Geochemistry, Geophysics, Geosystems

High fluid pressures and high fluid flow rates in the Megasplay Fault Zone, NanTroSEIZE Kumano Transect, SW Japan

Authors


Abstract

[1] Annular pressure while drilling data shows high fluid overpressures at Site C0001 in part of the megasplay fault zone of the NanTroSEIZE transect across the subduction zone of SW Japan. Mostly standard annular pressures while drilling occur at three other sites, including two penetrating major faults. The two holes at Site C0001 show a step up to lithostatic annular pressure at about 500 mbsf, following initial indicators of overpressure at about 375 mbsf (meters below seafloor). The annular pressure remains high and increasing to total depth of 1000 mbsf. Seismic lines through the site show bright reflectors in the zone of initial annular pressure increase. Borehole images, sonic velocities, and resistivity all suggest a zone of fractures, from about 490 to 630 mbsf. A hydraulic model of the fluid system explains the observed pressures by influx of formation fluid at about 500 mbsf. The combination of a natural influx of 3300 l/m plus 2200 l/m from the drilling system can explain the observed annular pressures. The highly fractured zone that bleeds fluids to the borehole may be sealed by a localized zone of compressive stress or by overlying gas hydrates.

1. Introduction

[2] Because high pore pressures decrease effective stress and enable fault slip [Hubbert and Rubey, 1959], knowledge of this parameter is critical in understanding deformation. The deformation of porous sediments at subduction zones is an especially prospective environment for high fluid pressures. Accordingly, the drilling and observatory programs along the Integrated Ocean Drilling Program (IODP) NanTroSEIZE transect spanning the Nankai Trough of SW Japan (Figures 1 and 2) have placed special emphasis on pore pressure measurements. To date fluid pressure measurements have been made in the central forearc basin and in the seaward-most branch of the mega splay fault system [Kopf et al., 2011; Saffer et al., 2010]. The pressure measurements recorded to date are only modestly above hydrostatic. This paper reports on pressure measurements taken in the annulus of the borehole shortly after drilling. Specifically, we analyze measurements taken in an environment of difficult drilling conditions, which revealed annular pressures that are distinctly higher than those recorded elsewhere along the NanTroSEIZE transect.

Figure 1.

Locations of Sites C0001, C0002, C0003, C0004, C0006, C0007, and C0008 [after Tobin et al., 2009].

Figure 2.

Interpreted seismic data showing tectonic setting of sites in Figure 1 [Moore et al., 2009]. Note that Site C0001 is located in the upper portion of the megasplay fault zone.

2. Nature of Annular Pressure While Drilling (APWD) Measurements and Tools

[3] The oil industry has developed tools for measuring fluid pressure in the borehole annulus while drilling, primarily for safety reasons, and to properly balance fluid pressures within boreholes to those in the adjacent formation [Desbrandes, 1990]. The data on fluid pressure is measured between 2 m and 15 m above the bit and recorded downhole (Figure 3). Commonly the Measurement While Drilling (MWD, Table 1) tool is part of a series of Logging While Drilling (LWD) tools that provide a complete set of logs, including borehole resistivity imaging (Figure 3).

Figure 3.

Tool schematic for Hole C0001D [Expedition 314 Scientists, 2009a]. Hole C0001A included only the MWD tool and therefore the APWD measurement was located 6.45 m above the base of the bit rather than 13.2 m. Because C0001A was a pilot hole the geologically relevant data only included APWD, annular temperature while drilling and natural gamma ray.

Table 1. Acronyms
AcronymExplanation
APWDAnnular Pressure While Drilling
BHABottom Hole Assembly: Logging tools, large diameter heavy pipe
IODPIntegrated Ocean Drilling Program
LWDLogging While Drilling
mbsfmeters below seafloor
MWDMeasurement While Drilling
NanTroSEIZENankai Trough Seismogenic Zone Experiment
RABResistivity-at-Bit

[4] During IODP Expedition 314 the Measurement While Drilling tool recorded annular pressure (APWD) in all boreholes. The annular pressure measured in a riserless borehole is the sum of the density of the fluid in the borehole and up through the oceanic water column, plus any pressure created by the viscous resistance of fluid flow up through the annulus. Using appropriate models this pressure can be related to a volumetric flow rate [Bourgoyne et al., 1986]. Any differences in flow from the known pumped rate must come from or be lost to the formation [Ward and Beique, 2000].

[5] Seawater was the drilling fluid in all the holes but cuttings load increases its density and therefore borehole pressure. The geometry of the drill string, and especially its bottom hole assembly (BHA) can affect fluid flow and pressure; accordingly, the specifics for the holes considered in this study are provided in Table 2.

