Asymmetric melt sills and upper crustal construction

beneath overlapping ridge segments: Implications for

the development of melt sills and ridge crests

C. H. Tong 1, J. W. Pye, P. J. Barton, R. S. White, M. C. Sinha 2, S. C. Singh 3, R. W. Hobbs

Bullard Laboratories, Department of Earth Sciences, University of Cambridge, Madingley

Road, Cambridge, CB3 0EZ, United Kingdom.

S. Bazin 3, A. J. Harding, G. M. Kent, J. A. Orcutt

Cecil and Ida M. Green Institute of Geophysics and Planetary Physics, Scripps Institution of

Oceanography, University of California, San Diego, CA 92093, USA.

To be submitted to Geology

Corresponding author: ch.tong@ic.ac.uk

Present addresses: 1 T. H. Huxley School, Imperial College of Science, Technology and

Medicine, Prince Consort Road, London, SW7 2BP, United Kingdom.

2 School of Ocean and Earth Science, University of Southampton, Southampton, SO14

3ZH, United Kingdom.

2

3 Laboratoire de Géosciences Marines, Institut de Physique du Globe de Paris, 75252 Paris

Cedex 05, France.

Abstract

A new 3-D tomographic velocity model and depth-converted image of the melt sills beneath

the 9°03'N overlapping spreading center (OSC) on the East Pacific Rise shows that the

upper crustal construction at this ridge discontinuity is highly asymmetric with reference to

the bathymetric ridge crests of the overlapping limbs. Despite the similarly curved ridge

crests, the asymmetries are markedly different under the two limbs, and appear to be related

to the contrasting evolutionary history of the limbs. The overlap basin is closely related to

the propagating eastern limb in terms of its seismic structure. By contrast, the self-

decapitating western limb forms a distinct morphological region that displays little structural

relationship with the adjacent overlap basin and other relict basins. The observed structural

asymmetries may arise because the rotational effects of the stress field in the overlap region

act on the upper, brittle part of the crust, while the melt sill geometry is controlled by the

deeper, ductile behavior of the plate boundary. As the OSC is migrating southward, the

differential development of melt sills and ridge crests may be inferred from the results of this

study. Ridge propagation appears to involve two major processes: the advancement of the

melt sill at the ridge tip, and the development of ridge crest morphology in the northern part

of the overlap basin region near the existing propagating limb. The eastward displacement of

the melt sill with respect to the bathymetric ridge crest of the self-decapitating limb suggests

that the rotational motion of the ridge crest is not primarily driven by the melt sill, but may

be associated with the rotational stress field in the overlap region.

3

Introduction

Our understanding of the evolution of the geometry of overlapping spreading centers,

commonly found on fast- and intermediate-spreading mid-ocean ridges (Macdonald and

Sempere, 1984; Macdonald et al., 1991), has hitherto been based on their morphological

expression at the seafloor. Crack propagation theory has been applied to explain the

formation of the curved shape of overlapping en-echelon ridge segments (e.g. Pollard and

Aydin, 1984; Baud and Reuschle, 1997). Numerical modeling has successfully described the

evolution of OSCs based on magnetic and bathymetric data (e.g. Wilson, 1990; Carbotte and

Macdonald, 1992; Cormier et al., 1996), suggesting that many OSCs are migrating relative to

their nearest transform offsets. Generally, one of the ridge segments is propagating, while

the other segment repeatedly cuts itself off from the main ridge axis and is rafted off to the

flank of the plate boundary. This process is known as self-decapitation (Macdonald et al.,

1987). Although crack propagation theory predicts the observed configuration of OSCs (e.g.

Macdonald and Sempere, 1984), it does not explain how these overlapping “cracks” are

formed at crustal levels. Furthermore, no crustal constraints have been placed on the

kinematic reconstruction of the evolution of ridge segments.

With extensive geophysical studies conducted in its vicinity (e.g. Carbotte and

Macdonald, 1992; Kent et al., 1993, 2000; Harding et al., 1993), the 9°03'N OSC on the East

Pacific Rise is one of the best places to investigate the crustal structure related to ridge

propagation and self-decapitation. The northern ridge segment of this OSC is propagating

southward, while the southern ridge segment has undergone repeated episodes of self-

decapitation, leaving relict basins and relict ridge segments to the west of the ridge axis

(Carbotte and Macdonald, 1992). We here use refraction and reflection seismic data to

4

construct a 3-D tomographic velocity model and a depth-converted image of the melt sills to

investigate the upper crustal construction at the OSC.

