Seismic profiles show that a thick sequence of sediments filled the rift basins in the late Miocene and early Pliocene (Fig. F8) (e.g., Goodliffe et al., 1999), though Sites 1108, 1114, and 1116 did not penetrate to these levels. On the northern margin, shallow-marine mixed carbonate-siliciclastic sediments from Sites 1109 and 1115 record gradual subsidence of shelf basins in the early Pliocene (Shipboard Scientific Party, 1999). The northward-flowing Miocene drainage was reversed as the margin thinned and subsided southward toward the active rifts. The transition at ~50 mbsl from inner to outer neritic paleowater depths occurred first at Site 1109 (~4.9 Ma) and slightly later at Site 1115 (~4.6 Ma) (Fig. F7).
In the early Pliocene, however, basins on the northern shelf were somewhat isolated from the rift basins farther south, as evidenced by sediments lapping onto the Site 1118 basement high (and lateral equivalents). The shelf deposits formed lensoid stratal packages until ~3.2 Ma (Fig.F8) (Goodliffe et al., this volume; Fang, 2000). We mentioned above that thermal subsidence following previous forearc extension may have produced this accommodation space. Alternatively, or in addition, the flanks of the active rifts may have been uplifted and back-tilted as the footwalls to the graben-bounding faults were unloaded, thereby providing a southern border to the shelf basins that were gradually subsiding concomitant with regional crustal thinning. The paleochannel drilled at Site 1109 (Goodliffe et al., 1999) transected this rift-flank and outer-shelf high, locally connecting the shelf and rift basins.
Benthic foraminifers provide a detailed record of the continued margin subsidence, with Site 1109 passing ~150 mbsl (i.e., from outer neritic to upper bathyal) at ~4.2 Ma, followed by Site 1115 at ~3.8 Ma (Fig. F7) (Shipboard Scientific Party, 1999). Site 1118 was finally drowned at this time (~3.8 Ma) and quickly passed to upper bathyal depths (~3.6 Ma) as the margin deepened southward (Shipboard Scientific Party, 1999). As Site 1118 drowned, shallow-water bioclastic debris with Sr isotopic ages of 11.3-13.6 Ma was redeposited in early Pliocene limestones there (Allan et al., this volume) above fluvial conglomerates with prevalent dolerite and rare gabbro clasts (Shipboard Scientific Party, 1999; Cortesogno et al., this volume). Rapid slope sedimentation and even faster subsidence occurred at all three sites after 3.8 Ma (Fig. F7). The pattern of shingled onlap of near-horizontal and thicker rift-proximal sediments with the northern margin slope sediments was established by 3.8 Ma and continued to ~1.2 Ma (Fig. F8) (Goodliffe et al., this volume). The northern margin sediments record a marked decrease in magnetic susceptibility and the concentration of magnetite at 3.8 Ma, coincident with the increased sedimentation rates associated with the onlapping sediments (Ishikawa and Frost, in press).
The paleorift basin Sites 1108, 1114, and 1116, as well as the northern margin Sites 1109, 1115, and 1118, yielded their most complete stratigraphic records from middle and late Pliocene times. The siliciclastic deposition was largely from turbidity currents. Higher-energy environments in the rift basins than on the margins are evidenced by high-angle tabular cross-lamination in sandstones and intraformational rip-up clasts, whereas classic Bouma sequences and cross/convolute laminations in clay to fine sandstones are common to both regions (Shipboard Scientific Party, 1999; Robertson et al., 2001; Cortesogno et al., this volume). Sites 1108, 1114, and 1116 were all at mid-bathyal water depths (>500 ± 100 mbsl) at 3.4 Ma, whereas the northern margin sites did not subside through these water depths until 2.6 Ma (Fig. F7). Sites 1114 and 1116 remained mid-bathyal throughout the Pliocene. Site 1108 passed into lower-bathyal depths by 3.2 Ma (>900 mbsl) (J. Resig, pers. comm., 2001), some 500 m deeper and ~5 km southeast along the Moresby rift axis from the section depicted in Figure F8.
Awadallah et al. (this volume) performed a detailed analysis using Formation MicroScanner (FMS) and core data of the turbidite facies and bed-thickness characteristics of the Pliocene sections in the three northern margin sites. They found that the number of sand and silt turbidites decreases approximately exponentially with increasing bed thickness and that there is a significant "tail" of relatively thick beds. The frequency of sand and silt turbidite deposition decreases upslope from Site 1118 to 1109 to 1115, probably as a result of lateral and distal fining (and subsequent bioturbation of the fines). The turbidite bed thicknesses fit a power-law model, with exponential distributions of the number of beds thicker than a constant times the exponent thickness. Awadallah et al. propose that the underlying control on this pattern is the well-known power-law distribution of earthquakes—a model that seems very feasible in this region of active rifting (Fig. F2). Sites 1109 and 1118 show segmented power-law fits, with the exponent constant being greater for thicker than thinner beds. They explain this following Malinverno (1997), who showed that natural bed-thickness data sets can be expected to plot as segmented linear trends with different slopes if there is a relationship between bed length and bed thickness that depends on the bed volume. For example, smaller flows and their deposits might be confined by topographic features such as the channel in which Site 1109 was drilled, whereas larger flows might be free to spread over wider areas.
