.9 Ma (Fig. 3) when the Easter Island ages are included. In contrast, an excellent fit is achieved for EVC ages I 14.9 Ma when Easter Island age data are excluded (Fig. 3). This strongly suggests that the EVC plume is upwelling in the region of Sala y Gomez Island or somewhere between it and Easter Island. Another problem with an Easter Island plume is that it leads to a predicted linear velocity of - 13 cm/yr (Fig. 3a and Table 4) for recent Nazca plate motion (i.e., - 2 1 Ma), much faster than the rate of 4.7 cm/yr (at the latitude of the EVC) predicted by, for exam- ple, the NUVEL-1A global plate model for the past 3.2 Ma [29].

Fig. 3b shows the time-space distribution of EVC volcanism based on the assumption that the EVC plume is upwelling under Sala y Gomez Island. Two regression lines are shown, one assuming that a

&&au Calculatiort Isochron Calculation

Sample Age (Ma) i IO Steps Used (“0 Total % jpAr Age (Ma) + la MSWD Trapped JOAr/‘6Ar

Nazca Ridge

SO 80 14DS-4 25.8 + 0.6 950, Iooo, 1050 63 25.6 ‘- 1.6 0.01 296i 16

Sala y Gomez Ridge

so 80 12 lx-1 22 0 * 0.5 800, 900. 950. 1000 81 21.6k 1.1 0.07 304+ 22

1050. 1100. 1200

SO I’IDS-IA 14.9 * 0.2 1000. 1025. IOSO. 1075 95 14.9 + 0.3 0.14 295.4 f 0 3

1100. 117.5

SO80 IIDS-IA 11.5* 0.2 800. 850. 900. 950. 1000 78 II.75 0.3 0 59 295 + 2

1050, 1100

Crough Seamount

SO80 7lDS-2 7.6 f 0.2 700, 800. 875, 925, 70 7.7 f 0.5 0.01 293 f 15

1075

SO80 67DS-I 8.4 k 0.1 800. 875, 925, 975, 65 8.3 f 0.2 0.08 298f 3

1025. I075

Easter Island Region

El9209 0.13 f 0.02 800. 925. 1250. 1450 100 0.13 f 0.02 0.03 299? 15

SO80 29DS.4 0.23 + 0.08 SM), 900. IOLW 70 0.22 It 0.22 0.01 296r IO

SO80 35DS-1A 0.63 f 0.18 lOGI. 1125. 1200 52 0.25 * 098 0.04 317 f 91

SO80 93 DS-8 2.4 5 0.5 900. 975. 1040. 1100 48 3.0 + 1.1 0.2 282 f 24

‘Steps used’ refers to heating steps included in plateau and isochron age calculations. %j9Ar indicates the portion of the total “9Ar released

used in calculating plateau and isochron ages. Trapped indicates the initial @Ar/ ‘6Ar in samples (atmospheric).

J.M. O’Connor et al./Earth and Planetary Science Letters 136 (199.5) 197-212 203

change in Nazca plate velocity occurred at _ 15 Ma (i.e., incorporating the Nazca Ridge age) and the other assuming constant plate velocity for the past 22 Ma. In contrast to a model assuming an Easter Island plume (Fig. 3a), this model is more compatible with a recent slowdown in Nazca plate velocity (Fig. 3b). The apparent non-zero K-Ar ages for Sala y Gomez Island might be explained by the significant slowdown in Nazca plate speed, resulting in an increase in the life-span of volcanism at individual locations along the EVC. This is supported by the fact that very young fresh lava flows were observed during the first expedition to the island 1301. An alternative explanation is that the hotspot might be located further to the east.

3.1. Submarine volcanism west of Easter Island

40Ar/ 39Ar ages are reported here (Fig. 2 and Table 2) for a chain of three seamounts extending * 200 km to the west of Easter Island (i.e., Moai,

Pukao, and Umu seamounts [21,22], Fig. 1). These ages increase from east (Moai) to west (Umu> (Fig. 3b), opposite to that expected if the Nazca plate were moving to the east over a plume upwelling near Easter Island. The relationship between Easter Island to Umu Seamount volcanism and the neighboring spreading boundary over the past N 3 Ma is discussed further in a later section.

3.2. Other hotspot traces on the Nazca plate

The Galapagos (e.g., [31-331) and Juan Feman- dez chains [34] are two previously studied hotspot lineaments on the Nazca plate. The time-space distribution of published 40Ar/ 39Ar ages for the seamount section of the Galapagos Chain [32] is shown in Fig. 4a. The distribution of K-Ar ages [33] for the younger ocean island section of the Galapa- gos Chain is much more scattered, yet clearly suggests a slowdown in Nazca plate velocity (e.g., [32]) between - 5 and 4 Ma. This slowdown in Nazca

30 . Sala y Gomez lrland

l Easter Island

25 0 K-Ar age 121 _ . . . ..I I Fig. 2. Table 2

q Fig. 2. Table 2 20 -

3

H

0 ‘co 800 1200 1600 zoo0 2400 2800 3200

Distance from Easter Island (km)

Easter Island Scamounu (hg 2. Table 2)

.&I -300 200 700 1200 1700 2& 2700

Distance from Sala y Gomez Island (km)

Fig. 3. (a) Time-space (distance from Raster Island) distribution of ages of volcanism along the EVC. Superscript numbers indicate age data

included in fitted regression lines used to determine Nazca plate speed as follows: 5 14.9 Ma; (2) =

(I) = Sala y Gomez Island and Sala y Gomez Ridge Sala y Gomez Island and Sala z Gomez Ridge I 20 Ma; (‘) - Easter Island, Sala y Gomez Island, Sala y Gomez Ridge -

$ 20 Ma; (4) = Sala y Gomez Ridge (1 14.9 Ma) and Nazca Ridge (25.8 Ma). Also shown is the 0 to I 3.2 Ma linear velocity of the

