IODP Proceedings    Volume contents     Search


Site summaries

Site C0010 riserless observatory

  • Hole C0010A = 33°12.5981′N, 136°41.1924′E

  • Water depth = 2523.7 m

  • Drilling depth = 555 mbsf (Table T1)

Site C0010 is one of the key observatory sites because it intersects a branch of the megasplay fault at ~410 mbsf (see Saffer, McNeill, Byrne, Araki, Toczko, Eguchi, Takahashi, and the Expedition 319 Scientists, 2010) (Figs. F3, F4). During Expedition 332, the site was revisited about 15 months after the hole was drilled and cased. The two major objectives this time were (1) to recover the SmartPlug temporary observatory installed in 2009 and (2) to replace it with a GeniusPlug temporary observatory. No drilling, LWD/MWD, or other operations were performed at this site, which is the planned location for a LTBMS.

A number of science objectives for Site C0010 were achieved during Expedition 319 in 2009, including drilling with LWD/MWD across the megasplay fault to TD of 560 mbsf, casing the borehole (with casing screens at the fault), and installation of a simple pore pressure and temperature monitoring system (SmartPlug). It was possible to (1) define major lithologic unit boundaries of the shallow megasplay fault zone and (2) determine the preferred placement of screened casing joints spanning the megasplay fault zone for observatory monitoring (Fig. F3). Through comparison with previously drilled Site C0004 located ~3.5 km along strike (Kinoshita, Tobin, Ashi, Kimura, Lallemant, Screaton, Curewitz, Masago, Moe, and the Expedition 314/315/316 Scientists, 2009) these data also provide insights into along-strike differences in the architecture of the megasplay fault and hanging wall (Flemings et al., 2009).

Although the SmartPlug is a relatively simple instrument, it marks the first observatory placement during the NanTroSEIZE project. The instrument contains a cylindrical steel hull with an internal metal frame that hosts the electronics and transducers (see Fig. F5). The inner section is cushioned against shock and is therefore suitable for installation in the Kuroshio Current and other high-current areas where VIV during deployment may be a concern. The instrument package includes a data logger, a temperature sensor within the data logger housing, a self-contained temperature sensor, and two pressure gauges (one “upward looking” and one “downward looking”). These sensors monitor pressure (1) below the packer seal in a screened interval that is open to the fault zone and (2) above the packer seal to serve as a hydrostatic reference open to the overlying water column. In addition to the thermistors incorporated with the pressure transducers for thermal compensation, two independent temperature sensors are included in the SmartPlug. These two temperature sensors are at a depth just below the packer in the upward-looking section of the casing packer. The retrieval of the SmartPlug allows the collection of the first observatory data (pressure and temperature) from the shallow section of the megasplay fault, as well as an evaluation of observatory installation procedures in the Kuroshio Current. These data will be invaluable in planning for future permanent observatory installation.

After reentry into Hole C0010A, the latching tool connected to the retrievable bridge plug without much effort, and the bridge plug and SmartPlug package was safely pulled out of the hole only a few hours after reentry (Fig. F6). On a long but careful recovery trip through the water column, the SmartPlug remained undamaged and was recovered safely on deck on 7 November (Fig. F6). Once the tack-welding points of the SmartPlug-to-bridge plug connection were cut, the SmartPlug was unscrewed; corrosion was found on the outside and inside of the instrument, most likely a result of the rusted inner casing of the borehole.

The SmartPlug sensor carrier with the pressure transducers, miniature temperature logger (MTL), and pressure housing containing the data logger was cleaned prior to hooking it to a computer via the AQW connector for data download. A full record of the deployment period (23 August 2009–7 November 2010) plus ~2 weeks prior to the deployment was recovered from all transducers and the MTL, attesting to the robust design of the instrument. Full-time series records are shown for the upward-looking (hydrostatic reference) and downward-looking (megasplay fault zone) pressure (Fig. F7) and the platinum chip and MTL temperature (Fig. F8). The full records clearly show an increase in pressure and decrease in temperature as the instrument entered the water and was lowered toward and into the seafloor. Thereafter, equilibration began, but both records suggest that the borehole had not fully equilibrated by the time we recovered the instrument (see Figs. F7, F8). The data also show that the bridge plug effectively sealed the borehole because upon reentry of the drill string and latching onto the device during Expedition 332, the upward-looking pressure sensor shows a strong fluctuation owing to displacement of borehole fluid, whereas the downward-looking pressure sensor recorded a near-constant signal (Fig. F9).

In order to illustrate some of these findings, we provide a few preliminary data examples of pressure and temperature. Figure F10 provides an example of the tidal forcing seen in the upward- and downward-looking pressure transducers. As is consistent with the expectation from a subseafloor formation, the response of formation pore pressure to ocean tidal loading is characterized by diminished amplitude and a small phase shift relative to the seafloor tide (Wang and Davis, 1996).

