IODP Proceedings    Volume contents     Search

doi:10.2204/iodp.proc.319.103.2010

Downhole measurements

Modular Formation Dynamics Tester results

The MDT wireline logging tool assembly was run to measure in situ pressure and stress (Tables T14, T15, T16; Fig. F59). The tool was operated in two modes: single probe and dual packer. The single probe test (SPT) pushes a probe against the borehole wall, extracts fluid to reduce pore pressure, and records the subsequent pressure recovery. The results are used to estimate in situ pore pressure and permeability. The dual packer test isolates an interval of the formation (1 m at Site C0009) to either draw down the pressure (to determine in situ pressure and permeability over a larger volume than in SPTs) or increase the borehole pressure to create a hydraulic fracture and measure the in situ least principal stress (See "Modular Formation Dynamics Tester" in "Downhole measurements" in the "Methods" chapter). The results are presented below in three sections: SPT results, dual packer drawdown results, and dual packer hydraulic fracture test results (see also C0009MDT.PDF and C0009_T1.XLS in DOWNHOLE in "Supplementary material").

Single probe test results

For each SPT, we plotted the measured pore pressure and fluid withdrawal rate (Figs. F81, F82, F83, F84, F85, F86, F87, F88, F89). A summary of events for each deployment is included in Table T16. The SPTs can be divided into two types: rapid pressure recovery and slow pressure recovery. Test MDT_059 illustrates a rapid recovery (Fig. F81). Initially, the pore pressure recorded mud pressure in the borehole; after this, the probe was set, resulting in a small pressure increase (45–114 s; Table T16; Fig. F81). Three drawdown tests were then performed (drawdown in Fig. F81). In each drawdown, between 4.8 and 9.9 cm3 of formation fluid was extracted at 60 cm3/min (Table T16) and pore pressure declined by ~0.2 to 0.5 MPa. After drawdown, pore pressure rapidly recovered (within 10 s) to a constant value of 28.23 MPa. This final value is lower than borehole pressure (31.20 MPa) (mud pressure in Fig. F81) and is ~0.5 MPa higher than hydrostatic pressure (blue dashed line in Fig. F81). Test MDT_078 exhibits similar recovery behavior (Fig. F88).

In contrast, Test MDT_60 illustrates slow recovery behavior (Fig. F82). In two drawdowns, ~5 cm3 was extracted at 30 and 50 cm3/min, resulting in a total pressure drop of ~10–11 MPa. Pore pressure then gradually rose during the recovery phase. Even after 120 s, pore pressure was still rising (Table T16; Fig. F82).

We used the final build-up pressure at the end of the drawdown cycles as an estimate of in situ (formation) pore pressure (Table T15; Fig. F90A). In the case of slow recovery, the last pressure value is unlikely to reflect the true in situ pore pressure and is probably most appropriately considered a minimum. In a pressure-depth plot, the final build-up pressures generally plot on a straight line that is very slightly above estimated hydrostatic pressure (compare red triangles to solid blue line in Fig. F90A) (Table T15). The deepest SPT measurement has an estimated in situ pore pressure that is less than the hydrostatic pressure. However, pore pressure was still rapidly rising at the end of the test (Fig. F86) and we infer that pore pressure did not reach the in situ pore pressure during the recovery. Hydrostatic pressure is calculated by assuming a pore fluid density of 1023 kg/m3. The slight apparent overpressure could be due to (1) the influence of drilling mud on the surrounding formation pressure, (2) an underestimate of hydrostatic pressure, or (3) the presence of actual slight overpressure.

Fluid mobility is calculated according to Equation 36 in "Downhole measurements" in the "Methods" chapter. It is important to note that Schlumberger suggests that fluid mobility values <10–15 m2/cp may be unreliable. The drawdown mobilities are presented in Table T15 and Figure F90, where mobility is defined as the ratio of permeability to viscosity. Mobility ranges from <10–16 to 10–14 m2/cp (Table T15; Fig. F90). Generally, mobility is lower where gamma ray values are higher (Fig. F59A). For example, five SPT measurements were made in Unit III (Fig. F59A). Three of these were made at approximately the same depth (~875 m WMSF), have low mobility, and are in a zone of relatively high gamma ray values (Fig. F59A). Near the base of Unit III, one SPT measurement records a relatively high mobility and is in a relatively low gamma ray interval (Test MDT_065). Test MDT_065, which has high mobility, was conducted at a depth (1175 m WMSF) where there was an anomalously low sonic velocity (Fig. F59A). The underlying measurement (Test MDT_078) at 1217 m WMSF records low mobility and is in a zone of high gamma ray values. The three SPTs in Unit IV all record low mobility and have generally high gamma ray values (Fig. F59A).