Nazca plate (at the latitude of the EVC) as predicted by the NUVEL-IA global plate motion model [29]. The regression line connecting the

Sala y Gomez Ridge (2 14.9 Ma) and Nazca Ridge (25.8 Ma) predicts that Nazca plate speed was much slower (- 5.2 cm/yr) than during

the subsequent phase of EVC formation ( - 13 cm/yr). The short horizontal arrow connecting the 22 Ma EVC age with this regression line indicates the offset of this sample site from middle of the very broad volcanic swath that probably formed during the complex transition

from Nazca Ridge to Sala y Gomez Ridge formation (Fig. 1). &Ar/ j9Ar ages (Fig. 2 and Table 2) are shown as solid or crossed boxes or (errors are smaller than size of symbols, with the exception of 14DS and IZDS samples for which error bars are shown), and earlier K-Ar

ages as open boxes. Sources of K-Ar age data are as follows: Sala y Gomez Island [2,8]; Easter Island this study, [8,9]. (b) Time-space

(distance from Sala y Gomez Island) distribution of volcanism along the EVC.

204 J.M. O’Connor et al. / Earth and Planetary Science Letters 136 11995) 197-212

plate velocity is reflected in global models for plate motions in the hotspot reference (i.e., NUVEL-1A [29] and AMl-2 1351, Fig. 4b). The Juan Femandez Chain also records this slowdown in Nazca plate velocity, as is shown on combining the time-space distribution of volcanism along the Juan Femandez Chain [34] with that of the Galapagos Seamount Chain (Fig. 4b). However, in order to make such a direct comparison, an adjustment has to be made for the - 33“ distance between the Galapagos and Juan Femandez chains. This is because the linear velocity of a plate (i.e., the length of hotspot trails) increases with greater distance (to a maximum at 90”) from the rotation poles reconstructing its motion. We have used the published poles for recent Nazca motion, NUVEL-1A [29], AMl-2 [35] and 3Pl [ll], to make this adjustment (Fig. 4b).

3.3. EVC, Galapagos and Juan Fernandez hotspot chains

The Juan Femandez (0 to - 4 Ma) and Galapa- gos Seamount ( - 5 to 9 Ma) chains have been used

* Galapagos Islands K-AI ages (331

0 loo 200 30 400 SO0 6fx 700 8M)

Distance from active hotspot (km)

to evaluate the time-space distribution of EVC volcanism (Fig. 5). As discussed above, a suitable rotation pole for Nazca plate motion must be used to adjust the lengths of hotspot trails to a common latitude so that they can be compared directly. The older section of Sala y Gomez Ridge trends at - 90”, in contrast to the - 70” trend of the more recent Sala y Gomez Island-Sala y Gomez Ridge section of the EVC (Fig. 1). A pole located between AMl-2 and 3P1 (i.e., - 59%; lOloW) best reconstructs Nazca motion in accord with this older section of the EVC. In contrast, the 3Pl [ll] rotation pole best agrees with the trend of the younger Sala y Gomez Island-Sala y Gomez Ridge section of the EVC and that of the Juan Femandez Chain. This apparent change in pole location occurred sometime after 8 Ma (Fig. 1) and may well have coincided with the slowdown of the Nazca plate at - 4.5 to 5.5 Ma (Table 4). For the purposes of our study we have used the pole at 59%; 101”W to adjust the lengths of the Galapagos Seamount Chain. However, using the 3Pl pole to make this adjustment produces a basically similar result (i.e., 12.7 cm/yr, as op-

q Juan Femandez Chain K-Ar ages 1341

(b)

I I I I I I 1

0 loo 200 300 4al 500 600 7cu 800

Distance from active hotspot (km)

Fig. 4. (al Time-space distribution of published 4oAr/ s9Ar ages for Galapagos seamounts [28] and K-Ar ages for Galapagos Islands [33].

The location of the active Galapagos hotspot is assumed to be at the Island of Femandina (0’2l’S; 91’30’W). Our preferred linear regression

tit to Galapagos Seamount data (solid line) excludes some data with large errors (shown as vertical lines). The more recent linear velocity of

the Naxca plate at the latitude of the Galapagos Chain has been calculated using the formula V= o sin 4 (w= angular velocity,

4 = angular distance from rotation pole, and V = linear plate velocity) and rotation poles for Naxca plate motion over plumes; i.e.,

NUVEL-IA [29], AMl-2 1351 and 3Pl 1111. (b) The recent velocity of the Nazca plate has also been calculated on the basis of the

time-space distribution of volcanism along the Juan Femandez Chain, as defined by published K-Ar age data [34] and the assumption that

the hotspot is currently active at Domingo seamount (33”57.5’$ 8l”SO’W) [22]. Three published rotation poles (NUVEL-IA (291, AMl-2 [35] and 3Pl (111) have been used to adjust the length of the Juan Femandez Chain so that it is equivalent to that which would have formed

at the latitude of the Galapagos Chain.

J.M. O’Connor et al./Earth and Planetary Science Letters 136 (1995) 197-212 205

posed to 12.9 cm/yr predicted by the 59”s; 101”W pole).

Combining EVC, Galapagos and Juan Femandez data (Fig. 5a) shows that the length of the EVC, calculated on the basis of an Easter Island plume, is not similar to the Galapagos and/or Juan Femandez chains. In contrast, assuming that there is a Sala y Gomez Island plume results in much better agreement (Fig. 5b). Recent GLORIA-B imaging to the northeast of Easter Island has shown the existence of

0 500 looO 15w 2M)o 2500 3lmJ

Distance from Easter Island /hotspots (km)

-8W -300 2w 700 1200 17W 22W 2700

Distance from Sala y Gomez Islatxlhotspots

0 500 Ical 1500 2ow 2500

Distance from 104” 20’ W

a great number of fresh lava flows and recent volcanoes on the seafloor at 25”5O’S; 104”2O’W l-201. This indicates that a plume might also be located N 110 km to the east of the island of Sala y Gomez (following the suggestion that there is a plume N 3.3” east of Sala y Gomez Island [3]). Based on this assumption, excellent agreement can be achieved between the EVC, Galapagos and Juan Femandez hotspot trails (Fig. 5~). This result is compatible with an upwelling EVC plume forming a broad region of hotspot volcanism, encompassing at least Sala y Gomez Island and young recent volcanism N 110 km to the east (Fig. 1).

