Extension Across The Indian-Arabian Plate Boundary: The Murray Ridge

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 17

Geophys. J. Int.

(2000) 142, 461–477

Extension across the Indian–Arabian plate boundary:


the Murray Ridge

R. A. Edwards, T. A. Minshull* and R. S. White


Department of Earth Sciences, University of Cambridge, Bullard L aboratories, Madingley Road, Cambridge, CB3 0EZ, UK.
E-mail: edwards@esc.cam.ac.uk

Accepted 2000 March 10. Received 2000 March 10; in original form 1997 December 30

SU M MA RY
Seismic reflection profiles from the Murray Ridge in the Gulf of Oman, northwest
Indian Ocean, show a significant component of extension across the predominantly

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


strike-slip Indian–Arabian plate boundary. The Murray Ridge lies along the northern
section of the plate boundary, where its trend becomes more easterly and thus allows
a component of extension. The Dalrymple Trough is a 25 km wide, steep-sided half-
graben, bounded by large faults with components of both strike-slip and normal
motion. The throw at the seabed of the main fault on the southeastern side of the half-
graben reaches 1800 m. The northwest side of the trough is delineated by a series of
smaller antithetic normal faults. Wide-angle seismic, gravity and magnetic models show
that the Murray Ridge and Dalrymple Trough are underlain by a crystalline crust up
to 17 km thick, which may be continental in origin. Any crustal thinning due to
extension is limited, and no new crust has been formed.
We favour a plate model in which the Indian–Arabian plate boundary was initially
located further west than the Owen Fracture Zone, possibly along the Oman continental
margin, and suggest that during the Oligocene–Early Miocene Indian Ocean plate
reorganization, the plate boundary moved to the site of the present Owen Fracture
Zone and that motion further west ceased. At this time, deformation began along the
Murray Ridge, with both the uplift of basement highs, and subsidence in the troughs
tilting the lowest sedimentary unit. Qalhat Seamount was formed at this time.
Subsequent sediments were deposited unconformably on the tilted lower unit and then
faulted to produce the present basement topography. The normal faulting was
accompanied by hanging-wall subsidence, footwall uplift, and erosion. Flat-lying recent
sediments show that the major vertical movements have ceased, although continuing
earthquakes show that some faulting is still active along the plate boundary.
Key words: Arabian Sea, crustal structure, extension, faulting, plate boundary.

change in direction occurs to a more easterly trend. This marks


IN TR O DU C TI O N
the southern end of the Murray Ridge. The northern end of
The Murray Ridge forms the northernmost extension of the the Murray Ridge is buried beneath the thick sediments of the
Owen Fracture Zone in the Arabian Sea, and as such comprises Indus Fan. Here the Indian–Arabian plate boundary terminates
part of the boundary between the Indian and Arabian plates at a triple junction located near Karachi. To the west, the
(Fig. 1). At the southern end of this plate boundary the seafloor Makran subduction zone marks the convergence of the Arabian
spreading centres marked by the Sheba and Carlsberg ridges and Eurasian plates, while the Indian–Eurasian plate boundary
are offset approximately 300 km along the Owen Fracture extends to the north along the strike-slip Ornach-Nal fault
Zone. Rifting along the Sheba Ridge extends into the Gulf of system.
Aden and separates the African and Arabian plates. The Owen Present-day extensional tectonics along the Murray Ridge,
Fracture Zone strikes NNE to a latitude of 22°N, where a where a series of fault-bounded grabens are offset by bathy-
metric highs, has been recognized from both seismic reflection
* Now at: School of Ocean and Earth Science, University of data (White 1984; Minshull et al. 1992) and earthquake focal
Southampton, Southampton Oceanography Centre, European Way, mechanisms (Quittmeyer & Kafka 1984). Global and local
Southampton, UK. plate circuits suggest that the Murray Ridge is undergoing

© 2000 RAS 461


462 R. A. Edwards, T . A. Minshull and R. S. W hite

40O 50O 60O 70O 80O Early Miocene (23 Ma), major changes in plate geometry
resulted in the evolution of the Sheba Ridge and opening of
the Gulf of Aden. This change in plate motions coincided with

ult
EURASIAN

Fa
al
30
O
PLATE a brief period of compression and uplift along the Owen

ch-N
30O Fracture Zone (Whitmarsh 1979).

Orna
PE
R
Two main hypotheses for the evolution of the Gulf of Oman
SI
AN
MAKR

G
AN and the Owen Basin have been proposed, and these are shown
U
LF
KARACHI
ARABIAN GULF
OF OM
AN in Fig. 2. Whitmarsh (1979) suggests that the plate boundary
O PLATE RAY
MUR GE
20 RID has been located along the Owen Fracture Zone since India

S IN
20O

NE
BA
first started to move northwards at 90 Ma. In this model, the

ZO
EN

RE
oceanic crust in the Gulf of Oman and Owen Basin must be
OW

TU
SHEBA R
AC

ID late Jurassic–Early Cretaceous in age, having been formed


FR

GE
GULF OF ADEN INDIAN
when east and west Gondwanaland first separated (Fig. 2a).
EN

O
PLATE
OW

10
10
O In the second model, proposed by Mountain & Prell (1990),
C
AFRICAN the plate boundary initially lay further west along the Oman
AR
L

continental margin, while the Mascarene Basin opened and the


SB

PLATE
ER

G RI
DG
E Carlsberg Ridge developed (Fig. 2b). During the Oligocene–
Early Miocene the plate boundary then switched to the present-

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


O O
50 60 70O
day Owen Fracture Zone when the Sheba Ridge was formed,
Figure 1. Location map of the Arabian Sea showing the schematic and motion ceased along the Oman continental margin. In
plate tectonic setting. Oceanic spreading ridges are marked by double this model the crust in the Owen Basin and Gulf of Oman
lines, other extensional plate boundaries by thick dashed lines, strike- would be Palaeocene–Eocene in age.
slip faults by single lines, ocean–continent collision by solid triangles, The data currently available in the area do not strongly
and continent–continent collision by open triangles. Adapted from
favour either one of these hypotheses. Heat-flow data from the
Jacob & Quittmeyer (1979).
Gulf of Oman (Hutchison et al. 1981) suggest that the oceanic
crust is Late Cretaceous in age. Seafloor spreading magnetic
anomalies are not observed in the Gulf of Oman, giving further
oblique extension at a very slow full rate of ~2 mm yr−1
support to a Late Cretaceous age for the oceanic crust since
(Gordon & DeMets 1989; DeMets et al. 1990). In this paper
it would have formed during the Cretaceous magnetic quiet
we present recently processed multichannel seismic (MCS)
zone. Preliminary interpretation of a wide-angle seismic line
lines across the Murray Ridge to provide better constraints
on the present-day tectonic regime and the tectonic history of across the southeast Oman continental margin by Barton et al.
the region. (1990) favours some strike-slip motion along the margin,
parallel to the Owen Fracture Zone, lending some support to
the model of Mountain & Prell (1990).
EVO LU TI O N O F TH E G UL F O F O M A N,
AR A B IA N SE A A N D O W EN BA S I N
The Gulf of Oman occupies the roughly triangular region S EI S M IC D ATA S E TS AVAI LA B LE I N TH E
bounded to the north by the Makran subduction zone, to the A R EA
southwest by the Oman continental margin, and to the south-
The principal data set in the area comprises a suite of multi-
east by the Murray Ridge and Owen Fracture Zone (Fig. 1). This
channel seismic (MCS) reflection lines acquired across the
area of the northwestern Indian Ocean has had a complicated
southern Murray Ridge on the RRS Charles Darwin (Fig. 3).
tectonic history since the break-up of Gondwanaland.
The evolution of the Indian Ocean began with the separation The data were collected using a 48-channel, 2.4 km streamer
of west and east Gondwanaland during the Mesozoic (Besse and a 1600 in3 airgun array. The airguns were fired every 20 s,
& Courtillot 1988). The continental margins of eastern Africa which, with a speed of 4.5–5.0 kt, gave a shot spacing of ~50 m.
are marked by a series of Late Jurassic–Early Cretaceous basins As a result of the calm sea conditions the data are generally
formed by the separation of India–Madagascar–Antarctica of excellent quality. The processing sequence for these data
from Africa–Arabia (Mountain & Prell 1990). At this time, included deconvolution before stack, bandpass filtering, velocity
India and Madagascar were moving south with respect to Africa analysis and normal moveout correction prior to stacking. An
(Fig. 2). This separation ceased at approximately anomaly f–k migration with a constant velocity of 1500 m s−1 was also
M0 time (102 Ma). By 90 Ma, a new spreading centre had performed. In addition to the MCS data, an ocean-bottom
developed between Madagascar and India–Seychelles to form seismograph (OBS) wide-angle seismic profile was also acquired
the Mascarene Basin. At 65 Ma, spreading jumped to the north along the axis of the Dalrymple Trough, coincident with MCS
to leave the Seychelles on the African plate. This coincided line CAM28.
with an increase in the spreading rate and with extensive Other single-channel seismic data are also available in the
volcanic activity which formed the Deccan Traps (e.g. White research area. The second main data set consists of a series of
& McKenzie 1989). India continued its rapid movement north, eight single-channel seismic lines which cross the Murray Ridge
at rates between 150 and 200 mm yr−1, until the Palaeocene– (Fig. 3). These were acquired on the RRS Shackleton in 1980.
Eocene boundary (55 Ma) when it is possible that initial Further single-channel seismic data are available from the US
contact had been made between India and the continental Navy Ship W ilkes, which shot a number of east–west-orientated
blocks to the north (Powell 1979). During the Oligocene and lines across the Gulf of Oman.