Table 2. Borehole and Bottom Hole Assembly (BHA) Characteristics
 C0001AC0001D
Bit Diameter8.5 in.8.5 in.
BHA Diameter6.25 to 7.25 in., excepting 0.7 m of 8.5 in. and 0.7 m 8.25 in. stabilizers.6.25 to 7.00 in., excepting for two stabilizers at (8.06 and 8.25 in.) and large diameter sensor pads on the seismic vision and density tools (8.5 and 8.25 in. respectively)
BHA Length165 m (note C0001A is both significantly shorter and has no large diameter measurement tools)360 m (including 186 m of high weight drill pipe at 7.25 in.)
Distance from Bit to APWD Measurement6.45 m13.2 m

[6] In this paper we analyze pressure data from two holes, C0001A, and C0001D, separated by 65 m and located at Site C0001 (Figures 1 and 2). Both holes are 8.5 inches in diameter. C0001A was a pilot hole that was drilled with only the PowerPulse MWD tool whereas C0001D included the full suite of LWD tools shown in Figure 3. The annular pressure measurement is made at 13 m above the bit in Hole C0001D and 6 m above the bit in Hole C0001A. The LWD and MWD tools are larger diameter than the drill pipe and include stabilizers for centralization of the tools. Some of the LWD tools include large diameter sensor pads that are used to obtain measurement in close proximity to the borehole wall. Therefore the full combination MWD/LWD string of tools provides more resistance to uphole flow of fluids than when only the MWD tool is utilized. The full string of tools is also more likely to have its annulus restricted by a surge of cuttings or cavings into the borehole. Known as a pack-off event, a restriction of the borehole annulus would manifest itself as a sharp spike in the APWD reading with a corresponding spike in torque; if the cuttings or cavings get stuck between the larger diameter components of the BHA and the borehole wall rotation and annular flow is restricted. As the “blockage” clears from around the BHA, both torque and APWD will return to their baseline values. If the blockage does not clear this can lead to the drill string becoming stuck. The PowerPulse MWD tool only records pressure when pumping is turning a turbine in the tool and supplying power. Regrettably we have no record of pressure when the rig floor pumps are off, for example during a pipe connection.

3. Geological and Geophysical Setting

[7] The subduction of the Philippine Sea plate beneath southwest Japan forms the Nankai Trough and the associated accretionary prism. In the Kumano region estimates of the rate of subduction range from 4 to 6.5 cm/year along azimuths of 300 to 315° [Miyazaki and Heki, 2001; Seno et al., 1993; Zang et al., 2002]. Along the NanTroSEIZE transect, the subduction direction is rotated about 15 to 30 degrees counterclockwise from the dip azimuth of the margin, so there is a component of obliquity (Figure 1). The continental margin originated both from the deposits on the oceanic plate, that were structurally emplaced to form the accretionary prism, and sediments accumulated on the slope and in the forearc basin on top of the accretionary prism (Figure 2) [Kinoshita et al., 2009; Saito et al., 2010]. Large-scale thrust faults dominate the structure of the margin [Moore et al., 2007, 2009]. Large apparently active megasplay (out-of-sequence thrust) faults emerge along the inner trench slope [Moore et al., 2007]. The preferred slip surface of the 1944 M 8.1 earthquake also follows the boundary of the forearc basin and upper trench slope region [Baba and Cummins, 2005]. Accordingly, Site C0001 not only lies in this megasplay fault zone (Figure 2), but also in the region of the surface projection of the 1944 earthquake.

[8] Drilling at Site C0001 included both logging holes during Expedition 314 and coring holes during Expedition 315, two months later. Mud and mudstone dominate the lithology of Site C0001 in the accretionary prism at depths greater than 207 mbsf [Expedition 314 Scientists, 2009c; Expedition 315 Scientists, 2009]. Cores from Site C0001 to 458 mbsf show small-scale faults of normal, thrust, and strike-slip displacement with normal faults being most dominant and latest [Expedition 315 Scientists, 2009; Kinoshita et al., 2009]. At Hole C0001I, attempts to core between 458 extended to 520 mbsf were frustrated by high torque and stuck pipe ultimately forcing the abandonment of Site C0001 [Kobayashi et al., 2008]. The persistently unstable hole conditions at Site C0001 occurred at about the same depths (∼400 to >600 mbsf) during both Expedition 314 and 315, over more than two months.

[9] The logs recorded during Hole C0001D provide a geologic context for the pressure data that is the focus of this paper (Figure 4). The interval between 490 and 630 mbsf is characterized by anomalously low velocity and resistivity. This interval is marked by dm to m-thick zones of anomalously low resistivity of variable orientation. These low resistivity zones may represent shear zones filled with low resistivity, water-rich sediment or extensional fractures associated with shearing [e.g.,Twiss and Moores, 2007, p. 71]. Overall, the lower velocity and resistivity of the 490 to 630 mbsf-zone is consistent with high water content.