Dual packer drawdown test results

Test MDT_073 was a dual packer drawdown test (Tables T14, T16; Figs. F91, F92). At 1539.69 m WMSF (tool zero depth), the MDT's dual packers were inflated to isolate a 1 m section where the borehole was relatively smooth. The test location was selected based on criteria summarized in the "Methods" chapter. The packers were first inflated (1630–2000 s; Fig. F91). Then pressure generated by the packer inflation was bled off (2000–2800 s) and a drawdown test was performed (~3200–3450 s). It was planned to draw the pressure down by 5 MPa. However, because of initial concern over borehole stability conditions, the drawdown was stopped after pressure had declined by 3.52 MPa (Figs. F91, F92). Then pressure was allowed to recover for 180 s (Figs. F91, F92).

Pressure recovery from the MDT dual packer drawdown test is analyzed using the curve-matching approach of Papadopulos et al. (1973), which uses an analytical solution in dimensionless coordinates wherein fractional pressure dissipation ([PPd]/[PiPd]) is plotted against the log of dimensionless time (β) for different dimensionless storage parameters (α) (Fig. F93). P is pore pressure during the recovery, Pd is pore pressure at the start of the recovery, and Pi is in situ pore pressure. The dimensionless time (β) is:

β = πkLtC, (3)

where

  • k = permeability,

  • L = length of the open section,

  • t = time since the onset of the buildup,

  • µ = fluid viscosity, and

  • C = system compressibility.

The dimensionless storage parameter (α) is:

α = πrw2LSs/C, (4)

where

  • rw = well bore radius, and

  • Ss = specific storage.

The data from the recovery part of the drawdown test are matched with type curves (plotted in log time) from the Papadopulos et al. (1973) solution. We assume Pi = 39.321 MPa, Pd = 35.621 MPa, and other parameters are as presented in Table T17. Our data most closely match when α = 10–1 and β = t/1000, which define a storage coefficient (Ss) of 5.2 × 10–11 Pa–1 (Equation 4) and a permeability (k) of 1.3 × 10–17 m2 (Equation 3). These calculations should be viewed with caution because system compressibility (C) was calculated only from the water stiffness (Kf = 2 GPa): the water volume (V) in the 1 m long 12¼ inch borehole is 0.0760 m3; thus, C = V/Kf = 3.8 × 10–11 m3/Pa. If the tool itself had significant compressibility, the true storativity would be lower and the resulting permeability would be higher. In addition, in situ pressure (Pi) was assumed to equal pressure in the borehole (39.21 MPa). This pressure is most likely too high (Fig. F92), and if a lower pressure were used, it would also result in a higher calculated permeability.

Dual packer hydraulic fracture test results

Hydraulic fracture tests using the MDT dual packer were carried out near the top (Test MDT_080) and bottom of the 12¼ inch open hole section (Test MDT_074) (Fig. F59). As for the dual packer drawdown test, test locations were selected on the basis of other logging data (see the "Methods" chapter). Test MDT_074 is located within the cored section (Core 319-C0009A-3R) of Unit IV at 1532.7 m WMSF (Tables T14, T16). Test MDT_080 at 873.7 m WMSF was the last test in wireline logging Run 3.

Figure F94 illustrates the pressure and injection rate during hydraulic fracture Test MDT_074. From 1400 to 3200 s, fluid was injected in a series of pulses. After each injection, the pumps were stopped and pore fluid pressure fell. At ~3200 s, pressure reached 41.9 MPa and then the pumps were stopped for 600 s to change the setting of pumping because the pressure continued to increase without breakdown and the rate of pressure increase was too slow. After changing the pumping parameter to use dual pump, we resumed pumping with an interruption at 3960 s before the final injection from 4017 to 4207 s. Pressure immediately after this final injection was 41.6 MPa. We take this pressure to be an instantaneous shut in pressure (ISIP) and thus tentatively represents the least principal stress magnitude (Fig. F94). However, it should be emphasized that the ISIP is not reproducible, and therefore we do not have confidence in this estimate of least principal stress.

