Apparent overconsolidation of mudstones in the Kumano Basin of southwest Japan: Implications for fluid pressure and fluid flow within a forearc setting

Authors


Abstract

[1] The Kumano Basin is located in the Nankai Trough subduction zone of southwest Japan. During the past 1.6 million years, approximately 800 meters of sandy turbidites and hemipelagic mud were deposited near the distal edge of the forearc basin, at Site C0002 of the Integrated Ocean Drilling Program. Constant-rate-of-strain consolidation tests yield estimates of in situ permeability that range from 2.6 × 10−17 m2 to 2.5 × 10−18 m2; overconsolidation ratios range from 1.7 to 2.6, and values of the compression index range from 0.39 to 0.78. Several processes contributed to the apparent overconsolidation. Strata dip toward land, and pore fluids probably migrate up-dip and vent along a bathymetric notch near the seaward edge of the basin. Efficient lateral drainage through sandy turbidites has kept pore pressures within interbeds of mudstone at (or close to) hydrostatic. In addition, alteration of dispersed volcanic glass, precipitation of authigenic clay minerals, and collapse of random grain fabric has probably strengthened the bonding among grains. Cementation is particularly likely within the lower basin (unit III), where values of porosity remain anomalously high. If fluid overpressures (and underconsolidation) exist anywhere within the basin, the most likely loci are where sandy turbidites terminate against impermeable mudstones along landward-dipping on-lap surfaces. Those types of on-lap geometries, in addition to structural closures, might provide promising targets for oil/gas accumulation in other forearc basins, particularly where petroleum source rocks have been buried to the optimal depths of catagenesis.

1 Introduction

[2] Most subduction zones contain forearc basins between the magmatic arc and the frontal accretionary prism [e.g., Dickinson and Seely, 1979; Dickinson, 1995]. Their architecture and evolution are influenced by a complicated interplay of protracted, subduction-induced deformation, erosion and transport of sediment from the nearby arc, eustatic sea-level fluctuations, subsidence, and uplift along imbricate and out-of-sequence thrust faults [Gulick et al., 2002; Susilohadi et al., 2005; Melnick and Echtler, 2006; Berglar et al., 2008; Contardo et al., 2008; Paquet et al., 2011]. Typically, the lowermost forearc deposits rest above a regional-scale unconformity. Tectonism often expands accommodation space for prograding deltas and turbidite fans, and episodes of rapid subsidence can lead to thick accumulations of sediment behind the outer-arc high (otherwise known as the trench-slope break). The sedimentary history of these thrust systems is difficult to predict in detail, however, because climate, sediment supply, and tectonics interact in many different ways [Stolar et al., 2006; Roe et al., 2008]. In fact, modeling indicates that the rate of forearc sedimentation exerts more influence on wedge taper and seismic coupling of the subduction megathrust [Fuller et al., 2006] than the other way around.

[3] Although exceptions do exist [Magoon, 1994], few submarine forearc basins have been singled out as viable targets for conventional oil and gas exploration. That tepid reputation is largely because of their low levels of thermal maturity, high sedimentation rates, and low concentrations of organic matter. On the positive side, deep burial of source rocks can compensate for low heat flow [Peters et al., 1994], and stratigraphic traps are common in turbidite systems with channel-levee complexes and pinch-outs of sandy depositional lobes [e.g., Godo, 2006; Gardiner, 2006]. Thus, forearc basins are gaining more attention at the frontiers of oil and gas exploration [Struss et al., 2008; Lutz et al., 2011], and patterns of fluid migration and pore pressure become more significant when viewed within that broader context.

[4] There are several ways to evaluate bulk hydrogeological properties of forearc deposits. The consolidation state of mudstone provides a proxy for in situ effective stress, pore fluid pressure, and stress history [Saffer, 2003; Long et al., 2011]. Overpressured conditions (i.e., fluid pressure greater than hydrostatic) can arise from a wide variety of mechanisms, including compaction disequilibrium, thermal expansion of interstitial fluids, transfer of water from mineral dehydration reactions (opal to quartz, smectite to illite), and catagenesis of organic matter [Osborne and Swarbrick, 1997; Swarbrick et al., 2002]. Occurrences of underconsolidation in subduction zones are often attributed to supra-hydrostatic pore pressures and partly undrained conditions during burial [von Huene and Lee, 1983; Shephard and Bryant, 1983; Saffer, 2003]. Conversely, occurrences of true (or apparent) overconsolidation provide evidence for erosion and unroofing of overburden, stronger bonding or cementation of mineral grains, and/or tectonic deformation during which the orientation of maximum effective stress is non-vertical [Shephard and Bryant, 1983; Morgan and Ask, 2004].

