Earth and Planetary Science Letters 258 (2007) 61 – 72
www.elsevier.com/locate/epsl
Discrete plumbing systems and heterogeneous magma sources of
a 24 km 3 off-axis lava field on the western flank of
East Pacific Rise, 14 °S.
Nobuo Geshi a,⁎, Susumu Umino b , Hidenori Kumagai c , John M. Sinton d ,
Scott M. White e , Kiyoyuki Kisimoto a , Thomas W. Hilde f
a
Geological Survey of Japan, AIST, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8567, Japan
Department of Biology and Geosciences, Graduate School of Science, Shizuoka University, 836, Ohya, Suruga-ku,
Shizuoka, 442-8529, Japan
Institute for Research on Earth Evolution, Japan Agency for Marine-Earth Science and Technology (JAMSTEC), 2-15,
Natsushima, Yokosuka, Kanagawa 237-0061, Japan
d
Department of Geology and Geophysics, University of Hawaii, 1680 East-West Road, Honolulu, HI 96822, USA
e
Department of Geological Science, University of South Carolina, Columbia, SC 29208, USA
f
Department of Geology and Geophysics, Texas A and M University, College Station, TX 77843-3115, USA
b
c
Received 2 August 2006; received in revised form 7 March 2007; accepted 7 March 2007
Available online 16 March 2007
Editor: R.W. Carlson
Abstract
The largest known mid-ocean-ridge off-axis lava flow field occurs slightly off-axis near 14 °S along the East Pacific Rise
(EPR). It comprises at least four volcanologically discrete units. We collected lava samples from 24 sites within the off-axis lava
field and 7 sites on the adjacent ridge crest using the Shinkai-6500 submersible. The lava field comprises at least three distinct
MORB compositions, all of which are different from the lavas collected from the nearby ridge axis. The east and west cones and
the northern lobe of the lava field consist of normal MORB (N-MORB) lavas with a low concentration of incompatible elements
and low LILE/HFSE and LREE/HREE ratios. By contrast, the samples from the west plain of the field have a higher concentration
of incompatible elements and higher LILE/HFSE and LREE/HREE ratios indicating T-MORB character. The lava samples
collected from the summit of the east cone show the highest concentration of LILE elements and LREEs among the lava field (EMORB lava). The N-MORB of the off-axis lava field are more depleted in incompatible elements than the adjacent EPR axis lavas,
possibly reflecting the re-melting of the residual mantle in the off-axis region. The E-MORB lava was probably derived from fertile
mantle that did not undergo melting beneath the spreading center. T-MORB, which occupies the main part of the 14 °S lava field, is
the product of magma mixing between N-MORB and E-MORB magmas.
© 2007 Elsevier B.V. All rights reserved.
Keywords: East Pacific Rise; off-axis volcano; mid-ocean ridge; lava flow; MORB; Shinkai 6500
⁎ Corresponding author. Tel.: +81 298 61 3599; fax: +81 298 56 8725.
E-mail address: [email protected] (N. Geshi).
0012-821X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsl.2007.03.019
62
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
1. Introduction
Recent surveys along the mid-ocean ridge system
have revealed the existence of many “off-axis” lava
fields close to the East Pacific Rise (EPR), the fastest
spreading center on the earth [1–4]. The presence of
these off-axis fields indicates that the magmatic
activity of the mid-ocean ridge system is not limited
to the ridge axis but is spread over several kilometers
extending away from the spreading center and
includes off-axis volcanism. Off-axis volcanic activity
adds to the pile of lava layers at the off-axis area.
More than half of the extrusive layer of the EPR was
possibly emplaced in the off-axis region [5,6],
suggesting the importance of off-axis magmatism in
the development of fast-spread ocean crust. Understanding emplacement mechanisms and the petrology
and geochemistry of off-axis lava flows is therefore
essential to comprehensively understand the evolution
of the oceanic crust and the magmatic plumbing
systems in fast-spreading ridges.
To investigate the off-axis magma plumbing
process in the EPR region, we surveyed the distribution of an off-axis lava flow field on the western flank
of the EPR 14 °S region and collected lava samples
from it during the NIRAI-KANAI Cruise (YK04-07)
in July–August 2004 using the R/V Yokosuka and
Shinkai 6500 submersible of JAMSTEC [7]. We
conducted four dives on the 14 °S lava flow field and
two dives on a nearby ridge crest ca. 30 km south of
the 60 center of the 14 °S lava field (Fig. 1). In
addition, we conducted one submersible dive on an
off-axis seamount at 16 °S. These dives provided the
first bottom observations of large off-axis lava field
and the volcano-tectonic features of the ridge axis in
this region. In this paper, we present the results of
the petrological examinations of the 14 °S lava flow
field.
2. Geographical and geological features
The 14 °S off-axis lava flow field occurs along the
western flank of the EPR, 2–19 km from the nearby
ridge axis. The lava field spreads from 13°40′S to
14°07′S and covers an area of approximately
49 km × 16 km (Fig. 1A, B). The total area of the lava
field is 420 km2 and its volume exceeds 24 km3. Sidescan sonar images reveal the 14 °S lava field is
recognized as a region of higher sonar backscatter as
compared to the surrounding seafloor (Fig. 1C) [7]. This
indicates that a thinner cover of soft sediment exists on
this field.
