DRILLING STRATEGY

Prior CORKs in Accretionary Settings and the Advanced CORK Concept
The multizone ACORK represents a significant improvement and completely new engineering approach compared to the original single-seal CORK (Fig. 6). Four of the original 13 CORKs were deployed in accretionary settings, two were deployed on the Cascadia margin, and two were deployed at the Barbados prism. Results from two of these CORKs, at Hole 892B on the north end of the "Hydrate Ridge" on the Oregon margin and Hole 949C through the Barbados décollement, particularly exemplify the potential of the planned Nankai Trough ACORKs. Close to the toe of the Barbados prism, the CORK at Hole 949C monitored in situ conditions and processes via a screened section that spans most (but not all) of the décollement zone, where it produces a moderate-amplitude negative-polarity seismic reflection (Shipley et al., 1994) interpreted to result from modest fluid overpressure (Moore et al., 1995, 1998). Located in a setting higher up the Oregon margin, the CORK at Hole 892B monitored conditions and processes via a perforated section spanning a thrust fault that intersects a bottom seismic reflector (BSR) updip—a BSR that is isolated behind solid casing at the CORK itself. Results from the two sites (Table 1) illustrate well the suite of investigations that are possible with CORKs (Fig. 6), including documenting in situ baseline conditions, capturing signals of fluid flow transients, sampling in situ fluids, directly testing the hydrogeological properties of the formation, and determining key elastic and fluid transport properties of the formation with the tidal-loading method of Wang and Davis (1996) and Davis et al. (2000). The last is a powerful method that strongly complements logging and traditional pressure-testing measurements for determination of formation properties at a range of spatial scales. It requires the sustained long time-series observations of subsurface pressures possible with a CORK in order to allow proper spectral analysis to determine formation response to the various components of the tidal signal.

Despite such successes, the original CORK design has one key scientific limitation: with a single seal near the top of casing, the CORK essentially integrated signals of hydrogeological processes from the entire open-hole section or screened interval. At the original CORK sites, this corresponded to a first-order natural hydrogeological stratification (i.e., relatively impermeable sediments overlying more permeable upper basement or fault zone). However, growing evidence indicates that more detailed subsurface sampling is needed for comprehensive understanding of in situ hydrogeological systems, which commonly have complex structures and more than one active zone in nature; this is certainly the case at Nankai. The planned Nankai Trough ACORKs are the first scheduled deployment of a new multizone CORK. This concept was developed in a scientific workshop and engineering meeting in 1997 and 1998 (Becker and Davis, 1998 [www.JOI-ODP.org/USSSP/Workshops/AdvancedCORKS/Advanced_CORK_report.html]). The ACORKS thus represent an important engineering and scientific investment for future ODP and Integrated Ocean Drilling Program (IODP) in situ hydrogeological observations.

Figure 7 illustrates a composite concept for a multilevel isolation system in an accretionary prism in which one generic hole represents the range of possible hydrogeologically active formations that might require a sustained time-series observational approach for comprehensive understanding of the whole hydrological system. This figure illustrates four principal zones requiring separate isolation in this generic section: a gas hydrate horizon shallow in the section, a thrust fault within the accretionary prism, the décollement zone, and the subducting oceanic basement. The planned Nankai Trough ACORK program includes the two deepest zones illustrated in Figure 7, so it is appropriate to review the generic objectives of the packer configurations shown, realizing this is highly schematic. Short intervals are shown bracketing a thrust fault within the accretionary prism, as well as a primary detachment fault, to monitor activity in the parts of the system that are probably hydrologically most active. Multiple short intervals above and below a hydrate/gas phase boundary and the primary detachment boundary allow hydromechanical properties and vertical pore pressure gradients to be determined. Broader intervals within the prism, the underthrust sediments, and igneous basement allow definition of the overall thermal, fluid pressure, and compositional regimes. Finally, paired instrumented holes will allow lateral gradients and transient events to be characterized.