[5] The Kumano Basin offshore southwest Japan (Figure 1) provides an excellent natural laboratory to study linkages among forearc tectonic, sedimentation, physical properties, and hydrogeology. The basin is one of several in the forearc of the Nankai Trough subduction zone. Terrigenous sediment supplies and presence or absence of slope gullies and submarine canyons control depositional styles across the Nankai forearc [Blum and Okamura, 1992]. Depositional activity, moreover, has changed profoundly during eustatic sea-level fluctuations [Blum and Okamura, 1992; Omura and Ikehara, 2010]. The Kumano Basin was created or accentuated by uplift along a prominent out-of-sequence thrust, termed the megasplay fault [Moore et al., 2007; 2009] (Figure 1). Strata within the basin dip to the northwest [Gulick et al., 2010], and a prominent bathymetric “notch” has been eroded along the basin's outer edge [Martin et al., 2010]. Hypothetically, if landward-dipping sand beds crop out within the notch, then efficient discharge through seafloor vents should favor hydrostatic pore pressures and normal consolidation of the mudstone interbeds. Conversely, if high-permeability conduits pinch out or terminate by on-lap along the migration routes, then localized overpressures and underconsolidation might develop in the mudstones near the pinch-outs.

Figure 1.

Map showing the Kumano transect region of Nankai Trough with drill sites of the Integrated Ocean Drilling Program (IODP). Seismic inline section from 3-D survey shows principal tectono-stratigraphic domains and locations of drill sites. Modified from Moore et al. [2009].

[6] To test these ideas, we conducted 1-D consolidation tests on whole-round core samples of mudstone from Site C0002 of the Integrated Ocean Drilling Program (IODP); the site is located ~4 km from the southeastern edge of the Kumano Basin (Figure 2). Our primary purpose was to reveal the sediment's past state of in situ stress from estimates of preconsolidation pressures (i.e., the pre-test maximum effective stress obtained from each consolidation response). We also report values of hydraulic conductivity, intrinsic permeability, and compressibility because of their importance for basin modeling [e.g., Aplin and Vasseur, 1998; Bjørlykke et al., 2010]. Our interpretations are placed in a context of lithostratigraphy and basin-scale deformation to build a more holistic reconstruction of the basin's evolution, including some implications for assessments of petroleum systems within subduction zones.

Figure 2.

Interpreted seismic-reflection profile crossing IODP Sites C0009 and C0002, showing seismic surfaces (S1, S2, A, B, C, D, E), unconformities (UC1 and UC2), and provisional correlation of lithostratigraphic units (I, II, III, and IV). BSR stands for bottom simulating reflector, and KBEFZ stands for Kumano Basin Edge Fault Zone. Thin dashed lines highlight locations of normal faults. Modified from Expedition 319 Scientists [2010].

2 Geologic Background and Geographic Setting

[7] The Nankai Trough Seismogenic Zone Experiment was designed to investigate several aspects of subduction dynamics and fault-zone behavior through a combination of ocean drilling and deployment of long-term borehole observatories [Tobin and Kinoshita, 2006]. IODP has drilled thirteen sites along the so-called Kumano transect offshore the Kii Peninsula of Honshu (Figures 1 and 2). These boreholes are grouped into three broad domains: sedimentary strata and upper igneous crust on the incoming plate (subduction inputs), fault zones and wallrocks near the toe of the accretionary prism, and the shallow megasplay fault system [Underwood and Moore, 2012].

[8] Kumano is the largest of several forearc basins in the Nankai Trough subduction zone; it measures about 100 km from east to west and 70 km from north to south (Figure 1). Total sediment thickness near the drill sites varies from ~1000 m to ~2500 m (Figure 2). The basin first formed as a structural depression during activity on the megasplay system at 1.67 to 1.56 Ma [Strasser et al., 2009]. A phase of landward tilting from 1.3 to 1.0 Ma was probably caused by slip on the megasplay, and an extensive system of normal faults post-dates that tilting [Gulick et al., 2010]. The Kumano Basin Edge Fault Zone (KBEFZ), along the seaward edge of the basin (Figure 2), is thought to accommodate strain partitioning due to the oblique convergence of the Philippine Sea plate, and faulting may be responsible for creating a 750-m-deep, 3-km-wide bathymetric “notch” that extends ~100 km along-strike [Martin et al., 2010].

[9] Site C0002 was logged during Expedition 314 to a depth of 1401 meters below seafloor (mbsf) using logging while drilling (LWD) [Expedition 314 Scientists, 2009]. It was then cored discontinuously to a depth of 1052 mbsf during Expedition 315 [Expedition 315 Scientists, 2009a]. Site C0009, which is located approximately 20 km farther landward (Figure 2), was sampled for cuttings during riser drilling operations and cored over a limited depth interval (1510 to 1591 mbsf) across the base of the forearc deposits [Expedition 319 Scientists, 2010]. Expedition 315 Scientists [2009a] divided the cores from Site C0002 into four lithologic units, all of which correlate with the logging units interpreted from LWD (Figure 3). Extension of age-depth relations from Site C0002 to C0009 was hampered by the poor precision of cuttings analysis, but the unit boundaries and some prominent depositional surfaces appear to be time-transgressive [Expedition 319 Scientists, 2010; Hayman et al., 2012]. The age control from C0002 cores is not precise enough to resolve eustatic cycles (Figure S1), but it stands to reason that sediment influx to the Kumano Basin accelerated during each of the Pleistocene glacial lowstands, particularly after the onset of major Northern Hemisphere glaciation at 1.4 Ma [e.g., Miller et al., 2005; Lisiecki and Raymo, 2007]. Those eustatic cycles affected oceanographic circulation (i.e., Kuroshio and Tsushima Currents) [Iwatani et al., 2012] and forearc stratigraphy [e.g., Pickering et al., 1999].