The lava flow field can be divided into four volcanic
domains based on topographic character: the east cone,
west plain, west cone, and northern lobe (Fig. 1A, B).
The east cone is a volcano with flat top and surrounded
by arcuate terraces. The summit of the east cone is
2.5 km from the present spreading axis. The volcano is
approximately 6 km across at the base and 350 m in
height. Small knolls, several tens of meters high, exist
on the flat top of the volcano. The surface of the cone is
covered with pillow lava running down from the
summit. Lavas erupted from the east cone cover an
area of approximately 72 km2 and the estimated volume
of the east cone is 7.2 km3.
The west plain comprises the central portion of the
14 °S off-axis lava field, covering an area of
approximately 251 km2, which is larger than that of
the previously largest-known off-axis lava on the EPR
[8]. Assuming the average thickness of the lava of the
west plain to be 50 m, the total volume is estimated to
be 12.6 km3. An alignment of the low ridges running
along the NNE–SSW direction occurs along the
highest region of the west plain. From the topographic
character and flow structure of the lava, this ridge is
likely to be an eruptive fissure of the central lava flow.
The main part of the west plain is characterized by
smooth-surface sheet flows with many collapse pits
and fissures.
The west cone is a circular flat-top seamount with a
diameter of 4 km and height of nearly 350 m with
smaller satellite cones aligned parallel to the present
ridge axis. The lava flows from the west cone are best
exposed at the western foot of the west cone and the
eastern half is probably covered by the lavas of the west
plain. The lavas of the west cone and the minor satellite
cones are exposed over an area of about 74 km2. The
total volume of the west cone is estimated to be
approximately 4 km3. The sides of the west cone were
covered by elongated pillow lava running down the
slope. The sonar backscatter is the weakest on the west
cone of the 14 °S off-axis lava field (Fig. 1C). This
suggests that the west cone has the thickest sediment,
and is likely the oldest part of the 14 °S off-axis lava
field.
The northern lobe, which is characterized by a shieldlike rise, is 7.5 km long and 4 km wide with a narrow
branch extending along a graben to the north. The
northern lobe covers an area of 22.3 km2. Its volume is
estimated to be 2.1 km3. In the 14 °S lava field, the sonar
backscatter is the highest for the northern lobe (Fig. 1C),
which suggests that this region might be the youngest
part of the lava field. The surface of the northern lobe is
covered by pillow lava.
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
63
Fig. 1. (A) Bathymetry map of the 14 °S off-axis lava field. Counter interval is 50 m. The outlines of the each flow unit of the off-axis lava are shown
by black line. WC: west cone, WP: west plain, EC: east cone, NL: northern lobe. Broken lines show the submersible track of each dive and the
numbers attached each line are dive number. (B) Relief map of the 14 °S off-axis lava field. (C) Backscatter intensity image of the side scan sonar of
the 14 °S off-axis lava field. Higher intensity is shown as darker area.
64
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
In order to estimate the emplacement age of the flow
field, sediment thickness was measured along the survey
track by the sub-bottom sonic profiler of GSJ/AIST
attached to the Shinkai 6500 submersible. The seafloor age at the edge of the lava field is estimated as
120 ± 5 ka, based on a linear interpolation of the local
plate spreading rate between the ridge axis and the edge
of the Jaramillo chron, where it is covered with 2 m of
sediment. Based on the thickness of the sediment and
the age of the basement, we estimate the average
sedimentation rate in this area as approximately
0.017 m/kyr. Using this sedimentation rate we calculate
an emplacement age of the lava of the west plain that is
covered with 0.3–0.4 m of sediment to be 18–24 ka.
Thinner sediment cover on the east cone and north lobe
suggests even younger ages for these areas.
3. Sampling and methods
We collected lava samples from 24 sites within the
lava field and 7 sites on the ridge crest in the 14 °S area.
Also, we conducted one submersible dive on an off-axis
seamount near 16 °S where 6 lava samples were
collected. The tracks of submersible dives in the 14 °S
area are shown in Fig. 1A. One or two pieces of lava were
collected from each sampling site. The whole-rock
compositions of the lavas were determined at the
Geological Survey of Japan/AIST using X-ray fluorescence (XRF) and inductively-coupled plasma mass
spectrometry (ICP-MS). Rock chips (ca. 50 g) were
ultrasonically cleaned and then soaked in distilled water
at 70 °C for 4 to 5 days to remove seawater contamination. The rock chips were then pulverized using an agate
mortar. The major elements were analyzed by a Philips
PW1404 spectrometer using glass beads prepared by the
fusion of 0.4 g of the samples and lithium tetraborate in the
ratio of 1:10 [9]. In general, the external error and accuracy in the process were< 2% for the major elements. The
concentrations of trace elements and rare-earth elements
(REE) were analyzed by ICP-MS on the Micromass
Platform ICP using a VG Plasmaquad PQ2+ instrument.
Approximately 200 mg of powder was dissolved in HFHNO3. After evaporation to dryness, the samples were redissolved in 2% HNO3 [10]. The reproducibility is better
than ± 4% (2 S.D.) for the REE, Rb, and Nb and better
than ± 6% (2 S.D.) for other elements. These results are
reported in supplementary Table 1.