© 2000 RAS, GJI 142, 461–477


T he Indian–Arabian plate boundary 463

(a) 141 Ma 65 Ma 53 Ma 23 Ma

ARA EUR
AFR EUR
ARA ARA
IND I
ARA
SEY
IND SEY
AFR SEY AFR AFR
IND SEY
MAD

MAD
MAD

MAD
ANT

(b) 141 Ma 65 Ma 53 Ma 23 Ma

ARA EUR
AFR EUR

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


ARA ARA
IND I
ARA
SEY
IND SEY
AFR SEY AFR AFR
IND SEY
MAD

MAD
MAD

MAD
ANT

141 Ma 65 Ma 53 Ma 23 Ma

Figure 2. Cartoon showing the two main hypotheses for the evolution of the Gulf of Oman: (a) the model of Whitmarsh (1979), and ( b) the model
of Mountain & Prell (1990). See text for further explanation. Grey shaded regions show the possible ages of oceanic crust in the two models. The
locations of the boundaries between the different ages of crust are intended for guidance only and are not accurate. The black dot shows the
approximate location of the Murray Ridge. AFR, African plate; ANT, Antarctic plate; ARA, Arabian Plate; EUR, Eurasian plate; IND, Indian
plate; MAD, Madagascar; SEY, Seychelles microplate.

grabens here are generally wider than the Dalrymple Trough


MO R P H O LO GY O F T HE M U R RAY R ID GE
to the south, with widths of 30–55 km, and are delineated by
From the available bathymetric and seismic data, the Murray faults with smaller surface throws. The trend of these northern
Ridge can be considered to extend from Qalhat Seamount at grabens is 055°, i.e. more easterly than the Dalrymple Trough
22°N to approximately 24°N, where the topographic expression but the same as the southeastern ridge.
of the ridge is buried beneath the thick sediments of the Indus
Fan. The ridge itself can be divided into two main segments,
Southern Murray Ridge
which have different morphologies.
The southern Murray Ridge consists of a 150 km long, The Dalrymple Trough is crossed by MCS lines CAM16,
25 km wide, fault-bounded trough, the Dalrymple Trough CAM17 and CAM29, and single-channel seismic lines SH2
(Figs 3 and 4), which is flanked on its southeastern side by and SH3 (Fig. 3). Lines CAM16 and CAM17 also cross the
a steep-sided ridge. To the northwest, a low, broad swell is topographic high of the Murray Ridge at their southeastern
apparent from the bathymetry. The floor of the Dalrymple ends. All the lines exhibit a similar morphology (Fig. 4). The
Trough reaches depths of 4400 m, some 1000 m deeper than most striking feature on the seismic sections is the large normal
the seafloor in the Gulf of Oman to the northwest. The fault which bounds the southeastern edge of the Dalrymple
Dalrymple Trough trends 045° while the southeastern ridge Trough (Fig. 5). The throw of the fault at the seabed is 1800 m
trends 055°. The two are separated by a 2000–2500 m deep on CAM17. The Dalrymple Trough contains sediments with
sediment-covered platform, the width of which increases from a thickness of more than 2 s TWT (two way traveltime)
10 km in the south to 35 km in the north. Profiles across the (approximately 2 km). Acoustic basement in the trough can-
southern Murray Ridge are therefore very asymmetric. not be observed on any of the reflection profiles. Lines from
The northern portion of the Murray Ridge is much more CAM16 to CAM29 show the sediments of the trough dipping
symmetric in character (Figs 3 and 4). Fault-bounded grabens gently to the southeast with dips of up to 2° (Fig. 6). The large
are again observed but these are flanked to the northwest southeastern fault is therefore the main fault of a half-graben,
and southeast by broader, lower-lying topographic highs. The and the sediments in the hanging wall have been rotated

© 2000 RAS, GJI 142, 461–477


464 R. A. Edwards, T . A. Minshull and R. S. W hite

60o E 61o E 62o E 63o E 64o E 65o E 66o E


PAKISTAN

o
25o N 25 N
1000
00
2000 MAKRAN 20

SH
SH
SH

8
7
6
SH
3000

5
24o N 24o N

SH
GULF OF OMAN

4
SH

00
30
3
CA
SH2
SH

M2
30

1
00
E
23o N 23o N

9
CA G
M1 CA 0 ID
3 M1 00 R

3000
4
CA 7 DT AY
M1 R
R

28
C
4 U

AM

AM
M

16
2000 INDUS

C
QALHAT FAN
o o
22 N SEAMOUNT 22 N
3000

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


CA
M1
0 9
00 o
21o N 2 21 N
Z
OF

100 km
o
20 N 20o N
60o E 61o E 62o E 63o E 64o E 65o E o
66 E

Figure 3. Bathymetry of the Gulf of Oman and Murray Ridge. Bathymetric contours in metres are taken from a 5×5 min grid of all the available
shipboard data in the area. Solid lines mark the multichannel seismic profiles available from RRS Charles Darwin cruise 18, broken lines show the
single-channel seismic profiles from RRS Shackleton cruise 1/80. DT, Dalrymple Trough; OFZ, Owen Fracture Zone. White stars show the location
of two dextral strike-slip earthquakes (Quittmeyer & Kafka 1984). Locations of Figs 6, 7, 8 and 9 are shown by the bold portions of line.