Figure 3.

Seismic-reflection image of the region near IODP Site C0002 with details of core-log-seismic integration, from Expedition 314 Scientists [2009] and Expedition 315 Scientists [2009a]. Dashed lines highlight unit boundaries and unconformities. White circles show approximate positions of samples tested in this study. BSR stands for bottom simulating reflector.

[10] Beginning at the top of the C0002 section (i.e., the seafloor), unit I (upper forearc-basin facies) consists of dark olive-gray to greenish gray hemipelagic mud, abundant sand and silt turbidites, and thin beds of light gray volcanic ash. Based on LWD logs, the unit thickness is 136 m. The siliciclastic sand beds that were recovered using a hydraulic piston coring system are typically <15 cm thick, with sharp bases and normal grain-size grading. The thickest sand bed is ~1.86 m. The ages of these strata range from Holocene to ~1 Ma. Approximately 900,000 years ago (equivalent to ~100 mbsf), the rate of sedimentation at Site C0002 slowed from ~350 m/Myr to ~107 m/Myr (Figure S1). Overall, the stratigraphy of unit I thins and fines upward, and the facies character is typical of sheet-flow turbidites in a basin-plain environment [e.g., Pilkey et al., 1980; Ricci Lucchi and Valmori, 1980]. Turbidity currents reached the distal margin of Kumano Basin after flowing through submarine canyons and gullies that were eroded into the upper slope and shelf [Blum and Okamura, 1992; Omura and Ikehara, 2010].

[11] The first core from lithologic unit II (lower forearc-basin facies) was recovered at 150 mbsf, although the boundary between logging units was set at 136 mbsf. That boundary relates more to a decrease in porosity than it does to a difference in lithofacies. The dominant lithology recovered by coring is hemipelagic mud. The compacted mudstone is locally structureless but more commonly shows plane-parallel laminae and incipient fissility. Orientation of this fabric is horizontal to gently inclined. Secondary lithologies include thin turbidites of sand, sandy silt, and silt. Volcanic ash is rare. There is a substantial coring gap between 203 mbsf and the first rotary-core-barrel (RCB) core at 479 mbsf (Figure S1). Use of the RCB coring system resulted in very poor recovery of unconsolidated sand. The LWD logs, on the other hand, captured a much higher proportion of sandy layers (Figure S2). On a 10-m scale, the logs also reveal numerous depositional cycles (e.g., upward thinning and fining) and systematic changes in the frequency of occurrence of turbidites [Expedition 314 Scientists, 2009]. The unit's lower boundary at 830 mbsf coincides with the oldest silty turbidite, with a corresponding age of ~1.6 Ma (Figure S1). Rates of sedimentation within unit II ranged from ~560 m/Myr to ~1690 m/Myr. The abrupt acceleration of turbidite sedimentation evidently hinged on two factors: incision of through-going erosional conduits to funnel sediment supplies efficiently from the shoreline [Underwood and Moore, 2012] and sufficient uplift along the basin's seaward edge to create more accommodation space for the turbidites [Strasser et al., 2009; Gulick et al., 2010]. Collectively, the logs and cores from unit II define a coarsening and thickening-upward trend near the distal margins of a sheet-like turbidite system.

[12] The condensed section of lithologic unit III (starved basin facies) is composed mostly of Miocene to Pliocene mudstone. This unit extends from 830 to 922 mbsf with sedimentation rates ranging from 17 to 60 m/Myr (Figure S1). Calcareous nannofossils are unusually abundant in the mudstone, and bulk powder X-ray diffraction reveals an average calcite content of 16 wt.% [Expedition 315 Scientists, 2009a]; those values are significantly higher than the concentrations within unit II (above) and unit IV (below). The abrupt increase in carbonate content across the unit's lower boundary is consistent with rapid uplift of the seafloor from a position below the carbonate compensation depth (CCD) (unit IV) to a position above the CCD (unit III). The same trend in carbonate content is evident, although more gradual, in the slope-basin deposits at Ocean Drilling Program (ODP) Sites 1175 and 1176 in the central Nankai Trough [Underwood et al., 2003a]. Seismic reflection profiles show angular discordance and complicated geometries (Figure 3), and a possible unconformity occurs within the unit at ~865 mbsf (Figure S1). Bedding steepens relative to the nearly flat dips in unit II [Expedition 314 Scientists, 2009; Expedition 315 Scientists, 2009a]. Unit III is probably the product of incipient forearc and/or lower trench-slope deposition [Expedition 315 Scientists, 2009a; Underwood and Moore, 2012].

