The lower Shikoku turbidite facies (Unit III at Site 1177) is of early-late Miocene age (~6-16 Ma) and therefore represents a major period of deposition in the Shikoku Basin although at a relatively low sedimentation rate of ~30 m/m.y. (Shipboard Scientific Party, 2001b). Sandstones at Site 1177, sand interval D, are latest early-middle Miocene age and are older than all other medium to coarse sand encountered during Leg 190 (Fig. F2) (Shipboard Scientific Party, 2001b). These sands are dominated by quartz and feldspar with lower lithic fragments that in two samples are mainly felsic volcanic fragments and in the other are mainly sedimentary rock fragments (Table T1). The greater quartz and feldspar content of these sands is consistent with derivation from a plutonic terrane; the source is difficult to identify on the basis of sandstone petrography alone. Further constraints from sediment geochemistry, isotopic data (e.g., U/Pb zircon ages), and zircon/apatite fission tracks are required to identify an appropriate source. A source is tentatively suggested as the inner zone of southwest Japan (Fig. F4). At this time emergent southwest Japan was dominated by Cretaceous and younger plutonic and volcanic rocks of the inner zone. This would imply a transverse delivery of sediment from the northwest into the Shikoku Basin (Fig. F4).
Initiation of deposition of the lower Shikoku Basin turbidite succession (Unit III at Site 1177) coincides with several major events in the Japanese region including rapid opening of the Japan Sea at 18-15 Ma (Otofuji, 1996; Lee et al., 1999). Also a phase of major igneous activity affected the outer zone of southwest Japan at 13-15 Ma with dominantly felsic plutonic rocks and associated volcanics (Fig. F4) (Hasebe et al., 1993) and with less volumetrically significant high-Mg andesite magmatism of the Setouchi volcanic belt (Figs. F4, F5) at ~13-14 Ma (Tatsumi et al., 2001). The Miocene outer zone igneous activity produced the rhyolitic tuffs discovered near the base of Site 808 (Shipboard Scientific Party, 1991), but no equivalents are present at Site 1177. The low content of volcanic material in the turbidites of sand interval D contrasts with the tuffaceous units at Site 808 and the underlying Unit IV at Site 1177.
Subduction may have been at least episodically active during the Miocene along the proto-Nankai Trough as indicated by (1) collision between the Honshu arc and the Izu-Bonin island arc in four main phases over the last 12 m.y. (Amano, 1991), (2) crustal shortening along the north coast of southwest Japan in the late Miocene (5-8 Ma) (Itoh and Nagasaki, 1996), and (3) resumption of widespread calc-alkaline volcanic activity in Kyushu at ~6 Ma (Kamata and Kodama, 1994). The continuing influx of sandy sediment into the Shikoku Basin at Site 1177 indicates that if subduction was occurring throughout deposition of the turbidite succession (6-16 Ma), then the trench and/or trough did not prevent sediment transport. If subduction occurred at the present rate of 40 mm/yr, then the turbidites at Site 1177 were derived from settling from turbidity currents that traveled as far as 600 km beyond the deformation front at the base of the proto-Nankai Trough accretionary prism. Deposition of the upper part of the turbidite succession at Site 1177 was synchronous with deposition of the axial and outer trench wedge facies at Site 1178 (Shipboard Scientific Party, 2001c). Thus, the Nankai Trough must have existed as a paleogeographic feature in the late Miocene (cf. Shipboard Scientific Party, 1975b) and this indicates that subduction was active from at least 7.5 Ma.
Late Miocene sandstone and sand from the upper part of Site 1177 (sand intervals A and B) and Site 1178 are similar to each other but contrast with those from Sites 1175 and 1176 in their significantly lower quartz content (Table T1; Figs. F3, F5). The lower quartz content is displaced by greater abundance of sedimentary rock fragments (Fig. F3; Table T1). Their quartz content is consistent with derivation from a continental land mass. This is also indicated by the abundance of woody plant material, especially in the upper parts of turbidite layers (Shipboard Scientific Party, 2001b). The commonly lower chert content of these sands compared with those from Sites 1175 and 1176 indicate that the Shimanto Belt was an unlikely source of these sediments (see below). Significant sedimentary rocks must have been present in the source area in addition to the uplifted subduction complexes and metamorphic rocks present throughout southwest Japan (Fig. F4) (Taira et al., 1988; Isozaki, 1996).
