PUBLICATIONS
Geochemistry, Geophysics, Geosystems
RESEARCH ARTICLE
Helium isotopic textures in Earth’s upper mantle
10.1002/2014GC005264
Key Points:
3
4
He/ He along ocean ridges carries
information about the upper mantle
flow field
Long-wavelength variations are
similar for Indian, Atlantic, and Pacific
ridges
Short length scale variations result
from melting of heterogeneous
mantle
Supporting Information:
ReadMe
AuxiliaryMaterial FS1-FS4
AuxiliaryMaterial Text1
AuxiliaryMaterial S1
Correspondence to:
D. W. Graham,
[email protected]
Citation:
Graham, D. W., B. B. Hanan, C.
Hemond, J. Blichert-Toft, and F.
Albarède (2014), Helium isotopic
textures in Earth’s upper mantle,
Geochem. Geophys. Geosyst., 15, 2048–
2074, doi:10.1002/2014GC005264.
Received 24 JAN 2014
Accepted 23 APR 2014
Accepted article online 27 APR 2014
Published online 30 MAY 2014
mond3, Janne Blichert-Toft4,
David W. Graham1, Barry B. Hanan2, Christophe He
4
and Francis Albarède
1
College of Earth, Ocean, and Atmospheric Sciences, Oregon State University, Corvallis, Oregon, USA, 2Department of Geological Sciences, San Diego State University, San Diego, California, USA, 3Domaines Oc
eaniques, Institut Universitaire Europ
een de la Mer, Place Nicolas Copernic, Plouzan
e, France, 4Laboratoire de G
eologie de Lyon, Ecole Normale Sup
erieure de
Lyon and Universit
e Claude Bernard Lyon 1, CNRS UMR 5276, Lyon, France
Abstract We report 3He/4He for 150 mid-ocean ridge basalt (MORB) glasses from the Southeast Indian
Ridge (SEIR). Between 81 E and 101 E 3He/4He varies from 7.5 to 10.2 RA, encompassing more than half the
MORB range away from ocean island hot spots. Abrupt transitions are present and in one case the full range
occurs over 10 km. Melting of lithologically heterogeneous mantle containing a few percent garnet pyroxenite or eclogite leads to lower 3He/4He, while 3He/4He above 9 RA likely indicates melting of pyroxenitefree or eclogite-free mantle. Patterns in the length scales of variability represent a description of helium isotopic texture. We utilize four complementary methods of spectral analysis to evaluate this texture, including
periodogram, redfit, multitaper method, and continuous wavelet transform. Long-wavelength lobes with
prominent power at 1000 and 500 km are present in all treatments, similar to hot spot-type spectra in Atlantic periodograms. The densely sampled region of the SEIR considered separately shows significant power at
100 and 30–40 km, the latter scale resembling heterogeneity in the bimodal distribution of Hf and Pb
isotopes in the same sample suite. Wavelet transform coherence reveals that 3He/4He varies in-phase with
axial depth along the SEIR at 1000 km length scale, suggesting a coupling between melt production,
3
He/4He and regional variations in mantle temperature. Collectively, our results show that the length scales
of MORB 3He/4He variability are dominantly controlled by folding and stretching of heterogeneities during
regional (1000 km) and mesoscale (100 km) mantle flow, and by sampling during the partial melting
process (30 km).
1. Introduction
The origin, scale, and survival of mantle heterogeneities, and the link to mantle convection have been of
considerable interest for 50 years [e.g., Gast et al., 1964; Tatsumoto, 1978; Polve and Allègre, 1980; Allègre
et al., 1980; Zindler et al., 1982; Hart, 1984; White, 1985; Zindler and Hart, 1986; Allègre and Turcotte, 1986; Sun
and McDonough, 1989; Hofmann, 1997; Albarède, 2005; Anderson, 2006; Iwamori et al., 2010, and many
others]. During the past decade, our understanding of mantle geochemistry has grown tremendously,
mainly through the rich picture that is emerging by detailed mapping of isotopic variability in oceanic
basalts, which are the partial melting products of the mantle. Computational modeling of mantle dynamics
has also begun to quantify the range and distribution of geochemical variability produced during mantle
convection and tectonic plate recycling [Christensen and Hofmann, 1994; Kellogg et al., 2002, 2007; Brandenburg et al., 2008]. These models emphasize the need to quantify isotopic variability at localities such as midocean ridges and areas of hot spot-ridge interaction, and at length scales that range from dense sampling
within individual ridge segments up to regional (ocean basin) scales [Rubin et al., 2009]. There is clear evidence that stretching and folding of upper mantle material during convection [e.g., Olson et al., 1984] produces periodic patterns in mantle heterogeneities that are sampled by melting beneath mid-ocean ridges.
Indeed, the spatial patterns in MORB isotope compositions reflect blobs and streaks that originate from
both the deep mantle and from tectonically recycled crust and lithosphere [Gurnis, 1986a, 1986b; Agranier
et al., 2005; Meyzen et al., 2007; Graham et al., 2006; Hanan et al., 2013].
Magmas produced beneath mid-ocean ridges represent our best probes of upper mantle composition and
the conditions of melting. Spatial variations in melt generation beneath ridges linked to intrinsic variations
in mantle heterogeneity or to lateral variations in mantle temperature are further revealed by the relations
between basalt composition and physical attributes of ridges such as axial depth, ridge morphology, and
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2048
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Figure 1. Location map of basalt glasses from the Southeast Indian Ridge analyzed for He isotopes. Individual ridge segments, the Amsterdam-St. Paul (ASP) Plateau, and the AustralianAntarctic Discordance (AAD) are indicated. Red circles are the GEISEIR I and II sample suites from this study. Other samples in yellow have 3He/4He data reported in Graham et al. [1999,
2001], Mahoney et al. [2002], and Nicolaysen et al. [2007]. The base map was made using GeoMapApp (http://www.marine-geo.org/geomapapp), which is supported by the National Science Foundation and internal funding from Lamont-Doherty Earth Observatory.
the depths of seismically imaged magma lenses [Klein and Langmuir, 1987; Shen and Forsyth, 1995; Rubin
et al., 2009; Russo et al., 2009; Carbotte et al., 2013]. Despite the rich picture of upper mantle dynamics
emerging from detailed isotope studies of MORBs, the characteristic scales of mantle heterogeneity and
their linkage to mantle convection and melting remain poorly understood for large sections of the midocean ridge system. The best-characterized region to date is the northern Mid-Atlantic Ridge [Agranier et al.,
2005], especially where the MAR crosses the Iceland hot spot and its geologic features can be sampled at
very high resolution [Maclennan et al., 2003; Shorttle and Maclennan, 2011]. The fast-spreading East Pacific
Rise between 9 N and 10 N also stands in sharp contrast to the rest of the mid-ocean ridge system in
terms of its dense sampling, although the overall length sampled at high resolution is relatively short (100
km). For much of the mid-ocean ridge system, our appreciation of mantle heterogeneity and melting at
length scales shorter than ridge segmentation scales remains modest, mostly due to inadequate sampling
density.
This study describes 3He/4He variations along the intermediate spreading rate Southeast Indian Ridge
(SEIR), focusing primarily on the region between 81 E and 100 E. The SEIR contains significant along-axis
physical and chemical variability that can be used to constrain mantle melting and source lithology [e.g.,
Russo et al., 2009]. It also samples the Indian Ocean upper mantle, which is known to be isotopically distinct
from the mantle beneath the Pacific and North Atlantic [Dupre and Allègre, 1983; Hamelin et al., 1986; Hart,
1984; Dosso et al., 1988; Mahoney et al., 2002]. The SEIR stretches from the Rodrigues Triple Junction (25.6 S,
70.1 E) to the Macquarie Triple Junction (62 S, 151 E; Figure 1). Between 76 E and 78 E it crosses the
Amsterdam-St. Paul Plateau (ASP), a pronounced bathymetric swell associated with relatively hot mantle
upwelling beneath Amsterdam and St. Paul islands, while between 120 E and 128 E it crosses the
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2049
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Australian-Antarctic Discordance
(AAD), a region of deep bathymetry (>4000 m) associated with
2009-2010
relatively cold mantle and low
GEISEIR
12
Detailed Sampling
melt production. Notably, over a
distance of 2500 km, between
86 E and 120 E, there is a regular
10
eastward decrease in axial depth
from 2300 to 5000 m, and a mor8
phological transition from axial
high to axial valley due to
decreasing melt production rate
and crustal thickness. This depth
Australiangradient occurs at intermediate
2
Antarctic
Discordance
and uniform spreading rate (68–
3
75 mm/yr full rate) and in the
absence of large transform off4
Amsterdamsets and nearby mantle hot
St. Paul
spots. Axial depth and ridge morPlateau
phology encompass most of the
-1000
0
1000
2000
3000
Along-Axis Distance (km)
global range for spreading ridges
away from hot spots, making the
Figure 2. Variations along the Southeast Indian Ridge in 3He/4He (R/RA) and axial depth
SEIR a regional-scale analogue of
(km). Data sources are the same as for Figure 1. Analytical uncertainties are smaller than
the 60,000 km long global ocean
the size of the data points. The highlighted box outlines the GEISEIR I and II sampling
areas between 81 E and 100 E (n 5 175 sample locations, for segments K, L, and C13–
ridge system. Detailed geophysiC17). Along-axis distance in km is calculated from the pole of plate rotation (13.2 N,
cal surveys have revealed three
38.2 E) and is expressed relative to the location of St. Paul Island (38.70 S, 77.55 E).
axial morphological zones
between 81 E and 118 E. Most
ridge segments in the western zone (K, L, and C17–C13, which recently were resampled and are the focus
of the current study; Figure 1) have axial-high morphology, similar in dimension and shape to segments of
the fast-spreading East Pacific Rise [Cochran et al., 1997; Sempere et al., 1997; Baran et al., 2005]. Segments
C12-C4 to the east have a transitional morphology typical of other intermediate-spreading ridges.
Southeast Indian Ridge
Axial Depth (km)
3
He/4He (R/RA)
14
Earlier work by Graham et al. [1999, 2001] discovered localized peaks in 3He/4He (and Fe8, a proxy for the
depth of mantle melting) [Klein and Langmuir, 1987] having basal widths between 200 and 500 km that are
superimposed on a long-wavelength west-to-east decrease in 3He/4He (Figure 2). These helium isotope
peaks are most prominent near the ASP Plateau, and at 88 E, 97 E, and 111 E (at along-axis distances from
St. Paul island of 910, 1800, and 2850 km, respectively). The latter three peaks are located near the centers
of regional segmentation of the ridge defined by the persistence of fracture zone anomalies in satellite
gravity data. This zeroth-order segmentation is comprised of several coherent tectonic units, up to 900
km in length along-axis, that have been stable since reorientation of the SEIR when it migrated over the Kerguelen mantle plume at 38 Ma [Small et al., 1999]. Graham et al. [2001] used a spatial correlogram analysis
to evaluate the length scales of helium isotope variability along the SEIR. The correlogram revealed crude
structures in the He isotope variations at 150 and 400 km length scales which they interpreted to relate
to upper mantle heterogeneity and secondary convection scales. The new, dense sampling for the current
study by the French GEISEIR expeditions (GEochimie Isotopique de la SEIR) represents a suite of more
closely spaced (5–10 km) basalt glasses that stretch over a distance of more than 1600 km. This sampling
provides a unique opportunity to quantify 3He/4He variations at moderately high resolution, for more than
1500 km along this mid-ocean ridge spreading center where systematic geochemical variations are present.
2. Field Sampling and Analytical Methods
The GEISEIR I and II expeditions took place aboard the R/V Marion Dufresne, in austral summers of 2009
and 2010. GEISEIR I sampled the ridge axis between 88.5 E and 99.3 E (along-axis distance from St. Paul of
875–2010 km; Figure 1) at varying sample density. Between 88.5 E and 96 E the sampling density was
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2050
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Table 1. 3He/4He, He and CO2 Concentrations for Indian Ocean MORB Glasses
Sample
Segment
Location
GSR2 WC A1-1
GSR2 WC A1-2
GSR2 WC A1-3
GSR2 WC A1-5
GSR2 WC A1-5/6
GSR2 WC A1-6
GSR2 WC A1-7
GSR2 WC A1-8
GSR2 WC A1-9c
GSR2 WC A1-10
GSR2 WC A1-11
GSR2 WC A1-12b
GSR2 WC A1-13b
GSR2 WC A1-14
GSR2 WC A1-15
GSR2 WC A1-16
GSR2 WC A1-16/17
GSR2 WC A1-17
GSR2 WC A1-18
GSR2 WC A1-19
GSR2 WC A1-20
GSR2 WC A1-21
GSR2 WC A1-22
GSR2 WC A1-23
GSR1 WC-A2
GSR2 WC A2-1
GSR2 WC A2-2
GSR2 WC A2-3
GSR2 WC A2-4
GSR2 WC A2-4 replicate
GSR2 WC A2-5
GSR2 WC A2-6
GSR2 WC A2-7
GSR2 WC A2-8
GSR2 WC A2-9
GSR2 WC A2-10
GSR2 WC A2-11
GSR2 WC A2-12
GSR2 WC A2-13
GSR2 WC A2-14
GSR2 WC A2-15
GSR2 WC A2-16
GSR2 WC A2-17
GSR2 WC A2-18
GSR2 WC A2-19
GSR2 WC A2-20
GSR1 WC B7
GSR1 DR B4-1
GSR1 WC B5
GSR1 DR B3-1(interior glass)
GSR1 DR B3-1 (interior-repicatel)
GSR1 DR B3-1(exterior glass)
GSR1 DR B3-5
GSR1 WC B2
GSR1 DR B2-9
GSR1 DR B1-14
GSR1 WC B21
GSR1 WC B20
GSR1 WC B19b
GSR1 DR B8-1
GSR1 WC B18
GSR1 DR B7-2 (exterior glass)
GSR1 WC B16
GSR1 WC B15b
GSR1 WC B14
GSR1 WC B13
GSR1 DR B6-2
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
C17
C17
C17
C17
C17
C17
C17
C17
C17
C17
C17
C17
C16
C16
C16
C16
C16
C16
C16
C16
C16
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
GRAHAM ET AL.
Longitude
( E)
Latitude
( S)
Depth
(m)
Distance
(km)
80.975
81.069
81.238
81.373
81.471
81.583
81.662
81.777
81.910
81.982
82.060
82.152
82.221
82.317
82.397
82.485
82.563
82.623
82.713
82.827
82.932
83.037
83.123
83.203
87.724
87.977
87.881
87.801
87.627
87.627
87.530
87.438
87.348
87.262
87.163
87.033
86.938
86.847
86.756
86.636
86.531
86.439
86.339
86.243
86.155
86.061
88.699
88.835
89.057
89.153
89.153
89.153
89.153
89.412
89.540
89.694
89.797
89.878
89.977
90.172
90.242
90.335
90.512
90.595
90.677
90.843
90.953
41.105
41.170
41.299
41.463
41.538
41.587
41.660
41.745
41.801
41.883
41.963
42.050
42.114
42.172
42.242
42.304
42.366
42.427
42.492
42.538
42.593
42.660
42.729
42.790
41.893
42.057
42.005
41.951
41.833
41.833
41.773
41.713
41.650
41.588
41.531
41.553
41.494
41.430
41.366
41.366
41.313
41.254
41.192
41.137
41.072
41.013
41.704
41.803
41.921
41.960
41.960
41.960
41.960
42.169
42.230
42.327
42.360
42.421
42.482
42.560
42.581
42.652
42.758
42.784
42.828
42.927
42.984
2600
2612
2688
2563
2520
2537
2524
2558
2690
2763
2716
2510
2473
2620
2769
2735
2705
2577
2540
2575
2616
2635
2680
2812
2427
2363
2398
2434
2453
2453
2455
2442
2439
2458
2385
2280
2359
2296
2356
2506
2493
2498
2547
2585
2648
2743
2620
2560
2392
2243
2243
2243
2243
2350
2400
2350
2531
2486
2505
2790
2530
2530
2672
2430
2380
2255
2360
395.3
405.9
426.1
446.8
458.5
468.9
479.3
492.7
505.0
515.6
526.4
538.6
547.6
557.7
567.8
577.9
587.3
595.5
605.7
616.1
626.6
637.9
648.3
657.7
874.5
902.1
892.3
883.4
864.0
864.0
853.6
843.4
833.2
823.3
812.9
806.1
795.8
785.5
775.1
767.3
756.9
746.8
735.9
725.9
715.5
705.3
924.9
940.6
963.1
972.1
972.1
972.1
972.1
1003.1
1015.6
1032.3
1041.2
1050.6
1061.1
1079.0
1085.1
1095.9
1114.6
1121.7
1130.0
1147.4
1158.4
C 2014. American Geophysical Union. All Rights Reserved.
V
3
He/4He
(R/RA)
2-sigma
(R/RA)
Vesicle [He]
(ccSTP/g)
Vesicle [CO2]
(ccSTP/g)
7.63
7.85
7.67
7.63
7.83
7.60
7.67
7.63
7.64
7.59
7.60
7.75
7.61
7.61
7.56
7.53
8.07
7.92
8.39
8.24
8.05
7.96
7.77
7.92
9.41
9.51
9.38
9.30
10.05
10.04
9.60
9.51
9.65
9.52
8.92
8.53
8.39
8.72
8.72
8.65
8.63
8.73
8.58
8.24
8.64
8.56
9.40
9.51
9.48
7.63
7.66
7.64
7.66
8.92
9.52
8.43
8.68
8.65
8.46
8.81
8.43
8.53
8.30
8.33
8.52
8.20
8.41
0.05
0.06
0.06
0.08
0.05
0.05
0.07
0.05
0.06
0.06
0.06
0.11
0.09
0.06
0.06
0.07
0.06
0.05
0.06
0.08
0.07
0.05
0.06
0.07
0.06
0.07
0.07
0.06
0.07
0.07
0.06
0.05
0.07
0.06
0.06
0.05
0.07
0.06
0.06
0.05
0.06
0.06
0.06
0.05
0.06
0.06
0.07
0.09
0.06
0.06
0.06
0.06
0.07
0.06
0.06
0.13
0.06
0.07
0.12
0.06
0.06
0.06
0.21
0.07
0.06
0.06
0.07
3.89E-06
2.42E-06
8.83E-06
1.03E-06
4.04E-06
4.33E-06
8.14E-06
5.97E-06
2.34E-06
1.22E-05
2.46E-06
5.28E-07
8.26E-07
2.30E-06
3.00E-06
6.34E-06
1.20E-05
4.87E-06
2.46E-06
1.13E-06
7.16E-06
6.08E-06
2.35E-06
6.80E-06
1.33E-05
1.05E-05
8.71E-06
1.33E-05
9.44E-06
9.82E-06
1.79E-05
1.82E-05
1.13E-05
1.07E-05
1.16E-05
5.71E-06
8.34E-06
1.03E-05
1.80E-05
4.59E-06
1.72E-05
3.39E-06
1.34E-05
5.82E-06
1.70E-05
3.96E-06
1.11E-05
1.05E-06
1.64E-05
4.30E-06
1.24E-05
9.74E-06
7.94E-06
2.50E-06
7.75E-06
5.07E-07
1.39E-05
2.64E-06
4.28E-07
5.07E-06
2.64E-06
1.78E-05
1.37E-07
1.43E-06
3.48E-06
7.32E-06
6.56E-06
3.53E-02
1.87E-02
1.13E-01
6.99E-03
2.41E-02
4.55E-02
7.74E-02
6.51E-02
1.65E-02
1.49E-01
9.03E-02
1.79E-04
5.89E-03
1.74E-02
2.29E-02
4.30E-02
5.17E-02
1.14E-02
2.76E-02
6.95E-03
4.97E-02
5.80E-02
8.83E-02
3.47E-02
1.01E-01
6.13E-02
1.61E-02
9.63E-02
nd
1.78E-02
9.47E-02
1.09E-01
5.56E-02
6.00E-02
7.22E-02
3.89E-02
6.78E-02
7.36E-02
1.53E-01
2.39E-02
5.37E-02
7.33E-02
8.51E-02
3.97E-02
1.20E-01
2.04E-02
7.44E-02
5.60E-04
1.37E-01
2.66E-02
1.03E-01
7.44E-02
5.51E-02
1.74E-02
4.29E-02
8.42E-04
6.66E-02
1.98E-02
1.53E-03
5.13E-02
1.99E-02
1.76E-01
6.60E-04
1.33E-02
3.78E-02
7.41E-02
4.72E-02
2051
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Table 1. (continued)
Sample
Segment
Location
GSR1 WC B12
GSR1 WC B8
GSR1 DR B5-1
GSR1 DR B11-2
GSR1 WC B36
GSR1 DR B10-2
GSR1 WC B33
GSR1 WC B32
GSR1 WC B30
GSR1 WC B29
GSR1 WC B28
GSR1 WC B27
GSR1 DR B9-1
GSR1 WC B26b
GSR1 WC B25
GSR1 WC B24b
GSR1 WC B23
GSR1 WC B22
GSR1 DR B12-1
GSR1 WC B52
GSR1 DR B15-2
GSR1 WC B51
GSR1 DR B14-1
GSR1 WC B49b
GSR1 WC B47
GSR1 DR B13-2
GSR1 WC B45
GSR1 WC B44b
GSR1 WC B43b
GSR1 WC B42
GSR1 WC B41
GSR1 WC B40
GSR1 WC B39
GSR1 WC B38
GSR1 WC B55
GSR1 WC B54
GSR1 WC B53
GSR1 DR C8-1
GSR1 DR C7-3
GSR1 DR C6-2
GSR1 WC C49
GSR1 WC C48
GSR1 DR C5-12
GSR1 WC C46
GSR1 WC C45
GSR1 WC C43
GSR1 WC C42
GSR1 WC C38
GSR1 WC C36
GSR1 WC C35
GSR1 WC C34
GSR1 WC C33
GSR1 DR C4-1
GSR1 WC C31
GSR1 WC-C30
GSR1 WC C29
GSR1 WC C28
GSR1 WC C27
GSR1 WC C26
GSR1 WC C25
GSR1 WC C24
GSR1 WC C23
GSR1 WC C22
GSR1 WC C20
GSR1 WC C19b
GSR1 WC C18
GSR1 WC C17
C16
C16
C16
C16
C16
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C13/14
C13/14
C13/14
C13/14
C13/14
C13/14
C13/14
C13/14
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
Smt
Smt
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
GRAHAM ET AL.