The MDT tool was moved to 878.7 m WMSF for Test MDT_080. In this test, the pressure cycle was repeated five times maintaining the same flow rate (Fig. F95). The pressure-time curve for the last four cycles did not change significantly. For each of these cycles, we report the ISIP in Table T18. Figure F95B illustrates an example of these values for one of the cycles. ISIP was determined by finding the break in slope on a pressure-time plot after injection ceased. Based on these four values, we estimate ISIP to be 34.8 MPa. ISIP is commonly interpreted to record the least principal stress (σ3).

Least principal stress magnitudes for Tests MDT_074 and MDT_080 are shown in Figure F90A. We place much more confidence in the measurement from Test MDT_080 (873.7 m WMSF) than the measurement from Test MDT_074. For each measurement, we calculate the effective stress ratio (K) by

K = (σ3Ph)/(σvPh), (5)

where

  • σ3 = least principal stress,

  • Ph = hydrostatic pressure (calculated assuming a pore fluid density of 1023 kg/m3), and

  • σv = overburden stress (see Fig. F90A caption for explanation of overburden calculation).

For Tests MDT_080 and MDT_074, K = 0.82 and 0.44, respectively. We do not know the orientation of fractures induced or activated by hydraulic fracturing because no borehole images were taken after the hydraulic fracture tests. However, we can verify that σ3 < σv at both test depths (Figs. F94, F95).

Leak-Off test

As a standard part of drilling operations, two LOTs were performed at the base of the 20 inch casing (2786.2 m DRF, 703.9 m DSF). As defined in "Logging and data quality" (Fig. F82), the base of cement lies at 2790.9 m DRF (708.6 m DSF). The bottom of the 17 inch hole at this time (after drilling out the cement plug) was at 2798 m DRF (715.7 m DSF). The mud density in the hole was 1080 kg/m3. The LOT was performed with the outer annulus closed by the BOP, and pressure was measured at the cement pumps. A summary of operations is included in Table T19, and the pressure-time data are included in a supplementary data table (see C0009_T2.XLS in DOWNHOLE in "Supplementary material").

Figure F96 illustrates volume injected versus pressure during the LOTs. We interpret leak-off pressure to be at the break in slope on the pressure-time plot (Zoback, 2007). Leak-off pressures were found to be 100 psi (0.689 MPa) and 105 psi (0.724 MPa) for LOT 1 (Fig. F96A) and LOT 2 (Fig. F96B), respectively. The pressure of the mud column to the leak-off depth (2790 m DRF, 712.9 m DSF) is 29.53 MPa. Thus, downhole leak-off pressures are 30.22 and 30.25 MPa, respectively, for LOTs 1 and 2. These values are interpreted to equal least principal stress (Fig. F90). We note that there is considerable uncertainty in picking the slopes of the lines to determine least principal stress (σ3) (e.g., Fig. F96B). The stress ratio (K) is 0.44 for both LOT measurements of σ3.

Downhole temperature

Borehole temperature logs were acquired during each wireline logging run (drilling Phases 5, 6, and 7) (Fig. F97; see also C0009_T3.XLS in DOWNHOLE in "Supplementary material"). The EMS tool (see "Logging" in the "Methods" chapter) recorded borehole temperature during wireline logging Runs 1 and 2 from the 20 inch casing shoe (703.9 m WMSF) to ~1585 m WMSF (Fig. F97). Borehole temperatures are higher in Run 2 than in Run 1 throughout the borehole. Wireline logging Run 2 was initiated ~14.5 h after Run 1. There was no circulation between the two logging runs. After wireline logging Run 2 and before Run 3, there was circulation for a wiper trip to clean the hole. The MDT tool was run during wireline logging Run 3 (see "Operations," Table T1), and temperature was recorded for each MDT test (circles and squares in Fig. F97). Temperatures recorded as the MDT was run into the hole are lower than temperatures recorded when the tool was being pulled out of the hole (compare circles to squares in Fig. F97).