downwards (Minshull et al. 1992). The sediments show slight from the unmigrated single-channel profiles were corrected for
upward folding due to either drag on the fault plane or migration. The antithetic faults on the northwestern flank of
differential compaction from the centre to the edge of the the trough dip at 30°–50°, but the main fault to the southeast
basin. The northwestern wall of the Dalrymple Trough is dips at only ~20°. Towards the northern end of the Dalrymple
bounded by several antithetic normal faults, the throws of Trough, where the structure becomes more symmetrical,
which increase from the southern end of the trough towards both bounding faults dip at only ~25°. The total horizontal
the northern end. Individual reflectors are difficult to trace extension may be estimated by summing the horizontal offsets
across the faults even though the throw of the fault may be corresponding to each seabed fault scarp. This approach
small, suggesting a component of strike-slip motion (Fig. 6). assumes that the tilting of individual blocks is small and that
On profile SH3 the structure is more reminiscent of a full there has been no syn- or post-rift sedimentation, and no
graben, with the downfaulted sediments now dipping towards erosion of the footwall. A more rigorous approach is not
the northwest, suggesting a switch in the main fault from the possible because the stratigraphy cannot be correlated reliably
southeastern wall to the northwestern wall of the trough across the major faults. The tilt of the bedding is generally
(Fig. 4). Profile CAM28, which lies along the axis of the small (Fig. 4), but on the northern Murray Ridge there is
Dalrymple Trough (Fig. 3), shows gentle folding of the lower evidence for considerable synrift sedimentation. In addition,
sedimentary sequences at its northeastern end (Fig. 7), indicating only faults with vertical offsets of greater than ~200 m could
a small component of compression along the plate boundary. be measured reliably. Any such estimate of the total extension
Irregularities in the shape of the plate boundary could lead to is therefore a minimum. Application of this approach to the
localized compression as the two plate slide past each other. southern five profiles of Fig. 4 yields a total extension of
The synclines of the folds contain ponded sediments. Acoustic 5–7 km.
basement cannot be observed on profile CAM28. The main topographic high of the Murray Ridge is located
Profiles CAM16, CAM17, CAM29 and SH3 run almost to the southeast of the Dalrymple Trough. Here the footwall
exactly orthogonal to the walls of the Dalrymple Trough, so of the main normal fault shows 2 s TWT of sediments overlying
may be used to estimate fault dips if the faults are assumed to a rough acoustic basement (Fig. 5). The sediments are back-
run parallel to these major scarps. Dips were estimated by tilted away from the fault as unloading of the footwall has
fitting straight lines to the faults imaged on the profiles and led to uplift (White 1984). This sediment-covered platform
assuming a mean seismic velocity of 2 km s−1, appropriate for increases in width by 25 km from the southwest to the northeast
the faulted upper part of the sediment column. Dips measured (Fig. 4). The thick, relatively flat-lying sediments in the footwall

© 2000 RAS, GJI 142, 461–477


T he Indian–Arabian plate boundary 465

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021

Figure 4. Line drawings of the seismic profiles that cross the Murray Ridge. Profiles are aligned along the axis of the ridge. The stippled region
marks acoustic basement. Individual sedimentary reflectors shown are not necessarily consistent between different profiles or across major faults,
but are intended to show sediment dip and unconformities only. DT denotes the Dalrymple Trough. The vertical axes show two-way traveltime
in seconds.

© 2000 RAS, GJI 142, 461–477


466 R. A. Edwards, T . A. Minshull and R. S. W hite

500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000
0
NW SE
CAM17

2
3
4
TWT [s]

v v
v v
6
1

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


8

10 km

10

Figure 5. MCS profile CAM17 across the southern Murray Ridge showing the Dalrymple Trough. The section has been f–k-migrated with a con-
stant velocity of 1500 m s−1. Every fourth trace is plotted. Labelled sedimentary units correspond to: (1) turbidites deposited on basement—these
have then been deformed and uplifted prior to the deposition of (2); (3) flat-lying recent sediments.

suggest that these are older sediments which have been elevated rather enigmatic structures which may represent volcanic flows
by faulting, rather than sediments deposited after the faulting. interfingering with the sediments. At 6.5 s TWT, beneath the
These sediments must be relatively rigid, in other words old sediments, an interface with multiple hyperbolic reflections
and lithified, to maintain the shape of the fault plane. This can be seen. This interface has a similar character to acoustic,
‘hanging basin’ is bounded to the northeast by the outcropping volcanic basement. However, this reflector is underlain by a
basement of the Murray Ridge. The sediment layers terminate strong reflection at just over 7 s TWT (Fig. 8). It is possible
abruptly against the acoustic basement, suggesting that the that the deeper reflector represents the true basement which
basement has either been faulted upwards, or that it formed a has been overlain by volcanic or debris flows, the tops of which
topographic high prior to deposition of the sediments. Along are represented by the shallower interface with hyperbolic
line CAM17 the Murray Ridge appears to be tilted towards reflections.
the southeast, suggesting that the ridge may be a tilted fault To the northwest of the Dalrymple Trough, the thick sedi-
block (Fig. 5). There is no evidence for uplift of the Murray ments and acoustic basement on the Arabian plate are well
Ridge coeval with deposition of the sediments to the northwest. imaged on MCS lines CAM16, CAM17 and CAM29. Basement
The elevated topography of the Murray Ridge might be highs can be observed towards the northwestern end of all
explained by a phase of compression across the plate boundary. three lines (Fig. 4), and these mark the boundary between the
Similar features are observed at the Mendocino Transform undisturbed sediments of the Gulf of Oman to the west and
Fault, which separates the Juan de Fuca/Gorda plate and the the faulted sediments of the Dalrymple Trough and Murray
Pacific plate (Godfrey et al. 1998). The Mendocino Transform Ridge to the east. To the southeast of these basement highs,
Fault is flanked by the Mendocino Ridge, an elevated tilted the sediments dip at a shallow angle towards the Dalrymple
block which slopes away from the plate boundary, similar to Trough and contain numerous minor normal faults. The
the geometry of the Murray Ridge. Godfrey et al. (1998) basement also deepens towards the southeast. The acoustic
interpret the Mendocino Ridge as a pop-up feature caused by basement is overlain by a thin layer (up to 0.5 s TWT) of
compression across the plate boundary. Sediment cores from highly reflective sediments which infill the topographic lows
DSDP sites 223 and 224, located at 18°N and 16°N, respectively, of the basement and onlap the basement highs (Fig. 5). This
on the Owen Fracture Zone, show evidence of uplift in the suggests that these sediments are turbidite flows. The basement
Early Miocene (Whitmarsh 1974). This uplift has been attri- has subsequently been tilted to the southeast and the base-
buted to the plate reorganization at that time, as the Gulf of ment high uplifted, causing uplift of the turbidite unit on its
Aden started to open. A small change in the plate motions southeastern flank.
could lead to compression across the Owen Fracture Zone The younger sediments are deposited unconformably on
and Murray Ridge. Any uplift due to compression across the the underlying turbidites. These units thicken towards the
Murray Ridge is therefore probably also Miocene in age. Dalrymple Trough and onlap onto the lower unit towards the
The southeastern side of the Murray Ridge is less sharply northwest, suggesting continued subsidence of the Dalrymple
delineated. To the east the thick sediments of the Indus Fan Trough. The sediments have then been faulted, possibly by
can be observed (Fig. 8), but the flank of the ridge exhibits reactivation along existing faults.

© 2000 RAS, GJI 142, 461–477


T he Indian–Arabian plate boundary 467

NW CDP SE
3000 2500 2000 1500 1000
0

CAM29 10 km
2

4
TWT [s]

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


10

CDP
3000 3500 4000 4500 5000 5500
0

CAM17
2

4
TWT [s]

10

CDP
4500 4000 3500 3000 2500 2000
0

CAM16
2

4
TWT [s]

10
Figure 6. Part of MCS profiles CAM16, CAM17 and CAM29 showing the Dalrymple Trough. The sections have been f–k-migrated with a
constant velocity of 1500 m s−1. Every second trace is plotted. See Fig. 3 for locations.

© 2000 RAS, GJI 142, 461–477


468 R. A. Edwards, T . A. Minshull and R. S. W hite

SW CDP NE
6500 7000 7500 8000
4
CAM28 10 km

6
TWT [s]

10
Figure 7. Part of MCS profile CAM28 at the northeastern end of the Dalrymple Trough showing gentle folding along the axis of the trough. The

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


section has been f–k-migrated with a constant velocity of 1500 m s−1. Every second trace is plotted. See Fig. 3 for location.

