The use of petrographic discrimination diagrams such as those of Dickinson et al. (1983) to distinguish the tectonic setting of sediment source areas is well established. Interpretation of these diagrams is based on the assumption that the nature and availability of source lithologies is intimately related to tectonic processes controlling the development of a basin margins. Problems with interpretation can arise if much detritus was derived from a tectonically unrelated setting (e.g., older orogen). Also, the method implicitly assumes an ideal tectonic arrangement (e.g., similar to the type area in the western United States [arc, forearc basin, and trench]). However, the Woodlark Basin is unusual in several respects, notably the presence of arc-type volcanics well out into the forearc region, which complicates the interpretation of petrographic results.
The most commonly used diagram is the QFL plot with the provenance fields of Dickinson et al. (1983) superimposed (Fig. F18). The dominance of feldspar and volcanic lithic grains and the rarity to complete absence of quartz ensures the sandstones and sands plot within the magmatic arc field. Further division on the basis of the abundances of feldspar and lithic fragments results in the majority of the sandstones (four of five) from the upper Miocene Trobriand forearc basin plotting within the undissected-magmatic arc field, which confirms the validity of this model (Fig. F18). The sands and sandstones of the Woodlark rift basin, when divided according to these same abundances, plot within the undissected- to transitional-magmatic arc fields (Fig. F18). This highlights the evolving nature and the complex local sediment sources and pathways of the Woodlark rift basin between the late Miocene to Pleistocene, which will be discussed below, according to the stages of tectonic evolution proposed for the basin by Robertson et al. (2001) (Figs. F19, F20).
Southwestward subduction of oceanic crust of the Solomon Sea initiated the Miocene-Holocene Trobriand volcanic arc (Lock et al., 1987; Davies et al., 1987), which is well developed on the Papuan Peninsula. The front of the Trobriand Arc was located in the vicinity of the Moresby Seamount and D'Entrecasteaux Islands (Robertson et al., 2001). This is supported by drill evidence, with a middle Miocene forearc sequence recovered only at the most northern site (1115). Based on seismic stratigraphic records (Taylor et. al., 1995), Robertson et al. (2001) (Fig. F20) further suggested that the cored interval here represents only the upper section of a kilometer-thick forearc succession. The composition of sandstone near the base of the cored interval is dominated by colorless vitric and felsitic volcanic lithic fragments. Biotite and amphibole are the dominant detrital ferromagnesian minerals. This suggests that volcanics of intermediate to silicic composition were the main source of detritus. In addition, the presence of alkalic volcanic clasts in a packstone of middle Miocene age was reported by Cortesogno et al. (this volume). The lithic detritus within the remainder of the sandstones of the Trobriand forearc sequence is dominated by lathwork grains, and clinopyroxene is the dominant and, in some cases, the exclusive detrital ferromagnesian mineral. Felsitic detritus is present in most sandstones. Therefore, the sandstone detritus during the middle Miocene was dominantly sourced from pyroxene basalts and minor silicic volcanics (Fig. F19). Calc-alkaline volcanics of the Trobriand Arc (Lock et al., 1987; Davies et al., 1984) to the south are thought to be the source for this detritus (Fig. F20). However, Robertson et al. (2001) suggested that similar calc-alkaline volcanoes located within the present forearc region (e.g., Woodlark Island) (Ashley and Flood, 1981) could also have provided an additional source (Fig. F20). The presence of silicic and alkalic volcanics suggests that the contemporary Trobriand Arc had a complex volcanic history.
During this period, the Trobriand outer arc/forearc was subaerially exposed and eroded and the initial sediments of the rift basin prograded northward over the forearc succession (Robertson et al., 2001). A shallowing-upward succession at Site 1115 (>5.4 Ma) culminated in emergence and erosion of an unconformity, which is overlain by an upper Miocene shallow-marine succession (Robertson et al., 2001) (Fig. F19). These authors further recognized nonmarine counterparts of these shallow-water deposits at Sites 1109 and 1118, although these remain undated. Conglomerates at Sites 1109, 1115, and 1118 contain clasts of dolerite, basalt, and gabbro (Taylor, Huchon, Klaus, et al., 1999). The dolerite is of inferred ophiolitic origin (Robertson et al., 2001). The lowermost upper Miocene sandstones at Site 1115 are dominated by lathwork grains, with clinopyroxene the most abundant and, in some samples, exclusive detrital ferromagnesian mineral. Lowermost upper Miocene sandstone from Site 1109 contains goethite/limonite concretions and chlorite and clinopyroxene as the predominant ferromagnesian minerals. The composition of the sandstones and conglomerate thus suggests a basinwide dominantly mafic volcanic (basalt to dolerite) source at this time (Fig. F19). During late Miocene time this was the exposed Trobriand forearc, which included Paleogene ophiolitic rocks (Fig. F20). Derivation from the hinterland of Papua New Guinea, including the Paleogene Papuan ophiolite belt, is unlikely as other metaophiolite rocks (e.g., serpentinite) are absent.