A cursory review of the data identified multiple pressure and—to a lesser extent—temperature excursions that may be related to seismic events, although further detrending and processing of the data are required to filter the tidal signal and resolve pressure anomalies. Nevertheless, several earthquakes that took place during the deployment period generated pressure signals that were resolvable by the SmartPlug despite the sampling interval of 60 s for this deployment. Figure F11 shows the sinoidal records of the seafloor and formation pressure data, showing the signal from the M 8.8 Maule (Chile) earthquake event in the early part of the record. About 24 h after initial arrival of the pressure waves, the associated tsunami had crossed the Pacific Ocean and loaded the formation at Site C0010. Many similar examples are found in the time series record (Kopf et al., 2011). In contrast to the pressure responses recorded, temperature signals are less pronounced and, thus, more ambiguous. There are numerous examples where temperature remains unchanged while pressure shows excursions related to seismic or wave loading. On the other hand, significant changes in temperature were recorded on one occasion on 29 May 2010 (Fig. F12). Interestingly, no significant change in pressure was recorded around that time. Also, it remains unresolved why three thermistors show an increase in temperature while the platinum chip temperature drops slightly. At this point, neither a seismic event nor any other natural explanation for the observed shifts in temperature can be provided. Presumably, artifacts such as settling of the instrument or fluid circulation may help explain the temperature changes but would have ultimately caused the pressure to vary as well (which is not the case).

The SmartPlug record contains a number of events seen in both the pressure and temperature data, which will be studied in more detail after Expedition 332. At first glance, however, the following summary statements can be made:

  1. The plug apparently settled about 50 cm over the first 2 months, or some of the thread grease or anticorrosion additive to the borehole fluid was mucking the inside of the instrument (and hence the upward-looking pressure transducer) and loaded the seafloor sensor.

  2. The formation appears to be overpressured by ~10 kPa and was still recovering at the end of the record.

  3. The pressure records captured pressure and tsunami waves associated with several earthquake events. The tsunami waveform related to the M 8.8 Maule, Chile, earthquake is similar to existing records from, for example, the NEPTUNE seafloor cabled network (E. Davis, pers. comm., 2010) but has larger amplitude and a more persistent pressure signal lasting 3–4 days (see Fig. F11, and Kopf et al., 2011).

  4. The temperature records are well resolved but show yet-to-be-explained offsets, which may either have something to do with a change in the thermal insulation around the sensors or the instrument pressure case, the changing nature of heat dissipation, or hydrogeologic events such as earthquake loading of the formation. Postcruise research may shed light on the observed problem of opposite temperature trends from the different sensors during the same event (Fig. F12).

  5. The loading efficiency of ~0.8 at tsunami frequency is very similar to that at tidal frequency. Rayleigh and pressure waves are much larger in the formation relative to the water column, whereas for the oceanographic loading from tides and the Chile earthquake tsunami, the formation signal is damped compared to that measured at the seafloor reference port (Fig. F10).

More information regarding both the instrument and the data set and the interpretation of the results can be found in Kopf et al. (2011).

One additional data record collected during the SmartPlug deployment at Site C0010 is triaxial accelerometer data monitoring drill string VIV. This measurement was taken after the problems associated with some of the observatory installation procedures tested during Expedition 319 (for details see Saffer, McNeill, Byrne, Araki, Toczko, Eguchi, Takahashi, and the Expedition 319 Scientists, 2010). In 2009, an instrument carrier was damaged and two borehole instruments were lost while another’s internal circuitry was destroyed by drill string VIV-induced shocks. As a result, from fall 2009 to spring 2010 the Japan Agency for Marine-Earth Science and Technology (JAMSTEC) tested various techniques on land to minimize/reduce VIV and demonstrated that VIV related to strong currents can be efficiently reduced by attaching ropes along the axis of the drill string. This was successfully tested in the field during a Chikyu shakedown cruise in March 2010, and this technique was made a routine part of observatory operations in the high-velocity regions of the Kuroshio Current.

During Expedition 332, a set of four ropes was attached in the moonpool area when the drill string was passing from the rig floor into the water (Fig. F13). The attachment was carried out in the upper 500 m of the drill string where the current velocities are the highest. Figure F14 demonstrates how the ropes serve to reduce VIV along the drill string where the sensitive observatory instruments are located. VIV in the upper part of the drill string increased slightly when compared to a deployment free of ropes; however, such vibrations are tolerable given that no electronics or other sensitive parts are housed there. During drifting for SmartPlug retrieval at Site C0010 (see Fig. F14), we encountered accelerations of ~0.5 g or less. This value was even lower at slower drift speeds (e.g., at Site C0002), where acceleration did not exceed 0.2 g. The angle between drifting direction and sea current was ~45°, which also helped to reduce VIV. This finding has important repercussions for future observatory installations.