Rates of clastic input and sedimentation decrease after 3.2 Ma at the northernmost Site (1115), and there is a concomitant increase (to the present) in carbonate content with the increasing proportion of pelagics (Shipboard Scientific Party, 1999; Cortesogno et al., this volume; Robertson and Sharp, this volume). At Site 1109 the corresponding change (increasing pelagics and hence carbonate) did not occur until ~1.2 Ma. A marked increase in magnetic susceptibility and the concentration of magnetite at 3.2 Ma at Site 1115 reflects a change in source material and/or supply route (Ishikawa and Frost, in press). This change appeared slightly earlier at Sites 1118 (3.4 Ma) and 1109 (3.3 Ma), indicating northward onlap (Ishikawa and Frost, in press).
The provenance of the mixed clastic sediments is multifold. The lithic mineralogy of the sandstones and geochemical data on clay-mudstones and tephra reveal that the clastics were derived from calc-alkaline volcanics plus pelitic metamorphic rocks and lesser ultramafic rocks (Shipboard Scientific Party, 1999; Lackschewitz et al., 2001; Robertson et al., 2001; Cortesogno et al., this volume; Robertson and Sharp, this volume; Sharp and Robertson, this volume). These components have nearby source exposures on the Papuan Peninsula and D'Entrecasteaux Islands (Fig. F2). There are also Pliocene andesites on the Amphlett Islands and Egum Atoll, and 1- to 2-Ma trachytes on the Lusancay Islands (Smith and Milsom, 1984). Coalified wood fragments in the recovered sediments reveal that source terranes included eucalyptus forested uplands subject to forest fires (Cook and Karner, this volume). In addition, high-K volcanic ash layers and muds rich in tephra that appear at Sites 1109 and 1115 at <2.3 Ma are indicative of explosive eruptions from rhyolitic volcanoes of the D'Entrecasteaux Islands, both peralkaline (Dawson Strait) and calc-alkaline (Moresby Strait) (Lackschewitz et al., 2001; Robertson and Sharp, this volume; Stoltz et al., 1993).
The emergence of the D'Entrecasteaux Islands, associated with the buoyant emplacement of metamorphic core complexes in a continental arc rifting environment (Baldwin et al., 1993; Martinez et al., 2001), shed voluminous coarse debris along strike from the Moresby rift, as evidenced by the Pliocene conglomerates and sandstones at the proximal Goodenough 1 well and the sandstones, siltstones, and interlayered limestones at the more distal Nubiam 1 well (Fig. F4) (Francis et al., 1987). Most of the ophiolitic cover rocks were eroded during the unroofing and exposure of the metamorphic (and granodioritic) cores (Davies and Warren, 1988; Hill, 1994).
Interpretation of seismic profiles between Sites 1109 and 1115 indicates that another source of sediment since ~3.2 Ma was slumping of and erosion from the northern margin slopes (Goodliffe et al., this volume). Approximately 40 m of Pliocene sediment was redeposited in Pleistocene slumps, and there is a significant component of reworked slope sediments at Site 1109 (Resig and Frost, 2001). These processes continue today, as evidenced by headwall scars and submarine channels eroding the area (Fig. F2). In addition to this downslope sedimentation, there is evidence for dominantly along-basin axis flow in the Pliocene. The orientation of the subhorizontal maximum axes of the ellipsoids of the magnetic susceptibility (corrected for bedding dip and core orientation) between 490 and 680 mbsf (middle Pliocene) at Site 1118 suggest an east-southeast-west-northwest-directed paleocurrent during sedimentation, almost perpendicular to the present-day slope (Shipboard Scientific Party, 1999).
Siesser (this volume) studied the relative abundance of two temperature-diagnostic nannofossil species to identify Pliocene surface water temperature trends at Site 1115. He shows that surface waters were mostly warm during the early Pliocene until a cool interval at ~3.2 Ma. A return to warm conditions at 3-2.8 Ma was followed by another cool interval at 2.6-2.4 Ma, a warm event at 2.3 Ma, and then a decline at 2.2 Ma to sustained cool conditions reflecting Northern Hemisphere glaciation.