Figure 4.

RAB image of Hole C0001D with resistivity and sonic velocity curve data. The image represents the unwrapped circumference of the borehole and is about 0.7 m wide. Note zone of anomalously low and irregular velocity and resistivity in the curve data (∼490 to 630 mbsf) encompasses dark, low resistivity interval of the RAB image between about 525 and 630 mbsf. Enlarged portion (∼520–550 mbsf) of that image shows low resistivity zones in a variety of orientations. The low resistivity areas in this image could represent fluid-rich shear zones, extensional structures developed due to shearing, or extensional hydrofractures. LVRs represent thin intervals of decreased velocity and resistivity that correlate with bands of low resistivity in the RAB image outside or bordering the 490–630 mbsf zone of the RAB image. Horizontal exaggeration of the main resistivity image is 70:1 and 14:1 in the enlarged portion of the image.

4. Observations of Measurements of Annular Pressure While Drilling

4.1. Overview of Measurements of Annular Pressure While Drilling

[10] The Sites at which APWD measurements were obtained extend from the forearc basin to the base of the inner trench slope (Figure 2). The pressure curves (Figure 5) fall into two categories: (1) Sites C0002, C0004, and C0006 show a gradual increase in pressure along a gradient that is moderately above hydrostatic; this represents a standard riserless APWD gradient for the holes drilled on Expedition 314. This standard APWD gradient is estimated at 10.6 kPa/m, as measured from the forearc basin section (0–936 mbsf) of Site C0002. (2) Sites C0001 and C0003 also show initial pressure gradients that follow the standard gradient and that are interrupted by sharp increases in pressure that in the case of site C0001 exceed lithostatic values. Our analysis of APWD data investigates the geologic significance of these sharp pressure increases. We concentrate on Site C0001 because of the presence of two holes with pressure measurements; C0003 was not examined in detail because the logging tools, and most of the downhole data, were lost when the pipe became stuck and an equipment malfunction on the rig caused the pipe to separate.

Figure 5.

Annular pressure while drilling for all sites plotted with a common pressure origin. Note the standard APWD gradient shown by Sites C0002, C0004, C0006, and the shallower portions of Sites C0001, and C0003. The deeper parts of Sites C0001 and C0003 diverge from this standard gradient.

4.2. Pressure Anomalies at Site C0001

[11] Site C0001, located in the upper portion of the megasplay fault zone, recorded similar pressure surges at Holes C0001A and C0001D (Figure 5). Hole C0001D is located about 65 m landward (NNW of Hole C0001A). Drilling began at the mudline at Hole C0001A 7 days before C0001D. We focus mostly on the pressure data at C0001A; having been drilled first, it is not subject to possible cross-hole interference. However, because C0001D had a full complement of logging tools, we rely on data from this hole to characterize aspects of the drill site (Figure 4).

4.3. Time Data Displayed in Depth

[12] APWD, and most log measurements, are recorded every 10 s, along with the depth at the particular time. Typically log data is displayed in depth and only shows data points measured the first time the tool passed a given depth (e.g., Figure 6). Alternatively if you plot all time data in depth it is possible to display repeat measurements at the same depth. For example, repeat measurements occur if annular pressure is recorded as the hole is drilled and then later at the same depth when the pipe is raised for hole cleaning, a “wiper trip.” The particular MWD tool (PowerPulse) utilized on Exp. 314 requires that the rig floor pumps be operating for data to be recorded. Thus, not all pipe movements have an APWD record. The APWD data in Figure 7is one these messier plots of all time data in depth that allows us to understand pressure variations through time and repeated pressure measurements at particular depths. All other depth-pressure plots in this paper are plotted in the traditional depth format, with no repetition and reoccupation of previously measured depths.

Figure 6.

APWD depth curves for Holes C0001A and C0001D. Note that APWD for C0001A follows the standard APWD gradient until about 413 mbsf, from this depth to 530 mbsf the pressure spikes to and locally above lithostatic before settling into a stable pressure regime that is initially lithostatic but follows the standard APWD gradient to the base of the hole. The Hole C0001D pressure curve follows the standard gradient to just above 400 mbsf. Between 400 and 522 mbsf the pressure oscillates from the standard gradient with pulses toward the lithostatic curve. At 522 mbsf the Hole C0001D pressure surges above lithostatic and remains so until 663 mbsf. During this depth interval the drill pipe became stuck [Kobayashi et al., 2008] and torque values were very high (Figure 8). At 690 mbsf the Hole C0001D annular pressure falls back to a trend similar to that in the C0001A hole. Note that to a depth of about 500 mbsf the C0001D pressure curve is generally lower than the C0001A curve over the same depth interval. Because C0001A was drilled at a faster rate over the first 500 m (60 m/hr rather than 30 m/hr), the cutting load was probably greater in C0001A, increasing the annular pressure. Both holes were drilled at 30 m/hr below 500 m.

