Continental breakup requires thinning the lithospheric mantle and the crust to the point where seafloor spreading can initiate. Constraints on how this thinning has been achieved in the western Woodlark Basin, where Moresby rift is at the point of breakup immediately adjacent to a spreading center (Figs. F1, F2, F3), are provided by measurements of crustal structure, strain rate, thermal regime, and upper crustal faulting and subsidence. These data require that a major component of crustal thinning results from flow of the lower crust, that this flow is synrift, and that it may be activated more by fluids than temperature, as discussed below.
That the crust has been thinned is confirmed by a seismic receiver function study that indicates that the crust is 32-42 km thick beneath the Papuan Peninsula and northern Woodlark Rise, but only 20-26 km thick beneath the Goodenough Basin and D'Entrecasteaux Islands (Ferris et al., 2000). The crust thins further to an average thickness of 15-20 km flanking Moresby rift (Zelt et al., 2001).
Based on the Woodlark Basin opening history, strain rates are estimated to have averaged 10-14/s at 151.7°E and could have been one or two orders of magnitude faster when focused on a particular rift (such as immediately preceding breakup) (Taylor et al., 1999). Rifting in the western Woodlark Basin is as fast or faster than anywhere else on Earth today (Abers, 2001) and at least three times as fast as in the Cretaceous separation of Newfoundland from Iberia, for example (Wilson et al., 2001).
In situ measurements of thermal gradients combined with shipboard core measurements of thermal conductivities revealed low heat flow at the northern margin Sites 1115 and 1109 (~30 mW/m2) that increased rapidly at the rift flank Site 1118 (60 mW/m2) to relatively high values in Moresby rift Sites 1111 and 1108 (95-100 mW/m2). These values are consistent with the organic petrology and geochemistry results discussed above. They were confirmed and significantly extended by a pogo-probe heat flow transect, >100km long, that revealed a quasisymmetric heat flow anomaly peaking at ~250 mW/m2 centered over the basin south of Moresby Seamount and due west of the second spreading segment (Figs. F1, F2, F3) (Martinez et al., 2000; Goodliffe et al., 2000). The heat flow high (where values exceed 100 mW/m2) is 30-40 km wide and tapers southward to values of 40-50 mW/m2 on Pocklington Rise (Martinez et al., 2000; Goodliffe et al., 2000).
The locus of breakup and rifting is presently focused on Moresby rift and during the Brunhes Chron included the basin south of Moresby Seamount where the thermal anomaly peaks. Regionally, this location is along the volcanic front of the Trobriand arc, which explains the cooler thermal regime to the north (forearc) than to the south (backarc) (Fig. F3). The reconstructions in Figure F8 summarize the structural and stratigraphic constraints on how the upper crust has stretched in the region of the Leg 180 drilling transect. MCS profiles reveal that the upper crust of the northern margin (Woodlark Rise) has experienced very little extension since rift onset (~1% strain) (e.g., Goodliffe et al., 1999, this volume; Fang, 2000). The small-offset normal faults there occurred late in the rifting history (during the Quaternary) accompanying the long-wavelength downflexing of the margin toward Moresby rift. Nevertheless, the basement, which was near sea level at rift onset at 8.4 Ma, has subsided to >3 km south of Site 1109, equivalent to nearly 3 km of tectonic subsidence if the ~800 m of sediments were to be unloaded (Fig. F8). Simple isostatic calculations show that this requires ~10 km of crustal thinning, which equates to 25%-33% strain for crust initially 40-30 km thick. The total strain since 8.4 Ma over the section shown in Figure F8 is ~33%. Only ~1% strain is evidenced north of Moresby rift, even though the crust may be thinned to 15-20 km (Zelt et al., 2001). How was the long-wavelength subsidence and crustal thinning, which is beyond the influence of faults bounding Moresby rift, achieved in the absence of significant upper crust brittle deformation?
This is another example of the "upper-plate paradox" in which many conjugate rifted margins show evidence of large regional subsidence with little attendant brittle deformation (Driscoll and Karner, 1998). The significant difference in this example is that we know that this phenomenon occurs synrift, not just postbreakup. Lower crustal ductile extension can explain this paradox but, for the Woodlark Rise, we also know that the ductile flow is not thermally enabled. The northern margin is cold (~30 mW/m2), having been chilled by the southward subduction of the Solomon Sea plate at the Trobriand Trough. We infer that it is the presence of fluids in the crust (whether derived from Neogene subduction or inherited from the Paleogene ophiolites and subduction) that allows the lower crust to flow out from under the northern margin, causing the northern margin to subside synrift. The lower crustal flow would thus be on a more regional scale than the upper crustal brittle deformation, which is presently focused at the Moresby rift and normal fault.