3.4. Possible underestimation of EVC ages

Samples from seamounts and ridges discussed in this study were recovered by dredging. Therefore, the possibility exists that these samples represent younger surface volcanism, whereas the oldest phases of volcanism at a particular sample site could be I 3 Ma older. We have considered the possibility that such an underestimation might account for the diffi- culty in explaining formation of the EVC by an Easter Island plume (Fig. 6). Although the time- space distribution of adjusted EVC ages (i.e., +3 Ma, Fig. 6) does, indeed, suggest that the EVC plume could be located under Easter Island, there

Fig. 5. (a) Comparison of the time-space distribution of volcan-

ism along the EVC (assuming an Easter Island plume) (Fig. 3a),

and along the Galapagos and Juan Femandez chains (Fig. 4). The

time-space distribution of Galapagos age data (5) have been

changed ‘@ such that it is equivalent to that which would have

been formed had the Galapagos formed at the same latitude as the

EVC. This adjustment has been made using a rotation pole located

midway between 3Pl [Ill and AMl-2 [35] (i.e., 59%; lOloW);

see text for further discussion. The Juan Femandez Chain has

been similarly adjusted using the NUVEL-IA [29] rotation pole

for recent Nazca plate motion over plumes (using 3Pl or AMl-2

would produce a similar result). Also shown is the predicted

velocity of the Nazca plate at the latitude of the EVC predicted by

the NUVEL-IA model [29]. (b) Similar to (a), except that the

EVC plume is assumed to be under Sala y Gomez Island. (cl

Similar to (a) with the exception that the EVC plume is assumed

to be located at 25”5O’S; 104’2O’W. the location of recently

mapped fresh lava flows and recent volcanoes [20]. The long solid

regression line shown (‘) has been fitted to Sala y Gomez Ridge data >8Maand 114.9Ma.

206 J.&f. O’Connor et al. / Earth and Pianerary Science Letters 136 (19951197-212

are, nonetheless, a number of serious problems with this model. We must make the special, and perhaps unreasonable, assumption that the sampling of the Galapagos Seamount Chain has not been similarly biased towards young volcanism. In addition, the recent slowdown in Nazca plate velocity (e.g., [29]) is not compatible with, this model and a m 3 Ma upper age for Easter Island (Fig. 6). We therefore consider that a plume in the Sala y Gomez Island region better explains formation of the Sala y Gomez Ridge-Nazca Ridge section of the EVC.

3.5. Nazca plate motion for the past - 26 Ma

The time-space distribution of EVC volcanism and linear velocities of the Nazca plate are in Tables 3 and 4. The speed of the Nazca plate between N 15 Ma and when the Nazca plate slowed down is about 13 cm/yr, irrespective of where the Easter plume is currently upwelling (Table 4). However, estimating when this slowdown in Nazca plate motion began depends very much on where the EVC plume is located and on the assumed speed of the Nazca plate following this latest velocity change (Table 4). Our preferred model (Fig. 5c) points to this slowdown having started at about 4.5 Ma, assuming a NUVEL-

30 @ EVC ages plus 3 Ma R

25

5

0

0 500 iOo0 1500 2ca 2500 3000

Distance from Easter Island /hotspots

Fig. 6. The time-space distibution of EVC volcanism, assuming

an Easter Island plume (Fig. Sa), along with EVC ages that have

been systematically increased by 3 Ma (as indicated by vertical dashed arrows). The regression line shown above (*’ has been

fitted to Sala y Gomez Ridge I 14.9 Ma (+3 Ma) and to Sala y

Gomez Island ages ( + 3 Ma).

Table 3

Time-space distribution of EVC volcanism

Time span Distance (km) f~;t~5~landplume (Fig. 5a)

2690 0 to 14.9 2120 oto 11.5 1620 0 to 8 1200 OtO-l(NtJVFLtA) -50

ptt;50mez Island plume (Fig. 56) 2280

0 to 14.9 1720 oto 11.5 1220 0to8 800 0 t0 2.5 (NUVEL-IA) 129 0 10 3.3 (Juan Femandez chain) 211

104” 2O’W plume (Fig. SC) 0 to 25.8 2170 0 to 14.9 1610 oto11.5 1110 0 to 8 690 0 t0 4.5 (NW&L-IA) 233 0 to 5.5 (Juan Fcmandez Chain) 3.57

Easter l&mdplume & adjusted EVCages (Fig. 6) 0 to 28.8 Ma 2690 Oto 17.9Ma 2120 0 to 14.5 Ma 1620 Oto IlMa 1200

1A estimate of Nazca speed (Table 4). An earlier (2 15 Ma to < 26 Ma) slowdown in Nazca speed is also apparent (Fig. 3). Our estimates of Nazca pIate rotation angles subtended since u 26 Ma (i.e., the distance between dated EVC samples sites and zero age volcanism above the plume) also vary according to where the plume is located (Table 3).

The rotation pole that best reconstructs Nazca motion in accord with the w 90” trending section of the EVC (i.e., 59%; lOloW) is located - 86” from the EVC. Therefore, estimates of the linear velocity of the Nazca plate based on the time-space distribution of volcanism along this section of the EVC must also approximate the angular velocities of the Nazca plate (Tables 3 and 4).