Figure 7.

Torque and Annular Pressure at Hole C0001A. Annular pressure is displayed in depth from a time recording. Unfortunately, the torque data was not available in a time file and is displayed in the traditional depth format. With the annular pressure data all times are displayed and the data curves may repeat for a specific depth. Repetition of depths occurs due to raising the pipe in preparation for making a connection or for short wiper trips where the pipe may be raised more than one stand (39 m) for hole cleaning purposes. At 526 mbsf the pipe became stuck. Short wiper trips were started at the addition of each stand of drill pipe below 526 mbsf. The hole remained in good condition until total depth (1000 mbsf) [Kobayashi et al., 2008]. Below about 530 mbsf multiple passes through each depth interval document the development of a consistent deep pressure baseline with stable torque values. The deep pressure baseline is shifted approximately 3 MPa from the shallow baseline. The torque values peak near the baseline shift and become less variable beyond 550 mbsf. Some of the pressure oscillations from about 420–530 mbsf and the stuck pipe at 526 mbsf are probably due to restriction of the annulus by collapse of sediment into the borehole. The consistency of the deep baseline pressures during the multiple occupations of depths greater than 530 mbsf indicate a stable borehole lacking temporally variable annular restriction.

4.4. APWD Anomalies at Hole C0001A

[13] The APWD data follows a standard AWPD gradient to 380 mbsf (Figures 6 and 7), establishing a shallow pressure baseline. From 380 to 530 mbsf the pressure curve fluctuates between slightly above hydrostatic to near and locally over lithostatic values (Figures 6 and 7). Below 530 mbsf the curve begins at a lithostatic value but follows a uniform gradient that is elevated and increases along the standard APWD gradient, to a total depth of 1000 mbsf.

[14] We call the stable APWD curve below 530 mbsf the deep pressure baseline. Below 530 mbsf the annular pressure values do not fall back toward hydrostatic. The multiple paths of data points along this deep pressure baseline confirm its validity (Figure 7). Projection of the shallow and deep pressure baselines to the sharp pressure offset at 490 mbsf indicates a 3 MPa offset (Figure 7).

[15] Although Figure 7 shows all APRS time data plotted in depth, the torque values, experienced by the drill string from the rig floor, are plotted as a normal depth file (due to the unavailability of time data). The overall view shows that the torque values are low above 414 mbsf, have the highest variations in the transition zone between the shallow and deep baseline trends, and show less variation at depths greater than 530 mbsf. High values in torque from 410 to 510 mbsf roughly correlate with some of the annular pressure extremes.

4.5. Anomalies in APWD at Hole C0001D

[16] Although not analyzed in detail the pressure curve from C0001D does offer valuable additional insights. The APWD of holes C0001A and C0001D show considerable similarity (Figure 6). Hole C0001D maintains a standard APWD gradient to about 370 mbsf, from 370 to 522 mbsf the pressure curve oscillates from the standard APWD gradient to values about 1.5–3 MPa higher. Deeper than 522 mbsf the C0001D pressure curve surges to exceed lithostatic values by more than 2 MPa. Below 672 mbsf the pressure curves falls below lithostatic and assumes a trajectory of increasing pressure that is subparallel to the curve at this depth in C0001A. Examination of the time plot (similar to Figure 7) indicates that the pressure never falls toward hydrostatic below this depth. Thus the annular pressure has reached a new stable baseline curve that persists until the bottom of the hole. Overall C0001D shows an upper level of the standard APWD gradient, a zone of instability culminating in a supra-lithostatic pressures, and a lower zone of elevated but stable annular pressure that increases along the standard APWD gradient, and is consistent with the lower pressure baseline identified at Hole C0001A.

4.6. Summary of Drill-String Torque for All Sites

[17] Drill-string torque provides a view of the tightness of the hole and possible blocking of the annulus by cuttings or cavings. (Figure 8). Figure 8 shows all torque data from holes crossing either splay faults or the frontal thrust. Hole C0001D shows a peak of torque at about 530 mbsf which gradually decreases to values slightly higher than those observed in Hole C0001A at about 700 mbsf. The torque value in Hole C0006B is anomalously low in the 700–900 mbsf range where the hole intercepted a sandy channel deposit with enlargement verified by caliper data [Expedition 314 Scientists, 2009b]. Both the hole enlargement and a less sticky sandy lithology decreased the torque.

Figure 8.