Longitude
( E)
Latitude
( S)
Depth
(m)
Distance
(km)
91.034
91.401
91.503
91.858
91.941
92.228
92.306
92.398
92.630
92.764
92.845
92.933
93.017
93.086
93.234
93.355
93.456
93.547
93.648
93.853
93.940
94.032
94.149
94.310
94.336
94.426
94.604
94.680
94.769
94.925
95.010
95.088
95.258
95.340
96.034
96.217
96.313
96.120
95.867
95.726
95.773
95.831
95.932
96.112
96.189
96.059
96.117
96.298
96.451
96.512
96.571
96.627
96.692
96.764
96.896
96.948
97.009
97.089
97.160
97.225
97.262
97.319
97.378
97.554
97.603
97.677
97.740
43.038
43.272
43.337
43.450
43.503
43.347
43.386
43.438
43.557
43.633
43.692
43.743
43.796
43.828
43.897
43.922
43.980
44.023
44.078
44.158
44.188
44.231
44.274
44.413
44.569
44.600
44.724
44.762
44.805
44.884
44.923
44.957
45.043
45.082
45.156
45.250
45.303
45.833
46.215
45.947
46.018
46.047
46.088
46.217
46.257
46.624
46.656
46.746
46.834
46.872
46.896
46.925
46.952
47.053
47.120
47.160
47.191
47.210
47.242
47.268
47.315
47.356
47.385
47.476
47.500
47.533
47.569
2380
2790
2670
2737
2910
2400
2583
2557
2540
2558
2484
2509
2580
2531
2673
2634
2657
2721
2703
2906
2800
2926
2820
3042
2817
2760
2696
2667
2694
2779
2919
2727
2870
2960
2674
2772
2825
2000
1460
3500
3137
3195
2994
3110
3206
2517
2517
2465
2583
2511
2556
2580
2620
2576
2597
2603
2591
2578
2530
2545
2517
2448
2413
2434
2435
2477
2487
1167.2
1206.5
1217.5
1248.1
1257.0
1265.8
1273.4
1282.8
1305.7
1319.4
1328.5
1337.6
1346.4
1353.0
1367.1
1376.6
1387.0
1395.7
1405.8
1424.3
1431.9
1440.6
1451.0
1470.3
1481.7
1489.5
1508.9
1516.2
1524.6
1539.8
1547.8
1554.9
1571.3
1579.1
1629.3
1646.9
1656.5
1675.5
1682.1
1656.8
1664.1
1669.6
1678.7
1698.1
1705.4
1719.3
1724.9
1741.8
1756.9
1763.1
1768.3
1773.6
1779.4
1790.1
1802.4
1808.1
1813.9
1820.1
1826.5
1832.3
1837.4
1843.4
1848.9
1865.5
1870.0
1876.7
1882.8
C 2014. American Geophysical Union. All Rights Reserved.
V
3
He/4He
(R/RA)
2-sigma
(R/RA)
Vesicle [He]
(ccSTP/g)
Vesicle [CO2]
(ccSTP/g)
8.32
8.19
8.53
8.36
8.33
8.52
8.43
8.51
8.37
8.62
8.64
8.49
8.53
8.49
8.52
8.01
8.00
8.10
7.97
8.18
8.43
8.10
8.23
8.15
7.75
7.68
7.78
7.73
7.86
7.80
7.84
7.77
7.58
7.69
8.18
8.09
8.08
7.88
7.98
7.82
7.78
7.90
7.35
7.83
7.75
8.50
7.86
8.48
8.34
8.32
8.19
8.37
8.27
8.70
8.39
8.31
8.35
8.52
8.31
7.96
8.16
8.20
8.11
7.88
7.91
7.61
7.70
0.05
0.08
0.06
0.07
0.17
0.06
0.07
0.06
0.06
0.05
0.06
0.06
0.07
0.07
0.05
0.06
0.07
0.07
0.07
0.07
0.06
0.05
0.06
0.05
0.05
0.05
0.07
0.06
0.06
0.07
0.06
0.05
0.05
0.05
0.06
0.05
0.07
0.18
0.06
0.07
0.06
0.06
0.06
0.06
0.05
0.08
0.06
0.07
0.05
0.14
0.05
0.05
0.05
0.06
0.06
0.05
0.05
0.06
0.05
0.06
0.07
0.06
0.06
0.06
0.05
0.04
0.05
4.72E-06
1.15E-06
1.15E-05
7.84E-06
3.11E-07
2.55E-06
8.94E-06
1.16E-05
1.17E-05
4.08E-06
1.27E-05
1.07E-05
8.12E-06
2.52E-06
4.94E-06
3.97E-06
1.93E-06
7.01E-06
5.91E-06
1.80E-06
2.29E-06
5.08E-06
2.31E-06
5.30E-06
3.30E-06
3.86E-06
7.56E-06
9.06E-06
9.51E-06
1.72E-06
8.62E-06
1.40E-05
5.10E-06
2.58E-06
1.17E-05
1.01E-05
2.06E-06
2.22E-07
2.40E-06
1.19E-06
6.69E-06
1.14E-05
1.92E-06
8.82E-06
3.70E-06
1.39E-06
8.51E-06
7.00E-06
5.61E-06
3.50E-07
1.29E-05
1.45E-05
5.85E-06
1.00E-05
9.84E-06
3.99E-06
1.73E-05
3.28E-06
1.16E-05
1.04E-05
6.33E-06
8.82E-06
8.34E-06
1.15E-05
4.14E-06
6.78E-06
1.34E-05
4.43E-02
7.69E-03
7.82E-02
6.30E-02
3.26E-03
2.40E-02
1.03E-01
1.31E-01
9.78E-02
7.50E-03
7.29E-02
4.56E-02
8.66E-02
2.17E-02
5.05E-02
3.09E-02
1.49E-02
3.88E-02
5.27E-02
8.70E-03
1.48E-02
3.86E-02
1.81E-02
4.00E-02
4.22E-03
3.65E-02
3.48E-02
6.49E-02
6.84E-02
1.33E-02
7.89E-02
1.04E-01
5.68E-02
2.67E-02
6.16E-02
1.05E-01
1.49E-02
3.67E-04
7.50E-03
8.13E-03
5.73E-02
8.10E-02
1.56E-02
6.63E-02
3.52E-02
6.85E-03
1.01E-01
3.83E-02
4.08E-02
4.89E-04
1.07E-01
1.02E-01
5.74E-02
9.89E-02
8.10E-02
2.63E-02
1.44E-01
2.22E-02
1.05E-01
1.01E-01
4.20E-02
6.07E-02
7.92E-02
9.36E-02
3.56E-02
4.34E-02
6.33E-02
2052
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Table 1. (continued)
Sample
Segment
Location
GSR1 WC 15b
GSR1 WC C13b
GSR1 WC C11b
GSR1 WC C9
GSR1 DR C3-4
GSR1 WC C7b
GSR1 DR C2-2
GSR1 WC C4
GSR1 WC C2b
GSR1 DR C1-1
SEIR-WW10 Samples
WW10 66-9
WW10 67-13
WW10 72-2
WW10 79-19
WW10 83-17
WW10 85-19
WW10 88-1
WW10 105-1
WW10 132-1
SWIR
MD34-7
MD34-6
MD34-5
MD34-4
MD34-3
MD34-2
MD34-1
Agulhas 53-1-32
Agulhas 53-3-3
Protea5 D37-2
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C17
C17
C16
C14
C14
C13/14
C13
C11
C5
He/4He
(R/RA)
2-sigma
(R/RA)
Vesicle [He]
(ccSTP/g)
Vesicle [CO2]
(ccSTP/g)
1893.7
1904.1
1912.8
1927.3
1941.1
1946.3
1953.1
1972.2
1996.9
2006.7
8.01
7.80
7.85
7.86
7.84
7.80
7.98
7.86
7.77
7.82
0.05
0.05
0.05
0.05
0.08
0.05
0.05
0.06
0.05
0.09
3.84E-06
4.96E-06
1.50E-05
1.42E-05
9.54E-07
3.98E-06
1.80E-05
1.03E-05
1.29E-05
7.83E-07
2.02E-02
4.47E-02
1.47E-01
1.25E-01
6.45E-02
4.74E-02
1.25E-01
1.13E-01
1.41E-01
5.93E-03
965.9
968.5
1208.3
1585.1
1605.1
1637.7
1795.8
2238.1
2930.6
10.22
9.70
8.34
7.66
8.00
8.12
9.46
7.30
7.23
0.08
0.06
0.09
0.07
0.24
0.07
0.05
0.04
0.06
1.48E-06
5.77E-06
8.68E-07
1.82E-06
4.06E-08
2.51E-06
7.34E-06
3.53E-06
9.26E-07
2.40E-04
2.62E-02
4.21E-04
1.78E-02
4.21E-04
2.38E-02
4.10E-02
2.77E-02
4.11E-03
7.59
8.47
7.00
7.68
9.62
9.10
8.49
8.05
8.48
7.15
0.06
0.06
0.05
0.06
0.07
0.06
0.06
0.04
0.04
0.04
nd
nd
nd
nd
nd
nd
nd
6.94E-07
7.97E-07
2.86E-07
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
Longitude
( E)
Latitude
( S)
Depth
(m)
Distance
(km)
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
97.862
97.995
98.094
98.237
98.415
98.453
98.545
98.770
99.035
99.174
47.623
47.656
47.697
47.790
47.832
47.880
47.897
47.978
48.116
48.133
2537
2696
2778
2700
2607
2741
2640
2830
2811
2730
Smt
Smt
Smt
axis
axis
axis
axis
axis
axis
89.313
89.183
91.283
95.409
95.764
96.127
96.833
103.038
111.783
41.713
41.878
43.417
45.106
45.051
45.195
47.076
47.765
50.212
1375
1850
1470
3020
3069
2555
2568
2783
3328
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
36.30
38.80
40.65
43.70
49.88
56.27
57.84
32.67
34.07
45.77
44.81
44.18
43.89
40.98
37.71
33.76
31.69
47.17
46.03
40.13
1500
2190
2550
3700
3260
3800
4050
3250
3250
3075
3
0.1 km21 (1 sample per 10 km), and between 96 E and 99.3 E it was 0.2 km21 (1 sample per 5 km). GEISEIR II sampled the ridge axis between 81 E and 88 E at sampling density of 0.1 km21. Recovery was
extremely good at >150 sites using the newly designed French wax corer (nicknamed Hellboy) supplemented with 30 dredge localities. For example, along the midsection of the 96 E–99 E sampling, recovery
was typically 50–100 g of extremely fresh glass at every locality and sample appearance suggests that this
section of the ridge is currently active or has been active very recently.
Large, fresh glass chunks from the wax core and dredge collection were first cleaned ultrasonically in ethanol and acetone, followed by drying in air. Pieces free of surface alteration and obvious cracks were handpicked under a binocular microscope. Helium isotope analyses were carried out by crushing in vacuum,
liberating helium and associated gases (primarily CO2) from vesicles trapped within the glass. Typically 200–
300 mg of fresh glass (several mm-sized chips) were loaded into stainless steel crushers between a thin,
removable stainless steel disk anvil at the tube bottom and an overlying magnetic piston. The magnetic piston was lifted using a system of external solenoids and actively forced onto the sample 75 times to produce
high impact strokes. The sample was crushed to a powder while still connected to the vacuum line. The
released CO2 and any minor H2O were condensed together during the crushing at 77 K (liquid N2 temperaR Zr-Al getters.
ture) in a stainless U-tube section. Noncondensable reactive gases were removed using SAESV
Noble gases were separated onto a cryogenically controlled charcoal trap at 10 K. The trap was warmed to
45 K and helium was admitted directly to the mass spectrometer for isotope ratio and peak height determination, while Ne and other noble gases were held behind. Line blanks were always performed before sample analysis and were typically <5 3 10211 cm3 STP 4He.
R noble gas mass spectrometer at Oregon
Helium isotope analyses were performed using a Nu InstrumentsV
State University (OSU). Helium concentrations were determined by peak height comparison to standards,
and CO2 concentrations were determined by capacitance manometry following distillation of CO2 from H2O
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2053
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
into a calibrated volume of the
sample extraction line using liqGEISEIR
30
uid nitrogen and frozen isoproall SEIR
panol traps. The mass
25
spectrometer is fitted with a
high-sensitivity ion source (Nier20
type) and is operated under static
15
vacuum. It is all-metal construction and can be baked to reach
10
low background levels and low
static gas rise rates. The instru5
ment is fitted with a single faraday cup and three ion-counting
0
6
7
8
9
10
11
12
13
14
15
detectors. The high mass ion
3
He/4He (R/RA)
counter position is shared with
the faraday with switching via an
Figure 3. Histogram of 3He/4He for Southeast Indian Ridge basalts. Replicate analyses
electrostatic deflector. The cenhave been averaged. The dashed red curve depicts a unimodal Gaussian distribution for
tral and low mass ion-counting
the densely sampled region between 81 E and 100 E (segments K, L, and C13–C17.
Range 5 7.18–10.22; Mean 5 8.24, s.d. 5 0.58, skewness 5 1.04, median 5 8.18; n 5 175
detectors have a special filter to
sample locations). The solid blue curves depict a mixing analysis describing the histoeliminate scattered low energy
3
4
gram as a bimodal distribution, where one subpopulation has mean He/ He 5 8.12
ions that may interfere with small
RA 6 0.37 (1 s.d.) and a proportion of 0.90, while the other subpopulation has mean
3
He/4He 5 9.57 RA 6 0.26 and a proportion of 0.10.
ion beams (which can be especially important when trying to
measure very small amounts of 3He in the presence of large amounts of 4He, for example). The OSU mass
spectrometer is a variable dispersion instrument that employs specially designed electrostatic lenses. It performs the 3He/4He measurement in a pseudohigh-resolution mode that partially resolves 3He from
1
1
HD 1 1H1
3 . During the course of this study, on peak HD 1H3 count rates were typically 25 cps during the
3
course of the analysis. Beneath the peak-flat where He was measured the HD 1 1H1
3 count rate was effectively zero.
Number
35
Gas standards to calibrate the instrument’s sensitivity and mass discrimination include both marine air and
a high 3He/4He gas standard (the helium standard of Japan, HESJ) [Matsuda et al., 2002]. The HESJ used as a
running standard during this study was subsampled from the same batch provided to the National Oceanic
and Atmospheric Administration/Pacific Marine Environmental Laboratory [Lupton and Evans, 2004]. It was
calibrated with the OSU noble gas mass spectrometer against marine air collected in 2007 at the Oregon
coast, and found to have 3He/4He (20.4 RA) identical to the HESJ subsample routinely used at NOAA/PMEL.
Several samples have been analyzed in duplicate or triplicate during this study in order to assess sample
reproducibility at OSU and to compare the results to earlier analyses [Graham et al., 2001]. Two samples
from the WW10 expedition (72-2 and 88-1) previously analyzed at NOAAA/PMEL were reanalyzed and the
results agree within analytical uncertainty (8.34 6 0.09 versus 8.21 6 0.06 and 9.46 6 0.05 versus 9.38 6 0.05,
respectively). One of the newly collected basalts having elevated 3He/4He (GSR II A2-4) was analyzed twice
in the OSU lab and the results agree within uncertainty (10.04 and 10.05 RA). Another sample (GSR I DR B3-1
from segment C17), in which the interior and exterior portions of thick glass from a single basalt could be
separated, was analyzed three times and each time gave the same isotopic result (7.63, 7.66, and 7.64 RA).
Lastly, one sample (GSR I DR C5–12 from segment C13/14) was found to have a 3He/4He ratio (7.35 RA) similar to that for a previously analyzed sample (WW10 WC48 with 3He/4He 5 7.31 RA) collected from the same
location. This last comparison indicates that the low 3He/4He anomaly in the midsection of segment C13/14
is confined to a restricted area (<10 km length), similar to what is observed for sample DR B3-1 along segment C17.
3. Results
Results for vesicle 3He/4He, and He and CO2 concentrations in the new sample suite are reported in Table 1.
Table 1 also includes analyses of nine WW10 basalts from the SEIR to supplement the new GEISEIR results.
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2054
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Off-axis Seamounts
E-MORB (K/Ti>0.2)
9
3
He/4He (R/RA)
10
8
K
L
C17
C16
C15
C14
C13
C12
7
500
1000
1500
2000
Along-Axis Distance (km)
Figure 4. 3He/4He along the densely sampled portion of the SEIR axis between 81 E and 101 E (distance is shown in km from the location
of St. Paul island). Vertical dashed lines delineate boundaries to the first-order and second-order tectonic segmentation of the ridge axis.
We also report data for 10 basalt glasses from the Southwest Indian Ridge (SWIR, from expeditions of Marion Dufresne 34, Agulhas 53, and Protea 5) for which Pb, Nd, Sr, and Hf isotope compositions have been previously published [Hamelin and Allegre, 1985; Mahoney et al., 1989, 1992; Chauvel and Blichert-Toft, 2001;
Janney et al., 2005].
Vesicle He concentrations range between 0.2 and 18 lccSTP/g (Table 1), and values at the high and low
ends of this concentration range occur thoughout the study area. There is no obvious covariation of
3
He/4He with He concentration, MgO content or Fe/MgO ratio (not shown). Likewise there are no simple
relationships in the overall data set, or in the data grouped by individual segments, between He concentration or He/CO2 ratio and axial depth. The helium isotope variability in these SEIR basalt glasses is therefore not explained by shallow-level processes operating in the crust (e.g., radiogenic ingrowth of
degassed magma during prolonged storage in the crust is insignificant in producing the observed
3
He/4He variations).
A histogram of 3He/4He (Figure 3) compares the new results to the rest of the SEIR. The samples collected
between 81 E and 100 E (n 5 175) show a 3He/4He range of 7.2–10.2 RA, which is more than half the range
in MORB glasses (6–11 RA) erupted away from the influence of ocean island hot spots having high
3
He/4He [Graham, 2002; Kurz et al., 2005]. 3He/4He ratios for samples between 81 E and 100 E have an
approximate Gaussian distribution. Its median is 8.17 RA, identical to the value of 8.1 RA previously computed for MORBs from each of the Atlantic, Indian, and Pacific Ocean basins [Graham, 2002]. In detail, there
may be a slightly increased proportion of 3He/4He values above 9 RA (10% of the samples analyzed) compared to what is expected for a single Gaussian population. A mixture analysis of the new 3He/4He data
shows that when it is deconvolved into two distinct normal populations, population 1 has mean
3
He/4He 5 8.12 RA 6 0.37 (1 s.d.) and a proportion of 0.90, while population 2 has mean 3He/4He 5 9.57
RA 6 0.26 and a proportion of 0.10 (Figure 3). This possible bimodality has implications for estimating the
upper limit of 3He/4He in depleted upper mantle away from high-3He/4He hot spots, and where recyled
low-3He/4He material is absent (discussed further below).
3
He/4He variations are evident at a variety of length scales (Figure 4), ranging from long (1000 km), to
intermediate (100 km) and to short (20–40 km) distances. Numerous examples of abrupt, along-axis
transitions in 3He/4He can be found. Some of these abrupt changes occur near tectonic segment
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2055
Geochemistry, Geophysics, Geosystems
11
Fe8
boundaries (e.g., K/L, C15/14,
C13/12). However, there are clear
examples of abrupt changes not
associated with such boundaries
that sometimes occur in the middle of a segment (e.g., L, C15,
C14, C13).
Segment K
Segment L
Segment C17
Segments C12- C16
Segments C2-11
Segments J2-J4
Segments ASP
Segment H
Segments B-G
12
10
9
8
7
6
A
14
10
3
He/4He (R/RA)
12
10
GEISEIR I & II
ASP Segments
9
8
0.1
0.2
0.3
0.4
0.5
8
B
7
0.0
0.1
0.2
0.3
0.4
0.5
10.1002/2014GC005264
0.6
K/Ti
Figure 5. (a) Fe8 and (b) 3He/4He (R/RA) versus K/Ti (elemental weight ratio) for SEIR
basalts. Data are from this study plus Douglas Priebe [1998], Graham et al. [1999, 2001],
Johnson et al. [2000], and Nicolaysen et al. [2007]. The two highlighted samples are the
enriched basalt GSR I DR-B3-1 (green) and the depleted basalt WW10 66-9 (yellow) from
segment C17 that are separated by <10 km. For reference, the gray line depicts the linear
regression for regional averages of the global MORB data set from Shen and Forsyth
[1995].
Abrupt excursions to low
3
He/4He (7.65 RA) ratios are
notable in four particular E-MORB
glasses. E-MORBs are defined
here as having an elemental K/Ti
ratio >0.2. The notable cases are
samples DR B3-1 (analyzed in
triplicate) and B3-5 along segment C17, and samples DR C5-12
and WW10 WC48 from the same
area in the middle of the short
segment C13/14. These E-MORBs
occur as isolated examples far
from ridge segment boundaries.
Seamounts are present both
north and south of segment C13/
14 at the same longitude as the
central portion where its EMORBs were recovered. Each of
those seamounts (samples DR
C7-3 and C8-1) erupted low-K/Ti
basalts that have higher 3He/4He
ratios than the E-MORBs, and are
similar to the depleted MORBs
recovered elsewhere along C13/
C14. This further confirms that
the low-3He/4He signature along
C13/14 is restricted to a small
area (<10 km).
The low 3He/4He in the E-MORBs from segments C17 and C13/14 (Figure 4) are also the lowest ratios found
for hundreds of kilometers along the SEIR. The value of 7.5 RA along C17 corresponds precisely to the low
(and uniform) values observed along segments K and much of C14, each of which is 400–500 km away.
The E-MORB signature along C13/14 has a slightly lower 3He/4He than along C17, and it resembles the low
values observed to its east along C12 which is also 400 km away. While segments K and C12 show relatively low and uniform 3He/4He they do not typically erupt E-MORBs. This pattern of occurrence suggests
that enriched mantle heterogeneities may be widespread in the subridge mantle, but only occasionally do
they give rise to E-MORB chemical compositions. In contrast, the presence of enriched heterogeneities may
be the primary cause of downward variations in 3He/4He, from typical ambient values of 8.5 RA or higher.
Abrupt excursions to high 3He/4He also occur throughout the study area. Notably, along segment C13
where our sampling density is the highest (1 sample per 7 km of ridge), 3He/4He increases abruptly to 9.4
RA (in sample WW10 88-1) from background values of 8.2 to 8.4 RA that extend over >100 km of this ridge
segment This abrupt change occurs over 15 km or less. Similarly, the region of broadly elevated 3He/4He
along the eastern portion of segment L and the western portion of segment C17 shows 3He/4He spike-like
increases (up to 10 RA in sample GSR II WC A2–4). These spikes occur over distances of 10 km or less.
The highest 3He/4He ratios in the 81 E–100 E region are generally found in depleted basalt glasses characterized by low K/Ti (Figure 5b). This contrasts markedly with the basalts from the Amsterdam-St. Paul
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2056
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
vicinity, where 3He/4He and K/Ti show a generally positive relationship due to the enriched hot spot effect on
depleted mantle beneath the ridge axis (Figures 2 and 5b). Between 81 E and 100 E, highly depleted lavas
occur along the ridge axis and also just outside the axial region. The highest 3He/4He occurs in a low-K/Ti
basalt (WW10 66-9) recovered from 1375 m water depth near the summit of a large seamount immediately
north of the spreading ridge at 89.3 E (Figure 5). The high 3He/4He in this seamount basalt also occurs within
10 km of the E-MORBs along segment C17 that have distinctly low 3He/4He (DR C3-1 and C3-5).