[13] Unit IV (accreted turbidite facies) contains deformed interbeds of indurated mudstone, siltstone, and sandstone with steep bedding dips and low carbonate contents. There is a pronounced angular unconformity at a coring depth of ~922 mbsf that marks the boundary between units III and IV [Expedition 315 Scientists, 2009a]. The associated hiatus lasted from ~5.0 Ma to ~3.8 Ma, and the oldest nannofossil datum for unit IV is 5.9 Ma (Figure S1). Acoustic character is chaotic, with discontinuous reflectors and erratic dips (Figure 3). The unconformity at the top of unit IV is thought to record the initial uplift and erosion of the previously deformed accretionary prism along a system of out-of-sequence (splay) faults [Strasser et al., 2009].

3 Materials and Methods

3.1 Sampling Protocol

[14] We obtained ten whole-round (WR) core samples from Site C0002 at depths of 553.8 to 920 mbsf, including five from the lower forearc-basin facies and five from the starved-basin facies (Figure 3). One sample was tested twice. We did not sample WRs from unit I because of time restrictions during the final days of Expedition 315. The WR samples were cut on the drillship within several hours of recovery, capped and taped, sealed with wet sponges in aluminum vacuum bags, and maintained at temperature of ~4 °C during shipment and storage. Immediately prior to lab testing, the samples were extruded from the core liners, and a trimming jig was used to shape cylindrical specimens for the consolidation tests.

3.2 Index Properties

[15] While trimming each WR, we retained two or three pieces of mudstone for water content measurements. Measurements of water content and calculations of void ratio and porosity followed ASTM [2006] and Blum [1997], with a correction for salt in the pore water. In some instances, the shore-based value of void ratio (e) is higher than the closest shipboard value [Guo et al., 2011]. Such discrepancies can be attributed to sample disturbance during shipment (e.g., creation of microcracks), the effects of trimming, or subtle differences in composition and texture between the nearby sampling intervals.

3.3 Compositional Analyses

[16] The procedures for shipboard analysis of bulk powders by X-ray diffraction (XRD) were described by Expedition 315 Scientists [2009b], with relative abundances of total clay minerals + quartz + feldspar + calcite equal to 100%. Shore-based analyses of the clay-size fraction (<2 µm equivalent spherical diameter) followed the methods of Underwood et al. [2003b], where the relative abundance of smectite + illite + chlorite + kaolinite + quartz is equal to 100% [Guo and Underwood, 2011a]. The kaolinite/chlorite ratio was determined from the overlapping peaks between 24° 2θ and 26° 2θ [Guo and Underwood, 2011b].

3.4 Consolidation Tests

[17] Constant-rate-of-strain (CRS) consolidation tests utilized nearly identical oedometer systems at the University of Missouri (MU) and the Pennsylvania State University (PSU). The equipment and testing protocols were fully described by Guo et al. [2011] and Saffer et al. [2011]. The diameter of the confining ring at MU is 41.4 mm; the maximum axial load is 44 kN, and the maximum total axial stress is 33 MPa. The maximum axial load at PSU is 50 kN; using ring diameters of either 36.6 mm or 50 mm, the corresponding maximum vertical total stress is either 50 MPa or 20 MPa. Specimens were backpressured for 24 h to ~200–400 kPa (MU) or 300 kPa (PSU) using deaired synthetic seawater (1.75 g NaCl in 500 mL distilled water). The vertical loads were applied with the sample base undrained, the sample top open to the backpressure, and at a specified rate of axial displacement [ASTM, 2006]. Displacement rates correspond to strain rates of 0.23 to 0.5% h−1. Tests were extended to peak axial stresses of 20 MPa (MU) and to either 20 MPa or 40 MPa (PSU) with continuous monitoring of sample height (mm), applied vertical total stress (kPa), and basal pore pressure (kPa).

[18] Figure S3 shows a representative example of CRS test results. The compression index (Cc) refers to the slope of the virgin portion of the compression curve on a plot of e versus log (σv), where σv is the vertical effective stress. We estimate Pc, the maximum past effective stress (or preconsolidation pressure), using two techniques: the strain energy density (SED) method [Becker et al., 1987] and the Casagrande [1936] method. Comparisons were made between each test-derived value of Pc and a calculated value of in situ hydrostatic vertical effective stress (σvh) at the equivalent burial depth, assuming monotonic and uniaxial loading [Holtz and Kovacs, 1981]. The in situ vertical effective stress was calculated by subtracting the hydrostatic pore pressure from the total overburden pressure, as constrained by the shipboard bulk-density profiles [Expedition 315 Scientists, 2009a].

[19] After each test, the overconsolidation ratio (OCR) was determined using the ratio of Pc to σvh. An OCR value of unity (Pc = σvh) indicates normal consolidation. Values of OCR < 1 indicate underconsolidation, whereas OCR > 1 indicates overconsolidation. Sample disturbance was assessed qualitatively by comparing the difference between the initial (pre-test) void ratio and the void ratio at the point of Pce) to the initial void ratio (ei) [Lunne et al., 1997]. We rated most of the samples as poor to very poor quality using this criterion [Guo et al., 2011] (Table S2).