Torque curves for all holes penetrating major faults as seen on the seismic data (Figure 2). C0001A and D show the highest values probably due to penetration of an unstable zone at about 400 to 530 mbsf. Both boreholes probably suffered annular constriction at these depths resulting in lithostatic to supra-lithostatic annular pressures. The extreme torque values in C0001D are probably due to the large diameter and considerable length of the BHA including large diameter stabilizers and sensor pads that restricted the annulus (Table 2). The low values of torque in C0006B are probably due the sandy lithology encountered at depth that led to some annular erosion, increasing the clearance around the BHA and therefore decreasing torque. Holes C0004 and C0006 crossed major faults at 315 and 711 mbsf, respectively, with no anomalous torque response. Neither C0004 nor C0006 showed large excursions in APWD that is consistent with their low torque values.

5. Interpretation of Pressure Anomalies at Site C0001

5.1. Interpretation of Pressure Anomalies at Hole C0001A

[18] The basic patterns of pressure values at C0001A show an initial standard APWD gradient which we call the shallow pressure baseline (Figure 7). The trend to higher pressures starts at 375 mbsf with an initial anomaly (Figures 6 and 7), culminating at 490 mbsf where the annular pressures reach lithostatic. The sharp surge at the base of the transition zone suggests substantial fluid flow from a localized interval. An elevated annular pressure curve is maintained along the deep pressure baseline to the bottom of the hole. Although the rig floor pumping rate increased from 1900 to 2200 lt/m across the shallow to deep pressure baseline shift, this increase in flow is insufficient to explain the observed pressure increase (See “Hydraulic Model of Fluid Flow System,” below). Therefore a major fluid influx probably occurs at about the 500 mbsf depth, continues for the duration of the hole, and supports the deep pressure baseline.

5.2. Interpretation of Pressure Anomalies at Hole C0001D

[19] Although similar to the pressure curve at C0001A, the pressure curve from C0001D could be affected by the previously drilled hole, which is 65 m to the SSE. Like C0001A, C0001D shows the standard APWD gradient, here called the shallow pressure baseline with a transition zone to high pressures starting below 370 mbsf. Both holes show a sharp increase in pressure in the 490 to 522 mbsf range, which we interpret as flow from a fluid filled conduit, probably a fault or fracture zone. The BHA utilized at C0001D was longer and locally larger in diameter than that used at C0001A. The extremely high torque values below 535 mbsf may reflect the constriction caused by this extensive BHA (Figure 8). At 725 mbsf the torque returns to a value similar to that seen in the C00001A hole: at this depth the annular pressure is no longer above lithostatic and has returned to a deep pressure baseline similar to that seen in C0001A.

5.3. Interpretation of Torque Data With Respect to Annular Pressure

[20] At Site C0001 the increases in torque correlate reasonably well with the increases of annular pressure. The increases in torque in the transition zones at Holes C0001A and C0001D might be due to restriction of the annulus due to shedding of sediment into the borehole. At Site C0001A both the pressure and torque stabilize below 530 mbsf (Figure 8). The latter behavior is interpreted as cleaning of the borehole with continuing pipe rotation and pumping. Thus the new higher-pressure baseline is interpreted as a response to fluid flow into the borehole, not annular occlusion due to cuttings load or hole collapse. The pressure curve at Hole C0001D also returns to near this baseline at 690 mbsf (Figure 6), after a dramatic drop in torque (Figure 8), suggesting the sustained influence of an additional influx of fluid into the borehole. At both Hole C0001A and Hole C0001D these influxes are interpreted to be at about 490 and 522 mbsf, respectively, where the major pressure surges are recorded.

[21] A possible cause of the extended greater than lithostatic pressure interval (522–672 mbsf) at C0001D is the longer, mostly larger diameter BHA. The BHA in Hole C0001D included tools to measure sonic velocity, for resistivity imaging, for receiving seismic pulses from the surface and for measurement of density and porosity—all in addition to the MWD tool, solely run on C0001A. In C0001D, the array of stabilizers used to centralize the pipe plus additional large diameter sensor pads could restrict the annulus considerably. These large diameter tools extend up 174 m from the bit (Table 2). Once these tools clear the posited zone of major fluid input (∼522 mbsf), the bit depth would be 696 mbsf. It is about this depth that the annular pressure in C0001D drops below lithostatic and joins the pressure curve of C0001A. Moreover, the torque curve for C0001D ends its high excursion at about 700 mbsf and stabilizes for the remainder of the hole.