NW CDP SE
Northern Murray Ridge
1000 500
2 The northern portion of the Murray Ridge has a very different
CAM16 morphological expression from its southern counterpart. The
10 km steep-sided graben of the Dalrymple Trough and the high
elevation of the ridge are replaced by much more subdued
4 topography to the north (Fig. 4). Single-channel seismic pro-
files SH4 and SH5 show a wide graben without elevated flanks.
Again the graben is filled with thick sediments and the acoustic
basement cannot always be detected. The faults which define
TWT [s]

the graben have much smaller seabed throws than those of the
6 Dalrymple Trough, and individual faults cannot be traced
from profile to profile. The southeastern end of line SH4 shows
thick sediments terminating abruptly against possibly faulted
acoustic basement to the southeast, while the sediments them-
8 selves sag downwards into a broad basin shape. This suggests
that the main locus of extension may have originally been
located at the southeastern end of the line but has switched to
its current location, leaving a relict basin on the flanks of the
current graben. The northwestern end of line SH4 shows
10 the acoustic basement clearly. Like the case for the southern
Figure 8. Part of MCS profile CAM16 showing the southeastern Murray Ridge, the oldest sediments appear to have been
flank of the Murray Ridge and the adjacent thick sediments of the uplifted along with the acoustic basement, and more recent
Indus Fan. The section has been f–k-migrated with a constant velocity sediments are unconformable on the deeper, uplifted unit
of 1500 m s−1. Every second trace is plotted. See Fig. 3 for location. (Fig. 4). A similar pattern can also be observed at the south-
eastern end of line SH7. Lines SH5 and SH6 show the widest
grabens, and there is some evidence, particularly on line SH5,
All of the seismic reflection profiles across the southern for an uplifted horst in the centre of the graben.
Murray Ridge show a thin veneer (0.2 s TWT) of flat-lying Moving to the north, the character of extension changes
recent sediments. Although these sediments are faulted across again, and structures more reminiscent of the southern Murray
the main normal faults, they post-date the major fault move- Ridge are observed along lines SH7 and SH8, with a steeper-
ments along the ridge, such as the footwall uplift along the sided, narrow graben (Fig. 4). The sediments in the graben on
main normal fault (Fig. 5). The main phase of deformation along line SH7 dip towards the southeast. The floor of the graben
the southern Murray Ridge appears to have ceased, and any shallows towards the north due to increased sediment input
current deformation is limited to small-scale fault reactivation. from the Indus Fan.
The Murray Ridge is, however, seismically active, showing The wide grabens observed in the northern portion of the
that faulting is still occurring. Focal mechanisms determined Murray Ridge suggest that there has been greater extension
for two earthquakes along the northern Murray Ridge (Fig. 3) here than to the south. This agrees well with the strike of the
have predominantly dextral strike-slip motion (Quittmeyer & ridge, which becomes more easterly towards the north (Fig. 3).
Kafka 1984). The ridge is therefore at a greater angle to the strike-slip Owen

© 2000 RAS, GJI 142, 461–477


T he Indian–Arabian plate boundary 469

Fracture Zone, and a greater component of extension can be and then subsided. The unconformity slopes gently towards
expected. Like the case for the southern Murray Ridge, the the southeast, suggesting that the subsidence is not uniform.
deformed sediments to the north are overlain by a thin layer This increased subsidence towards the plate boundary may be
of flat-lying recent sediments. explained by either downward bending of the plate, due to the
load of the seamount, increasing towards the edge of the plate
(assuming that the Arabian and Indian plates are not coupled
Qalhat Seamount and the transition to the Owen Fracture
across the Owen Fracture Zone) or by extension and there-
Zone
fore crustal thinning and associated subsidence at the plate
Qalhat Seamount is located at 22°N, 62°E, at the intersection boundary.
of the Murray Ridge and the Owen Fracture Zone. This is Moving south of Qalhat Seamount, the transition from the
the northerly of two closely spaced seamounts, which sit on the Murray Ridge to the Owen Fracture Zone can be considered
Arabian plate. The seamount is crossed by two MCS profiles, complete. There is little evidence of extension, although the
CAM14 and CAM19 (Fig. 3). From the bathymetric and plate boundary remains an abrupt feature only 20 km wide.
MCS data, Qalhat Seamount is identified as a steep-sided, The Owen Fracture Zone is characterized morphologically by
flat-topped, symmetrical feature which rises 2000 m above the a ridge comprising a steep easterly facing scarp and a more
abyssal plain to depths of just 300 m below the sea surface. gentle western scarp, located on the Arabian plate. The plate
Examination of MCS line CAM14, which crosses the boundary itself is located at the foot of the steep eastern scarp
northern edge of the seamount (Fig. 9), shows the upper 2.5 s (Whitmarsh 1979; Mountain & Prell 1990). This topographic

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


TWT of flat-lying sediments in the Gulf of Oman onlapping relief forms a continuous barrier to the turbidites, and hence
the northwestern flank of the seamount. The seamount must marks the western limit of the Indus Fan.
therefore pre-date the deposition of these sediments. The base
of this sedimentary package is marked by an unconformity
with highly reflective layered sediments beneath. The reflective C O M PA R I SO N S WI TH TH E M O RP H O LO G Y
character of this lowermost unit appears to be the same as
O F M I D- O C EA N R ID GE S A N D
that of the turbidite unit observed on line CAM17 (Fig. 5). To
CON TI NE NTA L R IF TS
the northwest this lower sediment package infills the topo-
graphy of the undulating acoustic basement, but towards the The Murray Ridge exhibits extension in a transtensional
southeast it appears to be interfingered with rough volcanic regime, and as such might be expected to show morphological
(debris flow?) units on the seamount’s flank. The acoustic, similarities to either, or both, continental rifts or mid-ocean
presumed oceanic, basement of the Gulf of Oman can be ridges. The topography of both the Mid-Atlantic Ridge and the
traced under the flank of the seamount. East African Rift show similarities to the Dalrymple Trough:
The top of the seamount is marked by a flat subhorizontal a narrow, steep-sided deep flanked by elevated ridges (Mutter
unconformity which separates 0.3 s TWT of sediments above & Karson 1992; Hayward & Ebinger 1996). Mutter & Karson
from rough acoustic basement below. This flat unconformity (1992) relate the morphological similarities of slow spreading
indicates that the seamount was once subaerial, was eroded, mid-ocean ridges and continental rifts to their dependence on

3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000 8500 9000 9500 10000
0
NW CAM14 SE

2
PLATE
BOUNDARY

4
TWT [s]

10 km

10
Figure 9. Part of MCS profile CAM14 showing the thick sediments in the Gulf of Oman and Qalhat Seamount. The section has been
f–k-migrated with a constant velocity of 1500 m s−1. Every fourth trace is plotted. The plate boundary is located between CDPs 8300 and 9000.
The Arabian Plate extends to the northwest; the Indian Plate extends to the southeast. See Fig. 3 for location.

© 2000 RAS, GJI 142, 461–477


470 R. A. Edwards, T . A. Minshull and R. S. W hite

mechanical processes rather than magmatic processes. The dip plate from the Israel–Sinai subplate (Wdowinski & Zilberman
on the main fault of the Southern Murray Ridge is, however, 1997; Ben-Avraham 1985). Seismic reflection profiles across
smaller than for mid-ocean-ridge normal faults, which typically the Gulf of Elat show a series of asymmetric, steep-sided, fault-
dip at 30°–60° (e.g. Huang & Solomon 1988) and may have bounded basins containing thick sediments that tilt either to
even steeper dips at the seabed (Karson et al. 1987). Dips the west or to the east (Ben-Avraham 1985). The Gulf of Elat
below 30° have been found in continental settings such as the therefore shows strong similarities to the Dalrymple Trough.
Woodlark Basin, where extension is occurring along a low- Ben Avraham (1992) suggests that the asymmetric basins are
angle (~27°) basin-bounding normal fault (Taylor et al. 1999), caused by strike-slip motion along one bounding fault (the
and on rifted margins (Reston et al. 1996). deeper side), while the other side of the basin is defined by a
Plate circuit calculations show that the present-day motion series of normal faults. If this model is applied to the Murray
along the Owen Fracture Zone is 2 mm yr−1 of right-lateral Ridge it predicts that the main strike-slip motion is along the
strike-slip (Gordon & DeMets 1989), making this the slowest- southeastern boundary of the Dalrymple Trough. From the
slipping plate boundary on Earth today. From plate tectonic seismic reflection profiles, sedimentary units cannot be traced
reconstructions there is no evidence for changes in the plate across all the faults on the northwestern side of the trough,
motions since the Oligocene–Early Miocene plate reorganization also suggesting that there is some strike-slip motion along this
(Besse & Courtillot 1988). Prior to this time the motion along side of the basin.
the Owen Fracture Zone is not well constrained, the two main
hypotheses suggesting either no motion, or considerable strike-
W I DE- A N GL E S EI S MI C DATA

