JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 102, NO. B7, PAGES 15,463-15,487,JULY 10, 1997
The Southeast Indian Ridge between 88øE and 118øEß
Gravity anomalies and crustal accretion
at intermediate spreading rates
James R. Cochran
Lamont-DohertyEarthObservatoryColumbiaUniversity,Palisades,
New York
Jean-Christophe
Semp6r61,2
Schoolof Oceanography,
Universityof Washington,Seattle
SEIR Scientific Team 3
Abstract. Althoughslowspreadingridgescharacterized
by a deepaxial valley andfastspreading
ridgescharacterizedby an axial bathymetrichigh havebeenextensivelystudied,the transition
betweenthesetwo modesof axialmorphologyis notwell understood.We conducted
a geophysical surveyof the intermediatespreadingrateSoutheast
IndianRidgebetween88øEand 118øE,a
2300-km-longsectionof the ridgelocatedbetweenthe Amsterdamhot spotandthe AustralianAntarcticDiscordancewheresatellitegravitydatasuggestthatthe Southeast
IndianRidge(SEIR)
undergoesa changefrom an axial high in the westto an axial valley in the east. A basicchangein
axial morphologyis foundnear 103ø30'Ein the shipboarddata;the axisto the westis markedby
an axial high,while a valleyis foundto the east. Althougha well-developedaxialhigh,characteristic of the East Pacific Rise (EPR), is occasionallypresent,the more commonobservationis a
rifted high that is lower andpervasivelyfaulted,sometimeswith significant(> 50 m throw) faults
within a kilometerof the axis. A shallowaxial valley (< 700 m deep)is observedfrom 104øEto
114øEwith a suddenchangeto a deep(>1200 m deep)valleyacrossa transformat 114øE. The
changesin axial morphologyalongthe SEIR are accompanied
by a 500 m increasein near-axis
ridge flank depthfrom 2800 m near88øEto 3300 m near 114øEandby a 50 mGal increasein the
regionallevel of mantleBouguergravityanomaliesoverthe samedistance.The regionalchanges
in depthandmantleBougueranomaly(MBA) gravitycanbe bothexplainedby a 1.7-2.4 km
changein crustalthickness
or by a mantletemperature
changeof 50øC-90øC.In reality,melt supply (crustalthickness)andmantletemperature
arelinked,sothat changesin bothmay occur
simultaneously
andtheseestimates
serveasupperbounds.The along-axisMBA gradientis not
uniform. Pronouncedstepsin the regionallevel of the MBA gravityoccurat 103ø30'E-104øEand
at 114øE- 116øEandcorrespond
to thechangesin thenatureof the axialmorphologyandin the
amplitudeof abyssalhill morphology
suggesting
thatthedifferentformsof morphologydo not
gradeinto eachotherbutratherrepresentdistinctlydifferentformsof axial structureandtectonics
with a sharptransitionbetweenthem. The changefrom an axialhighto an axial valleyrequiresa
thresholdeffectin whichthe strengthof thelithosphere
changes
quickly. The presence
or absence
of a quasi-steady
statemagmachambermay providesucha mechanism.The differentformsof
axialmorphology
arealsoassociated
with differentintrasegment
MBA gravitypatterns.Segments
with an axial highhavean MBA low locatedat a depthminimumnearthe centerof the segment.
At EPR-like segments,
the MBA low is about10 mGal with along-axisgradientsof 0.15-0.25
mGal/km, similarto thoseobservedat the EPR. Rifted highshavea shallowerlow andlower
gradients
suggesting
anattenuated
composite
magmachamber
anda reducedandperhaps
episodic
melt supply. Segmentswith a shallowaxial valleyhaveveryflat along-axisMBA profileswith
little correspondence
betweenaxialdepthandaxialMBA gravity.
Introduction
1Also
at Groupe
deRecherche
enG6od6sie
Spatiale,
UMR-5566,
ObservatorieMidi-Pyr6n6es,Toulouse,France.
The morphology
of fastspreading
andslowspreading
midoceanridgesdiffersdramatically.Fastspreading
ridges,such
astheEastPacificRisespreading
at 65-150mrn/yr,arecharacterizedby an approximately
20-km-wide,
200- to 400-m-high
ridge axishigh and relativelylow relief (50-150m) abyssal
hills on the ridgeflanks[Menard, 1960;Macdonald,1989]
Copyright
1997by theAmerican
Geophysical
Union.
Papernumber97JB00511.
0148-0227/97/97JB-00511 $09.00
2Now
atExxon
Production
Research
Company,
Houston,
Texas.
3Southeast
Indian
Ridge
(SEIR)
Scientific
Team
(D.Christie,
College
of OceanicandAtmospheric
Sciences,
OregonStateUniversity;
M.
Eberle,Department
of Geological
Science,
BrownUniversity;
L. Gel/,
D6partment
deG6osciences
Marines,
!FREMER;
J. A. Goff,University
of TexasInstituteof Geophysics,;
H. Kimura,Earthquake
Research
Institute,Universityof Tokyo; Y. Ma, Lamont-Doherty
Earth
Observatory
of Columbia
University;
A. Shah,
School
of Oceanography,
University
ofWashington;
C. Small,Lamont-Doherty
EarthObservatory
of ColumbiaUniversity;B. Sylvander,Collegeof Oceanicand
Atmospheric
Sciences,
OregonStateUniversity;
B. P. West,Schoolof
Oceanography,
Universityof Washington;
andW. Zhang,LamontDohertyEarthObservatory
of Columbia
University).
15,463
15,464
COCHRAN ET AL.: SOLJTHEAST INDIAN RIDGE GRAVITY ANOMALIES
COCHRAN ET AL.: SOUTHEASTINDIAN RIDGE GRAV1TYANOMALIES
Slow spreadingridges, such as the Mid-Atlantic Ridge spreading at 20-40 mm/yr, are associatedwith 1- to 2-km deep, 25to 40-km-wide axial rift valleys and much rougher(> 500 m
relief) abyssal hills on the ridge flanks [Heezen, 1960;
Macdonald, 1986].
Although the basic difference in the morphology of
"Atlantic-type"and "Pacific-type"ridges has been known for
over 35 years, it is still not completely understood. How the
same basic processesof plate separation,mantle upwelling,
and decompression
melting resultin suchfundamentallydifferent morphology and tectonics is a basic question in understandingcrustal accretion. Intermediatespreadingrate ridges
can provide information on crustal accretion in conditions
other than those representedby the well-studied end-members
and in particular allow the nature of the transitionbetweenthe
very different slow spreading and fast spreadingend-member
axial structures to be examined.
The nature of the transition
is
an important observationin constrainingmodels for the creation of axial morphology [e.g., Chen and Morgan, 1990;
Phipps Morgan et al,, 1987]. However, in spite of their potential importance in understandingcrustal accretion and the
processes that create seafloor morphology, intermediate
88 ø
spreadingrate ridgeshave not been studiednearly as intensely
as have fast and slow spreadingridges.
Field
Program
During the Austral summer of 1994/1995, a geophysical
survey of the intermediate spreading rate Southeast Indian
Ridge(SEIR) from 88øEto 118øEwascarriedout on two legsof
the "Westward" expedition on R/V Melville (Figure 1). The
objectives of the field program were (1) to investigate
changes in axial morphology accompanyingthe west to east
increasein ridge axis depth, including in particular the transition from axial high to axial valley observed near 101øE
[Small and Sandwell, 1992; Ma and Cochran, 1996], (2)to
investigate possible changesin melt supply and distribution
within segments(as evidencedby gravity anomalies)accompanying the changein axial morphologyand (3) to determine
geochemical variations along the ridge which were hypothesized to accompanythe transition in morphology.
The first leg of the field program conducteda geophysical
survey of the ridge axis and flanks between 91øE and 118øE.
The second leg was primarily devoted to rock sampling but
also obtained additional geophysical coverage between 88øE
90 ø
89 ø
15,465
91 ø
92 ø
II
_42 ø
_42 ø
10B
_43 ø
_43 ø
M3
88 ø
89 ø
90 ø
91 ø
92 ø
•;i•...i,.::Z.L?:•i...i.
7•'"'"'•?-'-3•!
•'""
"••••••••••.17'•.'•'•.
.....
'•'"•'
•'•'
"'•""•'""'""•'•::•'•:".
.....................
7:7.....
0
2500
3•0
3500
4•0
4500
5000
6•0
Figure 2. Bathymetricmap of detailedsurveybox 0 at the very westernend of the survey area (Figure 1).
The bathymetryis contouredat 250-m intervalsand color changesoccurat 500-m intervals. The detailedsurvey extendsout to approximately1.2 Ma seaflooron both flanks. Solid line showsthe location of center
beam bathymetryand free-air gravity profile shownin Figure 10 (10B). This map includessecond-ordersegmentsM1, M2, and M3 (Figures7 and 8) which are the shallowestin the surveyarea (Figure 7). The next two
segmentsto the west (K and L) are the shalloweston the SEIR [Johnsonet al., 1996; Ma and Cochran, 1996].
The central portionsof segmentsM1 and M2 have EPR-like axial highs, but a rifted axial high characterizes
most of the survey box.
15,466
COCHRAN
ET AL.:SOUTHEAST
INDIANRIDGEGRAVITYANOMALIES
1O0ø
101 ø
102ø
103ø
1.04ø
105 ø
10A
_47 ø
10c
P1
_48 ø
_48 •
P2
P3
100ø
1O1ø
0
102ø
2500
3000
3500
103 ø
4000
4500
104ø
5000
105ø
6000
Figure 3. Bathymetricmap of detailedsurveybox 1. The bathymetryis contouredat 250-m intervals,and
shadingchangesoccurat 500-m intervals. The detailedsurveyextendsout to approximately
1.2 Ma seafloor
on both flanks. Solid lines show the location of centerbeam bathymetry and free-air gravity profiles shown
in Figure 10 (10A and 10C). This map includessecond-ordersegmentsP1, P2, P3, and P4 (Figures 7 and 8).
The central portion of segmentP1, immediately to the east of the left-stepping 100øE transform,has an EPRlike axial high. SegmentP2 is characterizedby a rifted high and is typical of the majority of segmentswest of
103øE. SegmentS4 marksthe westernend of a regionextendingeastwardto 129øEin which the SEIR is characterizedby an axial valley. Note that the transformat 102ø45'Ehas evolved from a nontransformoffset since
1 Ma.
and 91øE. A 2300-km-long segmentof the ridge axis was surveyed during the two legs including four detailed study "boxes"
which extend for a total of 1100 km along the axis (Figure 1)
and include eight second-orderridge segments. Within each
box, Seabeam 2000 bathymetry, gravity, and magnetics lines
perpendicularor slightly oblique to the ridge axis extending
roughly 45 km (1.2 Myr) away from the axis on each flank
were obtainedat 9-km intervals. This spacingprovidesessentially completebathymetriccoverageand gravity and magnetics lines at closeenoughspacingto allow confidentinterpola-
Gaussian filter to remove ship motions. These total field
measurementsare then sampled at 1-min intervals, merged
with the Global Positioning System (GPS) navigation to
depth, and long-wavelengthtemperaturevariationsalong the
SEIR. In this paper, we will concentrateon the pattern of
gravity anomalies,their relationshipto axial morphologyand
what they reveal aboutcrustalaccretionat intermediatespread-
wavenumbers
of lessthan0.1 km-1 (wavelengths
> 10 km)
determine
the EOtvOs correction
and the free-air
anomalies
calculatedby removal of the predictedgravity at the measurement latitude using the 1967 International Gravity Formula.
Since most applicationsof shipboardgravity data use spatial
rather than temporal dimensions,we used an Akima spline to
resamplethe data at 0.2-nauticalmile (370 m) intervals.
