Tomographic studies using earthquake waves propagating through the Earth's interior have revolutionized our understanding of mantle structure and dynamics. High-quality digital seismic data obtained from seismic stations on land, for example, have been used to identify zones of anomalous velocity and anisotropy in the mantle, and from these, to determine patterns of mantle flow. In particular, Tanimoto (1988) has demonstrated the existence of a strong pattern of deep (>550 km) high-velocity anomalies in the western Pacific, suggesting complex interaction between subducting slabs and the surrounding mantle, whereas more recent studies in areas of dense seismic coverage have provided crude images of subducting plates extending to the 670-km discontinuity (van der Hilst et al., 1991; Fukao et al., 1992) and of deep velocity anomalies extending beneath ridges (Zhang and Tanimoto, 1992; Su et al., 1992).
One of the critical problems facing seismologists who wish to improve such tomographic models is the uneven global distribution of seismic stations. Few seismic stations are located on the 71% of the Earth's surface covered by oceans, and this problem is particularly acute in large expanses of ocean such as the Pacific. The scientific importance of establishing long-term geophysical observatories at deep ocean sites to understand the dynamic processes occurring in the Earth's interior through seismic imaging has long been recognized by the earth science and ocean drilling communities (COSOD II, 1987; JOI-ESF, 1987; Purdy and Dziewonski, 1988; JOI/USSAC, 1994; Montagner and Lancelot, 1995). The International Ocean Network (ION) project has identified the western Pacific and the Philippine Sea as a particularly important gap in the global seismic network. By installing long-term borehole seismic observatories in the seafloor in this region, the ION project is attempting to fill this gap so that high-resolution tomographic images of slab-mantle interaction zones can be obtained at great depth in the most active system of subduction complexes in the world. To this end, borehole seismic observatories were installed at ODP Sites 1150 and 1151 (Stations JT-1 and JT-2 in Fig. F14) on the inner wall of the Japan Trench during Leg 186 (Suyehiro, Sacks, Acton, et al., 2000) and another observatory was successfully installed at Site 1179 in the western Pacific (Station WP-2 in Fig. F14) during ODP Leg 191 (Kanazawa, Sager, Escutia, et al., 2001).
A major objective of the ION project was to establish a borehole seismic observatory at a quiet site in the middle of the Philippine Sea on the upper plate of the Mariana subduction system to determine whether the Pacific plate is penetrating into the lower mantle below the 670-km discontinuity under the Mariana Trench but not under the Izu-Ogasawara (Bonin) Trench. A high-quality digital seismic observatory was thus installed during Leg 195 at Site 1201 in the West Philippine Basin west of the Kyushu-Palau Ridge between existing stations at Inuyama (IMA) and Taejon (TJN) to the north, Minami Torishima (MCSJ) and Chichijima (OGS) to the east, Ponphei (PATS) and Jayapura (JAY) to the south, and Ishigakishima (ISG) and Baguio (BAG) to the west (Fig. F14). The observatory is designed as a stand-alone system with its own battery pack and recorder at the seafloor so that it can be serviced and interrogated by an ROV. Like the borehole observatories installed at Sites 1150 and 1151, however, there is a coaxial transoceanic telephone cable (TPC-2) near Site 1201 that can be used eventually for data recovery and power. There are plans to connect data, control, and power lines to the TPC-2 cable, which is owned by the University of Tokyo, after confirmation of data retrieval. This is done under the auspices of the Ocean Hemisphere Network Project, a national program from 1995 to 2001 in Japan. The data will eventually become accessible worldwide through the Internet.
Site 1201 is located in the West Philippine Basin in 5711 m of water ~100 km west of the inactive Kyushu-Palau Ridge and 450 km north of the extinct Central Basin Fault (Fig. F15). Early interpretations of magnetic lineations (Hilde and Lee, 1984) indicated that the site lies on 49-Ma crust near Chron 21 and formed by northeast-southwest spreading on the Central Basin Fault. The spreading direction then changed to north-south at ~45 Ma, and spreading finally ceased at ~35 Ma as volcanism stopped on the Kyushu-Palau Ridge. Because the earliest magnetic anomalies in the region predate the initiation of subduction at ~45 Ma along the Kyushu-Palau Ridge, Hilde and Lee (1984) considered that the Philippine Sea initially formed by entrapment of an older Pacific spreading ridge. More recent bathymetric and magnetic surveys (Okino et al., 1999) show that the site lies at the transition from well-defined anomalies south of the Oki-Daito Ridge to more complicated anomalies to the north, which implies that the crust to the north may have formed at a different spreading center. Analysis of paleolatitude and declination data from the Philippine plate and its margins suggests that the plate has drifted about 15° to the north and rotated clockwise by up to 90° since the middle Eocene (Hall et al., 1995).