Prior to wireline logging, circulation during drilling cooled the borehole and the near-borehole region. We interpret that the three wireline logging runs record the borehole temperature gradually equilibrating with the surrounding formation. The low values recorded by the MDT during tool lowering are due to either circulation during the wiper trip between Runs 2 and 3 or to the thermal inertia of the tool. Measured temperature increases with depth in all three runs, and there are two changes in temperature gradient at ~720 and 1300 m WMSF (Fig. F97). These depths are very close to the boundaries of logging and lithologic units (see "Lithology"), and the changes in gradient may be related to thermal conductivity contrasts at unit boundaries.

Vertical seismic profile

Walkaway VSP experiment at Site C0009

The two-ship, or walkaway, VSP experiment differed from most studies in industry and from preceding walkaway experiments in scientific ocean drilling in that it was the first walkaway experiment in IODP riser drilling and it employed an array of seismometers. In IODP, a wireline downhole seismic receiving system with up to 20 geophones is available, more geophones than commonly used during industry applications. The scientific objective during Expedition 319 was to illuminate deep reflectors with long offsets using wide-angle (refraction) information to measure bulk physical properties below the borehole. Improved constraints on physical properties of the lower plate are desired to better constrain earthquake seismographs. In industry, reflection seismic imaging in a volume around a well is the common objective.

Because the use of new techniques involved unanticipated problems during riser drilling, flexibility and multiple changes in the scheduling of the shooting ship Kairei were required and impacted the Kairei's other scheduled cruises. These changes were necessary because the Kairei was the preferred shooting ship for the walkaway VSP experiment with its powerful tuned array of air guns. The scheduling delays were linked to unanticipated operational problems and delays due to weather (see "Operations"). Therefore, the time window for the walkaway VSP shifted up until the final cementing of the 13⅜ inch casing. After tools for the walkaway VSP were in the hole, cable failures between seismometers caused an additional 7 h delay while the cable was pulled out of the hole to make repairs. The Kairei also experienced problems with the strong Kuroshio Current that slowed its operations so as to stay within the required array tow speed. During operations, the plan for the walkaway VSP experiment was continually revised. Ultimately, we obtained 880 shot records in Line 1 and 275 shots in circular line Circle 1 (Fig. F98).

Seismometer orientation

For the walkaway VSP, the VSI seismometers were clamped to the casing (Fig. F99). The horizontal first motion records from circle-shooting data (Fig. F100) clearly show a dependency of polarity and amplitude on the direction of the air gun shot (Fig. F101). The orientation of each VSI seismometer was not controlled because the seismometers were connected by wireline. However, using first-motion polarity and amplitude, the orientation of each seismometer can be estimated by finding the direction of maximum negative amplitude of the first motion in the x- and y-components. Preliminary results of this estimation are shown in Table T20. The results suggest that the upper 11 seismometers were rotated relative to one another and that below this the seismometers (12–20) were consistently oriented. All seismometers showed similar response except for one at 3126 m WRF (1044.9 m WMSF) (Seismometer 10). This seismometer had a much smaller response to the air gun signals for all components, and irregular responses to signals were observed, suggesting poor coupling of the seismometer to the casing (or of the casing to the formation). The seismometers exhibited relatively stable responses over a wide range of shot directions, so first-motion amplitude followed a cosine relationship very closely (Fig. F101) and the differences in estimated azimuth of seismometer axes (x- and y-) were in most cases close to 90° (Table T20). It is not clear why the x-azimuth is estimated to be 90° clockwise from y, whereas Schlumberger claims the opposite (i.e., that the y-azimuth is 90° clockwise from x).

