Geological setting

The Nankai Trough is a subducting plate boundary, where the Philippine Sea plate (PSP) underthrusts the southwestern Japan margin at a rate of ~4.1 cm/y along an azimuth of 310°–315°N (Seno et al., 1993) down an interface dipping 3°–7° (Kodaira et al., 2000). The subducting lithosphere of the Shikoku Basin was formed by backarc spreading at 15–25 Ma (Okino et al., 1994). The Nankai subduction zone forms an "end-member" sediment-dominated accretionary prism. In the toe region off Muroto, a sedimentary section ~1 km thick is accreted to or underthrust below the margin (Moore, Taira, Klaus, et al., 2001).

The three major seismic stratigraphic sequences identified in the northern Shikoku Basin are the lower and upper Shikoku Basin sequences and the Quaternary turbidite sequences (Fig. F3). The upper Shikoku Basin facies off Kumano decreases modestly in thickness toward the north, whereas the lower Shikoku Basin facies displays a much more complicated geometry as a result of the effects of basement topography (Le Pichon et al., 1987a, 1987b; Mazzotti et al., 2000; Moore et al., 2001). Seismic thickness decreases above larger basement highs and a more transparent acoustic character indicates local absence of sand packages that characterize most other parts of the lower Shikoku Basin. The mechanical differences between subducting basement highs and subducting basement plains could be significant for fault zone dynamics and earthquake rupture behavior.

The deformation front behavior off Kumano is fundamentally different than it is at Muroto or Ashizuri. Seismic reflection data off Kumano delineate the frontal fault clearly near the prism toe; however, there is little evidence for seaward propagation of the décollement within deeper Shikoku Basin strata (see Proposal 603A-Full2 One interpretation of the seismic profile is that the décollement steps up to the seafloor, thereby thrusting older accretionary prism strata (upper Shikoku Basin facies?) over the upper Quaternary trench-wedge facies (Fig. F3). Submersible observations also indicate that semilithified strata of unknown age have been uplifted and exposed along a fault scarp at the prism toe (Ashi et al., 2002). Farther inboard, the fault ramps down into the lower Shikoku Basin facies (Park et al., 2002).

The lower forearc slope consists of a series of thrust faults that have shortened the accreted sedimentary units of the accretionary prism. A combination of swath-bathymetric and multichannel seismic (MCS) data show a pronounced, continuous outer ridge (outer arc high) of topography extending >120 km along strike, which may be related to megasplay fault slip, including the 1944 Tonankai M 8.2 earthquake and repeated previous earthquakes. Remotely operated vehicle (ROV) and manned submersible diving surveys along this feature reveal a very steep slope on both sides of the ridge (Ashi et al., 2002, unpubl. data). The outer arc high coincides with the updip end of the splaying system of thrust faults that branch from a strong seismic reflector interpreted by Park et al. (2002) as a major OOST, which we term the "megasplay" because it is a feature that traverses the entire wedge and has had a protracted history shown by the thick forearc basin trapped behind its leading edge. The megasplay is hypothesized to represent the mechanical boundary between the inner and outer accretionary wedge and between aseismic and seismogenic fault behavior (Wang and Hu, 2006). At depth this megasplay is a high-amplitude reflector (Fig. F2), and it branches into a family of thrust splays in the upper few kilometers below the seafloor. Drilling into, sampling, and instrumenting this splay fault system at several locations downdip is a major goal of the NanTroSEIZE effort. Slip on the megasplay fault may thus be an important mechanism that accommodates strain resulting from relative plate motion and is the locus of some or all coseismic fault displacement.

The most direct evidence for recent megasplay fault activity comes from stratigraphic relationships at the tips of the faults in young slope sediments. Direct fault intersections with the seafloor are not observed; however, the thrust sheets wedge into these deposits, causing tilt and slumping of even the deposits nearest to the surface.

Other direct evidence that the megasplay fault has been active in geological to recent times comes from Kumano forearc basin stratigraphy. The Kumano Basin is characterized by flat topography at ~2000 m depth and is filled with turbiditic sediments to a maximum thickness of ~2000 m. Little is known regarding the detailed stratigraphy of the Kumano Basin, but several remarkable features are recognized in the seismic profiles (Fig. F4). The overall sedimentary sequences filling the basin can be divided into four main units by unconformities based on seismic reflection stratigraphy. The sediments in the southern part of the basin are tilted northward, truncated by a flat erosional surface, and subsequently cut by normal faults (Park et al., 2002). The depositional center appears to have migrated northward after each successive unconformity. The sequences above the unconformities are tilted less than those below the unconformities. All of the sediments pinch out toward the north. All of these features appear to be caused by uplift of the outer rise and potentially by postseismic relaxation after coseismic slip on the splay faults (Park et al., 2002).