The sediment section at Site 1201 was predicted to be ~400 m thick based on recent seismic reflection surveys showing a two-way traveltime to basement of 0.45 s (Fig. F16). Drilling at other sites in the region during Deep Sea Drilling Project (DSDP) Legs 31 and 59 (Karig, Ingle, et al., 1975; Kroenke, Scott, et al., 1981) recovered a relatively barren deepwater section dominated by Holocene to Eocene-Paleocene(?) brown pelagic clays overlying basement near the Oki-Daito Ridge (DSDP Sites 294 and 295). At DSDP Sites 290 and 447 to the south, the section consists of a barren interval of Pliocene clays underlain by Oligocene nannofossil-bearing silty clays mixed with ash. This was underlain by a thick section of polymict and volcanic breccia presumably derived from the Kyushu-Palau Ridge to the east. The underlying basement consists of 80% basalt pillows and 20% diabase. Because Site 1201 lies in a similar setting at the foot of the Kyushu-Palau Ridge, it was considered likely that the section would be similar to that at Sites 290 and 447.
The principal objective at Site 1201 was to install a long-term borehole seismic observatory in the middle of the Philippine plate to improve global seismic coverage, to study the structure of the upper mantle under the Philippine Sea, and to study plate interactions in the western Pacific. It was also expected that drilling at Site 1201 would provide samples representative of the Eocene/Paleocene crust of the northern West Philippine Basin. Results from this site would thus augment those obtained during DSDP Legs 31 and 59, which were the first legs to sample and estimate the age of basement in the region and to confirm that the seafloor formed by backarc spreading. Results from this site will also add to our knowledge of backarc crustal structure and geochemistry, microplate tectonics, magnetic lineations, and sedimentation. Because core quality and dating techniques have vastly improved since these early legs, it was also anticipated that drilling at Site 1201 would provide better age control on backarc spreading as well as detailed records of Northern Hemisphere climate change, eolian transport, and arc volcanism in the region during the Tertiary.
As outlined above, one of the main reasons for installing a borehole seismic observatory in the middle of the Philippine plate was to achieve homogeneous seismic coverage of the Earth's surface with at least one station per 2000 km in the northwestern Pacific area (Fig. F14). Aside from plugging an important gap in the global seismic array, the Site 1201 observatory will produce high-quality seismic data. Tests with other borehole seismometers show that the noise level for oceanic borehole instruments is much lower than for most land stations (e.g., Stephen et al., 1999). High-quality seismic data from this site will be used for several purposes.
First, an observatory at Site 1201 will provide data from the backarc side of the Izu-Ogasawara and Mariana Trenches, giving greater accuracy and resolution of earthquake locations and source mechanisms. The observatory will also be valuable for resolving events in the Ryukyu and Philippine Trenches because its location is analogous to that of station WP-2 off the Japan Trench.
Observations of seismic surface waves as well as various phases of body waves from earthquakes along the margins of the Philippine plate will provide sufficient data to map differences in plate structure among the different basins comprising the plate (e.g., the West Philippine, Shikoku, Japan, and Parece Vela Basins). Only a few previous studies with limited resolution exist on the lithospheric structure of these areas (Seekins and Teng, 1977; Goodman and Bibee, 1991). Surface wave data suggest that the plate is only ~30 km thick (Seekins and Teng, 1977). Such a value is inconsistent with predicted values from age vs. heat flow and age vs. depth curves (Louden, 1980). A long-line (500 km) seismic refraction experiment in the West Philippine Basin could not image the lithosphere/asthenosphere boundary (Goodman and Bibee, 1991).