The gravity data resultingfrom this standardprocessingare
tion between tracks. Bathymetricmaps of the four survey characterizedby 2-5 mGal variations with wavelengthsof a
boxes are shown in Figures 2-5.
few kilometers (Figure 6a). This short-wavelength noise,
This is one of two paperspresentingpreliminarygeophysi- which can not be due to sources at or below the seafloor, is
cal results from the two cruises. The companion paper by characteristic of marine gravity data collected with BGM-3
Sempereet al. [this issue]discussesthe segmentation,
axial gravimeters. The spectrumof our resampleddata is linear at
Gravity data were collected throughoutthe survey of the
SEIR using a Bell Aerospace BGM-3 gravimeter [Bell and
Watts, 1986]. The meter output (counts) are acquiredat 1-s
(Figure 6b), as is typical of marine gravity spectra. At higher
wavenumbers(shorterwavelengths),the spectrumis much flatter, representinga noise floor. We attemptedto remove the
short-wavelengthnoise by designinga digital low pass filter
to retain all wavelengthsgreater than 10 km at full power and
roll off to full stop by 5 km (C. Small, manuscriptin preperation). The spectrumof the filtered data is shownin Figure 6b,
and the filtered and unfiltered data can be comparedin Figure
6a. Application of the filter resulted in a significant improvement of crossover errors with the mean crossovererror
intervals, converted to acceleration and smoothed with a 180-s
reduced from 2.0 mGal to 1.1 mGal.
ing rates.
Gravity
Data
Collection
and
Reduction
COCHRANET AL.:SOUTHEAST
INDIANRIDGEGRAVITYANOMALIES
106 ø
107 ø
108'
109'
I10 ø
Ill
15,467
ø
-48 ø
-48 ø
15B
13A
-49 ø
-49 ø
13B
I3C
$2
-50 ø
s1
106 ø
107 ø
0
108 •
2500
3000
109 ø
3500
4000
lIO*
4500
5000
lllø
6000
Figure 4. Bathymetric map of detailed survey box 2 (Figure 1). The bathymetryis contouredat 250-m intervals, and shadingchangesoccur at 500-m intervals. The detailed survey extendsout to approximately 1.2
Ma seaflooron both flanks. Solid lines show the location of centerbeam bathymetryand free-air gravity profiles shown in Figures 13 and 15 (15A, 15B, 13A, 13B, 13C). The map includessegmentsR, S1, S2, and S3
(Figures7 and 8). The ridge axis for 50 km nearthe centerof segmentR is a narrowvolcanicridge on top of a
20-km-wide fault boundedplateau. The plateaudies away at both endsof the segmentwhere the axis is within a
shallow axial valley. Short segmentS2 has nearly been cut off by the propagatingrift at the easterntip of S1.
,
The SoutheastIndian Ridge
The SEIR forms the plate boundarybetweenthe Australian
and Antarctic plates from the Indian Ocean Triple Junction
located near 25øS, 70øE to the Macquarie Triple Junctionnear
63øS, 165øE. Rifting betweenAustralia and Antarcticabegan
in the Late Cretaceouswith slow seafloor spreadingbeginning
by about96 Ma [Cande and Mutter, 1982; Veevers,1986]. The
SEIR has been spreadingat intermediaterates of about 55-75
mm/yr since the Oligocene (Anomaly 13 ) [Royer and
Sandwell, 1989]. Present-dayspreadingrates increaserapidly
from 57.5 mm/yr at the Indian Ocean Triple Junction to 68
mm/yr at Amsterdamand St. Paul Islandsand then more slowly
to a maximum of 76 mm/yr near 50øS, 114øE, before decreasing gradually to 72 mrn/yr at the George V fracture zone near
change in axial morphology near 101øE from an axial high
toward the west to an axial valley toward the east is of particular interest because it occurs well away from anomalousfeaturessuchas hot spotsor the AAD in an area wherethe spreading rate is nearly constant.
There is, however, a significant depth gradient acrossthe
region (Figure 7) with axial depthsincreasingfrom about 2300
m near 88øE to nearly 4500 m eastof 114øEnear the AAD [Ma
and Cochran, 1996]. Ridge flank depthsincreaseby about500
m from 2800 m to 3300 m over the same region. [Ma and
Cochran, 1997]. Forsyth et al. [1987] determined that the
shear wave velocity beneath the AAD is considerably faster
than for similar age mantle in other parts of the SEIR and the
Pacific, which can be interpreted to result from colder than
50øS,
139øE
[DeMets
etal.,1990].Spreading
rates
inthesur- normal
mantle
in thatregion.Majorelement
chemistry
imveyareaarenearly
constant
at72-76mm/yr,
within
therange-pliesshallower
andsmaller
extents
of melting
beneath
the
AAD than in adjacent areas, also suggestingthe presenceof
colder mantle beneath the AAD [Klein et al., 1991; Pyle,
1994]. This is further supportedby interpretation of gravity
Macdonald, 1986; Malinverno, 1993].
data
[West et al., 1994] and by recentseismicrefractionresults
Satellite altimetry data [Small and Sandwell, 1994] and the
from
the AAD which show thin crust (4.2 km), indicating a low
sparsepre-1995 shipboarddata from the SEIR [Cochran, 1991;
Ma and Cochran, 1996] show that transitions from axial lows melt supply consistent with a colder than average mantle
to axial highsoccur at a numberof placeson the SEIR, specifi- [Tolstoyet al., 1995]. The SEIR between88øE and 120øEthus
cally near 82øE, near 101øE, and near 127øE at the eastern presentsthe opportunity to conduct a "controlled experiment"
boundaryof the Australian-AntarcticDiscordance(AAD). The in which the effects on the crustal accretionprocessesof varyof 70-80 mm/yr where the transitionfrom an axial high to an
axial valley is observedto occur [Small and Sandwell, 1989;
15,468
COCHRANET AL.:SOUTHEAST
INDIANRIDGEGRAVITYANOMALIES
112ø
_49ø
113ø
114ø
115 ø
1t6 ø
117•
-49 ø
.....
..
ß
ß
.......
:•.:...:::.
•..•.............,•
S4
_50
o::•:,:•?::::•:
S3....
-..'...
-,ß
"
.•......•
....
.
........
5•'••.:.'Y•
.:,.x•).
•'..::,:.
ß
-
z:..::, .:•::.;.
ß
112'
113ø
114ø
115ø
116ø
117ø
!
0
2500 3000 3;00 4000 4500 5000 6000
Figure 5. Bathymetric map of detailed surveybox 3 (Figure 1). The bathymetryis contouredat 250-m intervals, and shadingchangesoccur at 500-m intervals. The detailed survey extendsout to approximately1.2
Ma seaflooron both flanks. Solid lines show the location of the centerbeam bathymetryand free-air gravity
profiles shownin Figure 14. The map includessegmentsS4 and T, as well as the easternend of S3 (Figures7
and 8). The axial morphology changesfrom a shallow valley to a deep valley acrossthe 114øE transform,
which appearsto form the effective westernboundaryof the AAD. An eastwardpropagatingrift has recently
stalledjust west of the 114øEtransformand hascut an en echelonseriesof shearsto reachthe transform. The
recentlyeliminatedridge tip can be identifiednear 50øS, 114øE.
20
I
,
I
,
I
,
I
•
I
,
I
•
I
20
-
10
-
0
%1 i !
-10
-10
-30
-20
- 10
0
10
20
30
2400
t I , ! • I • I , I , I • I t 2400
.
2800
2800
3200
3200
3600
3600
-30
-20
-10
0
10
20
30
Kilometers
Figure6a. Center
beamSeabeam
bathymetry
andfree-air
gravity
profile
across
theSEIRaxisat
105ø33.1'E.
Thedotted
gravity
curve
shows
theraw(unfiltered)
data.Thesolidcurveshows
thesame
data
afterapplication
ofthefiltershown
inFigure
6btoremove
short-wavelength
noise.
COCHRANET AL.:SOUTHEAST
INDIANRIDGEGRAVITYANOMALIES
15,469
FrequencyResponse
BGM-3 Gravity Filters
10km5kin
I
:
:
ß
0.15
-
:
.
m
0.10
-
0.05
-
0.00
-
!i"
f/•'"'!'["
p•"I'
t.i:•",..i..•
'!..]Fit-te
....
_
-4-
I
-10
-5
0
5
10
km
0.5
0.0
1.0
k (km-1)
Figure6b. (left)Filterconvolved
withfree-air
gravity
datato remove
short-wavelength
noise.Filterfully
retains
all wavelengths
greater
than10km androllsoff to full stopat 5 km. (right)Spectra
of therawandfilteredfree-airgravitydata. Thefilterremoves
the"noisefloor"present
at wavelengths
shorterthan10 km in
the raw data.
ing the mantletemperature
at a constant
spreading
ratecanbe
examined.
The first-order
segmentation
of theSEIRasdetermined
by
constructed
by digitizingthe locationof the ridge axis at 1 arc
min (- 1.2 km at this latitude)intervalsof longitudeon detailed Seabeammapsand extractingthe depthat thoseloca-
C. Small et al. (The structureand segmentation
of the tionsfrom the griddedbathymetry.The along-axis
depthproSoutheast
IndianRidge,submitted
to Marine Geology,1997) file in Figure7 confirmsthe generalincreasein axial depth
is shownin Figure8. Theletternamesassigned
to theseg- from westto eastdeterminedfrom olderwidely spacedtracks
ments continuethe conventionestablishedby Royer and
[Ma and Cochran, 1996]. These data also confirm the observa-
Schlich[1988]for theportion
of theSEIRfromtheRodriques tion made by Ma and Cochran [1996] that the rate of axial
TripleJunction
to Amsterdam
Islandandextended
by Johnson deepening
increases
to the eastof about100øE. Axial depths
et al. [1996] to the region from AmsterdamIsland to 88øE.
increasegradually from about 2400 m to about 2800 m be-
Thereare eight-firstordersegments
in the surveyareawhich tween89øEand 102øEandthen moresteeplyto about3700 m
rangein along-axislengthfrom 55 km (segmentN) to over near 113øE. There is a sharp,step-likeincreasein the axial
700km (segment
M). Segment
boundaries
weredetermined
by depthsto about4500 m acrossa transformfault at 114øE which
mapping stable, long-lived fracture zone traces in satellite marksthe transition
from a shallowto a deeprift valley.
gravitydata(C. Smallet al., submitted
manuscript,
1997) All
The along-axisdepthprofile in Figure7 alsoshowsthat the
of thefirst-order
segment
boundaries
aremarkedby transform increasein axial depth is not monotonic. At an intermediate
faultswith present-day
offsetsof 37 to 130 km.
scaleof 300-500 km the eastwardincreasein depthis moduThe first- andsecond-order
ridgesegmentation
is alsoindi- lated by the major long-livedtransforms.Axial depthsbecatedon Figure7. The second-order
segments
are lessstable come shallower by a few hundred meters from west to east
over time, and in several cases the nontransform offsets acrosstransforms
at 96øE,100øE,and 106ø30'Eanddeepenby
bounding
themappearto be propagating
to theeast. The east- nearly 1000 m acrossthe transform at 114øE.
ernboundaries
of segments
M3, M5, S1, andS3areall propa- The along-axisvariationin ridge flank depths[Ma and
gatingto the east(Figures4, 5, and9). A propagating
rift Cochran,1997]is muchlessthanthevariationin axialdepth
whichformedthe easternboundaryof segmentS4 hasre6ently (-500 m ratherthan -2100 m) (Figure7) which is to be excollided with the 114øE transform. The remnants of the elimipectedsincethe axial depthreflectsnot onlythe changein the
nated ridge segmentcan be seenjust west of the 114øEtrans-
depthof the ridgebut alsochangesin the form andrelief of the
formnear50øS(Figure5). The boundarybetweensegments
P3 axialtopography.The ridgeflank depthsalsoappearto vary
andP4 alsoconsistsof propagating
rifts, but herethe propa- more smoothly and continually than the axial depths.
gationappearsto be oscillatoryor "dueling,"so that therehas However,we can only determineridge flanksdepthfor the
beenlittle net motionsince1 Ma (Figure3).
eight segments
for whichwe possess
significantoff-axisdata
(Figures
2-5). Thuswe cannotdetermine
whether
is a stepin
Morphologyof the SoutheastIndian RidgeAxis
An along-axisprofile of axial depth on the SEIR from
89.5øEto 115.5øEis shownin Figure 7. The profile was
ridge flank depthsacrossthe transformsat 96øE, 100øE,and
106ø30'E. Ridgeflank depthsdo not changesignificantly
across
thetransform
at 114øEin spiteof the 1000m change
in
axial depth(Figures5 and 7).