The strontium and helium isotopic compositions for
representative samples were determined. A VG sector
mass spectrometer at the Geological Survey of Japan/
AIST was used for the analysis of the strontium isotopic
composition according to the procedure described by
Nakajima et al. [11]. The 87Sr/86Sr ratios were normalized to 86Sr/88Sr = 0.1194. Repeated analyses of the
NBS987 standard during this study gave a 87Sr/86Sr ratio
of 0.71025 ± 0.00002 (1σ). The helium isotope compositions were determined at IFREE/JAMSTEC using both
crushing and powder melting gas-extraction techniques
with a GVI5400 noble gas mass spectrometer. Typically,
0.2–1.0 g of glass chunks were selected and ultrasonically cleaned with 1% HNO3, acetone, ethanol, and deionized water to remove surface contamination. After the
crushing gas extractions, the crushed samples were
retrieved and wrapped in Al-foil for melting gas
extractions (powder melting). The helium amounts were
calibrated by repeated measurements of diluted air whose
volume was also calibrated by EB-1 radiogenic argon
standard [12]. The measured 3He/4 He ratios were
calibrated by HESJ helium gas standard and corrected
for instrumental mass discrimination [13]. All the
reported data were blank corrected. Further instrumental
details are reported in [14].
4. Results
4.1. Petrography
All the rocks are relatively unaltered in hand sample
and thin section, although minor weathered skins are
observed on some glassy surfaces. A major portion of
the lava of the west plain is almost aphyric with a very
minor volume of plagioclase phenocryst of up to 1.5 mm
in length. Some plagioclase-phyric lavas were found at
the east and west cones. These contain up to 5 vol.% of
plagioclase crystals (An = 69–83) several mm in length.
Further, minor olivine microphenocrysts (Fo = 80–89)
less than 1 mm in length are present.
4.2. Whole-rock compositions
4.2.1. Off-axis lava
The samples collected from the off-axis lava field
have whole-rock MgO contents ranging from 6.5 to
8.8 wt.% and whole-rock FeO⁎/MgO ratios ranging
from 1.0 to 2.0. The 14 °S off-axis lavas can be divided
into distinct compositional groups, corresponding to the
morphological divisions of the lava field described
earlier (Fig. 2). The lava of the west plain ranges from
6.5 to 7.8 wt.% MgO, while that of the east and west
cones ranges from 7.2 to 8.8 wt.% MgO, showing that
the lava of the west plain is more fractionated compared
to the east and west cones. It appears that the 14 °S offaxis lavas can be divided into three groups based on the
whole-rock K2O contents: low potassium N-MORB,
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
65
Fig. 2. Whole rock compositions of the 14 °S off-axis lavas and ridge crest lava plotted against whole-rock MgO content. Dashed lines show the range
of the reported composition of the EPR ridge-axis lava [9,10]. Compositional range of N-MORB, T-MORB and E-MORB groups of 14 °S off-axis
lava are shown in TiO2, CaO and K2O diagrams.
higher potassium T-MORB, and very high potassium EMORB (Fig. 2). The N-MORB group has higher CaO
and lower TiO2 and K2O concentrations at a given MgO
concentration (CaO = 11.3–12.5 wt.%, TiO2 = 1.1–
1.6 wt.%, K2O = 0.06–0.16 wt.%) and was found on
the east cone, west cone, and northern lobe. The TMORB group was found on the west plain and has lower
CaO and higher TiO2 and K2O concentrations at a given
MgO concentration (CaO = 10.7–11.2 wt.%,
TiO2 = 1.8–2.3 wt.%, K2O = 0.16–0.21 wt.%). Samples
828-5 and 828-6 collected from the summit of the east
cone comprise the E-MORB group having the highest
K2O content (0.46 wt.%) among the 14°S lava samples.
These three groups reveal discrete compositional trends
that are unlikely to have evolved from a common
parental magma by fractional crystallization of plagioclase and olivine.
N-MORB lavas exhibit a lower concentration of
REEs compared to the T-MORB (Fig. 3A). Primitive
mantle-normalized abundances of REEs in the NMORB lavas are less than 9, whereas the corresponding
HREE values in the T-MORB lava exceed 9. All the offaxis lava samples, except those of the E-MORB lavas
(samples 828-5 and 828-6), show LREE-depleted
patterns (Fig. 3). The range of primitive mantlenormalized La/Sm ratios is 0.73–1.01 for the NMORB lavas (except for 828-1) and 0.95–1.11 for the
T-MORB. The E-MORB lavas exhibit a LREE-enriched
character with La/Sm ratio of 2.1.
4.2.2. Ridge-axis lava
The lava samples collected from the ridge axis
adjacent to the 14 °S off-axis lava field have whole-rock
MgO contents ranging from 7.1 to 7.5 wt.% (Fig. 2), and
66
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
Fig. 3. A: Primitive mantle normalized trace element abundances of the 14 °S off-axis lava, B: 14 °S ridge-axis lava, C: trace element abundance of
the 14 °S off-axis lava normalized by the averaged composition of the 14 °S ridge-axis lavas.
have relatively high concentrations of incompatible
elements – especially K, Rb, and Nb – compared to the
reported N-MORB lavas of the EPR axis [15] (Fig. 2).