Walkaway VSP records

Seismic records from walkaway VSP Line 1 exhibit various types of seismic waves. The range of horizontal offset between sensors and air gun shooting spans from 400 m to 29.3 km. In these ranges, we are able to identify direct wave arrivals, seismic phases associated with multiples in the ocean, refractions from accretionary prism layers, and reflections from interfaces below the seismometers within the accretionary prism as well as from the splay fault and presumably from the deeper décollement (Figs. F102, F103). Apparent velocities of major refracted waves (which correspond to regions immediately below velocity discontinuities) are 1.75–1.8, 1.89–1.9, 2.45–2.6, 3.05, and 4.0 km/s. Seismic phases from deeper boundaries are not clear from records for single seismometers because of noise, but reflections of apparent velocity of 6.6–6.8 km/s are present. We may be able to identify seismic phases from the deeper décollement by stacking the borehole seismic array records because arrivals of upgoing refracted and reflected waves from the shallower boundaries in the array record (indicated by Areas A and B in Figs. F104, F105) show good coherence over the length of the array. We can identify deep seismic phases better in shot records from the southern part of the linear transect than the northern part. This is probably due to (1) increased background noise level and reduced source strength in the northern shots and (2) the possible effect of boundaries dipping to the northwest. Background noise appeared to increase in northern shots of offset >10 km, but it is not known why. The amplitude of the air gun signal was reduced for shots north of 10.7 km offset because Kairei experienced air gun problems in one of the four subarrays and continued shooting with the remaining three subarrays.

Deeper and weaker seismic phases appear clearer in horizontal component records than the vertical records because of the effects from seismic waves traveling vertically along the casing, as explained in detail below. Horizontal records from each seismometer were rotated to decompose the radial and transverse directions from the air gun shot. Figure F103 shows rotated horizontal records from the borehole seismometer at 3004 m WRF (922.9 m WMSF). The amplitude of signals in the transverse component is much smaller than in the radial component for most arrivals. In radial horizontal component records, some phases are clearly observed ~1 s after refracted P-wave arrivals, which have similar apparent velocities to the preceding P phases. We interpret these later arrivals as P to S converted waves from boundaries below the borehole.

Records from the seismic array provide a distinct advantage over single seismometer records, in that phase identification of different seismic waves is possible by differentiating their propagation direction and speed. In the walkaway VSP data set from 16 seismometers spaced at ~15 m intervals, upgoing and downgoing waves are clearly distinguishable (Figs. F104, F105). In these figures, upgoing refracted waves dominate in Area A, whereas in Area B upgoing reflected waves and downgoing direct waves are mixed. Velocity filtering on the data set would help illuminate seismic phases of primary interest, including upgoing reflections and refractions from deep boundaries. These seismic phases are apparent at frequencies <20 Hz for vertical, and <30 Hz for horizontal records (e.g., index A and B in Figs. F104, F105).

In addition to these phases, there are other types of waves observed in the walkaway VSP records, which are generally treated as noise in data processing. In near-offset recordings, we can identify an irregular phase dominated in high frequencies (>20 Hz) preceding the direct wave arrivals through the formation (C in Fig. F104). This phase has high apparent velocity in the array (~6 km/s) and is coherent in amplitude over the array for each shot but changes irregularly in amplitude for different shots. It probably corresponds to subordinate seismic waves generated by direct waves hitting the structure of the borehole in some places, but it is not clear why these waves behave so differently for air gun shots separated by only 60 m. We also observe later phases abundant in high-frequency energy prior to the direct wave arrival (D in Fig. F104). These phases have similar amplitude between adjacent shots and also a similar apparent velocity of ~6 km/s. The same type of seismic phase was also observed in the zero-offset VSP experiment. From the observations in zero-offset VSP data, it seems likely that this signal is generated at several boundaries in the borehole from direct waves in the formation. Some of these boundaries correspond to mechanical elements of the borehole construction, such as the wellhead and the bottom of the 36 inch casing, but others are not correlated to casing shoes or other borehole boundaries and might be related to formation boundaries. The phase (D in Fig. F104) also has horizontal energy as indicated by D in Figure F105. On the other hand, we observe an earlier phase (E in Fig. F105), which is prominent only in the horizontal component. We also observed a later phase (F in Fig. F105) in the horizontal component, which has an apparent velocity similar to that for water. The later phase may be a sound wave traveling down to the bottom of the casing, and we see a similar phase in zero-offset VSP records. In the zero-offset VSP records we also identified seismic phases that reflected back from a depth close to the bottom of the 13⅜ inch casing, with an apparent velocity near that of water. We also observed that seismic waves are sometimes generated within the interval of the seismic array and propagate both upward and downward with an apparent velocity near that of water; we interpret this as a wave in the mud within the hole (G in Fig. F105).