4. Plume hotspot versus plume channeling volcanism along the EVC

The tectonic setting of EVC plume volcanism over the past N 26 Ma is shown in Fig. 7. “oAr/ 39Ar dating (Fig. 2, Table 2) indicates that Crough Seamount [22], part of the EVC west of the Easter

J.M. O’Connor et al. / Earth and Planetary Science Letters 136 (I 995) 197-212 207

microplate (Fig. 11, is between 7.6 f 0.2 Ma and 8.4

Ma + 0.1 Ma (Fig. 2, Table 2). Comparing Pacific seafloor and Crough Seamount ages clearly indicates

that this section of the EVC formed in a tectonic setting near the spreading boundary (Fig. 7). In contrast, the section of the Sala y Gomez Ridge

contemporaneous with Crough Seamount, formed in

an intraplate tectonic setting (Fig. 7). Clearly, this is

unlikely to have happened had the EVC plume been

upwelling close to the spreading axis between the Pacific and Nazca plates (i.e., an Easter Island

plume). This suggests that the model of a plume

upwelling under Easter Island is not appropriate. We

interpret this as supporting evidence for a plume in

the Sala y Gomez Island region that has been chan-

neling plume material to the Nazca-Pacific spreading boundary [5,6] for at least the past 7.5-8.3 Ma.

Limited K-Ar age data show that Sala y Gomez

Island formed at N < 2 Ma [2,8] in an intraplate

setting (Fig. 7). In contrast, initial construction of Easter Island would appear to have occurred close to a spreading axis (Fig. 7). Volcanism has continued to erupt on Easter Island until at least 0.13 + 0.02 Ma

(Fig. 2, Table 21, in a progressively more intraplate setting (Fig. 7). Near to Easter Island, Umu Seamount

began to form at N 2.4 + 0.5 Ma, most likely very

near to the East Rift of the microplate (Fig. 7). In contrast, Moai and Pukao Seamounts erupted in a

Table 4

Linear velocity of the Nazca plate at the latitude of the EVC

Time span linear velocity (cm/yr) Easter Island plume (Fig. 50) oto-1 4.1 -1 to 14.9 13.3 14.9 to 25.8 5.2

Ma y Gomez Island plume (Fig. 5b) 0 to 2.5 (NLR?ZL-IA) 4.1

i:o 3.3 (Juan Femandez Chain) 5.8

2.5 or 3.3 to 14.9 12.7 14.9 to 25.8 5.1

104” 20’W piume (Fig. SC) 0 to 4.5 QWVELIA) 4.1 0 to 5.5 (Juan Femandez Chain) 5.8 4.5 or 5.5 to 14.9 13.3 14.9 to 25.8 5.1

Easter Island plume & adjusted NC ages (Fig. 6) 0 to 17.9 12.7 17.9 to 28.8 5.2

-3000 -2000 -1000 0 1000 2000 3000

Distance from Sala y Gomez Island (km)

Fig. 7. Comparison of the time-space evolution of EVC volcan-

ism (Fig. 5b) with that of surrounding Nazca and Pacific seafloor.

Seafloor ages have been derived from both average rates of Nazca

and Pacific relative plate motions [IO] and anomaly age picks

located to the north and south of the EVC [7] [lo]. Age range for

Crough Seamount are from Fig. 2 and Table 2.

more intraplate setting between u 0.23 + 0.08 Ma and - 0.6 + 0.2 Ma, respectively. So, although we cannot establish exactly how the EVC between Sala

y Gomez Island and the Easter microplate formed, the interaction between channeled EVC plume material and spreading axes/microplates has been an important factor.

5. Discussion

The location of the upwelling EVC plume most in agreement with all available age data is in the general region of Sala y Gomez Island. This is compatible with an upwelling EVC plume forming a broad

region of hotspot volcanism encompassing at least

Sala y Gomez Island and young recent volcanism about 110 km to the east. Combining information

about the time-space distribution of volcanism along the EVC, Juan Femandez and Galapagos chains

allows us to reconstruct Nazca plate motion relative to an assumed fixed hotspot for the past N 26 Ma (Fig. 5c and Table 4). This result provides new supporting evidence for the permanence of mantle plumes, at least under the Nazca plate. The possibility that ages reported here significantly underesti- mate the duration of EVC volcanism seems remote.

208 J.M. O’Connor et al./ Earth and Planetary Science Letters 136 (1995) 197-212

Irrespective of where we assume the plume to be upwelling along the EVC, we calculate a speed of _ 13 cm/yr between N 15 Ma and until the start of the most recent slowdown of the Nazca plate [29]. This robust determination of linear plate velocity can also be taken as an estimate of angular plate velocity during the formation of the section of the EVC orientated at - 90”. It is unclear as to when the trend of the EVC changed from N 90” to N 70”. How- ever, it must have occurred at < 8 Ma, and probably in association with the most recent slowdown of the Nazca plate. Determining when this slowdown began is dependent on where the EVC plume is upwelling and the estimated speed of the Nazca plate following slowdown. Our preferred model (Fig. 5c and Table 4) suggests that the latest slowdown began at either 5.5 or 4.5 Ma, depending on whether we use Nazca speeds predicted by the NUVEL-1A model 1291 or the time-space distribution of Juan Femandez Chain volcanism. As the NUVEL-1A is a globally tested model, we prefer to take N 4.5 Ma as our best working estimate.

West of the proposed plume, EVC formation (on both the Nazca and Pacific plates) is unrelated to plate motion over a plume and is best explained by plume channeling. The actual mechanisms by which this section of the EVC formed are not yet clear, but interaction between plume material and spreading axis/microplate boundaries apparently played a significant role.

A variation on the plume channeling idea is that the impacted EVC mantle plume might have spread out into a broad region of up to 1000 km in diameter [36,13]. In addition, the suggestion that a second hotspot is located in the Crough Seamount region (i.e., hotspot 2 [4]> is not ruled out completely by

this study. Although our results are most compatible with the model of an upwelling plume in the region of Sala y Gomez Island, we have, as yet, no direct evidence to support a vertically rising plume. In the absence of such direct evidence, the possibility still exists that a very broad region of anomalously hot mantle, unconnected to the lower mantle via a plume conduit, exists under the EVC region of the Pacific (e.g., [37,38]). This would require that this region be fixed in the mantle over long periods of time and that formation of the section of the EVC that records Nazca plate velocity is, in some, as yet undeter- mined, fashion, controlled by weaknesses in the lithosphere.