In contrast to the above examples of significant 3He/4He variability over short length scales in the study
area, there are also examples where large sections of the ridge axis have uniform 3He/4He. The best case is
the 180 km section of segment K (between 395 and 578 km distance from St. Paul). Here the total
3
He/4He range (for 18 samples collected at approximately 10 km intervals) is 7.53–7.85 RA, and 16 of the
samples lie between 7.53 and 7.75 RA. The mean and standard deviation for this section of the SEIR are
7.65 6 0.086 RA, which corresponds to a total variability of 61.1%. This is similar to the variability observed
in the Lucky Strike area of the Mid-Atlantic Ridge [Moreira et al., 2011], which is currently the most densely
sampled portion of the mid-ocean ridge system for 3He/4He. Although the ridge length covered by that
study was very short (13 km) the sampling density was very high (25 samples corresponding to a mean
sample spacing of 0.5 km) and 3He/4He averaged 8.28 6 0.07 RA (60.9%). Moreira et al. [2011] attributed
the uniformity in 3He/4He to lava eruptions from a well-mixed crustal magma lens. The limited variability
observed here along much of segment K is comparable to that observed by Moreira et al. [2011], making
our new results remarkable because they represent an order of magnitude increase in ridge length compared to Lucky Strike and these SEIR lavas are unlikely to be derived from a common magmatic system.
4. Discussion
4.1. Melting of Lithologically Heterogeneous Mantle and 3He/4He in SEIR Basalts
Previous studies along the SEIR have inferred that the He, Pb, Sr, Nd, and Th isotope variations are controlled by variations in the depth and extent of melting of heterogeneous mantle [Graham et al., 2001;
Mahoney et al., 2002; Nicolaysen et al., 2007; Russo et al., 2009]. The melting process was thought to sample
lithologically heterogeneous mantle in response to a lateral temperature gradient, extending from hotter
mantle near the ASP Plateau to colder mantle beneath the Australian-Antarctic Discordance. U-series disequilibria in the SEIR basalts further revealed the connection between melting rate (which is strongly controlled by the lithologic constitution of the mantle), the extent of melting, mantle porosity, and the depth at
which melting initiates [Russo et al., 2009]. However, there are no simple correlations between indicators of
degree of melting (such as Na8) and K/Ti in SEIR basalts, indicating that the K/Ti variability largely represents
compositional heterogeneity in the mantle source region. Nor are there any simple correlations between
K/Ti and Pb isotope ratios or eHf, suggesting that each subpopulation that is distinguishable in terms of Pb
and Hf isotope compositions involves derivation from both N-MORB and E-MORB mantle sources [Hanan
et al., 2013].
In terms of 3He/4He, however, we noted earlier that a mixture analysis of the 3He/4He histogram for the 81 E–
101 E region could be deconvolved into two distinct subpopulations having 3He/4He 5 8.1 6 0.8 (2r) RA and
3
He/4He 5 9.6 RA 60.6 (2 r) RA (Figure 3). The relationship between 3He/4He and K/Ti also bears on this observation. Low 3He/4He occurs in SEIR basalts over the full range of K/Ti (Figure 5). However, all E-MORBs having
K/Ti > 0.3 have low 3He/4He ratios (segments C17 and C13/14, plus two samples west of the study region
toward the AAD). In contrast, the highest 3He/4He ratios (>9 RA) typically occur in the most depleted lavas (15
of 17 samples have K/Ti < 0.1), and the highest ratio (10.2 RA) is found in a depleted lava (K/Ti 0.05) from the
seamount immediately adjacent to the C17 ridge axis (Figure 5). Notably, 3He/4He ratios above 9 RA are absent
from E-MORBs altogether. As well, the negative 3He/4He – K/Ti correlation shows quite different behavior
from that observed along the ASP-influenced section of the SEIR (Figure 5), where elevated 3He/4He (>14 RA)
is associated with higher K/Ti [Graham et al., 1999]. This different relationship between He isotopes and minor
element variations for the ASP region versus the rest of the SEIR reinforces the notion that source heterogeneity, as revealed by variability in He isotopes, should be considered within the framework of lithologic variability (and its effect on partial melting), as revealed by variability in K/Ti.
It is well recognized that a small amount of garnet pyroxenite/eclogite embedded in peridotite may be
responsible for a host of geochemical signatures in basalts. The landmark paper by Hirschmann and Stolper
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2057
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
[1996] paved the way to the idea that compositional variability in MORBs at short length scales can be
related to the presence of small amounts of pyroxenite/eclogite within a mantle source comprised mostly
of a peridotitic matrix. Pyroxenite has a lower solidus temperature, and primarily because it is pyroxene rich
it has a higher melt productivity than peridotite. For example, the presence of 5% pyroxenite could contribute as much as 40% to the total melt of a peridotite-pyroxenite mixture. Unfortunately, the major element signature of pyroxenite in the ensuing basaltic melt is difficult to detect because it is similar in major
element composition to a peridotite melt. In general, the melting of pyroxenite-peridotite mixtures in which
the two components have different trace element and isotope compositions leads to a number of important complexities. Two of these are relevant to the present study; (1) the isotopic range observed in basalts
is unlikely to represent the full range of heterogeneity in the mantle source region—i.e., the depleted peridotite composition may not be represented by even the most depleted MORB magma—and (2) isotopeisotope arrays do not necessarily point directly toward the mantle end-members [Kellogg et al., 2007; Stracke
and Bourdon, 2009; Rudge et al., 2013].
Eclogite or garnet pyroxenite in the mantle may originate either as subducted ocean crust, delaminated
lower continental crust, or trapped basaltic melt. Melts of these mafic lithologies alone typically have elevated FeO and no enrichment in K2O [Lambart et al., 2013]. However, Kogiso et al. [1998] showed that melting a basalt-peridotite mixture may generate alkali-rich melts. Such mixtures can be formed by invasion of
peridotite with partial melts derived from eclogite or by hybridization with peridotite as the recycled crust
is thinned during convective stirring in the mantle. This yields olivine-bearing, silica-deficient pyroxenite
[Kogiso et al., 2003] and changes the melting situation dramatically. Melts of such hybrid mantle will be
alkalic and have low SiO2 and CaO, and high Al2O3 and TiO2. While the E-MORBs from segments C17 and
C13/14 are not alkalic lavas, they do lie close to the alkaline/subalkaline boundary. Along with their elevated
K2O (0.7–1.2 wt %) and K/Ti and low 3He/4He (Figure 5), they have very low CaO/Al2O3 (0.60–0.65) at moderate values of MgO (6.5–8.0 wt %). Because eclogite and garnet pyroxenite are olivine-free lithologies, primary magmas derived from them may have lower MgO than those derived from peridotite. The extent to
which the crystallization of clinopyroxene at depth could be responsible for the lower CaO/Al2O3 traits
requires a full assessment of the major element variations in the GEISEIR basalt glass suite and is beyond
the scope of the present paper. We simply point out here that the E-MORB character of some of the lavas
from the present study area appears to be consistent with the melting of a peridotite-eclogite mixture that
may account for the low 3He/4He character.
There is a prevailing negative covariation of K/Ti with Fe8 along the SEIR (Figure 5). This relationship is also
observed in MORBs globally [Shen and Forsyth, 1995]. The observed negative covariation along the SEIR further supports the notion that the high-K/Ti mantle source is enriched in a ‘‘basaltic’’ component [Langmuir
and Hanson, 1980]. This makes it difficult to know the extent to which one can attribute Fe8 variations to
changes in the initial depth of melting (or the mean depth of melting) as is commonly assumed. The broad
covariation between 3He/4He and Fe8 for the SEIR led Graham et al. [2001] to attribute the helium isotope
variations to melting of heterogeneous mantle containing a small amount of garnet pyroxenite. Such lithologic heterogeneity should have lower 3He/4He if it originates from recycled crust or ancient trapped melt.
Garnet pyroxenite or eclogite has a lower solidus temperature and will melt preferentially to surrounding
mantle peridotite. It will also contribute a significantly smaller proportion to the total magma when melting
begins at greater depth (at higher mantle temperature and pressure) because the pyroxenitic melts become
increasingly diluted by partial melts of the peridotite at shallow depths. Following this reasoning, Graham
et al. [2001] suggested that hotter mantle would begin to melt deeper, produce a shallower ridge axis, and
produce melts having higher Fe8 and 3He/4He. This explanation implies that highly depleted (pyroxenite or
eclogite free) regions of the upper mantle will have 3He/4He ratios of 9 RA, or higher.
Russo et al. [2009] found that both N-type and E-type MORBs record a west-to-east decrease in (230Th/238U)
activity ratio along the SEIR from 90 E to 118 E, with no systematic differences in Th-U disequilibria
between the two MORB types. They carried out forward modeling of mixtures of melts derived from pyroxenite 1 peridotite to examine likely fractions of each melt type that would be consistent with the basalt
observations. Using simple but reasonable assumptions, Russo et al. [2009] showed that garnet pyroxenite
or eclogite comprises 3–5% of the mantle beneath the easternmost SEIR, and lesser amounts in the 81 E
and 100 E region. One possibility then is that the elevated 3He/4He ratios of 9–10 RA along segments L and
C17 approach the intrinsic He isotope composition for pyroxenite-free (harzburgite) upper mantle. It seems
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2058
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
doubtful though that melting of harzburgite alone would be sufficient to produce the shallower axial ridge
in the 88 E region, as well as the relatively large seamounts immediately adjacent to the ridge (which rise
to a water depth of 1400 m).
The axial location of the E-MORBs is an additional observation that bears on any melting scenario for lithologically heterogeneous mantle. The two E-MORBs having the most extreme 3He/4He are the GSR I DR3
samples from segment C17, and the DR5 sample (plus WW10 WC48) from segment C13/14. Samples DR B31 and B3–5 were recovered from a small knob atop the axial volcanic ridge in segment C17; remarkably this
locality corresponds to the shallowest depth of the axial ridge in the whole 81 E–100 E study area. In comparison, samples DR C5-12 and WW10 WC48 were recovered from the center of the small segment C13/14.
This second example of low 3He/4He in E-MORBs also appears to be confined to a small and shallower section of the ridge axis. As noted earlier, seamounts are found just north and south of segment C13/14 at the
same longitude as the central portion where the E-MORBs were recovered. Each of these seamounts (samples DR C7-3 and C8-1) erupted low K/Ti basalts that have 3He/4He ratios similar to other depleted MORBs
from this area. The low 3He/4He signatures in the E-MORBs are, therefore, restricted to small areas (<10 km)
in each case. These restricted occurrences along segments C17 and C13/14 imply that mantle heterogeneity
is sampled almost exclusively by vertical melt transport beneath these areas of the ridge.
The presence of such enriched compositions at segment centers and shallow locations is not what would
be expected from the Klein and Langmuir [1987] model of melt generation beneath ridges. That model predicts that the largest extents of melting and consequently the most depleted MORB magmas will be present beneath the shallowest sections of ridge. The nearby MORBs from segments C17 and C13/14 do have
these requisite features. The rare occurrences of E-MORBs at the very shallowest locations suggest a linkage
between melt generation from enriched mantle and rapid magma transport via formation of permeable
conduits [Katz and Weatherley, 2012], and that this is sometimes more discernible beneath central axial
highs. Perhaps the inward heat flux that drives the melting of the heterogeneity [Katz and Rudge, 2011]
leads to a more discrete sampling of the enriched material when surrounding mantle regions have experienced extensive melting, thereby leading to E-MORB association with a local axial high. These E-MORB
melts appear to have ascended from depth without significant interaction with the surrounding (and perhaps barren) peridotite or with depleted melts that may have been present in a magma lens. Elsewhere
along the ridge, chemical evidence for the presence of enriched source material is more cryptic because
the enriched melts become diluted en route with melts from surrounding peridotite. The presence of
enriched heterogeneities appears to be responsible for much of the variability in basalt 3He/4He, producing
downward excursions from a background depleted peridotite value that is greater than 9 RA beneath much
of the SEIR.
4.2. Crustal Magma Lens Filtering of 3He/4He Variability
Magma redistribution complicates the simple mapping of surface variability in 3He/4He to mantle-scale variability. Both dike propagation within the crust and lateral mixing in crustal magma lenses or near the
MOHO may be important. Along the fast-spreading East Pacific Rise (EPR), 5–15 km long crustal magma
lenses are responsible for the fine-scale tectonic and geochemical segmentation [Carbotte et al., 2013].
Upper crustal accretionary units at this scale along the EPR have distinct geochemical and morphological
characteristics that reflect near vertical transport from the underlying magma lens. Our sampling interval in
the GEISEIR study area (Figure 1) ranges from 10 km in the west (along segments K, L, and C17 to C13/14)
to 5 km in the east (along segment C13). To a first order, this sampling density allows for tests of possible
along-axis crustal magma lens continuity and a comparison with first-order and second-order tectonic segmentation. Addressing questions associated with finer scales of segmentation requires a higher sampling
density than is currently available. Currently, there also are no seismic investigations of the GEISEIR region
to provide a context for the location of possible crustal magma lenses.
The restricted presence of E-MORBs in the SEIR sample suite is relevant to discerning the level of crustal filtering. The occurrence of E-MORBs at a ridge segment center was noted previously along the Mid-Atlantic Ridge
[Gale et al., 2011]. In that case, they were associated with transitional type basalts that occur throughout the
ridge segment, and this cooccurrence was interpreted to reflect temporal variability with the E-MORBs derived
from an earlier pulse of magmatism [Gale et al., 2011]. Temporal changes in erupted magma composition
must also occur along the SEIR, but our understanding of these changes is very limited. Given the present
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2059
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
observations it seems likely this temporal variability is subordinate to mantle melting effects in producing
much of the 3He/4He variation. Nearly the full range in 3He/4He is found over a distance of 10 km in the case
of the C17 basalts, and a number of the most depleted MORBs along the SEIR also come from the area immediately adjacent to the E-MORB locations. Given our sampling density, it seems likely that spatial variability in
3
He/4He associated with the different MORB types (Figures 4 and 5) results from quasi-discrete vertical supply
of melt from the mantle to crust. Although there is considerable uncertainty in the amount of filtering by
crustal magmatic processes, an issue that requires further evaluation through major and trace element modeling, these crustal-scale processes currently appear to be of lesser importance in controlling the helium isotope
variability along the SEIR than melting of the underlying heterogeneous mantle. In this context, we do note
that the dominant 3He/4He-K/Ti relationship for SEIR basalts is a large variability of 3He/4He at low K/Ti, while
most of the higher K/Ti lavas have low 3He/4He ratios (Figure 5b). Mixing of depleted and enriched magmas,
in which the depleted magma end-member has a wider range of He concentrations relative to the enriched
magma, could lead to large variability in 3He/4He at low and relatively uniform K/Ti. One possible cause might
be protracted storage and degassing of volumetrically more abundant depleted magmas in the crust, and the
occasional intrusion and mixing with enriched magmas. This idea can be further evaluated when additional
volatile data (CO2 and Ar concentrations, plus Ar isotopes) become available.
4.3. Isotope Geochemistry of the Indian Ocean Mantle
The full database for He, Pb, and Hf isotopes in SEIR basalts is provided in the supporting information (Table
S1) where the along-axis variations in Pb and Hf isotopes are compared to those for 3He/4He (Figure S1).
4.3.1. He-Pb-Hf-Nd-Sr Isotope Relations
Figure 6 displays the relations between 3He/4He and selected radiogenic isotope ratios of Pb and Hf for the
same samples from the Southeast Indian Ridge. The isotope relationships show several features, including:
1. Overall, the full data SEIR set shows no systematic relationships between He isotopes and those of Pb or
Hf.
2. The ocean island and ridge localities having high 3He/4He ratios show a considerable range in the isotope
compositions of Pb and Hf. All SEIR lavas west of the AAD have a DUPAL signature [Hart, 1984] based on
their elevated 208Pb/206Pb ratios [Mahoney et al., 2002].
3. The highest 3He/4He ratios extend to 18 RA and occur at the distant Heard Island/Kerguelen Plateau hot
spot, >1500 km from the SEIR. Lower ratios (8–10 RA) are also observed at those localities. 3He/4He ratios
are bimodal at Heard Island, with high values in the Laurens Peninsula volcanic series and low values in the
Big Ben volcanic series [Hilton et al., 1995].
4. Elevated 3He/4He ratios, >14 RA, are also associated with the Amsterdam-St. Paul hot spot which influences ridge segments atop and immediately northwest of the ASP Plateau [Graham et al., 1999]. Along some
of these segments the isotope trends involving He approximate binary mixtures but they can have distinctly
different Sr-Nd-Pb-Hf isotope compositions for the hot spot (high-3He/4He) end-member [Nicolaysen et al.,
2007] (cf. segment H versus the ASP segments in Figure 6).
5. Neglecting the E-MORBs along segment C17, basalts from segments L and C17 in the western sector of
the GEISEIR area show a slightly increasing trend of 3He/4He with radiogenic Pb isotope ratios (Figure 6).
These trends along L and C17 could be viewed as extensions of the He-Pb isotope trends produced by mixing with the high-3He/4He ASP hot spot (as exhibited by segments H and those atop the ASP plateau), but
extending to even more depleted mantle compositions than are found near the hot spot. However, segments L and C17 are located 1000 km away and there is little if any overlap in Sr-Nd-Pb-Hf isotope space
between these different groupings, which makes a contiguous trend linking them questionable. In this context, based on Pb-Nd-Sr isotope variations in the vicinity of segments L and C17, Mahoney et al. [2002] suggested there was little if any mantle flow from the ASP or Kerguelen-Heard hot spots toward the eastern
SEIR. The eastward changes in some isotopic parameters are in fact the opposite of those expected from
hot spot material flowing into and mixing with depleted MORB mantle beneath the ridge axis.
6. The E-MORBs along segment C17 having high K/Ti (e.g., GSR1 DR B3 highlighted in green in Figure 6) lie
away from any possible He-Pb-Hf isotope trends shown by other basalts from the same ridge segment. The
E-MORBs are similar in He and Hf isotope compositions to some lavas from Heard Island and Kerguelen (Figure 6), and to the anomalous basalt MD34 D5 from the Southwest Indian Ridge (Table 1). This indicates that
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2060
Geochemistry, Geophysics, Geosystems
20
18
16
14
A
B
Plume
Kerguelen
SWIR
12
10
3
He/4He (R/RA)
Plume
Segment K
Segment L
Segment C17
Segments C12- C16
Segments C2-11
Segments J2-J4
Segments ASP
Segment H
Segments B-G
Heard Island
DM
DM
8
D5
6
0.90
0.95
1.00
208
RCL
20
1.05
15.45
206
He/4He (R/RA)
15.55
15.60
15.65
204
Pb/ Pb
C
Heard
Island
16
15.50
207
Plume
18
RCL
D5
Pb*/ Pb*
3
10.1002/2014GC005264
Plume
D
14
12
10
DM
DM
8
D5
RCL
6
-5
D5
RCL
ROC
Heard
0
5
εHf
10
15
20
17.5
18.0
206
18.5
19.0
19.5
204
Pb/ Pb
Figure 6. 3He/4He versus (a) 208Pb*/206Pb*, (b) 207Pb/204Pb, (c) eHf, and (d) 206Pb/204Pb. The two highlighted samples are the enriched basalt GSR I DR-B3-1 (green) and the depleted basalt
WW10 66-9 (yellow) from segment C17 that are separated by <10 km. Data are from this study, plus Graham et al. [1999, 2001, 2006], Hamelin and Allègre [1985], Mahoney et al. [1989,
1992, 2002], Johnson et al. [2000], Janney et al. [2005], Nicolaysen et al. [2007], and Hanan et al. [2013]. The full SEIR isotopic data set is reported in the supporting information (Table S1).
Heard Island data are from Barling and Goldstein [1990], Hilton et al. [1995], and Barling et al. [1994]. Because Hf isotope compositions have not been reported for the same samples from
Heard Island analyzed for 3He/4He, the range of expected compositions for the Big Ben (high 3He/4He) and Laurens Peninsula (low 3He/4He) volcanic series were estimated from Nd isotopes and the global Hf-Nd isotope correlation, and are shown as boxes. Kerguelen picrite and high-MgO basalt data are from Doucet et al. [2005, 2006]. Where their olivine separated
from individual samples was analyzed multiple times the 3He/4He results have been averaged. Approximate mantle end-members are shown as Plume (having elevated 3He/4He but not
necessarily the most extreme Pb and Hf isotope compositions, due to multicomponent mixing with recycled materials), DM (depleted mantle), RCL (recycled subcontinental lithosphere),
and ROC (recycled ocean crust/subduction mantle wedge).
basalt K/Ti enrichment along the SEIR results from melting of mantle containing ancient heterogeneities
that vary significantly in their abundance over length scales of 10 km or less.
7. Samples from the SWIR are few in number (Table 1) but they show both elevated (>9 RA) and low (<7
RA) 3He/4He ratios. These SWIR samples resemble some of the SEIR basalts, most notably the depleted and
enriched MORBs from segments L and C17. In particular, sample MD34 D5 resembles the E-MORBs from
C17 (DR B3) in both its low 3He/4He (7 RA) and low eHf. The SWIR D5 sample carries the well-known
extremely low-206Pb/204Pb composition measured in some Indian MORBs (Figure 6), as well as characteristically low eNd and eHf [Hamelin and Allègre, 1985; Mahoney et al., 1992; Chauvel and Blichert-Toft, 2001; Janney
et al., 2005]. Previous work by Mahoney et al. [1992] showed that the SWIR segment from which MD34 D5
was recovered, between 39 S and 41 S, represents a tectonic corridor containing stranded fragments of
continental lithosphere derived by thermal erosion beneath Madagascar during the mid-Cretaceous.
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2061
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
4.3.2. Melting of Isotopically Heterogeneous Mantle and 3He/4He in SEIR Basalts
Collectively, the relations involving 3He/4He shown in Figure 6 show that the low-3He/4He signature in the
E-MORBs cannot be explained by derivation from an older blob of plume-derived (initially high-3He/4He)
mantle that was passively entrained into the upper mantle flow. While old blobs would be expected to
show a negligible aging effect in 207Pb/204Pb, the E-MORB D3B-1 having low 3He/4He also has very low eHf,
and significantly higher 208Pb*/206Pb* than other basalts from the same ridge segment. This suggests that
the low 3He/4He is a signature of ancient (U1Th)/He fractionation. However, whether 3He/4He of 7.5 RA
reflects the end-member composition of the heterogeneity giving rise to the E-MORB signature is uncertain.
The in situ 3He/4He ratio could be significantly lower than this value, because the extent to which the
3
He/4He in a basalt reflects a discrete mantle heterogeneity depends on the size of the heterogeneity relative to the volume of mantle sampled during the melting process. This volume may be significantly larger
for He than for lithophile tracers such as Hf and Pb, which have diffusivities in mantle phases that are
1032105 times smaller than He at typical temperatures for the upper mantle (1300–1400 C; see Hart et al.