The timing of the provenance change between the lower-middle Miocene sandstone of interval D at Site 1177 to the upper Miocene sands/sandstones at Site 1177 is poorly constrained because of the limited data from sands in intervals C and much of B at Site 1177. On Shikoku Island, cooling of the Shimanto Belt at ~10 Ma is indicated by apatite fission track data and is interpreted to result from exhumation of the accretionary prism triggered by growth related to the Outer Zone felsic igneous activity supplying sediment to the trench (Tagami et al., 1995, p. 227). Sand derived from Outer Zone magmatism was not detected in samples from Site 1177.
Upper Miocene sands from both the Nankai Trough and the turbidite succession of the Shikoku Basin contain only minor felsic volcanic detritus, which indicates that they are unlikely to derive from either the Kii Peninsula or Kyushu, where 13- to 15-Ma felsic igneous rocks are more abundant (Fig. F4). It is tentatively suggested, based solely on the sedimenticlastic sand composition and the lower chert content compared to Site 1175 and 1176 sands, that these upper Miocene units reflect erosion of lower-middle Miocene forearc basin deposits on Shikoku. Only small remnants of these units are preserved in the Shimanto Belt on Shikoku (Taira et al., 1988), but they are more widespread on the Kii Peninsula (Hisatomi, 1988). It is inferred that these deposits were formerly more widespread on Shikoku but were largely removed by erosion. Determination of sediment geochemistry and isotopic signatures of Leg 190 sediments and comparison with forearc basin sedimentary successions on Shikoku Island and the Kii Peninsula is required to substantiate this inference.
Pleistocene-Pliocene sands at Sites 1175 and 1176 are dominantly quartzose and lack mafic to intermediate volcanic rock fragments thought to characterize sand derived from the collision zone in central Honshu (Taira and Niitsuma, 1986; Marsaglia et al., 1992; Underwood et al., 1993). The abundance of sedimentary rock fragments and chert is consistent with derivation from the nearby uplifted subduction complexes of the Shimanto Belt in Shikoku (Fig. F4) (see fig. 22 in Taira and Niitsuma, 1986). The sands were presumably transported from the north-northwest down the inner trench slope rather than by axial transport along the Nankai Trough. Submarine canyons developed along the inner trench slope of the Nankai Trough extend up onto the continental shelf (Okino and Kato, 1995). These features or ones of similar geometry probably existed on the inner trench slope in the Pliocene-early Pleistocene when these sands were deposited. This confirms shipboard observations and the inference that a local southwest Japan source provided the trench fill to slope transition at Sites 1175 and 1176 (Moore, Taira, Klaus, et al., 2001).
The upper Pleistocene volcaniclastic sands at Sites 1173 and 1174 contain abundant mafic to intermediate volcanic fragments with less common sedimentary and metamorphic fragments and are consistent with derivation from the east in the collision zone between the Izu-Bonin and Honshu island arcs (Figs. F4, F5) (Taira and Niitsuma, 1986). Marsaglia et al. (1992) noted that sands from east to west along the Nankai Trough plotted on a mixing line between an undissected magmatic arc component and an accretionary prism component on a QFL triangular diagram. Sand from Sites 582 and 583 lay in the transitional arc field, whereas sand from Site 298 lay in the dissected arc field and sand from Site 297 lay on the boundary between the dissected arc field and the recycled orogen field (Fig. F3D). The implication is that along the trough to the west the uplifted subduction complex (recycled orogen) provided sediment to the trough and that this increased with distance from the Izu collisional zone. Site 1175 and 1176 sands represent the Shikoku end-member on the mixing line for Pleistocene sands in the Nankai Trough (Marsaglia et al., 1992).