Acknowledgements

We gratefully thank the Bundesministerium fur Forschung und Technologie (BMFT) for funding the SO8Oa cruise and post cruise work. Captain Ku11 and crew of the R/V Tonne are thanked for their help during the SOBOa cruise. B. Hacker wrote the soft- ware used for the reduction of the age data and gave helpful advice to JMO during the analytical phase of this project. P. van den Bogaard and V. Schenk provided preparation facilities at GEOMAR and the Mineralogy Institute, University of Kiel, respectively. M. Pringle kindly sent us TCR monitor mineral. A.G. Johnson and colleagues at the Oregon State University Radiation center provided excellent service. H. Jade, A. Strum and M. Maase assisted with the early stages of sample preparation. We thank four anonymous reviewers for comments and suggestions that helped to improve the final version of this paper. [ VCI

J.M. O’Connor et al./ Earth and Planetary Science Letters 136 (1995) 197-212

Appendix A. Data tables

209

Sample: SO80 14D.S4 J=O.O003985 k O.OC0OOO4 Temp 40139 37139 36i39 KlCa z39Ar %4OAr’ 950 35.9035 308.2376 0.1551 0.001 0.237 56.6 IO00 36.361 I 307.1193 0.2517 O.OQl 0.452 36.4 1050 36.1072 298.6537 0. I326 0.001 0.627 64.5 1100 69.0593 289.7801 0.1117 0.001 0.756 84.2 1150 197.8228 261.6270 0.1230 0.002 0.853 91.1 1200 368.3563 284.5006 0.1632 0.001 0.902 92.0 1400 361.2223 275.5951 0.5422 0.001 I.000 68 .o Total fusion age. TFA= 73.5 f 3.4 Ma (including J) Weighted mean plateau age. WhIPA= 25.8 f 0.6 Ma (including I) Inverse isochron age =25.6 + 1.6 Ma. (MSWD =O.Ol: 4oAr/36Ar=295.8 f 16. I) Steps used: 950. IOOO, 1050, or 63% 2 ‘PAr

Sample: SO80 l2DS-I J=O.O003951 ? O.OOOOOO38 Temp 40139 37139 36J39 KG Z 39Ar % 40Ar’ 800 32.6056 31.3481 0.2017 0.015 0.043 35.8 900 30.9664 32.1849 0.0823 0.015 0.204 58.1 950 31.4547 31.8158 0.0701 0.015 0.321 62.8 loo0 30.8ooO 28.1589 0.04 I4 0.017 0.456 75. I IO50 30.8185 30.1126 0.0279 0.016 0.579 63.7 II00 25.1 I04 28.3532 0.0405 0.017 0.690 71.6 1200 33.0376 26.7456 0.0144 0.018 0.808 93.7 I250 18.6451 11.0109 0.03 I6 0.044 0.869 68.6 1300 26.8748 12.6486 0.0680 0.038 0.927 58.2 1400 30.3697 I I .6087 0.1120 0.042 1.000 48.3 Total fusion age. TFA= 20.98 + 1.72 Ma (including J) Weighted mean plateau age. WMPA= 22.0f0.5 Ma (includmg J) Inverse isochron age =21.6 f 1.1 Ma. (MSWD =0.07; 4oAr/36Ar=304 k 22) Steps used: 800. 900,950. IOOO, 1050, 1100, 1200, or 81% L 39Ar

Sample: SO80 I7DS-IA J&o003967 C O.OOOCGO31 Temp 40/39 37139 36/39 K/G Z 39Ar % 40&* 1000 20.8685 21.5922 0.0257 0.022 0.863 77.7 1025 21.1312 18.6013 0.05 I7 0.026 0.883 60. I I050 2 I .4970 17.8046 0.0902 0.027 0.901 45.7 1075 20.2921 18.7746 0.0794 0.026 0.913 47.7 II00 20.7033 18.8415 0.0666 0.026 0.925 52.9 II75 20.5481 18.1496 0.1212 0.027 0.946 37.1 1300 22.8101 18.6638 0.1316 0.026 0.966 37.6 I400 21.0303 19.1601 0.1294 0.025 I.000 36. I Total fusion age, TFA= 15.0 f 0.18 Ma (including 1) Weighted mean plateau age. WMPA= 14.9 f 0.2 Ma (including I) Inverse isochron age =l4.9 f 0.3 Ma. (MSWD =O.l4; 40Ar/‘6Ar=295.4 f 0.3) Steps used: IOCO, 1025. 1050. 1075, 1100. ll75,or95% Z 19Ar

Sample: SO80 18DS- IA J=O.o003943 + 0.00#3034 Temp 40139 37139 36139 WCs Z39Ar %4OAr* 800 15.8818 84.2649 0.5218 0.005 0.075 9.2 850 16.2384 84.6170 0.2638 0.005 0.175 17.7 900 16.5695 84.3006 0.3461 0.005 0.297 14.0 950 16.0372 83.7877 0.3449 0.006 0.434 13.7 loo0 15.7619 8 I .6358 0.2065 0.006 0.581 21.4 IO50 I6 3788 79.4156 0.1297 0.006 0.698 32.5 II00 16.781 I 75.6048 0. I332 0.006 0.777 32.2 1150 24.5233 70.4065 0. I845 0.007 0.827 32.3 1200 15.5530 72.2307 0.2408 0.006 0.855 18.4 1250 17.3538 67.0242 0.2906 0.007 0.877 17.0 1300 20.9577 62. I873 0.1919 0.008 0.895 27.9 1400 18.4491 69.3076 0.1169 0.007 0.976 37.6 1475 18.8315 71.8119 rOT74 0.006 1.000 76.6 Total fusion age, TFA= 12.0 -k 0.2 Ma (including I) Weighted mean plateau age. WbiPA= I I .5 + 0.2 Ma (including J) Inverse isochron age =I 1.7 + 0.3 Ma. (MSWD =0.6; @A#Ar=294.5 f 1.6) Steps used: 800.850.900,950, IOOO, 105O.llC0, or 78% 2 39Ar