5.4. Interpretation of Borehole Images From Hole C0001D

[22] In most of the interval where APWD exceeded lithostatic at Hole C0001D the borehole is characterized by anomalously low and irregular velocity and a lowered resistivity (Figure 4). Details of the image in this zone show low-resistivity bands that have been interpreted as hydrofractures (Figure 4) [Chang et al., 2010]. They may also represent shear zones filled with low resistivity, water-rich sediment. Similarly the more irregular zones of low resistivity could represent injected water-rich sediment. Overall, the lower velocity and resistivity of the 490 to 630 mbsf-zone is consistent with high water content. This dilated interval may represent a fault zone or fault damage zone. The dilational features of this low resistivity zone could have been exacerbated by packoffs and fluid injection during drilling. However, the high annular pressures occurred at the same depth in both holes C0001A and C0001D, suggesting a depth-controlled geologic cause for this phenomenon. Finally the minimum and maximum horizontal stress magnitudes estimated from breakouts exceed lithostatic suggesting that the low velocity, low resistivity zone from 490 to 630 mbsf is in compression [Chang et al., 2010].

6. Geologic Model of Fluid Flow System at Site C0001

[23] Holes C0001A and C0001D show pressure surges beginning around 375 mbsf and culminating at about 500 mbsf. When projected on the seismic line for Site C0001, the 500 mbsf depth lies on top of a zone of high amplitude reflectors that are about 150 m thick at this location (Figures 9a and 9b). On the seismic lion the zone of high amplitude reflectors consists of one NW dipping subzone and two crosscutting SE dipping subzones (Figure 9b). These subzones could reflect fluid-engorged damage zones of a complicated (conjugate?) thrust fault system and are geometrically similar to published seismic interpretations of faults here [e.g.,Expedition 315 Scientists, 2009; Moore et al., 2009]. The high amplitude reflectors also correlate with the 144 m-thick interval in C0001D where APRS is above lithostatic pressure (Figure 6). The depth of the pressure surges could represent a cap to the zone of high pressure. Accordingly our preferred interpretation of the high pressure is an overpressured fracture zone, probably associated with a fault zone (Figure 9b). The cores taken during IODP Expedition 315 at Site C0001 show a decrease in chloride in the pore water below 400 mbsf to the base of the cored hole at 458 mbsf [Expedition 315 Scientists, 2009]. Low chloride fluids are typical of fluids migrated from deep sources, commonly along faults and fracture zones [Kastner et al., 1991]. Thus we envision this fracture zone to be fully developed at 500 mbsf with an upper transition zone extending to 375 mbsf where the first of the APWD anomalies occurred.

Figure 9.

(a) Seismic Line through Holes C0001A and C0001D showing how a zone of bright reflectors intersects the above holes at about 500–600 mbsf. Seismic data from [Moore et al. 2009]. (b) Interpretation of Figure 9a showing a possible fluid flow through a zone of bright reflectors intersecting Site C0001 to cause instability and overpressuring at about 500 mbsf. Continued overpressure during deepening of the hole to 1000 mbsf occurs due to a combination of natural flow from the formation and from fluids pumped from the rig floor. The transition from the standard APWD gradient to the overpressured regime lies at the base of the shallow pressure baseline. Full transition to the deep pressure baseline occurs over the depth interval of 375 to 530 mbsf (Figure 6).

Figure 9.

(continued)

7. Hydraulic Model of Fluid Flow System

[24] Our results suggest that Holes C0001A and C0001D penetrated an overpressured fracture zone and perhaps associated fault zone. The pressure surge at about 500 mbsf may represent an influx of fluids. Because we know the volume of fluids being pumped into the hole, the added volume necessary to cause the pressure surge can be estimated by a hydraulic model of the borehole. We have used industry-standard hydraulic modeling software, WellPlan by Halliburton. This software can be used to estimate annular pressures created by fluid influxes into the borehole. These fluid influxes, either due to pumping or flow from geological features, are resisted by the density of the fluid and viscous resistance to flow in the annulus. The consequent overpressures are recorded by the MWD tool. Inputs to WellPlan include the hole size and the details of the outside dimensions of the bottom hole assembly and drill pipe, thus defining the annular geometry. The program also requires information on the density and rheology of the drilling fluid, in this case seawater. The model assumes an in-gauge borehole diameter that has neither been enlarged through erosion nor constricted by sloughing.

[25] Our model examines the pressure increase in Hole C0001A across the 490 mbsf depth interval, which is the major dividing line between the shallow and deep pressure baselines (Figure 10). The rig-floor pumping rate prior to the surge was 1900 lt/min, which predicts a annular pressure of 27.86 MPa that lies close to the extension of the shallow baseline (Figure 10). This match provides confidence that the modeling software is accurately simulating the pressure and flow rate in the borehole, both of which are known. The shallow and deep baseline trends show a shift of approximately 3 MPa at the 490 mbsf depth. To achieve a 3 MPa shift, the model requires an increase in flow rate of 3628 lt/m. During the shift from a shallow to deep baseline, the rig floor pumping rate was increased about 300 lt/m. Thus, the estimated fluid influx (3628 lt/m), less rig floor pumping rate increase (300 lt/m), suggests a fluid influx of more than 3300 lt/min from the formation over the transition from the shallow to deep baseline pressures. We have developed a hydraulic model of C0001D which indicates higher flow rates. However, because of the potential for annular constriction due to the complex bottom hole assembly, we have more confidence in the C0001A model.