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


slip motion as India moved northwards. There is no evidence
for extension along the Murray Ridge/Dalrymple Trough before Limited wide-angle seismic data are available from a short
23 Ma. Thus, the current extension rate across the Dalrymple reversed profile located along the axis of the Dalrymple
Trough must be extremely slow. At such slow rates, conductive Trough, coincident with MCS line CAM28. The data were
cooling of the upwelling mantle beneath the thinning litho- recorded by two ocean-bottom seismographs (OBS) and shot
sphere would prevent decompression melting such as usually using a single 700 in3 airgun. The source-to-receiver ranges
occurs beneath oceanic spreading centres (e.g. Bown & White were determined from the navigation and checked by ray
1994). At the Southwest Indian Ridge, where the full spreading tracing the direct water wave. Because the airgun source
rate is 12–16 mm yr−1 (Chu & Gordon 1999), this cooling volume is small the data quality is rather poor and it is difficult
effect results in only ~4 km of melt generation compared to identify arrivals at offsets greater than 20 km (Fig. 10).
to the ~7 km normally observed (Muller et al. 1997); the Traveltimes of first and second arrival phases were picked, and
Dalrymple Trough probably marks a more extreme example the picks were assigned uncertainties of 50–75 ms. The data
of the same process. From the magnetic and seismic data there were iteratively forward modelled using the  code of
is no evidence for any magmatism related to extension within Zelt & Smith (1992). Fig. 11 shows the observed and computed
the Dalrymple Trough. traveltimes and the best-fit velocity model obtained. The root-
From the seismic reflection profiles discussed above it is mean-square misfit between the observed and calculated travel-
clear that the Murray Ridge exhibits segmentation along its times is 45 ms, which is comparable with the traveltime pick
length; different portions of the ridge have different structural uncertainty: the model is therefore the smoothest that can be
styles. The southern and northern Murray Ridge form two produced within the resolution of the data. The upper two
clearly separate segments, and the northern Murray Ridge sedimentary layers, with velocities of 1.75 and 2.15 km s−1,
could be divided again between lines SH6 and SH7 where the were determined from the stacking velocities of MCS line
graben width changes rapidly (Fig. 4). From the bathymetry, CAM28. The final model comprises an almost 1-D structure
the main graben can be seen to be offset to the east from line with a crystalline crustal thickness of 12–14 km, overlain by
SH3 to SH4 (Fig. 3). It appears that the northern segment of ~3 km of sediments. The upper crust beneath the sediments
the Murray Ridge has undergone more extension than the is modelled by a layer with velocities of 4.6–5.4 km s−1, which
southern segment, as the graben is wider and the surface are constrained by first arrivals observed out to 25 km offset.
throws of the faults are smaller. The northern segment has the These velocities are comparable to those found in oceanic
appearance of a full graben, while the southern segment shows layer 2, or in the upper part of continental crust. The velocity
the morphology of a half-graben. This is a pattern observed of the lower crust is unconstrained, as refracted arrivals are
on continental rifts, where as the rift develops the morphology not observed. Velocities in the lower crust of 5.6–6.7 km s−1
changes from a half-graben to a full graben with smaller fault were included in the model. If these are the correct lower-
lengths and throws as the crust becomes thinner (Hayward & crustal velocities, the crustal thickness has to be of the order
Ebinger 1996). The coverage of seismic data over the Murray of 12–14 km to fit the reflected Moho arrival observed at long
Ridge is not sufficiently dense to determine the form of the offsets. Alternatively, the lower crust could be modelled with
segment ends, so we cannot tell whether these are marked by higher velocities of 6.6–7.2 km s−1, more reminiscent of oceanic
normal faults or by purely strike-slip motion layer 3, which would require a slightly thicker crust, or with
Although the extensional deformation along the Murray lower velocities which should allow the crust to be thinner
Ridge shows similarities to both continental rifts and mid- and still fit the observed Moho reflection.
ocean ridges, comparisons are best made with similar tectonic
settings where both strike-slip and extensional tectonics are
G R AVI TY DATA A N D M O DE LLI N G
present within pull-apart basins along ‘leaky’ transform faults.
A good example of rift basins along a transform fault is Fig. 12(a) shows the satellite gravity data over the Gulf of
provided by the Dead Sea Transform, which runs from the Oman. The Murray Ridge and the Dalrymple Trough are
Gulf of Elat, north as far as Turkey and separates the Arabian marked by very large positive and negative free-air gravity

© 2000 RAS, GJI 142, 461–477


T he Indian–Arabian plate boundary 471

Offset [km]
-30 -20 -10 0 10 20
0

6
t - x/6.0 [s]

10

12

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


14

OBS9 Hydrophone
Figure 10. Wide-angle seismic data recorded on OBS 9 at the northeastern end of the line along the Dalrymple Trough. The data have been
bandpass filtered from 6 to 15 Hz, and are plotted with a reduction velocity of 6 km s−1.

anomalies, respectively. The anomalies range from 230 to is 9.7 mgal. The greatest misfit occurs at 140–160 km along
−150 mgal. To the south, Qalhat Seamount also shows a large the model, over the crest of the Murray Ridge. The calculated
positive anomaly. The large magnitude of these anomalies anomaly contains a short-wavelength feature which is not
suggests that the Murray Ridge and Dalrymple Trough are observed in the satellite gravity data. This may be due to either
not isostatically compensated. a 3-D effect from the Murray Ridge, or the long-wavelength
To help constrain the crustal structure under the Murray nature of the satellite data, which will not resolve anomalies
Ridge and Dalrymple Trough, a simple 2-D gravity model was with wavelengths of less than 20 km (Marks 1996). The final
computed. The free-air gravity anomaly was extracted from gravity model has a crystalline crust 13 km thick under the
gridded satellite data (Smith & Sandwell 1995) along MCS Dalrymple Trough. If our assumptions about densities are
line CAM17 (Fig. 3). The profile was also extended 120 km to correct, the Murray Ridge and the low basement high to the
the northwest, into the Gulf of Oman (Fig. 12). The bathymetry northwest of the Dalrymple Trough are both underlain by
and depth to the top of basement were taken from MCS line crust with a thickness of ~17 km. To the northwest, the Gulf
CAM17. As further constraints are not available, the top of of Oman is underlain by a crust 6 km thick. Calculation of the
basement was assumed to extend horizontally into the Gulf pressure along the base of the model shows relatively large
of Oman from the end of CAM17, at a depth of 7.5 km. The lateral variations in pressure over distances of several tens of
sediments overlying the basement were modelled as a single kilometres (Fig. 13). This suggests that the model is not in
isostatic balance, and some form of additional dynamic support
layer. From the wide-angle seismic modelling, the average
may be necessary for the lithosphere to withstand the lateral
interval velocity for the whole sediment column was deter-
pressure changes. More accurate crustal thickness information
mined to be 3.2 km s−1, and this corresponds to a density of
will come from new wide-angle seismic data, which suggest
2.4 g cm−3, according to the relationship of Hamilton (1978).
slightly lower densities, velocities and crustal thicknesses
Densities in the crust were derived from the velocities obtained
(Edwards et al. 1998).
during the wide-angle seismic modelling using the velocity–
density relationship of oceanic crust of Carlson & Herrick
(1990). The crustal structure at 110 km along the gravity model M A GN ET IC D ATA A ND M O D EL LIN G
is tied to the wide-angle seismic model, with 3 km thick A compilation was made of all the shipboard magnetic data
upper crust (density 2.61 g cm−3) and 10 km thick lower crust available in the Gulf of Oman. After examination of track
(density 2.84 g cm−3). The upper mantle was assigned a density crossovers and the elimination of bad tracks, the data were
of 3.33 g cm−3. As no constraints are available for the crustal reduced to IGRF95, and the anomalies were gridded on a
structure other than the tie with the wide-angle seismic profile, 5×5 min grid (Wessel & Smith 1991). Fig. 12(b) shows the
the upper crust was assumed to have a constant thickness magnetic anomalies in the study area. Perhaps the most
across the model. interesting feature of the contour plot is the lack of any strong
Fig. 13 shows the best-fit gravity model. The root-mean- magnetic anomaly associated with the topographic high of
square misfit between the observed and calculated anomalies the southern Murray Ridge. This suggests that the ridge does

