Previous   |   Next

doi:10.2204/iodp.proc.319.103.2010

Discussion and conclusions

Geomechanics: stress and pore pressure

We collected several data sets during riser drilling at Site C0009 that constrain both in situ stress orientation and magnitude. We also identified faults and fractures in cores and resistivity images that provide insight into deformation history (see "Structural geology"). Borehole breakouts inferred from wireline calipers in Unit IV (1285–1579.9 m WMSF) indicate that the maximum horizontal stress (S_Hmax) is oriented northwest–southeast (Fig. F38). DITFs in borehole resistivity images ~300–500 m uphole in Unit II (~800–1000 m WMSF) are also compatible with a northwest–southeast oriented S_Hmax. This orientation for S_Hmax is ~90° to that at Site C0002, located ~20 km seaward in the Kumano Basin (Fig. F115) (Kinoshita, Tobin, Ashi, Kimura, Lallemant, Screaton, Curewitz, Masago, Moe, and the Expedition 314/315/316 Scientists, 2009). The emerging picture of stress state across the margin is that S_Hmax is slightly oblique to the plate convergence direction (but approximately perpendicular to the trench) in the outer accretionary wedge (Fig. F38) and rotates ~90° within the seaward-most 15–20 km of the Kumano Basin where there is active northwest–southeast extension. Landward of this, S_Hmax rotates back to an orientation nearly perpendicular to the trench and similar to that on the outer slope.

We also obtained direct measurements of least principal stress magnitude (σ₃) from a MDT hydraulic fracture test at 879 m WMSF and an LOT at 704 m WMSF. In both cases, the vertical stress (σ_v) is greater than the least principal stress. If we assume that the principal stresses are horizontal and vertical, then σ₃ is the minimum horizontal principal stress and the vertical stress is either the maximum or intermediate principal stress, depending on the value of the maximum horizontal stress. Taken together, the breakouts, DITFs, and in situ stress magnitude data indicate either a normal or strike-slip faulting regime. In the case of normal faults, the dominant strike would be northwest–southeast. The ratio of effective stresses (σ′_hmin/σ′_v) is significantly greater for the MDT measurement than the LOT, but we consider the MDT measurement to be slightly more reliable.

In addition to in situ stress indicators, we used resistivity images, seismic reflection data, and cores to document fault types and orientations as constraints on stress and deformation history (Fig. F42). The relative timing of different phases of faulting can be determined for some structures, but recent fault activity cannot be confirmed, except potentially where faults cut the seafloor in seismic reflection data. Fractures in FMI resistivity images trend northeast–southwest (including one documented normal fault) and we documented a range of fault types, crosscutting relationships, and orientations in cores within Unit IV. In seismic reflection data close to Site C0002, normal faults trending northeast–southwest are common, and a second less prevalent set trending northwest–southeast is also observed; many of these offset the seafloor (Moore et al., 2009) (Fig. F115). Recently active normal faults are less common in the landward part of the Kumano Basin near Site C0009, and the northwest–southeast trending set becomes slightly more common relative to the northeast–southwest trending set. The presence and orientation of normal faults in seismic reflection data (northwest–southeast trend dominant) are generally consistent with the in situ stress magnitude and orientation data. However, the dominant northeast–southwest strike of faults and fractures we documented in FMI images is not: if these structures are normal faults, they are inconsistent with the orientation of S_Hmax; whereas, if they are thrusts, they are inconsistent with the fact that S_hmin < S_v.

There are several possible explanations for this range of structural and geomechanical observations. One possibility is that the structures identified in FMI images are inactive and therefore do not relate directly to the present-day stress regime. This explanation is consistent with the fact that most of the fractures and faults measured in FMI data are resistive, suggesting they may not be active or critically stressed (e.g., Barton et al., 1995). A difference between past and in situ stress states could be related to variations in stress during the earthquake cycle or to long-term processes related to the migration of deformation. At Site C0002, long-term strain and modern stress indicators are in agreement and indicate extension normal to the margin (northwest–southeast). This state of stress could have existed at Site C0009 during an earlier phase and would explain the orientation of faults observed in FMI data.

A second explanation is that structural observations and downhole measurements at different depths represent real changes in stress regime downhole. In this case, the stress regime would be consistent with normal faulting in the upper ~900–1200 m WMSF and transition to one of lateral compression below this, perhaps across the boundary into Unit IV. Third, it is possible that the two horizontal principal stresses are very close in magnitude in the landward portion of the basin near Site C0009, such that σ_hmin ≈ σ_Hmax < σ_v (i.e., σ₃ ≈ σ₂ < σ₁). This stress state would allow normal faulting on structures as observed in the seismic reflection data and FMI images, while also honoring the stress measurement data that indicate σ_hmin < σ_v. However, this hypothesis would not explain the observation of only one dominant (northeast–southwest) trend for structures in the FMI resistivity data.