Sample: E19209 J=O.O003641 f O.oooooO3 Temp 40139 37139 36i39 KKa Z39Ar % 4OAr*

E 0.2173 16.3884 0.0153 0.030 0.077 6.2 0.1937 15.9554 0.0091 0.030 0.25 I 11.7

I250 0.2179 15.6631 0.0086 0.03 I 0.547 14.1 1450 0.1945 15.5668 0.0053 0.031 l.CQO 35.8 Tocal fusion age, TFA= I33 f 22 ka (including J) Weighted mean plateau age. WMPA= I31 f 19 ka (including J) InvE?se isochron age =I26 f I3 ka. (MSWD =0.03; 4oAr/36Ar=299 k IS) Stepyused: 800. 925. 1250, 1450, or 100% Z 39Ar

Age (Ma) 25.6 + 0.9 26.0 + I.1 25.8 + I.1 49.0 f I.5 136.9 + 2.3 247.1 k 4.9 242.6+ 31.2

Age (Ma) 23.1 f 2.9 21.9 f 0.7 22.3 f. 1.2 21.8 + 0.9 21.8 f I.8 17.8 f 10.9 23.4 f 5.4 13.2 f 2.0 19.1 f 14.0 21.5 i 7.5

Age (Ma) 14.9 f 0.2 15.1 k 0.8 15.4 f 0.8 14.5 f 1.2 14.8 f 1.2 14.7 f 0.8 16.3 f 0.8 15.0 f 0.5

Age (Ma) I I.3 f 0.7 II.5 + 0.4 ll.7f0.4 ll.4f0.4 I I.2 f 0.3 Il.6 f 0.3 ll.9f0.4 17.4 f 0.5 I I .o * 0.9 12.3 f I .4 14.8 f 1.4 13.1 f 0.4 13.3 f I.1

Age (ka) I36 + 129 l27+66 143 f 39 I28 f 23

210 J.M. O’Connor et al./Earth and Planetary Science Letters 136 (1995) 197-212

Sample: SO80 29DS-4 J=O.o003743 r o.cwOBO41 Temp 40139 37139 36139 KG E39Ar % 4OAr* 800 0.3554 56.4790 0.0664 0.008 0.170 2.2 900 0.3383 56.4364 0.0365 0.008 0.413 4.9 lcoo 0.3481 55.7339 0.0406 0.008 0.698 4.2 1 loo 3.7685 54.8798 0.0357 0.009 0.864 36.6 1200 17.7247 48.6938 0.041 I 0.010 0.923 67.3 I400 48.4806 54.7504 0.0699 0.009 1.000 74.0 Total fusion age, TFA= 3807 f 76 ka (including J) Weighted mean plateau age. WMPA= 234 f 75 ka (includin

1 I)

Inverse isochron age =224 * 223 ka. (MSWD = 0.01 ; 40Ad 6Ar=296 f IO) Steps used: 800.900. lOC0, or 70% Z )9Ar

Sample: SOS0 35DS. I A J=O.O003766 f O.OOLXNO46 Temp 40139 3IJ39 36139 KlCa z39Ar c4oAr* 1000 0.8765 119.0520 0.0510 0.004 0.165 12.2 1125 0.88 I8 115.4828 0.0548 0.004 0.398 10.2 1200 I.1318 114.4154 0.0592 0.004 0.516 10.9 1400 3.2019 113.7500 0.0603 o.co4 1.000 24.9 Total fusion age. TFA= 1381 i I I3 ka (including J)

Weighted man plateau age. WMPA= 625 * I82 ka (including I) Inverse isochron age =249 zb 983 La. (MSWD = 0.04; 4”Ar/~“A~317 f 91) Steps used: 1000, 1125. 1200. or 52% X 39Ar

Sample: SO80 93DS-8 1=0.cO03850 * 0.oooo0041 Temp 4OJ39 37139 36139 KG z 39Ar % 40Ar* 900 2.3339 302.0529 0.1838 0.001 0.108 5.7 975 3.5855 296.5534 0.1186 0.001 0.236 19.4 1040 3.2687 291.6463 0.099 1 0.001 0.368 28.7 1100 4.5807 295.545 1 0.1121 0.001 0.484 26.7 1150 23.9388 294.6146 0.1044 0.001 0.604 71.0 1200 20.2576 283.8228 0.1169 0.001 0.691 56.9 1400 22.8294 280.1642 0.1152 0.001 l.ooo 60.3

Total fusion age, TFA= 9.25 f 0.33 Ma (including I) Weighted mean plateau age, WMPA= 2.4 + 0.5 Ma (including J) Inverse lsochron age =3.0 f I. I Ma. (MSWD = 0.2 ; 40Ar/36Ar=282 k 24) Steps used: 900.975, 1040. I IOO. or 48% Z 39Ar

Sample: SO80 7lDS-2 (plag) J=O.O+03948 ? O.oooooOl7 Tcmp 40139 37139 36139 K/G Z39Ar % 40Ar’ 700 10.5934 169.9957 0.1328 0.003 0.101 26.5 800 10.5882 169.8785 0.0841 0.003 0.215 44.7 875 10.6916 165.8843 0.0727 0.003 0.367 52.7 925 10.7044 171.8742 0.0797 o.ocl3 0.463 48.3 1075 10.7603 158.1977 0.0714 0.003 0.700 52.4 1175 12.8489 153.0432 0.078 I 0.003 0.818 50.8 1425 16.8244 137.5799 0.0715 o.DO3 I ._300 59.5 Total fusion age. TFA= 8.57 ? 0. I6 Ma (including I) Weighted mean plateau age. WMPA= 7.6 f 0.2 Ma (including J) Inverse isochron age =I.7 ? 0.5 Ma. (MSWD =O.Ol; ““A&Ar=293 + IS) Steps used: 7OO.800. 875.925, 1075. or 70% 1 “‘AI