© 2000 RAS, GJI 142, 461–477


472 R. A. Edwards, T . A. Minshull and R. S. W hite

SW NE

OBS 13 OBS 9
(a)
2

4
T - X/6.0 (s)

9
0 10 20 30 40 50 60 70 80

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


CAM17 DISTANCE (km)

0 10
10 20 30 40 50 60 70 80
(b) 0
1.51
2.60 1.75
1.75
1.75

3.50 5 2.15

4.10
4.60 4.60
DEPTH (km)

5.40 5.40
10
5.60 5.60 5.60

15

6.70
20 6.70 7.80
7.80

8.00 8.00 8.00


25

Figure 11. (a) Observed and computed traveltimes for the two OBHs used to record the wide-angle seismic data in the Dalrymple Trough, plotted
with a reduction velocity of 6 km s−1. ( b) The best-fit velocity model for the wide-angle seismic line showing the rays traced. Note the poor ray
coverage of the lower crust. Velocities are shown in km s−1.

not have a volcanic origin, although basaltic tuffs have been non-unique, and although the preferred final model was chosen
dredged from the northeast end of the ridge (Barker 1966). because it reproduced well the main features of the observed
Likewise, the Dalrymple Trough does not have a magnetic data (i.e. the prominent high over the basement high to the
signature. northwest of the Dalrymple Trough, the weak low over the
The strongest magnetic anomalies are observed to the Dalrymple Trough, and the weak high over the Murray Ridge),
northwest of the Murray Ridge. A linear positive anomaly is alternative models may also provide a good fit. The palaeo-
flanked on its northwestern side by a parallel negative anomaly, inclination and -declination of the oceanic crust in the Gulf
the location of which coincides with the Little Murray Ridge, of Oman were set at −19° and −10°, respectively. These
a buried basement high (White 1983). These two anomalies values were estimated from the plate reconstructions of Bess
appear to be offset twice along their length: once at 22°N and & Courtillot (1988), assuming the crust is Eocene in age.
again at 24°N. These offsets may be related to the segmentation Whole-crust magnetization has been assumed in the model,
observed along the Murray Ridge. The magnetic data show and the sediments are assumed to be effectively non-magnetic.
no evidence for seafloor spreading anomalies in the Gulf of The final model shows reversely magnetized crust, with a
Oman. Any magnetic anomalies in the Gulf of Oman are NRM of −2.5 A m−1, extending under the Gulf of Oman
parallel to the Murray Ridge and the Little Murray Ridge and (Fig. 14). More weakly reversely magnetized crust is included
hence are probably related to the evolution of these ridges. in the model to 90 km, the northwest edge of the Dalrymple
A 2-D magnetic model was constructed along the same Trough. The crust under the Dalrymple Trough and Murray
profile as the gravity model in Fig. 13, using layer thicknesses Ridge has been modelled with a weak induced magnetization
from the gravity model and varying only the magnetizations of 0.2–0.9 A m−1. The magnetic anomaly was also calculated
of the five blocks shown. The magnetic anomaly was taken assuming a strong NRM of −5.0 A m−1 in the upper 3 km of
from the gridded magnetic data set. The observed and calcu- crust under the Gulf of Oman, as might be expected for oceanic
lated magnetic anomalies and the preferred final magnetic crust. This model also provides an adequate fit to the observed
model are shown in Fig. 14. Magnetic models are notoriously anomaly (Fig. 14).

© 2000 RAS, GJI 142, 461–477


T he Indian–Arabian plate boundary 473

(a) o
o
60 E
o
62 E
o
64 E
o
66 E o

26 N 26 N

200
175
o o
150

free air gravity anomaly (mGal)


24 N 24 N 125
100
75
50
25
o o 0
22 N 22 N -25
-50
-75

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


-100
-125
o o
-150
20 N 20 N
-175
-200
o o o o
60 E 62 E 64 E 66 E

(b) o
o
60 E
o
62 E
o
64 E
o
66 E o

26 N 26 N

400
350
o o
300
24 N 24 N 250
200
150
100 magnetic anomaly (nT)
50
o o 0
22 N 22 N -50
-100
-150
-200
-250
o o -300
20 N 20 N
-350
-400

o o o o
60 E 62 E 64 E 66 E
Figure 12. (a) Satellite free-air gravity anomalies in the Gulf of Oman. The location of the gravity model in Fig. 13 is shown by the white line.
Bathymetric contours are shown in black, and seismic profiles by the thicker black lines. ( b) Magnetic anomalies in the Gulf of Oman. The
anomalies are gridded on a 5×5 min grid from all available shipboard data. The lower right-hand box shows the location of tracks. The white
line shows the location of the magnetic model in Fig. 14.

© 2000 RAS, GJI 142, 461–477


474 R. A. Edwards, T . A. Minshull and R. S. W hite

NW SE

Pressure

Pressure
900 900

(MPa)

(MPa)
800 800

Free-air Gravity Anomaly (mgal)


Free-air Gravity Anomaly (mgal)
100 100
observed anomaly
calculated anomaly

0 0

-100 -100

0 0

2.40

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


2.40 2.61

10 10

Depth (km)
Depth (km)

2.84

20 3.33 20

30 30
-100 -50 0 50 100 150
Distance (km)

Figure 13. 2-D gravity model across the Murray Ridge. The middle panel shows the observed (thick line) and calculated (thin line) anomalies; the
lower panel shows the best-fit gravity model. The upper panel shows the pressure along the base of the model. Densities are labelled in kg cm−3.
The model is tied to the wide-angle model at 110 km model distance. Model location is shown in Fig. 12.

The absence of strong magnetic anomalies over the Murray


IN TER P R ETATI ON O F TH E M O DE LS —
Ridge (Fig. 12b) suggests that the ridge is not volcanic in
CR US TA L S TR U CTU RE
origin, although the seismic reflection data and dredged tuffs
The wide-angle seismic, gravity and magnetic models have do suggest limited volcanism along the southeastern margin
provided information about the nature and thickness of the of the ridge (Fig. 8). The preferred magnetic model in Fig. 14
crust under the Murray Ridge and Dalrymple Trough (Figs 11, includes weakly magnetized, 0.2–0.9 A m−1, crust underlying the
13 and 14). The most striking result of the modelling is that Murray Ridge and Dalrymple Trough. If the crust had been
the crust is much thicker than would be expected in a normal heavily intruded by volcanic material, a higher magnetization
oceanic setting. The wide-angle seismic and gravity data have would be expected. From the magnetic and seismic reflection
data the Murray Ridge does not appear to be volcanic in origin.
suggested a crystalline crustal thickness of ~13 km under the
The plate tectonic model of Whitmarsh (1979), which
Dalrymple Trough, and ~17 km under the Murray Ridge.
postulates that the Indian–Arabian plate boundary has lain
Normal oceanic crust has a thickness of ~7 km (White et al.
along the Owen Fracture Zone since India first started to
1992), and this is similar to the 6 km thick crust (excluding
move northwards at 90 Ma, requires considerable compression
sediments) which is modelled under the Gulf of Oman using
along the Murray Ridge due to its more easterly strike (Fig. 2).
the gravity data. The magnetic model includes reversely mag- However, this is difficult to reconcile with the seismic reflection
netized crust with a relatively high NRM of −2.5 A m−1 in data which show relatively undeformed uplifted sediments
the Gulf of Oman, again suggesting oceanic crust in this region. on the northwestern side of the Murray Ridge, suggesting that
While the Gulf of Oman appears to be underlain by normal the uplift has occurred as a discrete event following sediment
oceanic crust, the crust under the Dalrymple Trough and deposition, rather than coeval with the deposition of the sedi-
Murray Ridge is very different in character. The thick crust ments over an extended period of time. Uplift along the Owen
under the Dalrymple Trough and Murray Ridge may be Fracture Zone to the south, resulting in a change in deposition
explained by three different hypotheses: (1) the crust is oceanic from lower Miocene turbidites to middle Miocene pelagic
in origin and has been thickened by the addition of volcanic sediments at DSDP site 224, has been described by Whitmarsh
material; (2) the crust is oceanic in origin and has been (1974), who relates the uplift to a small change in the plate
thickened by compression and crustal shortening; and (3) the motions. A short phase of compression and uplift along the
crust is continental in origin. These three hypotheses are Murray Ridge (possibly during the Miocene) is supported by
discussed below. the seismic reflection data.