Forearc basin development and correlation to Site C0002: depositional and tectonic environment

Our interpretation of new data from Site C0009, evaluated in the context of previous results from drilling in the Kumano Basin (Ashi et al., 2008), parallels the interpretation of geological and tectonic evolution initiated by the Expedition 314 and 315 scientists (Expedition 314 Scientists, 2009; Expedition 315 Scientists, 2009). Four lithologic units were described at both sites. These units (Units I and II taken together, Unit III, and Unit IV) comprise three depositional sequences (Fig. F115, F116; see "Background and objectives" and "Cuttings-Core-Log-Seismic integration" for discussion of regional seismic surfaces).

At both Sites C0009 and C0002, Unit IV is composed of mudstone with thin-bedded, fine-grained turbidites. At Site C0002, the unit is significantly deformed and has limited evidence of carbonate deposition, interpreted as deposition near or below the CCD. This unit was interpreted as accretionary prism material by Expedition 314 and Expedition 315 scientists (Kinoshita, Tobin, Ashi, Kimura, Lallemant, Screaton, Curewitz, Masago, Moe, and the Expedition 314/315/316 Scientists, 2009). At Site C0009, Unit IV is weakly deformed and contains carbonate. Despite the presence of carbonate deposition and only modest deformation at Site C0009, Unit IV could represent frontally accreted prism sediment. Alternatively, this unit may represent trench-slope deposits or sediments deposited in the early Kumano Basin. Above Unit IV, Unconformity UC2 has >1000 m of relief between Sites C0009 and C0002 and marks a hiatus of approximately equal age and duration at both sites (~5.6–3.8 Ma) (Figs. F115, F116). This suggests a tectonic event of regional significance potentially related to the onset of out-of-sequence thrusting in the prism, underplating or ridge subduction.

Unit III at Site C0009 is ~5 times the thickness of Unit III at Site C0002, although it spans only about twice the duration (Fig. F116). At Site C0002, Unit III is interpreted as a condensed mudstone section deposited in the early forearc basin (Ashi et al., 2009). At Site C0009, Unit III is composed of two subunits (IIIA and IIIB) and is distinguished from Unit III at Site C0002 by the presence of silt and ash beds and abundant wood and lignite (Figs. F13, F19). Based primarily on the biostratigraphy, we interpret Subunit IIIB at Site C0009 as laterally equivalent to Unit III at Site C0002, representing early (unconformable) forearc basin or slope deposits.

Seismic Surface S2 separates laterally continuous downlapping reflections above from a more acoustically transparent unit of variable thickness below (Fig. F116A, F116B). Downlap proceeded with an apparent direction from the south-southeast to the north-northwest. Based on this observation and biostratigraphic constraints that indicate the underlying sediments are younger at Site C0009 than at Site C0002, we interpret Surface S2 as a time-transgressive downlap surface, which youngs to the northwest (landward). The apparent downlap angle is steepest to the south-southeast (near Site C0002) and gradually flattens to the north-northwest (toward Site C0009). Seismic Reflectors S-A and S-B downlap Surface S2 near Site C0009 (Fig. F116B). Above this, Units I and II define a package of sediments grading upward from mudstone (Unit II) to interbedded mudstone and sandstone (Unit I). This succession was deposited in both locations; however, it is greatly expanded at Site C0009 (Fig. F115). These strata record infilling of the Kumano Basin and the progressive migration of sediment deposition to the northwest, possibly related to slip on the megasplay fault.

Plate boundary structure from the walkaway VSP experiment

A collaborative effort between IODP CDEX and Japan Agency for Marine-Earth Science and Technology enabled a long-offset two-ship walkaway VSP experiment using an air gun source towed by the Kairei, along with a zero-offset VSP using a source at the drillship, in both cases shooting to receivers within the borehole. The walkaway VSP trackline consisted of a single line crossing the borehole, with offsets up to 30 km, and a circular trackline around the borehole of ~3.5 km radius to investigate anisotropy. The long offsets allowed refractions and reflections from the accretionary wedge, plate boundary, and subducting plate to be recorded at the wireline seismometers within the borehole to ~1200 m DSF. Recording arrivals in the borehole environment provides a higher resolution image than surface ship and OBS acquisition because the seismometers are coupled to stiff and less attenuative formation; this configuration also allows high-fidelity measurement of shear waves converted from P-waves at formation boundaries. The data will allow seismic analyses of the velocity structure of the subduction zone forearc and the seismic attributes of the plate boundary in the region beneath the borehole at a depth of ~10–12 km.

Insights from scientific riser operations

Expedition 319 was noteworthy in that it marked the first riser drilling in IODP history. As noted above, this allowed us to conduct several scientific operations new to IODP, including measurement of in situ stress, permeability, and pore pressure; real-time analysis of mud gases; and analysis of cuttings to obtain sedimentological, chemical, and physical property data. Here, we briefly discuss key insights gained from each of these operations that may provide guidance in planning future scientific riser drilling.

