Background

Borehole Seismic Observatories
The scientific importance of establishing long-term geophysical stations at deep ocean sites has been acknowledged by the earth science and ODP communities and is detailed in various reports (JOI-ESF, 1987; Purdy and Dziewonski, 1988; JOI/USSAC, 1994; Montagner and Lancelot, 1995; ODP Long Range Plan, 1996). The objective is to understand the processes driving Earth's dynamical systems from a regional to global scale by imaging the Earth's interior with seismic waves. Unfortunately, few seismometers are located on the 71% of the Earth's surface covered by oceans. The asymmetry and nonuniformity of seismic station distribution makes high-resolution imaging of some parts of the mantle nearly impossible. Many new ocean-bottom seismometers, whose locations have been carefully selected to optimize imaging (Fig. 1), are needed to accomplish the goals of international geoscience programs that use earthquake data. Aside from Site 1179, which was drilled and instrumented during Leg 191, several other western Pacific sites have also been selected for instrumentation. Observatories at Sites 1150 and 1151, located on the inner wall of the Japan Trench, were installed during Leg 186 (Suyehiro, Sacks, Acton, et al., 2000). In addition, proposed site WP-1, located in the Philippine Sea, is scheduled for drilling and instrumentation during Leg 195. Downhole instruments for these western Pacific borehole observatories have been developed under an ongoing national program within Japan (Ocean Hemisphere Network Project [OHP]). Data from these observatories will eventually become accessible worldwide through the OHP data center.

Aside from plugging an important gap in the global seismic array, the Site 1179 observatory will produce high-quality digital seismic data. Tests with other borehole seismometers show that the background noise level for oceanic borehole instruments is much less than most land counterparts (e.g., Stephen et al., 1999). Recent studies that exploit high-quality digital seismic data obtained on land have shown exciting new results pertaining to mantle flow. In the western Pacific, for example, Tanimoto (1988) showed that there exists a strong l = 2 (angular order) pattern of deep (>550 km) high-velocity anomalies from waveform inversions of R2, G1, G2, X1, and X2 surface waves. This suggests a complex interaction of subducting slabs with the surrounding mantle, including the 670-km discontinuity in the region (Tanimoto, 1988). However, because of sparse global coverage by existing seismic stations, current seismic wave resolution is insufficient to image the actual interaction of the plates with the mantle. More recent studies show the potential of new mantle imaging techniques, with finer scale images having been obtained in certain locations where high-quality data are dense. Two examples are the deep extension of the velocity anomaly beneath ridges (Zhang and Tanimoto, 1992; Su et al., 1992) and the fate of subducted plates at the 670-km discontinuity (van der Hilst et al., 1991; Fukao, 1992). These detailed results are possible because of the extraction of detailed information from existing seismograms. Such studies are limited by sparse data coverage, a barrier that new ocean-bottom stations can help break.

Scientific ocean drilling was introduced to ocean borehole seismometers during DSDP Leg 88 in 1982 when the Glomar Challenger drilled a cased hole at Site 581, ~320 km due north of Site 1179, and emplaced a seismometer built by the Hawaii Institute of Geophysics (Duennebier, Stephen, Gettrust, et al., 1987). The experiment confirmed that the deep-sea ocean crust is a quiet environment for seismic observatories (Duennebier et al., 1987), and it recorded a number of teleseismic events (Butler and Duennebier, 1987) including a bodywave magnitude (Mb = 6.8) temblor in Japan (Duennebier, 1987).

In September 1989, a feedback-type accelerometer capsule was installed in Hole 794D in the Japan Sea during Leg 128 (Ingle, Suyehiro, von Breymann, et al., 1990; Suyehiro et al., 1992, 1995). The instrument recorded a teleseismic event (Mb = 5.4 at a ~4000-km epicentral distance) that clearly showed a surface wave dispersion train (Kanazawa et al., 1992). In May 1992, a comparison of seafloor and borehole (Hole 396B) sensors was made using a deep-sea submersible for installation and recovery (Montagner et al., 1994). Another borehole seismometer was installed in 1998 225 km southwest of Oahu (Site 843) and has been used to better understand the deep-sea seismic noise environment (Stephen et al., 1999). In August 1999 during Leg 186, seismometers, strainmeters, and a tiltmeter were emplaced in boreholes at Sites 1150 and 1151 in the deep-sea terrace of the Japan Trench (Suyehiro, Sacks, Acton, et al., 2000). Data from these observatories have only recently been recovered by remotely operated vehicle (ROV). Although at this stage there is no consensus as to how seafloor seismic observatories should be established, it is becoming clearer that oceans can provide low-noise environments, especially when seismometers are placed into the igneous crust inside a borehole.