[20] CRS testing also provides a continuous record of hydraulic conductivity (K) along the specimen axis as a function of consolidation state. We used the values of applied strain rate (dε/dt), specimen height (H), initial specimen height (Ho), unit weight of pore water (γw), excess pore pressure at the undrained sample base (∆u), and the following equation [ASTM, 2006; Long et al., 2008]:

display math

[21] To calculate intrinsic permeability (k) from hydraulic conductivity under laboratory testing conditions, we assumed a fluid viscosity (ν) of 0.001 Pa·s for seawater at 20 °C, and a fluid density (ρ) of 1027 kg/m3:

display math

[22] Test results generally define a relation between intrinsic permeability and porosity (n) of the form: log k = log ko + a ∙ n, where a is a constant and ko is the initial value of permeability [e.g., Neuzil, 1994]. These relations can be used to estimate the in situ hydraulic conductivity for each specimen, by extrapolating the porosity-log permeability relation to the in situ porosity at a given depth [Long et al., 2008]. We assume that the in situ void ratio (en) for each specimen is equal to the value of e at Pc on the experimental compression curve (Figure S3). To compute hydraulic conductivity from intrinsic permeability, we account for the effects of temperature and salinity on viscosity and density, assuming a pore-water salinity of 35‰ [Fofonoff, 1985; El-Dessouky and Ettouny, 2002], and by extrapolating the shallow thermal gradient at Site C0002 [Expedition 315 Scientists, 2009a] to the depth of each sample.

4 Results

4.1 Sediment Composition

[23] Bulk-powder XRD provides relative abundances (wt.%) of total clay minerals, quartz, feldspar, and calcite [Expedition 315 Scientists, 2009a]. The mineral assemblage at Site C0002 is dominated by total clay minerals, followed by quartz, feldspar, and calcite. Concentrations of calcite increase significantly within the starved basin facies (unit III) relative to the overlying forearc basin deposits (Figure S4). Relative percentages of calcite within unit II range from 0.1 to 10.8 wt.% and average 2.1 wt.%, whereas carbonate in unit III ranges from 0.1 to 28.2 wt.% and averages 16.5 wt.%. The specimens that we tested (Table S1) are typical of the bulk mudstone compositions within each lithologic unit.

[24] XRD analyses of the clay-size fraction show that illite is the most abundant clay mineral within units I and II, followed closely by expandable clays of the smectite group and chlorite [Guo and Underwood, 2011a]. The clay mineral assemblage, however, changes as a function of stratigraphic position (Figure 7). Within unit II, the relative percentage of illite ranges from 31 to 40 wt.%, with an average of 35 wt.%; smectite ranges from 19 to 32 wt.% and averages 25 wt.%. Percentages of illite within unit III range from 30 to 40 wt.% and average 35 wt.%, whereas smectite increases to a range of 30 to 49 wt.% and an average of 36 wt.%. These results are consistent with what has been documented for coeval (Pliocene to Quaternary) deposits throughout the Nankai subduction system [Underwood and Pickering, 1996; Underwood and Fergusson, 2005; Guo and Underwood, 2011a]. Miocene mudstones within unit IV, in contrast, tend to have higher contents of smectite, typically greater than 40% of the clay-size fraction.

4.2 Consolidation Tests

4.2.1 Compression Index

[25] Figure S5 shows all the virgin compression curves obtained from CRS consolidation tests (see Guo et al. [2011] for complete documentation of each test). On each virgin curve, the smallest value of effective normal stress coincides with the estimate for Pc using the Casagrande method. Values of Cc range from 0.39 to 0.78 and average 0.55 (Table 2). Steeper slopes (higher Cc values) mean that the mudstones are more compressible. As depth increases in unit II, we notice a systematic decrease of Cc (Figure 4), but the average Cc value for unit III (0.605) is significantly higher than a comparable value for unit II (0.491) (Table S2). We ran linear regressions to see whether basic lithologic variables (e.g., void ratio, total clay minerals, and total organic carbon) have a meaningful influence on Cc (Figure 4). The correlation between Cc and void ratio (r = 0.59) is not statistically significant at a confidence level of 95%; neither is the correlation between Cc and burial depth, wt.% smectite, and wt.% calcite. The correlation versus total clay minerals (r = 0.77) is statistically significant at a confidence level of 99.5%, and the correlation coefficient for total organic carbon (r = 0.69) is statistically significant at a confidence level of 97.5%.

Figure 4.

Profile of compression index (Cc) versus depth at IODP Site C0002. Mudstone compressibility decreases with depth within unit II (blue dots, dashed arrow), but then increases within unit III (red dots). Linear regressions on the right show Cc vs. void ratio, total clay minerals (from bulk powder X-ray diffraction), and total organic matter, with values of the correlation coefficient (r).