Method

Our 3-D tomographic analysis used 69,796 crustal first arrivals from seismograms collected

by an array of 19 ocean bottom seismometers (Fig. 1) in the ARAD experiment (Singh et al.,

1999). The maximum source-receiver offset is ~10 km and the mean pick uncertainty is 28

ms. We used a travel-time inversion algorithm that jointly minimizes travel-time misfits and

model roughness (Hobro et al., in press). The velocity model is densely parameterized with a

horizontal node spacing of 500 or 494 m and a vertical spacing of 195 m.

Normal moveout correction and CDP stacking (with trace-offset range of 0-3 km)

were applied to the 3-D seismic reflection data collected using a 3.1 km streamer with 124

channels. After two-pass Kirchhoff time migration, the resulting data were interpolated to a

final bin size of 25 m x 25 m with amplitude compensation and filtering applied. The two-

way travel-times of the horizons which were interpreted as the melt sills were picked. The

velocity model obtained from tomographic inversion was used for depth–converting the

horizons by integrating along vertical ray paths. Due to the complex local bathymetry, depth

conversion using vertical rays is superior to ray-tracing for producing a sharper image of the

reflectors (Tong, 2000). Details of the tomographic and reflection analyses can be found in

Tong (2000).

Results

The 4.5 km/s velocity contour, which generally coincides with the depths of sharpest change

in velocity gradient (Fig. 3: A2, B2 and C2), is chosen to represent the layer 2A/2B

5

boundary. This is compatible with definitions used in previous studies (e.g. Harding et al.,

1993). The shallowest points of the 4.5 km/s contour under the spreading limbs are

identified as the neovolcanic axis (Fig. 2c), which is a good proxy for the locations of highest

magmatism (Perfit and Chadwick, Jr., 1998). (The precise choice of velocity contour has little

effect on the results.) The lateral resolution of the velocity model is estimated as ~2.5-5 km,

and the uncertainty in the depth of the melt sill is about ±110 m (Tong, 2000). There are

significant differences in geometry between the depth-converted melt sills and those

observed on the time-migrated seismic sections (Fig. 3: A1–C1, A4–C4). This highlights the

importance of velocity control in the depth-conversion of reflectors. Compared with the

central parts of ridge segments well removed from major discontinuities (e.g. Toomey et al.,

1994; Hooft et al., 1997), a relatively thick layer 2A and deep melt sills are imaged in this

study. These results are consistent with those reported in previous investigations near major

ridge discontinuities (e.g. Tolstoy et al., 1997).

The asymmetric crustal structures beneath the limbs are likely to be related to their

evolutionary history. The thick low-velocity layer under the overlap basin and the tip of the

propagating eastern ridge crest may be caused by lateral injection of melt beyond the tip of

the melt sill (Fig. 3: C1), a mechanism proposed in similar geological settings elsewhere (e.g.

Cormier et al., 1996). The rotational effect of the stress field in the overlap region (Perram

and Macdonald, 1990) may lead to the inwardly curved ridge tip seen in the bathymetry and

the thick low-velocity layer 2A that covers the eastern ridge tip and overlap basin. The

“unrotated” plunging tip section of the propagating eastern melt sill may provide a reference

against which the extent of rotation can be measured. Although the thick layer 2A may

represent a frozen lava pond (Bazin et al., in press), it is also possible that the tensional

component of the rotation induces fractures, leading to the observed negative velocity

6

anomaly and the shallow layer with distinct lower velocity-gradients (Fig. 3: C2 and C3). The

relatively extensive region of low velocity-gradient layer 2B (Fig. 3: C2) may be caused by

fractures or by a positive thermal anomaly created by the propagating tip of the melt sill,

where a high-magnetisation zone (Fe-Ti basalts) is found (Sempere and Macdonald, 1986;

Bazin et al., in press).

