Earth and Planetary Science Letters 257 (2007) 391 – 406
www.elsevier.com/locate/epsl
The role of lithospheric gabbros on the composition
of Galapagos lavas
A.E. Saal a,⁎, M.D. Kurz b , S.R. Hart c , J.S. Blusztajn c ,
J. Blichert-Toft d , Y. Liang a , D.J. Geist e
a
b
Department of Geological Sciences, Brown University Providence, Rhode Island 02912, United States
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543, United States
c
Department of Geology and Geophysics, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543, United States
d
Ecole Normale Supérieure de Lyon, CNRS UMR 5570, 69364 Lyon, France
e
Department of Geological Sciences, University of Idaho, Moscow, Idaho 83844, United States
Received 20 November 2005; received in revised form 24 February 2007; accepted 27 February 2007
Available online 6 March 2007
Editor: R.W. Carlson
Abstract
New geochemical data combined with a compilation of all published data for the Galapagos basalts indicate that the Galapagos
Archipelago can be divided into two very different volcanic areas, east and west of the 91° W fracture zone. In both regions, the
interaction between plume melts and depleted upper mantle lavas most likely controls the composition of the erupted basalts.
However, we find important differences in the trace element compositions between lavas from the two regions. First, Ti/Gd and Nb/
La ratios correlate with 3He/4He, and all three ratios decrease with increasing distance from Fernandina. These correlations suggest
that the high Ti/Gd and Nb/La ratios are typical of the plume mantle source that carries the high 3He/4He signature. Second, there is
an increase in the Ba/Th, K/La, Sr/Nd, Eu/Eu⁎, Pb/Nd ratios, and a higher 87Sr/86Sr at similar 143Nd/144Nd (high ΔSr) from west to
east across the fracture zone, and from north to south within the eastern region. These geochemical characteristics resemble those of
plagioclase-rich cumulates from the oceanic crust and ophiolite complexes. The compositions of the lavas from the eastern region
can be explained by interaction of basalts with plagioclase-rich cumulate during melt percolation through the oceanic lithosphere.
We argue that kinetic interaction of magmas with plagioclase-rich cumulates, previously formed either in the Galapagos Spreading
Center or beneath the leading edge of the plume on the western region, are responsible for the observed composition of basalts
erupted in the eastern region.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Galapagos; basalt; geochemistry; gabbro
1. Introduction
⁎ Corresponding author. Tel.: +1 401 863 7238; fax: +1 401 863
2058.
E-mail address: [email protected] (A.E. Saal).
0012-821X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsl.2007.02.040
The geochemistry of oceanic basalts has been a
powerful tool in unraveling the composition, the scale of
heterogeneity, and the dynamics of the mantle in areas
complicated by the juxtaposition of mid-ocean ridges
and hot spot magmatism. One of the best known
392
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
examples of hot spot–ridge interaction is the Galapagos
Archipelago, whose location adjacent to the Galapagos
Spreading Center (GSC) prompted numerous petrological–geochemical [1–8] geophysical [7–14], and fluid
dynamics [15–18] studies on the archipelago and the
spreading center.
The compositional variation of lavas from the
Galapagos Islands partly reflects the mantle heterogeneity
beneath the archipelago. Previous studies have invoked
the existence of four main mantle components contributing to the magmatism of the islands: two different types of
enriched mantle represented by basalts from Pinta and
Floreana Islands, a high-3He/4He (22 to 29 Ra; [19])
source corresponding to lavas from Fernandina, and a
depleted mantle component exemplified by basalts from
Genovesa Island [[1,5,8] and references therein]. Although the origin of the enriched and high-3He/4He mantle components has been discussed extensively, the role of
the depleted component in the Galapagos lavas has received comparatively less attention; it has been interpreted to reflect either the entrained upper mantle [1,5,8]
or an intrinsic part of the plume itself [4]. Meanwhile, the
impact of the oceanic lithosphere on the geochemistry of
the Galapagos basalts has mostly been ignored.
The present study focuses on the depleted mantle
component observed in the central region of the Galapagos Archipelago, and the effect that the oceanic lithosphere exerts on the erupted lavas. Although the basalts
from the central region of the archipelago most probably
represent entrained asthenospheric mantle into the plume,
our results indicate that these lavas have unusual trace
element signatures compared to normal MORB and suggest the interaction between the lavas and plagioclase-rich
cumulates during melt migration through the oceanic
lithosphere.
2. Geodynamic background
The Galapagos Archipelago is located in the eastern
equatorial Pacific, 1000 km west of South America, and
approximately 150 km south of the active Galapagos
Spreading Center (GSC). The Galapagos Islands constitute one of the most active volcanic regions in the world
with 21 emergent volcanoes, of which 16 have had historic
to Holocene eruptions within an area of 45,000 km2 [1,8].
The islands and numerous seamounts are constructed
mainly upon a shallow submarine volcanic platform overlying a young lithosphere b15 Ma old [5,7,9,20].
The Galapagos Islands have been interpreted as the
surface manifestation of a mantle plume. Seismic tomography images showing a slow velocity anomaly
extending into the lower mantle [11,14] and an anoma-
lously thin mantle transition zone [12] beneath the
Archipelago support a hot spot origin for the plume.
However, geological, geophysical and geochemical
studies have found it challenging to fit all aspects of the
Galapagos Archipelago to a “standard” plume model [5].
In contrast to Hawaii, the “archetype” of the plume model,
the Galapagos Islands do not form a linear chain; the
volcanism is not strictly time-transgressive in the direction
of plate motion, nor is there a consistent petrological or
geochemical evolution of the volcanism such as seems to
be the case for Hawaii. Nevertheless, the ages of the oldest
lavas dated from each volcano form an overall progression from young in the west to old in the east [1,2,21]. The
islands, as they drift away from the hot spot, have had
intermittent eruptions for approximately 3 Ma, as for
example in the case of San Cristobal Island [1,2,7,21].
The continuous volcanism in this most easterly island
indicates either an extended zone of melting downstream
from the plume, or the existence of post-shield extensional
volcanism across the Galapagos platform.
The analysis of bathymetric and gravity data by
Feighner and Richards [9] indicates that the crustal
thickness in the Archipelago increases from approximately 10 km at the margin of the Galapagos platform to as
much as 18 km beneath the southeastern end of Isabela
Island. Furthermore, their study reveals that the Galapagos
platform is built on two lithospheric sectors, west and east
of the 91° W fracture zone, which represents an estimated
∼3 Ma age discontinuity across the lithospheric boundary. The eastern sector of the Archipelago is underlain by a
weak lithosphere with an elastic thickness of 6 km or less,
and close to Airy compensation. In contrast, the western
sector is flexurally supported by a lithosphere with an
elastic thickness of 12 km. Surface wave tomographic
study of the upper mantle beneath the Galapagos
Archipelago confirms the different thickness of a “highvelocity lid” across the 91° W fracture zone reaching
depths of ∼70 km beneath Fernandina, and ∼40 km
beneath the eastern region [14]. Although Feighner and
Richards [9] suggest that the age offset across the 91° W
fracture zone extends south of the Galapagos Islands, a
compilation of magnetic data [20,22] suggests that the age
offset terminates within the Archipelago.
The present-day volcanic and seismic activities are
also distinctive in both areas of the Archipelago, west
and east of the 91° W fracture zone. In the western part
of the Archipelago (Fernandina and Isabela Islands)
the eruptive edifices are large central shield volcanoes
of unusual shape (“inverted soup-bowl”) with welldeveloped calderas. These volcanoes have the most
frequent historic volcanic and seismic activity due to
their location close to what is considered the leading
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
(western) edge of the plume with respect to plate motion
at Fernandina, the most active volcano in the Archipelago [2,23]. In contrast, the volcanoes in the eastern,
central and southern areas (hereafter referred to as the
“eastern sector”) are seismically less active and present
more varied morphologies and eruptive histories.
Thicker lithosphere and the present-day higher magma
supply rate in the western sector probably are responsible for yielding large volcanic structures [3].
The composition of the Galapagos lavas also varies
within the Archipelago, consistent with the differences in
volcanic activity. Lavas erupted from the large western
volcanoes are moderately fractionated tholeiites that
cooled and fractionated at low pressure (1–3 kb) in the
crust. In contrast, lavas erupted from small cones and
vents in the eastern sector are compositionally diverse,
ranging from picrites to basanitoids; they cooled and
fractionated below the Moho at pressures N 5 kbar [[3]
and references therein]. Several studies have used major
and trace element contents, to suggest that the onset of
melting is deeper (garnet stability field) and more extensive in the western part of the Archipelago, and less
extensive and shallower (spinel stability field) in the
eastern sector [1–3,5,24]. Finally, the Sr, Nd, Pb and Hf
isotopic composition of the Galapagos lavas indicate that
both enriched and depleted mantle components are present in the western and eastern sectors across the 91° W
fracture zone [1,5,8]. The exception is the high-3He/4He
mantle component found in lavas from Fernandina island
(3He/4He 22 to 29 Ra; [19]), which decreases steadily as
the distance from Fernandina increases [2].
There is a general agreement that the basalts from a
heterogeneous Galapagos plume are diluted with upper
mantle melts in the central part of the Archipelago [[1,3],
and references therein]. Two different geodynamic
models have been proposed to explain this pattern.
Richards and Griffiths [25] presented a model that
involves entrainment of asthenospheric mantle into a
plume subject to bending and velocity shear by the effect
of the regional mantle flow. In contrast, Bercovici and
Lin [26] proposed that the plume cools by contact with
the lithosphere producing gravity currents that cause the
plume to take on a mushroom shape, with the hot asthenosphere being entrained in the center, while the plume
component is located at the colder margins of the plume.
3. Sampling and analytical techniques
3.1. The data considered in this study include
1) A compilation of all published major, trace element,
and isotopic compositions reported for lavas with
393
MgO ≥ 5 wt.% from the Galapagos Archipelago.
This compilation comprises: data from the GEOROC
database, the latest published data from the literature
not yet in the database, and our own new unpublished
data.
2) New sampling of historic to Holocene lavas from
numerous volcanoes and islands from the Galapagos
Archipelago. Most of the selected samples are younger
than 50,000 year old, dated by surface exposure
techniques using cosmogenic 3He [2,27–29].
The composition and the young ages of the samples
provide a “present-day” picture of the distribution of the
compositionally distinctive sources beneath the Archipelago. The dataset reported in Table 1 is only a summary of selected representative samples from each
island. The complete dataset is presented in Table 1 of
the Supplementary Information.
Major and trace element data reported here were
measured by XRF and ICP-MS at the GeoAnalytical
Lab, Washington State University. Precisions for most
major and trace elements are better than 3% and 10%
(2σ), respectively. Sr, Nd and Pb isotopic measurements
were carried out on the VG-354 TIMS multi-collector at
the Woods Hole Oceanographic Institution. External
reproducibilities of the standards were: 16 ppm (2σ,
n = 35) for 143 Nd/ 144 Nd, 26 ppm (2σ, n = 35) for
87
Sr/86Sr and 0.05% per amu (2σ, n = 60) for Pb isotopes. The Sr, Nd and Pb total procedure blanks were less
than 40 pg, 70 pg and 150 pg, respectively, and are all
negligible. Hf isotopes were measured using the MCICPMS Plasma 54 at the Ecole Normale Supérieure in
Lyon. The external reproducibility of 176 Hf/177Hf on
replicate Hf standard solutions was better than 30 ppm,
and the total procedural Hf blank was less than 25 pg,
and thus also negligible (see Supplementary Information
for further details on all analytical techniques employed).
4. Results
For this study we selected all samples with MgO
contents ≥5 wt.%. All of the lavas are basalts, mostly
tholeiitic with only a few alkali basalts. SiO2 contents and
Mg# range from 51.9 to 44.19 wt.% and from 0.76 to 0.40,
respectively. The Mg# indicates that most of the samples
have undergone fractional crystallization and some have
been affected by olivine accumulation (Fig. 1).