15,470
COCHRAN
ETAL.:SOUTHEAST
INDIANRIDGEGRAVITYANOMALIES
-2500
-2500
,
I
-3000
-3000
-3500
-3500
-4000
-4000
E
•..
-4500
10
o
-lO
03
,3,
,I[,4,
,I,5,[611
I,,•,,'[l,1
,[,2'
I•l,3,
I[''' i....
t
N
-4500
10
-10
-2o
•E -30
-30
r/'t -4o
-40
-
-50
-60
-60
-70
-50
•
'
'
AxialHigh
• [ ', ,•
"-
I
90
'
'
'
I
92
'
'
'
i
94
'
'
'
I
96
'
'
'
I
98
'
'
'
I
100
'
'
'
i
'
102
'
'
I
104
'
'
'
I
106
Axial Valley
'
'
'
I
'
-70
'
108
Longitude
Figure 7. (top) Axial andridgeflank depthas a functionof longitudealongthe SEIR. The axial depthprofile, shownas a heavysolidline, wasconstructed
by digitizingthe locationof the ridge axis at 1 arc sec(- 1.2
km) longitudeintervalson detailedSeabeam
mapsandextractingthedepthat thoselocationsfrom the gridded
bathymetry. The ridge flank depths,shownas light lines (solid for the north flank and dashedfor the south
flank) for the eight segments
containedwithin the detailedstudyboxes,were determinedby Ma and Cochran
[1997] from the deterministiccomponentof an empiricalorthogonalfunctionanalysisof the bathymetric
data. Axial depthincreases
from west to eastfrom 2350 m in segment
M1 to 4000-4500m in segment
T. The
accompanying
changein ridgeflank depthis -500 m (2800-3300m). First- and second-order
segmentation
of
the axisof the SEIR is shownalongthe bottom. (bottom)Axial mantleBougueranomaly(MBA) as a function
of longitudealongthe SEIR. Within the four detailedstudyboxes,the griddedMBA gravitywassampledat
1.2-kmintervalsalongthe ridgeaxis. Betweendetailedsurveyboxes,spotvalues(dots)were obtainedwhere
the shiptrackcrossedthe ridgeaxis. Thin solidlinesshowthe regionaltrend. Note the abruptstepsin the
axial MBA level near 104øEand 115øE,whichcorrespond
to changesin axial morphology.
There is also a segmentscale(-100 km) axial depthvariation superimposedon the longer-wavelengthpattern. The
ridge axis is generally shallowesttoward the centerof segments with axial depthsincreasingtoward segmentbounding
offsets. This type of segmentscale depth variation is also
observedat both fast and slow spreadingridgesandappearsto
be in large part relatedto segmentscaleupwellingand melt
distribution [Macdonald et al., 1988, 1991; Sempereet al.,
1990; Lin et al., 1990]. Intrasegmentaxial depthvariations
are generallylessthan 500 m to the westof 103øEbut may be
as great as 1000 m in segmentsboundedby propagatingrifts
and in two segments(P3 andR) in which thereis a changefrom
axial high to valley within the segment.
Satellite altimetry data suggestthat a changefrom an axial
high to an axial valley occurson the SEIR between 100øE and
105øE [Small and Sandwell, 1992]. The shipboardsurvey
determinedthat the transitionoccurswithin segmentP3 which
extendsfrom 47ø43'S, 102ø52'Eto 48ø11'S, 104ø22'E(Figures
3 and 7). All segmentsfor 1900 km west of 103øEto 81øE near
AmsterdamIslandare characterized
by axial highs(Figures2, 3
and 10). The only exception is segmentN, a 55-km-long
segmentnear 95ø45'E, bounded at both ends by transforms
with offsets of 93 and 37 km, which has a 200-500 m deep
valley at its axis (Figure 9). All segmentsfor 1450 km eastof
104ø30'E to the easternboundary of the AAD at 129øE are
characterizedby axial valleys (Figures4, 5, and 11-14). There
is again a single exceptionto this pattern. For about 50 km in
the centerof segmentR, locatedimmediatelyto the eastof the
130 km offset 106.5øEtransform,the axis is formedby a linear volcanic ridge along the apex of a broad fault-bounded
plateau (Figures 4 and 15). The plateau becomesless distinct
to the east and the easternend of the segmenthas a shallow
axial valley (Figure 15).
Althougha transitionfrom an axial high to an axial valley
can be defined near 103ø30'E, the transition is not between
well-developed, end-member morphologies. Both "East
COCHRANET AL.: SOUTHEASTINDIAN RIDGEGRAVITYANOMALIES
95 ø
I
.•;.: :ß . --...•
110'
I
.....
15,471
115'
t• •
•.....
120'
l
l
l
..:.•.•:-•....:::.
:.:..
i
..
":'
!
-50'
::
'
•
•'?.:"'..:::
ß :'
"•" •
..... ,
.•'•
."•:•
......
'•'.'
.... r•?
';,.•;•
•...• • •
%::
.....
( -::.
• •..
i •'
.........
•'•:':i •':"'
:•-,•
:• f•:"";';i :*•'•"
.• '•• .......
•
85'
90'
:"""
...•:.
•: ........
/•.. •4.•:"•"•.•...:?".::"-,
'•:' -"..-•-,
f ;•"::4:.. •.
,• .........
• :•" ?.:.?•-'•
½.
• :r:•
"•
.•-:.'"•'.•.'..
"':
.....
•'?•:&
...... .,•:..:':,/•:.
• :'..: ....." "'?'
:•_. ""...."• .
:....:'..:•'.-•.
'.'27,,•,• ?: ,-.. .-:• •
:•'...'.
.......
•.
.?.' ......
'•
•::
'•
•.':
........
'%..........
'•'•i ".'":'•
'":'"i '• ":'•'""I-::"•
•:':•:i
95'
, •.....
.
•:.'
ß .....
!
-50'
•'•";':-:'/'
•,•:?•-:
.? .&:•'..
:. ':"':•'•"
.:•:.
'!"?'H':•?i:-•½
•:..-:"•;•::'•./;•'
•..... ' ......
.........
"•'"':•
.....
.•(•'.:•;:.
?' ,--•"f•::.,.'•::;.•".qlt,:•.
..•:5'•L•.
' '•: ..•:'•.•
:•':••..•....
.::.• .•;'?''?'•:•:•,.'.....
• ':•
.:•::•:•.:
.½.
;•. -•:.. ...•::
................
'•:i .•:
........'•::'
•
l•'
•
. ...
.......
' '•
105'
•:•:•.
.....
....'•i•
•'.........
•':
I10 •
'•'•......
' ..,• •' •
:?
..........
•
.....
• '•
.-
f
[
':-;•:
k
.. ::•
• '?•
•. •
•'
:-. f
• •-•...•.,. ..• ..
......• ' .......•;.
:....... "'.•
..... i
115•
120'
Figure 8. First-order segmentationof the SEIR superimposed
on a band-passedsatellite gravity image.
Surveytracksof the Westward9 and 10 cruisesof R/V Melville are shownin white. The first-ordersegmentation was establishedby C. Small et al. [submittedmanuscript,1997] throughidentificationof stable longlived segmentboundariesin satellitegravity images. Figure is from C. Small et al. [submittedmanuscript,
1997].
Pacific Rise-type" axial highs and deep "Mid Atlantic Ridge- Gravity Anomalies at the Southeast
type" rift valleys do exist on the SEIR, but they are found only Indian Ridge Axis
in a few specific areas. Well-developed deep rift valleys (>
Free-air and mantle Bouguer gravity anomaly maps of the
1000 m deep) occur only to the east of the 114øE transform
four detailed study boxes, contoured at 5 mGal intervals are
(Figures 5 and 14). The change in morphology across the
shownin Figures 16-19. The free-air anomalymapswere con114øEtransformis quite dramatic(Figure 5), and the transform
structedby gridclingthe filtered shipboardgravity data on a 1.5
appearsto mark the effective western boundaryof the AAD.
arc minx 1 arc min (-1.8 km x 1.8 km) grid and fitting a
Axial valleys on the SEIR to west of 114øEare generally< 500
smooth surfaceusing a continuouscurvaturespline in tension
m deep (Figures 4, 5, and 11-13) except near segmentends
[Smith and Wessel, 1990; Wessel and Smith, 1995]. For calwhere they may deepento 700 m or to about 1000 m at propaculation of mantle Bouguer anomaliesthe gravity effect of the
gating rift tips (Figures4, 5, and 13).
topographicand the assumedMoho relief were determinedfrom
Well-developed axial highs similar to those characteristic
gridded bathymetricdata using the techniqueof Parker [1973],
of the EPR are found only in segmentsS1 and S2 immediately
retaining four terms in the series expansion. A constant
eastof the 88øEtransform(Figure 2) and in the centralportion
crustal thicknessof 6 km was assumedwith density contrasts
of segmentP1 immediately to the east of the 100øE transform
of
1700and600 kg/m3 assumed
acrossthe water/crust
and
(Figures 3 and 10). Otherwise, axial highs on the SEIR are
crust/mantleinterfaces respectively. A 0.5 arc minx 0.5 arc
"rifted highs" which are lower and less well defined and are
min (-600 m x 900 m) topographicgrid was usedin the calcucharacterizedby faults with throws of 50-100 m very near the
lations. The grids of the gravity effectsof the topographicand
axis, often within about 1 km (Figure 10). The easternend of
assumedMoho relief were sampled at the locations of the
S2 and both ends of P1 also more rifted.
shipboardmeasurementsand the values obtainedat those locaShaliow rift valleys and rifted highsthus appearto form a
tions subtractedfrom the coincidentshipboardfree-air gravity
distinctiveand characteristicintermediaterate axial morpholanomalies. The resultingMBA anomalieswere then gridded
ogy which is prevalenton the SEIR. These morphologiesdifand contoured in a manner identical to that used for the free-air
fer from the well-studied, end-membermorphologiesnot only
gravity.
in the amplitudeof the axial relief but also in the natureof the
neovolcanic zone. Faulting is evident throughout the axial
Free-Air and Mantle Bouguer Anomalies
region in areas with these morphologies,and in many areasa
traditional morphologically defined neovolcanic zone is not
Free-air anomalies on the ridge flanks away from large
evident. In fact, the axis can not alwaysbe accuratelydefined topographicfeatures are generally in the range of-5 to 10
on the basis of morphologyalone.
mGal and only exceed+30 mGal over the largesttopographic
15,472
COCHRANETAL.:SOUTHEAST
INDIANRIDGEGRAV1TY
ANOMALIES
95 ø
96 ø
97 •
The free-air anomaliesgenerally reflect the topographicrelief. Most of the smaller-scalefeatlaresrelatedto bathymetric
relief in the free-air anomalies are removed in the MBAs
which
thus appear much smootherat the 5-mGal contour interval of
our maps (Figures 16-19). The three large seamountsin the
westernsurveybox show 5-10 mGal MBA lows (Figure 16)
reflectingisostaticsupportthi'oughcrustalflexureresultingin
slightly thicker crust underneaththe seamounts. The large
free-air anomaliesover transformsare almostentirely removed
by the topographiccorrection.For example,the very complicated free-air anomalies associated with the combination trans-
form/non-transformridge offset at 108.5øE(Figure 18) are replaced in the MBA map by a simple gravity high trending
alongthe spreading
direction
in whichindividual
bathymetric
featurescannotbe identified. In general,transformsappearas
10-15 mGal MBA gravity highs probablyreflectingthinned
crustbeneaththe transforms[Prince and Forsyth, 1988].