By contrast, the high velocity-gradient region that characterizes layer 2A at the tip of

the western limb is localized under the ridge crest as it forms an isolated morphological

region (Fig. 3: A2). The northern part of the overlap basin to its east forms part of the

eastern limb system with its continuous crustal and melt sill constructions (Fig. 3: A2), and

the relict basins to its west display structures similar to those found under the central and

southern parts of the overlap basin (Fig. 3: A3–C3). Our results confirm the magmatically

depleted nature of the western ridge tip (Sempere and Macdonald, 1986). The thick layer 2A

at the western limb tip does not show the positive velocity anomaly observed above the

western melt sill (Fig. 3: A3 and C3). This positive velocity anomaly may reflect the thin on-

axis layer 2A, similar to those observed along mid-segment regions (e.g. Toomey et al., 1994).

The lack of this positive anomaly at the ridge tip may indicate the pattern of on-axis

hydrothermal alteration (Sohn et al., 1996), and possible faulting induced by the dwindling

melt supply. If the depth of the melt sill is related to the heat balance of the crust (e.g.

Henstock et al., 1993), the comparable depths of the melt sills may imply that the apparent

difference in melt supply between the limbs has little effect on their bulk thermal structure.

Implications

As the OSC is migrating southward, we may regard the along-axis variations of the

asymmetric melt sills and the upper crustal layers as snapshots detailing the stages of the

7

construction of the fuller and wider ridge crest that is observed away from major ridge

discontinuities. The advancement of the melt sill may be accompanied by lateral melt

injection at the ridge tip. As the OSC migrates further south, the plunging tip section of the

eastern melt sill may eventually acquire a geometry and dimensions similar to that of the

same melt sill imaged further north. A shallower, elongated region on the melt sill may

eventually appear, marking the location of the neovolcanic axis. The effect of the rotational

stress field on the upper, brittle part of the crust is significant, resulting in the “rotated”

extrusives observed beyond the ridge tip.

The melt supply from the melt sill beneath the overlap basin gradually transforms the

northern part of the basin region into part of the fuller ridge crest that extends from the

existing curved limb. The development of a new neovolcanic axis under the overlap basin

may lead to the abandonment of the current neovolcanic axis under the central part of the

eastern limb, where the neovolcanic axis is closely related to the similarly curved ridge crest.

The observed eastward displacement of the western melt sill from the bathymetric

high of the ridge axis may imply that the surficial expression of self-decapitation is

influenced primarily by the rotational effect of the stress field in the overlap basin, rather

than being driven by the melt sill. Equivalently, the geometry and location of the melt sill

may lag behind the anti-clockwise rotational motion of the western limb that forms the initial

stages of self-decapitation.

In conclusion, the asymmetric upper crustal features under the two limbs presented

in this study are closely related to the OSC evolution. Change in plate motion is suggested as

one of the causes of plate boundary re-orientation (e.g. Pockalny et al., 1997), and our

constraints on the upper crustal evolution at the OSC offer new insights into the

investigation of its underlying mechanisms.

8

Acknowledgement

We thank the Natural Environment Research Council (U.K.), National Science Foundation,

and Newton Trust for funding the ARAD experiment. Department of Earth Sciences,

University of Cambridge Contribution no. ES.6413.

References cited

Baud P. and Reuschle T., 1997, A theoretical approach to the propagation of interacting

cracks: Geophysical Journal International, v. 130, p. 460-468.

Bazin, S., A. Harding, G. Kent, J. Orcutt, C. H. Tong, J. Gee, S. Singh, P. Barton, M. Sinha,

R. White, and H. Van Avendonk, in press, 3-D crustal emplacement at the 9°03'N

overlapping spreading center on the East Pacific Rise I: Correlations between magnetization

and tomographic imaging: Journal of Geophysical Research.

Carbotte, S. and Macdonald, K., 1992, East Pacific Rise 8°-10°30'N: Evolution of Ridge

Segments and Discontinuities From SeaMARC II and Three-Dimensional Magnetic Studies:

Journal of Geophysical Research, v. 97, p. 6959-6982.

Cormier M.-H., D. S. Scheirer, and K. C. Macdonald, 1996, Evolution of the East Pacific

Rise at 16°-19°S since 5 Ma: Bisection of overlapping spreading centers by new, rapidly

propagating ridge segments: Marine Geophysical Researches, v. 18, p. 53-84.