We divide the lavas from the Galapagos Archipelago
into two main groups: The “western group”, formed by
basalts erupted from large shield volcanoes (volcanoes
on Fernandina and Isabela Islands) close to the leading
edge of the plume on an older and thicker lithosphere
394
Major, trace element and isotopic composition of representative Galapagos lavas
Sample ID
NSK97-215
NSK97-238
NSK97-252
NSK97-254
SG93-19
GI92-10
GI92-1
SN 91-60
GI92-16
GI92-18
GI92-19
W95-84
SKV98-108
SKV98-101
Island/volcano
Fernandina
Santiago
San
Cristobal
San
Cristobal
Floreana
Cerro Azul
Santa Cruz
Sierra
Negra
Alcedo
Darwin
Ecuador
Wolf
Pinta
Marchena
Lat. south
0
22.59
0
14.25
0
43.54
0
45.82
1
13.722
0
58.97
0
44.54
0
56.30
0
23.51
0
12.70
0
1.57
0
Long. west
91
33.93
90
51.03
89
24.63
89
27.31
90
25.465
91
26.63
90
17.53
90
56.30
91
13.15
91
23.10
91
31.14
91
SiO2
Al2O3
TiO2
FeO
MnO
CaO
MgO
K2O
Na2O
P2O5
Total
XRF
Ni
Cr
Sc
V
Ba
Rb
Sr
Zr
Y
Nb
Ce
ICPMS
Cs
Rb
0.08
0
0.30
90
36.52
0
21.43
44.41
90
32.18
48.54
13.95
2.72
10.42
0.175
10.68
10.39
0.36
2.49
0.269
100.00
48.40
14.85
2.13
10.67
0.178
10.25
10.22
0.31
2.74
0.242
100.00
47.30
16.77
1.01
8.83
0.165
12.37
11.17
0.12
2.16
0.104
100.00
47.44
16.40
0.98
9.04
0.170
11.55
11.80
0.19
2.31
0.112
100.00
46.70
14.16
1.74
9.49
0.176
11.25
13.64
0.48
2.17
0.190
100.00
48.39
13.97
3.73
13.69
0.223
9.90
5.31
0.78
3.59
0.433
100.00
47.44
16.13
1.72
10.60
0.171
9.84
10.55
0.22
3.08
0.255
100.00
48.97
14.13
3.48
13.29
0.21
10.59
5.62
0.50
2.85
0.36
100.00
49.64
14.00
3.06
12.36
0.204
10.82
6.11
0.50
2.91
0.408
100.00
48.80
16.74
2.58
10.05
0.164
12.39
6.07
0.32
2.64
0.240
100.00
47.93
14.32
2.29
11.00
0.170
9.90
10.90
0.52
2.69
0.278
100.00
48.58
15.21
2.88
10.97
0.18
11.20
7.09
0.50
3.05
0.35
100.00
48.53
16.84
2.04
9.13
0.15
12.77
6.96
0.53
2.79
0.26
100.00
48.73
14.42
2.51
12.37
0.20
11.49
6.31
0.32
3.38
0.26
100.00
226
643
35
327
40
7
302
135
24
18.3
22
241
495
34
246
27
5
234
139
29
11.1
26
211
481
40
228
28
3
173
66
21
5.2
9
259
479
40
210
44
3
173
67
21
6.5
14
336
828
34
264
145
9
338
93
22
16.2
26
27
46
30
422
160
16
374
226
39
40
60
236
406
30
198
0
2
374
158
26
6.2
50
40
101
30
385
102
8
298
192
36
27.6
29
44
98
38
351
50
8
284
180
35
21.5
40
56
152
35
309
24
3
382
155
26
12.8
38
265
499
27
226
95
10
344
150
25
22.1
38
89
248
28
343
58
3
374
216
34
17.3
51
61
78
35
247
124
8
340
143
26
20.9
34
38
132
44
335
41
3
221
164
39
12.5
23
0.06
5.96
0.05
5.09
0.02
1.21
0.04
2.28
0.14
9.42
0.16
14.66
0.03
2.25
0.11
9.8
0.09
7.94
0.06
3.61
0.11
9.21
0.06
5.9
0.06
8.6
0.05
5.16
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
Table 1
73.47
1.05
0.30
16.89
1.24
11.83
27.00
1.15
3.66
289.08
17.11
3.35
4.88
1.78
5.09
0.87
5.13
24.32
0.99
2.46
0.34
2.00
0.29
33.62
0.703264
0.512922
0.283054
19.109
15.548
38.758
0.947
48.64
0.76
0.21
9.60
0.65
9.00
20.96
1.07
2.99
229.61
14.86
3.53
4.90
1.79
5.61
0.94
5.86
29.18
1.12
2.93
0.39
2.37
0.35
32.35
0.703036
0.512998
0.283033
19.015
15.561
38.657
0.946
23.89
0.24
0.06
3.57
0.25
3.38
8.46
0.90
1.31
173.67
6.79
1.65
2.40
0.96
3.16
0.61
4.08
22.96
0.89
2.56
0.37
2.36
0.37
45.35
0.702991
0.513041
0.283144
18.704
15.555
38.386
0.948
38.49
0.37
0.10
4.94
0.31
4.31
9.53
0.75
1.36
167.06
6.58
1.54
2.18
0.89
2.84
0.55
3.87
22.08
0.83
2.37
0.34
2.32
0.36
41.56
0.703086
0.513042
0.283139
18.783
15.546
38.416
0.943
180
0.89
0.17
16.14
0.91
11.99
21.86
1.72
2.69
338
11.88
1.85
3.39
1.22
3.69
0.66
4.14
22.46
0.82
2.41
0.31
1.85
0.3
0.703446
0.512945
178.98
2.48
0.68
35.00
2.32
26.32
54.42
2.06
6.91
377.61
30.59
5.72
8.01
2.71
8.00
1.33
7.83
39.40
1.55
3.88
0.54
3.22
0.48
34.13
0.703321
0.512934
0.702952
0.513099
19.614
15.599
39.300
0.953
19.337
15.561
38.955
0.945
18.521
15.497
38.038
0.929
For a complete dataset of the samples analyzed See Table 1 in the Supplementary Information.
34
0.37
0.09
5.71
0.36
7.92
20.5
0.89
3
374
14.62
3.37
4.37
1.64
4.48
0.84
5.17
26.89
1.04
2.9
0.4
2.38
0.38
120
1.64
0.48
26.97
1.89
19.03
42.35
1.43
5.53
308
25.68
5.09
7.22
2.52
7.76
1.31
7.56
37.05
1.47
3.74
0.51
3.04
0.45
34.4
0.703397
0.512925
0.283034
19.357
15.591
39.022
0.950
85
1.25
0.3
21.04
1.35
15.78
35.45
2.9
4.76
284
22.19
4.46
6.71
2.32
6.94
1.23
7.2
35.59
1.39
3.71
0.5
2.94
0.44
44
0.66
0.18
11.9
0.8
11.52
26.76
2.27
3.68
382
17.43
3.36
5.2
1.89
5.36
0.89
5.29
25.42
1.01
2.7
0.35
2.01
0.29
120
1.27
0.28
20.54
1.2
15.73
32.27
1.68
4.09
344
18.84
3.36
5.23
1.81
5.22
0.89
5.12
25.99
0.99
2.67
0.35
2.04
0.3
0.703174
0.512962
0.702856
0.513020
0.283068
18.810
15.536
38.268
0.925
0.703200
0.512951
0.283051
19.304
15.600
39.059
0.959
19.116
15.552
38.699
0.940
60
1.03
0.35
15.74
1.14
15.74
37.81
1.38
5.22
388
24.56
5.46
7.1
2.49
7.4
1.24
7.06
34.04
1.37
3.36
0.46
2.71
0.4
36.9
0.702747
0.513041
0.283104
18.897
15.542
38.377
0.928
130
1.66
0.39
19.98
1.39
15.69
33.48
0.95
4.29
349
19.68
3.78
5.36
1.94
5.62
0.95
5.51
26.31
1.08
2.71
0.37
2.2
0.33
32.4
0.703288
0.512909
0.283012
19.156
15.594
39.003
0.967
51
0.73
0.21
10.78
0.76
10.01
24.70
0.85
3.62
235
18.46
4.39
6.07
2.23
7.33
1.29
7.65
40.25
1.56
4.10
0.57
3.41
0.52
40.23
0.702843
0.513028
0.283109
18.893
15.563
38.500
0.941
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
Ba
Th
U
Nb
Ta
La
Ce
Pb
Pr
Sr
Nd
Hf
Sm
Eu
Gd
Tb
Dy
Y
Ho
Er
Tm
Yb
Lu
Sc
87
Sr/86Sr
143
Nd/144Nd
176
Hf/177Hf
206
Pb/204Pb
207
Pb/204Pb
208
Pb/204Pb
208⁎ 206⁎
/
395
396
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
Fig. 1. A) TiO2 and B) Al2O3 versus MgO contents of Galapagos lavas. The analyses used are a compilation of all the published data available
through the GEOROC database complemented with the data presented in this study. Only samples with MgO content N5 wt.% have been considered
in our work. Dashed arrow indicates the olivine control line. We also plot the median for gabbros from Annieopsquotch [30] the oceanic crust (two
median values are presented one for “protoliths" and one for “strips") [31] and Oman [[32] and reference therein]. Symbols in black and red represent
lavas from the “western" and “eastern" groups respectively.
located to the west of the 91° fracture zone [12,14,28].
The second group, the “eastern group”, represents lavas
erupted from small cones and vents located in islands
further away from the leading edge of the plume on a
relatively thin and young lithosphere. Most of the lavas
in the eastern group are located to the east of the 91°
fracture zone, with the exception of the basalts from
Darwin and Wolf Islands, which actually erupted to the
west of the fracture zone. It is important to emphasize
that the present-day eruption rates are different between
both groups, with the western group having higher
eruption frequency than the eastern group [2,3,28].
The compilation of all the existing data reveals that
even though lavas from both groups share similar ranges
in Sr, Nd, Pb, and Hf isotopes (with the eastern group
displaying a larger range), they have important differences in their major and trace element compositions.
Lavas from the eastern group range from similar to
lower TiO2 and higher Al2O3 contents, lower incompatible trace element abundances, and lower MREE/
HREE ratios (e.g., Sm/Yb, Hf/Lu) than those from the
western group at similar MgO contents (Figs. 1 and 2).
Moreover, while lavas from the western volcanoes are
characterized by positive Nb and Ti anomalies, and
negative K and Pb anomalies in primitive-mantle-normalized plots, many basalts from the eastern group are
distinguished by negative Th anomalies, positive Ba, Sr
and small Eu anomalies, and small or no negative
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
397
Fig. 2. Primitive mantle-normalized diagram for representative samples from the Galapagos Archipelago: A) The western region (Fernandina, Cerro
Azul and Sierra Negra volcanoes). B) The eastern region (Floreana, Santa Cruz and San Cristobal Islands). For comparison, we included
representative analyses of gabbros from the oceanic crust [31] and Oman [[32] and reference therein]. Normalizing values are from McDonough and
Sun, [55]. Symbols as in Fig. 1.
anomalies for K and Pb. These anomalies represent
ratios involving pairs of trace elements with similar
incompatibilities, which are not easily fractionated
during mantle melting. Thus, high Ba/Th, K/La, Sr/Nd
(∼ Sr/Sr⁎), Eu/Eu⁎ and Pb/Nd ratios characterize the
source of many lavas from the eastern group, while high
Ti/Gd and Nb/La (∼ Nb/Nb⁎), and low K/La and Pb/Nd
ratios distinguish the source of basalts from the western
group (Fig. 2, and Supplementary Information Fig. 1
and Table 2).
To avoid the problems of temporal variation between
lavas from different volcanoes/islands, the lack of interlaboratory calibrations, and the differences in analytical
techniques inherited in the compilation of all the existing Galapagos data, we plot key trace element ratios
using only our own new data set (Fig. 3). Most selected
samples are less than 50 ka old, with the exception of
three samples from Floreana (100–200 ka) and two
samples from Santa Cruz, (100 ka and 500 ka), providing a present-day picture of the compositional variation across the archipelago. All samples were analyzed
in the same laboratory, and batches of samples from
both eastern and western groups were analyzed during
the same analytical sessions. Our data confirm the observations made using the compilation of all the existing
data, and demonstrate that the different trace element
compositions between eastern and western groups are
not an artifact of the data compilation.
The relative high Ba/Rb, Ba/Th (Th/Th⁎), K/La, Sr/
Nd, Eu/Eu⁎ and Pb/Nd ratios in some lavas from the
eastern group resemble those of plagioclase-rich cumulates from the oceanic crust and ophiolite complexes
(hereafter referred to as the “plagioclase-rich gabbro”
signature). Therefore, we compared the composition of
398
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
Fig. 3. Ti/Gd, Ba/Rb, K/La, Sr/Sr⁎, and Th/Th⁎ versus Nb/Nb ⁎ (La) for individual Galapagos lavas reported in Table 1 of the Supplementary
Information. Note that most of the samples are younger than 50 ka with the exception of three samples from Floreana (100–200 ka) and two samples
from Santa Cruz, (100 ka and 500 ka). We plot the median for gabbros from Annieopsquotch [30] the oceanic crust [31] and Oman [[32] and reference
therein]. Nb/Nb ⁎ (La) = [Nb / ((Ba + La) / 2)] and Sr/Sr⁎ = [Sr / ((Pr + Nd) / 2)] and Th/Th⁎ = [Th / ((Ba + La) / 2)] using normalized values. Normalizing
values are from McDonough and Sun [55]. Symbols as in Fig. 1.
the Galapagos lavas with the median composition of
gabbroic cumulates from ophiolites and the oceanic
crust [30–32] and reference therein]. Figs. 1, 2 and 3
(see also Supplementary Information Fig. 1 and Table 2)
clearly show that lavas from the eastern group have
intermediate trace element compositions between those
from the western group and the plagioclase-rich
cumulates.
The Sr, Nd, Pb, Hf and He isotope data of lavas from
the Archipelago have been discussed extensively
[1,2,5,6,8,28,33,34]. Our compilation confirms the
conclusions previously reached in those papers. Namely, most of the isotopic variation within the archipelago
occurs between volcanoes/islands, and generates a
geographic distribution of the isotopic variations that
defines a horseshoe pattern [1,2,5,33,34]. Four main
components are needed to bracket the isotopic composition of lavas from the archipelago: a MORB component defined by the Genovesa lavas, two enriched
components defined by Floreana and Pinta basalts, and
the high-3He/4He component characterized by Fernandina lavas. Two main observations arise from dividing
the Archipelago into the western and eastern groups: 1)
Lavas from the eastern group tend to have higher
87
Sr/86Sr at similar 143Nd/144Nd than those from the
western group. To illustrate this observation Fig. 4
shows a third-order polynomial fit to the data for the
western group lavas on the 143Nd/144Nd–87Sr/86Sr
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
399
Fig. 4. 143Nd/144Nd versus 87Sr/86Sr. The analyses used are a compilation of all the published data available through the GEOROC database
complemented with the data presented in this study. The curve is the fitted polymomial of the third order to the data of the western group lavas. ΔSr is
the deviation in 87Sr/86Sr of each sample to that curve at equal 143Nd/144Nd. ΔSr = {87Sr/86Sr − [2.969 − 4.4170 × Nd + 8327.8535 × (Nd − 0.51299)2 +
51,322,354 × (Nd − 0.51299)3]} × 10,000. The inset shows a plot of the ΔSr versus Sr/Sr⁎. We used the median for Sr/Sr⁎, ΔSr from each island/
volcano. Error bars represent 2σ standard errors; we did not report the error bars when the number of samples analyzed per island/volcano was ≤3.
Samples with no reported major element compositions were not considered in the calculation of the median to avoid the effect of crystal fractionation.
Sr/Sr⁎ = [Sr / ((Pr + Nd) / 2)] using normalized values. Normalizing values are from McDonough and Sun [55]. Symbols as in Fig. 1.
diagram; the deviation from this curve toward higher
87
Sr/86Sr at a given 143 Nd/144Nd for each sample is
referred to as ΔSr. The ΔSr is not only higher in lavas
from the eastern group, but also correlates with the Sr/
Sr⁎ ratios (Fig. 4 inset). 2) 3He/4He values are higher in
lavas from the western group, reaching a maximum in
basalts from Fernandina Island. Also, 3He/4He ratios
define a positive correlation with Ti/Gd and Nb/La
ratios, suggesting that high Ti/Gd and Nb/La ratios are
typical of the mantle source carrying the high-3He/4He
signature (see [2] and Fig. 5).
5. Discussion
5.1. Compositional variation across the Archipelago;
plume, entrained MORB and lithospheric signatures
Geological, geochemical and geophysical information point to the existence of an upwelling thermal plume
beneath the Galapagos Archipelago. The Rayleigh wave
tomographic study of Villagomez et al. [14] shows a
continuous low-velocity volume beneath the western
volcanoes (Fernandina and Isabella Islands), which
flattens against the base of the “high-velocity lid”, tilts
in a northerly direction and spreads eastward and westward. These results are consistent with the anomalously
thin mantle transition zone observed beneath the western
volcanoes [12], the age progression of the oldest lavas
dated from each volcano, from young in the west to old
in the east [1,2,21], and with the high eruption frequency
and the high 3He/4He ratios measured for Fernandina
basalts [2,28].
The interaction between plume melts and MORB
most likely controls the isotopic composition of the lavas
from both, the western and eastern, regions ([1,2,5]).