[2008] for discussion). This means that as melt transport evolves from a tree-like structure at the base of the
melting regime [Hart, 1993] upward to the hydrofracture/diking regime, the volume of solid mantle
sampled for He can be larger by two orders of magnitude or more (i.e., approximately the reciprocal of the
square root of the diffusivity ratio, (104)1/2). The process is complex because other factors are also important, including the size of the heterogeneity, the relative concentrations of He versus Pb and Hf in the surrounding depleted mantle and in the enriched heterogeneity, stretching and folding of the heterogeneity
during solid state deformation, the geometry of melting, the melting rate of the heterogeneity versus surrounding peridotite, and magma ascent rate. Hart et al. [2008] showed that the diffusivity of helium prevents long-term survival (>109 years) of a closed system He isotope signature within radiogenically
modified material when it is smaller than a few kilometers. However, the effect of a small heterogeneity on
the ambient mantle can persist for long time periods. The extreme He-Pb-Hf isotope signatures for the
E-MORBs along segment C17 occur over a distance of <10 km, and are consistent with the presence of a
small, ancient (1082109 years) heterogeneity in the underlying mantle. This heterogeneity is embedded in
depleted mantle peridotite but it is enriched in K, U, and Th. For a slab geometry with a thickness of 1 km,
the distance over which diffusion produces a lower 3He/4He in the ambient mantle is <5 km when
DHe 5 10210 m2/s and times are in excess of 200 Myr [Hart et al., 2008]. This leads to 3He/4He of 7.5 RA for
melts sampled over a distance of 10 km or less, similar to the E-MORBs erupted along segment C17.
Differences in the volume of mantle equilibration for He versus lithophile elements during the melt generation process also extend to the presence of refractory mantle domains. Such refractory domains might be
produced by ancient melting events and have been hypothesized to be distributed within the upper mantle
based on ancient Os isotope model ages in abyssal peridotites [Harvey et al., 2006; Liu et al., 2008]. These
refractory domains would be carried passively in the upper mantle flow field and interwoven with both
enriched (recycled) heterogeneities and fertile (lherzolitic) upper mantle but they might not typically contribute to melt generation. In addition, Salters et al. [2011] observed different stacked subparallel Hf-Nd isotope arrays for individual sections of the global ridge system which they suggested originate through
incongruent melting of mantle with a variable lithological makeup. When a heterogeneous mixture comprised of lherzolite, recyled mafic crust and refractory depleted mantle/recycled lithosphere undergoes
melting (with each having a different solidus temperature), the refractory component contributes lithophile
elements through melt-rock reaction rather than via directly producing melt. Given that the volume
sampled for helium is 100 times that for lithophile elements during melting, if such refractory domains
are present then they will be most effectively sampled for 3He/4He compared to any other isotopic tracer.
The possible bimodal 3He/4He distribution in Figure 3 takes on added significance in this regard. The two
3
He/4He subpopulations may represent magmas generated from peridotite 1 pyroxenite melting (<9 RA)
versus pyroxenite-free melting (>9 RA).
4.3.3. Origin and Distribution of Isotopic Heterogeneity Beneath the SEIR
We attribute the lower 3He/4He signatures observed in our study area, and much of the variability downward from 3He/4He 9 RA, to the influence of melting of streaks of enriched mantle embedded within
upper mantle peridotite. This enriched material may have a recycled origin, such as ancient subducted
ocean crust. It may have also been generated during rift initiation and continental breakup as delaminated
lithospheric mantle/lower crust [Mahoney et al., 1992; Janney et al., 2005; Hanan et al., 2004, 2013], or as
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2062
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
small-degree melts trapped in the upper mantle [Mahoney et al., 2002; Russo et al., 2009]. The low eHf of
samples such as GSR I D3B-1 and the He-Hf-Pb isotope relations (Figure 6) point to the involvement of
recycled continental lithosphere. Much of this material was incorporated during the rifting of Gondwana at
180 Ma and in association with mantle plume-continental lithosphere interaction during the eruption of the
Karoo flood basalt province [Janney et al., 2005; Hanan et al., 2013]. This continental component can also be
marked by relatively low 206Pb/204Pb, exemplified best by the MD34 D5 sample (Figure 6), although in the
case of GSR I D3B-1 its 206Pb/204Pb is not distinct from the surrounding MORBs on segment C17 despite its
very low eHf [Hanan et al., 2013]. The most radiogenic Pb isotope signatures, notably the elevated 206Pb/204Pb
(up to 19.6) and 207Pb/204Pb (up to 15.67) in the vicinity of the ASP plateau [Doucet et al., 2004; Nicolaysen
et al., 2007; Janin et al., 2012], are commonly associated with recycled ocean crust [Hanan and Graham, 1996],
or with mantle wedge that was contaminated with melts during subduction [Donnelly et al., 2004]. In those
cases where the SEIR lavas show this radiogenic Pb isotope signature most strongly, they have elevated
3
He/4He rather than the low ratios expected from the presence of ancient recycled material. This may be
accounted for by hybridization of streaks of the recycled material with deep, less degassed mantle. For basalts
from the Austral Islands, Parai et al. [2009] showed that nucleogenic 21Ne/22Ne correlates with radiogenic Pb
(designated HIMU by Zindler and Hart [1986]), but the source 21Ne/22Ne is lower than for the depleted upper
mantle and resembles values for less degassed mantle. Parai et al. [2009] therefore suggested that the
recycled material responsible for the radiogenic Pb isotope signature behaves as an open system, acquiring a
large part of its noble gas signature from equilibration with the surrounding deep mantle (3He/4He at HIMU
localities is typically lower than in MORBs, between 5–7 RA) [Graham et al., 1992; Hanyu and Kaneoka, 1998;
Hanyu et al., 1999; Parai et al., 2009]. The C isotope signature of the ASP hot spot is not as radiogenic in its Pb
isotope character as the HIMU signature of the Australs due to its younger recycling age (Proterozoic or
younger for C versus Archean for HIMU: Hanan and Graham [1996]; Cabral et al. [2013]). We suggest that the
Hanyu and Kaneoka [1998] and Parai et al. [2009] model can be extended to account for the cooccurrence of
elevated 3He/4He in the ASP-influenced lavas having radiogenic Pb isotope compositions.
The present He isotope results considered in light of the isotope variations in Pb-Hf-Nd-Sr therefore indicate
the involvement of four discernible geochemical end-members. These components are depleted (ambient)
upper mantle, recycled ocean crust/mantle wedge (C component), a deep plume component (having elevated 3He/4He), and recycled continental lithosphere (distinguished most readily by low eHf).
A cartoon of the mantle flow field responsible for the isotope signatures along the SEIR is shown in Figure 7.
One possibility is that small domains (blobs) of plume material escape the deep upwelling beneath the ASP
Plateau and eventually become incorporated into the melting region 1000 km away beneath segments L
and C17. While we cannot exclude this model due to possible temporal variability in blob compositions associated with hot spot volcanism, we do not favor it because it does not accurately account for the collective SrNd-Pb isotope characteristics of the lavas in the 88 E region of the SEIR [Mahoney et al., 2002]. A more plausible model is the rare injection of blobs from the mantle transition zone or deeper mantle, which become passively entrained into the upper mantle flow. In this regard, it is noteworthy that the dominant wavelength of
the He isotopic features along the SEIR is 1000 km (see below). This length scale resembles that for undulating, low-viscosity plume-like features deduced from seismic tomography that are present near 400 km depth
and rooted in the deep mantle [French et al., 2013].
4.4. Spectral Analysis of the Length Scales of Helium Isotope Variability
Upper mantle heterogeneities originate through continuous mantle differentiation over Earth history and
are acted upon by mantle stirring. Below we adopt a comparative strategy to deduce quantitatively the
scales of these heterogeneities in the upper mantle. Because the local isotopic composition of the mantle
depends mainly on the advection of heterogeneities by the mantle flow field [Richter and Ribe, 1979], our
analysis has implications for mantle convective processes. This convection is vigorous on geological timescales (occurring at high Rayleigh number), three dimensional and time dependent. It is forced at plate
scales but nonlinear interactions can spread the energy out over a range of intermediate and smaller scales.
To evaluate the length scales of helium isotope variability in a quantitative way, we adopted four different
approaches; (1) the periodogram [Lomb, 1976; Scargle, 1982], (2) red-noise spectral estimation (REDFIT of
Schulz and Mudelsee [2002]), (3) the multitaper method (MTM) of spectrum estimation [Thomson, 1982; Lees
and Park, 1995], and (4) continuous wavelet transform [Grinsted et al., 2004]. Details are described in the
supporting information.
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2063
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
4.4.1. Periodogram
Periodogram analysis of the iso1000
0
tope variations along midL / C17
ocean ridges from the Atlantic
10
ASP
Ocean was previously shown to
Plateau
C13
contain two spectral types
0
[Agranier et al., 2005]. The first
is a hot spot type, containing
distinct lobes at low wave number k (large distance), plus a
sharp cutoff near k0.2 ( )21
(550 km). Beyond this wave
number (at higher frequency),
the signal appeared to be
dominated by white noise. This
spectral type is exhibited by
3
He/4He, 87Sr/86Sr, and
400
206
Pb/204Pb [Agranier et al.,
red-melt
3
4
orange-primitive (high He/ He, deep mantle) component
2005] along mid-ocean ridges
grey-recycled (ocean crust, continental lithosphere) components
from the Atlantic. The second
light green-ambient mantle
spectral type was referred to as
the Batchelor regime. It exhibits
Figure 7. A cartoon of one possible model for mantle heterogeneity, mantle flow, and melta long-wavelength structure
ing beneath the SEIR. Length scales are approximate. Note the break in scale between the
crust and mantle. The light background depicts ambient mantle (peridotite) and orange
similar to type 1, but it has a
depicts domains of mantle having elevated 3He/4He (>15 RA, as measured at the ASP Plateau)
continuous decrease with
that are ultimately derived from the deep mantle. Gray folded strands depict ancient compoincreasing wave number (which
nents, either recycled (continental and oceanic) lithosphere or frozen melt (trapped during
initial ocean basin rifting), attested to by the isotope relationships for Hf-Pb-Sr-Nd described
can be approximated by a
in Mahoney et al. [2002], Nicolaysen et al. [2007], Graham et al. [2006], and Hanan et al. [2013].
power law exponent of 21).
Red depicts melt and is highly schematic.
This spectral type is exhibited
by Nd and Hf isotopes, and the second component of a Pb isotope principal component analysis [Agranier
et al., 2005]. The type 2 spectrum was hypothesized to be produced by continual reduction in the size of
mantle heterogeneities caused by the stretching and refolding that occurs during mantle convection.
Distance (km)
Depth (km)
2000
The Lomb-Scargle periodogram for 3He/4He variations along the SEIR is shown in Figure 8. We performed
the analysis using the full SEIR data set (248 sample locations), data from the GEISEIR I and II study area
(81 E–101 E, 179 locations), and the most densely sampled area (89.7 E–101 E, 111 locations). Similar to
what was observed previously along the Mid-Atlantic Ridge [Agranier et al., 2005], lobes are present for
3
He/4He at wave numbers of 0.001 and 0.002 km21 (1000 and 500 km, respectively) in both the full data
set and the GEISEIR study region. In the most densely sampled region (Figure 8c), the periodogram also
begins to reveal possible structures at larger wave number (0.003 and 0.004, i.e., 330 and 250 km, respectively). It is difficult to interpret the meaning of the higher wave number peaks in Figure 8c as they do not
appear to be statistically significant in the full data analysis. They may be artificial overtones of the lower
wave number peaks near 1000 and 500 km.
4.4.2. Red-Noise Spectral Estimation (REDFIT)
Figure 9 shows a REDFIT analysis of 3He/4He along the SEIR. The data have been analyzed in two ways. The
first treatment focuses on the GEISEIR I and II study region (n 5 172), while the second is for the full data set
(248 sample locations). The results in Figure 9 were obtained using an overlapping segment approach, but
other choices for number of segments and window shape (rectangular, triangular) did not produce outcomes that were significantly different. The results of this analysis reveal a significant peak at large length
scale. For the full data set, this power is mostly concentrated in a 1000 km peak (significant at >99% confidence level), much like the Lomb-Scargle periodogram. An 500 km peak is also present in this case but it
is not as statistically significant (Figure 9b). Significant (>95% confidence level) peaks are also observed in
the full data set at length scales of 50–70 km. For the more restricted, densely sampled data (1825 km of
ridge, see Figure 9a) the 500 km peak is significant at the 99% confidence level. Peaks are significant at
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2064
Geochemistry, Geophysics, Geosystems
Length Scale (km)
1000
500
250
100
10
All Data
n=248
z=4600 km
(1 per 18.5 km)
10
A
10
0.0009
1111 km
0.0021
476 km
10
0
0.001
0.002
500
250
10
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
50
100
Moderate Density
n=179
z=1825 km
(1 per 10.3 km)
10
10.1002/2014GC005264
500 and 30 km for the
densely sampled region.
Overall, the REDFIT results
resemble those obtained
from the periodogram analysis. Notably, when only the
GEISEIR study area is considered we observe a statistically significant peak near 30
km, identical to the characteristic length scale deduced
from the bimodal distribution of Hf and Pb isotopes in
the GEISEIR I region [Hanan
et al., 2013].
4.4.3. Multitaper Method
(MTM) of Spectral
10
Estimation
0.0020
Application of MTM requires
498 km
equally spaced data. To
10
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
0.018
0.02
achieve this, we preprocessed the helium isotope
500
250
50
100
10
data series utilizing a GausHigh Density
n=111
sian interpolation applied
z=1000 km
10
over three different subsets
(1 per 9 km)
of the data in order to
C
reduce effects of aliasing
10
0.0029
369 km 0.0041
plus any data gaps on spec0.0018
242 km
565 km
tral estimation. The prepro10
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
0.018
0.02
cessing is illustrated
graphically in Figure 10a. In
Wave Number (km-1)
treatment (a), the He isotope
Figure 8. Periodograms of the SEIR 3He/4He variations. (a) All Data (248 locations). (b) Samples
data were interpolated
from 80.98 E–101 E (179 locations). (c) Samples from 89.69 E–101 E (111 locations). The
along-axis between 2300
dashed lines represent the 95% confidence level above which the peaks in the spectrum are
deemed significant.
and 13140 km at an interval
i of 40 km using a Gaussian
half-width w of 60 km. The total number of interpolated locations n is 87. In (b), the along-axis distance
ranged from 380 to 2140 km, w 5 40 km, i 5 20 km, and n 5 89. In (c), the along-axis distance ranged from
700 to 2010 km, w 5 20 km, i 5 10 km, and n 5 132. The application of the Gaussian filter leads to a relatively sharp cutoff in spectral estimation at a wave number between the full-width and half-width. Examples
for the estimated power spectrum and F-test values for each of the three cases (case (a) using five tapers,
cases (b) and (c) using three tapers) are illustrated in Figures 10b–10d.
0.0007
1538 km
B
For the longest part of the data series (subset a), the spectrum estimation reveals periodicities in relative
power at 1000 and 500 km (Figure 10b). The 500 km peak is not well resolved in this estimation of the
spectrum but the F-value suggests it is significant at >95% confidence. Although a broad spectral peak also
occurs near 200 km, it is not statistically significant. Elevated F-values also occur at higher wave number
(wave number 0.4, length scale 100 km) but this is close to the Gaussian filter cutoff length scale (60–
120 km) and the associated spectral peaks are also quite subdued. The occurrence of relatively high
F-values at higher wave number where peaks are absent in the spectrum can be caused by random noise
related to peculiarities in sampling [Lees and Park, 1995]. From case (a), we conclude that significant power
is found in the overall data series at 1000 km, while a weaker peak may also be present near 500 km.
Spectrum estimation for the shorter treatments of the data series (subsets b and c) reveal significant relative
power at a length scale between 425 and 650 km in each case (Figures 10c and 10d). Evidence for an
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2065
Geochemistry, Geophysics, Geosystems
Length Scale (km)
200 100
50
20
25
81°-100°E (n=172)
Power
Theor.AR(1)
chi2 95%
chi2 99%
τ=21 km
Power
10
1
bandwidth
A
0.1
0.00
1,000
1000
0.01
200
0.02
100
0.03
0.04
0.05
50
All SEIR (n=248)
τ=31 km
bandwidth
Power
100
10
B
1
0.00
0.01
0.02
-1
Wave Number (km )
Figure 9. Red-noise spectral estimation for 3He/4He (after Schulz and Mudelsee [2002]). (a)
Upper plot shows results for the densely sampled portion of the SEIR (z 5 1825 km, from
80.98 E–100 E; n 5 172 locations). (b) Lower plot shows results for all SEIR data (z 5 4600
km, n 5 248 locations). Power is expressed in units proportional to the square of the amplitudes of the sinusoids present in the data. The illustrated example uses a Welch overlapping segment average (WOSA) with 50% overlap for each of three segments (see
supporting information for details). Other choices for window shape (rectangular, triangular) and number of segments do not produce results very different from those depicted
here. For the cases here, in the dense area s 5 21 km with bandwidth 5 0.00197 km21. For
all ridge data s 5 31 km and bandwidth 5 0.00070 km21. The solid gray curve depicts the
theoretical autoregressive (AR(1)) red-noise model (the null hypothesis against which peak
significance is tested). The dashed lines depict 95% and 99% significance levels.
10.1002/2014GC005264
500 km peak is, therefore,
found for all the cases we investigated but it is not well resolved
in the treatment using subset (a).
The peak near 500 km in the
more densely sampled data
series suggests it may be a fundamental length scale of He isotope variability in the mantle,
rather than a harmonic overtone
of the 1000 km scale that was
produced from the choice of
data windows or interpolation
schemes.
In case (b), the strongest peaks
are observed at 150 and 100
km, with greater than 99% confidence level for each (Figure 10c).
For the most densely sampled
case (c) explored here we also
observe peaks at 100 and 40
km. The latter length scale exhibits the most significance in this
case (c) in terms of its high Fvalue, and it resembles the 30
km signal previously observed in
the REDFIT spectrum estimation
using unequally spaced data. This
40 km peak is also close to the
Gaussian filter cutoff length (20–
40 km). In addition, in cases (b)
and (c), the exact position and significance of the peaks at 150,
100, and 40 km can shift
somewhat with the choice of the
number of tapers, indicating that
the data series show some tendencies to nonstationarity, i.e.,
they have changing 3He/4He
means and variances along the
length of the SEIR. This is an inherent limitation in our ability to
make more extensive quantitative
arguments from these simple
spectral analysis approaches.
We conclude from the MTM analysis that fundamental peaks in the He isotope variability occur along the
SEIR at approximately 1000, 500, 150–100, and 40 km length scales.
4.4.4. Continuous Wavelet Transform
We employed the algorithm developed by Grinsted et al. [2004] to estimate the CWT for 3He/4He variations
along the SEIR. Data were interpolated at 10 km intervals along the SEIR using a piecewise cubic hermite
spline. The CWT for 3He/4He along the SEIR is shown in Figure 11. Pervasive power is observed at 1000
and 500 km length scales along much of the SEIR. These length scales are similar to those discerned from
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2066
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Distance (km)
a
MTM
Smoothed
Periodogram
10-2
Power
He/4He (R/RA)
Gaussian Interpolation
a. w=60 km, i=40 km
b. w=40 km, i=20 km
c. w=20 km, i=10 km
100
200
800
10-3
3
b
10-4
c
14
F values
A
12
10
B
95% c.l.
4
2
8
6
-1000
0
1000
2000
0.0
3000
Distance (km)
100
10-3
0.4
50
0.5
25
MTM
Smoothed
Periodogram
10-2
Power
Power
MTM
Smotthed
Periodogram
0.3
Distance (km)
200
50
10-2
0.2
Wave Number
Distance (km)
400
0.1
10-3
10-4
10-4
C
D
F values
F values
25
99% c.l.
20
15
10
95% c.l.
10
95% c.l.
5
5
0.0
0.1
0.2
0.3
0.4
0.5
Wave Number
0.0
0.1
0.2
0.3
0.4
0.5
Wave Number
Figure 10. MTM analysis of the SEIR He isotope variations. (a) Gaussian interpolation of the variations. The data series was treated in three
different ways. In treatment (a), the He isotope data were interpolated along-axis between 2300 and 13140 km, at an interval (i) of 40 km
using a Gaussian filter having a half-width (w) of 60 km. The total number of interpolated locations (n) is 87. The MTM applied to treatment
(a) shown here utilized five tapers. In treatment (b), the along-axis distance ranged from 380 to 2140 km, w 5 40 km, I 5 20 km, and
n 5 89. The MTM applied to treatment (b) shown here utilized three tapers. In treatment (c), the along-axis distance ranged from 700 to
2010, w 5 20 km, i 5 10 km, and n 5 132. The MTM applied to treatment (c) shown here utilized three tapers. (b–d) show the spectrum
estimates from the MTM, and the corresponding confidence levels from the computed F-test values. The smoothed periodograms for the
same data treatments are shown for comparison. Wave number range is from zero to the Nyquist limit (5 sampling interval divided by
0.5), corresponding to 80, 40, and 20 km for series a, b, and c, respectively.
the various periodogram and REDFIT treatments discussed earlier. One advantage of the wavelet analysis is
that it provides a depiction of how power at different periods (length scales) is localized across the spatial
domain. In addition to the power at 500 and 1000 km length scales, the CWT for 3He/4He displays localized power at length scales of 75–250 km in the vicinity of the ASP hot spot, and at 50–100 km for
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2067
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
32
Length Scale (km)
32
16
8
4
64
2
1
128
256
1/2
1/4
1/8
1/16
1/32
512
1024
12
10
3
He/4He (R/RA)
14
8
6
-1000
0
1000
2000
3000
Along-Axis Distance from St. Paul (km)
Figure 11. Continuous wavelet transform (CWT) and the 3He/4He variations along the SEIR. The vertical axis on the upper plot is a natural
logarithm scale of the length scale (km) with each unit corresponding to a doubling in size. Signal power is displayed in color according to
the scale shown at the right. The cone of influence of boundary effects is outlined in gray. Thick contours outline 5% significance levels.
basalts erupted at along-axis distances of 800–1100 km. This latter area corresponds to the local 3He/4He
peak that is clearly visible in the along-axis data itself (Figures 2 and 4).
In our treatment, we also determined the Cross-Wavelet Transform (XWT) and Wavelet Transform Coherence (WTC) between 3He/4He and Pb and Hf isotopes (see supporting information Figures S2–S4) and
between 3He/4He and axial depth. These additional visualizations can be used to investigate power and
phase relationships between isotopes and bathymetry along the SEIR. The XWT is constructed from two
CWT’s [Grinsted et al., 2004]. In a case where the XWT phase relationship is locked within a narrow range, it
seems likely that variations at that length scale should have a similar mechanistic origin.
Figure 12a shows the XWT between 3He/4He and axial depth. A significant common power is observed at
500 km and between 75 and 250 km in the region of the ASP plateau (Figure 12a). At these shorter periods, it is notable that the relationship between He isotopes and axial depth is not simply in-phase or out-ofphase. This contrasts with the 1000 km length scale along nearly the entire SEIR in which there is a remarkable in-phase relationship (Figure 12a). The WTC is plotted in Figure 12b and helps resolve more clearly the
1000 and 500 km scale relationships between He isotopes and axial depth hinted at by the XWT. The
1000 km length scale coherence pervades the SEIR, although it is rendered less certain close to the boundaries of the sampling domain by the edge effect delineated by the cone of influence (COI). The 500 km scale
encompasses the western one third to one half of the study region. The coherence at 1000 km length scale
between 3He/4He and axial depth is well defined in the WTC, while at 500 km length scale the phase relationship is shifted by about 45 . Local islands of coherence at hundred km length scales may also be present
in the distal eastern sections of the SEIR at along-axis distance of 2500 km (Figure 12b). These longwavelength scales are similar to the very largest scales of ridge segmentation, which is defined by the persistence of fracture zone anomalies in satellite gravity data [Small et al., 1999]. Those features appear to delineate
tectonic corridors that have been stable since 38 Ma when the SEIR migrated over the Kerguelen hot spot.