Finally, Site 1201 will provide higher seismic resolution of mantle and lithosphere structures in key areas that are now poorly imaged. There are indications that the subducting Pacific plate does not penetrate below the 670-km discontinuity and that it extends horizontally (Fukao et al., 1992; Fukao, 1992), but the resolution of these studies is poor (>1000 km) beneath the Philippine Sea and the northwestern Pacific, especially in the upper mantle, where significant discontinuities and lateral heterogeneities exist (Fukao, 1992). Data from Site 1201 will be crucial in determining whether the Pacific plate is penetrating into the lower mantle in the Mariana Trench but not in the Izu-Ogasawara (Bonin) Trench (van der Hilst et al., 1991; Fukao et al., 1992; van der Hilst and Seno, 1993) and in determining how the stagnant slab eventually sinks into the lower mantle (Ringwood and Irifune, 1988). Detailed images of mantle flow patterns may also help explain how backarc basins open and close and explain the mantle heterogeneity that causes the basalts sampled from western Pacific marginal basins to have Indian Ocean Ridge isotopic characteristics (Hickey-Vargas et al., 1995).
In addition to the seismic objectives at Site 1201, we recognized that coring at the site might accomplish a number of important geologic objectives.
Although the age of the basement in the northern West Philippine Sea has been estimated from magnetic anomalies, paleontologic confirmation has been imprecise because of spot coring, core disturbance, and poor preservation of microfossils. By continuous coring to basement using modern coring techniques, we hoped to obtain an accurate basement age from undisturbed microfossils, magnetostratigraphy, or radiometric dating of ash horizons. This information would be of considerable importance in constraining models of backarc spreading.
Recent studies on the relationship between mid-ocean-ridge basalt (MORB) chemistry and crustal thickness indicate that the degree of partial melting is strongly controlled by the temperature of the upwelling mantle at the ridge. The volume of the melt (represented by the crustal thickness) and its chemical composition are sensitive to the temperature. This means that a knowledge of crustal thickness in an ocean basin makes it possible to estimate the temperature at which the crust was formed and the concentration of major and minor chemical elements in the resulting basalts (e.g., Klein and Langmuir, 1987; White and Hochella, 1992). To date, these studies have concentrated on young MORBs. The chemical model on which these predictions are based still has large uncertainties, partly because there are few cases off ridge where rock samples and high-quality seismic data have been collected at the same location. Chemical analysis of the basalt samples from Site 1201 should provide clues as to why the crust in the Philippine Basin is 3 to 4 km thinner than normal.
Previous drilling in the West Philippine Sea was conducted during DSDP Legs 31 and 59 before the advent of piston coring, and many of the holes were only spot cored. As a consequence, the available core from the region is almost useless for stratigraphic and paleontologic reconstructions. By obtaining a continuous, high-quality record of pelagic sedimentation supplemented by high-quality logs, we hoped to obtain a proxy record of Tertiary climate change for the region. It was anticipated that the upper levels of the section might also contain a record of eolian transport from Eurasia.
Although ash and tuff were present in the sediments recovered in the region during previous legs, it was impossible to reconstruct the ash fall stratigraphy because of core disturbance and the discontinuous nature of the coring. By continuous coring using APC and XCB techniques and correlation with high-resolution Formation MicroScanner (FMS), natural gamma spectrometry tool (NGT), and ultrasonic borehole imager (UBI) logs, we hoped to obtain a detailed record of arc volcanism around the Philippine Sea.
Paleomagnetic measurements of sediments and basalt cores are important because oriented samples are difficult to obtain from the oceans. The basalts record the direction of the magnetic field at the time the basalts were emplaced and can be used to infer the paleolatitude of the site (e.g., Cox and Gordon, 1984). Although it was unlikely that enough flow units would be cored at Site 1201 to average secular variation adequately, it was thought that the results would be useful in determining a Paleogene paleomagnetic pole for the Philippine plate. Sediments are typically a good recorder of the Earth's magnetic field and should contain a continuous record of the movement of the Philippine plate through the Cenozoic. By collecting oriented sediment cores, we hoped to study the rotation of the Philippine plate and the initiation of subduction of the Pacific plate.
After arriving on site, we planned to drill and core two pilot holes to determine the geology of the formation and to establish the casing requirements for the reentry hole that was to house the borehole seismic observatory. The first was to be APC cored to refusal (estimated at 200 mbsf) with the Tensor core orientation tool then XCB cored to basement, estimated at ~370 mbsf. After a jet-in test, a second pilot hole was to be drilled to basement and then cored 100 m into basement using the RCB to determine the nature of the basement. This hole would then be logged to identify a suitable interval for setting the seismometer package.