Fluid dynamic models have shown that depleted upper
mantle can be entrained into the plume flow either
by significant lateral entrainment of ambient mantle into
the plume conduit, or as an ambient mantle viscously
coupled to the plume flow [35]. However, even though
the lavas from both regions share similar range in Nd, Hf
and Pb isotopes, their trace element compositions are
significantly different. Such differences in the basalt
compositions are characterized by a decrease in the Nb/
Nb⁎ and Ti/Gd ratios, and an increase in the Ba/Th, K/
La, Sr/Sr⁎, Eu/Eu⁎ and Pb/Nd ratios from west to east
(Fig. 3 and Supplementary Information Fig. 1 and Table
2). For example, when lavas from Wolf and Darwin
volcanoes (western group) are compared with lavas from
Santa Cruz and San Cristobal Islands (eastern group),
both suite of basalts have similar Nd, Hf and Pb isotopic
compositions (e.g., similar Nd isotopes in Fig. 4), but
400
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
Fig. 5. 3He/4He versus Ti/Gd, and Nb/Nb ⁎ (La) for Galapagos lavas.
Data from Kurz and Geist [2], Kurz unpublished data, and this study.
Symbols as in Fig. 1.
lavas from Santa Cruz and San Cristobal have a
“plagioclase-rich gabbro” signature in their trace element
composition and high ΔSr that are absent from Wolf
and Darwin's basalts (Figs. 3, 4). Similar observations
can be made by comparing lavas from Sierra Negra and
Cerro Azul volcanoes (western group), with lavas from
Floreana Island (eastern group).
The 3He/4He ratios and the ΔSr represent the main
isotopic difference between lavas from the western and
eastern groups. 3He/4He correlates with Ti/Gd and Nb/La
ratios (Fig. 5), and all three ratios decrease with increasing
distance from Fernandina. These correlations suggest that
the high Ti/Gd and Nb/La ratios are typical of the plume
mantle source that carries the high 3He/4He signature. In
contrast, the ΔSr is characteristic of lavas with a
“plagioclase-rich gabbro” signature. This is best exemplified by the correlation of Sr/Sr⁎ with ΔSr in basalts
from the eastern group (Fig. 4 inset). Therefore, the
process responsible for the generation of the “plagioclaserich gabbro” signature affected both, the trace element and
the Sr isotopic composition of the lavas.
The variations in Nb/Nb⁎, Ti/Gd, Ba/Th, K/La, Sr/
⁎
Sr , Eu/Eu⁎ and Pb/Nd ratios between lavas from the
eastern and western groups have to be controlled by
either variations in the mantle source compositions or
some type of interaction process within the lithosphere.
The similar incompatibility of the trace elements used
in each ratio precludes significant variations of these
ratios during mantle melting. Furthermore, plagioclase
accumulation during magma differentiation could produce basalts with trace element ratios similar to those
observed in lavas from the eastern group. However,
inspection of thin sections from samples with “plagioclase-rich gabbro” signature indicates that those samples
have less than 1% plagioclase phenocrysts. Therefore,
plagioclase accumulation is not the process responsible
for the unusual trace element ratios observed in many
basalts from the eastern group. In contrast, plagioclase
fractionation during basalt differentiation could erase
the “plagioclase-rich gabbro” signature, making it difficult to determine whether originally a lava had such a
signature or not (see an example in Fig. 2 of the
Supplementary Information). Basalts from the northern
area of the archipelago (e.g., lavas from Pinta,
Marchena, Wolf and Darwin Islands) do not show the
“plagioclase-rich gabbro” signature in their trace
element compositions; because most of these lavas are
quite evolved, it is not clear whether the lack of such a
signature is an original characteristic of these lavas, or is
related to plagioclase fractionation.
5.2. The origin of the “plagioclase-rich gabbro”
signature
Two main hypotheses have been proposed to explain the “plagioclase-rich gabbro” signature in oceanic island basalts: 1) the presence of an ancient recycled
plagioclase-rich cumulate (now eclogite), as a component intrinsic to the mantle plume [36–38], and 2) the
interaction of basalts with plagioclase-rich cumulate
during melt percolation through the oceanic lithosphere
[39–41].
The “plagioclase-rich gabbro” signature in the Galapagos lavas cannot be explained by melts generated from
either an entrained upper mantle component, or an ancient
recycled plagioclase-rich gabbro component within the
plume. Melting of the upper mantle will produce basalts
similar to MORB, which lack the “plagioclase-rich
gabbro” signature. Meanwhile, melts generated from an
ancient recycled plagioclase-rich cumulate (now eclogite)
within the plume will be characterized by low 206Pb/204Pb
compared to those of present-day MORB. However, the
Galapagos lavas with a “plagioclase-rich gabbro” signature have 206Pb/204Pb that is either very similar to, or
much higher (as in the case of Floreana lavas) than those
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
of present-day MORB. Thus, the only alternative explanation for the origin of the “plagioclase-rich gabbro”
signature in the Galapagos lavas is the interaction of
basalts with plagioclase-rich cumulates during melt
percolation through the oceanic lithosphere.
There seems to be a geographic distribution of lavas
with a “plagioclase-rich gabbro” signature within the
archipelago. Most of these lavas are located to the east
of the 91° fracture zone, within the central-eastern and
southeastern sectors of the archipelago. Three main
factors seem to be associated with the geographic distribution of this group of lavas: 1) The composition of the
lavas, 2) the variation in the magmatic flux passing
through the lithosphere, and 3) the variation of the
lithospheric thickness. First, the low incompatible trace
element contents of the magmas generated from the
depleted mantle source erupting in the central-eastern
and southeastern sectors [1,5] make the lavas more
susceptible to modification during melt-cumulate interaction than incompatible trace element enriched melts
generated from the mantle plume. Second, the centraleastern and southeastern sectors are characterized by a
low present-day magmatic flux through the lithosphere,
as indicated by their low eruption frequency, which
decreases with increasing distance from Fernandina and
Isabela Islands [2,3,28]. Finally, seismic information
indicates that the “high-velocity lid” to the east of the
91° W fracture zone thickens from north to south with
increasing distance from the GSC [14]. The combination of a thicker lithosphere and low magmatic flux will
effectively reduce the melt to rock ratio creating an
environment conducive to the chemical modification of
the lava during transport through the lithosphere.
Therefore, the increase in the abundance of lavas with
depleted trace element compositions combined with the
thickening of the lithosphere and the decreasing magma
flux would result in basalts with a “plagioclase-rich
gabbro” signature within the central-eastern and southeastern regions of the archipelago.
Most of the plagioclase-rich cumulates beneath the
central-eastern region of the archipelago originated from
the crystallization of melts probably at the GSC during
the time the Galapagos plume was centered or close to
the ridge. Studies of ophiolites suggest that the crust–
mantle transition zone beneath ocean ridges is one of the
possible areas where such cumulates may be stored.
This zone, located between the underlying residual
peridotite and the dominantly gabbroic crustal section, is
a diffuse area formed by successive stages of basaltic
melt injections that extends sideways from the ridge axis
for tens of kilometers. The composition of the gabbroic
rocks in this zone indicates that they are cumulates
401
formed by partial crystallization of basalts, and demonstrates that plagioclase-rich cumulates can form far
below the level of magmatic neutral buoyancy [42–46].
Multichannel seismic studies of the crust–mantle
transition zone beneath mid-ocean ridges show that
this zone is characterized by low seismic velocities,
suggesting the presence of low melt fractions (a maximum of a few percent), and inefficient heat removal
[47–49]. The interaction of the Galapagos plume with
the GSC would certainly enhance and probably extend
the crust–mantle transition zone between the archipelago and the ridge.
Direct assimilation or partial melting of plagioclaserich cumulates followed by mixing with plume melts
cannot account for the observed major and trace element
composition of the lavas with a “plagioclase-rich gabbro”
signature. A direct assimilation process would require an
unreasonably large proportion of digested cumulate
material to produce the trace element composition of the
lavas from the eastern group (see Supplementary
Information Fig. 3). In addition, the existing experimental
data indicate that melt generated by partial melting of
gabbro have high SiO2 content [[50], and references
therein], but Galapagos lavas with “plagioclase-rich
gabbro” signature have relatively low SiO2 contents.
Therefore, neither assimilation nor partial melting of
gabbro cumulate is the process responsible for the trace
element composition of lavas in the eastern region. As
will be shown below, diffusive interaction between
plagioclase-rich cumulates and percolating basalts can
explain the observed “plagioclase-rich gabbro” signature.
5.3. A melt-cumulate diffusive interaction model
The basic idea behind this model is that cations of
different size and charge diffuse at different rates in
plagioclase, giving rise to kinetic fractions among the
incompatible trace elements (e.g., diffusivities of Sr2+
and Pb2+ are significantly larger than those of REE3+ in
plagioclase; see Supplementary Information Table 3).
As a first step towards testing this hypothesis, we
consider a simple end member case of a troctolitic mush
(plagioclase + olivine + melt) in which the variations in
the incompatible trace element content in the interstitial
percolating melt are due only to diffusive exchange
between the melt and its surrounding minerals. Since
olivine does not significantly contribute to the incompatible trace element budget, only providing a dilution
effect, our model evaluates the diffusive interaction
between a melt and a single solid cumulate phase,
plagioclase. For simplicity, we neglect the effects of
crystal dissolution, precipitation or crystallization that
402
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
may arise during crystal–melt diffusive re-equilibration
in a partially molten system [51,52]. A more general
case considering melt percolation, diffusive exchange,
and dissolution-re-precipitation processes between gabbro (plagioclase + clinopyroxene + olivine) and interstitial melt will be the focus of a future paper.
Since the exact solutions to the problem of diffusive
re-equilibration between a crystal and a well-stirred melt
pocket (plagioclase + melt) in a spherical, cylindrical, or
one-dimensional setup are rather complicated [53], we
use a simple approximate solution to the diffusion
equations in a finite crystal–melt geometry [51]. This
simple solution was obtained by approximating the
concentration gradient at the crystal–melt interface
using boundary layer theory. In the model, the solid is
assumed to be homogeneous in composition at the onset
of melt-cumulate interaction. During the interaction, the
interface between the solid and the melt is assumed to be
in chemical equilibrium at all times; the mineral grains
do not deform, and the concentrations in the interior of
the grains are only controlled by diffusion. Furthermore,
the diffusion in the melt is assumed to be instantaneous.
In the model presented here, the variation of the trace
element concentration in the melt with time, Cf(t), is
given by:
Cf ðtÞ ¼ Cfl þ ðCf0 −Cfl Þexpð−aK tÞ
ð1aÞ
qf /Cf0 þ qs ð/−1ÞCs0
qf / þ qs ð/−1ÞKD
ð1bÞ
3Ds ½qf / þ qs ð/−1ÞKD qf /
r2
ð1cÞ
Cfl ¼
aK ¼
where Cf0 and Cs0 are the initial trace element concentrations in the melt and solid, respectively; Cf∞ is the
equilibrium trace element concentration in the melt; ϕ is
the volume fraction of the melt (porosity); KD is the
solid-melt partition coefficient; r is the radius of the
mineral grain; Ds is the diffusion coefficient in the solid
phase, t is time; and ρs and ρf are the densities of the
solid and the melt, respectively.
The case presented here considers a melt with a composition similar to a primitive MORB interacting with
plagioclase-rich cumulate crystallized from a melt having
relatively higher incompatible trace element contents,
higher Sm/Yb ratios and 87Sr/86Sr ratios. Mixing of
plume and upper mantle melts at either the GSC or at the
leading edge of the plume in the western sector of the
archipelago could easily be responsible for the generation
of the enriched melts that produced the cumulates [54].
We performed a model run using the parameter values
listed in Table 3 of the Supplementary Information. We
assumed a troctolite comprised of an equal proportion of
olivine and plagioclase. Previous studies of the crust–
mantle transition zone constrain the grain size of the
crystal phases to 1–2 mm [45], the melt to rock ratio to
1–2%, and the temperature to approximately 1200 °C
[47]. We further assumed that the temperature remained
constant at 1200 °C, consistent with the inferred inefficient heat removal from the transition zone [47].
Fig. 6 displays the results of this simple diffusive
interaction model, which can easily explain the composition of the erupted lavas; in this specific example the
composition of a basalt from San Cristobal Island is
reproduced. Hence, diffusive interaction between oceanic basalts and plagioclase-rich cumulates can produce
significant changes in the trace element composition of
the percolating lavas. During melt-cumulate interaction,
the Ba/Th, K/La, Sr/Nd, Eu/Eu⁎ and Pb/Nd ratios in the
melt radically diverge from those initially produced
during mantle melting. The trace elements will diffuse
from the cumulate into the melt in an attempt to reach
equilibrium. The larger diffusion coefficients of Ba, K,
Sr, Eu+2, and Pb in plagioclase compared to those of Th,
La Nd, and Gd, will ultimately generate a “plagioclaserich gabbro” signature equivalent to that observed in
lavas from the central-eastern region of the archipelago.
Furthermore, because the depleted percolating melt has
a lower 87Sr/86Sr ratio than the cumulate, the diffusive
interaction process will also increase the 87Sr/86Sr ratios
in the melt at a rate that is significantly faster than that
required to modify the 143 Nd/144Nd ratios, resulting in a
high ΔSr values. Thus, contemporaneous variation in
the ΔSr and the Sr/Sr⁎ during melt-cumulate diffusive
interaction would provide an explanation for the observed ΔSr–Sr/Sr⁎ correlation defined by the Galapagos lavas with “plagioclase-rich gabbro” signature
(Fig. 6B).
As shown in Eqs. (1a)–(1c), the compositions of the
percolating melt and the cumulate, the interaction time,
the grain size, and the melt to cumulate ratio (or porosity) are among the most important factors influencing
the results of our model. Clearly, the larger the chemical
disequilibrium between the percolating melt and the
cumulate, the more susceptible the melt will be to the
diffusive interaction. For example barium, due to its
lower diffusion coefficient in plagioclase compared to
those of K, Sr, and Pb, would require either a longer
interaction time or a larger chemical disequilibrium
between the percolating melt and the cumulate to produce an anomaly similar to those of the other elements.