Gravity
_46ø
Anomalies
Over
Propagating
Rifts
.46 ø
A distinctive pattern of free-air anomalies is found at the
propagatingrifts at 111øE and 112ø45'E(Figures 18 and 19).
Prominent-25 to -35 mGal free-air anomaly lows are located
over both the propagatingand retreatingrift tips and over a
series of abandoned rift tips along the inner pseudofaults.
Large-amplitude(25 to 30 mGal) positiveanomaliesare found
associatedwith blocky topographichighs located betweenthe
!'•*'
...,:;-;%?--"
propagating and retreating rift tips and with similar features
which appearto be their fossil equivalentsalong the inner
o
pseudofault. The outer pseudofaultsappearas linear bandsof
positive and negative free-air anomalies with a relief of-30
mGal (Figures 18 and 19). This is particularly clear for the
111øE propagatingrift which is well exposedon the southern
ridge flank in our survey(Figures4 and 18). The crustcreated
95 •
9½
9T
at retreating rifts appearsto be uplifted by 200-400 m along
the outer pseudofaults,which are marked by scarps facing
0
2500
3000
3500
4•
4500
5000
6•0
toward the crust created at the propagatingrift. A shallow
Figure 9. Bathymetric
mapof theSEIRridgeaxisfrom95øE trough extends along the base of the scarp. The free-air
to 97øE. The bathymetwis contouredat 250-m interv•s and anomaliesat the outer pseudofaultsare largely removedby the
shading
changes
occurat 500-minterv•s. •e mapshowsthe topographic correction. The outer pseudofaults at the
ridgeaxisat thesoutheastern
endof segment
MS, •1 of seg- 112ø45'E and 114øE rifts are not apparentin the MBA map
mentsM6 and N, and the northwestern
end of segmentO. (Figure 19). There is a 10 mGal MBA high over part of the
SegmentN, whichextendsfor 55 km betweentwo tr•sforms, 111øE outerpseudofault(Figure18), but it is at the very edgeof
is theonlysegment
between
82øEand104ø•ch•acterized
by the survey and may representan artifact. The inner pseudoan axial valley morphology. The M5 ridge axis is propagat- faults appearin the MBA maps as linear-10 mGal positive
ing to the southeast.
anomalies.
.:
..:
...
...
features. The largestpositivefree-air anomalymeasuredin the
surveyarea (78 mGal) is found over a fracturezone ridge near
96øE (Figure 9) which reachesto within 700 m of the surface
and appearsto be slightly oblique to the spreadingdirection,
so that it is under compression[Sempereet al., this issue].
Free-air gravity anomaliesof 55-65 mGal were also measured
over three large seamountsat the very western end of the
survey area at 41ø53'S, 88ø09'E; 41ø45'S, 89ø16'E; and
43ø23'S, 91ø17'E (Figures 2 and 16). The largestmeasured
negative anomaly (-61 mGal) is at 49ø55'S, 106ø35'E over the
deep (4800 m) pull-apart basin within the 106.5øE transform
(Figures4 and 18). Free-air anomaliesof-50 to -60 mGal were
also measuredin three deepbasinswithin the complex116øE
transform (Figures 5 and19).
Gravity
Anomalies Over the SEIR
Axis
The ridgeaxis at segments
with an axial high,whetherEPRlike, rifted, or anomalous
(segmentR), is markedby a 10-15
mGal free-air anomalyhigh. The axial gravity high is best
developedwherethe ridge crestbathymetrichigh is well developedand appearsmagmaticallyvigorous(i.e., lessfaulted),
usuallynear the centerof segments(Figures10, 16 and 17).
The EPR-like segments,M1, M2, and P1, show a characteristic "bull's eye" patternwith MBA lows of 10 to 15 mGal cen-
tered over the most vigorousportionof the segment. The
along-axishalf-segment
MBA gradient(as definedby Lin and
Phipps Morgan [1992]) for segmentP1 is 0.24 mGal/km
(Figure 17b), which is similar to those observed at the fast
spreadingEPR [Wangand Cochran,1993, 1995].
COCHRANET AL.: SOLYI'HEAST
INDIAN RIDGEGRAVITY ANOMALIF_3
I
,
I
•
,I
,
I
,
•
I
2200
I
I
'
I
,
I
,
15,473
30
P1- 100ø
38'9'E
2or,.•
lO
I
'
I
'
I
'
I
'
I
'
o
2600
3000
3400
-30
-40
-20
-10
0
10
20
30
10
20
30
40
Kilometers
2200
2600 .
3000
3400
-30
-20
-10
0
Kilometers
I
,
I
,
I
,
,
I
,
I
,
•
'
•
'
30
P2-1
2200
'
I
'
I
'
I
'
I'
'
I
2600
"
20•
•0 E
I ' I ' 't0
3000
34OO
-40
-30
-20
-10
0
10
20
30
40
Kilometers
Figure 10. Center beam Seabeambathymetryand free-air gravity profiles acrossthe SEIR axis illustrating
observedaxial high morphologies. The location of the profiles is shownin Figures 2 and 3. Profiles have
been projectedperpendicularto the axis which is locatedat 0 km on the horizontal scale. The ridge axis was
locatedon the basisof an axial magneticanomalyhigh. The longitudeat which the profiles crossthe ridge
axis is shownon each profile along with the segmentin which it is located. The northernridge flank is to the
right. Well-developed"EPR-like" axial highs (Figures10a) are only observedimmediatelyeast of left-stepping transformsat 88øE and 100øE. Otherwiserifted axial highs (Figures 10b and 10c) are the characteristic
axial morphologyfrom 82øE to 103øE.
The two rifted high segments(M3 and P2) includedin our
detailed surveyboxes also have MBA lows over the centerof
ing all of 165-km-longsegmentS1 (--108ø40'Eto 110ø45'E)
(Figure 4). The axial valley has a relief of about 1000 m at
the segment,but theseare not as pronounced
as at the more both endsof the segmentnear the 108ø40'Enontransformoffvigoroussegments
(Figures16 and 17). The half-segment
gra- set and the 110ø45'Epropagatingrift. However,the axial valdientsfor thesetwo segments
are 0.10-0.14 mGal/kmwhichis ley shoalsrapidly away from both of thesefeaturesand has a
at the low end of the rangegenerallyobservedat segments depthof < 600 m fromabout109øEto 110ø25'E(Figures
4 and
13). The axisis markedby a free-airanomalyminimumwith
ThusvigorousEPR-likesegments
are markedby along-axis an amplitudeproportional
to the depthof the valley (Figures
MBA gradients similar to those observed at the EPR. The 13 and 18). Near the endsof the segment,the free-airminirifted high segments
havelower gradientsawayfrom major mum is about 40 mGal but decreases to a subtle 3 mGal low
ridge offsets,althoughthey are still characterized
by small near 109ø55'Ewherethe axial valleybecomesso shallowas to
MBA lows centeredover vigorous,apparentlyvolcanically be almostunrecognizable
on across-axis
profiles(Figures4
with axial highs[Wang and Cochran,1995].
active portionsof segments.
and 13).
Most of the region between 104øEand 114øEis characterIn spite of the great along-axisvariationin axial relief and
ized by a shallowaxial valley (Figures4, 5 and 11-13),includ- in free-air gravity anomalies,mantle Bouguer anomalies
15,474
COCHRANET AL.: SOUTHEASTINDIAN RIDGE GRAVITY ANOMALIES
105 ø
106 ø
107 ø
108 •
12A
-48 ø
-48 o
..,.
12B
:..
P4
12C
•49 ø
Q1
R
Q2
...........
104 ø
Zd]11111
105 •
:•-:•-:k;:*;•?•:::::•,;½•:.•,•,•:,,;;.•
•*..----•
0
......
106 ø
2500
3000
•":,.':
:.:..'.":'::'.:'
...........
: '.:.":'.'.::.:•
:• '
3500
4000
107 •
108 ø
ß': ..........
4500
5000
6000
Figure11. Bathymetric
mapof theSEIRridgeaxisfrom104øE
to 108øE.Thebathymetry
is contoured
at
250-mintervals,andshadingchanges
occurat 500-mintervals.The mapshowsthe ridgeaxisat the southeasternendof segment
P3, all of segments
P4, Q1, andQ2, andthenorthwestern
endof segment
R. The location of gravityand center-beam
bathymetry
profilesacrossthe axisfrom segments
P4, Q1, andQ2 shownin
Figure 12 (12A, 12B, and 12C) are indicated.
withinsegment
S1 showverylittle along-axis
variationaway
from the segmentboundaries. Axial MBA gravity valuesvary
by less than 3 mGal for a distanceof 110 km from 108ø50'E to
110ø14'E in spite of the fact that axial depth varies by about
450 m along this portion of the ridge (Figure 18) and the axial
morphology changes from a 600-m-deep valley to a practically flat axial profile and back to a 400-m-deepvalley (Figure
4). AxialMBAsdoincrease
by 10mGalovertheeastern
50 km
of the segmentas the axial depthincreasesby -1000 m toward
the 111øE propagatingrift. This probably reflects a decrease
in crustal thicknesstoward the propagatingrift.
A similar lack of correspondence
betweenaxial depth and
mantle Bouguer anomalies is evident in segmentS4. In this
segmentthe axial depth decreasessteadilyfrom -3900 m at the
retreating rift tip near 112ø55'E to -3000 m at 113ø37'E
(Figure5 and 19). As a result,the axial valley, which is 700 m
deep at the western end of the segment, becomes steadily
shallower until it evolves into a narrow, 150-m-high volcanic
ridgefor a few km near 113ø37'E. Thereare very steepMBA
gradientsin the immediate vicinity of the 108ø45'E retreating
rift tip and the 114øE transform, again probably reflecting
crustal thinning near those discontinuities. However, for 50
The shallowvalleymorphology
on the SEIR thusappearsto
be characterizedby a lack of correspondence
betweenaxial
depthand axial MBA valuesandby nearlyconstantaxial MBA
values away from major ridge offsets. This differs from the
pattern normally observedat ridge segmentswith either axial
highsor deep axial valleyswhich both generallyshowwelldefinedminimaassociated
with the shallowestor mostvigorous portion of the segment. The axial depth variation observedin the central portion of segmentsS1 and S4 must be
created and maintained by a non-isostatic mechanism and be
related to ridge axis dynamicssince it is not observedon the
ridge flanks.
The deep axial valley to the east of the 114øE transform
(Figure 5) is marked by large-amplitudefree-air anomalies
(Figures 14 and 19) with across-axisrelief of-60
regal. Both
the bathymetry and the free-air anomalies show a distinct
asymmetry across the axis such that the rift mountains and the
free-airgravityanomaliesoverthemare higherto the northat
both endsof the segmentand to the southin the centerof the
segment. The asymmetrydoes not appearto representan
"insidecomer"effect[Tucholke
andLin, 1994],particularly
at
the westernendof the segment,
wheretheasymmetry
is of the
km from 113ø05'E to 113ø44'E, there is less than 3 mGal vari-
opposite sense.
ation in the MBA values This region of flat MBA gravity
encompasses
700 m of relief in axial depthand the entire range
of axial morphologies.