Stress, permeability, and pore pressure from MDT and LOT measurements

In Hole C0009A, we deployed the MDT wireline logging tool to measure in situ formation pore pressure, permeability, and minimum principal stress magnitude (σ₃) at several depth intervals. This was the first time that the tool had been used in IODP drilling because it is currently not usable with the small-diameter riserless borehole drilled by IODP. We conducted nine SPTs to measure formation pore pressure and fluid mobility, and three dual packer tests (one drawdown test to measure formation hydraulic properties and two hydraulic fracture tests to measure least principal stress magnitude) (Fig. F90). Successful future deployment of this tool to measure in situ pore pressure and stress magnitude deeper within the accretionary prism and in the vicinity of major fault zones will constitute a major breakthrough in understanding subduction zone fault mechanics and is a critical part of the Nankai Trough Seismogenic Zone Experiment (NanTroSEIZE) program. However, there are limitations in using the MDT tool successfully that warrant consideration, most notably that pore pressure and permeability measurements may be unreliable in low-permeability formations (k < ~10^–15 m²) because the time required for pressure equilibration can greatly exceed deployment time. This also highlights the value of obtaining FMI or other borehole imaging data prior to running the MDT in order to select measurement targets that will yield meaningful data.

Sampling and analysis of cuttings and mud gas

Samples from riser drilling mud, including cuttings and mud gas, were collected for the first time in IODP history during riser drilling in Hole C0009A. The results of shipboard analyses may provide guidance for future IODP riser drilling by demonstrating the types and quality of data that can be obtained from these materials, especially for shallow sediments and sedimentary rocks. In addition to basic lithologic description and biostratigraphic analysis, for sufficiently lithified cuttings (deeper than ~1000 m DSF in ~2.5 Ma claystones) we were able to document deformation structures preserved in large (> ~2 mm) cuttings, measure physical properties including porosity and density, and quantify composition by XRD, XRF, and carbonate analyses. We also conducted several experiments on cuttings and core samples to understand the sensitivity of shipboard measurements to cuttings processing techniques (including the composition of fluid used for washing and soaking time), the cuttings size fraction(s) used for analysis, and drilling mode and mud composition.

We identified several significant challenges in conducting analyses on cuttings. Comparisons between cuttings (collected from 5 m intervals) and wireline log data indicate a relatively small depth uncertainty of ~10 m, but mixing during mud ascent and/or cavings from uphole may occur over larger distances. We also found that cuttings may be preferentially preserved from particular lithofacies (e.g., more lithified or cohesive claystones) and thus provide a biased sampling of the formation because material from unconsolidated sands or mudstones was lost by disaggregation and could not be separated from drilling mud. Finally, we noted that absolute values for many measurements, including porosity, bulk density, and sediment geochemistry, do not accurately reflect formation properties because of physical and/or chemical processes during circulation, retrieval, and laboratory handling. We conclude that sedimentological, physical property, and geochemical data from cuttings are useful for some applications such as defining overall lithofacies, age, characterization of mud rock provenance, basic chemistry, and first-order compositional variations. Data from cuttings are probably less meaningful for other applications such as grain size assessment, detailed characterization of lithofacies, porosity determination, or detailed chemical analyses.

As one example of these issues, the high pH of the drilling mud and the mud composition (including additives to reduce mud loss) strongly affect measurements of carbonate, calcium content, and some other elements, as observed in XRD and XRF analyses. This also impacts relative mineral and element abundances from XRD and XRF. These effects are particularly clear when comparing measurements from cuttings and cores within the cored interval. Chemical contamination appears to be correlated with the time cuttings were exposed to drilling mud in the borehole (Fig. F20). Similarly, porosity values from cuttings are anomalously high, both with respect to their depth of origin and in comparison to log and core data, and this difference is affected by sample handling and soaking procedures (Figs. F61, F62; see also "Physical properties"). We conclude that the absolute values of compositional and physical property data from cuttings should be treated with caution but that some downhole trends may still be reliable.

We also monitored mud gas chemistry throughout riser drilling operations to document the composition and concentration of gas released from the formation as it was drilled. This method, used previously in International Continental Scientific Drilling Program drilling (e.g., Erzinger et al., 2006), was used in scientific ocean drilling for the first time during Expedition 319. One example of the value of these data comes from comparison of the mud gas, cuttings, and wireline logging data in Subunit IIIB. Increased mud gas methane concentrations are clearly correlated with increased wood content in the cuttings and with several intervals of low V_P/V_S and Poisson's ratio observed in sonic velocity logs (Figs. F76, F77, F80; see also "Cuttings-Core-Log-Seismic integration"). Because pore water geochemical analyses are not possible on cuttings and are difficult on strongly lithified or low-porosity core samples, mud gas analysis is a promising approach for characterizing formation fluids in future deep riser drilling. Such data are important toward understanding hydrologic and geochemical processes associated with faulting and fluid flow.

Top of page   |   Previous   |   Next