This is an electronic version of "Taylor, B., A. Goodliffe, F. Martinez & R. Hey, Continental rifting and initial sea-floor spreading in the Woodlark basin, Nature, 374 (6522), 534-537, 1995"

Continental rifting and initial seafloor spreading in the Woodlark basin

Click HERE for full figure captions

Our understanding of the processes by which continents rift and seafloor spreading initiates is derived primarily from studies either of old passive margins and oceanic crust or of young regions of intra-continental extension where spreading has not yet started. It has been thought that continental rifting ceases when seafloor spreading begins (1,2), that oceanic fracture zones develop from transfer or transform faults within continental rifts (3,4), and that linear magnetic anomalies correlate with the onset of seafloor spreading during times of magnetic reversals (5,6). Here we present a marine geophysical survey of one of the few presently active examples of continental rifting and spreading initiation, the western Woodlark basin/Papuan peninsula region of New Guinea, which shows that in detail these assumptions do not hold. The data confirm models of the rifting to spreading transition that invoke both ridge propagation and nucleation of discrete spreading cells (7-10), and provide and unambiguous example of a spreading CENTER reorienting by synchronous jumping rather than propagation.

During the last 6 Myr the formerly contiguous, eastward extensions of the Papuan Peninsula (the Woodlark and Pocklington Rises) were stretched, separated and subsided as the spreading centre in the Woodlark basin extended westward between them (11-13) into orogenically thickened crust that is 25-50 km thick and rises 1-3 km above sea level (14) (Fig. 1). Structures associated with the continental rifting include numerous full and half grabens (15), and large metamorphic core complexes on the D'Entrecasteaux islands and the Papuan Peninsula (16,17). Normal movement along a ductile mylonitic shear zone (1-2 km thick) exposed on the islands resulted in the rapid uplift of deep metamorphic rocks (from ~30 km depth (7-11 kbar) in 4 Myr) and the juxtaposition of unmetamorphosed cover rocks (18-21). Granodioritic intrusion then focused uplift on several domes offset by strike-slip faults, forming topography up to 2.5 km. The extension is accompanied by peralkaline rhyolitic volcanism in the D'Entrecasteaux islands (22-24) and by crustal tensional seismicity as far west as 148E (25). Earthquake source parameters (25) (Fig. 2a) indicate that low angle (10-25 degrees) normal faulting is active in the region of incipient continental separation.

In April-May 1993 we conducted a geophysical and HAWAII MR1 sidescan survey of the western Woodlark basin (26). The sidescan survey provided almost total coverage bathymetry and acoustic imagery of the basin and its margins inside the bounding reefs and islands (Fig. 2a,b). Together with a magnetization map (Fig. 2c) that we derived from a 3-D inversion of the magnetic anomalies and bathymetry, as well as seismic reflection and gravity data shown elsewhere (26,27), these data provide an exceptionally clear view of the transition from intra-continental rifting to seafloor spreading.

The V-shaped Woodlark basin shallows towards its western apex (Fig. 2a). Its neovolcanic zone, defined by the strong acoustic backscatter (Fig. 2b) within the Brunhes magnetic anomaly (Fig. 2c), extends westward to the spreading tip at 9.8S, 151.7E where it abuts Moresby Seamount, a crustal block that dredging (28) and geophysics indicates may be a metamorphic core complex. The seafloor spreading centre in our survey area is divided into three first order segments by the 50-km-long Moresby transform fault at 154 12'E and by a 30 km non-transform offset at 152 50'E (Figs. 1, 2). The first (western) segment is subdivided into three overlapping second order segments arranged in a right-stepping en echelon pattern. It is characterized by numerous small seamounts and axial ridges with low relief. In cross-section, the second (central) segment has a 3 to 4-km-wide rift valley within which there is a small axial ridge. It is subdivided by a third order (1 km) left-stepping, non-transform offset at 153 30'E. The third (eastern) segment, like others further east, trends ENE and has an axial valley that is 6-7-km-wide and a 3.5- to 4.0-km deep. In along axis profile, each first order spreading segment is shallowest in the CENTER and asymmetrically deepest at the ends (Fig. 1). Several young off-axis seamounts occur south of the CENTER of the second segment (Fig. 2). The oceanic crust has a characteristic abyssal hill fabric, with numerous superposed seamounts of varying sizes and on- as well as off-axis origins. Notwithstanding complexities in the spreading history, sea floor along a given fabric element is isochronous. For example, adjacent to segment two, magnetic anomaly 2 corresponds to an abyssal hill on both sides of the axis (Fig. 2c).