This process can also explain the observation that most
Galapagos lavas with “plagioclase-rich gabbro”
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
403
Fig. 6. Calculated results of diffusive interaction between a melt with a composition similar to a primitive MORB interacting with plagioclase-rich
cumulate crystallized from a melt with relatively higher incompatible trace element contents, higher Sm/Yb ratios and 87Sr/86Sr ratios. The results are
displayed as A) Primitve mantle-normalized trace elements and B) ΔSr versus Sr/Sr⁎. We compare the results of our model with the actual trace
element composition of a basalt from San Cristobal Island, NSK97-252. The arrow indicates the direction of progressive diffusive interaction with
time. For the example presented here, we assumed a troctolite comprised of an equal proportion of olivine and plagioclase (grain size = 2 mm), melt to
rock ratio of 1%, at 1200 °C. We display five different interaction times 5, 10, 20, 40 and 300 years. The input and output data of the model are shown
in Table 3, Supplementary Information.
signature do not show significant Eu anomalies. The
small positive Eu anomaly in samples with large
positive Sr anomalies would suggest that either Eu is
present mostly as Eu+ 3, which has a lower diffusion
coefficient than Sr (see Table 3 of the Supplementary
Information), or the chemical disequilibrium between
the percolating melt and the cumulate is significantly
smaller for Eu+2 than for Sr. Furthermore, a percolating
melt could develop both a positive Sr anomaly and a
negative Eu anomaly during diffusive interaction with a
plagioclase-rich cumulate. This case requires a percolating melt with lower Sr and higher Eu contents than
those of the melt responsible for the generation of the
cumulate. This would occur when an incompatible trace
element depleted melt, generated in the spinel stability
field, interacts with a plagioclase cumulate crystallized
from an incompatible trace element enriched melt
generated in the garnet stability field. Scenario that is
quite likely for the lavas erupted on the central-eastern
region of the archipelago. Thus, a precise knowledge of
the percolating melt and cumulate compositions, and the
partition and diffusion coefficients would be required to
determine the interaction time between the melt and the
plagioclase cumulate. Even though the interaction time
cannot be precisely defined, it is evident from the results
of our simple model that very short times, on the order
404
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
of a few years, are required to produce the “plagioclaserich gabbro” signature in the percolating basalt.
Another factor affecting the interaction time is the
grain size of the cumulate. Increasing the grain size is
equivalent to reducing the surface area over which the
diffusive interaction takes place. Thus, the result of
increasing grain size will be comparable to decreasing
the melt-cumulate interaction time. Increasing the
plagioclase grain size will require longer interaction
times to achieve the same degree of diffusive exchange.
Furthermore, we have to take into account the effect of
the melt fraction in our model. Our results indicate that
the larger the melt to plagioclase-cumulate ratio, the less
pronounced the effects of diffusive interaction on the
melt. Seismic data constrain the porosity in the crust–
mantle transition zone beneath mid-ocean ridges to, at
most, a couple of percent [47–49]. Therefore, the melt to
cumulate ratio will be small and the effect of diffusive
interaction within the crust–mantle transition zone thus
may have a large influence on the trace element and
isotope ratios of the percolating melt.
The model presented here assumes that the percolating melt is saturated with plagioclase, and therefore there
is only diffusive interaction between the melt and the
cumulate without dissolution or re-crystallization of the
plagioclase. Interaction of plagioclase-undersaturated
melts with plagioclase-rich cumulates would cause the
dissolution of plagioclase from the cumulate. Due to the
compatible behavior of Sr in plagioclase, dissolution of
such mineral would generate an enhanced “plagioclaserich gabbro” signature in the melt, with an even larger
Sr/Sr⁎ anomalies than those produced by simple
diffusive interaction processes. The high Al2O3 content
of the Galapagos lavas with a “plagioclase-rich gabbro”
signature suggests that dissolution of plagioclase occurred during the percolation of the lavas through the
oceanic lithosphere. Therefore, our simple diffusive
interaction model represents a conservative approach to
evaluating the melt-cumulate interaction processes that
affected the composition of the Galapagos basalts.
6. Conclusions
The Galapagos Archipelago can be divided into two
distinct volcanic areas, east and west of the 91° W fracture
zone. We find important differences in the compositions
of the lavas from both regions. First, the Ti/Gd and Nb/La
ratios correlate with 3He/4He, and all three ratios decrease
with increasing distance from Fernandina. These correlations suggest that the high Ti/Gd and Nb/La ratios are
typical of the plume mantle source that carries the high
3
He/ 4 He signature. Second, there is an increase in the
Ba/Th, K/La, Sr/Nd, Eu/Eu⁎, Pb/Nd ratios, and ΔSr in
lavas erupted from west to east across the fracture zone,
and from north to south within the eastern region. The
high Ba/Th, K/La, Sr/Nd, Eu/Eu⁎, and Pb/Nd ratios in
the Galapagos basalts from the eastern region resemble
the composition of plagioclase-rich cumulates from the
oceanic crust and ophiolite complexes. The compositions
of these lavas can be explained by interaction with
plagioclase-rich cumulate during melt percolation
through the oceanic lithosphere. We argue that diffusive
interaction of basalts with plagioclase-rich cumulates,
formed either in the Galapagos Spreading Center or
beneath the leading edge of the plume on the western
region, are responsible for the observed trace element
compositions of lavas in the eastern region.
Acknowledgements
We thank M. Perfit, W. White and J. Sinton for their
thoughtful reviews that considerably improved the
manuscript. We are grateful to the Charles Darwin
Research Station for the logistical support during the
field trips to the Galapagos Islands. This work has
benefited greatly from discussions with E. Hauri, J. Van
Orman and E. Drenkard. The support for this work was
provided by the US National Science Foundation.
Appendix A. Supplementary data
Supplementary data associated with this article can
be found, in the online version, at doi:10.1016/j.
epsl.2007.02.040.
References
[1] W.M. White, A.R. McBirney, R.A. Duncan, Petrology and
geochemistry of the Galapagos Islands: portrait of a pathological
mantle plume, J. Geophys. Res. 98 (1993) 19533–19563.
[2] M.D. Kurz, D.J. Geist, Dynamics of the Galapagos hotspot from
helium isotope geochemistry, Geochim. Cosmochim. Acta 63
(23–24) (1999) 4139–4156.
[3] D.J. Geist, T. Naumann, P. Larson, Evolution of Galapagos
magmas: mantle and crustal fractionation without assimilation,
J. Petrol. 39 (5) (1998) 953–971.
[4] K. Hoernle, R. Werner, J. Phipps Morgan, D. Garbe-Schönberg,
J. Bryce, J. Mrazek, Existence of complex spatial zonation in the
Galapagos plume for at least 14 Ma, Geology 28 (5) (2000)
435–438.
[5] K.S. Harpp, W.M. White, Tracing a mantle plume: isotopic and
trace element variations of Galapagos seamounts, Geochem.
Geophys. Geosyst. 2 (2001) (paper number 2000GC000137).
[6] J.-G. Schilling, D. Fontignie, J. Blichert-Toft, R.H. Kingsley, U.
Tomza, Pb–Hf–Nd–Sr isotope variations along the Galapagos
Spreading Center (101–83 W): constraints on the dispersal of the
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
Galapagos mantle plume, Geochem. Geophys. Geosyst. 4 (2003)
(paper number 2002GC000495).
D.M. Christie, R.A. Duncan, A.R. McBirney, M.A. Richards, W.M.
White, K.S. Harpp, C.G. Fox, Drowned islands downstream from the
Galapagos hotspot imply extended speciation times, Nature 355
(1992) 246–248.
J. Blichert-Toft, W.M. White, Hf isotope geochemistry of the
Galapagos islands, Geochem. Geophys. Geosyst. 2 (2001) (paper
number 2000GC000138).
M.A. Feighner, M.A. Richards, Lithospheric structure and compensation mechanism of the Galapagos Archipelago, J. Geophys.
Res. 99 (1994) 6711–6729.
J.P. Canales, G. Ito, R. Detrick, J. Sinton, Crustal thickness along
the western Galapagos Spreading Center and the compensation
of the Galapagos hotspot swell, Earth Planet. Sci. Lett. 203
(2002) 311–327.
R. Montelli, G. Nolet, F.A. Dahlen, G. Masters, E.R. Engdahl, S.-H.
Hung, Finite-frequency tomography reveals a variety of plumes in
the Mantle, Science 303 (2004) 338–343.
E.E. Hooft, D.R. Toomey, S.C. Solomon, Anomalously thin
transition zone beneath the Galapagos hotspot, Earth Planet. Sci.
Lett. 216 (2003) 55–64.
R. Werner, K. Hoernle, U. Barckhausen, F. Hauff, Geodynamic
evolution of the Galapagos hot spot system (Central East Pacific)
over the past 20 m.y.: constraints from morphology, geochemistry, and magnetic anomalies, Geochem. Geophys. Geosyst. 4
(2003) (paper number 2003GC000576).
D.R. Villagomez, D.R. Toomey, E.E. Hooft and S.C. Solomon,
Upper mantle structure beneath the Galapagos Archipelago from
surface wave tomography, J. Geophys. Res. (submitted for
publication).
M.A. Feighner, M.A. Richards, The fluid dynamics of plumeridge and plume–plate interactions: an experimental investigation, Earth Planet. Sci. Lett. 129 (1995) 171–182.
G. Ito, J. Lin, C.W. Gable, Interaction of mantle plume and
migrating mid-ocean ridges: implications for the Galapagos
plume-ridge system, J. Geophys. Res. 102 (1997) 15403–15417.
M.G. Braun, R.A. Sohn, Melt migration in plume-ridge systems,
Earth Planet. Sci. Lett. 213 (2003) 417–430.
P.S. Hall, C. Kincaid, Melting, dehydration, and the dynamics of
off-axis plume–ridge interaction, Geochem. Geophys. Geosyst.
2 (2003) (paper number 2000GC000137).
M.D. Kurz, D.J. Fornari, J.M. Curtis, M. Perfit, D. Scheirer, D.
Geist, The leading age of the Galapagos hotspot: new submarine
evidence from Fernandina volcano, Eos, Trans. - Am. Geophys.
Union 81 (48) (2000) 1282–1283.
D.S. Wilson, R.N. Hey, History of rift propagation and
magnetization intensity for the Cocos–Nazca spreading center,
J. Geophys. Res. 100 (1995) 10041–10056.
C.W. Sinton, D.M. Christie, R.A. Duncan, Geochronology of
Galapagos seamounts, J. Geophys. Res. 101 (1996) 13689–13700.
U. Barckhausen, C.R. Ranero, R. von Huene, S.C. Cande, H.A.
Roeser, Revised tectonic boundaries in the Cocos Plate off Costa
Rica: implications for the segmentation of the convergent margin and
for plate tectonic models, J. Geophys. Res. 19 (2001) 19207–19220.
T. Simkin, L. Siebert, Volcanoes of the World, Geoscience Press,
Tucson, Arizona, 1994.
D.J. Geist, An appraisal of melting processes and the Galapagos
hotspot: major and trace-element evidence, J. Volcanol.
Geotherm. Res. 52 (1992) 65–82.
M.A. Richard, R.W. Griffiths, Thermal entrainment by deflected
mantle plumes, Nature 342 (1989) 900–902.
405
[26] D. Bercovici, J. Lin, A gravity current model of cooling mantle
plume heads with temperature-dependent buoyancy and viscosity, J. Geophys. Res. 101 (1996) 3291–3309.
[27] M.D. Kurz, D. Colodner, T.W. Trull, R.B. Moore, K. O'Brien,
Cosmic ray exposure dating with in situ produced cosmogenic
3He: results from young Hawaiian lava flows, Earth Planet. Sci.
Lett. 97 (1990) 177–189.
[28] M.D. Kurz, S.K. Rowland, J. Curtice, A.E. Saal and T. Nauman,
Eruptions rates at Fernandina volcano, Galapagos Archipelago,
from cosmogenic helium, Submitted to Geology, now under
review, 2007.
[29] M.D. Kurz, T. Kenna, D. Kammer, J.M. Rhodes, M.O. Garcia,
Isotopic Evolution of Mauna Loa Volcano: a View from the
Submarine South West Rift, , 1995 289–306 pp.
[30] C.J. Lissenberg, J.H. Bedard, C.R. van Staal, The structure and
geochemistry of the gabbro zone of the Annieopsquotch
ophiolite, Newfoundland: implications for lower crustal accretion
at spreading ridges, Earth Planet. Sci. Lett. 229 (2004) 105–123.
[31] S.R. Hart, J. Blusztajn, H.J.B. Dick, P.S. Meyer, K. Muehlenbachs, The fingerprint of seawater circulation in a 500-meter
section of ocean crust gabbros, Geochim. Cosmochim. Acta 63
(23–24) (1999) 4059–4080.
[32] C.J. Garrido, P.B. Kelemen, G. Hirth, Variation of cooling rate
with depth in lower crust formed at an oceanic spreading ridge:
plagioclase crystal size distributions in gabbros from the Oman
ophiolite, Geochem. Geophys. Geosyst. 2 (2001) (Art. No.
2000GC000136).
[33] D.J. Geist, W.M. White, A.R. McBirney, Plume-asthenosphere
mixing beneath the Galapagos Archipelago, Nature 333 (657–660)
(1988).
[34] D.W. Graham, D.M. Christie, K.S. Harpp, J.E. Lupton, Mantle
plume helium in submarine basalts from the Galapagos Platform,
Science 262 (1993) 2023–2026.
[35] E.H. Hauri, J.A. Whitehead, S.R. Hart, Fluid dynamic and
geochemical aspects of entrainment in mantle plumes, J. Geophys.
Res. 99 (1994) 24275–24300.
[36] C. Chauvel, C. Hemond, Melting of a complete section of
recycled oceanic crust: trace element and Pb isotopic evidence
from Iceland, Geochem. Geophys. Geosyst. 1 (2000) (Paper
number 1999GC000002).
[37] A.W. Hofmann, K.P. Jochum, Source characteristics derived
from very incompatible trace elements in Mauna Loa and Mauna
Kea basalts, Hawaiian Scientific Drilling Project, J. Geophys.
Res. 101 (1996) 11831–11839.
[38] A.V. Sobolev, A.W. Hofmann, I.K. Nikogosian, Recycled
oceanic crust observed in ‘ghost plagioclase’ within the source
of Mauna Loa lavas, Nature 404 (2000) 986–990.
[39] L.V. Danyushevsky, R.A. Leslie, A.J. Crawford, P. Durancei,
Melt inclusions in primitive olivine phenocrysts: the role of
localized reaction processes in the origin of anomalous
compositions, J. Petrol. 45 (2004) 2531–2553.
[40] H.-J. Yang, F.A. Frey, D. Weiss, A. Giret, D. Pyle, G. Michon,
Petrogenesis of the flood basalts forming the northern Kerguelen
Archipelago: implications for the Kerguelen plume, J. Petrol. 39
(4) (1998) 711–748.
[41] A.A. Gurenko, A.V. Sobolev, Crust–primitive magma interaction
beneath neovolcanic rift zone of Iceland recorded in gabbro
xenoliths from Midfell, SW Iceland, Contrib. Mineral. Petrol.
151 (2006) 495–520.