© 2000 RAS, GJI 142, 461–477


T he Indian–Arabian plate boundary 475

NW SE

200 200
observed anomaly

Magnetic Anomaly (nT)

Magnetic Anomaly (nT)


calculated anomaly

100 100

0 0

-100 -100
0 0

0.9
10 -2.5 -2.0 -1.5 0.2 10
Depth (km)

Depth (km)

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


20 20

30 30
-100 -50 0 50 100 150
Distance (km)

Figure 14. 2-D magnetic model. The upper panel shows the observed (thick line) and calculated (thin line) anomalies; the lower panel shows the
preferred crustal model. The dashed line shows the calculated anomaly for a model with a strongly (−5.0 A m−1) magnetized upper 3 km of crust
for −120 to −90 km model distance, over the Gulf of Oman. Magnetizations are labelled in A m−1. Model location is shown in Fig. 12.

The weak induced magnetization of the crust under the From the magnetic model (Fig. 14) the Indian–Arabian plate
Murray Ridge and Dalrymple Trough suggests that the crust boundary appears to be located along the northwestern edge
may not be oceanic in origin. Miles & Roest (1993) modelled of the Dalrymple Trough, where a large change in magnetic
seafloor spreading magnetic anomalies in the Arabian Sea and properties from crust with a strong negative magnetization
placed the ocean–continent boundary off western India at to crust with a weak positive magnetization is observed. This
~19° north, along the Laxmi Ridge. From their interpretation, observation does not agree with the model of Ben-Avraham
the Indus Fan and Murray Ridge are underlain by thinned (1992), which predicts that the plate boundary is located on
continental crust. Malod et al. (1997) and Miles et al. (1998) the southeastern boundary of the Dalrymple Trough. It is
have also proposed that the Indus Fan is underlain by thinned probable that strike-slip motion is occurring along both
continental, or transitional, crust. The velocity structure of the margins of the trough.
lower crust is commonly used as a means of distinguishing
between oceanic crust, which exhibits characteristic layer 3
CON CLU SION S
velocities of 6.6–7.2 km s−1, and thinned continental crust,
where these velocities will be absent (Whitmarsh et al. 1986). Interpretation of a series of multichannel seismic lines has
Unfortunately, refracted arrivals from the lower crust are not shown the form of deformation of the Murray Ridge and
seen in the wide-angle data, making the distinction between has provided clues as to the evolution of this feature. The
oceanic and thinned continental crust very difficult. However, Dalrymple Trough is clearly an extensional feature and the
the weak induced magnetization, and the thick crystalline crust seismic reflection data show a series of grabens delineated by
do favour a continental origin for the crust. large-offset normal faults. Seabed offsets of resolvable faults
The gravity model shows that any crustal thinning due suggest a minimum total extension of 5–7 km. The seismic
to extension under the Dalrymple Trough itself is limited reflection and magnetic data from the Dalrymple Trough show
(Fig. 13). The trough appears to be a down-faulted block, with no evidence for the generation of new igneous crust. This can
any thinning confined to the crust immediately beneath the be attributed both to the limited amount of extension, and to
bounding faults. As the Murray Ridge/Dalrymple Trough the slow extension rate, which allows for conductive cooling
forms the plate boundary, the faults must penetrate the entire of the upwelling mantle, thus preventing significant mantle
crust, and this is supported by the gravity model. Thick crust melting as it decompresses. The morphology of the Dalrymple
under the Dalrymple Trough was also proposed by Whitmarsh Trough bears a strong resemblance to continental rifts, and
(1979), who modelled a gravity profile across the Dalrymple particularly to pull-apart basins along transform faults.
Trough and southern Murray Ridge using a ~10 km thick Wide-angle seismic, gravity and magnetic modelling suggest
low-density (2.35–2.70 g cm−3) crust beneath the Dalrymple that the Murray Ridge and Dalrymple Trough are underlain
Trough to fit the observed gravity low. by 13–17 km thick crust. Any localized crustal thinning due

© 2000 RAS, GJI 142, 461–477


476 R. A. Edwards, T . A. Minshull and R. S. W hite

to extension appears to be very limited. The origin and nature Barton, P.J., Owen, T.R.E. & White, R.S., 1990. The deep structure
of the crust cannot be reliably determined from the above of the east Oman continental margin: preliminary results and
data. Although the thickness and low magnetization of the Interpretation, T ectonophysics, 173, 319–331.
crust under the Murray Ridge and Dalrymple Trough suggest Ben-Avraham, Z., 1985. Structural framework of the Gulf of Elat
a continental origin, the possibility that the ridge and trough (Aqaba), northern Red Sea, J. geophys. Res., 90, 703–726.
Ben-Avraham, Z., 1992. Development of asymmetric basins along
are underlain by deformed oceanic crust cannot be ruled out.
continental transform faults, T ectonophysics, 215, 209–220.
The seismic reflection data suggest that, although most of
Besse, J. & Courtillot, V., 1988. Paleogeographic maps of the continents
the features along the Murray Ridge can be explained by bordering the Indian Ocean since the Early Jurassic, J. geophys.
extension, another mechanism is needed to account for all Res., 93, 11 791–11 808.
the elevated topography observed. The elevated flanks of the Bown, J.W. & White, R.S., 1994. Variation with spreading rate of
graben can easily be accounted for by uplift due to extension oceanic crustal thickness and geochemistry, Earth planet. Sci. L ett.,
by, for example, isostatic unloading of the footwall. However, 121, 435–449.
the high elevation of the southern Murray Ridge is more Carlson, R.L. & Herrick, C.N., 1990. Densities and porosities in the
difficult to explain by extension alone. oceanic crust and their variations with depth and age, J. geophys.
We propose the following speculative geological history of Res., 90, 9153–9170.
the Murray Ridge. As India moved northwards during the Chu, D. & Gordon, R.G., 1999. Evidence for motion between Nubia
Cretaceous, the Indian–Arabian plate boundary was located and Somalia along the Southwest Indian Ridge, Nature, 398, 64–67.
to the west of the Owen Fracture Zone, possibly along the DeMets, C., Gordon, R.G., Argus, D.F. & Stein, S., 1990. Current