Zr/Nb ratios of the ridge-axis lavas near 14 °S range
from 13.7 to 34.7, suggesting a T-MORB-like character
(Supplementary Table 1). Within the 14 °S region, the
concentrations of Ti, Y, and Zr in the ridge-crest lavas
are comparable to those in the N-MORB of the east and
west cones, and lower than those of the T-MORB of the
west plain. The concentrations of K, Ba, Rb, and Nb in
the ridge-crest lavas of the 14 °S region are higher than
those in the N-MORB lavas of the east cone and similar
to or higher than those in the T-MORB lavas of the west
plain. The Nb/Y and K/Ti ratios in all the ridge lava
samples, except for sample 830-1, are higher than those
in the T-MORB of the west plain. The primitive mantle
normalized REE concentrations in the ridge-axis lavas
range from 5.7 to 10.6, except for sample 830-1 (Fig. 3B).
The primitive mantle normalized La/Sm ratios in the
ridge-axis lavas, except for sample 830-1, range from
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
67
1.05 to 1.36, which is higher than the average value for
the 14 °S off-axis lavas (Supplementary Table 1).
4.2.3. Isotopic compositions
The whole-rock 87Sr/86Sr ratio of six representative
lava samples of the 14°S off-axis lava field ranges
between 0.70241 and 0.70269 (Supplementary Table 1).
The E-MORB sample (828-6) shows higher 87Sr/86Sr
ratio (= 0.70269) compared to the other N-MORB and TMORB samples (< 0.70256) of the 14°S off-axis lava
field. The 87Sr/86Sr ratios of the two T-MORB lavas of
the west plain (832-5 and 833-2) are 0.70249 and
0.70252, significantly higher than that of the two NMORB lavas (833-1 from the east cone and 833-7 from
the west cone). Samples 828-1 and 834-4, which were
collected from the east cone and the north lobe with
higher K, Rb and LREE content, has higher 87Sr/86Sr
ratio (= 0.70256) than other N-MORB lava. The
87
Sr/86Sr ratio of the off-axis lava show good negative
correlation with Ti/K, Zr/K and Y/K ratios (Fig. 4).
The volatiles extracted from glass collected from the
chilled margins of the lava are fairly abundant in helium:
up to 3 × 10− 6 cm3 STP/g by crushing and 1 × 10− 5 cm3
STP/g by melting. The helium gas concentration is fairly
high in the case of the N-MORB lavas (e.g. [17]). All six
glass samples collected in the off-axis lava field exhibit
3
He/4He ratios that are 8.13–8.98 times the atmospheric
value (RA) (Table 1), and the ridge-axis sample (830-2)
has 3He/4He ratios of 8.74 RA. These values are well
within the global variation of MORB and below the
values observed at some high-He hotspots [18]. Except
for the E-MORB sample (828-6), no clear sign of the in
situ ingrowth of 3He/4He ratio was obtained from powder
melting gas extraction. This appears to be consistent with
fairly young age of the lava field, ca. 20 ka, obtained from
sediment thickness calculation. On the surface of sample
828-6, an abundant reddish-brown colored rusty adherent
and apparent devitrification were observed; this suggests
that the glass chunk of 828-6 under study was subjected
to hydration, which resulted in the loss of He and
seawater contamination. Both seawater contamination
and helium loss prior to the in situ ingrowth of 4He
decrease the 3He/4He ratio; however, the accumulation of
4
He in 20 kyr does not exceed 1 × 10−9 cm3 STP/g for 0.8
ppm Th. This accumulation is not sufficient to decrease
the 3He/4He ratio of sample 828-6, extracted by the
powder melting method, from 8.8 RA to 8.1 RA.
4.2.4. Comparison with other off-axis lavas of EPR
Many lava compositions of the off-axis volcanoes in
the EPR region have been reported [19–21]. Although
the compositions of these volcanoes show wide
Fig. 4. Whole-rock Ti/K, Zr/K and Y/K ratios of the representative
samples of the off-axis lava field plotted against their whole-rock
87
Sr/86Sr ratio. Filled circle: N-MORB, open square: T-MORB; gray
circle: E-MORB.
variations from N-MORB to E-MORB, the Ti, K, Rb,
Y, and Zr concentrations in most of the off-axis
volcanoes are comparable to those in the N-MORB
lavas and lower than those in the T-MORB lavas at a
given MgO concentration [19]. The major element
compositions of the 8°17S lava [16], which is another
example of the large off-axis lava field at EPR, are
similar to those of the N-MORB lavas of the 14 °S offaxis field (Fig. 2). One of the important characteristics of
he 8 °S off-axis lava is compositional uniformity: a
major part of the 8 °S lava consists of N-MORB-like
lava although up to four distinct compositional groups
can be recognized [16]. The 14 °S lava is, by contrast,
complex containing N-MORB, T-MORB, and EMORB lavas. While the 8 °S lava exhibits a compositional similarity to the adjacent EPR ridge lava [16]; this
is a clearly not the case with the lavas from the west
plain of the 14 °S lava field, which have higher Ti and K
concentrations than the adjacent axial lavas.