[42] A. Nicolas, A. Prinzhofer, Cumulative or residual origin for the
transition zone in ophiolites: structural evidence, J. Petrol. 24
(1983) 188–206.
406
A.E. Saal et al. / Earth and Planetary Science Letters 257 (2007) 391–406
[43] A. Nicolas, I. Reuber, K. Benn, A new magma chamber model
based on structural studies in the Oman ophiolite, Tectonophysics
151 (1988) 87–105.
[44] K. Benn, A. Nicolas, I. Reuber, Mantle–crust transition zone and
origin of wehrlitic magmas: evidence from the Oman ophiolite,
Tectonophysics 151 (1988) 75–85.
[45] F. Boudier, A. Nicolas, I. Benoît, Magma chambers in the Oman
ophiolite: fed from the top and the bottom, Earth Planet. Sci. Lett.
144 (1996) 239–250.
[46] P.B. Kelemen, K. Koga, N. Shimizu, Geochemistry of gabbro
sills in the crust–mantle transition zone of the Oman ophiolite:
implications for the origin of the oceanic lower crust, Earth
Planet. Sci. Lett. 146 (1997) 475–488.
[47] R.A. Dunn, D.R. Toomey, S.C. Solomon, Three-dimensional
seismic structure and physical properties of the crust and shallow
mantle beneath the East Pacific Rise at 9°30′ N, J. Geophys. Res.
105 (2000) 23,537–23,555.
[48] R.A. Dunn, D.R. Toomey, R.S. Detrick, W.S. Wilcock,
Continuous mantle melt supply beneath an overlapping spreading
center on the East Pacific Rise, Science 291 (2001) 1955–1958.
[49] R.A. Dunn, D.W. Forsyth, Imaging the transition between the
region of mantle melt generation and the crustal magma chamber
beneath the southern East Pacific Rise with short-period Love
waves, J. Geophys. Res. 108 (B7) (2003), doi:10.1029/
2002JB002217 (Article # 2352).
[50] J. Koepke, S.T. Feig, J. Snow, M. Freise, Petrogenesis of oceanic
plagiogranites by partial melting of gabbros: an experimental
study, Contrib. Mineral. Petrol. 146 (2004) 414–432.
[51] S.R. Hart, Equilibration during mantle melting: a fractal tree
model, Proc. Natl. Acad. Sci. 90 (24) (1993) 11914–11918.
[52] Y. Liang, Kinetics of crystal–melt reaction in partially molten
silicates. 1. Grain scale processes, Geochem. Geophys. Geosyst.
4 (2003), doi:10.1029/2002GC000375.
[53] J. Crank, The Mathematics of Diffusion, Clarendon Press,
Oxford, 1975.
[54] J.-G. Schilling, D. Fontignie, J. Blichert-Toft, R. Kingsley, U.
Tomza, Pb–Hf–Nd–Sr isotope variations along the Galapagos
Spreading Center (101–83 W): constraints on the dispersal of the
Galapagos mantle plume, Geochem. Geophys. Geosyst. 4
(2003), doi:10.1029/ 2002GC000495.
[55] W.F. McDonough, S.-S. Sun, The composition of the Earth,
Chem. Geol. 120 (1995) 223–253.
Supplementary Information
Sampling and Analytical Techniques
The data considered in this study include:
1) The compilation of all published major, trace elements and isotopic
compositions reported for lavas with MgO ≥ 5 wt% from the Galapagos archipelago
(http://georoc.mpch-mainz.gwdg.de/georoc/) is complicated by the lack of interlaboratory calibrations and differences in the analytical techniques. We therefore were as
careful as possible in evaluating the data produced by different analytical techniques and
different laboratories. Two main discrepancies should be mentioned here. The Nb
contents (XRF) for Marchena lavas reported by Vicenzi et al. [1] are higher than those
determined in this study by ICPMS for comparable samples (see Table 1), and the Y
contents (ICP-MS) of Volcan Ecuador lavas reported by Geist et al. [2] are higher than
the ICP-MS values we obtained for comparable samples from the same volcano. The
reasons for these discrepancies are unknown, but it may be due to either one of the
factors mentioned above or simply reflect sample heterogeneity. Because the data by
Vicenzi et al. [1] and Geist et al. [2] represent the majority of the data for Marchena and
Volcan Ecuador respectively, we included the data as reported by these authors and the
discrepancies with respect to our own data set are expressed in the standard errors (2σ)
reported for the median in each island/volcano. When the major element composition for
a particular sample was not reported, the trace elements were excluded from the
calculation of the median because the effect of fractional crystallization/accumulation
could not be evaluated.
2) The new data in this study followed the strategy of sampling basalts that define
the observed geographical so-called “horse shoe” pattern in isotopes and incompatible
element ratios for the Galapagos archipelago [3-5]; in our case the selection of samples
was constrained primarily by the age data available for the lavas, selecting the youngest
lavas in each island/volcano. The sampling was simplified by the fact that the observed
isotopic variations in the Galapagos lavas occur either between islands or between
volcanoes on the same island; the intra-volcano compositional variations are generally
smaller than the inter-volcano variations. We selected samples from the volcanoes A)
1
Fernandina, Cerro Azul, Wolf, Sierra Negra, Darwin, and Ecuador, which represent the
large shield volcanoes erupted on the older and thicker lithosphere to the west of the 91o
W fracture zone; and B) San Cristobal, Santa Cruz, Santiago, Floreana, Pinta, and
Marchena representing lavas erupted far from Fernandina on a thinner and younger
lithosphere, east of the fracture zone. Many of the samples are from well-recorded
historic eruptions, such as those from Fernandina and Marchena. With the exception of
these, most of the samples have been dated by surface exposure dating techniques using
cosmogenic 3He [6-8].
The rock samples were rough-crushed in plastic, and the cleanest possible chips
were handpicked and powdered in agate. Major and trace element data reported here were
done by XRF and ICP-MS at the GeoAnalytical Lab, Washington State University. The
precision for XRF data (major and trace elements) for seven repeat analyses of BCR-P
was < 1% (2σ) for all major elements, except MnO, K2O, Na2O and P2O5 with <3%
(2σ). The precision for XRF analyses for trace elements was <5% (2σ) for most
elements, <10% (2σ) for Rb, Ga, Cu and Zn, <20% (2σ) for Sc, and <30% (2σ) for Ba.
The precision for ICP-MS data on a single sample BCR-P run over a four month period
(n = 24) was <5% (2σ) for most trace elements, <10% (2σ) for Cs, Rb, Pb and <20%
(2σ) for Th and U.
The analytical techniques and uncertainties for Sr, Nd, Pb and He isotope
measurements have been previously reported by Kurz et al., [9-10] and Kurz and Geist,
[6]. Powders for Sr, Nd and Pb isotopic analyses were leached in warm 6N HCl for
1hour; Sr and Nd separation chemistry was done with conventional ion chromatography,
using DOWEX 50 cation exchange resin for Sr and REE, followed by HDEHP-coated
Teflon beads for Nd purification. Pb separation was undertaken with HBr–HNO3 on
AG1X8 anion exchange resin [11]. Isotopic measurements for all three elements were
carried out by thermal ionization mass spectrometry (TIMS) using the VG-354 multicollector instrument at the Woods Hole Oceanographic Institution. The 143Nd/144Nd ratios
were fractionation-corrected to
146
Nd/144Nd ratio of 0.7219 using a power law and
normalized to the La Jolla value of 0.511847. The
corrected to
86
87
Sr/86Sr ratios were fractionated-
Sr/88Sr of 0.1194 using a power law and normalized to the NBS-987
2
standard value of 0.71024. Some of the samples seem to have unusually high 87Sr/86Sr at
a given
143
Nd/144Nd. Different aliquots of the samples leached several times in 6.N hot
HCl for a period of 1 hour each time produced overlapping values within the 2σ error.
The Pb isotope ratios were fractionation-corrected to the reported values for the NBS-981
standard of
206
Pb/204Pb = 16.9356, 207Pb/204Pb = 15.4891, and 208Pb/204Pb = 36.7006 [12]
using linear law. External reproducibility of the standard are 16 ppm (2σ, n = 35) for
143
Nd/144Nd of the La Jolla standard, 26 ppm (2σ, n = 35) for
87
Sr/86Sr of the NBS-987
standard and 0.05% per amu (2σ, n = 60) for the Pb isotope ratios of NBS-981 standard.
The Sr, Nd and Pb total procedure blanks typically were less than 40pg. 70 pg and 150 pg
respectively, which are negligible for all the measured isotope ratios.
Separation of Hf and subsequent mass spectrometric analysis using the multicollector inductively-coupled plasma mass spectrometer (MC-ICP-MS) VG Plasma 54
were undertaken at the Ecole Normale Supérieure in Lyon following the procedures
described by Blichert-Toft et al. [13]. In order to monitor machine performance, the
JMC-475 Hf standard was run systematically after every two samples and gave 0.282160
± 0.000010 (2 sigma) for
176
Hf/177Hf during the single run session in which the samples
reported here were analyzed, corresponding to an external reproducibility of 30 ppm. The
internal run precision was half that or better for all the samples.
normalized for mass fractionation relative to
179
176
Hf/177Hf was
Hf/177Hf = 0.7325 using an exponential
law. The total procedural Hf blank was less than 25 pg and thus negligible.
Table Caption
Table 1: Major, trace element, and isotopic compositions of Galapagos lavas. The dataset
reported in Table 1 is partially from Kurz and Geist, [2], complemented with some new
unpublished data from M. Kurz, A. Saal and J. Blichert-Toft. Most samples are younger
than 50Ka, with the exception of three samples from Floreana (100-200 Ka) and two
samples from Santa Cruz (100 Ka and 500Ka). See section on analytical techniques.
Table 2: Correlation matrix of median values of specific trace element ratios in each
Volcano. Values are based on the compilation of all published major, trace elements and
isotopic compositions reported for lavas with MgO ≥ 5 wt% from the Galapagos
Archipelago. Samples with no reported major element compositions were not considered
3
in the calculation of the median to avoid the effect of crystal fractionation. TiO28 content
represent the TiO2 content normalized to 8 wt% MgO [14]. Nb/Nb*(La) =
[Nb/((Ba+La)/2)] and Sr/Sr* = [Sr/((Pr+Nd)/2)] using normalized values. Normalizing
values are from McDonough and Sun [15]. Note the significant correlation between the
trace element ratios indicative of the “plagioclase-rich gabbro” signature and with the
ΔSr.
Table 3: Input and output data for model runs of diffusive interaction between basalt and
a plagioclase-rich cumulate. The composition used for the initial percolating melt is a
representative value for a primitive MORB [16]. The trace element composition of the
melt that generated the plagioclase-rich cumulate has been assumed to be more enriched
than the percolating melt. We used experimentally determined partition coefficients [1718], and diffusion coefficient [19-25] for plagioclase at 1200 oC. To obtain diffusion
coefficients that have not been determined experimentally, we made estimates based on
cation size and charge considerations [26]. We assumed that the diffusion coefficients for
Nb, Ta, Zr, Hf, Ti are similar to those of Th and U. We also assumed that Eu is 50%
Eu2+, and assume a diffusion coefficient for Eu2+ similar to that of Sr.
Figure Caption
Figure 1: Nb/Nb*(La), Ti/Gd, Ba/Th, K/La and Sr/Sr* versus TiO28. We plot the median
for each trace element ratio and TiO28 from each island/volcano and for gabbros from
Annieopsquotch [27], Oman [28 and reference therein] and the oceanic crust [29]. Note
that for the oceanic crust we plot the median values from the “protholith” and “strips” as
defined by Hart et al., [29]. Error bars represent 2σ standard errors; we did not report the
error bars when the number of samples analyzed per island/volcano was ≤ 3. Samples
with no reported major element compositions were not considered in the calculation of
the median to avoid the effect of crystal fractionation. TiO28 content represent the TiO2
content normalized to 8 wt% MgO [14]. Nb/Nb*(La) = [Nb/((Ba+La)/2)] and Sr/Sr* =
[Sr/((Pr+Nd)/2)] using normalized values. Normalizing values are from McDonough and
Sun [15]. Symbols as in Figure 1 of the manuscript.
Figure 2: Primitive mantle-normalized trace element diagram of lavas with different
extent of differentiation from Cerro Azul volcano. Note that the Sr/Sr* in the lavas varies
4
from positive to negative with increasing extent of differentiation (decreasing MgO
content). This observation suggest that plagioclase fractionation during magma
differentiation could have erased the “plagioclase-rich gabbro” signature from primitive
basalts, making it difficult to determine whether differentiated melts originally had or not
such signature. Normalizing values are from McDonough and Sun, [15].
Figure 3: Sr/Sr* versus Nb/Nb*(La) for individual Galapagos lavas reported in Table 1.
We also plot the median for cumulate gabbros from Annieopsquotch [27], Oman [28 and
reference therein] and the oceanic crust [29]. Blue dashed lines represent bulk mixing
trends between representative lava from the western region of the Galapagos archipelago
and the median value for the reported cumulate gabbros. Tick marks represent the percent
of gabbroic component in the mixture. Direct assimilation process would require an
unreasonable large proportion of digested cumulate material in order to produce the trace
element composition of the lavas from the eastern group. Nb/Nb*(La) =
[Nb/((Ba+La)/2)] and Sr/Sr* = [Sr/((Pr+Nd)/2)] using normalized values. Normalizing
values are from McDonough and Sun [15]. Symbols as in Figure 1 of the manuscript.
References
[1] E.P. Vicenzi, A.R. McBirney, W.M. White and M. Hamilton, The geology and
geochemistry of Isla Marchena, Galapagos Archipelago: an ocean island adjacent to a
mid-ocean ridge., Journal of Volcanology and of Geothermal Research 40, 291-315,
1990.
[2] D. Geist, W.M. White, F. Albarede, K. Harpp, R. Reynolds, J. Blichert-Toft and M.D.
Kurz, Volcanic evolution in the Galapagos: The dissected shield of Volcan Ecuador.,
Geochemistry, Geophysics, Geosystems 3(paper number 2002GC000355), 2002.
[3] D.J. Geist, W.M. White and A.R. McBirney, Plume-asthenosphere mixing beneath
the Galapagos archipelago, Nature 333(657-660), 1988.
[4] W.M. White, A.R. McBirney and R.A. Duncan, Petrology and geochemistry of the
Galapagos Islands: Portrait of a pathological mantle plume., Journal of Geophysical
Research 98, 19533-19563, 1993.
[5] D.J. Geist, An appraisal of melting processes and the Galapagos hotspot: major and
trace-element evidence, Journal of Volcanology and Geothermal Research 52, 65-82,
1992.
[6] M.D. Kurz and D.J. Geist, Dynamics of the Galapagos hotspot from helium isotope
geochemistry, Geochimica et Cosmochimica Acta 63(23-24), 4139-4156, 1999.