Sample: SO80 67DS-I J=O.o003852 i O.OOOC0O43 Temp 40139 37139 36139 KG Z 39Ar % 4OAr’ 800 12.4680 155.6927 0.2509 0.003 0.080 15.3 875 12.2018 160.6763 0.1170 0.003 0.199 33.1 925 12.0764 157.8225 0.0853 0.003 0.310 43.6 975 I I .9345 155.1674 0.1028 o.cil3 0.428 36.9 1025 I I .9572 153.4687 0.0637 0.003 0.549 61.0 1075 12.0596 148.4866 0.0616 0.003 0.653 62.2 1150 26. I737 149.3594 0.0836 0.003 0.776 64.3 I225 69.2740 143.3852 0.093 I 0.003 0.850 79.3 1400 68.2830 134.9530 0. I834 0.w3 0.981 58.7 I475 213.9087 127.7946 0.6606 0.004 l.OtXJ 51.3 Total fusion age, TFA= 20.23 f 0.14 Ma (including J) Weighted mean plateau age. WMPA= 8.4 f 0.1 Ma (Including I) Inverse isochron age =8.3 f 0.2 Ma. (MSWD =O.OS; 40Ar/36Ar=298 f 3) Steps used: 800,875,925,975. 1025. 1075, or 65% Z ‘9Ar

Age (ka) 240 + 169 228 + 129 235C II0 2543 + 170 I I929 f 484 32442 f 426

Age (ka) 595 + 322 599 ? 252 769 * 452 2174 + 124

Age (Ma) I.6 f 1.2 2.5 + I.1 2.3 i 1.0 3.2 f I.1 16.5 + 0.6 14.0 f 1.6 15.8 _+ 0.4

Age (Ma) 7.5 i 0.6 7.5 f 0.5 7.6 _+ 0.4 7.6 C_ 0.6 7.6 + 0.2 9.1 f 0.5

Il.9 f 0.3

Age (Ma) 8.6 f 0.4 8.5 + 0.2 8.4 + 0.2 8.3 + 0.2 8.3 f 0.2 8.4 f. 0.2 18.1 + 0.2 47.5 + 0.4 46.8 f 0.4 142.8 f 2.8

xx9Ar is cumulative, “OAr l = rad fraction.

Decay constants: I = 5.53 X lo-“/yr; .“Ar/ mAr listed above have been comected for decay since irradiation (h”Ar = 1.975

10-2/&y). Correction factors used: 2.64E - 4 ( f 0.017) = (36Ar/ “‘Ar)Ca [39]. * 8.6E - 4 ( f 0.00007) = (“Ok/ “Ar)K. 6.73E - 4 ( iO.037) =

(‘9Ar/ “Ar)Ca [25].

* CLICIT (Cd Shielded), 12 hour irradiation at OSU TRIGA [261.

J.M. O’Connor et al./ Earth and Planetary Science Letters 136 (1995) 197-212 211

References

[l] W.J. Morgan, Convection plumes in the lower mantle, Na-

ture 230, 42-43, 1971.

[2] E. Bonatti, C.G.A. Harrison, D.E. Fisher, J. HOMO~~Z, J.-G.

&hilling, J.J. Stipp and M. Zentilli, Easter volcanic chain

(Southeast Pacific): a mantle hot line, J. Geophys. Res. 82,

2457-2478, 1977.

[3] R.H.J. Pilger and D.W. Handschumacher, The fixed-hotspot

hypothesis and origin of the Easter-Sala y Gomez-Nazca

trace, Geol. Sot. Am. Bull. 92, 437-446, 1981.

[4] E.A. Okal and A. Cazenave, A model for the plate tectonic

evolution of the east-central Pacific based on SEASAT inves-

tigations, Earth Planet. Sci. Len. 72, 99-116, 1985.

[5] J.G. Schilling, H. Sigurdsson, A.N. Davis and R.N. Hey,

Easter microplate evolution, Nature 317, 325-331, 1985.

[6] B.B. Hanan and J.-G. Schilling, Easter microplate evolution:

Pb isotope evidence, J. Geophys. Res. 94, 7432-7448, 1989.

[7] D.F. Naar and R.N. Hey, Tectonic evolution of the Easter

microplatc, J. Geophys. Res. 96, 7961-7993, 1991.

[8] J.G. Clark and J. Dymond, Geochronology and petrochem-

istry of Easter and Sala y Gomez islands: implications for the

origin of the Sala y Gomez Ridge, J. Volcanol. Geotherm.

Res. 2, 29-48, 1977.

[9] P.E. Baker, F. Buckley and J.G. Holland, Petrology and

geochemistry of Easter Island, Contrib. Mineral. Petrol. 44,

85-100, 1974.

[lo] C.L. Mayes, L.A. Lawyer and D.T. Sandwell, Tectonic

history and new isochron chart of the South Pacific, J.

Geophys. Res. 95, 8543-8567, 1990.

[ 1 I] D.F. Naar and R.N. Hey, Recent Pacific-Easter-Nazca plate

motions, in: Evolution of mid-ocean ridges, J.M. Sinton, ed.,

Geophys. Union Geophys. Monogr. 57, 9-30, 1989.

[I21 J.M. O’Connor and R.A. Duncan, Evolution of the Walvis

Ridge-Rio Grande Rise hot spot system: Implications for

African and South American plate motions over plumes, J.

Geophys. Res. 95, 17.475-17.502, 1990.

[13] J.M. O’Connor and A.P. le Rocx, South Atlantic hot spot-

plume systems: 1. Distribution of volcanism in time and

space, Earth. Planet. Sci. Lett. 113, 343-364, 1992.