9

Harding, A. J., G. M. Kent, and J. A. Orcutt, 1993, A multichannel seismic investigation of

the upper crustal structure of at 9°N on the East Pacific Rise: Implications for crustal

accretion: Journal of Geophysical Research, v. 98, p. 13925-13944.

Henstock, T. J., A. W. Woods, and R. S. White, 1993, The accretion of oceanic crust by

episodic sill intrusion: Journal of Geophysical Research, v. 98, p. 4143-4161.

Hobro, J. W. D., S. C. Singh, and T. A. Minshull, in press, Three-dimensional inversion of

combined seismic reflection and refraction data: Geophysical Journal International.

Hooft, E. E. E., R. S. Detrick, and G. M. Kent, 1997, Seismic structure and indicators of

magma budget along the Southern East Pacific Rise: Journal of Geophysical Research, v.

102, p. 27319-27340.

Kent, G. M., S. C. Singh, A. J. Harding, M. C. Sinha, J. A. Orcutt, P. J. Barton, R. S. White, S.

Bazin, C. H. Tong, and J. W. Pye, 2000, Evidence from three-dimensional seismic reflectivity

images for enhanced melt supply beneath mid-ocean-ridge discontinuities: Nature, v. 406, p.

614-618.

Kent, G. M., A. J. Harding, and J. A. Orcutt, 1993, Distribution of magma beneath the East

Pacific Rise near the 9°03'N overlapping spreading center from forward modeling of

common depth point data: Journal of Geophysical Research, v. 98, p. 13971-13996.

10

Macdonald, K. and Sempere, J.-C., 1984, East Pacific Rise from Siqueiros to Orzco Fracture

Zones: Along-strike continuity of axial neovolcanic zone and structure and evolution of

overlapping spreading centers: Journal of Geophysical Research, v. 89, p 6049-6069.

Macdonald, K. C., J.-C. Sempere, P. J Fox, and R. Tyce, 1987, Tectonic evolution of ridge-

axis discontinuities by the meeting, linking, or self-decapitation of neighboring ridge

segments: Geology, v. 15, p. 993-997.

Macdonald, K. C., D. S. Scheirer, and S. M. Carbotte, 1991, Mid-ocean ridges:

Discontinuities, segments and giant cracks: Science, v. 253, p. 986-994.

Perfit, M. R. and Chadwick, Jr.,W. W., 1998, Magmatism at mid-ocean ridges: constraints

from volcanological and geochemical investigations: Faulting and Magmatism at mid-ocean

ridges, Geophysical Monograph 106, American Geophysical Union, p. 59-115.

Perram, L. J. and Macdonald, K. C., 1990, A one-million-year history of the 11°45'N East

Pacific Rise discontinuity: Journal of Geophysical Research, v. 95, p. 21363-21381.

Pockalny, R. A., P. J. Fox, D. J. Fornari, K. C. Macdonald, and M. R. Perfit, 1997, Tectonic

reconstruction of the Clipperton and Siqueiros Fracture Zones: Evidence and consequences

of plate motion change for the last 3 Myr: Journal of Geophysical Research, v. 102, p 3167-

3181.

11

Pollard, D. D. and Aydin, A., 1984, Propagation and linkage of oceanic ridge segments:

Journal of Geophysical Research, v. 89, p. 10017-10028.

Sempere, J.-C. and Macdonald, K. C., 1986, Deep-tow studies of the overlapping spreading

centers at 9°03'N on the East Pacific Rise: Tectonics, v. 5, 881-900.

Singh, S. C., M. C. Sinha, A. J. Harding, G. M. Kent, P. J. Barton, J. A. Orcutt, R. S. White,

and R. W. Hobbs, 1999, Preliminary results from mid-ocean ridge three-dimensional seismic

reflection survey: EOS, v. 80, p. 181, 185.

Sohn, R. A., S. C. Webb, J. A. Hildebrand, and B. D Cornuelle, 1997, Three-dimensional

tomographic velocity structure of upper crust, CoAxial segment, Juan de Fuca Ridge:

Implications for on-axis evolution and hydrothermal circulation: Journal of Geophysical

Research, v. 102, p 17679-17695.

Tolstoy, M., A. J. Harding, and J. A. Orcutt, 1997, Deepening of the axial magma chamber

on the southern East Pacific Rise toward the Garrett Fracture Zone: Journal of Geophysical

Research, v. 102, p. 3097-3108.