4.2.2 Preconsolidation Pressure

[26] Values of Pc for units II and III range from 6.66 MPa to 14.60 MPa using the Casagrande method. The values are slightly higher (6.81 MPa to 15.88 MPa) using the SED method (Figure 5). In all cases, the preconsolidation pressure is greater than the in situ vertical effective stress (i.e., all of the specimens that we tested are overconsolidated). The differences between Pc and σvh range from 2.78 MPa to 9.14 MPa, and the corresponding OCR values range from 1.69 to 2.60 (Figure 5). Except for the upper two sample depths, which yielded significantly lower OCR values, there is no systematic trend of OCR with increasing depth, and we see no obvious differences in OCR between the deeper parts of unit II and unit III (Figure 5).

Figure 5.

Plots of in situ permeability, overconsolidation ratio, pre-test maximum effective normal stress, and porosity as a function of depth at IODP Site C0002. The bounds of each permeability value are from CRS permeability-porosity curves at the corresponding point for each Pc (Figure S6). Dashed green reference line depicts normally consolidated mud (OCR = 1). Dashed blue reference line depicts in situ effective vertical stress (σvh), which was derived by subtracting hydrostatic pore pressure from overburden pressure. Overburden pressure was constrained by integrating shipboard values of bulk density and depth [Expedition 315 Scientists, 2009a]. Porosity values from Expedition 315 Scientists [2009a].

4.2.3 Permeability

[27] Intrinsic permeability under laboratory testing conditions shows a log-linear decrease as vertical effective stress increases and porosity decreases (Figure S6). Under initial test conditions (σv = 1–2 MPa), permeability ranges from 6.5 × 10−17 to 9.8 × 10−16 m2 and porosities range from 52% to 54%. At loads of σv = 40 MPa, the values range from 1.5 × 10−19 to 3.7 × 10−19 m2 and the porosities range from 16% to 22%. We calculated the values of in situ permeability and hydraulic conductivity at each sampling depth using the Pc from each CRS test and permeant properties at the projected burial temperatures. Figure 5 reveals no obvious differences between unit II and unit III and no systematic trend with depth. The permeability values range from a maximum of 2.6 × 10–17 m2 (equivalent to an in situ hydraulic conductivity of 3.7 × 10–10 m/s) to a minimum of 2.5 × 10–18 m2 (equivalent to an in situ hydraulic conductivity of 3.1 × 10–11 m/s).

5 Discussion

5.1 Consolidation and Compressibility

[28] All of the specimens that we tested from Kumano Basin show evidence of apparent overconsolidation, with OCR values significantly greater than 1. Several factors need to be considered to explain this behavior. Erosion of overburden is the most common cause of true overconsolidation in soils [Skempton, 1970]. Overburden in marine environments can be stripped away by submarine slides and other types of mass wasting, as demonstrated by high OCR values in the upper part of IODP Site C0006 (Figure 1) [Dugan and Daigle, 2011]. At Site C0002, however, the unconformity within unit III (Figure 3) shows no geometric evidence of deep erosion, and with a sedimentation rate of only ~60 m/Myr, the Pliocene hiatus could not have altered the vertical effective stress by more than a few hundred kPa. Rapid burial beneath 830 m of Quaternary turbidites overprinted the earlier relict, and there are no unconformities within unit II. Thus, unloading was not responsible for the apparent overconsolidation.

[29] We expect sediment compaction in sedimentary basins to be uniaxial and vertical, but values of Pc will exceed in situ vertical effective stress if the stress path deviates from basinal condition [e.g., Shephard and Bryant, 1983]. On the basis of borehole-breakout data at Site C0002 [Chang et al., 2010] and recently active normal faults observed in seismic data [Gulick et al., 2010], the stress regime within Kumano Basin is extensional. The underlying accretionary prism, in contrast, is likely in a thrust-faulting regime with horizontal orientation of maximum principal stress [Chang et al., 2010]. Those deviations from the simulated stress path of simple basinal burial is one way to account for the OCR we observe in the experimental results.

[30] Collapse of microfabric during burial diagenesis is another cause of overconsolidation [Nygård et al., 2004; Aplin et al., 2006]. The apparent overconsolidation of mudstones from other parts of the Nankai subduction zone has been attributed to cementation by authigenic clay minerals [Morgan and Karig, 1995; Ujiie et al., 2003; Morgan and Ask, 2004]. We have no direct visual or geochemical evidence of such cementation at Site C0002, although based on the shallow geothermal gradient of 43 °C/km [Expedition 315 Scientists, 2009a], temperatures at the burial depths of our samples should be within the window of several relevant reactions. For example, the reaction temperature for conversion of opal-A to opal-CT ranges from 18° to 56 ° C [Pisciotto, 1981; Keller and Isaacs, 1985], and volcanic glass alters to smectite at temperatures as low as 20 °C [Hodder et al., 1993; Naish et al., 1993; Masuda et al., 1996]. In the nearby Shikoku Basin, traces of amorphous silica appear to have formed from alteration of dispersed volcanic glass [White et al., 2011], and small amounts of that cement may be responsible for inhibiting consolidation of the hemipelagic mud [Spinelli et al., 2007]. Porosity values in unit III are anomalously high and remain nearly unchanged with depth [Expedition 315 Scientists, 2009a], in a pattern reminiscent of the upper Shikoku Basin facies. We suggest a similar origin. On the other hand, higher rates of siliciclastic sedimentation within unit II diluted the concentrations of dispersed volcanic glass. Minute amounts of disseminated volcanic glass are difficult to quantify [Scudder et al., 2009], as are their alteration products. Thus, cementation still needs to be verified as a factor contributing to the apparent overconsolidation in the Kumano Basin.