Zero-offset VSP experiment at Site C0009

A zero-offset VSP (check shot) was carried out following the walkaway VSP. Zero-offset seismic data were obtained by moving the seismic array upward from the depth of the walkaway VSP experiment, at ~121 m intervals, until the bottom of the array reached a depth near the seafloor. Eight shuttles were used to acquire the waveforms. A three gun (each 250 in3) air gun array deployed from the Chikyu was used as the seismic source. A total of 94 shots were fired. Generally, 10 shots were fired for each seismic array position.

Zero-offset VSP records

The zero-offset VSP data contain many seismic phases related to the structure of the borehole. The earliest phases observed in the vertical component are vertically oriented waves propagating downward from the wellhead at an apparent velocity of 6 km/s (A in Fig. F106A). These probably correspond to subordinate seismic waves generated at the wellhead and propagating along the casing. The seismic wave may be reflected upward at the bottom of the 20 inch casing (B in Fig. F106A), but some of the energy propagates to the bottom of the hole. There is another seismic phase ~0.1 s later that propagates downward at a similar velocity (C in Fig. F106A). This phase also appears to be a converted wave originating deeper (probably at 2250 m WRF [168.9 m WMSF]), again from acoustic waves propagating inside the casing. Another set of similar strong phases (D in Fig. F106A) are generated at ~2550 m WRF (468.9 m WMSF) and comprise the strongest arrivals at these depths. We can also see strong phases (E in Fig. F106A) propagating upward. These phases may be reflections from the top of sections of casing that are well coupled to the formation by cement at ~3500 m WRF (1418.9 m WMSF). There is a lower velocity phase (F in Fig. F106A) between 2600 and 3217 m WRF (518.9 and 1135.9 m WMSF). This phase has an apparent velocity similar to that of water. This is probably an acoustic wave propagating in the mud inside the casing. A seismic wave reflecting back from ~3600 m WRF (1518.9 m WMSF) is also observed (G in Fig. F106A); this corresponds to the depth of float collar near the bottom of the 13⅜ inch casing.

These seismic phases are all very weak in horizontal component records (Fig. F106B, F106C). There are some clear phases propagating down with apparent velocities close to those obtained from sonic logs in the horizontal component records (thick line in Fig. F106B, F106C). We picked these arrivals as the formation interval velocities. It is not clear why these phases are apparent only in horizontal components and not in the vertical. In such near-offset VSP data (offset of 60 m), it is difficult to explain why compressional waves in the formation have such a large horizontal amplitude. Therefore, the picked interval velocities may not represent P-wave velocity of the formation but instead could represent a type of surface wave propagating along the borehole, although we deem this scenario unlikely because the velocities are too high to be surface waves.

At some depths, records are dominated by monotonic oscillation (H in Fig. F106A, F106B, F106C). This is probably due to poorly coupled sensors. The monotonic oscillation exhibits higher amplitude coincident with the arrival of acoustic waves in the casing (vertical component), and the oscillation is much weaker in the horizontal component. This observation shows that the seismometers are coupled to fluid inside the casing rather than to the casing itself.

Analysis on velocity pick for zero-offset VSP data

There was difficulty in picking the compressional wave arrival times. An initial report from Schlumberger (see C0009VSP.PDF in DOWNHOLE in "Supplementary material") described the zero-offset VSP and picked these arrival times. Although the report was very useful, it was determined that the velocities were far too high to be physically reasonable. Schlumberger revisited this issue and then provided the data in Table T21 as an estimate of the compressional wave traveltime. They interpreted the initial (very high) velocities (reported in C0009VSP.PDF in DOWNHOLE in "Supplementary material") as first arrivals of waves traveling along the casing. Schlumberger's new calculations were based on both the vertical and horizontal geophone records (see previous discussion).

The final velocity structure provided (Table T15; Figs. F107, F108) is in good agreement with wireline seismic data and the velocities from precruise seismic processing of the 3-D seismic reflection data.