A strong seismic reflector, correlated to the low-angle fault inferred from earthquake studies, dips 25 N and has an emergent segment along the northern flank of Moresby Seamount (27,29) (Fig. 3). Basement fault blocks overlain by only minor ponded sediments constitute the southern margin of the basin WNW of Misima island. In contrast the northern margin (above the low angle fault) has a down-flexed pre-rift sedimentary basin and basement sequence, unconformably onlapped by syn-rift sediments fed south by submarine channels, and cut by high-angle faults with a zig-zag pattern in plan view (27) (Figs.2, 3). Seismic reflection data indicate a remarkably sharp (<5 km wide at the surface) transition from oceanic crust to rifted crust in most areas (27,29), as has been observed in some other regions (5,30). There are no dipping reflector sequences indicative of excessive lava production and high degrees of mantle partial melting. (There are, however, numerous small volcanoes, a few kilometres in diameter, often erupted along margin faults.) Indeed, basalts from segment one have Na2O (3.1 wt% at 8.0 wt% MgO) and other geochemical characteristics31 indicative of low degrees of partial melting. This area would be classified 'non-volcanic' in the two-part division of passive margin types (32,33).

To test the applicability of various models of spreading propagation into continental lithosphere(7-10,34), we have used the distinctive seafloor fabric and seismic reflection characteristics of oceanic versus rifted continental crust to determine the boundary between them. This boundary was derived independent of magnetization information, but is plotted on the seafloor magnetization map (Fig. 2c) for comparison. Note that the landward boundary of oceanic crust and the magnetization contours are locally discordant. The distribution of oceanic crust reveals a segment due south of Woodlark island near 9 50'S that is surrounded by rifted crust. Like the westernmost neovolcanic subsegment, this isolated oceanic segment is direct evidence for nucleation of discrete spreading cells (and subsequent ridge jumping) within stretched continental crust(9,10). In other cases, the rifting to spreading transition involves ridge propagation (7-9). Evidence includes the V-shaped margins (pseudofaults) east of Moresby transform, against which oceanic crust of successively younger age abuts (Fig. 2c). A second example is the western V-shaped tip of the Brunhes anomaly for segment two: the propagator sliced off a sliver of rifted crust that is now nearly surrounded by oceanic crust. We do not find evidence for periodically spaced spreading CENTERs and 'punctiform' spreading initiation (34). Our data are best fit by a combined spreading centre nucleation and propagation model (9).

The Moresby transform fault at 154.2E is the only transform offset of the spreading axis west of 155E. It formed just prior to 1 Myr and the Jaramillo magnetic anomaly. Significantly, the transform does not extend to the southern continental margin and the aulacogen northeast of Rossel island (Fig. 2). Rather, it linked offset spreading segments that, at the time of magnetic anomaly 2 (~1.9 Myr) were overlapping but separated by rifted continental crust. The oceanic structural geometry was not inherited directly from the earlier rift geometry (35). The initial oceanic crust in the western Woodlark basin has no transform faults and the one transform that did develop, after ~0.9 m.y. of spreading, cut through intervening rifted crust to link overlapping spreading segments. The ability of the offset spreading segments to open initially without forming connecting transform faults appears to be related to the fact that rifting of the conjugate margins continues after spreading has started between them. We infer from two observations that extension does not immediately localize at the ridge axis, but instead is distributed between margin rifting and seafloor spreading. The first observation is that spreading on segment one overlaps in space and time with seismogenic faulting on the continental margins west of 153E (Fig. 2a). The second is that the seafloor fabric and magnetic anomalies that curve towards the spreading axis on both sides of segment three (Fig. 2) require non-rigid reconstructions of the adjacent margins, that is, that spreading propagation into continental crust is accommodated by progressively less crustal extension, instead of by the progressive cessation of spreading on an adjacent segment as in oceanic crust (7-9). The overlap in time between initiation of spreading and cessation of rifting along the margin can be estimated from both sets of observations to be 0.8-1 Myr. Spreading initiated west of 153E by the beginning of the Brunhes epoch (0.78 Myr) and seismogenic faulting of the margins continues today. Adjacent to segment three, the age difference between curved and linear conjugate sea floor fabrics at a given longitude is ~1 Myr.