Althoughthe 115øE segmentcloselyresemblestypical
MAR segments
on across-axis
bathymetry
andgravityprofiles
(Figure 14), along-axisdepthand gravityprofilesdo not re-
COCHRANET AL.:SO•AST
INDIANRIDGEGRAVITYANOMALIES
15,475
20
•oL•
oE
-10
2400
2800
3200
36OO
-30
-20
-10
0
10
20
30
Kilometers
30 •
:o •
•o E
2600
3000
3400
38O0
-30
-20
-10
0
10
20
30
Kilometers
20
C
Q2-105
ø33.1'E
2400
oE
-10
2800
3200
36OO
-30
-20
-10
0
10
20
30
Kilometers
Figure12. CenterbeamSeabeam
bathymetry
andfree-airgravityprofiles
across
theSEIRaxisfromsegmentsP4 (a), Q1 (b), andQ2 (c) wheretheaxisis characterized
by a shallow
valley.Thelocation
of theprofiles is shownin Figure11. Profileshavebeenprojected
perpendicular
to the axiswhichis locatedat 0 km on
thehorizontal
scale.Theridgeaxiswaslocated
onthebasisof anaxialmagnetic
anomaly
high. Thelongitudeat whichtheprofilescrosstheridgeaxisis shownon eachprofilealongwiththe segment
in whichit is
located.Thenorthern
ridgeflankis to theright. Segments
P4, Q1, andQ2 arethefirstsegments
to theeastof
the changein axial morphologywhich occursbetween103øEand 104ø30'E.
semblethose observedat the MAR [e.g., Lin et al., 1990; developedat segmentR (Figure 18). An MBA minimumis
Sempdrd
et al., 1993;Detricket al., 1995]. The axialdepth locatedover the shallowest
pointon the axis at 48ø27'S,
profileis nearlyflat in thewesternthirdof the segment
(Figure 107ø35'E.MBA gravityincreases
along-axis
by 12.5mGal
19b), so that the deepestportion of the axial valley is at witha gradient
of 0.35 mGal/kmto thewestasthedepths
114ø47'E near the center of the segmentrather than at the
boundingoffsets. Axial mantle Bougueranomaliesincrease
away from rather than toward the 114øEtransform. However,
the axial mantleBougueranomaliesdo roughlymirror the axial depthwith a local MBA high at the depthmaximumnear
114ø47'Eand a local low at the depthminimumnear 115ø06'E
plungeby 950 m over a distanceof 36 km towardthe 106.5øE
transform. The MBA axial gravityalso increases
eastward
from the depthminimumby 20.4 mGalover78 km (0.26
mGal/km)asaxialdepthsincrease
by 770 m towardthe21 km-
offset108ø45'E
ridgediscontinuity.
Thesteepaxialdepthand
MBA gradients
awayfrom107ø35'E
appear
to implya vigor(Figure 19).
ousmeltsupply
centered
at thatlocation.Thisis supported
by
The two segments,P3 at 104øE and R at 108øE, in which a thepresence
of a chainof largeseamounts
extending
away
transitionoccursbetweenan axial high and an axial valley fromthedepthminimum
onthenorthern
ridgeflank(Figure
(Figures4 and 5), both show distinctMBA bull's eye lows 4). However,off-axisridgeflankdepths
donotshowa miniwith large along-axisgravity gradients. This patternis best mumnearthe centerof the segment
but rathervarycontinu-
15,476
COCHRANET AL.:SOUTHEAST
INDIANRIDGEGRAVITYANOMALIES
30
20rO
lo E{
2600
o
300O
3400
3800
-40
-30
-20
-10
0
10
20
30
40
Kilometers
I
20
10
-10
2600
t •I •I •I •I •I
3000
3400
3800
-40
-30
-20
-10
0
10
20
30
40
10
20
30
40
Kilometers
2600
3000
3400
3800
-40
-30
-20
-10
0
Kilometers
Figure 13. CenterbeamSeabeam
bathymetry
andfree-airgravityprofilesacrossthe SEIR axisfrom segmentS1 whichis characterized
by an axialvalley. The locationof theprofilesis shownin Figure4. Profiles
havebeenprojected
perpendicular
to theaxiswhichis locatedat 0 km on thehorizontal
scale.The ridgeaxis
waslocatedon the basisof an axial magneticanomalyhigh. The longitudeat whichthe profilescrossthe
ridgeaxisis shownon eachprofilealongwiththesegment
in whichit is located.The northernridgeflankis
to the right.The 165-km-long
segment
is characterized
by an axialvalleyfor it entirelength. The valleyis
about1000m deepnearthecomplextransform/nontransform
offsetat its western
endandthepropagating
rift
at its easterntip (Figures13a and 13c). However,the axisshallowsrapidlyawayfrom its extremities,andthe
axialvalleyis lessthan600 m deepoverthemiddleof thesegment.Nearthesegment
centertheaxialvalley
is so shallowas to be almostunidentifiable(Figure 13b).
ously from -3050 m to -3200 m from west to east acrossthe
segment (Figure 7) [Ma and Cochran, 1997]. This implies
either that the strongfocusingof melt supplyin the centerof
the segmenthas occurredextremelyrecentlyor that it doesnot
result in a variation in crustal thicknesswhich is preserved
onto the ridge flank.
Both the axial depth and gravity are highly asymmetricin
segmentP3. The shallowestportion of the ridge axis is at
about 103øE, only about 10 km from the 102ø45'E transform
and the depths decreaseby less than 100 m from there to the
transform (Figure 17). The ridge axis in this area forms a
nearly flat plateau. The MBA gravity minimum is locatedat a
small (< 1 km) ridge offset near 103ø10'E, about 25 km from
the transform(Figures3 and 17). The changein gravityalong
the axis from 103ø10'E to the 102ø45'E transform is 7.5 mGal
with a gradient of-0.3 mGal/km. Axial MBA values increase
slowly eastwardby about 4 mGal from 103ø10'E to 103ø32'E
in an area where the axis is at a nearly constantdepth of
slightlyover 3000 m within a narrow,shallowvalley. East of
103ø32'E,the axial depthdropsprecipitouslyby -400 m in a
distanceof 17 km andthenmoregraduallyby another400 m to
reacha depthof 3800 m at the easternend of the segmentas
the axis evolvesinto an 1000 m deepvalley. The axial MBA
gravity anomaliesmirror the axial depth,increasingrapidly
by -10 mGal and then moreslowlyfor a total changeof-21
mGal from the minimumat 103ø10'Eto the easterntip of the
COCHRAN
ETAL.:SOLrDIEAST
INDIANRIDGEGRAVITYANOMALIES
30
T - 114 ø 56.2'E
10 •
-20
-30
2600
3000
3400
3800
4200
46OO
-40
-30
-20
-10
0
10
20
30
40
Kilometers
Figure 14. Center beam Seabeambathymetryand free-air gravity profile acrossthe SEIR axis at 114ø56.2'E
in segmentT, where the axis is characterizedby a deep axial valley. The location of the profile is shown in
Figure 5. The profile has been projectedperpendicularto the axis which is locatedat 0 km on the horizontal
scale. The northern ridge flank is to the right.
J
,
I
,
I
,
I
,
I
,
I
•
A
I
,
I
20
,
R-10V
ø36.4'E
lO o•
oE
2200
-10
2600
3000
34OO
-40
-30
-20
-10
0
10
20
30
40
Kilometers
,
I
,
I
,
•
,
I
,
I
,
I
,
I
,
20
10•
2400
'
I
'
I
'
I
'
I
'
I
'
I
'
I
'
-10
2800
3200
36OO
-40
-30
-20
-10
0
10
20
30
40
Kilometers
Figure 15. Center beam Seabeambathymetryand free-air gravity profiles acrossthe SEIR axis from segment R. The locationof the profilesis shownin Figure4. Profileshave been projectedperpendicularto the
axis which is locatedat 0 km on the horizontalscale. The ridge axis was locatedon the basisof an axial magnetic anomalyhigh. The longitudeat which the profilescrossthe ridge axis is shownon each profile. The
northernridge flank is to the right. (a) Centerof the segmentwhere the axis is markedby a small volcanic
ridge atopa broadfault-bounded
plateau. (b) Easternend of the segmentwherethe plateauhasdied away; the
axis is markedby a small valley with a pronouncedgravity minimum.
15,477
15,478
COCHRANET AL.:SOUTHEAST
INDIANRIDGEGRAVITYANOMALIES
88 ø
90'
89 ø
91 ø
in the west near 90øE to -3300 m in the east near 115øE [Ma
92'
FAA
•.":•;•....:
_42 •
ß""4,:.::!?
.-'";'3
.:?.!'?
:.•
-43 •
-43 ø
..:'7::.,,•.
......'. .
:}•L•;:.;:';';'""
-:-...:.
.....
.-•:'. •...
:..,.
*•,'a,•
Gravity Anomalies and Changes in Crustal
Thickness and/or Mantle Temperature
':: ..:1( "½
'4.:...-.':
-!
.........
.,
88 ø
90ø
89'
-lOO
-20
88'
-lO
89 ø
0
91ø
10
90'
20
and Cochran, 1997] (Figure 7). The axial morphologyvaries
from a 200- to 400-m-high axial high west of 103øE (Figures
2, 3 and 10) to a 1200 m-deep valley in segmentT [Sempereet
al., this issue] (Figures 5 and 14), accountingfor the remainder
of the 2100 m change in axial depth. The long-wavelength
variation in ridge flank depth records the change in mantle
temperature and melt supply (crustal thickness), while the
variation in the form and amplitude of the axial relief reflects
the responseof the ridge axis dynamics to these changesin
temperatureand melt supply.
92'
The regional variation in both ridge depth and MBA gravity
along the SEIR resultsfrom a combinationof crustalthickness
variations and changesin mantle temperature. If the 500-m
depth variation is assumedto result only from a change in
crustal thickness, simple isostatic considerations put the
change in crustal thickness at 1.8-2.4 km depending on the
30
91'
92 ø
MBA
-42'
assumed
crustaldensity(in therangeof 2700-2900
kg/m3).If
a 2-km decrease in crustal thickness is assumed, it would result
-43 ø
-43 ø
88 ø
89'
90 ø
91 ø
92 ø
::....•
:.:.::.•:•,•
.a
..........
*.........
.-••ii• ::'::•::''::":,.,i•:•i
:.i::
•
-100
-90
-80
-70
-60
-50
-40
100
in a 50-mGal increasein MBA gravity. This is approximately
the change in the level of the MBA gravity along the SEIR
from 88øE to 115øE(Figure 7).
The depth change can also be assumedto result from a
changein mantle density arising from a west-to east temperature gradientin the mantle with no changein crustalthickness.
The magnitude of the density change necessaryto create the
500 m differencein ridge depthdependson the depthto which
it is assumedto extend. If a compensationdepth of 125 km is
assumed[Parsonsand Sclater, 1977], then the requiredchange
in the averagemantledensityis closeto 10 kg/m3. A correspondingtemperaturedifferencecan be determinedas
Figure 16a. Free-air anomaly(FAA) and mantle Bouguer
anomaly (MBA) gravity maps of survey Box 0 (Figure 2).
Contour interval is 5 mGal, and shadingchangesat 10-mGal
intervals.
Area from -89ø10'E
to --89ø40'E
is masked out be-
causethe ship trackswere parallelto and a few kilometersfrom
the ridge axis (Figure 2) and thusdo not recordthe full amplitude of the axial anomaly.
AT= Ap/etpo
(1)
wherePoisa reference
mantle
density
andotis thevolumetric
coefficientof thermalexpansion.This expression
predictsa temperaturedifferenceof 90øC-100øCdependingon the exact
choiceof otandPo' If thetemperature
variationis assumed
to
extend to 200 km [Klein and Langmuir, 1987], then the den-
sityanomaly
is -6.2 kg/m
3 andthecorresponding
temperature
segmentnear 104ø18'E (Figure 17). As with segmentR, the
ridge flank depthsdo not demonstratethe large variationsseen
in the axial depthand in the axial MBA profile.
Discussion
Ridge Parallel
Changes in Depth
The -2100 m changein axial depth along the SEIR (Figure
7) consistsof two components;a regionalchangein the depth
of the ridge, and a changein the form and amplitudeof the
axial morphologywhich is superimposedon the ridge crest.
Ma and Cochran[ 1997] usedthe empiricalorthogonalfunction
(EOF) technique[Small, 1994] to decompose
the morphology
into a "deterministic"componentrepresentingthe basic shape
of the ridge axis and a "stochastic"componentrepresenting
local tectonic variability. Off-axis ridge-parallelprofiles of
the deterministiccomponenton both flanks show a regional
depthincreaseof 500 m acrossthe surveyarea, from -2800 m
differenceis -55øC. Thesetwo casespredictincreasesof 49
mGal and 48 mGal, respectively,in MBA gravity between
88øE and 114øE.
