The timing of, and vertical motions associated with, the onset of rifting are important parameters for models of continental rifting processes. In many areas, however, including the North Atlantic, these critical aspects are only broadly constrained (e.g., Wilson et al., 2001). It is important, therefore, to critically appraise the data bearing on this issue in the case of the western Woodlark Basin. Some framework information comes from the opening history of the Woodlark Basin, which we know had begun spreading by 6 Ma in the east and has since been propagating westward into rifting lithosphere (Taylor et al., 1999). Additional data comes from the metamorphic core complex on Misima Island, where high-grade metamorphic rocks occur beneath a low-angle (23°N) normal fault and low-grade cover rocks. Preliminary thermochronologic data indicate granite intrusion and deformation occurred at ~8 Ma, with subsequent hydrothermal alteration and ore mineralization from 4.0 to 3.2 Ma (Baldwin et al., 2000).
Ascribing the 8.4-Ma unconformity to rift onset is consistent with existing information, but it is not a unique interpretation of the margin stratigraphy and vertical motions, given the possible subduction and known eustatic effects discussed above. Indeed, the late Miocene subsidence and depositional history data from sites drilled on the Woodlark Rise alone are equivocal on this issue. Nevertheless, seismic horizons dated at these sites can be correlated into the Moresby rift and confirm that rift faulting began by ~8 Ma (see below).
A distinctive paleogeography and depositional facies characterizes the Trobriand forearc and the Woodlark Rise in the late Miocene above the 8.4-Ma unconformity. Everywhere in the forearc was either shallow-water (<50 m) or low-relief islands (such as the ~300-m basement high drilled at Site 1118, where lateritic paleosols are present in the interstices between dolerite and basalt clasts in a >50-m-thick conglomerate). The stratigraphic record at Sites 1109 and 1115 and Goodenough 1 and Nubiam 1 wells reflects a paralic region of islands, lagoons, algal reefs, swamps, and deltas fed by rivers draining upland to the south (Shipboard Scientific Party, 1999; Francis et al., 1987; Robertson et al., 2001). Sediment isopachs show that Site 1109 was drilled near the axis and Site 1118 is on the flank of one such paleochannel (Goodliffe et al., 1999). Conglomerates of basalt (at Site 1115) and dolerite and basalt (at Site 1109) that immediately overlie the unconformity were deposited in a fluvial to swampy setting (Shipboard Scientific Party, 1999). The overlying lagoonal facies is brackish at Site 1109 (with shell, plant, and wood fragments common) to initially relatively enclosed and then more open marine at Site 1115 (with abundant shell fragments) (Shipboard Scientific Party, 1999). Chemically distinctive claystones and siltstones (with higher Ti and Fe2O3 and lower K than calc-alkaline-sourced sediments) are compatible with derivation from an exposed and tropically weathered basic igneous basement (Robertson and Sharp, this volume).
What caused the initial accommodation space above the 8.4-Ma unconformity? One factor was the net ~120-m sea level rise between 8 and 5 Ma (Haq et al., 1988). Another was possibly residual thermal subsidence following earlier forearc basin extension. Seismic profiles show that lensoid stratal packages that onlap the forearc basin edges persist between Sites 1109 and 1115 and farther west until ~3.2 Ma (Fig. F8) (Goodliffe et al., this volume; Fang, 2000). This sag phase of shelf deposition can be modeled as thermal subsidence from the previous extension, and thus is not a reliable indicator of contemporary rifting.
The telltale sign of the onset of rifting is not to be found on the northern margin (the Woodlark Rise) but in the early graben such as Moresby rift, specifically in the earliest sediments that fill the space created by the graben-bounding faults. None of the Leg 180 sites targeted the earliest rift sediments because they occur at ~2500 mbsf in Moresby rift and at >1500 mbsf on Moresby Seamount (Fig. F8). We can, however, correlate the seismic horizons dated at Sites 1109, 1115, and 1118 around the regional grid of seismic profiles and into the Moresby rift. This correlation shows that the early rift sediments are older than 5.5 Ma but younger than the prograding clinoforms beneath the 8.4-Ma unconformity. This unconformity may therefore reasonably be correlated with rift onset, even though we recognize that it has a eustatic component.
Note that Moresby Seamount was not a topographic high in the late Miocene. It was part of a wide graben system that included what is now Moresby rift as well as the basin to the south of Moreseby Seamount (Figs. F3, F8). All the master faults that currently bound these basins were formed early in the rift history. Some differential sedimentation and tilting, characteristic of half-graben, are evidenced in the earliest rifts and also in the recent (sediment starved) phases of the deformation. At Sites 1114 and 1117 we drilled the south- and north-dipping faults, respectively, that initially delimited a horst block within the wider rift. The basement high drilled at Site 1118 was just north of the graben. Although there were several rifts active farther south at this time, there were none farther north except for a narrow half-graben near ~9°S that continued to reactivate a forearc structure (Fig. F3).