68
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
5. Discussion
5.1. Multiple plumbing systems
The co-existence of different types of volcanoes,
including conical volcanoes and flat lava, within the
14 °S lava field indicates that the lava field does not
comprise a single flow but rather is the product of at
least four different erupted units, each with its own
discrete vents and with different effusion styles. The
west and east cones and the northern lobe consist of a
pile of pillow lava, suggesting the lava extraction at a
low effusion rate. The west plain, which forms a major
portion of the off-axis lava field, has a low-angle shieldlike shape and is characterized by a widely distributed
sheet flow, suggesting that a short eruption occurred
with a high effusion rate [22,23].
The eruption of magma more than 10 km3 within a
short period for the west plain requires the storage of all
the erupted magma prior to the eruption. The evolved
compositional character of the lava of the west plain
(MgO = 6.5–7.8 wt.%) supports the existence of magma
storage which allowed crystal fractionation to occur. In
contrast, the continuous low effusion rate eruption for the
west and east cones does not require storage and
accumulation of a large magma reservoir prior to
eruption. The less-evolved petrological character of
these lavas (MgO > 7.2 wt.%) is consistent with the
absence of the salient evidence for crystal fractionation
process within the magma storage. Sinton et al. [24]
reported a similar relationship between the lavas from the
fissure eruptions and shield volcanoes in Iceland
volcanoes: lavas that erupt from fissures with a high
effusion rate have a more evolved composition than that
of shield volcanoes with a lower effusion rate, indicating
the presence of a large magma storage in the crust during
the fissure eruption. In the 14 °S region of SEPR, discrete
plumbing systems and probably separate events combined to form the off-axis lava complex shown in Fig. 1.
5.2. Depleted and heterogeneous sources
All the lava samples collected from the 14 °S off-axis
lava field are compositionally distinct from those
collected from the adjacent EPR axis, indicating that
the source of the 14 °S off-axis lava is independent from
the magma system beneath the EPR. The N-MORB and
T-MORB lavas that comprise the major portion of the
14 °S off-axis lava field, are characterized by lower
LILE/HFSE element ratios and LREE/HREE ratios, as
Table 1
Summary of helium measurement of 14°S off-ridge lava, EPR
Sample
Locality
Rock type
Extraction
[4He] (μccSTP/g)
3
Off-Axis
828-1
East cone
N-MORB
828-6
East cone
E-MORB
832-5
West plane
T-MORB
833-2
West plane
T-MORB
833-7
West cone
N-MORB
834-4
North lobe
N-MORB
Crush × 30
Powder melting
Total
Crush × 30
Powder melting
Total
Crush × 100
Powder melting
Total
Crush × 100
Powder melting
Total
Crush × 100
Powder melting
Total
Crush × 100
Powder melting
Total
0.89
4.54
5.43
0.11
1.61
1.72
2.23
2.16
4.39
0.75
4.06
4.81
1.95
1.27
3.22
2.17
7.46
9.64
8.92 ± 0.11
8.76 ± 0.07
8.79 ± 0.06
8.79 ± 0.10
8.09 ± 0.08
8.13 ± 0.07
9.12 ± 0.06
8.84 ± 0.07
8.98 ± 0.05
9.22 ± 0.10
8.55 ± 0.09
8.66 ± 0.08
8.56 ± 0.03
8.82 ± 0.05
8.66 ± 0.03
8.51 ± 0.03
8.80 ± 0.04
8.74 ± 0.03
On-Axis
830-2
Ridge axis
N-MORB
Crush × 30
Crush × 100
Powder melting
Total
0.75
0.39
4.39
5.54
8.84 ± 0.03
8.91 ± 0.05
8.72 ± 0.04
8.74 ± 0.03
Note: analytical uncertainties are generally 2% helium concentration, respectively.
He/4He(R/RA)
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
compared to the lavas collected from the nearby EPR
spreading axis during our cruise (Fig. 3). The depleted
compositional character of the 14 °S off-axis lavas is
consistent with formation by remelting of the residual
mantle previously melted beneath the spreading center.
Depleted off-axis lavas were also reported in the
northern EPR region [25] and Oman ophiolite [26];
and they suggest that the remelting of the residual
mantle may be a common source of off-axis lava. The EMORB lavas (samples 828-5 and 828-6), collected from
the summit of the east cone, with extremely high K and
Rb contents and enriched LREE, were most probably
derived from a relatively undepleted mantle that did not
undergo melting beneath the spreading center. The
difference of the strontium isotope composition between
the N-MORB and E-MORB (Fig. 4) also indicate that
they are derived from discrete mantle sources. The
existence of E-MORB lavas in the 14 °S off-axis lava
field indicates compositional heterogeneity within the
source mantle, as proposed in other EPR areas
[15,16,19–21].
The whole-rock composition of the aphyric lavas of
the N-MORB, T-MORB and E-MORB of the 14°S lava
field are plotted on a normative olivine–plagioclase–
quartz ternary projection [28] to estimate the melting
pressure [29] (Fig. 5). Lavas with evolved composition
(FeO⁎/MgO > 1.5) are excluded from the plot to avoid
the effect of crystal fractionation. Both N-MORB and TMORB lavas of the 14°S lava field plot close to the
69
isobaric line of 1 GPa of Hirose and Kushiro [29]
suggesting that there is no significant difference of
melting pressure for these lavas (Fig. 5).