The CWT results support our earlier inference based on the periodogram, REDFIT and MTM results that key
3
He/4He variability is found at 1000 and 500 km length scales. Significant localized variability also occurs
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2068
Geochemistry, Geophysics, Geosystems
16
32
Length Scale (km)
8
64
4
10.1002/2014GC005264
at 75–250 km length scales, primarily in the vicinity of the ASP
hot spot and at 900 km east of
St. Paul island.
2
4.5. Implications for Mantle
Convection
4.5.1. 3He/4He Textures in the
256
1/2
Upper Mantle
1/4
It has been recognized for 30
512
years that the Indian Ocean
1/8
upper mantle has a distinct long1024
1/16
A
term evolutionary history [Dupre
and Allègre, 1983; Hart, 1984;
-1000
0
1000
2000
3000
Hamelin and Allègre, 1985; HameAlong-Axis Distance from St. Paul (km)
lin et al., 1986]. The Pb-Nd-Sr iso1
topic variability for Indian Ridges
32
0.9
at a sampling interval of 50 km
0.8
64
was evaluated using the periodo0.7
gram approach and interpreted
0.6
128
by Meyzen et al. [2007] to result
0.5
from a refolded structure pro256
0.4
duced largely by refilling of the
0.3
512
upper mantle through localized
0.2
upwellings of deep mantle. The
1024
0.1
B
involvement of a deep mantle
0
source, inferred in Meyzen et al.
-1000
0
1000
2000
3000
[2007] and supported by the
Along-Axis Distance from St. Paul (km)
helium isotope results presented
here, is consistent with the comFigure 12. (a) Cross-wavelet transform (XWT) and (b) Wavelet transform coherence
(WTC) of 3He/4He and axial depth. The vertical axis is a natural logarithm scale of the
mon DUPAL nature of both
length scale (km) with each unit corresponding to a doubling in size. Signal power (a)
ocean island basalts (OIBs) and
and coherence (b) are contoured in color according to the scale shown at the right.
MORBs in the Indian Ocean [CasArrows depict the phase relationship between 3He/4He and axial depth as a function of
length scale (vertical axis) and location along the SEIR (horizontal axis). Right-pointing
tillo, 1988]. It is notable, however,
arrows indicate an in-phase relationship and left-pointing an antiphase relationship.
that this isotopic flavor is also
3
4
Downward pointing arrows correspond to where axial depth leads He/ He, while
demonstrably associated with
upward pointing corresponds to where depth lags. Thick contours outline 5% significance levels.
shallow recycling of continental
lithosphere and/or lower crust
elsewhere in the mantle, such as beneath the SWIR, the southern MAR, and the Arctic [Mahoney et al., 1992;
Janney et al., 2005; Kamenetsky et al., 2001; Class and le Roex, 2006; Goldstein et al., 2008].
128
Length Scale (km)
1
A basic question up to now has been whether the spectra of mantle heterogeneities beneath the Atlantic
[Agranier et al., 2005] are different from those beneath the Pacific and Indian oceans. One simple expectation might be that differences would be found that relate to tectonic history of the continents bordering
the basins or to the speed of mantle convection and the influence of mantle plumes. The 3He/4He character
for the Indian Ocean deduced from the spectral analysis of the SEIR (Figures 8–11) shows peaks near 1000
and 500 km. Because this large length scale resembles that in Atlantic periodograms it supports the notion
that these two distinct regions of the upper mantle have similar helium isotopic textures. The similar
3
He/4He textures further suggest that upwelling combined with stirring have produced broadly similar He
isotope dispersions in the upper mantle around the globe.
4.5.2. A Comparison to the Southern East Pacific Rise
We suggest that 3He/4He variability along the mid-ocean ridge system is marked by a similar longwavelength character and explore this further by an additional simple comparison. We choose the Pacific
because we have already shown through spectral analysis that ridge systems in the Indian and Atlantic
Oceans [Agranier et al., 2005] have similar long-wavelength He isotope textures. We compare the along-
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2069
Geochemistry, Geophysics, Geosystems
East Pacific Rise
14
He/4He (R/RA)
13
Easter
Microplate
12
11
10
3
9
8
ASP
Plateau
14
Southeast Indian Ridge
He/4He (R/RA)
13
12
11
3
10
9
8
-500
0
500
1000
1500
Distance (km)
Figure 13. A comparison of MORB He isotope variations from the southern East Pacific
Rise and the Southeast Indian Ridge. EPR data are from Poreda et al. [1993] and Kurz et al.
[2005]; SEIR data are from Graham et al. [1999, 2001], Nicolaysen et al. [2007], and this
study.
10.1002/2014GC005264
axis variations for the SEIR with
those for a section of the southern East Pacific Rise (Figure 13).
Although this 2000 km long
section of the EPR was sampled
at approximately 40% of the
sampling density of the SEIR, the
key long-wavelength features
are evident. The selected EPR
region encompasses the Easter
Microplate where MORB signatures in He-Pb-Sr-Nd-Hf isotopes
are influenced by input from the
nearby Easter mantle plume
[Hanan and Schilling, 1989; Fontignie and Schilling, 1991; Poreda
et al., 1993; Kingsley et al., 2007],
analogous to the ASP hot spot
along the SEIR. The selected EPR
section also includes the
superfast-spreading region from
13 S to 23 S with prominent HePb-Nd-Sr isotope anomalies
[Mahoney et al., 1994; Kurz et al.,
2005]. Figure 13 shows a remarkable similarity in both the amplitude and period of the 3He/4He
variations for these regions
despite the large difference in
their ridge spreading rates.
It could be a simple coincidence that these two sections of the mid-ocean ridge system seem so similar but
we consider that unlikely. Each area is marked by the upwelling of a mantle plume (ASP versus Easter) that
is sufficiently intense to produce ocean island volcanism. Each ridge section influenced by the nearby mantle plume shows elevated 3He/4He extending along the ridge for 300–400 km. Each ridge system also contains a broad auxiliary peak located 900–1000 km away from the primary hot spot region (the 88 E region
of the SEIR versus the superfast EPR). Each auxiliary region has an elevated 3He/4He anomaly (10–11 RA)
that is marked by a basal width of 200–400 km. Increased melt production, expressed as shallower axial
depth, plus the presence of several near-axis seamounts in the SEIR case, is also broadly associated with
each of these auxiliary regions.
There are also important differences between the two systems when they are compared in detail, and these
point to additional complexity. The Easter hot spot is 500–1000 km away from the ridge axis [Hanan and
Schilling, 1989], considerably farther than the case for the ASP hot spot (100 km). The tectonic settings are
different, with the Easter Microplate being comprised of two rift branches along its east and west boundaries while the ASP hot spot is located beneath a broad submarine plateau. The 3He/4He relationships with
Pb-Sr-Nd isotopes for each of the auxiliary regions in this comparison is not explained by simple mixing
models involving the respective hot spot sources and background MORB mantles [Mahoney et al., 2002;
Kurz et al., 2005]. It is further noteworthy that the 3He/4He-K/Ti relationships observed along the SEIR, in
which the E-MORB character is associated with lower 3He/4He, differ from those observed for the southern
EPR where E-MORBs sometimes carry elevated 3He/4He [Kurz et al., 2005].
Despite the aforementioned complexities, the SEIR-EPR similarity can be explained by similar auxiliary upwellings from depth. These upwellings may originate from either the lower mantle or deeper parts of the
upper mantle, and they appear to be located 1000 km away from the hot spots interacting with the
respective spreading ridge systems. These auxiliary upwellings may be mantle blobs [Allègre et al., 1984;
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2070
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Gurnis, 1986b; Manga, 1996; Becker et al., 1999; Vlastelic et al., 2002]. If the blob flux to the subridge mantle
is relatively low and blob size is of the order of a ridge segment or smaller then significant 3He/4He variability might occur on time scales of 1062107 years. This resembles the case study for MORBs and near-ridge
seamounts in the South Atlantic at 26 S, where 3He/4He along the ridge is low (7.3–7.7 RA) but older seamount lavas on both sides of the ridge axis extend to 11 RA [Graham et al., 1996]. In this context, the recent
waveform seismic tomography study by French et al. [2013] also reveals the global presence of undulating,
plume-like features that could produce blobs. These features have a wavelength of the order of 103 km and
are located at 400 km depth in the upper mantle but apparently are rooted in the deep mantle. They are
manifested as low-viscosity fingers and are present beneath the fast moving Pacific Plate where secondary
convection rolls are likely to develop, but also beneath the slower moving Atlantic and Indian Plates where
the presence of such rolls was thought to be unlikely.
5. Conclusions
Acknowledgments
This paper is dedicated to the memory
of John Mahoney. D.G. and B.H. were
supported by the Marine Geology &
Geophysics program of the U.S.
National Science Foundation. The OSU
Noble Gas Lab was established
through a Major Research
Instrumentation grant from MG&G.
The Keck Foundation contributed to
establishment of the Isotope
Geochemistry clean rooms at SDSU.
J.B.T. and F.A. received support from
the French Agence Nationale de la
Recherche (grant ANR-10-BLAN-0603
M&Ms—Mantle Melting—
Measurements, Models, Mechanisms).
C.H. acknowledges the French Institut
Paul Emile Victor (IPEV) for ship time
and staff onboard the ship, and the
French Institut National des Sciences
de l’Univers for financial support
through the program Campagne a la
mer. Anne Briais and Marcia Maia
provided superb bathymetric mapping
plus magnetics and gravity expertise
during GEISEIR I and II that led to the
highly successful rock sampling of the
ridge axis. We are grateful to captains
F. Duch^ene and C. Garzon of the R/V
Marion Dufresne, and their crews
during the MD171 (GEISEIR I) and
MD175 (GEISEIR II) scientific
expeditions. Figure 1 was produced
using GeoMapApp (http://www.
marine-geo.org/geomapapp), which is
supported by the National Science
Foundation and internal funding from
Lamont-Doherty Earth Observatory.
Some data analysis was carried out
using computer software written for
R plus the PAST software
MATLABV
package (PAleontological Statistics,
available at http://folk.uio.no/
ohammer/past/; Hammer et al. [2001]).
We thank Frank Spera and John
Lupton for valuable discussions, Alan
Mix for providing the Gaussian
interpolation software and help with
many aspects of the spectral analysis,
and Sujoy Mukhopadhyay and an
anonymous reviewer for constructive
reviews.
GRAHAM ET AL.
The SEIR is a regional-scale analogue of the global mid-ocean ridge system, and was previously known to
show significant isotopic variability due to ancient heterogeneities embedded in the upper mantle. A new
sample suite, collected at 5–10 km intervals for a distance of >1600 km between 81 E and 100 E, shows a
3
He/4He range from 7.5 to 10.2 RA that encompasses more than half the known range in MORBs erupted
away from ocean island hot spots. Abrupt transitions occur in this record over lateral distances of 10 km or
less, indicating significant small-scale (<10 km) helium isotope variability in the upper mantle beneath the
SEIR. The higher 3He/4He ratios in this densely sampled region are associated with depleted MORB compositions (low K/Ti), while the lower 3He/4He ratios are associated with E-MORB compositions (K/Ti > 0.2).
Relations between 3He/4He and the radiogenic isotopes of Pb and Hf reveal that the E-MORB He isotope signature is related to melting of ancient mantle heterogeneities that originated as recycled continental and
oceanic lithosphere, or as incipient melts trapped during continental rifting and the early stages of opening
of the Indian Ocean basin. Collectively the isotopes indicate involvement of at least four distinct mantle
end-members; depleted upper mantle peridotite, a deeply derived plume component (spatially restricted
to the influence of the Amsterdam-St. Paul hot spot), and recycled continental and oceanic lithosphere.
Melting of mantle containing a few percent garnet pyroxenite or eclogite leads to lower 3He/4He ratios,
while 3He/4He ratios above 9 RA may arise from melting of pyroxenite-free or eclogite-free mantle.
Spectral analysis of the He isotope variations reveals strong peaks at 1000 and 500 km, and less prominent peaks at 100–150 km and 30–40 km. Mesoscale flow features in the mantle may account for the
100–150 km length scales. The 30–40 km length scale is identical to that deduced from Hf and Pb isotope variations in the same sample suite, and appears best accounted for by melting of eclogite or garnet
pyroxenite that is heterogeneously distributed within the subridge mantle. The longest wavelength features
of 3He/4He variability in the Indian Ocean upper mantle resemble those beneath the Atlantic Ocean, previously described as a hot spot-type spectrum. 3He/4He variations along the SEIR are also remarkably similar
in amplitude and wavelength to those along the southern East Pacific Rise. The 3He/4He and axial-depth
variations along the SEIR are highly coherent at 500–1000 km. This indicates a close coupling between melt
production and 3He/4He, probably due to regional variations in upper mantle temperature driven by thermal upwelling from depth.
References
Agranier, A., J. Blichert-Toft, D. W. Graham, V. Debaille, P. Schiano, and F. Albarede (2005), The spectra of isotopic heterogeneities along the
Mid-Atlantic Ridge, Earth Planet. Sci. Lett., 238, 96–109.
Albarède, F. (2005), The survival of mantle geochemical heterogeneities, in Earth’s Deep Mantle: Structure, Composition and Evolution,
edited by R. D. van der Hilst, J. Bass, J. Matas, and J. Trampert, pp. 27–46, AGU, Washington, D. C.
Allègre, C. J., and D. L. Turcotte (1986), Implications of a two-component marble-cake mantle, Nature, 323, 123–127.
Allègre, C. J., O. Br
evart, B. Dupr
e, and J.-F. Minster (1980), Isotopic and chemical effects produced in a continuously differentiating convecting Earth mantle, Philos. Trans. R. Soc. London A, 297, 447–477.
Allègre, C. J., B. Hamelin, and B. Dupr
e (1984), Statistical analysis of isotopic ratios in MORB: The mantle blob cluster model and the convective regime of the mantle, Earth Planet. Sci. Lett., 71, 71–84.
Anderson, D. L. (2006), Speculations on the nature and cause of mantle heterogeneity, Tectonophysics, 416, 7–22.
Baran, J. M., J. R. Cochran, S. M. Carbotte, and M. R. Nedimovic (2005), Variations in upper crustal structure due to variable mantle temperature along the Southeast Indian Ridge, Geochem. Geophys. Geosyst., 6, Q11002, doi:10.1029/2005GC000943.
C 2014. American Geophysical Union. All Rights Reserved.
V
2071
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Barling, J., and S. L. Goldstein (1990), Extreme isotope variations in Heard Island lavas and the nature of mantle reservoirs, Nature, 348, 59–
62.
Barling, J., S. L. Goldstein, and I. A. Nicholls (1994), Geochemistry of Heard Island (Southern Indian Ocean): Characterization of an enriched
mantle component and implications for enrichment of the sub-Indian ocean mantle, J. Petrol., 35, 1017–1053.
Becker, T. W., J. B. Kellogg, and R. J. O’Connell (1999), Thermal constraints on the survival of primitive blobs in the lower mantle, Earth
Planet. Sci. Lett., 171, 351–365.
Brandenburg, J. P., E. H. Hauri, P. E. van Keken, and C. J. Ballentine (2008), A multiple-system study of the geochemical evolution of the
mantle with force-balanced plates and thermochemical effects, Earth Planet. Sci. Lett., 276, 1–13, doi:10.1016/j.epsl.2008.08.027.
Cabral, R. A., M. G. Jackson, E. F. Rose-Koga, K. T. Koga, M. J. Whitehouse, M. A. Antonelli, J. Farquhar, J. M. D. Day, and E. H. Hauri (2013),
Anomalous sulphur isotopes in plume lavas reveal deep mantle storage of Archaean crust, Nature, 496, 490–494, doi:10.1038/
nature12020.
Carbotte, S. M., M. Marjanovic, H. Carton, J. C. Mutter, J. P. Canales, M. R. Nedimovic, S. Han, and M. R. Perfit (2013), Fine-scale segmentation
of the crustal magma reservoir beneath the East Pacific Rise, Nat. Geosci., 6, 866–870, doi:10.1038/NGE01933.
Castillo, P. (1988), The Dupal anomaly as a trace of the upwelling lower mantle, Nature, 336, 667–670.
Chauvel, C., and J. Blichert-Toft (2001), A hafnium and trace element perspective on melting of the depleted mantle, Earth Planet. Sci. Lett.,
190, 137–151.
Christensen, U. R., and A. W. Hofmann (1994), Segregation of subducted oceanic crust in the convecting mantle, J. Geophys. Res., 99,
19,867–19,884.
Class, C., and A. P. le Roex (2006), Continental material in the shallow mantle—How does it get there?, Geology, 34, 129–132.
Cochran, J. R., J.-C. Semp
er
e, and S. S. Team (1997), The Southeast Indian Ridge between 88 E and 118 E: Gravity anomalies and crustal
accretion at intermediate spreading rates, J. Geophys. Res., 102, 15,463–15,487.
Donnelly, K. E., S. L. Goldstein, C. H. Langmuir, and M. Spiegelman (2004), Origin of enriched ocean ridge basalts and implications for mantle dynamics, Earth Planet Sci. Lett., 226, 347–366.
Dosso, L., H. Bougault, P. Beuzart, J.-Y. Calvez, and J.-L. Joron (1988), The geochemical structure of the South-East Indian Ridge, Earth Planet.
Sci. Lett., 88, 47–59.
Doucet, S., D. Weis, J. S. Scoates, V. Debaille, and A. Giret (2004), Geochemical and Hf-Pb-Sr-Nd isotopic constraints on the origin of the
Amsterdam-St. Paul (Indian Ocean) hotspot basalts, Earth Planet. Sci. Lett., 28, 179–195.
Doucet, S., J. S. Scoates, D. Weis, and A. Giret (2005), Constraining the components of the Kerguelen mantle plume: A Hf-Pb-Sr-Nd isotopic
study of picrites and high-MgO basalts from the Kerguelen Archipelago, Geochem. Geophys. Geosyst., 6, Q04007, doi:10.1029/
GC2005000806.
Doucet, S., M. Moreira, D. Weis, J. S. Scoates, A. Giret, and C. J. Allègre (2006), Primtive neon and helium isotopic compositions of high-MgO
basalts from the Kerguelen Archipelago, Indian Ocean, Earth Planet. Sci. Lett., 241, 65–79, doi:10.1016/j.epsl.2005.10.025.
Douglas Priebe, L. M. (1998), Geochemical and petrogenetic effects of the interaction of the Southeast Indian Ridge and the AmsterdamSaint Paul hotspot, MS thesis, 132 pp., Oregon State Univ., Corvallis.
Dupr
e, B., and C. J. Allègre (1983), Pb-Sr isotope variation in Indian Ocean basalts and mixing phenomena, Nature, 303, 142–146.
Fontignie, D., and J.-G. Schilling (1991), 87Sr/86Sr and REE variations along the Easter Microplate boundaries (South Pacific): Application of
multivariate statistical analyses to ridge segmentation, Chem. Geol., 89, 209–241.
French, S., V. Lekic, and B. Romanowicz (2013), Waveform tomography reveals channeled flow at the base of the oceanic asthenosphere,
Science, 342, 227–230, doi:10.1126/science.1241514.
Gale, A., S. Escrig, E. J. Gier, C. H. Langmuir, and S. L. Goldstein (2011), Enriched basalts at segment centers: The Lucky Strike (37 170 N) and
Menez Gwen (37 500 N) segments of the Mid-Atlantic Ridge, Geochem. Geophys. Geosyst., 12, Q06016, doi:10.1029/2010GC003446.
Gast, P. W., G. R. Tilton, and C. E. Hedge (1964), Isotopic composition of lead and strontium from Ascension and Gough Islands, Science,
145, 1181–1185.
Goldstein, S. L., G. Soffer, C. H. Langmuir, K. A. Lehnert, D. W. Graham, and P. J. Michael (2008), Origin of a ‘southern hemisphere’ geochemical signature in the Arctic upper mantle, Nature, 453, 89–94.
Graham, D. W. (2002), Noble gas isotope geochemistry of mid-ocean ridge and ocean island basalts; characterization of mantle source reservoirs, in Noble Gases in Geochemistry and Cosmochemistry, edited by D. Porcelli, R. Wieler, and C. J. Ballentine, pp. 247–318, Miner. Soc.
Am., Washington, D. C.
Graham, D. W., S. E. Humphris, W. J. Jenkins, and M. D. Kurz (1992), Helium isotope geochemistry of some volcanic rocks from Saint Helena,
Earth Planet. Sci. Lett. 110, 121–131.
Graham, D. W., P. P. Castillo, J. E. Lupton, and R. Batiza (1996), Correlated helium and strontium isotope ratios in South Atlantic near-ridge
seamounts and implications for mantle dynamics, Earth Planet. Sci. Lett., 144, 491–503.
Graham, D. W., K. T. M. Johnson, L. M. Douglas Priebe, and J. E. Lupton (1999), Hotspot-ridge interaction along the Southeast Indian Ridge
near Amsterdam and St. Paul Islands: Helium isotope evidence, Earth Planet. Sci. Lett., 167, 297–310.
Graham, D. W., J. E. Lupton, F. J. Spera, and D. M. Christie (2001), Upper mantle dynamics revealed by helium isotope variations along the
Southeast Indian Ridge, Nature, 409, 701–703.
Graham, D. W., J. Blichert-Toft, C. J. Russo, K. Rubin, and F. Albarede (2006), Cryptic striations in the upper mantle revealed by hafnium isotopes in Southeast Indian Ridge basalts, Nature, 440, 199–202.
Grinsted, A., J. C. Moore, and S. Jevrejeva (2004), Application of the cross wavelet transform and wavelet coherence to geophysical time
series, Nonlinear Processes Geophys., 11, 561–566.
Gurnis, M. (1986a), Quantitative bounds on the size spectrum of isotopic heterogeneity within the mantle, Nature, 323, 317–320.
Gurnis, M. (1986b), Stirring and mixing in the mantle by plate-scale flow: Large persistent blobs and long tendrils coexist, Geophys. Res.
Lett., 13, 1474–1477.
Hamelin, B., and C.-J. Allègre (1985), Large-scale regional units in the depleted upper mantle revealed by an isotope study of the SouthWest Indian Ridge, Nature, 315, 196–199.
Hamelin, B., B. Dupr
e, and C.-J. Allègre (1986), Pb-Sr-Nd isotopic data of Indian Ocean ridges: New evidence of large-scale mapping of mantle heterogeneities, Earth Planet. Sci. Lett., 76, 288–298.