Tong, C. H., 2000, Three-dimensional tomographic study of the 9°03'N overlapping

spreading centre on the East Pacific Rise, Ph.D. Dissertation, University of Cambridge, pp.

87, 157.

12

Toomey, D. R., S. C. Solomon, and G. M. Purdy, 1994, Tomographic imaging of the shallow

crustal structure of the East Pacific Rise at 9°30'N: Journal of Geophysical Research, v. 99,

24135-24157.

Wilson, D. S., 1990, Kinematics of overlapping rift propagation with cyclic rift failure: Earth

and Planetary Science Letters, 96(1990), p. 384-392.

Figure captions

Figure 1. Location of the 9°03'N overlapping spreading center and the data acquisition

geometry. Inset indicates location of study area. 160,000+ shots were fired within the box

shown in solid lines and recorded by 19 ocean bottom seismometers (stars). Broken lines

denote limits of the 3-D reflection survey. The origin of the local coordinate system is

indicated by orthogonal arrows, which correspond to the x- and y- axes. Cross-sections in

Fig. 3 are located by lines A, B and C.

Figure 2. a) Depth variation of the melt sills below the sea surface, and b) depth below

seafloor. The ridge crests of the eastern limb (EL) and western limb (WL) are indicated by

the solid contour of 2700 m. The overlap basin (OB) and relict basins (RB) (Carbotte and

Macdonald, 1992) are shown by the dotted 2900 m contour. Cross-sections in Fig. 3 are

located by lines A, B and C. c) Detailed bathymetry of the OSC. Heavy black lines mark

limits of the melt sills; thick grey lines show locations of the neovolcanic axes (see text for

details). Note that the neovolcanic axis and the melt sill under the western limb are offset

toward the eastern side of the ridge crest, and the neovolcanic zone is shifted toward the

central part of the ridge crest, where the melt sill vanishes (Fig. 2c). The other melt sill

13

extends to the northern part of the overlap basin from the eastern limb and consists of an

elongated shallow region near the outward edge of the ridge crest (Fig. 2a). Fig. 2b shows a

similarly elongated shallow structure on the western side of the melt sill under the overlap

basin. Noticeable increase in the depth of the melt sill in this region is restricted to its

western edge. A gradual decrease in width and an increase in depth of the melt sill are

observed toward its tip, where the neovolcanic axis deviates from the curved ridge crest (Fig.

2a).

Figure 3. A1, A2, A3 show the velocity, velocity-gradient, and velocity anomaly variations of

line A (Figs. 1 and 2). Regions that are not constrained by rays are shown in lighter shades.

Velocity-gradient at each velocity node is calculated by the velocity difference over the

vertical interval of the node spacing divided by the node spacing. Velocity anomaly plot

shows velocity deviations from the mean velocity at a given depth below seafloor. A4 shows

the time-migrated seismic section of line A with the picked horizon (melt sill) indicated by

the red line a. TWTT is two-way travel-time. The depth-converted horizon is represented by

the black lines in A1, A2 and A3. The red and blue lines in all plots show the locations of the

ridge crests of the western limb and eastern limb, respectively, based on the 2700 m contour.

B1–B4, and C1–C4 show the results corresponding to lines B and C, respectively. Note that

a thicker layer 2A, characterized by negative near-seafloor velocity anomalies and high

velocity-gradients, is observed in the tip region of the western limb (Fig. 3: A2 and A3). A

thicker layer 2A that extends from the overlap basin to the western part of the ridge crest is

imaged in the tip region of the eastern limb (Fig. 3: C1). A relatively extensive region of low

velocity-gradients in layer 2B is also found there (Fig. 3: C2). Also note the continuity in the

characteristics of upper crustal structures in the study area: a continuous layer 2A underlies

14

the northern part of the overlap basin and the eastern limb (Fig. 3: A2). By contrast, the

central and southern parts of the overlap basin separate the continuous high velocity-

gradient layers found under the adjacent ridge crests and off-axis regions (Fig. 3: B2 and C2).

A negative velocity anomaly is observed under the relict basins and the central and southern

parts of the overlap basin (Fig. 3: A3, B3 and C3).