The compositional heterogeneity between the ridgeaxis lava and off-axis lava in the 14 °S area and the
range of the helium isotope variations of these lavas
over the area rule out the possibility of the contribution
of a deeper mantle source such as a hot-spot-like
component, in contrast to the samples of the 17 °S
region [27]. These results suggest that the origin of the
magma comprising the lava field is essentially similar
to or well within the normal mid-ocean ridge system,
and the compositional diversity within this area represents the heterogeneity within the mid-oceanic ridge
system.
5.3. Origin of T-MORB lava
The T-MORB lava of the west plain is characterized
with higher concentration of K, Rb and LREE,
comparing to the N-MORB lava of the east cone, west
cone and northern lobe. The T-MORB lavas also have
higher 87Sr/86Sr ratio than that of the N-MORB lavas.
This lava composition can be modeled by mixing 10–
20% of E-MORB magma, such as 828-6, into N-MORB
magma (Fig. 6). The strontium isotope compositions
also support the idea that the T-MORB magma of the
west plain can be formed by the mixing of the depleted
N-MORB magma and enriched E-MORB magma (Fig.
Fig. 5. Whole-rock composition of the 14 °S off-axis lavas plotted on a normative olivine–plagioclase–quartz ternary projection. Isobaric lines are
after Hirose and Kushiro [29].
70
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
Fig. 6. Incompatible element concentration plotted against the K2O concentration. Symbols are in Fig. 2. Broken lines show the mixing relationship
between the averaged composition of N-MORB lava (K2O < 0.1 wt.%: gray star) and E-MORB lava (828-5 and 6; filled star) of the east cone. Open
star shows the assumed composition of the parental magma of T-MORB group.
4). Some lava samples (828-1 and 828-2) collected from
the east cone also have T-MORB like character and they
also can be formed by mixing of E-MORB magma with
N-MORB magma (Fig. 6).
Relatively constant incompatible element ratios, such
as K, Ti, Zr, Y and Nb, among the T-MORB lavas
suggest homogenization during magma storage. We
suggest that N-MORB and E-MORB magmas were
generated in discrete source regions, gathered into a
shallow reservoir, and mixed before eruption. Storage of
magma in a shallow reservoir also promotes cooling and
fractional crystallization. Evolved composition of the TMORB lava of the west plain, with low MgO
concentration (6.5– 7.8 wt.%), indicates olivine fractionation during their storage. Mass balance calculation
indicated that crystal fractionation of 4.9 wt.% of
plagioclase (An = 77) and 1.6 wt.% of olivine (Fo = 88)
from the hybrid magma (85% of the magma with the
averaged composition of the N-MORB lavas of the east
cone and 15% of the E-MORB with 828-6 composition:
shown in Fig. 6 with open star) can produce a magma
with averaged composition of the T-MORB lavas of the
west plain.
5.4. Comparison with the 8 °S off-axis lava
The only off-axis lava field comparable to the 14 °S
field described here is that at 8 °S [16]. The 8 °S lava
field is predominantly composed of sheet lava with a
flat surface that erupted from a vent on the eastern
slope of the ridge rise within 2.5 km from the spreading
center [8]. The 8 °S lava has a very low aspect ratio of
its topography with a smooth surface, suggesting a high
effusion rate during a short eruption like that of the
west plane of the 14 °S lava field. The petrological
character of the 8 °S lava is similar to that of the lavas
collected from the nearby ridge crest [16]; this suggests
that the 8 °S lava might have been directly supplied
from a magma reservoir beneath the spreading ridge
center. The 14 °S lava, by contrast, has a more depleted
composition as compared to those of the adjacent
ridge-crest lavas similar to the lavas of many off-axis
seamounts and flows reported in the EPR region. This
indicates that the plumbing system of the 14 °S off-axis
volcano is disconnected from the magma system of the
spreading center, although the eruption occurred only
3 km from the ridge center. Seismic surveys along the
southern EPR indicate that the sub-axial melt lens is
typically< 1.5 km wide [30] and this is consistent with
the idea that the roots of the 14 °S off-axis flow was
disconnected to the magma system of the adjacent
ridge axis.
One of the prominent differences between the 8 °S and
14 °S lava flow field is the geologic setting. Location of
the 14 °S off-axis flow field coincides with the
intersection of the EPR axis and the Sojourn Seamount
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
Chain– Sojourn Ridge, one of the major intraplate ridges
in the EPR region [7]. Upwelling of the mantle return
flow has been suggested as the cause of the volcanic
activity of Sojourn Seamount Chain– Sojourn Ridge
[7,31]. Fragments of fertile mantle carried by the
upwelling flow could produce the E-MORB magma
observed in the 14 °S off-axis lava. This upwelling would
also assist the remelting of the residual mantle to produce
N-MORB magma that is more depleted than that of the
adjacent spreading center. In this way, the mantle return
flow model also provides a mechanism to generate the
14 °S off-axis lava which is petrologically different from
that of the adjacent ridge axis. By contrast, no seamount
chain is associated with the 8 °S off-axis flow. The
petrological similarity between the 8 °S off-axis and its
adjacent spreading center [16] suggests that the source of
the off-axis lava flow is connected to the spreading center
and effect of any additional upwelling flow is negligible.