These simple calculationsvaried only crustalthicknessor
mantletemperature. They thusprovideupperboundson the
regionalvariationin crustalthicknessand temperaturesince,
in reality, changesin mantletemperatureand crustalthickness
(melt supply)are linked andboth occursimultaneously
across
the region. The fact that the long-wavelengthvariation in
MBA gravity is the same for the two end-membercalculations
meansthat the gravity data alone cannotdistinguishthe relative contributions of variations in crustal thickness and man-
tle temperature. The reasonwhy thesedifferent mechanisms
predictvirtually the samechangein MBA gravity is that the
sameanomalousmassis neededin eachcaseto isostatically
supportthe 500-m depth difference. The assumedvariation in
crustaland/or mantle structureis very long wavelengthand
therefore the gravity effect approachesthat of an infinite
"Bouguerslab"wherethe gravityeffectdependson the anomalous mass but not on distance from the source.
COCHRANET AL.' SOLJTt•ASTINDIAN RIDGEGRAVrI• ANOMALIES
89
-45
,
,
I
90
,
,
,
,
,
I
91
,
,
,
,
,
I
92
,
,
,
,
,
-45
-50
-50
-55
-55
-70
-70
-75
i
M1
i
'
M2
-75
i
M3
-2000
-2000
-2500
-2500
-3000
-3000
-35OO
15,479
,
89
,
90
'.
,
91
......
3500
92
DegreesEast
Figure 16b. (bottom) Along-axis depth and (top) along-axis mantle Bouguer anomaly profiles from
88ø41'E to 91ø51'E (box 0). Segmentsare identifiedbetweenthe two panels. The regionis boundedby a 70km-offset transformon the west and by a 30-km-offset eastwardpropagatingOSC on the east. The axis between 88ø50'E and 89ø40'E has an EPR-type axial high. A rifted high is found throughoutthe remainderof the
area (Figure 10b).
Gravity Anomalies and
in Axial Morphology
It is reasonable
Transitions
to assume that the lateral
variation
in as-
thenospherictemperatureis continuousand smooth. However,
the variation in the MBA gravity is not smoothand continuous (Figure 7). SegmentscaleMBA anomaliesare relatedto
the upwelling and melt distributionpattern within individual
segments [Macdonald, 1989; Wang and Cochran, 1993;
Tolstoyet al., 1993]. There are also a numberof distinctsteps
in the regional level of the MBA gravity. Although there is a
sectionof the ridge between92øE and 100øEwith insufficient
crossingsof the axis to confidently determine axial gravity
values except at a few locations(Figure 7), the axial MBA
values show a fairly uniform regional slope from 89øE to
103.5øE,increasingfrom -60 mGal to -40 mGal (Figure 7). A
10-mGal step in the gravity occurs from 103.5øE to 104øE
within segmentP3 (Figures7 and 17). From 104øEto 114øE,
there is again a steady,more gradualslopeto the east (albeit
with a large intrasegmentnegative anomaly within segment
R). The regionallevel increasesto the east from -30 mGal at
104øEto-20 mGal at 114øE. Another steepregionalMBA gradient occurs across and to the east of the 114øE transform.
MBA gravity increasesfrom -20 mGal to +5 mGal acrosssegmentT (Figures7 and 19).
The two areas of steep MBA gradients at 103ø30'E and
114øE correspondto basic changesin the form of the axial
morphology. The MBA gradient at 103ø30'Ecorrespondsto
the transitionfrom an axial high to a shallowaxial valley and
the gradientat 114øEcorresponds
to the transitionfrom a shal-
low axial valley to a deepMAR-like axial valley. The different
forms of axial morphologythus reflect basic changesin the
structureof the crust and upper mantle which are significant
enoughto create a 10-mGal differencein MBA gravity. The
step in the MBA level near 103ø30'E does not result from a
changein crustal thickness,since there is not an accompanying change in ridge flank depths [Ma and Cochran, 1997]
(Figure 7) as requiredby isostasy.
The presence of a basic change in ridge axis tectonics
accompanyingthe changesin morphology is also suggested
by analysisof the abyssalhills createdat segmentswith differing axial morphology[Ma et al., 1995; Goff et al., 1997; Ma
and Cochran, 1997]. In particularMa and Cochran [1997]
found that bathymetricroughnessdeterminedfrom the stochastic componentof their EOF analysisincreasesby more than a
factor of 2 from 88øE to 116øE (Figure 20). However, the increaseis not regular but occursin two discretestepsacrossthe
transformsat 102ø45'E and 114øE (Figure 20). Segmentsto
the west of 102ø45'Ewith an axial high, whetherrifted or EPRlike, form one group. Segmentslocatedbetween102ø45'Eand
114øE form a secondgroup. This region includesthe transitional segmentP3, the "anomaloushigh" segmentR, and
segmentsS1 and S4 characterizedby a shallow axial valley.
SegmentT eastof the 114øEtransformwith a deepaxial valley
forms a third distinctgroup. The populationof bathymetric
roughnessestimatesat all segmentswithin eachregion are statistically identical, but they differ significantly between
regions with a confidence level of >99% [Ma and Cochran,
1997]. Similar resultsare observedin a covarianceanalysisof
flank morphology[Goff et al., 1997]. The step-likeincreases
15,480
COCHRAN ET AL.: SOUTHEASTINDIAN RIDGE GRAVITY ANOMALIES
100ø
101ø
102ø
1.03ø
104'
105'
.................
.....
_47
ø
FAA
'•:...i:'•
.•::'!:):
'.......
??
.....
;:•':•?•
........
.
_47 '
•'•:'•i•-::;:
'•!•:i::'::•
.............
•'"11•:':•'•'•::!I!•':•
-48 o
:.:..•37t•::•:.•:•
•ß•.':
.....
::%:':::'•544
•-•"-.
.......
100ø
- 100
101'
-20
102'
- 10
0
103•
10
102 ø
20
..::
....":.-.,:,-
104ø
30
103 ø
105'
100
104 ø
105'
MBA
-48 ø
100'
-!00
101 ø
-60
102ø
-50
-40
103ø
-30
-20
..............................
:':':•1
104ø
-10
105ø
100
Figure17a. Free-airanomaly
(FAA)andmantleBouguer
anomaly
(MBA)gravitymapsof surveyBox 1
(Figure3). Contourintervalis 5 mGal,andshadingchanges
at 10-mGalintervals.
in bathymetricroughnesscorrespondto the changesobserved tendsto encourageformationof an even shallowermelt lens.
in the form of the axial morphologyand to the steps observed Below somecritical melt supply(crustalthickness),a crustal
melt lens can not be maintained and this mechanism can not
in the level of the MBA gravity.
function.Phipps Morgan and Chen [1993] calculatethat this
Model for the Transitions in Morphology
canlead to a majorchangein the thermalstructureandthatthe
An axial high is consideredto resultfrom the upwardflexure depthto the soliduscan increaseby severalkilometersas the
of weak axial lithosphere by low density asthenosphere result of a small change in crustal thickness across this
beneaththe axis [Madsenet al., 1984; Kuo et al., 1986; Wang
and Cochran, 1993]. The presenceof high temperatures
and a
thin lithosphericlid at axial highs is supportedby the common observationof a crustalmagmachamberat ridgeswith an
axial high [Detrick et al., 1987; Mutter et al., 1995] and the
lack of evidencefor a magma chamberat ridgeswith an axial
valley [Purdy and Detrick, 1986; Detrick et al., 1990].
The rapid transitionfrom an axial high to an axial valley
requiresa suddenchangein the strengthof the lithosphericlid
at the axis. Phipps Morgan and Chen [1993] have arguedthat
the presenceor absenceof a crustalmagma chamberis itself a
thresholdphenomenonwhich has a large effect on the temperature structureand strengthof the crustand uppermantle. This
occurslargely due to a feedbackmechanismin which the latent
heat releasedby freezingof the lower crustwithin the melt lens
threshold.
The transition
froman axialhighto an axialvalleycould
resultfromthis(ora similar)
threshold
mechanism.
Thegradient in asthenospheric
temperaturealong the SEIR and the
resulting
decrease
in meltsupply
meanthatat somepointthe
threshold
is reached
anda crustalmagmalenscannotbe main-
tained.Theresulting
sudden
increase
in lithospheric
strength
wouldsuppress
theflexureresponsible
fortheaxialhighand
alsoallowlargerfaultsto formcausing
theobserved
stepin
the bathymetric
roughness
(Figure20) andthe formationof a
shallowvalley(Figures3, 4 and11). The change
in thethermal structureneeds to be fairly substantial,however, to
accountfor the observedstepin the MBA gravity. If the 10mGal step in MBA gravity (Figures7 and 17) is due to an
increasein densityof a region5 km thick,thenthe average
COCHRANET AL.' SOUTHEASTINDIAN RIDGEGRAVITY ANOMALIES
101
102
103
15,481
104
105
-20
-20
-25
-25
-30
-30
-45
-45
-50
-50
P1
P3
P2
,
,
I
,
,
,
P4
,
,
I
,
,
,
,
,
2500
2500
3000
3000
3500
3500
.
101
102
103
104
,
,
105
DegreesEast
Figure 17b. (bottom)Along-axisdepthand (top) along-axismantleBougueranomalyprofilesfrom
100ø19'Eto 105øE(box 1). Segments
areidentifiedbetweenthetwo panels.The regionis bounded
on the
westby a 130-kmoffsettransform
(Figure3). Segments
P1 andP2, westof a 38-km-offset
transformnear
102ø45'Ehaveaxial highs(Figures10a and 10c). SegmentP3 from 102ø53'Eto 104ø18'Emarksthe basic
transitionfrom an axial high to the westto an axial valley to the east. SegmentP4 eastof 104ø20'Ehasa
shallow axial valley.
temperaturein that region would have to decreaseby -400øK.
Note that this large change in shallow temperature structure
must result from a very small change in asthenospheric
temperatureand melt supply.
The transition from "EPR-type" axial high to rifted high
appears to be gradational based on their similar ridge flank
roughness(Figure 20), gradation in intrasegment form of
axial MBA gravity (Figures 16 and 17) and the observation
the rigid lid from shearstressesarisingfrom divergentviscous
114øE.
1993].
mantleflow away from the axis. When the ductileregionin
the lower crust decreasesbelow some critical size the upper
crust becomescoupled to the mantle flow resultingin brittle
failure and the formation of a deeprift valley
On the other hand, compilations of axial relief [e.g.,
Macdonald, 1986; Malinverno, 1993; Small, 1994] do not
show a sharp break between shallow and deep rift valleys.
that the EPR-type segmentsbecomerifted at segmentends. if Rather axial relief appearsto increase continuouslywith dea quasi-steadystate magma chamberis presentat rifted highs, creasing spreading rate. It is thus possible that the sharp
but is deeper and/or attenuatedcomparedto EPR-type highs break from a shallow axial valley to a deep axial valley at
then the rigid lid will be thicker and stronger reducing the 114øEcould be the resultof a rapid changein mantletemperaupward flexure of the high and allowing faulting to begin near ture acrossthe westernboundaryof the AAD similar to that obthe axis.
servedacrossthe easternboundarywhere there is a very rapid
Our observationsalso suggest that a second sudden transi- change from an axial high to a deep axial valley acrossthe
tion occursfrom a shallow rift valley to a deep rift valley near easternboundarytransformof the AAD at 127øE [Palmer et al.,
This transition
could also result from a threshold
effect
in which a small changein mantle temperaturecausesa sudden Conclusions
change in axial tectonicsresulting in the formation of a deep
rift valley. For example, Chen and Morgan [1990] have pro1. Axial mantleBouguergravity anomaliesalongthe SEIR
posed a mechanism for the creation of a deep axial valley show a pronouncedregionalalong-axisgradient. Axial MBAs
which dependson ability of the ductile lower crustto decouple increase from --60 mGal near 88øE to --10 mGal at 115øE and
15,482
COCHRANET AL.: SOUTHEASTINDIAN RIDGE GRAVITY ANOMALIES
1• ø
107'
108'
110'
111'
MBA
ß
_4,-
:.