Hammer, O., D. A. T. Harper, and P. D. Ryan (2001), PAST: Paleontological statistics software package for education and data analysis, Palaeontol. Electron., 4, 9.
Hanan, B., J. Blichert-Toft, D. G. Pyle, and D. M. Christie (2004), Contrasting origins of the upper mantle revealed by hafnium and lead isotopes from the Southeast Indian Ridge, Nature, 432, 91–94.
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2072
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Hanan, B. B., and D. W. Graham (1996), Lead and helium isotope evidence from oceanic basalts for a common deep source of mantle
plumes, Science, 272, 991–995.
Hanan, B. B., and J.-G. Schilling (1989), Easter microplate evolution: Pb isotope evidence, J. Geophys. Res., 94, 7432–7448.
Hanan, B. B., J. Blichert-Toft, C. Hemond, K. Sayit, A. Agranier, D. W. Graham, and F. Albarede (2013), Pb and Hf isotope variations along the
Southeast Indian Ridge and the dynamic distribution of MORB source domains in the upper mantle, Earth Planet. Sci. Lett., 375, 196–
208, doi:10.1016/j.epsl.2013.05.028.
Hanyu, T., and I. Kaneoka (1998), Open system behavior of helium in case of the HIMU source area, Geophys. Res. Lett., 25, 687–690.
Hanyu, T., I. Kaneoka, and K. Nagao (1999), Noble gas study of HIMU and EM ocean island basalts in the Polynesian region, Geochim. Cosmochim. Acta, 63, 1181–1201.
Hart, S. R. (1984), A large-scale isotope anomaly in the Southern Hemisphere mantle, Nature, 309, 753–757.
Hart, S. R. (1993), Equilibration during mantle melting: A fractal tree model, Proc. Natl. Acad. Sci. U. S. A., 90, 11,914–11,918.
Hart, S. R., M. Kurz, and Z. Wang (2008), Scale length of mantle heterogeneities: Constraints from helium diffusion, Earth Planet. Sci. Lett.,
269, 507–516.
Harvey, J., A. Gannoun, K. W. Burton, N. W. Rogers, O. Alard, and I. J. Parkinson (2006), Ancient melt extraction from the oceanic upper mantle revealed by Re-Os isotopes in abyssal peridtites from the Mid-Atlantic Ridge, Earth Planet. Sci. Lett., 244, 606–621.
Hilton, D. R., J. Barling, and G. E. Wheller (1995), Effect of shallow-level contamination on the helium isotope systematics of ocean-island
lavas, Nature, 373, 330–333.
Hirschmann, M. M., and E. M. Stolper (1996), A possible role for garnet pyroxenite in the origin of the ‘‘garnet signature’’ in MORB, Contrib.
Mineral. Petrol., 124, 185–208.
Hofmann, A. W. (1997), Mantle geochemistry: The message from oceanic volcanism, Nature, 385, 219–229.
Iwamori, H., F. Albarède, and H. Nakamura (2010), Global structure of mantle isotopic heterogeneity and its implications for mantle differentiation and convection, Earth Planet. Sci. Lett., 299, 339–351, doi:10.1016/j.epsl.2010.09.014.
Janin, M., C. Hemond, M. Maia, P. Nonnotte, E. Ponzevera, and K. T. M. Johnson (2012), The Amsterdam-St. Paul Plateau: A complex hot
spot/DUPAL-flavored MORB interaction, Geochem. Geophys. Geosyst., 13, Q09016, doi:10.1029/2012GC004165.
Janney, P. E., A. P. le Roex, and R. W. Carlson (2005), Hafnium isotope and trace element constraints on the nature of mantle heterogeneity
beneath the central Southwest Indian Ridge (13 E to 47 E), J. Petrol., 46, 2427–2464.
Johnson, K. T. M., D. W. Graham, K. H. Rubin, K. Nicolaysen, D. Scheirer, D. W. Forsyth, E. T. Baker, and L. M. Douglas-Priebe (2000), Boomerang Seamount: The active expression of the Amsterdam-St. Paul hotspot, Southeast Indian Ridge, Earth Planet. Sci. Lett., 183, 245–259.
Kamenetsky, V. S., R. Maas, N. M. Sushchevskaya, M. D. Norman, I. Cartwright, and A. A. Peyve (2001), Remnants of Gondwanan continental
lithosphere in oceanic upper mantle: Evidence from the South Atlantic Ridge, Geology, 29, 243–246.
Katz, R. F., and J. F. Rudge (2011), The energetics of melting fertile heterogeneities within the depleted mantle, Geochem. Geophys. Geosyst.,
12, Q0AC16, doi:10.1029/2011GC003834.
Katz, R. F., and S. M. Weatherley (2012), Consequences of mantle heterogeneity for melt extraction at mid-ocean ridges, Earth Planet. Sci.
Lett., 335, 226–237, doi:10.1016/j.epsl.2012.04.042.
Kellogg, J. B., S. B. Jacobsen, and R. J. O’Connell (2002), Modeling the distribution of isotopic ratios in geochemical reservoirs, Earth Planet.
Sci. Lett., 284, 183–202.
Kellogg, J. B., S. B. Jacobsen, and R. J. O’Connell (2007), Modeling lead isotopic heterogeneity in mid-ocean ridge basalts, Earth Planet. Sci.
Lett., 262, 328–342.
Kingsley, R. H., J. Blichert-Toft, D. Fontignie, and J.-G. Schilling (2007), Hafnium, neodymium, and strontium isotope and parent-daughter
element systematics in basalts from the plume-ridge interaction system of the Salas y Gomez Seamount Chain and Easter Microplate,
Geochem. Geophys. Geosyst., 8, Q04005, doi:10.1029/2006GC001401.
Klein, E. M., and C. H. Langmuir (1987), Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness, J. Geophys. Res., 92, 8089–8115.
Kogiso, T., K. Hirose, and E. Takahashi (1998), Melting experiments on homogeneous mixtures of peridotite and basalt: Application to the
genesis of ocean island basalts, Earth Planet. Sci. Lett., 162, 45–61.
Kogiso, T., M. M. Hirschmann, and D. J. Frost (2003), High-pressure partial melting of garnet pyroxenite: Possible mafic lithologies in the
source of ocean island basalts, Earth Planet. Sci. Lett., 216, 603–617.
Kurz, M. D., M. Moreira, J. Curtice, D. E. Lott, J. J. Mahoney, and J. M. Sinton (2005), Correlated helium, neon, and melt production on the
super-fast spreading East Pacific Rise near 17 S, Earth Planet. Sci. Lett., 232, 125–142.
Lambart, S., D. Laporte, and P. Schiano (2013), Markers of the pyroxenite contribution in the major-element compositions of oceanic
basalts: Review of the experimental constraints, Lithos, 160–161, 14–36, doi:10.1016/j.lithos.2012.11.018.
Langmuir, C. H., and G. N. Hanson (1980), An evaluation of major element heterogeneity in the mantle sources of basalts, Philos. Trans. R.
Soc. London A, 297, 383–407.
Lees, J. M., and J. Park (1995), Multiple-taper spectral analysis: A stand-alone C-subroutine, Comput. Geosci., 21, 199–236.
Liu, C.-Z., J. E. Snow, E. Hellebrand, G. Br€
ugmann, A. von der Handt, A. B€
uchl, and A. W. Hofmann (2008), Ancient, highly heterogeneous
mantle beneath Gakkel ridge, Arctic Ocean, Nature, 452, 311–316.
Lomb, N. R. (1976), Least-squares frequency analysis of unequally spaced data, Astrophys. Space Sci., 39, 447–462.
Lupton, J., and L. Evans (2004), The atmospheric helium isotope ratio: Is it changing?, Geophys. Res. Lett., 31, L13101, doi:10.1029/
2004GL020041.
Maclennan, J., D. McKenzie, F. Hilton, K. Gronv€
old, and N. Shimizu (2003), Geochemical variability in a single flow from northern Iceland, J.
Geophys. Res., 108(B1), 2007, doi:10.1029/2000JB000142.
Mahoney, J., A. P. Le Roex, Z. Peng, R. L. Fisher, and J. H. Natland (1992), Southwestern limits of Indian Ocean Ridge mantle and the origin
of low 206Pb/204Pb mid-ocean ridge basalt: Isotope systematics of the Central Southwest Indian Ridge (17 –50 E), J. Geophys. Res., 97,
19,771–19,790.
Mahoney, J. J., J. H. Natland, W. M. White, R. Poreda, S. H. Bloomer, R. L. Fisher, and A. N. Baxter (1989), Isotopic and geochemical provinces
of the western Indian Ocean spreading centers, J. Geophys. Res., 94, 4033–4052.
Mahoney, J. J., J. M. Sinton, M. D. Kurz, J. D. Macdougall, K. J. Spencer, and G. W. Lugmair (1994), Isotope and trace element characteristics
of a super-fast spreading ridge: East Pacific Rise, 13–23 S, Earth Planet. Sci. Lett., 121, 173–193.
Mahoney, J. J., D. W. Graham, D. M. Christie, K. T. M. Johnson, L. S. Hall, and D. L. VonderHaar (2002), Between a hot spot and cold spot: Isotopic variation in the Southeast Indian Ridge asthenosphere, 86 –118 E, J. Petrol., 43, 1155–1176.
Manga, M. (1996), Mixing of heterogeneities in the mantle: Effect of viscosity differences, Geophys. Res. Lett., 23, 403–406.
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2073
Geochemistry, Geophysics, Geosystems
10.1002/2014GC005264
Matsuda, J., T. Matsumoto, H. Sumino, K. Nagao, J. Yamamoto, Y. Miura, I. Kaneoka, N. Takahata, and Y. Sano (2002), The 3He/4He ratio of
the new internal He Standard of Japan (HESJ), Geochem. J., 36, 191–195.
Meyzen, C. M., J. Blichert-Toft, J. N. Ludden, E. Humler, C. M
evel, and F. Albarède (2007), Isotopic portrayal of Earth’s upper mantle flow
field, Nature, 447, 1069–1074.
Moreira, M., J. Escartin, E. Gayer, C. Hamelin, A. Bezos, F. Guillon, and M. Cannat (2011), Rare gas systematics on Lucky Strike basalts (37 N,
North Atlantic): Evidence for efficient homogenization in a long-lived magma chamber system?, Geophys. Res. Lett., 38, L08304, doi:
10.1029/2011GL046794.
Nicolaysen, K. P., F. A. Frey, K. T. M. Johnson, J. J. Mahoney, and D. W. Graham (2007), Influence of the Amsterdam/St. Paul hotspot along
the Southeast Indian Ridge between 77 and 88 E: Correlations of Sr, Nd, Pb and He isotopic variations with ridge segmentation, Geochem. Geophys. Geosyst., 8, Q09007, doi:10.1029/2006GC001540.
Olson, P., D. A. Yuen, and D. Balsiger (1984), Convective mixing and the fine structure of mantle heterogeneity, Phys. Earth Planet. Inter., 36,
591–603.
Parai, R., S. Mukhopadhyay, and J. C. Lassiter (2009), New constraints on the HIMU mantle from neon and helium isotopic compositions of
basalts from the Cook-Austral islands, Earth Planet Sci. Lett., 277, 253–261, doi:10.1016/j.epsl.2008.10.014.
Polv
e, M., and C. J. Allègre (1980), Orogenic lherzolite complexes studied by 87Rb-87Sr: A clue to understand the mantle convection processes, Earth Planet. Sci. Lett., 51, 71–93.
Poreda, R. J., J.-G. Schilling, and H. Craig (1993), Helium isotope ratios in Easter Microplate basalts, Earth Planet. Sci. Lett., 119, 319–329.
Richter, F. M., and N. M. Ribe (1979), On the importance of advection in determining the local isotopic composition of the mantle, Earth
Planet. Sci. Lett., 43, 212–222.
Rubin, K. H., J. M. Sinton, J. Maclennan, and E. Hellebrand (2009), Magmatic filtering of mantle compositions at mid-ocean ridge volcanoes,
Nat. Geosci., 2, 321–328.
Rudge, J. F., J. Maclennan, and A. Stracke (2013), The geochemical consequences of mixing melts from a heterogeneous mantle, Geochim.
Cosmochim. Acta, 114, 112–143, doi:10.1016/j.gca.2013.03.042.
Russo, C. J., K. H. Rubin, and D. W. Graham (2009), Mantle melting and magma supply to the Southeast Indian Ridge: The roles of lithology
and melting conditions from U-series disequilibria, Earth Planet. Sci. Lett., 278, 55–66.
Salters, V. J. M., S. Mallick, S. R. Hart, C. H. Langmuir, and A. Stracke (2011), Domains of depleted mantle: New evidence from hafnium and
neodymium isotopes, Geochem. Geophys. Geosyst., 12, Q08001, doi:10.1029/2011GC003617.
Scargle, J. D. (1982), Studies in astronomical time series analysis. II. Statistical aspects of spectral analysis of unevenly spaced data, Astrophys. J., 263, 835–853.
Schulz, M., and M. Mudelsee (2002), REDFIT: Estimating red-noise spectra directly from unevenly spaced paleoclimatic time series, Comput.
Geosci., 28, 421–426.
Semp
er
e, J.-C., J. R. Cochran, and S. S. Team (1997), The Southeast Indian Ridge between 88 E and 118 E: Variations in crustal accretion at
constant spreading rate, J. Geophys. Res., 102, 15,489–15,505.
Shen, Y., and D. W. Forsyth (1995), Geochemical constraints on initial and final depths of melting beneath mid-ocean ridges, J. Geophys.
Res., 100, 2211–2237.
Shorttle, O., and J. Maclennan (2011), Compositional trends of Icelandic basalts: Implications for short–length scale lithological heterogeneity in mantle plumes, Geochem. Geophys. Geosyst., 12, Q11008, doi:10.1029/2011GC003748.
Small, C., J. R. Cochran, J.-C. Semp
er
e, and D. M. Christie (1999), The structure and segmentation of the Southeast Indian Ridge, Mar. Geol.,
161, 1–12.
Stracke, A., and B. Bourdon (2009), The importance of melt extraction for tracing mantle heterogeneity, Geochim. Cosmochim. Acta, 73,
218–238, doi:10.1016/j.gca.2008.10.015.
Sun, S.-S., and W. F. McDonough (1989), Chemical and isotopic systematics of oceanic basalts: Implications for mantle composition and
processes, in Magmatism in the Ocean Basins, edited by A. D. Saunders and M. J. Norry, pp. 313–345, Geol. Soc. Spec. Publ., London.
Tatsumoto, M. (1978), Isotopic composition of lead in oceanic basalts and its implication to mantle evolution, Earth Planet. Sci. Lett., 38, 63–
87.
Thomson, D. J. (1982), Spectrum estimation and harmonic analysis, Proc. IEEE, 70, 1055–1096.
Vlast
elic, I., H. Bougault, and L. Dosso (2002), Heterogeneous heat production in the Earth’s upper mantle: Blob melting and MORB composition, Earth Planet. Sci. Lett., 199, 157–172.
White, W. M. (1985), Sources of oceanic basalts: Radiogenic isotopic evidence, Geology, 13, 115–118.
Zindler, A., and S. R. Hart (1986), Chemical geodynamics, Annu. Rev. Earth Planet. Sci., 14, 493–571.
Zindler, A., E. Jagoutz, and S. L. Goldstein (1982), Nd, Sr and Pb isotope systematics in a three-component mantle: A new perspective,
Nature, 298, 519–523.
GRAHAM ET AL.
C 2014. American Geophysical Union. All Rights Reserved.
V
2074
Auxiliary Material for
Helium Isotopic Textures in Earth’s Upper Mantle
David W. Graham a, Barry B. Hanan b, Christophe Hémond c,
Janne Blichert-Toft d and Francis Albarède d
a
College of Earth, Ocean, & Atmospheric Sciences, Oregon State University, Corvallis, OR
97331, United States
b
Department of Geological Sciences, San Diego State University, San Diego, CA 92182-1020,
United States
c
Domaines Océaniques, Institut Universitaire Européen de la Mer, place Nicolas Copernic,
F-29280 Plouzané, France
d
Laboratoire de Géologie de Lyon, Ecole Normale Supérieure de Lyon and Université Claude
Bernard Lyon 1, CNRS UMR 5276, 46 Allée d’Italie, 69007 Lyon, France
Geochemistry, Geophysics, Geosystems, 2014
The Supporting Information contains:
1. text01. Background information on spectral analysis approaches used in the study
(Periodogram, Redfit, Multi-Taper Method, and Continuous Wavelet Transform).
2. fs01. Supplementary Fig. 1. Along-axis variations in 3He/4He, 206Pb/204Pb, 208Pb*/206Pb*
and εHf for basalt glasses from the Southeast Indian Ridge. Isotope data and sources are
given in Supplementary Table 1. 208Pb*/206Pb* is a proxy for the time-integrated Th/U
ratio of the mantle source; 208Pb*/206Pb* = (208Pb/204Pb-29.476)/(206Pb/204Pb-9.307). εHf =
(176Hf/177Hf/0.282772–1)x104. The Continuous Wavelet Transform for each isotopic
parameter is also shown for comparison.
3. fs02. Supplementary Fig. 2. Wavelet transform coherence (WTC) for 3He/4He 206
Pb/204Pb. Coherence is contoured in color according to the scale shown at the right.
Arrows depict the phase relationship between 3He/4He and 206Pb/204Pb as a function of
length scale (vertical axis) and location along the SEIR (horizontal axis). Right-pointing
arrow indicates an in-phase relationship and left-pointing arrow an anti-phase
relationship. Thick contours outline 5% significance levels against red noise.
4. fs03. Supplementary Fig. 3. WTC for 3He/4He - 208Pb*/206Pb*.
5. fs04. Supplementary Fig. 4. WTC for 3He/4He - εHf.
6. ts01. SEIR_Data.xls. Excel table of sample locations with He, Pb and Hf isotope data for
Southeast Indian Ridge basalts.
Background on Spectral Analysis Approaches
Different analytical approaches such as the correlogram [Graham et al., 2001] and
periodogram [Agranier et al., 2005; Meyzen et al., 2007] have been used previously to quantify
the length scales of isotope variability/convection in the upper mantle. To our knowledge, our
study is the first to apply red-noise and multi-taper spectral estimations and the continuous
wavelet transform to densely sampled MORB compositions. The length scales deduced from
these different approaches reflect different model assumptions against which the data series is
tested. We envision that analogous spectral approaches could be applied to the length scales of
heterogeneity observed in dynamical models of mantle convection using particle tracers
[Brandenburg et al., 2008]. We have not performed such convection modeling here as it is
beyond the scope of our study. We note, however, that adopting spatial analytical approaches is
logical and analogous approaches have been used previously in modeling of mantle
heterogeneity (e.g., Olson et al., [1984]; Todesco and Spera, [1992]). Applying these tools to
state-of-the-art convection models may allow closer connections to be made to observed isotopic
variations, and provide stringent constraints on key governing parameters that affect the folding
and stretching of material lines during mantle flow. Such parameters include the depth
dependence of viscosity and the maximum Lyapunov exponent defining the rate of material
stretching [Hunt and Kellogg, 2001; Ferrachat and Ricard, 2001; Farnetani and Samuel, 2003;
Schneider et al., 2007].
Periodogram
The periodogram is one way to estimate spectral density of a signal embedded in a data
series. The periodogram performs well with a spectrum that is a continuum, and when the
background is approximately white noise [Weedon, 2003]. The Lomb-Scargle periodogram is
commonly employed for a data series having an uneven sample spacing [Press et al., 1992]. The
data are typically detrended prior to analysis. The sum of the squared sine and cosine average
amplitudes yields the numerical value of the power at the frequency (or wave number) of
interest. A plot of power vs. frequency yields the periodogram. A compromise between
periodogram smoothness and spectral resolution is inherent in all spatially discrete data;
sinusoids of low frequency (or long period) are smoother in appearance than those of high
frequency (or short period). For a purely random data series (white noise), all of the sinusoids
should be of equal importance and thus the periodogram will vary randomly around a constant
value. If a data series has a strong sinusoidal signal for a particular frequency, there will be a
peak in the periodogram at that frequency.
Red-Noise Spectral Estimation (REDFIT)
A continuous and relatively flat spectrum, or “white noise”, describes a data series made
up of random numbers having oscillations of approximately equal amplitude over all length
scales. It is rare for natural systems to produce purely white noise data series. When the noise
contains an additional component that can be weighted (positively or negatively) from the
previous value(s) in the data series, it is referred to as an auto-regressive process [Weedon,
2003]. This type of system therefore has “memory” from nearby events. The second approach
we adopted for our analysis of the SEIR 3He/4He data therefore tests for this effect, following the
analytical methods of Schulz and Mudelsee [2002]. Their spectral analysis approach is designed
to test the variance of a data series against the null hypothesis that a first-order auto-regressive
(AR1) process is responsible for a continuous decrease of spectral amplitude (power) with
increasing frequency. An AR1 red-noise model is often a more appropriate null hypothesis than
the white noise model for the simple Lomb-Scargle periodogram described in the previous
section (e.g., “red noise” is commonly observed in paleoclimatic time series; Schulz and
Mudelsee, [2002]). AR1 parameters are estimated directly from the unevenly spaced data series
without interpolation, and the AR1 model is then transformed from the time (or in this case,
spatial) domain to the frequency domain. A characteristic scale (τ) is determined that represents
a measure of the memory effect in the AR1 process [Robinson, 1977]. The spectrum for the
irregularly spaced data series, estimated using the Lomb-Scargle periodogram, is then compared
to the AR1 model. A Monte Carlo approach is also used to determine a correction factor for the
periodogram; this allows adjustment of an inherent bias in the periodogram when there are
unevenly spaced data, because higher-frequency spectral amplitudes are overestimated due to a
non-independent distribution of Fourier components in the periodogram [Lomb, 1976; Scargle,
1982]. In addition, we adopted Welch-overlapped-segment-averaging (WOSA) to reduce
spectral leakage [Welch, 1967], the phenomenon whereby variance at some important frequency
may “leak” into other frequencies in the estimated spectrum due to the data windowing. WOSA
also helps to reduce noise but it reduces spectral resolution. The WOSA procedure splits the data
series into n50 segments (overlapping by 50%), and the final spectral estimate is then derived
from averaging the n50 periodograms. The statistical significance of a REDFIT spectral peak is
estimated using the upper confidence interval of the AR1 noise. A false-alarm level ([11/n]x100%, where n=number of data points within each WOSA segment) allows testing for
statistically significant peaks in the data series that arise from non-AR1 components [Schulz and
Mudelsee, 2002].
Multi-Taper Method (MTM) of Spectral Estimation
We adopted the MTM routine of Lees and Park [1995] to the He isotope results. MTM
employs a small number of window trials, in which the data are multiplied by mutually
orthogonal taper functions to compute independent estimates of the power spectrum, which are
then averaged. This minimizes spectral leakage that typically results from data windowing, and it
reduces spurious correlations that can emerge from using single data tapers. MTM also allows
confidence levels in spectral peaks to be computed directly through the computation of F-test
values. For a stationary data series MTM provides a stable spectrum estimation [Thomson 1982].