The magma source, plumbing system, and eruption style
are different for the two large off-axis lava fields in the
8 °S and 14 °S regions; this illustrates the diversity of the
near-axis lavas in a fast spreading system.
6. Conclusions
(1) The off-axis lava field discovered in the 14 °S area
of the EPR consists of at least four discrete areas
including the two volcanic cones and two flat lava
flows. This morphological difference among these
segments reflects the effusion rate and duration of
eruption.
(2) The 14 °S off-axis lava can be divided into three
discrete groups based on chemical composition:
N-MORB lava with low LILE/HFSE and LREE/
HREE ratios, comprising the east and west cones
and the northern lobe; T-MORB lava with high
LILE/HFSE and low LREE/HREE ratios comprising the west plain; and E-MORB lava with
extremely high concentrations of LREE and LILE
collected from the summit of the east cone.
(3) Depleted character of the N-MORB lavas in the
off-axis region, in comparison to the adjacent
ridge-axis lavas, suggests the remelting of mantle
previously melted beneath the EPR. A part of
magma was derived from a fertile mantle that
might have escaped melting beneath the ridge
axis. Helium isotope composition suggests no
contribution of hot-spot like deep source to these
magmas.
(4) The incompatible element concentration and the
strontium isotope composition indicate that the
mixing between N-MORB and E-MORB magmas
71
produced T-MORB magma. Crystal fractionation
within a shallow magma reservoir enhanced the
evolved character of T-MORB.
Acknowledgements
Our survey cruise was supported by Captain Sadao
Ishida and the crew of R/V Yokosuka and the operation
team of Shinkai 6500 submersible. The trace element
analysis using ICP-MS was supported by Dr. Osamu
Ishizuka and Ms. Miho Tanaka. The Sr isotope analysis
was supported by Dr. Takashi Nakajima. The insightful
reviews by R. W. Carlson and two anonymous reviewers
significantly improved the manuscript.
Appendix A. Supplementary data
Supplementary data associated with this article can
be found, in the online version, at doi:10.1016/j.epsl.
2007.03.019.
References
[1] D.S. Scheirer, K.C. Macdonald, Near-axis seamounts on the
flanks of the East Pacific Rise, 8 °N to 17 °N, J. Geophys. Res.
100 (1995) 2239–2259.
[2] R.T. Alexander, K.C. Macdonald, Small off-axis volcanoes on
the East Pacific Rise. Earth Planet, Sci. Lett. 139 (1996)
387–394, doi:10.1016/0012-821X(96)00028-3.
[3] S.M. White, K.C. Macdonald, D.S. Scheirer, M-H. Cormier, The
distribution of isolated volcanoes on the flanks of the Southern
East Pacific Rise, 15.3°–20 °S, J. Geophys. Res. 103 (1998)
30371–30384, doi:10.1029/98JB02791.
[4] S.M. White, C. Macdonald, M. Sinton, Volcanic mound fields on
the East Pacific Rise, 16–19 S: low effusion rate eruptions at
overlapping spreading centers for the past 1 Myr, J. Geophys
Res. 107 (2002), doi:10.1029/2001JB000483.
[5] E.E.E. Hooft, H. Schouten, R.S. Detrick, Constraining crustal
emplacement process from the variation in seismic layer 2A
thickness at the East Pacific Rise, Earth Planet, Sci. Lett. 142
(1996) 289–309.
[6] J.P. Canales, R.S. Detrick, S. Bazin, A.J. Harding, J.A. Orcutt,
Off-axis crustal thickness across and along the East Pacific Rise
within the MELT area, Science (1998) 1218–1221.
[7] S.M. White, S. Umino, S. Kumagai, Transition from seamount
chain to intraplate volcanic ridge at the East Pacific Rise,
Geology 34 (2006) 293–296.
[8] K.C. Macdonald, R. Haymon, A. Shor, 220 km2 recently erupted
lava field on the East Pacific Rise near lat 8 °S, Geology 17
(1989) 212–216.
[9] S. Togashi, S. Terashima, The behavior of gold in unaltered
island arc tholeiitic rocks from Izu-Oshima, Fuji, and Osoreyama
volcanic areas Japan, Geochim. Cosmochim. Acta 61 (1997)
543–554.
[10] O. Ishizuka, N. Taylor, J.A. Milton, R.W. Nesbitt, M. Yuasa, I.
Sakamoto, Variation in the mantle sources of the northern Izu arc
with time and space-constraints from high-precision Pb isotopes,
J. Volcanol. Geotherm. Res. 156 (2006) 266–290.
72
N. Geshi et al. / Earth and Planetary Science Letters 258 (2007) 61–72
[11] T. Nakajima, T. Shirahase, K. Shimata, Along-arc lateral variation
of Rb-Sr and K-Ar ages of Cretaceous granitic rocks in Southwest
Japan, Contrib. Miner. Petrol. 104 (1990) 381–389.
[12] N. Iwata, Geochronological study of the Deccan volcanism by
the 40Ar–39Ar method. PhD dissertation, University of Tokyo,
(1997) pp. 168.
[13] J. Matsuda, T. Matsumoto, H. Sumino, K. Nagao, J. Yamamoto,
Y. Miura, I. Kaneoka, N. Takahata, Y. Sano, The 3He/4He ratio of
the new internal He Standard of Japan (HESJ), Geochem. J. 36
(2002) 191–195.