During the late Miocene-early Pliocene, rifting led to a transition from terrestrial to marginal-marine and then shallow-marine deposition. At Sites 1109 and 1115, paralic deposits are overlain by upper Miocene-lower Pliocene lagoonal to shallow-marine rocks, whereas at Site 1118, lower Pliocene shallow-marine rocks abruptly overlie inferred upper Miocene fluvial conglomerates (Robertson et al., 2001) (Fig. F19). At Sites 1115 and 1109, the basin underwent gradual subsidence from neritic to bathyal conditions during the latest Miocene-middle Pliocene, whereas the marine transgression was delayed until latest early Pliocene at Site 1118 (Fig. F19). The beginning of extension was marked by a change to more explosive silicic volcanism, and the sandstones deposited during this interval were dominated by felsitic and colorless vitric grains, with amphibole and biotite the dominant detrital ferromagnesian minerals and clinopyroxene an insignificant component (Fig. F19). Mafic volcanics were still supplying some detritus to the basin, as shown by the presence of rare lathwork detritus within the sandstones and the presence of rare basalt and dolerite pebbles within shallow-marine paraconglomerate at Site 1118 (Taylor, Huchon, Klaus, et al., 1999).
Assuming sediment pathways were similar to today, it is suggested that the probable source of the Pliocene turbidites lay to the northwest (Robertson et al., 2001) (Fig. F20). In this interpretation, the main sources of ash and volcaniclastic turbidites were possibly the Amphlett Islands, Dawson Strait (e.g., Dobu Seamount), Moresby Strait, and surrounding areas where Pliocene-Pleistocene volcanic rocks occur (Smith and Milsom, 1984; Binns et al., 1987) (Fig. F20). Additional sources of volcanics could be the active Trobriand Arc volcanoes on the northern rift margin (i.e., the Luscany Islands, Trobriand Island, Woodlark Island, and Egum atoll), the eastern Papua Peninsula, and sediment reworked from the Cape Vogel Basin to the northwest (Fig. F20).
During this time at Sites 1118, 1109, 1118, and 1115, sandstones were deposited at bathyal depths, except for the lowermost 5 m at Site 1115, which was deposited in a neritic environment (Robertson et al., 2001) (Fig. F20). The lithic component of the sandstones that accumulated during this time at Sites 1115 and 1109 is dominantly felsitic with colorless vitric fragments. At Sites 1108 and 1118, the lithic component of the sandstone varies from rich in felsitic and colorless vitric material to rich in lathwork and microlitic grains. The sequences on Moresby Seamount at Sites 1116 and 1114 were deposited in bathyal depths and demonstrate similar changes in lithic detritus to those observed at Sites 1108 and 1118 during this period. At all sites except Site 1115, sand-sized metamorphic and serpentinite detritus makes its first appearance in the late Pliocene (~3 Ma) (Fig. F19).
The similarities in Pliocene turbidite sedimentation at Sites 1114 and 1116 on the southern margin and Site 1108 near the rift depocenter farther north have been taken to suggest that the intervening Moresby detachment fault was not then manifest as a major topographic feature (Robertson et al., 2001). Geochemical and mineralogical studies of the hemipelagic sediments (Robertson and Sharp, this volume) revealed very high abundances of Cr and Ni (and also locally Cu and Zn) sporadically throughout the Pliocene succession at all sites. This implies serpentinized ultramafic rock detritus was able to reach all sites in the basin as fine-grained sediment earlier than the abrupt influx in sandstone detritus after ~3 Ma. In addition, relatively high Al, K, Na, and minor elements Rb, Zr, and Y within lower-middle Pliocene hemipelagic sediments suggest terrigenous-derived sediments had access to the rift basin through low-density turbidity until late Pliocene time (Robertson and Sharp, this volume). Therefore, it is interpreted that all of the sites formed part of single turbiditic basin (or several subbasins) during this time.