[14] F.M. Richter, Convection and the large-scale circulation of

the mantle, J. Geophys. Res. 78, 8735-8745, 1973.

[15] F.M. Richter and B. Parsons, On the interaction of two scales

of convection in the mantle, J. Geophys. Res. 80, 2529-2541,

1975.

[ 161 W.F. Haxby and G. Weissel, Evidence for small-scale mantle

convection from Seasat Altimeter data, J. Geophys. Res. 91,

3507-3520, 1986.

[17] D.L. Turner and R.D. Jarrard, K-Ar dating of the Cook-

Austral island chain: a test of the hot-spot hypothesis, J.

Volcanol. Geotherm. Res. 12, 187-220, 1982.

[ 181 M. Maia and M. Diament, An analysis of the altimetric geoid

in various wavebands in the central Pacific Ocean: con-

straints on the origin of intraplate features, Tectonophysics 190, 133-153, 1991.

[l9] W.J. Morgan, Rodriguez. Darwin, Amsterdam..., a second

type of hotspot island, J. Geophys. Res. 83, 5355-5360,

1978.

[20] Z.J. Liu, Y. Rappaport and D.F. Naar, GLORIA-B side-scan

and swath bathymetry along the Easter Seamount Chain,

EOS Trans.Am. Geophys. Union 688, 1993.

[21] R.A. Hagen, N.A. Baker, D.F. Naar and R.N. Hey, A

SeaMARC II survey of recent submarine volcanism near

Easter Island, Mar. Geophys. Res. 12, 297-315, 1990.

[22] P. Stoffers, R. Hekinian and the Scientific Party, Cruise

report SONNE 80a - Midplate III oceanic volcanism in the

South Pacific, 58, Christian-Albrechts Univ., Kiel, Dept. of

Geology and Paleantology, 1992.

[23] P. Staffers, R. Hekinian, K.M. Haase and Scientific Party,

Geology of young submarine volcanoes west of Easter Is-

land, Southeast Pacific, Mar. Geol. 118. 177-185, 1994.

[24] R. Hekinian, P. Staffers, D. Ackermand, N. Binard, J.

Francheteau, C. Devey and D. Garbc-Schonberg, Magmatic

evolution of the Easter Microplate-Crough region (Southeast

Pacific), Mar. Geophys. Res., in press, 1995.

(251 G.B. Dahymple and W.A. Duffield, High precision

“Ar/ ‘9Ar dating of Oligocene rhyolites from the

Mogollon-Datil Volcanic Field, using a continuous laser

system, Geophys. Res. Lett. 14, 462-464, 1988.

[26] J.R. Wijbrans, M.S. Pringle and A.A.P. Koppers, Argon

Geochronology of small samples using the VULKAAN ar-

gon laserprobe, Scheveers. Proc. Kon. Ned. Akad. Wetensch.

98(2), 185-218, 1995.

[27] A.K. Baksi, K.A. Hoffman and M. McWilliams, Testing the

accuracy of the geomagnetic polarity time-scale (GPTS) at

2-5 Ma, utilizing “Ar/ j9Ar incremental heating data on

whole-rock basalts, Earth Planet. Sci. Lett. 118, 135-144,

1993.

[28] D. York, Least squares fitting of a straight line with corre- lated errors, Earth Planet. Sci. Lett. 5, 320-324, 1969.

[29] C. DeMets, R.G. Gordon, D.F. Argus and S. Stein, Effect of

recent revisions to the geomagnetic reversal time scale on

estimates of current plate motions, Geophys. Res. Len. 21,

2191-2194, 1994.

[30] R.L. Fisher and R.M. Norris, Bathymetry and geology of

Sala y Gomez, southeast Pacific, Geol. Sot. Am. Bull. 71, 497-505, 1960.

1311 R. Hey, Tectonic evolution of the Cocos-Nazca spreading

center, Geol. Sot. Am. Bull 88, 1404-1420, 1977.

[32] D.M. Christie, R.A. Duncan, A. McBimey, M.A. Richards,

W.M. White, K.S. Harpp and C.G. Fox, Drowned islands

downstream from the Galapagos hotspot imply extended

speciation times, Nature 355, 246248, 1992.

[33] W.M. White, A. McBimey and R.A. Duncan, Petrology and

Geochemistry of the Galapagos Islands: Portrait of a patho-

logical mantle plume, J. Geophys. Res. 98, 19,533-19,563,

1993.

[34] P.E. Baker, A. Gledhill, P.K. Harvey and C.J. Hawkesworth,

Geochemical evolution of the Juan Femandez Islands, SE

Pacific, J. Geol. Sot. London 144, 933-944, 1987.

[35] J.B. Minster and T.H. Jordan, Present-day plate motions, J.

Geophys. Res. 83, 5331-5354, 1978.

[36] R.C. Courtney and R.S. White, Anomalous heat flow and the

212 J.M. O’Connor et al. / Earth and Planetary Science Letters 136 (1995) 197-212

geoid across the Cape Verde Rise: evidence for dynamic support from a thermal plume in the mantle, Geophys. J. R. Astron. Sot. 87, 815-867, 1986.

[37] D.L. Anderson, T. Tanimoto and Y. Zhang, Plate tectonics and hotspots: The third dimension, Science 256, 1645- 165 1, 1992.

[38] D.T. Anderson, Y. Zhang and T. Tanimoto, Plume heads, continental Iithosphere, flood basalts and tomography, Mag-

matism and the causes of continental break-up, in: B.C. Storey, T. Alabaster and R.J. Pankhurst, eds., Geol. Sot. London Spec. Publ. 68, 99-124, 1992.

[39] G.B. Dahymple, E.C. Alexander, M.A. Lanphere and G.P. Kraker, Irradiation of samples for @Ar/ 39Ar dating using the Geological Survey TRIGA reacrtor, U.S. Geol. Surv. Prof. Pap. 1176, 1981.

Documents

Time-space mapping of Easter Chain volcanism