The number of tapers used is typically between 3 and 5, the exact choice of which represents a
tradeoff between spectral resolution and stability (variance). Increasing the number of tapers
weights the ends of a data series more heavily in the resulting spectrum estimation.
Continuous Wavelet Transform
The Continuous Wavelet Transform (CWT) allows data series signals to be “tiled” for
visualization. The tiling is performed over multiple scales ranging from coarse to medium to
fine, where at each scale the signal is resolved by averages [Torrence and Compo, 1998]. The
approach adopted here uses the Morlet “mother” wavelet; the chosen wavelet is multiplied with
the data series to produce the wavelet transform [Weedon, 2003]. Our approach employs the
algorithm developed by Grinsted et al. [2004] to estimate the CWT for He, Pb and Hf isotope
variations along the SEIR. Data were interpolated at 10 km intervals along the SEIR using a
piecewise cubic hermite spline. To be strictly valid, the histogram of the data series should be
roughly Gaussian. This is generally supported for 3He/4He by the results shown in Fig. 3,
although we noted earlier that it is possible that two sub-populations are present.
The CWT tiling procedure varies the size of the mother wavelet along the entire length of
the data series in a systematic way, filtering from highest frequency to lowest frequency
components in the data series [Weedon, 2003]. This allows different lengths of the data series to
be examined; consequently it leads to variable frequency resolution across the original data
series, unlike periodogram-based approaches. Long wavelengths cover larger proportions of the
original data series [Weedon, 2003]. This therefore leads to good frequency resolution for long
wavelengths, but comparatively poorer spatial resolution. In contrast, at lower frequencies the
frequency resolution is not as good, but the spatial resolution is better because fewer data are
needed in computing the shorter wavelength wavelets. In the CWT diagrams (Fig. S1), the
horizontal axis represents the original spatial (or time) domain, while the vertical axis is
portrayed on a natural logarithm scale. One unit on the vertical axis therefore corresponds to a
doubling in period size. Across the top of the diagram, therefore, a finer scale view is offered,
while across the bottom of the diagram the overview represents longer trends in the original data
series. The signal “power” (the squared correlation strength with the scaled mother wavelet) is
displayed in color to emphasize “strong” vs. “weak” aspects in the data. There is also an inherent
“cone of influence” (COI), where boundary effects become significant due to the finite length of
the data series [Torrence and Compo, 1998]. Outside of this COI the wavelet power is
considerably more uncertain. The significance level of the wavelet power is typically assessed
against the AR(1) red noise assumption [Grinsted et al., 2004].
Following the methods of Grinsted et al. [2004] we also determined the Cross Wavelet
Transform (XWT) and Wavelet Transform Coherence (WTC) between 3He/4He and Pb and Hf
isotopes (Supplementary Figs. 2 - 4), and between 3He/4He and axial depth (Fig. 12). The
computed WTC may be thought of as the localized correlation coefficient [Grinsted et al., 2004].
References
Agranier, A., J. Blichert-Toft, D. W. Graham, V. Debaille, P. Schiano, and F. Albarede (2005),
The spectra of isotopic heterogeneities along the Mid-Atlantic Ridge, Earth Planet. Sci.
Lett., 238, 96-109.
Brandenburg, J. P., E. H. Hauri, P. E. van Keken, and C. J. Ballentine (2008), A multiple-system
study of the geochemical evolution of the mantle with force-balanced plates and
thermochemical effects, Earth Planet. Sci. Lett., 276, 1-13, doi:10.1016/j.epsl.2008.08.027.
Farnetani, C. G., and H. Samuel (2003), Lagrangian structures and stirring in the Earth’s mantle,
Earth Planet Sci. Lett., 206, 335-348, doi:10.1016/S0012-821X(02)01085-3.
Ferrachat, S., and Y. Ricard (2001), Mixing properties in the Earth's mantle: effects of viscosity
stratification and of oceanic crust segregation, Geochem. Geophys. Geosyst., 2,
2000GC000092.
Graham, D. W., J. E. Lupton, F. J. Spera, and D. M. Christie (2001), Upper mantle dynamics
revealed by helium isotope variations along the Southeast Indian Ridge, Nature, 409, 701703.
Grinsted, A., J. C. Moore, and S. Jevrejeva (2004), Application of the cross wavelet transform
and wavelet coherence to geophysical time series, Nonlinear Proc. Geophysics, 11, 561566.
Hunt, D. L., and L. H. Kellogg (2001), Quantifying mixing and age variations of heterogeneities
in models of mantle convection: role of depth-dependent viscosity, J. Geophys. Res., 106,
6747-6759.
Lees, J. M., and J. Park (1995), Multiple-taper spectral analysis: a stand-alone C-subroutine,
Computers & Geosciences, 21, 199-236.
Lomb, N. R. (1976), Least-squares frequency analysis of unequally spaced data, Astrophys.
Space Sci., 39, 447-462.
Meyzen, C., M., J. Blichert-Toft, J. N. Ludden, E. Humler, C. Mével, and F. Albarède (2007),
Isotopic portrayal of Earth's upper mantle flow field, Nature, 447, 1069-1074.
Olson, P., D. A. Yuen, and D. Balsiger (1984), Convective mixing and the fine structure of
mantle heterogeneity, Phys. Earth Planet. Inter., 36, 591-603.
Press, W. H., S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery (1992), Numerical Recipes:
The Art of Scientific Computing, second edition ed., 933 pp., Cambridge Univ. Press,
Cambridge, UK.
Robinson, P. M. (1977), Estimation of a time series model from unequally spaced data,
Stochastic Processes and their Applications, 6, 9-24.
Scargle, J. D. (1982), Studies in astronomical time series analysis. II. statistical aspects of
spectral analysis of unevenly spaced data, Astrophys. J., 263, 835-853.
Schneider, J., J. Schmalzl, and T. Tel (2007), Lagrangian avenues of transport in the Earth's
mantle, Chaos, 17, 033115.
Schulz, M., and M. Mudelsee (2002), REDFIT: estimating red-noise spectra directly from
unevenly spaced paleoclimatic time series, Computers & Geosciences, 28, 421-426.
Thomson, D. J. (1982), Spectrum estimation and harmonic analysis, Proc. IEEE, 70, 1055-1096.
Todesco, M., and F. J. Spera (1992), Stability of a chemically layered upper mantle, Phys. Earth
Planet. Inter., 71, 85-99.
Torrence, C., and G. P. Compo (1998), A practical guide to wavelet analysis, Bull. Amer.
Meteorol. Soc., 79, 61-78.
Weedon, G. P. (2003), Time-Series Analysis and Cyclostratigraphy, 259 pp., Cambridge Univ.
Press.
Welch, P. D. (1967), The use of fast fourier transform for the estimation of power spectra: a
method based on tme averaging over short, modified periodograms, ISEE Trans. Audio
Electroacoust., AU-15, 70-73.
Length Scale (km)
Length Scale (km)
32
64
128
256
512
512
208Pb*/206Pb*
3 He/4 He
256
1.00
0.95
Length Scale (km)
8
64
128
256
512
32
64
128
256
512
1024
1024
19
15
Hf
206Pb/204Pb
128
1.05
12
Length Scale (km)
(R/RA)
14
32
64
1024
1024
10
32
18
10
5
-1000
0
1000
2000
Along-Axis Distance (km)
3000
-1000
0
1000
2000
Along-Axis Distance (km)
3000
Sample
03/01 D 1-22
05/02 D 1-3
05/03 D 1-6
07/04 D 1-2
BMRG06 76-2
BMRG06 76-5
BMRG06 77-2
BMRG06 77-3
22/07 D 2-3
22/07 D1-1
BMRG06 79-1
JC84 D4-3
BMRG06 75-1; -4
BMRG06
BMRG06
BMRG06
BMRG06
BMRG06
BMRG06
BMRG06
BMRG06
BMRG06
WC50
WC49
WC48
WC47 h
73-3
73-4
73-5
73-6
WC46
BMRG06 71-1
BMRG06 71-3
BMRG06 WC43
BMRG06 70-1
BMRG06 WC44
BMRG06 WC45
BMRG06 67-1
BMRG06 66-1
BMRG06 WC38
BMRG06 64-2
BMRG06 64-8
BMRG06 WC37
BMRG06 63-1
BMRG06 WC34
BMRG06 WC35
BMRG06 60-1
BMRG06 61-1
18/06 D3-3
18/06 D 2-3
BMRG06 58-1
BMRG06 58-4
BMRG06 59-1
BMRG06 55-1
Segment
Location Longitude Latitude Depth (m) Distance from StPaul 3He/4He 2-sigma 206Pb/204Pb 207Pb/204Pb 208Pb/204Pb
B
axis
72.240 -26.910
3625
-1291.1
8.26
0.05
17.666
15.437
37.670
D
axis
75.300 -29.700
3000
-860.2
8.91
0.07
17.817
15.447
37.750
D
axis
75.180 -29.810
4100
-859.8
8.77
0.07
17.635
15.457
37.493
F
axis
77.620 -32.670
2700
-471.7
8.58
0.06
18.000
15.489
38.040
F
axis
77.886 -32.764
2178
-446.6
8.49
0.04
F
axis
77.886 -32.764
2178
-446.6
8.50
0.04
17.998
15.449
37.909
F
axis
77.816 -32.945
2835
-437.3
8.55
0.04
F
axis
77.816 -32.945
2835
-437.3
8.57
0.05
17.877
15.445
37.788
G
axis
78.010 -34.350
3100
-315.6
8.53
0.07
18.155
15.466
38.103
G
axis
78.020 -34.380
3080
-312.6
8.51
0.05
18.400
15.524
38.490
G
axis
78.236 -34.570
3163
-283.7
8.60
0.05
18.27
15.495
38.287
G
axis
78.670 -34.910
3700
-228.9
8.25
0.04
Zeewolf FZ relay zone
78.595 -35.282
3269
-205.0
8.99
0.05
17.915
15.441
37.853
H
H
H
H
H
H
H
H
H
axis
axis
axis
axis
axis
axis
axis
axis
axis
78.354
78.464
78.595
78.709
78.825
78.825
78.825
78.825
78.953
-35.528
-35.641
-35.797
-35.939
-36.066
-36.066
-36.066
-36.066
-36.182
3413
3086
2900
2840
2659
2659
2659
2659
2734
-201.5
-185.6
-165.0
-146.7
-129.3
-129.3
-129.3
-129.3
-112.1
12.86
11.85
11.94
14.13
13.53
10.35
10.43
10.69
12.63
0.07
0.06
0.07
0.07
0.07
0.06
0.05
0.06
0.07
I1
I1
I1
I1
BMRG Smt
BMRG Smt
I2
I2
I2
I2
I2
I2
I2
St. Pierre
I2
J1
J1
J1
J1
J1/J2
J1/J2
J1
J1
axis
axis
axis
axis
off-axis
off-axis
axis
78.240
78.240
78.325
78.400
77.825
77.832
78.091
77.992
78.188
78.163
78.163
78.253
78.367
77.964
78.588
78.107
78.130
78.160
78.140
77.882
77.882
78.144
78.180
-37.056
-37.056
-37.121
-37.166
-37.721
-37.716
-37.578
-37.803
-37.860
-37.983
-37.983
-38.090
-38.199
-38.781
-38.433
-38.959
-38.950
-38.960
-38.980
-39.214
-39.214
-39.124
-39.290
1940
1940
2092
1850
875
679
1865
1935
1709
1805
1805
1780
1838
1678
2163
1978
2153
2020
2320
1370
1370
1903
2104
-88.9
-88.9
-78.4
-70.2
-61.8
-61.8
-56.8
-45.1
-28.4
-20.2
-20.2
-6.3
9.4
31.3
41.4
54.1
54.8
57.4
57.8
61.1
61.1
69.5
85.0
9.56
9.57
9.47
9.63
9.28
10.12
9.86
9.67
11.34
11.11
11.72
12.21
11.57
13.41
10.27
10.44
10.02
11.07
10.78
11.71
11.93
11.89
10.56
0.06
0.05
0.05
0.06
0.32
0.06
0.13
0.07
0.06
0.06
0.08
0.08
0.06
0.08
0.06
0.05
0.10
0.06
0.08
0.09
0.06
0.08
0.08
axis
axis
axis
off-axis
axis
axis
axis
axis
off-axis
off-axis
axis
18.722
19.407
18.957
18.111
15.562
15.646
15.594
15.494
38.923
39.664
39.196
38.237
18.552
19.204
15.551
15.622
38.798
39.453
18.245
18.344
18.302
19.139
18.554
18.757
19.097
19.143
19.075
18.751
18.917
18.523
17.858
18.507
18.019
18.122
17.777
18.051
19.13
15.521
15.553
15.54
15.618
15.586
15.578
15.590
15.593
15.599
15.568
15.556
15.561
15.516
15.554
15.541
15.572
15.549
15.564
15.594
38.509
38.639
38.621
39.472
39.036
39.131
39.411
39.429
39.441
39.113
39.223
38.940
38.287
38.913
38.603
38.684
38.477
38.680
39.441
18.189
17.843
15.557
15.521
38.860
38.445
176Hf/177Hf eps Hf
0.283211 15.52
0.283159 13.69
0.283226 16.06
0.283166 13.93
0.283205
15.31
0.283177
14.32
0.283047
9.73
0.283086
11.10
0.282935
5.76
0.282874
0.282967
3.61
6.90
0.282927
5.48
Sample
Segment
Location
BMRG06 54-3
BMRG06 48-2
13/05 D1-5
BMRG06 47-6; -1
BMRG06 51-1
BMRG06 46-1; -2
BMRG06 44-6
BMRG06 43-2
BMRG06 41-2
J2
J2
J3
J3
J2
J3
J4
J4
J4
axis
axis
axis
axis
GSR2 WC A1-1
GSR2 WC A1-2
BMRG06 39-1
GSR2 WC A1-3
GSR2 WC A1-5
BMRG06 WC09
GSR2 WC A1-5/6
GSR2 WC A1-6
GSR2 WC A1-7
GSR2 WC A1-8
GSR2 WC A1-9c
GSR2 WC A1-10
GSR2 WC A1-11
GSR2 WC A1-12b
GSR2 WC A1-13b
GSR2 WC A1-14
GSR2 WC A1-15
GSR2 WC A1-16
BMRG06 WC07
GSR2 WC A1-16/17
GSR2 WC A1-17
GSR2 WC A1-18
GSR2 WC A1-19
GSR2 WC A1-20
GSR2 WC A1-21
GSR2 WC A1-22
GSR2 WC A1-23
BMRG06 WC04
GSR2 WC A2-20
GSR2 WC A2-19
BMRG06 36-1
GSR2 WC A2-18
GSR2 WC A2-17
GSR2 WC A2-16
GSR2 WC A2-15
Longitude
Latitude
Depth (m)
Distance from StPaul 3He/4He 2-sigma
axis
axis
axis
axis
78.090
78.390
77.890
78.142
78.630
78.324
78.788
79.105
79.753
-39.450
-39.810
-40.360
-40.553
-40.213
-40.748
-40.970
-41.252
-41.767
2225
2479
3400
2845
3264
3014
2800
2785
2840
92.4
139.1
154.0
184.2
185.4
210.5
255.2
296.2
374.9
9.36
8.61
8.31
8.03
8.93
8.41
7.77
7.98
7.97
0.05
0.05
0.05
0.04
0.06
0.04
0.05
0.04
0.04
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
K
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
80.975
81.069
81.153
81.238
81.373
81.472
81.471
81.583
81.662
81.777
81.910
81.982
82.060
82.152
82.221
82.317
82.397
82.485
82.552
82.563
82.623
82.713
82.827
82.932
83.037
83.123
83.203
83.290
-41.105
-41.170
-41.242
-41.299
-41.463
-41.533
-41.538
-41.587
-41.660
-41.745
-41.801
-41.883
-41.963
-42.050
-42.114
-42.172
-42.242
-42.304
-42.352
-42.366
-42.427
-42.492
-42.538
-42.593
-42.660
-42.729
-42.790
-42.863
2600
2612
2776
2688
2563
2579
2520
2537
2524
2558
2690
2763
2716
2510
2473
2620
2769
2735
2645
2705
2577
2540
2575
2616
2635
2680
2812
2953
395.3
405.9
416.5
426.1
446.8
458.1
458.5
468.9
479.3
492.7
505.0
515.6
526.4
538.6
547.6
557.7
567.8
577.9
585.5
587.3
595.5
605.7
616.1
626.6
637.9
648.3
657.7
668.4
7.63
7.85
7.59
7.67
7.63
7.73
7.83
7.60
7.67
7.63
7.64
7.59
7.60
7.75
7.61
7.61
7.56
7.53
7.92
8.07
7.92
8.39
8.24
8.05
7.96
7.77
7.92
7.89
0.05
0.06
0.04
0.06
0.08
0.04
0.05
0.05
0.07
0.05
0.06
0.06
0.06
0.11
0.09
0.06
0.06
0.07
0.04
0.06
0.05
0.06
0.08
0.07
0.05
0.06
0.07
0.04
L
L
L
L
L
L
L
axis
axis
axis
axis
axis
axis
axis
86.061
86.155
86.230
86.243
86.339
86.439
86.531
-41.013
-41.072
-41.098
-41.137
-41.192
-41.254
-41.313
2743
2648
2500
2585
2547
2498
2493
705.3
715.5
722.3
725.9
735.9
746.8
756.9
8.56
8.64
8.50
8.24
8.58
8.73
8.63
0.06
0.06
0.05
0.05
0.06
0.06
0.06
206Pb/204Pb 207Pb/204Pb 208Pb/204Pb
18.298
18.156
18.012
17.947
18.144
17.996
17.694
17.617
17.600
15.538
15.511
15.490
15.464
15.522
15.506
15.465
15.441
15.436
38.785
38.465
38.090
37.933
38.518
38.034
37.715
37.524
37.572
17.584
15.460
37.627
17.591
15.438
37.687
17.675
15.447
17.735
15.487
176Hf/177Hf eps Hf
0.283124
12.45
0.283111
11.99
37.659
0.283202
15.21
37.760
0.283133
12.77
Sample
BMRG06 35-8
GSR2 WC A2-14
GSR2 WC A2-13
GSR2 WC A2-12
GSR2 WC A2-11
GSR2 WC A2-10
BMRG06 34-1
GSR2 WC A2-9
GSR2 WC A2-8
GSR2 WC A2-7
GSR2 WC A2-6
GSR2 WC A2-5
GSR2 WC A2-4
GSR1 WC-A2
GSR2 WC A2-3
GSR2 WC A2-2
GSR2 WC A2-1
BMRG06 33-1; -2
Segment
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
Location
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
Longitude Latitude Depth (m) Distance from StPaul 3He/4He 2-sigma 206Pb/204Pb 207Pb/204Pb 208Pb/204Pb
86.632 -41.322
2560
764.0
8.86
0.06
86.636 -41.366
2506
767.3
8.65
0.05
86.756 -41.366
2356
775.1
8.72
0.06
86.847 -41.430
2296
785.5
8.72
0.06
86.938 -41.494
2359
795.8
8.39
0.07
87.033 -41.553
2280
806.1
8.53
0.05
87.092 -41.517
2480
807.4
8.90
0.05
17.812
15.497
37.886
87.163 -41.531
2385
812.9
8.92
0.06
87.262 -41.588
2458
823.3
9.52
0.06
87.348 -41.650
2439
833.2
9.65
0.07
87.438 -41.713
2442
843.4
9.51
0.05
87.530 -41.773
2455
853.6
9.60
0.06
87.627 -41.833
2453
864.0
10.05
0.07
87.724 -41.893
2427
874.5
9.41
0.06
17.8987
15.4899
37.9453
87.801 -41.951
2434
883.4
9.30
0.06
87.881 -42.005
2398
892.3
9.38
0.07
87.977 -42.057
2363
902.1
9.51
0.07
88.042 -42.115
2450
910.3
9.70
0.05
17.970
15.501
37.979
GSR1 WC B7
GSR1 DR B4-1
WW10 69-1
GSR1 WC B5
WW10 66-9
WW10 67-13
GSR1 DR-B3-1
GSR1 DR-B3-5
GSR1 WC B2
GSR1 DR B2-9
GSR1 DR B1-14
GSR1 WC B21
GSR1 WC B20
C17
C17
C17
C17
C17
C17
C17
C17
C17
C17
C17
C17
C17
axis
axis
axis
axis
smt
smt
axis
axis
axis
axis
axis
axis
axis
88.699
88.835
88.922
89.057
89.313
89.183
89.153
89.153
89.412
89.540
89.694
89.797
89.878
-41.704
-41.803
-41.870
-41.921
-41.713
GSR1 WC B19b
GSR1 DR B8-1
WW10 70-32
GSR1 WC B18
GSR1 DR B7-2
GSR1 WC B16
GSR1 WC B15b
GSR1 WC B14
WW10 71-1
GSR1 WC B13
GSR1 DR B6-2; -1
GSR1 WC B12
WW10 65-9
GSR1 WC B8
C16
C16
C16
C16
C16
C16
C16
C16
C16
C16
C16
C16
C16
C16
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
89.977
90.172
90.187
90.242
90.335
90.512
90.595
90.677
90.795
90.843
90.953
91.034
91.095
91.401
-42.482
-42.560
-42.570
-42.581
-42.652
-42.758
-42.784
-42.828
-42.886
-41.878
-41.960
-41.960
-42.169
-42.230
-42.327
-42.360
-42.421
-42.927
-42.984
-43.038
-43.082
-43.272
176Hf/177Hf eps Hf
0.283097
11.49
0.283087
11.14
0.283042
9.55
2620
2560
2353
2392
1375
1850
2243
2243
2350
2400
2350
2531
2486
924.9
940.6
950.8
963.1
965.9
968.5
972.1
972.1
1003.1
1015.6
1032.3
1041.2
1050.6
9.40
9.51
9.14
9.48
10.22
9.70
7.63