Once the SmartPlug and accelerometer instruments were safely recovered, the final step of the temporary instrumentation operations was the first deployment of a GeniusPlug (Fig. F15) in IODP history. The GeniusPlug is similar to the SmartPlug, except that the instrument is ~30 cm longer, and it includes an OsmoSampler for fluid sampling over time as well as the osmotically driven microbiology chambers of the flow-through osmo colonization system (FLOCS) unit. For fluid sampling, it is especially important that final instrument placement is close to the screened interval; therefore, the bridge plug assembly was adjusted to include two joints of 3½ inch tubing, which were added between the instrument and the bridge plug (Fig. F4). The GeniusPlug was set up to sample fluids for up to 24 months and was deployed after a short drifting period with ropes but no accelerometer instruments attached to the drill string. The wellhead was protected using a corrosion cap and will be revisited for instrument recovery.

Site C0002 riserless observatory

  • Hole C0002G = 33°18.0126′N, 136°38.1488′E

  • Water depth = 1937.5 m

  • Drilling depth = 980 mbsf (Table T1)

Site C0002 is the centerpiece of the NanTroSEIZE project, intended to access the plate interface fault system at a location where it is believed to be capable of seismogenic locking and in a zone thought to have slipped coseismically in the 1944 Tonankai earthquake. The primary targets include both the basal décollement at ~6200 mbsf and the megasplay fault (Tobin and Kinoshita, 2006b). The megasplay fault reflector lies at an estimated depth of 5000~5200 mbsf, and the top of the subducting basement is estimated to lie at 6800~7000 mbsf (Figs. F1, F2).

At Site C0002, several scientific objectives were achieved during previous expeditions, during which a total of six holes have been drilled within a 100 m radius (Kinoshita, Tobin, Ashi, Kimura, Lallemant, Screaton, Curewitz, Masago, Moe, and the Expedition 314/315/316 Scientists, 2009). Operations included drilling and logging with a full suite of LWD/MWD to 1400 mbsf during Expedition 314 and coring two intervals from 0 to 204 mbsf and 475 to 1057 mbsf during Expedition 315 (Fig. F3). The succession explored comprises four units, the upper three of which are different lithofacies in the Kumano forearc basin. From ~920 to 936 mbsf, the transition zone to the underlying accreted strata was found. The primary objective at Site C0002 during Expedition 332 was to install the first permanent borehole observatory system during the NanTroSEIZE project into the basal forearc basin (Unit III, dominated by condensed mudstone) and the upper accretionary prism (Unit IV, containing interbeds of mudstone, siltstone, and sandstone) (Fig. F3, F16).

The suite of sensors for the downhole portion of the observatory includes (1) pressure ports, (2) a volumetric strainmeter, (3) a broadband seismometer, (4) a tiltmeter, (5) three-component geophones, (6) three-component accelerometers, and (7) a thermometer array (Fig. F16). For practical reasons, downhole components 4–7 are incorporated into one system, so the observatory includes

  • The hydrogeological component with a pressure unit at the CORK head and a flatpack umbilical with three stainless steel hydraulic lines running into the hole and terminated at a different depths,

  • The strainmeter,

  • The broadband seismometer, and

  • The tilt combo instrument with a thermistor string.

The set of sensors is designed to collect, as a whole, multiparameter observations at periods ranging from months to s and a wide dynamic range covering events from local microearthquakes, very low frequency earthquakes, to the largest earthquake slips of the Tonankai plate boundary 6 km below the sensors. The hydrogeologic component of the sensor set is further laid out to monitor formation pressures at three downhole levels and further records data from a seafloor pressure port for reference.

In preparation for LTBMS installation, Hole C0002G was drilled without coring but with three LWD runs (once with a 12¼ inch assembly and twice with a 10⅝ inch assembly), with a maximum rate of penetration of 20–30 m/h, in order to achieve a high data resolution in order to identify the most promising locations for LTBMS instrument locations. Both resistivity and natural gamma ray were measured to confirm the anticipated depths of lithologic boundaries identified during previous drilling of nearby Hole C0002A (Expedition 314; see Kinoshita, Tobin, Ashi, Kimura, Lallemant, Screaton, Curewitz, Masago, Moe, and the Expedition 314/315/316 Scientists, 2009). Changes in composition and texture were identified by variations in natural gamma values that coincided with changes in electrical resistivity. LWD data confirm that both the depth and character of the target formation (Unit III) were broadly consistent with previous logging (Fig. F3). In particular, a marked decrease in the resistivity fluctuations coupled with an increase in natural gamma values defined the Unit II/III boundary. The lower boundary of Unit III was also identified clearly from the gamma and resistivity logs. The strainmeter was positioned fully within the relatively homogeneous and continuous zone of elevated gamma ray and resistivity values, interpreted as homogeneous mudstone (Ashi et al., 2008).