ACORK Configuration for Leg 196
The concept shown in Figure 7 has been brought to engineering fruition at ODP, and a schematic of the system under development for the two Nankai holes to be instrumented is shown in Figure 8. The design centers around a fully sealed liner built in modules, with external annular packers that hydrologically isolate a number of intervals. These intervals will be accessed for pressure monitoring and fluid sampling via gravel-packed screened ports and small-diameter lines connected to the seafloor landing module. The inside diameter of the liner is sufficient to allow use of a mud-motor and underreamer for deployment of the string and a coring bit for deepening below the ACORK liner/packer string after the string is reamed in (e.g., to create a hydrologic connection to upper basement).

Note that Figure 8 shows the configuration with the option for a seismometer and thermistor string deployed at the time of ACORK installation, as is planned for Site 1173. The seismometer and thermistor string installation are contingent upon funding and JOIDES Advisory Structure approval. As a final step in the operations, the liner needs to be sealed. If a seismometer/thermistor string is installed as shown in Figure 8, the liner would be sealed near the top with a plug that allows electronic feed-throughs. If no instrument is installed in the central bore, as planned at Site 808, the bottom of the liner would be sealed with a bridge plug. This would allow the full length of the liner to be accessed, from either the drillship or from a wireline reentry vehicle, for subsequent deployment of independent sensor strings that do not require hydraulic access to the formation, including thermistor cables, hydrophones, seismometers, and tilt or strain sensors.

Note also that Figure 8 shows only two packers and screens in the ACORK. For the Leg 196 ACORKs, five packers and six screens are planned for Site 808, with four packers spanning décollement and one packer just above basement. For Site 1173, two or three packers would span the "protodécollement" (as defined by changes shown in the Leg 190 physical properties and wireline logging data as well as Leg 196 LWD data) and one packer would be deployed just above basement. To connect the screened ports to the seafloor, well-protected, robust hydraulic lines will be strapped onto the outside of the liner as the string is made up at the rig floor and passed sequentially through each packer and screened port above. The lines will terminate at pressure gauges at the seafloor, with final spurs running via valves to hydraulic connectors, where fluids can be collected or pumping or flow tests can be performed. Long-term (5-10 yr) data loggers and pressure gauges will be installed at the rig floor. The loggers will be accessible with a manned or unmanned submersible via underwater mateable connectors for periodic data downloads and reprogramming.

Although it will be possible to drill in a multipacker string directly, the most efficient and scientifically most sensible approach for Nankai will be to ream the CORK strings into LWD holes drilled prior to CORK operations. Previous experience during Legs 131 and 190 shows that ~200 m of surface casing will be required to stabilize unconsolidated sandy sediments in the frontal thrust region. This will be jetted or drilled in with small reentry cones, allowing the LWD and CORKing operations to be completed sequentially in the same holes. With co-located holes, it will be possible to "fine-tune" the precise spacing of the screens and packers along the modular liner assemblies on the basis of the core and LWD results.

Seismometer and Thermistor String Installation
One possible interpretation of existing temperature data at the top of the oceanic plate (i.e., basement of the Shikoku basin) is that it is approximately isothermal with a maximum temperature of ~140°C. This implies that the basement plays an important role for fluid paths. Scientific monitoring using CORKs has revealed a strong relationship between formation fluid pressures, regional stress fields, and Earth tides (Davis et al., 2000; Davis and Becker, 1999) and allows monitoring of in situ dynamic fluid behavior. As recent studies show (e.g., Endo et al, 1997, Kiguchi et al, 1996), short period fluid pressure fluctuations can be used to estimate formation permeability for which it is difficult to directly obtain in situ values. Therefore, the installation of either a strainmeter or seismometer could provide additional information to estimate the permeability of oceanic crust through the monitoring of in situ dynamic fluid behavior in relation to mechanical properties of the rocks (Mikada, 1999).

Tidal influence on formation fluid behavior has been also observed using thermistor strings (Kinoshita et al., 1998). An additional objective for deploying an ACORK is to insert a thermistor string inside the casing to monitor formation fluid temperature over time. Thermistors arranged close to the seismometer are expected to not only measure the "static" in situ heat flow of the Shikoku Basin but also to reveal the "dynamic" behavior of such fluids because of the change in stress of the surrounding media from the propagation of earthquake signals.