Figure 10.

Results of the hydraulic model of Hole C0001A. A pressure model at a depth of 490 mbsf and a known rig floor pumping rate of 1900 lt/min produces a pressure of about 28 MPa. This pressure is in good agreement with the value at the projection of the shallow pressure baseline to this depth. The observed 3 MPa pressure increase necessary to reach the projection of the deep pressure baseline at 490 mbsf requires a total increase in flow rate of 3628 lt/min. Of this, an increase pumping from the rig floor only accounts for 300 lt/min. Hence, a major increase in flow from the formation, about 3300 lt/min, is required.

8. Nature of Upper Fluid Seal

[26] At about 375 mbsf annular pressure pulses occur from the shallow baseline toward lithostatic values (Figure 7). Apparently the borehole emerged from a standard APWD pressure regime above 375 mbsf to a higher-pressure regime below. We consider two hypotheses that may explain this pressure transition: A gas hydrate seal, or a change in the state of stress from compressive to extensional or strike-slip conditions.

8.1. Gas Hydrate Seal

[27] The gas hydrate reflector extends seaward toward Site C0001 through slope sediments and is last recognized at about 400 mbsf, 6 km northwest of Site C0001 (Figure 9b) [Moore et al., 2009]. Above this reflector gas hydrate solid phases are stable and can seal porosity. The occurrences of anomalously low chloride contents in pore water at some depths in Site C0001 could be explained by the melting of gas hydrates in the cores. However, no other evidence of gas hydrate was observed in the cores [Expedition 315 Scientists, 2009].

8.2. Change in State of Stress

[28] Compressive states of stress with a horizontal maximum principal stress can retard vertical migration of fluid. Conversely extensional or strike-slip states of stress with horizontal minimum principal stresses can allow vertical migration of fluid at less than lithostatic pressures. Thus, the limiting stress to open fractures could control the fluid pressure in the formation.

[29] The stress profile determined by the study of breakouts and opening of fractures during drilling shows a shift from a normal faulting state of stress above 400 mbsf to a compressional state of stress at 525 mbsf to about 675 mbsf in Hole C0001D [Chang et al., 2010]. This profile in the modern state of stress is reflected, in part, by the pattern of faulting in cores at Site C0001 where the cored section, especially the upper 200 mbsf is dominated by normal faults. Below 200 mbsf approximately equal amounts of normal, strike-slip and thrust faults occur to 456 mbsf where drilling stopped due to unstable hole conditions [Expedition 315 Scientists, 2009].

[30] Changes from extensional to strike-slip, to compressive state of stress can occur most easily if there are small differences between the principal stresses. The natural occurrences of normal strike-slip and thrust faults, both at core scale and seismic scale suggests this occurs along the transect penetrated by the Sites C0001–C0006.

9. Discussion

9.1. Consistency of Deeper Baseline Pressure Trend Between Holes C0001A and C00001D

[31] We interpret the constancy of the deep baseline pressure trend below 530 mbsf at C0001A and below 690 mbsf at C0001D as due to relatively constant fluid flow into a borehole without annular constriction. Although, there is evidence for annular constriction at shallower depths where annular pressures exceed lithostatic and torque values are irregular and locally high. We believe the stable deep baseline condition has resulted from cleaning of the borehole during successive wiper trips as the bottom hole assemblies (of differing lengths and complexities) passed completely through the most unstable zone around 500 mbsf. In our opinion, the offset of the shallow to deep baseline resulted from consistent input of fluid at the 500 mbsf depth in both holes plus the constant pumping from the rig floor at about 2200 lt/min.

[32] Alternatively one could argue that the pressure surge at about 500 mbsf and the constant deep baseline gradient was somehow a function of a similar annular constriction during the deep drilling of both holes through fault zones. We believe the latter is an unlikely occurrence because Sites C0004 and C0006 penetrated faults and show no evidence of a permanent pressure surge nor large changes in annular pressures below the faults.

9.2. Similar Pressure Baseline Offsets Due to Drilling Through Overpressured Sands in the Gulf of Mexico

[33] The oil industry [Ostermeier et al., 2000] and IODP [Expedition 308 Scientists, 2006] have experienced substantial increases in annular pressure when drilling through overpressured sands at shallow depths in the Gulf of Mexico. According to Ostermeier et al. [2000], the step function in APWD, similar to that observed at Site C0001, is indicative of the voluminous and sustained flow, which can also be associated with high torque values and has resulted in loss of wells. The inferred fracture network that may source the fluid flow at Site C0001 apparently is a large reservoir, comparable to some of the thick and extensive overpressured sand units in the Gulf of Mexico.