Once the pilot holes had been completed, we planned to offset and jet in a reentry cone and ~60 m of 16-in casing. The hole would then be reentered and deepened to ~425 mbsf with a 14.75-in tricone bit to lower a 410-m string of 10.75-in casing ~40 m into basement and cement it in place. The hole would then be reentered again and deepened to ~100 m in basement for the seismometers. After the seismometer package had been made up, lowered into the hole, and cemented in place, a battery package would be lowered into the throat of the reentry cone and acoustically released. The pipe would then be tripped to the surface, bringing the deployment to completion so that it could be activated by an ROV at a later time. As at Site 1200, however, the actual operations at Site 1201 departed significantly from those that had been planned.
The JOIDES Resolution arrived on site at 1600 hr on 31 March 2001, following a 2-day transit from Guam. After the pipe was lowered to the seafloor, Hole 1201A was spudded with the APC/XCB to study the sediment section but the hole was abandoned after one core because of a premature APC shear pin failure. APC coring was then initiated in Hole 1201B at 1905 hr on 1 April and continued to a depth of 46.7 mbsf, after which the hole was deepened with the XCB to 90.3 mbsf (Table T1). The vessel was then offset 15 m to the west and a third APC hole, 1201C, was spudded and cored to refusal at 48.1 mbsf to provide a repeat section through the soft sediments. The Tensor core orientation tool was used on the third, fourth, and fifth cores in both Holes 1201B and 1201C, and a temperature measurement was taken with the Adara shoe at a depth of 44.6 mbsf in Hole 1201C.
A problem with the spooling of the coaxial cable used for the undersea television camera became apparent during the deepwater operations at the site. Since the undersea camera was required for reentry and the deployment of the seismic observatory, we decided to search for a deepwater pocket where we could fully extend and retension the cable in an attempt to correct the problem. After steaming 204 nmi northwest to a small basin indicated on Japanese hydrographic charts, we deployed the coaxial cable to a depth of 6183 m and fixed the spooling problems.
At 0718 hr on 5 April, the vessel was back at Site 1201 and the main pilot hole, Hole 1201D, was spudded 120 m south of Hole 1201C. The hole was drilled with a center bit to a depth of 80.4 mbsf, where RCB coring was initiated. Coring proceeded without incident to basement at 510 mbsf, which was considerably deeper than initially predicted, and then continued another 90 m into the basement to a total depth of 600 mbsf. After releasing the bit, Hole 1201D was logged with the triple combination (triple combo) tool from 80 mbsf to total depth. A second logging run with the FMS-sonic tool could not pass an obstruction at 366 mbsf because of deteriorating hole conditions. After completing the run in what was left of the open hole, the pipe was lowered again in an attempt to reopen the hole for logging but an impassable bridge was reached at a depth of 90 mbsf. At that point, a 50-m plug of cement was set to prevent future fluid communication with the cased reentry hole and seismometer.
The hole for the seismometer was initiated at 1600 hr on April 14, when Hole 1201E was spudded with a reentry cone and 16-in casing and jetted in to a depth of 39.1 m. The hole was then drilled to 543.0 mbsf and cased with 10.75-in casing to a depth of 527.0 mbsf, or 15 m into basement. After cementing the casing in basement, the 9.875-in ION installation hole was drilled to a total depth of 580 mbsf. By 1230 hr on 23 April, the seismometer instrument string was assembled and final electrical integrity checks were completed (Fig. F17). Hole 1201E was then reentered and the instrument package was lowered into the hole without incident. The seismometer package was cemented in place with the end of the stinger located at a depth of 568.7 mbsf (Fig. F18). The top of the uppermost seismometer was placed ~558.4 mbsf, or ~46.4 m below the basement contact. At 0830 hr on 24 April, the battery platform was lowered through the moonpool and landed in the reentry cone at 1400 hr (Fig. F19). A handheld acoustic command unit was used to release the platform. Proper platform installation was confirmed with the subsea television camera, and by 1100 hr on 25 April, the ship was secured for transit and under way to Site 1202 (alternate Site KS-1).