A 20 inch conductor pipe was jetted in to 41 mbsf, and then 9 inch casing was run to 887 mbsf. To ensure good coupling of the strainmeter to the formation and to eliminate local fluid motion around the seismic sensors and the tiltmeter, the sensors were cemented in the open hole section below the 9⅝ inch casing shoe in Unit III. One pressure port was installed in the open hole within Unit IV below the strainmeter to sample pore fluid pressure in the accretionary prism. This pressure port is terminated using miniscreens to minimize the risk of clogging or fouling the hydraulic lines (Fig. F17). Above the cemented sensors in the open hole, the 9⅝ inch casing has screened casing joints from 757 to 780 mbsf (Fig. F16) to monitor pore fluid pressure in Unit II; this monitoring interval is hydraulically isolated from the overlying ocean by a swellable packer (Fig. F18). The downhole sensors are digitally connected to the seafloor where power is supplied and data are recovered, whereas pore fluid pressure is transmitted through hydraulic tubing to be recorded by a data logger at the wellhead (Fig. F19).

The nonhydraulic components of the observatory are all located in the deeper portion of the LTBMS assembly string (Fig. F16). The location of the strainmeter, tiltmeter, seismometers, and pressure ports for pore fluid pressure monitoring were determined based on LWD information. In the lowermost portion of the sensor package, the strainmeter is installed in a fracture-free zone. It is a volumetric strain system with a small sensing volume (Fig. F20). Above the strainmeter is an instrument carrier (Fig. F21) on which a Guralp broadband seismometer and the tilt combo package (tiltmeter, geophone, accelerometer, and thermometer string digitizer) are mounted. The thermistor string has a total length of 146 m, with five nodes spaced over its length at predefined spacing. It is terminated at the bottom by a digitizing unit that is mounted to the top of the instrument carrier. Figure F22 shows one thermometer node protected by polyurethane and tie-wrapped to the 3½ inch tubing.

In total, three cables plus the flatpack umbilical run upward from the instrument carrier, together with the thermistor cable. Apart from the latter, everything was fed through four slits in the swellable packer’s rubber compound mantle. Further above, centralizers, tie wraps, and steel bands secure the cables and umbilical to the string and feed all lines into the CORK head (Fig. F23).

Figures F23, F24, and F25 illustrate the arrangement of the wellhead (CORK head and ROV platform), showing the orientation of the pressure recorder and transducers and Teledyne Ocean Design, Inc. (ODI) connectors. Before lowering the completion string into water, all the hydraulic tubes were equilibrated to seawater pressure by opening two-way valves on Bay 1 of the CORK head to allow fluid to enter the tube and displace trapped air (see also Fig. F19). All the pressure transducers were connected to ocean by three-way valves to protect them from excess pressures and damage. After landing on the borehole but prior to cementing, all two-way valves were closed to inhibit borehole fluid flow in the hydraulic tubes, and they remained closed until after the cementing operation. When cementing was completed and the completion string was released at the running tool, we switched the three-way valves to connect the three pressure transducers to start observation of borehole formation pressure. By ROV communication, we obtained the pore pressure records from the initial ~15 h of observation, as well as all pressure readings prior to switching the valves from the seawater port (i.e., hydrostatic pressure) to the ports at depth. The record (Fig. F26) clearly shows tidal variations with the pressure disturbance during circulation and cementing before the valves were closed. After switching the valves (after cementing was finished), we observed a shift in pressure and different tidal response from ocean, most notably for the bottom pressure port (PP1) below the cemented section near the strainmeter, suggesting that the cement has successfully isolated the bottom section of open hole from the upper portion of the casing and overlying ocean.

Before the ROV left the CORK head, the observatory was set to standby mode until future ROV visits. One exception is the CORK head–mounted pressure sensor unit, which was set to record at a 30 s interval; at the next site visit, a pore pressure record at four levels (3× formation pressure plus seafloor reference) will be available. In summary, the health of all observatory components installed with the LTBMS can be confirmed.

In the coming year, we will download pressure data and start long-term observation with all instruments by data recorder installed in the seafloor. The data recorder has a stack of batteries that can power the whole system for 1 y. We plan to connect the observatory to the undersea Dense Oceanfloor Network System for Earthquakes and Tsunamis (DONET) submarine cable observatory network so that measurements can be observed in real time from a shore-based monitoring station during the next visit to the Site C0002 observatory in 2011 or later.