5
[7] M.D. Kurz, D. Colodner, T.W. Trull, R.B. Moore and K. O’Brien, Cosmic ray
exposure dating with in situ produced cosmogenic 3He: Results from young Hawaiian
lava flows., Earth Planetary Science Letters 97, 177-189, 1990.
[8] M.D. Kurz, S.K. Rowland, J. Curtice, A.E. Saal and T. Nauman, Eruptions rates at
Fernandina volcano, Galapagos archipelago, from cosmogenic helium, Submitted to
Geology, now under review 2007.
[9] M.D. Kurz, T. Kenna, D. Kammer, J.M. Rhodes and M.O. Garcia, Isotopic evolutin
of Mauna Loa volcano: a view from the submarine south west rift., 289-306 pp., 1995.
[10] M.D. Kurz, T. Kenna, J.C. Lassiter and D.J. DePaolo, Helium isotopic evolution of
Mauna Kea volcano: First results from the 1 km drill core., Journal of Geophysical
Research 101, 11781-11791, 1996.
[11] G. Manhes, J. Minster and C.J. Allegre, Comparative uranium-thorium-lead and
rubidium-strontium of St. Severin amphoterite: consequences for early solar system
chronology., Earth Planetary Science Letters 39, 14-24, 1978.
[12] W. Todt, R.A. Cliff, A. Hanser and A.W. Hofmann, Evaluation of a 202Pb-205Pb
double spike for high-precision lead isotope analysis, 429-437 pp., 1996.
[13] J. Blichert-Toft, C. Chauvel, and F. Albarède,. Separation of Hf and Lu for highprecision isotope analysis of rock samples by magnetic sector-multiple collector ICP-MS.
Contribution to Mineralogy and Petrology 127, 248-260, 1997.
[14] E.M. Klein and C.H. Langmuir, Ocean ridge basalt chemistry, axial depth, crustal
thickness and temperature variations in the mantle, Journal of Geophysical Research 92,
8089-8115, 1987.
[15] W.F. McDonough and S.-s. Sun, The composition of the Earth., Chemical Geology
120, 223-253, 1995.
[16] R.K.Workman and Hart S.R. Major and trace element composition of the depleted
MORB mantle (DMM). Earth Planetary Science Letters 231, 1-2, 53-72, 2005.
[17] I.N. Bindeman, A.M. Davis and M.J. Drake, Ion microprobe study of plagioclasebasalt partition experiments at natural concentration levels of trace elements. Geochimica
et Cosmochimica Acta 62, 7, 1175-1193, 1998
[18] I.N. Bindeman and A.M. Davis, Trace element partitioning between plagioclase and
melt: investigation of dopant influence on partition behavior Geochimica et
Cosmochimica Acta, 64, 16, 2863-2878, 2000.
[19] D.J. Cherniak and E.B. Watson, A study of strontium diffusion in plagioclase using
Rutherford backscattering spectroscopy. Geochimica et Cosmochimica Acta, 58, 23,
5179-5190, 1994.
[20] D.J. Cherniak, Diffusion of lead in plagioclase and K-feldspar: an investigation
using Rutherford Backscattering and Resonant Nuclear Reaction Analysis. Contributions
to Mineralogy and Petrology 120, 3-4, 358-371, 1995.
[21] D.J. Cherniak, Ba diffusion in feldspar Geochimica et Cosmochimica Acta 66, 9,
1641-1650, 2002.
6
[22] D.J. Cherniak, REE diffusion in feldspar Chemical Geology 193, 1-2, 25-41, 2002.
[23] B.J. Giletti and J.E.D. Casserly, Strontium diffusion kinetics in plagioclase
feldspars. Geochimica et Cosmochimica Acta, 58, 18, 3785-3793, 1994.
[24] B.J. Giletti and T.M. Shanahan, Alkali diffusion in plagioclase feldspar. Chemical
Geology 139, 1-4, 3-20, 1997.
[25] H. Behrens, W. Johannes and H. Schmalzried, On the mechanisms of cation
diffusion processes in ternary feldspars. Physics and Chemistry of Minerals 17, 1, 62-78,
1990.
[26] A.E. Saal and J.A. Van Orman, The 226Ra enrichment in oceanic basalts: evidence
for melt-cumulate diffusive interaction processes within the oceanic lithosphere.
Geochemistry, Geophysics and Geosystems, 5, 2, doi:10.1029/2003GC000620, 2004.
[27] C.J. Lissenberg, J.H. Bedard and C.R. van Staal, The structure and geochemistry of
the gabbro zone of the Annieopsquotch ophiolite, Newfoundland: implications for lower
crustal accretion at spreading ridges., Earth and Planetary Science Letters 229, 105-123,
2004.
[28] C.J. Garrido, P.B. Kelemen and G. Hirth, Variation of cooling rate with depth in
lower crust formed at an oceanic spreading ridge: Plagioclase crystal size distributions in
gabbros from the Oman ophiolite. Geochemistry, Geophysics, Geosystems 2: Art. No.
2000GC000136, 2001.
[29] S.R. Hart, J. Blusztajn, H.J.B. Dick, P.S. Meyer and K. Muehlenbachs, The
fingerprint of seawater circulation in a 500-meter section of ocean crust gabbros,
Geochimica et Cosmochimica Acta 63(23-24), 4059– 4080, 1999.
7
Supplementary Information Table 1: Majors, trace element and isotopic composition of Galapagos lavas
Sample ID
island/volcano
Lat. South
Long. West
SiO2
Al2O3
TiO2
FeO
MnO
CaO
MgO
K2O
Na2O
P2O5
Total
XRF
Ni
Cr
Sc
V
Ba
Rb
Sr
Zr
Y
Nb
Ce
ICPMS
Cs
Rb
Ba
Th
U
Nb
Ta
La
Ce
Pb
Pr
Sr
Nd
Hf
Sm
Eu
Gd
Tb
Dy
Y
Ho
Er
Tm
Yb
Lu
Sc
87Sr/86Sr
143Nd/144Nd
176Hf/177Hf
206Pb/204Pb
207Pb/204Pb
208Pb/204Pb
208*/206*
NSK97-09 NSK97-201 NSK97-202 NSK97-204 NSK97-206 NSK97-211 NSK97-213 NSK97-214
Fernandina Fernandina Fernandina Fernandina Fernandina Fernandina Fernandina Fernandina
0 18.05
0 17.00
0 17.63
0 18.27
0 17.10
0 18.35
0 18.11
0 18.04
91 37.72 91 30.68
91 30.73
91 30.45
91 31.63
91 39.07
91 38.64
91 37.72
48.54
49.20
49.39
49.47
49.61
49.26
48.36
49.44
14.61
14.24
15.87
15.69
15.52
14.53
15.67
14.77
3.03
3.50
2.50
2.47
2.58
3.15
3.10
3.12
11.60
11.43
9.66
9.93
9.83
11.37
11.44
11.30
0.18
0.186
0.168
0.164
0.170
0.185
0.18
0.187
11.52
11.29
11.88
11.84
11.89
11.36
11.49
11.60
7.19
6.45
7.63
7.34
7.50
6.57
6.21
6.41
0.41
0.49
0.31
0.32
0.32
0.46
0.44
0.39
2.59
2.85
2.36
2.53
2.34
2.78
2.76
2.46
0.32
0.366
0.245
0.243
0.249
0.336
0.33
0.315
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
NSK97-215
NSK97-216
NSK97-220
NSK97-224
Fernandina Fernandina (1982) Fernandina Fernandina (1995)
0 22.59
0 23.67
0 21.15
0 27.37
91 33.93
91 33.07
91 34.00
91 36.65
48.54
48.88
48.88
49.18
13.95
15.05
16.15
15.71
2.72
3.42
2.47
3.09
10.42
11.46
9.62
10.22
0.175
0.188
0.164
0.176
10.68
11.34
12.68
11.75
10.39
5.84
7.17
6.21
0.36
0.54
0.31
0.47
2.49
2.91
2.32
2.86
0.269
0.376
0.244
0.337
100.00
100.00
100.00
100.00
NSK97-228
Fernandina
0 22.59
91 33.93
48.90
16.08
3.13
10.72
0.180
11.45
5.81
0.48
2.92
0.338
100.00
NSK97-229 NSK97-232
Fernandina Fernandina
0 25.33
0 26.36
91 23.29
91 24.35
49.14
49.53
15.59
16.15
2.69
2.42
10.62
9.63
0.175
0.168
12.09
12.33
6.43
6.77
0.38
0.31
2.60
2.46
0.271
0.237
100.00
100.00
81
298
36
356
85
6
336
155
29
20.5
30
54
119
36
388
33
8
368
178
28
26.2
49
97
323
35
309
35
6
311
126
23
17.4
42
87
311
33
302
13
6
313
126
23
16.6
32
87
288
34
316
37
6
304
128
23
17.3
36
55
166
32
353
49
7
333
175
30
24.6
33
52
181
30
355
87
6
341
171
30
22.6
41
49
155
33
366
37
6
321
155
29
21.0
38
226
643
35
327
40
7
302
135
24
18.3
22
37
143
35
383
41
9
350
196
31
26.0
37
77
341
32
294
10
5
327
125
22
16.5
27
51
186
35
357
58
10
344
171
29
23.8
43
37
185
30
351
29
5
354
172
29
22.3
45
57
187
33
330
29
5
335
145
26
17.3
35
62
242
34
316
26
4
315
124
23
15.6
35
0.08
7
94
1.76
0.38
19.31
1.5
13.85
31.8
1.06
4.27
332
20.09
4.02
5.74
2.09
6.05
1.05
5.99
27.6
1.18
2.96
0.4
2.4
0.35
29.8
0.09
9.25
113.89
1.74
0.50
26.18
1.88
17.77
40.51
1.18
5.22
396.10
23.23
4.57
6.51
2.25
6.40
1.11
6.41
32.33
1.29
3.17
0.42
2.51
0.37
45.45
0.703294
0.512928
0.06
5.39
67.93
0.97
0.28
15.18
1.08
10.51
24.27
1.42
3.22
305.49
15.62
3.06
4.50
1.60
4.74
0.80
4.69
23.87
0.93
2.27
0.31
1.82
0.27
35.61
0.703272
0.512921
0.283056
19.083
15.549
38.732
0.947
0.05
5.92
72.54
1.00
0.29
15.81
1.17
11.10
25.25
0.71
3.47
331.72
16.04
3.34
4.54
1.69
5.00
0.87
4.99
25.08
1.01
2.50
0.34
1.99
0.30
39.06
0.703281
0.512930
0.06
5.30
66.93
0.94
0.27
15.08
1.08
10.39
23.97
1.16
3.24
300.61
15.19
3.12
4.48
1.61
4.69
0.79
4.76
24.09
0.93
2.35
0.33
1.85
0.28
35.46
0.703289
0.512929
0.283038
19.105
15.552
38.760
0.948
0.08
7.65
95.09
1.33
0.39
20.70
1.49
15.24
34.20
1.39
4.55
318.56
21.23
4.27
5.98
2.11
6.12
1.04
6.02
30.40
1.14
2.94
0.41
2.41
0.36
38.42
0.703214
0.512940
0.09
7.5
100
1.52
0.4
20.5
1.57
15.11
34.66
1.07
4.65
343
21.89
4.43
6.28
2.21
6.57
1.12
6.55
30.28
1.27
3.21
0.45
2.6
0.39
29.1
0.07
7.59
88.93
1.31
0.38
20.02
1.50
14.10
32.27
0.93
4.23
351.39
19.54
4.15
5.77
2.04
6.01
1.03
6.08
30.14
1.21
3.04
0.41
2.46
0.36
44.25
0.703290
0.512937
0.283047
19.080
15.537
38.711
0.945
0.06
5.96
73.47
1.05
0.30
16.89
1.24
11.83
27.00
1.15
3.66
289.08
17.11
3.35
4.88
1.78
5.09
0.87
5.13
24.32
0.99
2.46
0.34
2.00
0.29
33.62
0.703264
0.512922
0.283054
19.109
15.548
38.758
0.947
0.08
9.67
114.20
1.64
0.49
24.31
1.75
17.88
40.24
1.10
5.27
370.66
23.91
4.96
6.92
2.37
7.04
1.17
6.94
34.54
1.37
3.42
0.47
2.80
0.41
42.20
0.703251
0.512958
0.283056
19.065
15.548
38.688
0.944
0.05
6.09
72.23
1.01
0.28
15.86
1.20
11.45
25.73
0.70
3.35
347.36
15.89
3.36
4.77
1.70
4.95
0.83
5.09
24.82
0.99
2.41
0.34
2.02
0.29
42.74
0.703272
0.512947
0.09
8.22
98.57
1.35
0.39
21.46
1.51
15.68
35.41
3.21
4.72
359.55
21.58
4.34
6.26
2.10
6.32
1.07
6.30
31.63
1.21
3.10
0.43
2.48
0.37
40.00
0.703200
0.512940
0.283061