[31] Fine-grained marine sediments yield a wide range of compression responses [e.g., Bayer and Wetzel, 1989; Aplin et al., 1995; Nygård et al., 2004]. Previous tests on mudstones from the Nankai Trough and Shikoku Basin (ODP Sites 1173 and 1177) yielded Cc values of 0.10 to 0.48 [Bellew, 2004]. In comparison, most samples from the Kumano Basin are more compressible (Cc = 0.39 to 0.78). One reason for the difference is the initial void ratio; compressibility will increase if the number and dimensions of particle contacts are smaller [Long et al., 2011]. We see a weak response of this type in the Kumano Basin, but the correlation is not statistically significant (Figure 4). As burial depth increases, the primary microfabric should collapse from a more random, flocculated or bioturbated arrangement of clay particles to an incipient, bedding-parallel fissility with lower porosity [Moon and Hurst, 1984]. Statistical analysis of particle alignment, as imaged by environmental scanning electron microscope, confirms this expectation [Guo et al., 2011]. Compressibility, therefore, should be lower in the deeper intervals of unit II, and lower still within unit III. The first part of that prediction holds, but the second does not (Figure 9). Another possible factor is the concentration of terrigenous organic material; macropores and micropores in woody tissue are easily crushed during loading [Wong et al., 2009]. This effect is also obvious in the Kumano Basin, although to a lesser extent than the influence of void ratio (Figure 9). Detrital clay-size particles are highly compressible [Aplin et al., 1995], and expandable clay minerals (smectite) are widely known to be more compressible than illite, chlorite, or kaolinite [Mitchell, 1993]. The clay-mineral effect (Figure S4) appears to supersede the expected effect of decreasing void ratio within unit III (Figure 4). Accordingly, in spite of deeper burial, mudstones in unit III exhibit higher compressibilities largely because of their differences in detrital clay mineralogy.

5.2 Permeability

[32] Siliciclastic mudstones and siltstones exhibit a very wide range of permeabilities, from 10−23 to 10−14 m2 [e.g., Neuzil, 1994]; that variability reflects differences in porosity, grain-size distribution, clay content, and pore-size distribution [Bennett et al., 1989; Dewhurst et al., 1998, 1999; Yang and Aplin, 2007]. The permeability values reported here are consistent with the results of Gamage et al. [2011] from the Nankai Trough (Figure 6), as well as the “global” permeability-porosity envelope established for mudstones by Neuzil [1994] (see also Dugan and Daigle [2011]; Rowe et al., [2011]; Ekinci et al., [2011]; Saffer et al., [2011]; and Yue et al. [2012]). The differences in permeability among our specimens are relatively small. The clay-size fraction does not change very much, relative to silt and sand [Kopf et al., 2011], and the correlation between permeability and total clay minerals (determined from bulk XRD) fails to meet the test for statistical significance. Similarly, linear regressions between intrinsic permeability and burial depth, void ratio, wt.% smectite, wt.% calcite, and TOC are not statistically significant. Evidently, the lithologic attributes of Kumano Basin mudstones do not change enough to create anything more than subtle differences in fluid transmission.

Figure 6.

CRS-derived permeability values for IODP Site C0002 (red dots) plotted as a function of porosity in Kumano Basin. Comparative results for other Nankai mudstones [Gamage et al., 2011] are superimposed on the global shale field of Neuzil [1994]. The regression equation (dashed black line) for Nankai mudstones [Gamage et al., 2011] has a correlation coefficient of 0.85.

5.3 Implications for Drainage and Overpressure

[33] Fluid overpressures build in sedimentary basins and subduction zones when excess pore pressure cannot dissipate rapidly enough to keep pace with sedimentation or burial [Bethke, 1986; Gordon and Flemings, 1998; Saffer and Tobin, 2011]. Overpressures can also be generated by diagenetic fluid production associated with clay-mineral dehydration and catagenesis of hydrocarbons [Osborne and Swarbrick, 1997; Swarbrick et al., 2002]. Underconsolidation (ORC < 1) is the most obvious sign of elevated fluid pressure, but we found no such examples in the Kumano Basin. To evaluate whether or not this outcome is consistent with known sedimentation rates and hydraulic properties, we used the Gibson [1958] analytical solution for sedimentation above an impermeable base to quantify expected overpressure in the basin.