.:..
-49 ø .,::
i
106 ø
1
107'
108'
107'
108'
Figure18a. Free-airanomaly
(FAA)andmantleBouguer
anomaly
(MBA)gravitymapsof surveybox2
(Figure4). Contourintervalis 5 mGal,andshading
changes
at 10-mGalintervals.
to +5 mGalat the 116øEtransform.
TheMBA gravitygradient 88øEto 114øE. In reality,changes
in melt supply(crustal
is accompanied
by a 500 m increase
in thedepthof theridge thickness)andtemperatureare linkedandwill occursimultaneflanksfrom-2800 m at 88øEto -3300 m at 114øE(Figure7). ouslyso thattheseestimates
serveas upperboundson along-
The regionalchangesin depthand gravitycan both be ex-
axis changes.
plained by either a 1.7-2.4 km west to east decreasein crustal
2. The along-axisMBA gravitygradientis not smoothand
thickness
or a 55øC-100øC(depending
on thedepthto which uniform.Distinctstepsin thelevelof theMBA gravityoccurs
it extends)westto eastdecrease
in mantletemperature
from at two locations, 103ø30'E-104øE and 114øE-116øE. The two
107
108
i
-15
109
,
,
,
,
,
i
110
,
,
,
,
,
i
111
,
,
,
,
,
-15
-20
-20
-25
-25
-40
-40
-45
-45
I
-3000
-3500
-3500
-4000
107
108
109
110
111
DegreesEast
Figure18b. (bottom)
Along-axis
depthand.(top)along-axis
mantle
Bouguer
anomaly
profiles
from
107ø06'E
to 110ø49'E
(box2). Segments
areidentified
between
thetwopanels.
Theridgeaxisforthecentral
50kmofsegment
R from-107ø20'E
to108øE
hasa blocky
axialhigh(Figure
15a).Thehighdiesoutintoan
axialvalley
toward
thebounding
transforms.
Segment
S1 hasa shallow
axialvalley
which
deepens
westof
109ø15'E
toward
a complex
combination
transform-nontransform
offset
andeast
of 110ø15'E
toward
apropa-
gatingrift tip at 111øE(Figure13).
COCHRAN ET AL.: SOUTHEASTINDIAN RIDGE GRAV1TYANOMALIES
112 ø
113 ø
114 ø
115 ø
116ø
117 ø
_49 ø
_49 ø
113 ø
112 ø
114 ø
115 ø
...
112ø
-lOO
113ø
114ø
I
i
-40
-30
115ø
116ø
ß
.............:.4'
..--.'•1:.•:,:•"-'-"'•
'•"
"••••"-'•-'•-•Y•'
'
•,•1 •
i
i
-20
-10
0
10
_50ø
_51 ø
_51 ø
I
20
1•
11T
"::*:•'
'::ii:::•i•
_51 ø
112 ø
117 ø
'1
116 •
-49'
MBA ?.......
_49 ø
_50 •
15,483
113 •
114 ø
115 ø
-100
-30
-20
116 ø
117 ø
•:•'"'•
..............................
•
:•:
......
•:::•
-10
0
1.0
20
100
Figure 19a. Free-airanomaly(FAA) and mantleBougueranomaly(MBA) gravitymapsof surveybox 3
(Figure 5). Contourintervalis 5 mGal, and shadingchangesat 10 mGal intervals.
acterized by shallow axial valleys and the 114øE transform
marks the change to a deep MAR-like axial valley. These
al., 1997; Ma and Cochran, 1997] (Figure 20) and correspond observationsimply that the different types of axial morpholto fundamentalchangesin the characterof the axial morphol- ogy do not simply grade into each other. Rather they repreogy. The ridge axis to the west of 103øEis markedby an axial sent different forms of axial structure and tectonics with a
high. Segmentsbetween 104øE and 114øEare generally char- sharp transitionbetween distinct modes of crustal accretion.
stepsin the level of the axial MBA gravity anomaliescoincide
with sharpstepsin ridgeflankbathymetric
roughness
[Goffet
113
114
116
115
i
-
5
0
-5 •
-5
-10•
-15
-20
-20
-25
-25
3O00
,
,
I
I
,
,
,
,
,
I
,
,
,
,
30O0
35OO
3500
4000
4000
45O0
4500
'
'
i
113
.....
i
.....
114
i
115
116
DegreesEast
Figure 19b. (bottom) Along-axis depth and (top) along-axis mantle Bouguer anomaly profiles from
112ø51'E to 115ø29'E (box 3). Segmentsare identified betweenthe two panels. SegmentS4 extendsfrom a
retreatingrift tip near 112ø50'Eto a propagating
rift tip whichhasrecentlycollidedwith the 114øEtransform
The axis shallowsfrom steadily from a 1000-m-deepvalley at the dying rift tip to a narrow 100-m-high volcanic ridge near 113ø50'E before deepeningslightly into a shallowvalley at its easternend. SegmentT is
characterizedby a deep (-1200 m) axial valley for its entirelength.
15,484
COCHRANET AL.: SOtYII-IEASTINDIAN RIDGEGRAVITY ANOMALrHS
i
_
o
o
o
o
o
o
o
o
o
o
o
oJ
o
o
o
co
o
(o
o
•1-
!
(LU)qldea le!X¾
(LU)sseuq6nobl
o
oJ
o
COCHRANET AL.: SOUTHEASTINDIAN RIDGEGRAVITYANOMALIES
The transition from an axial high to a shallow valley requires a rapid increasein the strength(thickness)of the lithosphericlid, both to suppressthe upward flexure responsiblefor
the axial high [Madsen et al., 1984; Kuo et al., 1986; Wang
and Cochran, 1993] and to allow the formationof largerthrow
faults. A threshold mechanism such as that proposed by
Phipps Morgan and Chen [1993] is necessaryto explain the
rapid, coincident changes in axial morphology, ridge flank
roughnessand axial MBA gravity. Phipps Morgan and Chen
[1993] argue that there is a critical melt supply (crustal thickness) below which a crustal melt lens cannot be maintained and
that the depth to the soliduscan increaseby several kilometers
as the result of a small decrease in crustal thickness across this
threshold. This changein thermal structurecould result in the
necessaryincrease in axial lithosphericstrength. The change
in thermal structure must be significant to account for the 10mGal stepin the MBA gravity. If the changein gravity results
from higher densitiesin the upper5 km of the mantle,then the
averagetemperaturein that region must decreaseby 400øK.
The transition from a shallow axial valley to a deep axial
valley at 114øE could result from a second threshold phenomenon,such as that proposedby Chen and Morgan [1990].
However, it might also resultfrom a very rapid changein mantle temperatureat the westernboundaryof the AAD, similar to
the change observed at the eastern boundary [Klein et al.,
1991; Sempereet al. , 1991; Palmer et al. , 1993].
3. The different forms of axial morphologyalong the SEIR
are associatedwith different intrasegmentMBA gravity pat-
terns.Segments
withanaxialhighhaveMBAlowslocated
at
a depth minimum near the center of the segment. Surveyed
segmentswith an EPR-like axial high have bull's eye MBA
anomalieswith 10-15 mGal lows and along-axis gradientsof
0.15-0.25 mGal/km, similar to what is observed at the EPR
[Wang and Cochran, 1993]. Segmentswith a rifted axial high
have shallower gravity minima and lower gradients. The gravity anomalies at rifted highs as well as the presenceof pervasive faulting acrossthese axial highs can be explained as the
result of a deeperor attenuatedmelt lens and compositemagma
chamber [Sinton and Detrick, 1992] in the lower crust and
upper mantle as a the result of a reducedand perhapsepisodic
melt supply.
Segmentswith a shallow axial valley are characterizedby a
lack of correspondencebetween axial depth and axial MBA
values and have a very flat (<3 mGal variation) along-axis
MBA profile away from segmentboundingridge offsets. The
exception to this pattern is segment R where the axis along
the central 50 km of the segment is a volcanic ridge at the
summitof a 20 km wide fault-boundedplateau. This segmentis
characterizedby very large along-axis changesin axial depth
(950 m) and in MBA gravity (20 mGal) with an MBA gradient
of 0.35 mGal/km. Both of these characteristicsindicate very
vigorous upwelling of melt beneath the segment. However,
this segmenthas ridge flank abyssalhill amplitudescharacteristic of segmentswith a shallow axial valley [Goff et al.,
1997; Ma and Cochran, 1997] (Figures7 and 20). It may represent the interaction of a ridge axis having a temperature
structurewhich would normally producea shallow axial valley
with a melt anomaly.
Only one segmentwith a deep MAR-like rift valley was
includedin the surveyarea (segmentT, near 115øEat the eastern end of the survey region). SegmentT does not however
have along-axis depth and gravity profiles characteristicof
segmentson the MAR. An 80 km-long segmentsuch as seg-
15,485
mentT wouldbe expectedto have-1400 m of along-axisdepth
relief associatedwith a 30 mGal MBA gravity low, according
to relationshipsdevelopedat the MAR [Lin et al., 1990;
Detrick et al., 1995]. Instead, it has only -450 m of along-
axis depth relief and no real intrasegmentMBA minimum
(Figure19). However,thissegment
is locatedin a regionwith
a very steepregionalMBA gradient (Figures7 and 19) and
probablymarksthe effectivewesternboundaryof the AAD.
The depthandgravitypatternat the segmentmaybe perturbed
by a rapid changein asthenospheric
temperatureacrossthe
boundaryof the AAD. We have insufficientdatato determine
whethersegmentsfartherto the eastwithin the AAD have a
gravitypatternmorecharacteristic
of the MAR.
Acknowledgments.We thankCaptainE. Buckandthe crewof the
R/V Melville for their assistance
duringthe field programwhich led to
the collectionof highqualitydataduringoftendifficultweatherandsea
conditions.We alsothankS. Smith,U. Albright,T. Porteus,E. Heckman
andJ. Skinnerof ScrippsInstitutionof Oceanography
for technicalhelp
duringcruisesWEST09MV andWEST10MV. The GMT [Wesseland
Smith, 1995] and MBsystem(D. Chayesand D. Caress,pers.comm.)
softwarepackageswereextensivelyusedin the preparation
of this paper. Carefulreviewsof the manuscript
by Jill Karsten,GregNeumann
and Paul Wessel were extremelyhelpful. This work was fundedby
National ScienceFoundationgrantsOCE-93-02091 (JRC) and OCE-9403583 (JCS). Lamont-Dohertycontributionnumber5638.
References
Bell, R. E., and A. B. Watts, Evaluation of the BGM-3
sea
gravity meter systemonboardR/V Conrad, Geophysics,51,
1480-1493, 1986.
Cande, S.C., and J. C. Mutter, A revised identification of the
oldest seafloor spreading anomaly between Australia and
Antarctica, Earth Planet. Sci. Lett., 58, 151-161, 1982.
Chen, Y. J., and W. J. Morgan, Rift valley/no rift valley transition at mid-ocean ridges, J. Geophys.Res., 95, 17,57117,581,
1990.
Cochran, J. R., Systematic variation of axial morphology
along the Southeast Indian Ridge (abstract), Eos Trans.
AGU, 72, 260, 1991.
DeMets, C., R. G. Gordon, D. F. Argus, and S. Stein, Current
plate motions, Geophys.J. Int., 101, 425-478, 1990.
Detrick, R. S., P. Buhl, E. E. Vera, J. C. Mutter, J. A. Orcutt, J.
A. Madsen, and T. M. Brocher, Multichannel seismicimaging of the axial magma chamberalong the East Pacific Rise
between 9øN and 13øN, Nature, 326, 35-41, 1987.