The principal objective at Site 1201 was to install a long-term borehole seismic observatory in the middle of the Philippine plate. Although this was successfully accomplished, the observatory will not be activated until its brains are installed during an ROV visit in the spring of 2002 and no data will be recovered until it is revisited in 2002 or 2003. In the meantime, the core recovered during the course of preparing the hole for the observatory produced striking and, in some cases, unexpected results.
Drilling at Site 1201 yielded a composite 600-m-thick section consisting of 510 m of Miocene through late Eocene sediments and 90 m of basalt. The sedimentary section consists of two lithostratigraphic units (Fig. F20). The uppermost unit (0-53 mbsf) consists of soft pelagic clays, cherts, and interbedded sandstones and silty claystones that contain significant amounts of red clay. The underlying unit (53-510 mbsf) is composed of a thick section of interbedded turbidites composed of detrital volcaniclastic material and traces of reef detritus from the Kyushu-Palau Ridge, which range in size from coarse sandstones and breccia through silty claystone to claystone. The individual turbidite layers range from tens of meters to a few millimeters in thickness and tend to decrease in thickness and grain size downsection (Figs. F21, F22), reflecting a gradual change from high-energy to low-energy deposition. The basal 20-30 m of the unit consists of interbedded turbidites and reddish tan to chocolate-brown claystones deposited in a quiet marine environment. One of the most striking features of the entire sediment section at Site 1201 is the color of the turbidites, which range from dark gray to dark greenish gray in the upper 240 m of the unit, where the volcaniclastics are (relatively) fresh, and then range from deep green to gray-green to the base of the unit. Thin section and XRD analyses show that these changes are related to progressive alteration with depth, including the devitrification of glass, the replacement of the calcic cores of plagioclase by clays, and the infilling of voids and vesicles by clays and zeolites in the upper part of the unit, and the wholesale replacement of volcaniclastic material in the lower part of the unit by smectite, chlorite, and zeolites (chabazite, erionite, heulandite/clinoptilolite, and analcime/wairakite) during diagenesis.
The composition of the interstitial water at Site 1201 is very unusual for deep-sea sediments and reflects the profound diagenesis that has occurred in the turbidites in the lower part of the section. The most striking feature is an extremely large increase in pH, Ca, and chlorinity with depth in the pore water; whereas seawater is mainly a sodium chloride solution, the altered seawater near the base of the sediments is mainly a calcium chloride solution (Fig. F23). Calcium increases to 270 mmol/kg, 27 times the concentration in seawater, by leaching from the volcaniclastic material. Similarly, chlorinity increases to 645 mmol/kg, 20% higher than seawater values, due to the removal of water during the formation of hydrous minerals such as clays and zeolites. The gain in Ca is balanced by the removal of 70% of the Na (to 140 mmol/kg) and the loss of nearly all of the Mg and K from the seawater during the formation of clay, smectite, and zeolites. Sulfate decreases as well, from 28 to 15 mmol/kg, by the precipitation of gypsum in response to the elevated Ca concentration. Alkalinity falls from the seawater value of 2.4 to <1 meq/kg as it is consumed by the precipitation of authigenic minerals. The rise in pH to 10.0 from the seawater value of 8.1 also reflects extreme alteration. Many of the pore water gradients in the top of the turbidite section can only be supported by ongoing reactions, which is consistent with the fact that the volcaniclastics at this level are not yet completely altered. Deeper in the section, however, many of the geochemical gradients approach zero, implying that equilibrium has been achieved and that the geochemistry observed is that of "fossil" pore water.
To reconstruct the geological history of the site and determine the timing of diagenesis, it is necessary to look at the microfossil and paleomagnetic record. The topmost (0-29 mbsf) and lowermost (462-509 mbsf) sections are barren of nannofossils, but moderately to poorly preserved nannofossils in the middle section allowed us to recognize six biozones spanning Zones NP19/NP20 to NP25 (Fig. F24). The turbidites between 53 and 462 mbsf represent an expanded sequence of late Eocene to early Oligocene age. Separated by a short hiatus and lying on top of the turbidites is a 25-m sequence of upper Oligocene (Zone NP25) red claystone. Compared to DSDP drilling results at Sites 290 and 447 (Karig, Ingle, et al., 1975; Kroenke, Scott, et al., 1981), the upper Eocene sediments (>34.3 Ma) recovered at this site are the oldest so far identified on the sedimentary apron of the Kyushu-Palau Ridge. Because the 47-m interval overlying basement at the site could not be dated on board ship, this is clearly a minimum age; dating of this critical interval must await the results of shore-based radiolarian studies.