Temperature is one of the key factors that affect frictional behavior at accretionary prism and thrusts. Heat flow data provide essential constraints on the thermal regime below the seafloor. Hyndman et al. (1997) estimate that the temperature near the updip limit of the Nankai asperity is ~150°C, based on heat flow data.

Heat flow on the trough floor offshore Kumano is ~100 mW/m2 (Fig. F5). This is consistent with the theoretical heat flow estimated from the age of the subducting Shikoku Basin (~20 Ma off Kumano). However, a very rapid sedimentation rate can reduce the surface heat flow by up to ~15%, which eventually would affect the temperature of the plate interface. Heat flow on the trough floor is uniform at ~50 mW/m2. There are a few locally elevated heat flow values at the cold seep community locations along OOSTs. The thermal regime in the accretionary prism is affected by the basal heat from the oceanic crust, thermal conductivity in the region, frictional heat along the megathrust, and radioactive heat sources. H. Hamamoto (pers. comm., 2007) numerically calculated the thermal structure using a code developed by Kelin Wang. He estimates the bottom hole temperature at proposed Sites NT2-03 (3500 meters below seafloor [mbsf]) and NT3-01 (6000 mbsf) to be 90°–100°C and 140°–150°C, respectively.

Previous drilling achievements

The Nankai accretionary prism has been one of the most intensively studied forearc regions in the world, by scientists using research vessels, submersibles, and scientific drilling. During the Deep Sea Drilling Project period (Legs 31 and 87) three sites were drilled off Cape Ashizuri, whereas in the ODP period, three legs (131, 190, and 196) were carried out in the Nankai Trough off Capes Muroto and Ashizuri (Moore et al., 2005). In 1990, seven holes were drilled at Site 808 during ODP Leg 131. During ODP Leg 190 (Moore et al., 2001), drilling was performed at six sites. During ODP Leg 196 (Shipboard Scientific Party, 2002a, 2002b), LWD operations and installation of a borehole observatory were conducted at revisited Site 808 (two holes) and at Site 1173 (two holes). The drilled holes provided essential information on the stratigraphy and physical properties of the strata deposited in the Shikoku Basin and initial accretionary processes.

As the dominantly hemipelagic strata are carried into the Nankai Trough, they are covered by a thick sequence of coarse terrigenous trench sediments, causing rapid consolidation of the Shikoku Basin strata. A décollement zone develops within the Shikoku Basin section. The upper Shikoku Basin section, along with the overlying trench sediments, are stripped off the PSP and added to the overriding plate, forming a wide accretionary prism. The lower Shikoku Basin strata are carried beneath the prism, where they continue to consolidate and dewater. Although the décollement zone could serve as a permeable channel along the subducting plate boundary, it also apparently forms a seal to the vertical transport of fluid, yielding a zone of overpressure at the top of the subducting section (Screaton et al., 2002).

The accreted strata form a classic fold and thrust belt at the toe of the prism. Approximately 75 km landward of the frontal thrust, a zone of OOST or splay faults cuts the prism. At this point, the décollement steps down to the top of the oceanic crust and the underthrusting Shikoku Basin strata may be added to the base of the prism (underplated). This point approximately coincides with the updip (seaward) limit of the seismogenic zone. Evidence of fluid migration up to the OOSTs, such as the presence of chemosynthetic clam colonies, has been found where the faults come to the surface (Ashi et al., 2002). Two holes that attempted to penetrate the faults (Leg 190 Sites 1175 and 1176) were drilled to hundreds of meters depth but were ultimately abandoned short of the fault reflectors because of the very poor core recovery in thick, poorly consolidated coarse sands.