We recognize several instances where curvilinear fault systems up to 100 km long are associated with magnetization boundaries landward of oceanic crust, particularly along the northern margin (Fig. 2c). Somewhat similar magnetic anomalies have been observed in rifted arc crust of the northern Mariana trough where their formation was ascribed to fault-controlled intrusions (36). An alternative explanation, applicable in regions of low-angle detachments such as the Woodlark basin, is that the magnetization contrasts could form between hanging-wall rocks and footwall rocks that acquire a magnetization as they are exhumed from mid-crustal levels and cool through their Curie temperature. Both explanations provide mechanisms for magnetic lineations to form landward of the oldest oceanic crust. Therefore caution must be exercised in attempting to locate the limits of old oceanic crust where the basement structure and crustal fabric are poorly imaged beneath sediment cover.

Our survey also provides clear evidence that a spreading centre can reorient by synchronous jumping rather than propagation. All segments other than the westernmost subsegment reoriented, as evidenced by their present trend being oblique to the Brunhes/Matuyama crustal boundary and sea-floor fabric within it (Fig. 2). A subsequent survey (37) confirmed the existence of similarly rotated spreading segments in the rest of the eastern Woodlark basin (Fig. 1). The 500-km-long Woodlark basin spreading axis changed orientation simultaneously, not by propagation, ~100 kyr ago. In addition to the gravitational collapse of orogenically overthickened lithosphere, causes for the opening of the Woodlark basin have been sought in the 'pull' of the Solomon Sea lithosphere which is being both subducted to the north at the New Britain Trench and dragged northwestwards by the overriding Pacific plate (11,13). Recent changes in the resultant of these additive forces, associated with spreading ridge and arc-continent collisions to the east and north respectively (38,39), may be responsible for the ridge reorientation.