Preliminary interpretation of the magnetic inclination record identified 64 reversals of the geomagnetic timescale in the sediment section. Although the Pliocene-Pleistocene section (0-5 m.y.) is apparently missing, the barren pelagic sediments in the top 29 mbsf provided an excellent record from the Thvera Subchron (C3n.4n) through the late and middle Miocene polarity intervals to Subchron C5Bn.1n or close to the base of the middle Miocene (Fig. F25). Major unconformities are present between 14.8 and 24.1 Ma and in the top section of Biozone NP24 at ~25-30 Ma. Surprisingly, the magnetic inclination record in the turbidites between 100 and 500 mbsf defines several long normal and reversed polarity chrons (C12n-C16n.2n) that are well constrained by biostratigraphic ages. The combined biostratigraphic and paleomagnetic results show that the sedimentation rates were moderate (35 m/m.y.) in the late Eocene, then very high (109 m/m.y.) in late Eocene-early Oligocene time, when the turbidites were being deposited, and then decreased to very low values (3 m/m.y.) during the Miocene, when the pelagic sediments at the top of the section were being deposited (Fig. F26).
Although the age of the basement could not be determined aboard ship, its composition and provenance are clear. The 90 m of basement drilled at Site 1201 consists of altered pillow basalts having a composition that is transitional between that of arc tholeiites and MORB and backarc basin basalts (Fig. F27). Geochemical and thin section analysis shows that the basalts have been strongly weathered, especially at the contact with the overlying sediments, where they show significant Na uptake and depletion in Ca. Hyaloclastites in the section have been palagonitized and altered to smectite, and interpillow sediments recovered from within the upper 10 m of basement contain marine microfossils, indicating eruption in a marine environment. Magnetic inclinations in the basaltic basement are shallow and indicate a position of the Philippine plate near the equator, at ~7° paleolatitude, during the Eocene.
From the data provided above, it is evident that the basement at Site 1201 formed near the equator by submarine eruption during the Eocene before 34.3 Ma. The composition of the basalts, which are transitional between island arc tholeiites and MORB or backarc basin basalts, suggests they erupted in an arc or backarc setting. The absence of calcareous nannofossils and the presence of siliceous microfossils in the interpillow sediments and pelagic sediments immediately overlying the basement suggests that the basement formed in a deep water environment below the carbonate compensation depth (CCD) (Fig. F28).
Beginning in the late Eocene and continuing into the early Oligocene (from ~35 to 30 Ma), pelagic sedimentation at the site became mixed with, and was finally overwhelmed by, increasingly thick, coarse, and energetic turbidites composed of arc-derived volcaniclastics and reef detritus. The composition and timing of the turbidites is consistent with a source to the east in the Kyushu-Palau Ridge, which was an active arc from ~48 to 35 Ma (Arculus et al., 1995) and only began to subside at ~28 Ma (Klein and Kobayashi, 1980). The presence of scoria and rounded lithic clasts in the volcaniclastic breccia at Site 1201 is consistent with subaerial erosion, but the absence of plutonic fragments indicates that the ridge remained undissected (Dickinson, 1985; Valloni, 1985). The upward coarsening of the turbidites can be attributed to many possible causes, including changes in arc elevation and erosion, sediment supply, proximity to source, tectonics, sea level, and slope gradient. Alteration of the volcaniclastics would have commenced immediately after deposition and is continuing to the present in the upper part of the section, but diagenesis would only have begun when the turbidite section became sufficiently thick for the temperature in the lower part of the section to reach 85° to 125°C (Fisher and Schminke, 1984).
Between the late Oligocene and early Pliocene, the Kyushu-Palau Ridge subsided, the deposition of turbidites came to an end, and pelagic sedimentation resumed at Site 1201 as the Parece Vela Basin opened and arc volcanism moved eastward relative to the Kyushu-Palau Ridge in response to plate reorganization. Finally, even pelagic sedimentation ceased at ~5 Ma, presumably in response to bottom currents caused by a change in bottom-water circulation.