19.068
15.541
38.672
0.942
0.08
8.57
101.04
1.36
0.40
21.58
1.59
15.46
34.81
0.96
4.65
383.16
21.46
4.50
6.13
2.18
6.50
1.09
6.40
32.13
1.25
3.21
0.44
2.57
0.38
39.10
0.703207
0.512947
0.283048
19.052
15.550
38.706
0.947
0.05
6.54
79.62
1.18
0.34
17.47
1.30
13.15
30.14
0.83
3.87
352.71
17.93
3.81
5.36
1.86
5.39
0.94
5.57
26.99
1.10
2.75
0.37
2.16
0.33
40.47
0.703242
0.512954
0.04
5.01
66.43
0.95
0.28
14.47
1.08
10.93
25.11
0.69
3.35
335.46
15.44
3.18
4.45
1.61
4.85
0.83
4.92
24.29
0.97
2.45
0.32
1.94
0.29
38.26
0.703260
0.512949
19.077
15.536
38.692
0.943
Supplementary Information Table 1: Majors, trace element and isotopic composition of Galapagos lavas
Sample ID
island/volcano
Lat. South
Long. West
SiO2
Al2O3
TiO2
FeO
MnO
CaO
MgO
K2O
Na2O
P2O5
Total
XRF
Ni
Cr
Sc
V
Ba
Rb
Sr
Zr
Y
Nb
Ce
ICPMS
Cs
Rb
Ba
Th
U
Nb
Ta
La
Ce
Pb
Pr
Sr
Nd
Hf
Sm
Eu
Gd
Tb
Dy
Y
Ho
Er
Tm
Yb
Lu
Sc
87Sr/86Sr
143Nd/144Nd
176Hf/177Hf
206Pb/204Pb
207Pb/204Pb
208Pb/204Pb
208*/206*
SG 93-16
GI92-26
Fernandina Fernandina
0 28.005 0 23.41
91 23.836 91 22.45
49.34
48.83
15.54
15.62
2.72
3.20
10.45
11.33
0.176
0.175
12.09
11.14
6.52
5.90
0.38
0.50
2.50
2.95
0.272
0.355
100.00
100.00
GI92-27
NSK97-238 NSK97-246 NSK97-126
NSK97-252
NSK97-254 NSK97-260 NSK97-262 SG93-19
SG93-22
SG93-23
SG 93-1
SG 93-5
SG 93-7
SG 93-8
Fernandina
Santiago
Santiago
San Cristobal San Cristobal San Cristobal
Floreana
Floreana
Floreana
Floreana
Floreana Cerro Azul Cerro Azul Cerro Azul Cerro Azul
0 26.00
0 14.25
0 21.95
0 47.46
0 43.54
0 45.82
1 14.00
1 18.88
1 13.722 1 16.288 1 16.729 0 59.021 0 52.089 0 50.495 0 48.866
91 23.53 90 51.03
90 35.24
89 28.17
89 24.63
89 27.31
90 25.35
90 30.50 90 25.465 90 29.285 90 29.317 91 26.817 91 29.877 91 28.7 91 26.835
48.96
48.40
47.26
47.09
47.30
47.44
46.78
48.19
46.70
46.40
47.13
47.97
48.09
50.93
48.89
15.11
14.85
16.95
17.11
16.77
16.40
14.30
16.16
14.16
14.39
14.04
15.47
14.79
13.60
15.49
3.32
2.13
2.25
1.13
1.01
0.98
1.75
1.85
1.74
1.67
1.27
2.28
2.04
1.45
2.66
11.16
10.67
10.72
9.04
8.83
9.04
9.09
9.80
9.49
9.78
9.15
10.35
9.96
9.47
10.22
0.187
0.178
0.180
0.16
0.165
0.170
0.172
0.174
0.176
0.169
0.164
0.174
0.175
0.155
0.185
11.34
10.25
11.21
11.33
12.37
11.55
11.68
11.72
11.25
10.60
10.80
12.67
11.75
9.90
11.28
6.13
10.22
7.93
11.16
11.17
11.80
13.21
8.48
13.64
13.30
13.54
7.83
10.20
11.84
7.03
0.49
0.31
0.30
0.34
0.12
0.19
0.54
0.55
0.48
0.83
0.89
0.41
0.34
0.18
0.66
2.93
2.74
2.92
2.49
2.16
2.31
2.29
2.88
2.17
2.66
2.79
2.59
2.45
2.34
3.23
0.360
0.242
0.281
0.15
0.104
0.112
0.193
0.201
0.190
0.218
0.223
0.256
0.210
0.137
0.350
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
100.00
58
193
38
330
40
6
330
146
25
19.3
17
38
163
28
334
76
9
364
185
33
26.2
46
41
172
38
382
60
7
337
179
30
24.5
36
241
495
34
246
27
5
234
139
29
11.1
26
111
174
31
266
29
4
312
143
27
14.0
36
239
447
38
217
66
4
229
82
23
11.8
8
211
481
40
228
28
3
173
66
21
5.2
9
259
479
40
210
44
3
173
67
21
6.5
14
329
824
35
255
170
12
330
88
21
17.1
30
118
273
37
296
170
10
301
89
22
17.0
26
336
828
34
264
145
9
338
93
22
16.2
26
324
698
33
247
257
20
390
91
21
23.5
18
346
1035
33
233
443
20
372
83
19
24
36
0.07
6.24
79.64
1.14
0.34
17.11
1.23
12.88
28.14
1.18
3.79
323.24
17.71
3.59
5.02
1.84
5.32
0.92
5.47
26.80
1.04
2.59
0.36
2.07
0.31
38.72
0.703205
0.512965
0.283069
19.036
15.549
38.645
0.942
0.1
8.91
107
1.3
0.31
24.11
1.55
16.73
37.25
1.56
4.94
364
22.6
4.23
6.41
2.18
6.4
1.07
6.52
32.57
1.26
3.34
0.45
2.54
0.38
0.06
4.8
74
0.99
0.18
8.89
0.63
6.95
14.63
0.72
1.94
229
9.04
2.06
2.82
1.08
3.43
0.67
4.31
22.77
0.96
2.73
0.39
2.48
0.39
32.3
0.702998
0.513028
0.283128
18.797
15.558
38.459
0.947
0.02
1.21
23.89
0.24
0.06
3.57
0.25
3.38
8.46
0.90
1.31
173.67
6.79
1.65
2.40
0.96
3.16
0.61
4.08
22.96
0.89
2.56
0.37
2.36
0.37
45.35
0.702991
0.513041
0.283144
18.704
15.555
38.386
0.948
0.04
2.28
38.49
0.37
0.10
4.94
0.31
4.31
9.53
0.75
1.36
167.06
6.58
1.54
2.18
0.89
2.84
0.55
3.87
22.08
0.83
2.37
0.34
2.32
0.36
41.56
0.703086
0.513042
0.283139
18.783
15.546
38.416
0.943
0.14
9.42
180
0.89
0.17
16.14
0.91
11.99
21.86
1.72
2.69
338
11.88
1.85
3.39
1.22
3.69
0.66
4.14
22.46
0.82
2.41
0.31
1.85
0.3
0.23
19.42
279
1.3
0.22
22.15
1.22
15.96
26.63
2.99
3.04
390
12.45
1.8
3.24
1.18
3.5
0.64
4.03
21.91
0.82
2.34
0.32
1.96
0.32
0.19
20.05
416
1.44
0.23
23.52
1.3
16.42
26.93
2.93
2.89
372
11.08
1.58
2.74
0.95
2.85
0.52
3.32
18.14
0.68
2.03
0.27
1.68
0.26
0.703446
0.512945
19.056
15.535
38.667
0.943
0.04
3.89
62.93
0.76
0.22
11.45
0.79
10.32
24.63
1.62
3.41
306.94
16.40
3.55
4.95
1.78
5.52
0.94
5.63
29.05
1.12
2.89
0.40
2.44
0.37
33.03
0.703089
0.513018
0.283071
19.016
15.587
38.701
0.950
0.12
10.07
166.80
0.86
0.19
13.76
0.84
9.92
18.85
2.28
2.33
284.96
10.35
2.10
3.12
1.15
3.78
0.67
4.35
22.73
0.88
2.31
0.34
2.07
0.31
36.30
0.703379
0.512943
19.049
15.541
38.687
0.945
0.05
5.09
48.64
0.76
0.21
9.60
0.65
9.00
20.96
1.07
2.99
229.61
14.86
3.53
4.90
1.79
5.61
0.94
5.86
29.18
1.12
2.93
0.39
2.37
0.35
32.35
0.703036
0.512998
0.283033
19.015
15.561
38.657
0.946
0.13
11.27
172.18
0.98
0.23
15.26
0.97
11.38
21.58
1.29
2.66
336.60
12.00
2.03
3.37
1.23
3.75
0.65
4.16
22.42
0.85
2.19
0.32
1.87
0.29
38.98
0.703211
0.512922
0.10
8.50
105.30
1.55
0.45
23.19
1.62
16.36
36.94
1.46
4.95
348.69
23.22
4.70
6.67
2.34
7.03
1.17
7.00
33.45
1.37
3.38
0.47
2.75
0.41
40.81
0.703208
0.512950
19.732
15.607
39.385
0.950
19.614
15.599
39.300
0.953
0.703415
0.512946
0.283111
19.794
15.639
39.534
0.959
117
356
39
321
95
6
348
141
24
24.3
59
0.04
5.10
113.84
1.42
0.37
20.01
1.33
15.95
33.34
4.85
4.14
329.94
18.79
3.31
4.93
1.71
5.08
0.85
4.98
24.62
0.95
2.42
0.34
2.03
0.31
36.41
0.703518 0.703300
0.512948
19.857
15.617
39.544
0.954
201
508
37
288
77
6
284
113
22
17.8
40
356
583
30
189
38
5
182
77
17
9.1
17
97
227
35
326
107
13
358
178
29
28.8
49
0.06
6.11
87.76
1.08
0.29
15.50
1.01
12.02
26.09
1.10
3.40
282.58
15.49
2.97
4.42
1.56
4.73
0.80
4.88
23.29
0.95
2.40
0.34
1.96
0.30
36.77
0.703323
0.512959
0.283069
19.358
15.585
39.002
0.948
0.04
3.61
43.63
0.57
0.16
7.66
0.52
5.75
12.88
0.99
1.82
186.90
8.87
2.10
3.05
1.13
3.60
0.64
3.77
19.69
0.75
1.93
0.26
1.55
0.23
31.74
0.703366
0.512958
0.283086
19.308
15.605
39.059
0.958
0.13
12.51
148.21
1.99
0.54
27.56
1.79
20.80
43.17
1.89
5.61
377.64
24.86
4.60
6.53
2.22
6.68
1.08
6.56
32.36
1.26
3.20
0.44
2.63
0.39
36.73
0.703288
0.512929
0.283057
19.314
15.611
38.932
0.945
Supplementary Information Table 1: Majors, trace element and isotopic composition of Galapagos lavas
Sample ID
island/volcano
Lat. South
Long. West
SiO2
Al2O3
TiO2
FeO
MnO
CaO
MgO
K2O
Na2O
P2O5
Total
XRF
Ni
Cr
Sc
V
Ba
Rb
Sr
Zr
Y
Nb
Ce
ICPMS
Cs
Rb
Ba
Th
U
Nb
Ta
La
Ce
Pb
Pr
Sr
Nd
Hf
Sm
Eu
Gd
Tb
Dy
Y
Ho
Er
Tm
Yb
Lu
Sc
87Sr/86Sr
143Nd/144Nd
176Hf/177Hf
206Pb/204Pb
207Pb/204Pb
208Pb/204Pb
208*/206*
SG 93-12
SG 93-13
Cerro Azul Cerro Azul
0 47.3
0 46.668
91 25.66 91 24.467
49.08
49.12
14.34
15.14
3.40
2.53
12.37
10.32
0.216
0.181
10.24
12.15
5.72
6.85
0.78
0.53
3.44
2.89
0.416
0.285
100.00
100.00
SG 93-15
GI92-10 CA 90-15
Cerro Azul Cerro Azul Cerro Azul
0 46.011 0 58.97
91 23.968 91 26.63
48.88
48.39
48.73
15.43
13.97
15.61
2.40
3.73
2.16
10.07
13.69
10.12
0.173
0.223
0.17
12.50
9.90
12.51
7.00
5.31
7.20
0.52
0.78
0.46
2.78
3.59
2.82
0.267
0.433
0.23
100.00
100.00
100.00
GI92-1
Santa Cruz
0 44.54
90 17.53
47.44
16.13
1.72
10.60
0.171
9.84
10.55
0.22
3.08
0.255
100.00
GI92-2A
Santa Cruz
0 29.35
90 15.73
48.82
14.99
1.78
10.48
0.170
10.71
9.72
0.33
2.77
0.220
100.00
GI92-4
Santa Cruz
47.19
16.03
1.28
10.35
0.163
10.81
11.40
0.15
2.46
0.159
100.00
GI92-5
Sierra Negra
0 38.11
90 55.27
49.33
13.97
3.01
12.20
0.195
11.19
6.50
0.44
2.86
0.289
100.00
GI92-8
Sierra Negra
0 55.08
90 53.42
49.06
14.18
3.36
12.71
0.206
10.62
5.99
0.49
3.03
0.344
100.00
SN91-05
Sierra Negra
0 49.45
91 09.25
49.40
13.77
3.30
12.92
0.21
10.42
5.88
0.55
3.19
0.35
100.00
SN91-32
Sierra Negra
0 58.20
91 03.15
48.61
13.21
3.98
14.69
0.23
9.61
5.11
0.70
3.40
0.46
100.00
SN91-40
Sierra Negra
0 47.45
90 49.25
48.40
13.62
3.68
13.61
0.22
10.42
5.68
0.61
3.33
0.42
100.00
SN 91-60
Sierra Negra
0 56.30
90 56.30
48.97
14.13
3.48
13.29
0.21
10.59
5.62
0.50
2.85
0.36
100.00
GI92-16
Alcedo
0 23.51
91 13.15
49.64
14.00
3.06
12.36
0.204
10.82
6.11
0.50
2.91
0.408
100.00
GI92-17
Darwin
0 17.67
91 20.92
48.98
16.61
2.79
10.37
0.168
11.88
5.65
0.39
2.89
0.266
99.99
GI92-18
Darwin
0 12.70
91 23.10
48.80
16.74
2.58
10.05
0.164
12.39
6.07
0.32
2.64
0.240
100.00
36
84
19
37
29
79
15
294
191
12
296
227
11
308
203
40
101
30
385
102
8
298
192
36
27.6
29
44
98
38
351
50
8
284
180
35
21.5
40
53
99
34
331
22
6
374
166
27
14.5
44
56
152
35
309
24
3
382
155
26
12.8
38
0.11
9.8
120
1.64
0.48
26.97
1.89
19.03
42.35
1.43
5.53
308
25.68
5.09
7.22
2.52
7.76
1.31
7.56
37.05
1.47
3.74
0.51
3.04
0.45
34.4
0.703397
0.512925
0.283034
19.357
15.591
39.022
0.950
0.09
7.94
85
1.25
0.3
21.04
1.35
15.78
35.45
2.9
4.76
284
22.19
4.46
6.71
2.32