LWD achievements during ODP Leg 196

During the first half of Leg 196, LWD operations were carried out at Sites 808 and 1173 (Shipboard Scientific Party, 2002a, 2002b). In Hole 1173B, LWD reached to basement at 737 mbsf. Here the LWD data verified a subtle porosity increase with depth from 122 to 340 mbsf, followed downhole by a sharp decrease in porosity and return to a normal consolidation trend. The sharp decrease in porosity correlated with the diagenetic transition from cristobalite to quartz in weakly developed grain cements and is marked by a strong seismic reflector that is reproduced well by a synthetic seismogram based on the LWD data. In Hole 1173B, resistivity-at-the-bit (RAB) images of the borehole show no evidence of a propagating protodécollement but, rather, reveal a basinal state of stress dominated by steeply dipping fractures and normal faults of variable strike (McNeill et al., 2004).

In Hole 808I, LWD reached just below the décollement zone (1035 mbsf), where poor drilling conditions precluded further penetration. Here, RAB images provide unparalleled structural and stratigraphic detail across the frontal thrust and décollement zones that indicate northwest–southeast shortening consistent with the seismic reflection data (Fig. F6A). RAB images also document borehole breakouts that show a northwest–southeast oriented maximum principal in situ stress direction, nearly parallel to the maximum principal stress direction inferred from microfaults in cores and from the plate convergence direction (Fig. F6B). Resistivity curves suggest that the frontal thrust zone has compacted, presumably a result of shearing. In contrast, the resistivity data suggest that the décollement zone is dilated (Bourlange et al., 2004). These resistivity anomalies in the frontal thrust and décollement zones cannot be explained by variations in pore water composition and need to be verified against the density and porosity logs after careful correction for borehole washouts.

Borehole hydrologic observatory at Nankai

Starting in the 1980s, ODP engineers and scientists developed instrumentation for long-term, in situ hydrological observatories called circulation obviation retrofit kits (CORKs) (e.g., Davis et al., 1992). CORKs have been deployed at the Middle Valley rift of the Juan de Fuca Ridge in 1991 (Davis and Becker, 1994) and the Cascadia and Barbados accretionary prisms.

Since CORKs allow only for a single seal measurement, scientific interest arose to include hydrological monitoring of multiple zones in a single hole, which has now been addressed with the advanced CORK (ACORK) (Becker and Davis, 1998). In 2001, ACORKS were installed for the first time in Holes 808I and 1173B during Leg 196 for a long-term monitoring experiment in the Nankai Trough accretionary prism (Shipboard Scientific Party, 2002a, 2002b). Drilling in Hole 808I penetrated the toe of the Nankai accretionary prism. Hole 1173B lies 13 km seaward and penetrates the sediments and uppermost igneous crust of the incoming plate. Based on the data obtained by ACORK, Davis et al. (2006) suggest a discrete episode of seismic and aseismic deformation of the Nankai Trough subduction zone accretionary prism.

Seismic studies/site survey data

The Kumano Basin region off the Kii Peninsula is among the best studied subduction zone forearcs in the world. A significant volume of site survey data has been collected in the drilling area over many years, including multiple generations of two-dimensional seismic reflection (e.g., Park et al., 2002), wide-angle refraction (Nakanishi et al., 2002; Nakanishi et al., submitted), passive seismicity (e.g., Obana et al., 2001), heat flow (Kinoshita et al., 2003), side-scan sonar and swath bathymetry, and submersible and ROV dive studies (Ashi et al., 2002). In 2006, Japan and the United States conducted a joint, three-dimensional (3-D) seismic reflection survey over a ~11 km x 55 km area, acquired under contract by Petroleum GeoServices, an industry service company (Fig. F1). The poststack trace spacing is 12.5 m in the inline direction and 18.75 m in the crossline direction. This 3-D volume—the first deep-penetration, fully 3-D marine survey ever acquired for basic research purposes—has been used to refine the selection of drill sites and targets in the complex megasplay fault region and to define the regional structure and seismic stratigraphy. As drilling proceeds, the 3-D seismic data will continue to be used to analyze physical properties of the subsurface through seismic attribute studies, to expand findings in the boreholes to wider areas, and to assess drilling safety. In addition, in early 2006, Japan Agency for Marine-Earth Science and Technology–Institute for Frontier Research on Earth Evolution (JAMSTEC-IFREE) collected a small area of narrow-width 3-D reflection data over the reference and frontal thrust site transect (Park et al., 2006) (Fig. F1). The supporting site survey data for the NanTroSEIZE Stage 1 expeditions are archived at the IODP-MI Site Survey Data Bank (