Detrick, R. S., J. C. Mutter, P. Buhl, and I. I. Kim, No evidence
from multichannelreflection data for a crustalmagma chamber in the MARK area on the Mid-Atlantic Ridge, Nature,
347, 61-64,
1990.
Detrick, R. S., H. D. Needham, and V. Renard, Gravity anomalies and crustal thicknessvariations along the Mid-Atlantic
Ridge between 33øN and 40øN, J. Geophys. Res., 100,
3767-3787,
1995.
Forsyth, D. W., R. L. Ehrenbard, and S. Chapin, Anomalous
upper mantle beneaththe Australian-AntarcticDiscordance,
Earth Planet. Sci. Lett., 84, 471-478, 1987.
Goff, J. A., Y. Ma, A. Shah,J. R. Cochran,and J.-C. Sempere,
Stochasticanalysis of seafloor morphologyon the flanks of
the SoutheastIndian Ridge: The influence of ridge morphology on the formation of abyssal hills, J. Geophys.
Res., in press, 1997.
Heezen, B.C.,
110, 1960.
The rift in the ocean floor, Sci. Am., 203, 99-
15,486
COCHRAN
ETAL.:SOIJFHEAST
INDIANRIDGEGRAVITYANOMALIES
Johnson, K. T. M., D. S. Scheirer, D. W. Forsyth, and D. W.
Graham,Cruisereport: Boomerangexpedition,Leg 06, R/V
Melville, 48 pp., Bishop Mus., Honolulu, Hawaii, 1996.
Klein, E. M., and C. H. Langmuir, Global correlationsof ocean
ridge basalt chemistry with axial depth and crustal thickness,J. Geophys.Res., 92, 8089-8115, 1987.
Klein, E. M., C. H. Langmuir, and H. Staudigel,Geochemistry
of basaltsfrom the southeastIndian Ridge, 115øE-138øE, J.
Geophys.Res., 96, 2089-2108, 1991.
Kuo, B. Y., D. Forsyth, and E. M. Parmentier, Flexure and
thickening of the lithosphere at the East Pacific Rise,
Geophys.Res. Lett., 13, 681-684, 1986.
Lin, J., and J. Phipps Morgan, The spreadingrate dependence
of three-dimensional mid-ocean ridge gravity structure,
Geophys.Res. Lett., 19, 13-16, 1992.
Lin, J., G. M. Purdy, H. Shouten,J. C. Sempere,and C. Zervas,
Evidence from gravity data for focusedmagmatic accretion
along the Mid-Atlantic Ridge, Nature, 344, 627-632, 1990.
Ma, Y., and J. R. Cochran, Transitions in axial morphology
along the Southeast Indian Ridge, J. Geophys. Res., 101,
15,849-15,866,
1996.
Ma, Y., and J. R. Cochran, Bathymetric roughness of the
SoutheastIndian Ridge: Implicationsfor crustalaccretionat
intermediate spreadingrate mid-ocean ridges, J. Geophys.
Res., in press, 1997.
Ma, Y., J. R. Cochran, C. Small, and J.-C. Sempere, The
SoutheastIndian ridge between88øE and 120øE: Ridge flank
morphologyand abyssalhills (abstract),Eos Trans. AGU,
76, F529, 1995.
Macdonald,K. C., The crestof the Mid-Atlantic Ridge: Models
for crustal generation processes and tectonics, in The
Geology of North America, vol. M, The Western North
Atlantic Region, edited by P. Vogt and B. Tucholke,pp. 5168, Geol. Soc. of Am., Boulder, Colo., 1986.
Macdonald, K. C., Tectonic and magmatic processeson the
East Pacific Rise, in The Geology of North America, vol. N,
The Eastern Pacific Ocean and Hawaii, edited by E. L.
Winterer, D. M. Hussong and R. W. Decker, pp. 93-110,
Geol. Soc. of Am., Boulder, Colo., 1989.
Macdonald, K. C., P. J. Fox, L. J. Perram, M. F. Eisen, R. M.
Haymon, S. P. Miller, S. M. Carbotte, M. H. Cormier, and
A. N. Shor, A new view of the mid-ocean ridge from the
behaviour of ridge-axis discontinuities,Nature, 335, 217225, 1988.
Macdonald, K. C., D. S. Scheirer, and S. M. Carbotte, Midocean ridges: Discontinuities, segmentsand giant cracks,
Science, 253, 986-994,
1991.
Madsen, J. A., D. W. Forsyth, and R. S. Detrick, A new isostatic model for the East Pacific Rise crest, J. Geophys.
Res., 89, 9997-10,016,
1984.
Malinverno, A., Transition between a valley and a high at the
axis of the mid-ocean ridges, Geology, 21, 639-642, 1993.
Menard, H. W., The East Pacific Rise, Science, 132, 17371746, 1960.
Mutter, J. C., S. M. Carbotte, W. Su, L. Xu, P. Buhl, R. S.
Detrick, G. A. Kent, J. A. Orcutt, and A. J. Harding, Seismic
images of active magma systemsbeneath the East Pacific
Rise between 17o05' and 17ø35'S, Science, 268, 391-395,
1995.
Palmer, J., J. C. Sempere, D. M. Christie, and J. Phipps
Morgan, Morphology and tectonics of the AustralianAntarctic
Discordance between 123øE and 128øE, Mar.
Geophys.Res., 15, 121-152, 1993.
Parker, R. L., The rapid calculationof potential anomalies,
Geophys.J. R. Astron.Soc.,31, 447-455, 1973.
Parsons,B., and J. G. Sclater,Ocean floor bathymetryand heat
flow, J. Geophys.Res., 82, 803-827, 1977.
PhippsMorgan,J., andY. J. Chen,Dependence
of ridge-axis
morphologyon magmasupplyand spreadingrate, Nature,
364, 706-708,
1993.
PhippsMorgan,J., E. M. Parmentier,andJ. Lin, Mechanisms
for the origin of mid-ocean ridge axial topography:
Implicationsfor the thermaland mechanicalstructureof accreting plate boundaries,J. Geophys.Res., 92, 12,82312,836,
1987.
Prince, R. A., and D. W. Forsyth, Horizontal extent of anomalously thin crustnear the Verna FractureZone from the threedimensional analysis of gravity anomalies, J. Geophys.
Res., 93, 8051-8063, 1988.
Purdy, G. M., and R. S. Detrick, Crustal structureof the MidAtlantic Ridge at 23øN from seismicrefraction,J. Geophys.
Res., 91, 3739-3762,
1986.
Pyle, D. G., Geochemistryof mid-ocean ridge basalt within
and surrounding the Australian-Antarctic Discordance,
Ph.D., Oregon State University, 178 pp., 1994.
Royer, J.-Y., and D. T. Sandwell, Evolution of the eastern
Indian
Ocean
since the late Cretaceous:
Constraints
from
Geosat altimetry, J. Geophys.Res., 94, 13,755-13,782,
1989.
Royer, J. Y., and R. Schlich, SoutheastIndian Ridge between
the Rodriquez triple junction and the Amsterdamand Saint
Paul Islands: Detailed kinematics for the past 20 m.y., J.
Geophys.Res., 93, 13,524-13,550, 1988.
Sempere, J. C., G. M. Purdy, and H. Schouten, Segmentation
of the Mid-Atlantic Ridge between 24øN and 30ø40'N,
Nature, 344, 427-431,
1990.
Sempere, J. C., J. Palmer, D. M. Christie, J. Phipps Morgan,
and A. N. Shor, Australian-Antarcticdiscordance,Geology,
19, 429-432,
1991.
Sempere,J. C., J. Lin, H. S. Brown, H. Schouten,and G. M.
Purdy, Segmentationand morphologic variations along a
slow-spreadingcenter: The Mid-Atlantic Ridge (24ø00'N30ø40'N), Mar. Geophys.Res., 15, 153-200, 1993.
Sempere, J. C., J. R. Cochran, and the SEIR Scientific Team,
The Southeast Indian Ridge between 88øE and 120øE:
Variations in crustal accretion at constantspreadingrate, J.
Geophys.Res, this issue, 1997.
Sinton, J. M., and R. S. Detrick, Mid-ocean ridge magma
chambers, J. Geophys.Res., 97, 197-216, 1992.
Small, C., A global analysisof mid-oceanridge axial topography, Geophys.J. Int., 116, 64-84, 1994.
Small, C., and D. T. Sandwell,An abrupt changein ridge axis
gravity with spreadingrate, J. Geophys.Res., 94, 17,38317,392,
1989.
Small, C., and D. T. Sandwell, An analysisof ridge axis gravity roughness and spreading rate, J. Geophys. Res., 97,
3235-3245,
1992.
Small, C., and D. T. Sandwell,Imaging mid-oceanridge transitions with satellite gravity, Geology, 22, 123-126, 1994.
Smith, W. H. F., and P. Wessel, Gridding with continuouscurvature splines in tension, Geophysics,55, 293-305, 1990.
Tolstoy, M., A. J. Harding, and J. A. Orcutt, Crustal thickness
on the Mid-Atlantic Ridge: Bull's eye gravity anomaliesand
focused accretion, Science, 262, 726-729,
1993.
Tolstoy, M., A. J. Harding, J. A. Orcutt, and J. Phipps
Morgan, Crustal thickness at the Australian Antarctic
COCHRANET AL.: SOUTHEASTINDIAN RIDGE GRAVITY ANOMALIES
15,487
Discordance and neighboring SoutheastINdian Ridge, Eos
Antarctic Discordance, Earth Planet. Sci. Lett., 128, 135-
Trans. AGU, 76(46), F570, 1995.
153, 1994.
Tucholke, B. E., and J. Lin, A geologicmodel for the structure
of ridge segments in slow spreading ocean crust, J.
Geophys.Res., 99, 11,937-11,958, 1994.
Veevers, J. J., Breakup of Australia and Antarcticaestimatedas
mid-Cretaceous(95+5 Ma) from magneticand seismicdata at
the continentalmargin,Earth Planet.$ci. Lett., 77, 91-99,
1986.
Providence, RI 02912
L. Geli, D6partment de G6osciencesMarines, IFREMER, B.P. 70,
Wang,X., andJ. R. Cochran,
Gravityanomalies,
isostasy
and
mantleflow at theEastPacificRisecrest,J. Geophys.
Res.,
98, 19,505-19,531,
D. Christieand B. Sylvander,Collegeof Oceanicand Atmospheric
Sciences,OregonStateUniversity,Corvallis,OR 97331-5503
J. R. Cochran,Y. Ma, C. Small, and W. Zhang, Lamont-Doherty
EarthObservatory
of ColumbiaUniversity,Palisades,
NY 10964
M. Eberle, Departmentof GeologicalScience,Brown University,
1993.
Wang,X., andJ. R. Cochran,Along-axisgravitygradients
at
mid-oceanridges: Implicationsfor mantle flow and axis
morphology, Geology,23, 29-32, 1995.
Wessel, P., and W. H. F. Smith, A new versionof the Generic
29280, Plouzan6, France
J. A. Goff, Universityof TexasInstituteof Geophysics,8701 MoPac
Expressway,Austin,TX 78759
H. Kimura,EarthquakeResearchInstitute,Universityof Tokyo, 1-1I Yayoi, Bunkto-ku,Tokyo 113, Japan
J.-C. Semp6r6,ExxonProductionResearchCompany,P.O. Box 2189,
Houston, TX 77252-2189
A. Shah and B.P. West, School of Oceanography,University of
Washington,Seattle,WA 98195-7940
MappingTools (GMT), Eos Trans.AGU, 76, 329, 1995.
West,B. P., J.-C.Sempere,
D. G. Pyle,J. PhippsMorgan,and
D. M. Christie,
Evidence
forvariable
uppermantle
tempera- (ReceivedAugust23, 1996;revisedJanuary20, 1997;
ture and crustal thickness in and near the Australian-
acceptedFebruary 13, 1997.)