6.94
1.23
7.2
35.59
1.39
3.71
0.5
2.94
0.44
0.05
4.72
51
0.78
0.2
13.36
0.91
12.95
30.14
2.49
4.1
374
19.02
3.61
5.78
2.06
5.75
0.98
5.81
27.81
1.14
2.95
0.39
2.25
0.33
0.06
3.61
44
0.66
0.18
11.9
0.8
11.52
26.76
2.27
3.68
382
17.43
3.36
5.2
1.89
5.36
0.89
5.29
25.42
1.01
2.7
0.35
2.01
0.29
0.703174
0.512962
0.702898
0.513011
0.283075
18.797
15.518
38.296
0.929
0.702856
0.513020
0.283068
18.810
15.536
38.268
0.925
46
90
34
412
138
15
356
211
35
36.4
75
74
119
37
330
105
11
329
152
26
23.3
41
86
138
41
309
92
11
342
143
24
22.2
46
27
46
30
422
160
16
374
226
39
40
60
73
147
41
313
80
9
342
135
25
236
406
30
198
0
2
374
158
26
6.2
50
229
362
29
212
31
5
259
128
27
9.4
32
234
454
34
199
1
2
298
93
21
5.7
19
53
158
40
367
63
8
296
160
32
25.3
45
40
104
36
362
92
9
302
183
35
28.5
38
0.18
14.91
188.27
2.35
0.64
33.88
2.19
25.00
51.64
2.27
6.55
361.68
28.90
5.45
7.74
2.59
7.71
1.27
7.77
37.13
1.48
3.74
0.51
3.00
0.45
34.86
0.703344
0.512917
0.283054
19.370
15.594
39.045
0.951
0.12
10.83
124.41
1.57
0.43
22.80
1.50
16.73
35.23
1.99
4.51
341.74
19.98
3.90
5.49
1.90
5.75
0.95
5.61
28.03
1.08
2.69
0.37
2.18
0.34
41.26
0.703372
0.512916
0.11
9.88
126.61
1.52
0.43
21.04
1.42
16.00
33.62
1.53
4.32
343.26
19.64
3.68
5.35
1.84
5.58
0.92
5.54
25.92
1.08
2.64
0.37
2.13
0.32
39.34
0.703336
0.512915
0.283045
0.16
14.66
178.98
2.48
0.68
35.00
2.32
26.32
54.42
2.06
6.91
377.61
30.59
5.72
8.01
2.71
8.00
1.33
7.83
39.40
1.55
3.88
0.54
3.22
0.48
34.13
0.703321
0.512934
0.09
9
80
1.06
0.33
3.54
14.56
31.33
1.14
4.03
342
18.4
3.45
5.07
1.79
5.3
0.88
5.2
25
1.02
2.68
0.35
2.07
0.31
41
0.703327
0.512930
0.03
2.25
34
0.37
0.09
5.71
0.36
7.92
20.5
0.89
3
374
14.62
3.37
4.37
1.64
4.48
0.84
5.17
26.89
1.04
2.9
0.4
2.38
0.38
0.05
4.13
63
0.54
0.14
9
0.55
8.66
19.9
0.72
2.76
259
13.15
2.94
4.08
1.56
4.53
0.86
5.3
27.7
1.07
3
0.38
2.38
0.36
0.01
1.34
22
0.22
0.07
3.44
0.22
4.46
11.45
1.08
1.78
298
8.76
1.94
2.93
1.14
3.22
0.6
3.87
20.49
0.78
2.24
0.3
1.86
0.28
0.08
7.48
97
1.13
0.25
23.26
1.5
15.63
34.26
0.98
4.68
296
21.39
3.94
6.19
2.21
6.12
1.1
6.56
31.72
1.28
3.35
0.43
2.59
0.39
0.1
8.71
115
1.39
0.36
26.82
1.75
17.94
39.41
1.38
5.31
302
24.45
4.53
6.99
2.46
6.95
1.24
7.25
35.73
1.4
3.75
0.5
2.91
0.44
0.16
15
194
2.7
0.74
49
2.5
34.2
73.4
2.2
9.4
294
42.5
6.8
11.2
3.6
11
1.88
11
57
2.21
5.9
0.77
4.68
0.7
0.14
12
137
1.8
0.47
34
1.7
22.2
47.9
1.4
6.1
296
28.1
4.9
7.68
2.6
7.5
1.35
8.1
40
1.58
4.2
0.54
3.29
0.49
0.12
11
129
1.6
0.42
33
1.6
21.2
46.3
1.5
6.1
308
28.2
4.8
7.68
2.6
7.5
1.35
8.1
41
1.57
4.2
0.54
3.27
0.48
0.702952
0.513099
0.703038
0.513018
0.702764
0.513087
0.703375
0.703420
19.438
15.588
39.101
0.950
18.521
15.497
38.038
0.929
18.819
15.536
38.417
0.940
18.617
15.541
38.153
0.932
0.703504
0.512941
0.283032
19.353
15.585
39.015
0.949
0.703392
19.337
15.561
38.955
0.945
0.703365
0.512904
0.283042
19.350
15.579
39.103
0.959
19.380
15.602
39.082
0.954
19.386
15.583
39.052
0.950
19.393
15.613
39.116
0.956
19.116
15.552
38.699
0.940
Supplementary Information Table 1: Majors, trace element and isotopic composition of Galapagos lavas
Sample ID
island/volcano
Lat. South
Long. West
SiO2
Al2O3
TiO2
FeO
MnO
CaO
MgO
K2O
Na2O
P2O5
Total
XRF
Ni
Cr
Sc
V
Ba
Rb
Sr
Zr
Y
Nb
Ce
ICPMS
Cs
Rb
Ba
Th
U
Nb
Ta
La
Ce
Pb
Pr
Sr
Nd
Hf
Sm
Eu
Gd
Tb
Dy
Y
Ho
Er
Tm
Yb
Lu
Sc
87Sr/86Sr
143Nd/144Nd
176Hf/177Hf
206Pb/204Pb
207Pb/204Pb
208Pb/204Pb
208*/206*
GI92-19
Ecuador
0 1.57
91 31.14
47.93
14.32
2.29
11.00
0.170
9.90
10.90
0.52
2.69
0.278
100.00
W95-25
Wolf
0 0.01
91 0.36
48.76
16.65
2.58
10.17
0.16
12.09
6.02
0.39
2.89
0.28
100.00
W95-54
Wolf
0 0.05
91 0.41
48.91
14.47
3.30
11.92
0.20
10.75
6.17
0.58
3.29
0.41
100.00
W95-59
Wolf
0 0.06
91 0.41
49.21
15.27
2.92
10.74
0.18
11.43
6.45
0.47
2.98
0.35
100.00
W95-74
Wolf
0 0.07
91 0.29
48.21
18.86
2.15
8.85
0.14
12.82
5.63
0.35
2.74
0.25
100.00
W95-75R
Wolf
0 0.07
91 0.29
48.90
18.20
2.30
8.83
0.14
12.39
5.86
0.37
2.73
0.28
100.00
W95-84
Wolf
0 0.08
91 0.30
48.58
15.21
2.88
10.97
0.18
11.20
7.09
0.50
3.05
0.35
100.00
265
499
27
226
95
10
344
150
25
22.1
38
59
141
32
311
40
3
387
178
29
14.5
40
50
105
38
374
54
6
374
250
37
18.8
46
73
130
37
332
95
5
380
205
32
16
73
65
196
29
248
50
3
428
160
25
12.4
37
72
205
35
273
82
3
418
170
27
13
102
89
248
28
343
58
3
374
216
34
17.3
51
61
78
35
247
124
8
340
143
26
20.9
34
38
132
44
335
41
3
221
164
39
12.5
23
0.11
9.21
120
1.27
0.28
20.54
1.2
15.73
32.27
1.68
4.09
344
18.84
3.36
5.23
1.81
5.22
0.89
5.12
25.99
0.99
2.67
0.35
2.04
0.3
0.05
4.8
53
0.85
0.27
12.59
0.92
12.5
30.24
1.5
4.2
404
20.31
4.64
6.07
2.2
6.54
1.11
6.32
29.89
1.22
3.04
0.41
2.43
0.36
34.4
0.702773
0.513039
0.283096
18.868
15.557
38.423
0.936
0.07
6.8
69
1.23
0.4
18.15
1.32
18.39
44.31
1.58
6.09
386
28.63
6.38
8.24
2.82
8.65
1.42
8.17
39.16
1.56
3.93
0.53
3.14
0.46
37.7
0.702744
0.513030
0.283097
18.889
15.541
38.373
0.928
0.06
5.9
60
1.03
0.35
15.74
1.14
15.74
37.81
1.38
5.22
388
24.56
5.46
7.1
2.49
7.4
1.24
7.06
34.04
1.37
3.36
0.46
2.71
0.4
36.9
0.702747
0.513041
0.283104
18.897
15.542
38.377
0.928
0.06
8.6
130
1.66
0.39
19.98
1.39
15.69
33.48
0.95
4.29
349
19.68
3.78
5.36
1.94
5.62
0.95
5.51
26.31
1.08
2.71
0.37
2.2
0.33
32.4
0.703288
0.512909
0.283012
19.156
15.594
39.003
0.967
0.05
5.16
51
0.73
0.21
10.78
0.76
10.01
24.70
0.85
3.62
235
18.46
4.39
6.07
2.23
7.33
1.29
7.65
40.25
1.56
4.10
0.57
3.41
0.52
40.23
0.702843
0.513028
0.283109
18.893
15.563
38.500
0.941
0.703200
0.512951
0.283051
19.304
15.600
39.059
0.959
15.76
39.21
5.70
380
26.82
6.88
2.26
7.15
1.10
6.22
1.19
3.12
0.42
2.56
0.39
0.702828
0.513030
0.283096
18.740
15.552
38.606
0.968
0.04
4.1
47
0.74
0.26
11.12
0.81
11.38
27.32
5.04
3.78
443
18.22
4.04
5.39
1.97
5.66
0.95
5.45
25.89
1.03
2.56
0.35
2.04
0.3
32.9
0.702766
0.513042
0.283105
18.895
15.558
38.452
0.936
15.18
38.12
5.56
418
26.21
6.63
2.16
6.86
1.07
6.03
1.15
3.03
0.42
2.53
0.39
0.702764
0.513038
SKV98-108
SKV98-101
Pinta
Marchena (1991)
0 36.52
0 21.43
90 44.41
90 32.18
48.53
48.73
16.84
14.42
2.04
2.51
9.13
12.37
0.15
0.20
12.77
11.49
6.96
6.31
0.53
0.32
2.79
3.38
0.26
0.26
100.00
100.00
Supplementary Information Table 2: Correlation Matrix of Median Values for each Volcano
Sr/Nd
Ba/Th
K/La
Sr/Sr*
Delta Sr
Pb/Nd
TiO2(8)
Nb/Nb*(La)
Sr/Nd
1
0.6590
0.6719
0.9779
0.8273
0.7718
-0.7701
-0.8139
Ba/Th
K/La
Sr/Sr*
Δ Sr
Pb/Nd
TiO2(8)
Nb/Nb*(La)
1
0.8771
0.6221
0.8215
0.8256
-0.5702
-0.6141
1
0.6251
0.7873
0.7728
-0.7704
-0.6206
1
0.8204
0.6550
-0.8015
-0.8353
1
0.7809
-0.7242
-0.5579
1
-0.5131
-0.5712
1
0.7348
1
Supplementary Information Table 3:
Input and Output of the Plagioclase-rich Cumulate-Basalt Diffusive Interaction Model
PM normalized input data
Ba
Th
U
Nb
Ta
K
La
Ce
Pb
Pr
Sr
Nd
Zr
Hf
Sm
3+
Eu
2+
Eu
Ti
Gd
Tb
Dy
Y
Ho
Er
Tm
Yb
Lu
Sr/Sr*
87
Sr/86Sr
ΔSr
Kd(plag-melt)
2.18E-01
8.00E-03
8.00E-03
1.00E-02
1.00E-02
1.31E-01
1.61E-01
1.02E-01
6.70E-01
1.13E-01
1.67E+00
8.00E-02
3.54E-03
3.54E-03
9.90E-02
3.96E-01
3.96E-01
6.11E-02
5.02E-02
PM normalized output data
Ba
Th
U
Nb
Ta
K
La
Ce
Pb
Pr
Sr
Nd
Zr
Hf
Sm
Eu
Ti
Gd
Tb
Dy
Y
Ho
Er
Tm
Yb
Lu
Sr/Sr*
87
Sr/86Sr
ΔSr
2
Enriched melt
5.30
5.33
5.97
10.18
11.08
4.70
9.70
9.37
6.48
8.72
8.40
8.07
7.75
7.57
6.97
6.77
6.50
6.22
5.64
5.29
4.64
3.75
4.17
3.66
3.29
3.08
3.01
1.00
0.7030
0
5 years
2.41
2.78
3.16
5.43
6.79
4.22
5.23
5.05
4.41
5.25
7.51
5.44
10 years
2.42
2.78
3.16
5.43
6.79
4.44
5.23
5.06
5.03
5.25
8.13
5.44
20 years
2.45
2.78
3.16
5.43
6.79
4.49
5.25
5.07
5.75
5.26
8.34
5.45
40 years
2.49
2.78
3.16
5.43
6.79
4.50
5.27
5.08
6.25
5.27
8.36
5.45
300 years
3.02
2.78
3.16
5.43
6.79
4.50
5.59
5.28
6.39
5.40
8.36
5.52
6.29
5.83
5.92
5.81
5.19
5.95
5.96
5.97
6.29
5.83
5.92
5.88
5.19
5.95
5.96
5.97
6.29
5.83
5.92
5.98
5.19
5.95
5.96
5.97
6.29
5.83
5.92
6.06
5.19
5.95
5.96
5.97
6.29
5.83
5.95
6.14
5.19
5.96
5.97
5.97
5.82
5.66
5.51
5.35
1.40
0.702905
2.72
5.82
5.66
5.51
5.35
1.52
0.702974
3.41
5.82
5.66
5.51
5.35
1.56
0.702995
3.62
5.82
5.66
5.50
5.35
1.56
0.702996
3.64
5.81
5.65
5.49
5.33
1.53
0.703000
3.64
2.93E-02
2.75E-02
1.81E-02
1.10E-02
D
1200oC
Initial melt
2.40
2.78
3.16
5.43
6.79
3.20
5.22
5.05
3.50
5.14
5.17
5.44
5.05
6.29
5.83
5.92
5.70
5.19
5.81
6.18
6.05
5.34
6.00
5.84
5.50
5.35
5.43
0.98
0.7025
0
Plag
[m /sec]
8.09E-19
3.10E-21
3.10E-21
3.10E-21
3.10E-21
4.38E-16
3.87E-19
3.87E-19
1.41E-17
3.37E-17
2.92E-19
3.10E-21
3.10E-21
2.92E-19
2.92E-19
3.37E-17
3.10E-21
3.24E-19
5.06E-19
A) 200
1110
150
D)
2
1.5
100
1
50
B)
700
1079
0.5
600
E)
500
2800
400
2400
300
C)
0
3200
200
2.8
2.4
10.8
4.5
3.8
2000
1600
0.5
2
1.5
2
2.5
3
3.5
median TiO28
1.6
1.2
0.8 median TiO28
0.5
1
1.5
1
Supplementary Information Figure 1
2
2.5
3
3.5
70
GI 92-10 (MgO 5.31)
50
SG 93-15 (MgO 7)
SG 93-5 (MgO 10.2)
30
SG 93-7 (MgO 11.84)
20
10
8
6
4
2
Cs
Ba
Rb
Ta
U
Th
Nb
La
K
Pb Sr
Hf Eu Gd Dy
Ho Tm
Lu
P
Ce Pr
Nd
Zr Sm Ti
Tb
Er
Yb
Y
Supplementary Information Figure 2
10
98
95
90
80
1
0
0.5
1
1.5
2
Nb/Nb*(La)
Supplementary Information Figure 3