[14] T. Hanyu, K.T.M. Johnson, N. Hirano, Z.-Y. Ren, Noble gas and
geochronology study of the Hana Ridge, Haleakala volcano,
Hawaii; implications to the temporal change of magma source
and the structural evolution of the submarine ridge, Chem. Geol.
238 (2007) 1–18, doi:10.1016/j.chemgeo.2006.09.008.
[15] J.M. Sinton, S.M. Smaglik, J.J. Mahoney, K.C. Macdonald,
Magmatic processes at superfast spreading mid-ocean ridge:
Glass compositional variations along the East Pacific Rise 13°–
23 °S, J. Geophys. Res. 96 (1991) 6133–6155.
[16] L.S. Hall, J.M. Sinton, Geochemical diversity of the large lava
field on the flank of the East Pacific Rise at 8°17′S, Earth Planet.
Sci. Lett. 142 (1996) 241–251.
[17] M. Moreira, J. Kunz, C. Allègre, Rare gas systematics in popping
rock: isotopic and elemental compositions in the upper mantle,
Science 279 (1998) 1178–1181, doi:10.1126/science.279.5354.
1178.
[18] D.W. Graham, Noble gas isotope geochemistry of Mid-Ocean
Ridge and Ocean Island Basalts: characterization of mantle
source reservoirs, in: D. Porcelli, C.J. Ballentine, R. Wieler
(Eds.), Noble gases—in geochemistry and cosmochemistry,
Reviews in mineralogy and geochemistry: Washington D.C.,
Mineral. Soc. Am., vol. 47, 2002, pp. 247–318.
[19] L.S. Hall, J.J. Mahoney, J.M. Sinton, R.A. Duncan, Spatial and
temporal distribution of a C-like asthenospheric component in
the Ramo Rahi Seamount Field, East Pacific Rise, 15°–19 °S,
Geochem. Geophys. Geosyst. 7 (2006) Q03009, doi:10.1029/
2005GC000994.
[20] Y. Niu, R. Batiza, Trace element evidence from seamount for
recycled oceanic crust in the Eastern Pacific mantle, Earth Planet.
Sci. Lett. 148 (1997) 471–483.
[21] Y. Niu, M. Regelous, J. Wendt, R. Batiza, J. O'Hara, Geochemistry of near-EPR seamounts: importance of source vs.
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
processes and the origin of enriched mantle component, Earth
Planet. Sci. Lett. 199 (2002) 327–345.
R.D. Ballard, R.T. Holcomb, T.H. van, lava lakes of the rift
valley, J. Geophys. Res. 84 (1979) 5407–5422.
R.W. Griffiths, J.H. Fink, Solidification morphology of submarine lavas: a dependence on extrusion rate, J. Geophys. Res.
97 (1992) 19729–19737.
J. Sinton, K. Grönvold, K. Sæmundsson, Postglacial eruptive
history of the Western Volcanic Zone, Iceland, Geochem. Geophys. Geosyst. 6 (2005) Q12009, doi:10.1029/2005GC001021.
M. Spiegelman, J.R. Reynolds, Combined dynamic and geochemical evidence for convergent melt flow beneath the East
Pacific Rise, Nature 402 (1999) 282–285, doi:10.1038/46260.
F. Einaudi, M. Godard, P. Pezard, J.J. Cochemé, C. Coulon, T.
Brewer, P. Harvey, Magmatic cycles and formation of the upper
oceanic crust at spreading centers: geochemical study of a
continuous extrusive section in the Oman ophiolite, Geochem.
Geophys. Geosyst. 4 (2003) 8608, doi:10.1029/2002GC000362.
M.D. Kurz, M. Moreira, J. Curtice, D.E. Lott III, J.J. Mahoney,
J.M. Sinton, Correlated helium, neon, and melt production on
the super-fast spreading East Pacific Rise near 17°S, Sci. Lett.
232 (2005) 125–142, doi:10.1016/j.epsl.2005.01.005.
D. Walker, T. Shibata, S.E. DeLong, Abyssal tholeiites from the
Oceanographer Fracture Zone II: phase equilibria and mixing,
Contrib. Mineral. Petrol. 70 (1979) 111–125.
K. Hirose, I. Kushiro, 1993, Partial melting of dry peridotites at
high pressures: determination of compositions of melts segregated from peridotite using aggregates of diamond, Earth Planet.
Sci. Lett. 114 (1993) 447–489.
C.H. Tong, W. Pye, J. Barton, S. White, C. Sinha, C. Singh,
W. Hobbs, S. Bazin, J. Harding, M. Kent, A. Orcutt, Asymmetric melt sills and upper-crustal construction beneath overlapping ridge segments: implications for the development of melt
sills and ridge crests, Geology 30 (2002) 83–86, doi:10.1130/
0091-7613.
W. Forsyth, N. Harmon, S. Scheirer, A. Duncan, Distribution of
recent volcanism and the morphology of seamounts and ridges in
the GLIMPSE study area: implications for the lithospheric cracking hypothesis for the origin of intraplate, non-hot spot volcanic
chains, J. Geophys. Res. 111 (2006) B11407, doi:10.1029/
2005JB004075.