9.3. Hole Abandonment

[34] Abandonment of both holes involved filling with 1.3 specific gravity mud [Kobayashi et al., 2008]. The 1.3 specific gravity mud produces a pressure at 500 mbsf that is probably less than that caused by a flow from the formation. The lesser pressure in the borehole relative to the formation would predict flow at the seafloor. However, remotely operated vehicle observations at both holes indicated no flow.

[35] This inconsistency with our argument for flow from the fractured zone at about 500 mbsf is troubling. Some potential explanations are as follows: It is possible that the holes collapsed, which is the fate of most holes in deformed sediments in accretionary prisms. Additionally, seepage forces caused by fluid flow into the borehole and swabbing effects of raising the pipe may have exacerbated collapse. Moreover, as the pipe was pulled out of the hole, no circulation nor hole cleaning was occurring. Cavings could have accumulated and sealed the borehole.

10. Conclusions

[36] Annular pressure while drilling data across the NanTroSEIZE transect of SW Japan indicates that holes at Site C0001 display large increases in annular pressure at about 500 mbsf that persist to about 1000 mbsf. Seismic data through the site show a series of bright reflectors at the depth of the step-up in annular pressure. The bright reflector zone dips landward (NNW) beneath the thickening accretionary prism. The bright reflectors may represent fluid accumulations tapped by the boreholes at Site C0001. Borehole imaging, sonic velocity and resistivity indicate a fractured zone correlating with the bright reflectors in the seismic data. A hydraulic model of the borehole explains the step up in APWD at Site C0001 by an influx of 3300 l/m of fluid at 500 mbsf. Evidence of high APWD is apparent initially at 375 mbsf and peaks at about 500 mbsf. The transition from the standard APWD gradient to nearly lithostatic APWD in the 375 to 500 mbsf interval may be due sealing by gas hydrate or due to a change to a compressive stress regime that would allow fluid accumulation.

Appendix: Pressure Measurement Variations

[37] The pressure measurements made with the MWD tool show some puzzling systematic variations. The variations do not change our conclusions because they are small relative to the features that we believe are geologically induced. Nevertheless, the issues surrounding these anomalies need to be addressed, as they are apparent when the data is examined closely.

[38] 1. The pressure measurement with the MWD tool (APWD) at the water bottom gives a water density that is about 4% less than a direct density measurement of water at the bottom using the density tool. We presume there is some systematic error in the APWD measurement. For consistency in our graphs, we have used the water bottom pressure from APWD and defined the hydrostatic gradient based on this pressure and depth. Thus, we are comparing downhole pressure data and water bottom/hydrostatic pressures with information from the same tool.

[39] 2. What causes variation in APWD in Hole C0001A verses Hole C0001D at depths less than 500 mbsf? APWD in C0001A is systematically higher that C0001D over this interval. C0001A was a pilot hole drilled with the MWD tool being the only downhole instrumentation, hence there was no requirement to drill slowly to acquire good RAB images. During the first 500 mbsf, C0001A was drilled about twice as fast as C0001D [Expedition 314 Scientists, 2009c] with a similar pumping rate. Thus, the higher-pressure curve at C0001A apparently results from a heavier cuttings load than at C0001D. Rates of penetration and pumping rates were similar below 500 mbsf at both holes allowing direct comparison of the data below this depth. Any additional viscous resistance from the longer BHA at C0001D apparently did not raise the APWD significantly.

[40] 3. What causes inflections of pressure curves toward lower values at 50 to 150 mbsf? Most of the pressure curves show initially higher pressures that inflect to lower values generally at about 50 to 70 mbsf. In order to avoid hole deviation at the start of drilling, the bit is initially rapidly “washed” down by pumping only, with no pipe rotation. Once the bit is seated normal rates of rotation, penetration, and pumping are established. The higher densities prior to the onset of rotation reflect a dense slurry of sediment and pumped fluid that decreases in density once the hole cleans up.

Acknowledgments

[41] We thank IODP member countries for supporting scientific ocean drilling and for their long-term commitment to the NanTroSEIZE project. We thank the drilling and logging personnel of the Chikyu for their careful efforts from which this paper is derived. We also appreciate the input of the engineering staff at the Center for Deep Earth Exploration, Japan Agency for Marine-Earth Science and Technology for their advice and education. Peter Flemings provided important suggestions on use of logs plotted in time to verify pressure patterns. Moore acknowledges the support of the U.S. Science Support Program in the USA for funding to participate in IODP Expeditions. We thank Hess Petroleum Corporation for access to hydraulic modeling software. Comments by Dan Moos, Demian Saffer, and an anonymous reviewer significantly improved the manuscript.

Ancillary