References

1. Falvey, D. A. Aust. Petrol. explor. Ass. J.. 14, 95-106 (1974).
2. Karner, G. D., Driscoll, N. W. & Weissel, J. K. Earth planet. Sci. Lett. 114, 397-416 (1993).
3. Klitgord, K. D. & Behrendt, J. C. in Geological and Geophysical Investigations of Continental Margins (eds Watkins, J. S., Montadert, L. & Dickerson, P. W.) 85-112 (Am. Assoc. Petrol. Geol., Tulsa, Oklahoma, 1979).
4. Lister, G. S., Etheridge, M. A. & Symonds, P. A. Geology 14, 246-250 & 891-892 (1986).
5. Montadert, L., de Charpal, O., Roberts, D., Guennoc, P. & Sibuet, J.-C. in Deep Drilling Results in the Atlantic Ocean: Continental Margins and Paleoenvironment (eds Talwani, M., Hay, W. & Ryan, W. B. F.) 154-186 (Am. Geophys. Un., Washington, D.C., 1979).
6. Tucholke, B. E., Austin, J. A. & Uchipi, E. in Extensional Tectonics and Stratigraphy of the North Atlantic Margins (eds Tankard, A. J. & Balkwill, H. R.) 247-263 (Am. Assoc. Petrol. Geol., Tulsa, Oklahoma, 1979).
7. Hey, R. N., Duennebier, F. K. & Morgan, W. J. J. geophys. Res. 85, 2647-2658 (1980).
8. Courtillot, V. Tectonics 1, 239-250 (1982).
9. Martin, A. K. Tectonics 3, 611-617 (1984).
10. Cochran, J. R. & Martinez, F. Tectonophysics 153, 25-53 (1988).
11. Weissel, J. K., Taylor, B. & Karner, G. D. Tectonophysics 87, 253-277 (1982).
12. Taylor, B. in Marine Geology, Geophysics and Geochemistry of the Woodlark basin - Solomon Islands (eds Taylor, B. & Exon, N. F.) 25-48 (Circum-Pacific Council for Energy and Mineral Resources, Houston, Texas, 1987).
13. Benes, V., Scott, S. D. & Binns, R. A. J. geophys. Res. 99, 4439-4455 (1994).
14. Finlayson, D. M., Muirhead, K. J., Webb, J. B., Gibson, G., Furumoto, A. S., Cooke, R. J. S. & Russel, A. J. Geophys. J. R. astr. Soc. 29, 245-253 (1976).
15. Smith, I. E. & Simpson, C. J. Bull. Bur. Min. Resour. Geol. Geophys. Aust. 125, 29-35 (1972).
16. Davies, H. L. Am. J. Sci. 280-A, 171-191 (1980).
17. Davies, H. L. & Warren, R. G. Tectonics 7, 1-21 (1988).
18. Davies, H. L. & Warren, R. G. Contr. Miner. Petrol. 112, 463-474 (1992).
19. Hill, E. J., Baldwin, S. L. & Lister, G. S. Geology 20, 907-910 (1992).
20. Hill, E. J. & Baldwin, S. L. J. metamorph. Geol. 11, 261-277 (1993).
21. Baldwin, S. L., Lister, G. S., Hill, E. J., Foster, D. A. & McDougall, I. Tectonics 12, 611-628 (1993).
22. Smith, I. E. M. & Milsom, J. S. in Marginal Basin Geology (eds Kokelaar, B. P. & Howells, M. F.) 163-171 (Geol. Soc. Spec. Publ., 1984).
23. Hegner, E. & Smith, I. E. M. Chem. Geol. 97, 233-249 (1992).
24. Stolz, A. J., Davies, G. R., Crawford, A. J. & Smith, I. E. M. Miner. Petr 47, 103-126 (1993).
25. Abers, G. A. Geology 19, 1205-1208 (1991).
26. Taylor, B., Martinez, F., Hey, R. & Goodliffe, A. (abstr.) Eos 74, 606 (1993).
27. Goodliffe, A., Taylor, B., Hey, R. & Martinez, F. (abstr.) Eos 74, 606 (1993).
28. Binns, R. A., Scott, S. D. et al. Proc. Pacif. Rim Congr. 87, 531-535 (1987).
29. Mutter, J. C., Mutter, C. Z., Abers, G. & Fang, J. (abstr.) Eos 74, 412 (1993).
30. Taylor, B. & Hayes, D. E. In The Tectonic and Geologic Evolution of Southast Asian Seas and Islands: Part 2 (Hayes, D. E.) 23-56 (Am. Geophys. Un., Washington, D.C., 1983).
31. Binns, R. A. & Whitford, D. J. Proc. Pacif. Rim Congr. 87, 525-534 (1987).
32. White, R. S. Eos 74, 58 (1993).
33. Mutter, J. C. Nature 364, 393-394 (1993).
34. Bonatti, E. Nature 316, 33-37 (1985).
35. Bosworth, W. Geology 14, 890-891 (1986).
36. Martinez, F., Fryer, P., Baker, N.A. & Yamazaki, T. J. geophys. Res. 100, 3807-3827, (1995).
37. Deep Ocean Resources Co., Southe Pacific Seafloor Atlas, sheets 10-11 (Jap. int. Coop. Ag./Metal Min. Ag. Jap./S. Pacific app. Geosci. Comm., Tokyo, 1995)
38. Crook, K. A. W. & Taylor, B. Mar. geophys. Res. 16, 65-89 (1994).
39. Abbott, L. D., Silver, E. A. & Galewsky, J. Tectonics 13, 1007-1034 (1994).

Acknowledgements

This research benefited greatly from the dedicated support of all onboard R/V Moana Wave cruise 9304, the services of the Hawaii Mapping Research Group, and the GMT software of Pål Wessel and Walter Smith. This work was supported by the US NSF.

Figure Captions

FIG. 1 Major physiographic features and active plate boundaries in the Papua New Guinea - Solomon Islands region. The stippled area encloses oceanic crust formed during the Brunhes chron, at spreading rates labelled in mm/yr. MT and ST = Moresby and Simbo Transform Faults, respectively; DE = D'Entrecasteaux Is. The depth profile at the bottom is along the axis of the Woodlark basin spreading CENTER, with the three first-order spreading segments numbered.

FIG. 2 Bathymetry, seismicity, acoustic imagery, and magnetization data for the western Woodlark basin. (a) INT= International Seismological Centre; PP = Papuan Peninsula; G = Goodenough island; F = Fergusson island; N = Normanby island; W = Woodlark island; M = Misima island; T = Tagula island, R = Rossel island; MS = Moresby Seamount; MT = Moresby Transform Fault. (c) Magnetization lineations corresponding to the Brunhes (B), Jaramillo (J) and magnetic anomalies 2 and 2' are labelled and the landward boundary of oceanic crust is outlined. The line spacing for the survey was 5 nautical miles, adequate to achieve complete coverage of the sidescan acoustic imagery and bathymetry as well as to interpolate the magnetic anomaly field.

FIG. 3 True scale meridional section through Moresby Seamount (27,29).