Tectonic Setting
The primary Leg 191 drill site (1179) is located in the northwest Pacific Ocean east of Japan. The Mesozoic M-series magnetic lineations in the region (Fig. 4) show that the lithosphere in this area was formed in Late Jurassic to Early Cretaceous time (Larson and Chase, 1972; Sager et al., 1988; Nakanishi et al., 1989). Paleomagnetic studies indicate that this part of the Pacific plate formed ~30° south of its present position, near or slightly north of the equator (Larson and Lowrie, 1975; Larson et al., 1992). The magnetic bight created by the intersection of "Japanese" and "Hawaiian" lineations implies that the spreading ridges that formed the lithosphere met at a triple junction that defined the northwest corner of the growing Pacific plate (Larson and Chase, 1972; Sager et al., 1988). Shatsky Rise, an oceanic plateau with an area about equal to Japan, began to form at the triple junction in latest Jurassic time coincident with a major reorganization of the spreading ridges and the triple junction (Sager et al., 1988; Nakanishi et al., 1989). Evidently the plateau formed rapidly at first, perhaps from a nascent mantle "plume head" (Sager and Han, 1993; Sager et al., 1999).

Sites 1180 and 1181 are located on an unnamed volcano located ~37 km west of the island of Rota in the Mariana Island arc, whereas Site 1182 is situated atop a volcano formed at the southern end of the Mariana Trough backarc spreading center. These sites were chosen for HRRS testing as backup to the original primary and alternate sites because it became necessary to go to Guam to pick up spare parts for the ship's drawworks. The Mariana arc is one of the major subduction zones of the western Pacific, where the Pacific plate converges with the trailing edge of the Philippine Sea plate (Karig, 1975; Hussong and Uyeda, 1981). Owing to divergence between the subducting Pacific plate and the retreating Philippine Sea plate, a series of several backarc basins has opened up between the two plates. The most recent of these is the Mariana Trough, which has formed since Miocene time and is located immediately westward of the Mariana arc (Karig, 1975). The volcano upon which Sites 1180 and 1181 are located is one that is known from bathymetry maps of the region near Guam and from a dredge taken from its northeastern flank for geochemical study. The dredge recovered pumice boulders and basaltic andesite rocks (Dixon and Stern, 1983; Stern et al., 1989). Although the precise age of the volcano is not known, it appears to have erupted in recent geologic time but is not currently active. The spreading center volcano that is the location of Site 1182 is one that was known from dredge studies of the Mariana Trough. It also must be geologically young owing to its position on the active Mariana Trough spreading center.

Sedimentary Setting
The history of the northwest Pacific plate since the formation of the lithosphere and Shatsky Rise seems to be one of northward drift and low sedimentation. Sediments atop Shatsky Rise are as thick as 1.2 km because the rise top remained above the carbonate compensation depth (CCD) and thick pelagic carbonate sediments were allowed to accumulate (Sliter and Brown, 1993). In contrast, sediments in the adjacent abyssal basins are thin, typically 300-500 m thick (Ludwig and Houtz, 1979), owing to seafloor depth and distance from major sediment sources.

Cores collected in the northwest Pacific basin by DSDP (Legs 6, 20, 32, and 86) and ODP (Legs 185 and 191) over the last 30 yr show a similar stratigraphy with three primary layers (Fisher et al., 1971; Heezen, MacGregor, et al., 1973; Larson, Moberly, et al., 1975; Heath, Burckle, et al., 1985; Plank, Ludden, Escutia, et al., 2000). A Miocene to Pleistocene blanket of siliceous clay and oozes is present from the seafloor downward to >200 m in places. In these sediments, diatoms and radiolarians are common to abundant but few calcareous microfossils are found. Ash layers are also common. Comparison with holes located southeast of Shatsky Rise (Fig. 5) indicates that this layer is largely absent or attenuated in that region. This observation implies that the thick Neogene layer results from productivity in waters of the western boundary currents. The gray to olive siliceous clays and oozes typically pass downward to barren brown or reddish brown clays. Although the age of these clays is often undetermined, at some sites they belong to the mid- to Late Cretaceous (e.g., Sites 51, 194, and 195) but this part of the column may contain a highly condensed Tertiary section as well (e.g., Site 576). Beneath the barren clays is an often poorly recovered layer consisting of calcareous oozes, chalk, or marl deposited soon after the formation of the crust while it was at a depth above the CCD. This layer has suffered poor recovery because it is associated with chert and porcellanite layers that are ubiquitous in the northwest Pacific. With rotary drilling using water as a flushing agent, the chert causes the formation to be ground up and the softer parts washed away, generally leaving only rounded chert fragments and slight traces of the softer matrix. In many holes, the top of the chert layer seems to correspond to the top of the calcareous section (Fig. 5) but this relationship is difficult to discern in some holes owing to poor recovery. In other holes, however, the chert appears higher in the section along with the barren brown clays.