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


southeast Oman continental margin. The oceanic crust in plate motions, Geophys. J. Int., 101, 425–478.
the Gulf of Oman, formed at the proto-Carlsberg Ridge, is Edwards, R.A., Minshull, T.A., Flueh, E., Kukowski, N., Reichert, C.
& the MAMUT working group, 1998. Crustal structure of the
Palaeocene–Eocene in age, older crust having been subducted
Dalrymple Trough: a strike-slip ocean–continent boundary in the
to the north beneath the Makran convergent margin. The first
north-west Indian Ocean (Abstract), EOS, T rans. Am. Geophys. Un.,
sediments deposited on the new oceanic crust were turbidite
79, F872.
flows, the source of which probably lay to the northeast as Godfrey, N.J., Meltzer, A.S., Klemperer, S.L., Tréhu, A.M., Leitner, B.,
continent–continent collision first began between India and Clarke, S.H. & Ondrus, A., 1998. Evolution of the Gorda
the continental blocks to the north. The basement high to the Escarpment, San Andreas fault and Mendocino triple junction
northwest of the Murray Ridge and the Murray Ridge itself from multichannel seismic data collected across the northern
were then uplifted, deforming the flat-lying turbidites above the Vizcaino block, offshore northern California, J. geophys. Res., 103,
basement. At this time the onset of extension was experienced 23 813–23 825.
and the basement started to subside towards the present-day Gordon, R.G. & DeMets, C., 1989. Present-day motion along the
Dalrymple Trough. This coincided with the plate reorganization Owen Fracture Zone and Dalrymple Trough in the Arabian Sea,
during the Early Miocene that resulted in the opening of J. geophys. Res., 94, 5560–5570.
the Gulf of Aden. Motion along transform faults between the Hamilton, E.L., 1978. Sound velocity-density relations in sea-floor
Owen Fracture Zone and the southeast Oman continental sediments and rocks, J. acoust. Soc. Am., 63, 366–377.
margin possibly ceased at this time, and any motion between Hayward, N.J. & Ebinger, C.J., 1996. Variations in the along axis
the Arabian and Indian plates was confined to the Owen segmentation of the Afar rift system, T ectonics, 15, 244–257.
Huang, P.Y. & Solomon, S.C., 1988. Centroid depths of mid-ocean
Fracture Zone. Also at this time the Qalhat seamounts were
ridge earthquakes: Dependence on spreading rate, J. geophys. Res.,
formed on the edge of the Arabian plate. As the Murray Ridge
93, 13 445–13 477.
was uplifted the sediments from the Indus Fan were diverted
Hutchison, I., Louden, K.E. & White, R.S., 1981. Heat flow and age
to the southeast. Subsidence continued in the Dalrymple of the Gulf of Oman, Earth planet. Sci. L ett., 56, 252–262.
Trough and thick sediments were deposited unconformably on Jacob, K.H. & Quittmeyer, R.L., 1979. The Makran region of Pakistan
the turbidite unit below. After the deposition of the majority and Iran: Trench-arc system with active plate subduction, in
of the sediments, the present-day large normal and strike- Geodynamics of Pakistan, pp. 305–317, eds Farah, A. & DeJong, K.A.,
slip faults were formed, giving a half-graben structure in the Geological Survey of Pakistan, Quetta.
Dalrymple Trough and a full-graben structure along the northern Karson, J.A. et al., 1987. Along axis variations in seafloor spreading
Murray Ridge. This was accompanied by further (footwall) in the MARK area, Nature, 328, 681–685.
uplift of the Murray Ridge and accompanying erosion. Finally, Malod, J.A., Droz, L., Mustafa Kemal, B. & Patriat, P., 1997. Early
large vertical movements ceased and a thin veneer of recent spreading and continental to oceanic basement transition beneath
sediments was deposited. the Indus deep-sea fan: northeastern Arabian Sea, Mar. Geol.,
141, 221–235.
Marks, K.M., 1996. Resolution of the Scripps/NOAA marine gravity
AC KN O WL ED GM E NTS field from satellite altimetry, Geophys Res. L ett., 23, 2069–2072.
Miles, P.R. & Roest, W.R., 1993. Earliest sea-floor spreading anomalies
We thank the officers and crew of RRS Charles Darwin cruise
in the north Arabian Sea and the ocean–continent transition,
CD18. This work was supported by the Department of Earth
Geophys. J. Int., 115, 1025–1031.
Sciences, Cambridge contribution 5592. RAE was supported Miles, P.R., Munschy, M. & Ségoufin, J., 1998. Structure and early
by NERC grant GR3/8838, and TAM by a Royal Society evolution of the Arabian Sea and East Somali Basin, Geophys.
University Research Fellowship. J. Int., 134, 876–888.
Minshull, T.A., White, R.S., Barton, P.J. & Collier, J.S., 1992.
Deformation at plate boundaries around the Gulf of Oman, Mar.
RE FE R ENC ES
Geol., 104, 265–277.
Barker, P.F., 1966. A reconnaissance survey of the Murray Ridge, Phil. Mountain, G.S. & Prell, W.L., 1990. A multiphase plate tectonic
T rans. R. Soc. L ond., A, 259, 187–197. history of the southeast continental margin of Oman, in T he

© 2000 RAS, GJI 142, 461–477


T he Indian–Arabian plate boundary 477

Geology and T ectonics of the Oman Region, pp. 725–743, eds Wdowinski, S. & Zilberman, E., 1997. Systematic analyses of the large-
Robertson, A.H.F., Searle, M.P. & Ries, A.C., Geol. Soc. Spec. scale topography and structure across the Dead Sea Rift, T ectonics,
Publ., 49. 16, 409–424.
Muller, M.R., Robinson, C.J., Minshull, T.A., White, R.S. & Wessel, P. & Smith, W.H.F., 1991. Free software helps map and display
Bickle, M.J., 1997. Thin crust beneath Ocean Drilling Program data, EOS, T rans. Am. geophys. Un., 72, 441 and 445–446.
borehole 735B at the Southwest Indian Ridge?, Earth planet. Sci. White, R.S., 1983. The Little Murray Ridge, in Seismic Expression of
L ett., 148, 93–107. Structural Styles, AAPG Stud. Geol., 15, 1.3.10–1.3.23.
Mutter, J.C. & Karson, J.A., 1992. Structural processes at slow- White, R.S., 1984. Active and passive plate boundaries around the Gulf
spreading ridges, Science, 257, 627–634. of Oman, north-west Indian Ocean, Deep Sea Res., 31,
Powell, C.McA., 1979. A speculative tectonic history of Pakistan 731–745.
and surroundings: some constraints from the Indian Ocean, in White, R. & McKenzie, D., 1989. Magmatism at rift zones: the
Geodynamics of Pakistan, pp. 5–24, eds Farah, A. & DeJong, K.A., generation of volcanic continental margins and flood basalts,
Geological Survey of Pakistan, Quetta. J. geophys. Res., 94, 7685–7729.
Quittmeyer, R.C. & Kafka, A.L., 1984. Constraints on plate motions White, R.S., McKenzie, D. & O.Nions, R.K., 1992. Oceanic crustal
in Southern Pakistan and the northern Arabian Sea from the thickness from seismic measurements and rare earth element
focal mechanisms of small earthquakes, J. geophys. Res., 89, inversions, J. geophys. Res., 97, 19 683–19 715.
2444–2458. Whitmarsh, R.B., 1974. Summary of general features of Arabian Sea
Reston, T.J., Krawczyk, C.M. & Klaeschen, D., 1996. The S-reflector and Red Sea Cenozoic history based on leg 23 cores, Init. Rep. Deep
west of Galicia (Spain): Evidence from prestack depth migration for Sea Drill. Proj., 23, 1115–1123.
detachment faulting during continental breakup, J. geophys. Res., Whitmarsh, R.B., 1979. The Owen Basin off the south-east margin of

Downloaded from https://academic.oup.com/gji/article/142/2/461/637288 by guest on 31 May 2021


101, 8075–8091. Arabia and the evolution of the Owen Fracture Zone, Geophys. J. R.
Smith, W.H.F. & Sandwell, D.T., 1995. Marine gravity field from astr. Soc., 58, 441–470.
declassified Geosat and ERS-1 altimetry (Abstract), EOS, T rans. Whitmarsh, R.B., Avedik, F. & Saunders, M.R., 1986. The seismic
Am. geophys. Un., 76, 156. structure of thinned continental crust in the northern Bay of Biscay,
Taylor, B., Goodliffe, A.M. & Martinez, F., 1999. How continents Geophys. J. R. astr. Soc., 86, 589–602.
break up: Insights from Papua New Guinea, J. geophys. Res., 104, Zelt, C.A. & Smith, R.B., 1992. Seismic traveltime inversion for 2-D
7497–7512. crustal velocity structure, Geophys. J. Int., 108, 16–34.

© 2000 RAS, GJI 142, 461–477

You might also like