The influx of sand-sized serpentinite and metamorphic detritus in the late Pliocene is thought to reflect a major change in the architecture of the Woodlark rift basin (Fig. F19). Prior to this time, fine-grained terrigenous-sourced sediments (including detritus from ultramafic rocks) derived from the Papua New Guinea mainland were able to reach the present northern margin of the Woodlark rift, as it did not then exist as a deep rift basin (Robertson and Sharp, this volume). However a discrete pulse of rifting in the late Pliocene resulted in the deepening of the Woodlark rift basin, and terrigenous input to the northern rift margin was cut off (Robertson and Sharp, this volume). The Paleogene Papuan ophiolite belt and the Owen Stanley metamorphics were unroofed as the southern margin of the rift was exhumed (e.g., Moresby Seamount) and, in places, subaerially exposed (e.g., D'Entrecasteaux Islands and onshore Cape Vogel Basin), resulting in new more proximal source of metamorphic, igneous, and ophiolitic detritus (Fig. F20).
The sources of the volcanic detritus in the early Pliocene (see above) continued to supply material throughout the Pliocene (Fig. F20). In addition, vitric fragments present that appear at <2.5 Ma (Sites 1109 and 1115) are indicative of high-K calc-alkaline volcanic centers, possibly located in the Dawson Strait (e.g., Dobu Seamount) and Moresby Strait, (Lackschewitz et al., 2001; Robertson and Sharp, this volume) (Figs. F19, F20). By this stage, the northern rift margin (Sites 1109 and 1115) was largely isolated from terrigenous sediment input (Fig. F20).
Sandstones during this time were deposited in bathyal depths as volcaniclastic turbidites and subordinate air-fall ash (Taylor, Huchon, Klaus, et al., 1999) (Fig. F19). Growth of a carbonate platform on the gently subsiding Trobriand Basin to the northwest (Tjhin, 1976) had the effect of markedly reducing clastic input to the Woodlark Basin during the Pleistocene at Sites 1108, 1109, and 1115 (Robertson et al., 2001).
At Sites 1115 and 1109, colorless vitric volcanic fragments represent the dominant lithic detritus, except for two limited time intervals at Site 1109 where lathwork fragments are predominant (Fig. F19). Sediments of this age recovered from Sites 1108 and 1110-1112 are interpreted as talus deposits, in part (Taylor, Huchon, Klaus, et al., 1999). Clasts recovered from Site 1108 included sandstone, greenschist, dark pelitic schist, and variably altered and deformed dolerite/gabbro. Rare clasts of acid-intermediate porphyry, quartz trachyte, lamprophyre, and sandstone were recovered from Sites 1110-1113. A sandstone clast from Site 1108 is of similar composition (metamorphic and volcanic detritus) to middle Pliocene or younger sandstones of the Woodlark rift basin, which suggests it was eroded from strata of that age within the basin. The abundance of lathwork fragments and detrital clinopyroxene in sandstone clasts from Site 1112 implies they were derived from Pliocene sandstones present in the Woodlark rift basin.
The source of the talus is considered to be due to the continued emergence of Moresby Seamount during the late Pliocene to Pleistocene, as bounded by a major inclined fault scarp (Fig. F20). The talus includes material derived from the Pliocene sedimentary cover of the seamount. Serpentinite was eroded from the seamount as extensional faulting unroofed deeper structural levels and continued unroofing exposed schistose and gneissic rocks of presumably the Owen Stanley metamorphics (Davies, 1980), which structurally underlie the Paleogene ophiolitic rocks on a regional scale.
Vitric fragments present at Sites 1109 and 1115 are similar in composition to those of late Pliocene age (see above) and are therefore indicative of high-K calc-alkaline volcanic centers possibly located in the Dawson Strait (e.g., Dobu Seamount) and Moresby Strait, (Lackschewitz et al., 2001; Robertson and Sharp, this volume) (Fig. F20). Other volcanic detritus of this age might have been sourced from Woodlark Island, the eastern Papua Peninsula, and/or the D'Entrecasteaux Islands (Fig. F20).