7.66
8.92
9.52
8.43
8.68
8.65
0.07
0.09
0.05
0.06
0.08
0.06
0.06
0.07
0.06
0.06
0.13
0.06
0.07
18.0498
15.4855
37.9225
0.283106
11.82
18.030
18.0245
18.1688
18.0577
18.1151
15.473
15.4859
15.4935
15.4831
15.5447
37.892
37.9059
38.0124
37.9073
38.0839
0.283108
0.283099
0.283119
0.282862
11.88
11.58
12.27
3.17
17.9807
17.9629
17.8959
17.9443
17.9423
15.4777
15.4950
15.4718
15.4723
15.4713
37.7676
37.6986
37.6701
37.7076
37.7055
0.283123
0.283086
0.283117
0.283131
0.283120
12.42
11.11
12.21
12.70
12.30
2505
2790
2575
2530
2530
2672
2430
2380
2350
2255
2360
2380
2494
2790
1061.1
1079.0
1080.7
1085.1
1095.9
1114.6
1121.7
1130.0
1141.6
1147.4
1158.4
1167.2
1174.1
1206.5
8.46
8.81
8.61
8.43
8.53
8.30
8.33
8.52
8.02
8.20
8.41
8.32
8.35
8.19
0.12
0.06
0.05
0.06
0.06
0.21
0.07
0.06
0.04
0.06
0.07
0.05
0.04
0.08
17.9199
17.9329
17.910
17.8969
17.9065
17.8785
17.8214
17.7824
17.863
17.8354
18.0420
17.8410
17.868
17.8527
15.4685
15.4734
15.474
15.4646
15.4696
15.4635
15.4608
15.4573
15.479
15.4633
15.4841
15.4635
15.480
15.4678
37.6191
37.6412
37.646
37.5682
37.6200
37.5556
37.5029
37.4902
37.456
37.4884
37.9118
37.5114
37.564
37.5135
0.283141
0.283125
0.283126
0.283153
0.283133
0.283142
0.283167
0.283171
0.283079
0.283151
0.283107
0.283150
0.283121
0.283137
13.03
12.48
12.52
13.49
12.77
13.08
13.98
14.11
10.86
13.40
11.85
13.37
12.34
12.90
Sample
WW10 72-2
GSR1 DR B5-1
WW10 73-8
GSR1 DR B11-2
GSR1 WC B36
Segment
C16
C16
C16
C16
C16
Location
smt
axis
axis
axis
axis
GSR1 DR B10-2
GSR1 WC B33
GSR1 WC B32
GSR1 WC B30
WW10 75-4
GSR1 WC B29
GSR1 WC B28
GSR1 WC B27
GSR1 DR B9-1
GSR1 WC B26b
WW10 76-1
GSR1 WC B25
GSR1 WC B24b
GSR1 WC B23
GSR1 WC B22
GSR1 DR B12-1
WW10 77-7
GSR1 WC B52
GSR1 DR B15-2
GSR1 WC B51
GSR1 DR B14-1
GSR1 WC B49b
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
C15
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
92.228
92.306
92.398
92.630
92.662
92.764
92.845
92.933
93.017
93.086
93.114
93.234
93.355
93.456
93.547
93.648
93.773
93.853
93.940
94.032
94.149
94.310
-43.347
-43.386
-43.438
-43.557
-43.575
-43.633
-43.692
-43.743
-43.796
-43.828
-43.882
-43.897
-43.922
-43.980
-44.023
-44.078
-44.117
-44.158
-44.188
-44.231
-44.274
-44.413
2400
2583
2557
2540
2610
2558
2484
2509
2580
2531
2653
2673
2634
2657
2721
2703
2795
2906
2800
2926
2820
3042
1265.8
1273.4
1282.8
1305.7
1309.0
1319.4
1328.5
1337.6
1346.4
1353.0
1358.3
1367.1
1376.6
1387.0
1395.7
1405.8
1416.4
1424.3
1431.9
1440.6
1451.0
1470.3
8.52
8.43
8.51
8.37
8.37
8.62
8.64
8.49
8.53
8.49
8.46
8.52
8.01
8.00
8.10
7.97
8.34
8.18
8.43
8.10
8.23
8.15
0.06
0.07
0.06
0.06
0.04
0.05
0.06
0.06
0.07
0.07
0.04
0.05
0.06
0.07
0.07
0.07
0.04
0.07
0.06
0.05
0.06
0.05
17.8681
17.8379
17.9037
17.8691
17.844
17.8674
17.8600
17.8789
17.8980
17.8554
17.874
17.8634
17.7708
17.8560
17.7982
17.8682
17.793
17.8109
17.8431
17.8284
18.4880
17.8194
15.4864
15.4828
15.4965
15.4834
15.478
15.4774
15.4806
15.4814
15.5186
15.4802
15.484
15.4772
15.4503
15.4662
15.4584
15.4681
15.465
15.4634
15.4734
15.4619
15.5157
15.4653
37.5861
37.5148
37.5892
37.6052
37.547
37.5737
37.5811
37.6216
37.6271
37.5567
37.586
37.5731
37.3351
37.4620
37.3963
37.4612
37.443
37.4491
37.5278
37.4537
38.2353
37.4629
0.283107
0.283114
0.283076
0.283096
0.283043
0.283117
0.283110
0.283098
0.283038
0.283112
0.283081
0.283111
0.283180
0.283156
0.283150
0.283148
0.283161
0.283155
0.283120
0.283152
0.283052
0.283152
11.85
12.08
10.76
11.44
9.58
12.20
11.94
11.52
9.40
12.04
10.93
12.00
14.42
13.59
13.38
13.31
13.76
13.55
12.29
13.44
9.92
13.42
GSR1 WC B47
GSR1 DR B13-2
GSR1 WC B45
GSR1 WC B44b
GSR1 WC B43b
WW10 78-2
GSR1 WC B42
GSR1 WC B41
GSR1 WC B40
GSR1 WC B39
GSR1 WC B38
WW10 79-19;17
WW10 81-1
WW10 82-35
WW10 83-17
WW10 84-7
GSR1 WC B55
WW10 85-19
GSR1 WC B54
GSR1 WC B53
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
C14
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
prop. tip
axis
fz
axis
axis
axis
axis
94.336
94.426
94.604
94.680
94.769
94.833
94.925
95.010
95.088
95.258
95.340
95.409
95.590
95.585
95.764
95.932
96.034
96.127
96.217
96.313
-44.569
-44.600
-44.724
-44.762
-44.805
-44.833
-44.884
-44.923
-44.957
-45.043
-45.082
-45.106
-45.174
-45.185
-45.051
-45.111
-45.156
-45.195
-45.250
-45.303
2817
2760
2696
2667
2694
2719
2779
2919
2727
2870
2960
3020
3190
3180
3069
2580
2674
2555
2772
2825
1481.7
1489.5
1508.9
1516.2
1524.6
1530.6
1539.8
1547.8
1554.9
1571.3
1579.1
1585.1
1601.1
1601.4
1605.1
1619.8
1629.3
1637.7
1646.9
1656.5
7.75
7.68
7.78
7.73
7.86
7.71
7.80
7.84
7.77
7.58
7.69
7.66
7.81
7.75
8.00
8.08
8.18
8.12
8.09
8.08
0.05
0.05
0.07
0.06
0.06
0.04
0.07
0.06
0.05
0.05
0.05
0.07
0.06
0.04
0.24
0.04
0.06
0.07
0.05
0.07
17.8647
17.8673
17.8721
17.8744
17.8877
17.871
17.8648
17.9117
17.8580
17.8609
17.8725
17.8662
17.883
17.926
18.1206
17.754
17.8807
17.8400
17.8163
17.7971
15.4660
15.4764
15.4824
15.4838
15.4877
15.489
15.4772
15.4837
15.4774
15.4861
15.4835
15.4879
15.479
15.479
15.4992
15.463
15.4751
15.4722
15.4705
15.4708
37.4748
37.5159
37.5190
37.5225
37.6340
37.537
37.5232
37.5800
37.5126
37.5371
37.5442
37.5548
37.546
37.550
37.9203
37.498
37.6461
37.5927
37.5781
37.5518
0.283157
0.283139
0.283148
0.283130
0.283092
0.283074
0.283142
0.283117
0.283148
0.283145
0.283139
0.283143
13.60
12.98
13.29
12.67
11.31
10.68
13.08
12.20
13.30
13.18
12.99
13.11
0.283153
0.283134
0.283139
0.283137
0.283143
0.283144
0.283144
13.47
12.81
12.98
12.92
13.12
13.15
13.15
C13/14
C13/14
C13/14
axis
axis
axis
95.726
95.773
95.831
-45.947
-46.018
-46.047
3500
3137
3195
1656.8
1664.1
1669.6
7.82
7.78
7.90
0.07
0.06
0.06
17.8729
17.8451
17.8852
15.4703
15.4655
15.4708
37.5551
37.5186
37.5909
0.283125
0.283146
0.283115
12.50
13.23
12.12
GSR1 DR C6-2
GSR1 WC C49
GSR1 WC C48
Longitude Latitude Depth (m) Distance from StPaul 3He/4He 2-sigma 206Pb/204Pb 207Pb/204Pb 208Pb/204Pb
91.283 -43.417
1470
1208.3
8.34
0.09
17.9660
15.4707
37.6750
91.503 -43.337
2670
1217.5
8.53
0.06
17.8894
15.4722
37.5998
91.684 -43.468
2738
1237.9
8.24
0.05
17.868
15.485
37.569
91.858 -43.450
2737
1248.1
8.36
0.07
17.8857
15.4703
37.5957
91.941 -43.503
2910
1257.0
8.33
0.17
17.8685
15.4646
37.5611
176Hf/177Hf eps Hf
0.283132 12.73
0.283135 12.82
0.283043
9.58
0.283152 13.44
0.283158 13.65
Sample
GSR1 DR C8-1
WW10 WC48
GSR1 DR C5-12
GSR1 DR C7-3
GSR1 WC C46
GSR1 WC C45
Segment
C13/14
C13/14
C13/14
C13/14
C13/14
C13/14
Location
smt
axis
axis
smt
axis
axis
Longitude Latitude Depth (m) Distance from StPaul 3He/4He 2-sigma 206Pb/204Pb 207Pb/204Pb 208Pb/204Pb
96.120 -45.833
2000
1675.5
7.88
0.18
17.7729
15.4586
37.3603
95.926 -46.090
2990
1678.4
7.31
0.04
18.494
15.519
38.250
95.932 -46.088
2994
1678.7
7.35
0.06
17.8229
15.4690
37.4754
95.867 -46.215
1460
1682.1
7.98
0.06
17.8279
15.4605
37.5359
96.112 -46.217
3110
1698.1
7.83
0.06
17.8554
15.4675
37.5352
96.189 -46.257
3206
1705.4
7.75
0.05
17.8430
15.4652
37.5114
176Hf/177Hf eps Hf
0.283172 14.16
0.283046
9.69
0.283140 13.00
0.283136 12.87
0.283147 13.25
0.283131 12.70
GSR1 WC C43
GSR1 WC C42
GSR1 WC C38
WW10 87-38
GSR1 WC C36
GSR1 WC C35
GSR1 WC C34
GSR1 WC C33
GSR1 DR C4-1
GSR1 WC C31
WW10 88-1
GSR1 WC-C30
GSR1 WC C29
GSR1 WC C28
GSR1 WC C27
GSR1 WC C26
GSR1 WC C25
GSR1 WC C24
GSR1 WC C23
GSR1 WC C22
WW10 89-4
WW10 89-107
GSR1 WC C20
GSR1 WC C19b
GSR1 WC C18
GSR1 WC C17
GSR1 WC 15b
GSR1 WC C13b
GSR1 WC C11b
WW10 90-1
GSR1 WC C9
GSR1 DR C3-4
GSR1 WC C7b
GSR1 DR C2-2
WW10 91-3
GSR1 WC C4
WW10 92-1
GSR1 WC C2b
GSR1 DR C1-1
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
C13
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
axis
96.059
96.117
96.298
96.368
96.451
96.512
96.571
96.627
96.692
96.764
96.833
96.896
96.948
97.009
97.089
97.160
97.225
97.262
97.319
97.378
97.512
97.512
97.554
97.603
97.677
97.740
97.862
97.995
98.094
98.157
98.237
98.415
98.453
98.545
98.601
98.770
98.943
99.035
99.174
-46.624
-46.656
-46.746
-46.778
-46.834
-46.872
-46.896
-46.925
-46.952
-47.053
-47.076
-47.120
-47.160
-47.191
-47.210
-47.242
-47.268
-47.315
-47.356
-47.385
-47.445
-47.445
-47.476
-47.500
-47.533
-47.569
-47.623
-47.656
-47.697
-47.711
-47.790
-47.832
-47.880
-47.897
-47.909
-47.978
-48.101
-48.116
-48.133
2517
2517
2465
2520
2583
2511
2556
2580
2620
2576
2568
2597
2603
2591
2578
2530
2545
2517
2448
2413
2450
2450
2434
2435
2477
2487
2537
2696
2778
2570
2700
2607
2741
2640
2900
2830
2668
2811
2730
1719.3
1724.9
1741.8
1748.2
1756.9
1763.1
1768.3
1773.6
1779.4
1790.1
1795.8
1802.4
1808.1
1813.9
1820.1
1826.5
1832.3
1837.4
1843.4
1848.9
1861.0
1861.0
1865.5
1870.0
1876.7
1882.8
1893.7
1904.1
1912.8
1917.6
1927.3
1941.1
1946.3
1953.1
1957.4
1972.2
1990.2
1996.9
2006.7
8.50
7.86
8.48
8.37
8.34
8.32
8.19
8.37
8.27
8.70
9.46
8.39
8.31
8.35
8.52
8.31
7.96
8.16
8.20
8.11
8.08
8.33
7.88
7.91
7.61
7.70
8.01
7.80
7.85
7.90
7.86
7.84
7.80
7.98
7.92
7.86
7.82
7.77
7.82
0.08
0.06
0.07
0.05
0.05
0.14
0.05
0.05
0.05
0.06
0.05
0.06
0.05
0.05
0.06
0.05
0.06
0.07
0.06
0.06
0.05
0.05
0.06
0.05
0.04
0.05
0.05
0.05
0.05
0.04
0.05
0.08
0.05
0.05
0.04
0.06
0.04
0.05
0.09
17.7245
18.1230
17.7448
17.731
17.8007
17.7847
17.7971
17.7773
17.8058
17.8703
17.841
17.8416
17.8871
17.8858
17.7622
17.9515
17.8890
17.9048
17.9081
17.9180
17.937
17.859
17.9202
17.9230
18.0536
17.9015
17.9280
17.9935
17.9210
17.843
17.8993
17.8966
17.8748
17.9041
17.914
17.9134
17.891
17.9009
17.9164
15.4775
15.4969
15.4779
15.470
15.4747
15.4768
15.4755
15.4760
15.4731
15.4707
15.481
15.4775
15.4784
15.4777
15.4709
15.4798
15.4703
15.4800
15.4807
15.4801
15.499
15.492
15.4806
15.4796
15.4804
15.4624
15.4797
15.4820
15.4811
15.483
15.4792
15.4794
15.4759
15.4805
15.483
15.4789
15.486
15.4783
15.4806
37.6277
37.9258
37.6538
37.596
37.6410
37.6341
37.5986
37.6161
37.6371
37.7432
37.786
37.6817
37.7149
37.7044
37.6068
37.7547
37.6584
37.7054
37.7278
37.7157
37.801
37.743
37.6915
37.6938
37.7766
37.5348
37.7114
37.7085
37.6625
37.614
37.6243
37.6981
37.5745
37.6684
37.682
37.6510
37.578
37.5939
37.6669
0.283112
0.283082
0.283108
0.283143
0.283106
0.283119
0.283110
0.283117
0.283112
0.283121
0.283137
0.283126
0.283116
0.283117
0.283124
0.283119
0.283106
0.283115
0.283114
0.283119
12.04
10.95
11.89
13.12
11.80
12.25
11.94
12.21
12.04
12.35
12.91
12.52
12.15
12.20
12.46
12.27
11.80
12.13
12.09
12.27
0.283078
0.283111
0.283128
0.283102
0.283117
0.283130
0.283134
0.283111
10.82
11.99
12.60
11.65
12.19
12.67
12.79
11.99
0.283114
0.283130
0.283128
0.283114
0.283126
0.283135
0.283074
0.283108
0.283092
12.10
12.66
12.59
12.09
12.52
12.82
10.68
11.87
11.31
WW10
WW10
WW10
WW10
96-1
98-3
100-1
103-4
C12
C12
C12
C12
axis
axis
axis
axis
100.672
100.961
101.530
102.540
-47.335
-47.458
-47.630
-48.019
2465
2583
2850
3230
2059.3
2085.0
2131.6
2218.4
7.36
7.33
7.32
7.18
0.04
0.04
0.04
0.04
18.133
18.257
18.105
17.951
15.483
15.512
15.485
15.465
37.923
38.102
37.944
37.719
0.283070
10.54
0.283060
10.18
WW10 105-1
C11
axis
103.038
-47.765
2783
2238.1
7.30
0.04
17.987
15.480
37.802
0.283087
11.14
Sample
WW10 106-4
WW10 110-4
Segment
C11
C11
Location
axis
axis
WW10 111-18
C10
axis
104.662
-48.212
3015
2368.3
7.16
0.04
18.322
15.506
38.217
0.283094
11.39
WW10
WW10
WW10
WW10
WW10
C9
C9
C9
C9/8
C9/8
axis
seamount
axis
transform
transform
105.224
105.588
105.868
106.494
106.494
-48.752
-49.116
-49.229
-48.873
-48.873
3630
2875
3696
3696
4835
2431.9
2473.4
2497.2
2521.6
2521.6
6.82
7.36
6.80
0.04
0.04
0.04
37.935
38.268
38.005
38.175
38.223
11.49
0.06
15.467
15.549
15.495
15.487
15.500
0.283097
6.82
18.144
18.351
18.060
18.233
18.251
0.283087
0.283130
0.283135
11.14
12.66
12.84
WW10 117-1
WW10 118-1
WW10 122-1
C8
C8
C8
axis
axis
axis
107.145
107.527
108.200
-48.348
-48.428
-48.700
3520
2673
3255
2540.6
2570.0
2627.5
6.88
6.80
6.65
0.04
0.04
0.04
18.284
18.365
18.282
15.508
15.503
15.491
38.249
38.353
38.273
0.283050
9.83
WW10
WW10
WW10
WW10
WW10
WW10
C7
C7
C7
C7
C7
C7
transform
axis
axis
axis
axis
axis
108.513
109.105
109.484
109.484
110.412
110.412
-49.025
-49.450
-49.527
-49.527
-49.828
-49.828
3935
3440
3240
3240
2745
2745
2662.9
2721.0
2749.5
2749.5
2823.9
2823.9
6.83
6.68
7.07
0.04
0.03
0.04
18.011
18.128
15.459
15.475
37.845
38.047
0.283119
12.27
7.13
7.70
0.05
0.04
18.198
18.132
15.512
15.533
38.262
38.193
WW10 130-1
C6
axis
111.133
-49.775
3515
2869.8
7.30
0.04
18.064
15.468
37.964
0.283146
13.23
WW10 132-1
WW10 134-1
WW10 135-8
C5
C5
C5
axis
axis
prop. tip
111.783
112.490
112.593
-50.212
-50.295
-50.307
3328
3620
3748
2930.6
2980.9
2988.2
7.23
7.20
7.20
0.06
0.04
0.04
17.972
17.964
17.931
15.478
15.454
15.455
37.829
37.819
37.785
0.283095
0.283185
11.42
14.61
WW10 138-1
WW10 140-5
WW10 141-1
C4
C4
C4
axis
axis
axis
112.855
113.453
113.615
-50.187
-50.300
-50.347
3820
3320
3002
3001.1
3045.3
3057.9
7.28
7.26
7.35
0.04
0.04
0.04
17.968
18.017
17.947
15.443
15.450
15.461
37.722
37.810
37.796
0.283126
12.52
0.283067
10.43
WW10 144-4
WW10 145-1
WW10 145-7
C3
C3
C3
axis
axis
axis
115.212
116.717
116.717
-50.007
-49.273
-49.273
3997
4610
4610
3153.1
3231.3
3231.3
6.84
6.53
7.20
0.04
0.03
0.05
17.894
17.594
17.362
15.459
15.473
15.443
37.887
37.612
37.320
0.283127
0.283274
12.55
17.75
WW10 146-1
C2
axis
117.182
-49.505
4755
3271.0
6.64
0.04
17.974
15.513
38.027
113-7
114-12
115-3
116-2
116-15
124-1
125-1
126-1
126-7
128-1
128-18
Longitude Latitude Depth (m) Distance from StPaul 3He/4He 2-sigma 206Pb/204Pb 207Pb/204Pb 208Pb/204Pb
103.353 -47.878
2950
2264.8
7.11
0.04
18.114
15.477
37.922
103.930 -48.103
3410
2314.3
7.25
0.04
18.062
15.466
37.844
176Hf/177Hf eps Hf
0.283150 13.37
0.283136 12.87
Sample
Segment
Location
Longitude
Latitude
Depth (m)
Distance from StPaul 3He/4He 2-sigma
206Pb/204Pb 207Pb/204Pb 208Pb/204Pb
176Hf/177Hf eps Hf
Data Sources
Hamelin, B., B. Dupré, and C.-J. Allègre (1986)
Pb-Sr-Nd isotopic data of Indian Ocean ridges: new evidence of large-scale mapping of mantle heterogeneities, Earth Planet. Sci. Lett. 76, 288-298.
Michard, A., R. Montigny and R. Schlich (1986)
Geochemistry of the mantle beneath the Rodriguez triple junction and the Southeast Indian Ridge. Earth. Planet. Sci. Lett. 78, 104-114.
Dosso, L., H. Bougault, P. Beuzart, J.-Y. Calvez, and J.-L. Joron (1988)
The geochemical structure of the South-East Indian Ridge, Earth Planet. Sci. Lett. 88, 47-59.
Graham, D. W., K. T. M. Johnson, L. M. Douglas Priebe, and J. E. Lupton (1999)
Hotspot-ridge interaction along the Southeast Indian Ridge near Amsterdam and St. Paul Islands: helium isotope evidence, Earth Planet. Sci. Lett. 167, 297-310.
Johnson, K. T. M., D. W. Graham, K. H. Rubin, K. Nicolaysen, D. Scheirer, D. W. Forsyth, E. T. Baker, and L. M. Douglas-Priebe (2000)
Boomerang Seamount: the active expression of the Amsterdam-St. Paul hotspot, Southeast Indian Ridge, Earth Planet. Sci. Lett. 183, 245-259.
Graham, D. W., J. E. Lupton, F. J. Spera, and D. M. Christie (2001)
Upper mantle dynamics revealed by helium isotope variations along the Southeast Indian Ridge, Nature 409, 701-703.
Mahoney, J. J., D. W. Graham, D. M. Christie, K. T. M. Johnson, L. S. Hall, and D. L. VonderHaar (2002)
Between a hot spot and cold spot: isotopic variation in the Southeast Indian Ridge asthenosphere, 86°-118°E, J. Petrol. 43, 1155-1176.
Chauvel, C., and J. Blichert-Toft (2001)
A hafnium and trace element perspective on melting of the depleted mantle, Earth and Planet. Sci. Lett. 190, 137-151.
Graham, D. W., J. Blichert-Toft, C. J. Russo, K. Rubin, and F. Albarede (2006)
Cryptic striations in the upper mantle revealed by hafnium isotopes in Southeast Indian Ridge basalts, Nature 440, 199-202.
Nicolaysen, K. P., F. A. Frey, K. T. M. Johnson, J. J. Mahoney, and D. W. Graham (2007)
Influence of the Amsterdam/St. Paul hotspot along the Southeast Indian Ridge between 77° and 88°E:
correlations of Sr, Nd, Pb and He isotopic variations with ridge segmentation, Geochem. Geophys. Geosyst. 8, doi:10.1029/2006GC001540.
Hanan, B. B., J. Blichert-Toft, C. Hemond, K. Sayit, A. Agranier, D. W. Graham, and F. Albarede (2013)
Pb and Hf isotope variations along the Southeast Indian Ridge and the dynamic distribution of MORB source domains in the upper mantle, Earth Planet. Sci. Lett. 375, 196-208, doi:10.1016/j.epsl.2013.05.028.