[34] In this model, the magnitude of overpressure is governed by a dimensionless time factor:

display math

where m is the average sedimentation rate, t is the total time, and cv is hydraulic diffusivity. The average sedimentation rate for units I and II is ~0.5 mm/yr, and the total time span is 1.6 Myr [Expedition 315 Scientists, 2009a]. Hydraulic diffusivity for the mudstones ranges from 3.91 × 10−8 to 9.76 × 10−7 m2/s [Guo et al., 2011]. The overpressure ratio is equal to normalized overpressure (Δu) divided by the hydrostatic effective vertical stress at the base of the stratigraphic column. Over a realistic range of Tg values, we predict overpressure ratios of 0.01 to 0.13 at the base of the forearc basin (Figure 7). Direct in situ measurements at Site C0009 showed that the stratigraphic column is hydrostatically pressured [Expedition 319 Scientists, 2010]. Using a Tg value of 0.04 (based on the average of laboratory-defined hydraulic diffusivities), the predicted overpressure ratio is merely ~0.02 at the base of Kumano Basin.

Figure 7.

Gibson [1958] one-dimensional solutions (upper plot) for overpressure at Site C0002, using constant sedimentation rate, no flow at the base, and different time factors (Tg = m2t/cv), where m is the sedimentation rate (mm/yr), t is total time (yr), and cv is hydraulic diffusivity (m2/s). Non-dimensional depth is equal to mbsf divided by depth at the base of the hole (h). Overpressure (Δu) is normalized to hydrostatic vertical effective stress at the base of the hole (σvh (h). Lower plot compares non-dimensional overpressure vs. non-dimensional depth using bulk hydraulic diffusivity for various composites of interbedded mudstone and turbidite sand.

[35] Several factors might drain Kumano Basin more efficiently and keep fluid pressures close to hydrostatic. Vertical drainage probably occurs through the pervasive normal faults that cut through units I and II (Figure 3). Fissility in shale usually leads to significantly higher anisotropies of permeability [Bolton et al., 2000; Clavaud et al., 2008], but the effect of incipient fissility in unit II mudstones [Expedition 315 Scientists, 2009a] is probably small. On the other hand, sandy turbidites and mudstone vary in terms of hydraulic diffusivity, so we estimated that effect using the formula of Bodvarsson [1970] and Talwani and Acree [1984]:

display math

where k = permeability, ν = dynamic viscosity, n = porosity, βf and βs are compressibilities of fluid and sediment, respectively. We assumed viscosity of seawater = 10−3 Pa∙s, porosity of sand = 0.44 [Expedition 315 Scientists, 2009a], effective compressibility of seawater = 4.6 × 10−10 Pa−1, sand permeability = 10−15 m2 [Saffer and Bekins, 2006], and compressibility of sand = 1.5 × 10−8 Pa−1 [Ostermeier, 2001]. The resulting value of hydraulic diffusivity for the sand is 1.7 × 10−4 m2/s. Iterations of the Gibson [1958] simulation for sand/mud composites show that >90% of the strata must be sand to sustain hydrostatic fluid pressures solely by vertical drainage (Figure 7). That proportion is more than the percent-sand deduced from LWD data (Figure S2).

[36] Overpressured fluids in the turbidites should drain laterally through the landward-dipping beds until trapped or discharged [Bredehoeft et al., 1988; Yardley and Swarbrick, 2000; Flemings et al., 2002]. The turbidites in the Kumano Basin appear to crop out within the bathymetric “notch” along the basin's outer ridge [Martin et al., 2010]. Rapid fluid escape through those sand layers (Figure 8) should keep fluid pressure in adjacent mudstone beds close to hydrostatic. The suppression of overpressure allows chemical diagenesis and fabric collapse to create the apparent overconsolidation. Without seepage through the notch, excess fluid pressures might build where the turbidite beds pinch out (i.e., at lobe or channel/levee margins) or where such beds terminate against the on-lap surfaces that formed during tilting of the basin (Figure 8). The best places to look for underconsolidation would be along the shallow on-lap surfaces, such as the S1 surface (Figure 2). For those seals to be effective, however, the on-laps must juxtapose sandy intervals against intervals of low-permeability mudstone. Thick mudstone intervals are non-existent within the upper parts of Kumano Basin, so overpressures seem unlikely.

Figure 8.

Conceptual model of subsurface hydrogeology near Site C0002, Kumano Basin. Blue arrows represent hypothetical fluid-flow directions. Lateral drainage is focused along landward-dipping turbidite sands. Venting occurs within the bathymetric “notch,” and vertical flow is enhanced along normal faults. The most likely places for fluid overpressures to build are where sandy turbidite intervals on-lap thick intervals of low permeability mudstone (yellow).

[37] The test results from Site C0002 provide important insights into behavior of the Kumano petroleum system. Turbidites were deposited in the forearc at high sedimentation rates (~830 m in 1.6 Myr), comparable to the rates recorded in other forearc basins [e.g., Paquet et al., 2011]. The prevalence of woody organic matter [Expedition 319 Scientists, 2010] shows that the basin is gas-prone. The shallow geothermal gradient is relatively low at Site C0002 (43 °C/km). Concentrations of dissolved methane, ethane, and propane [Expedition 315 Scientists, 2009a] seem to require some up-dip migration of thermogenic gases from sources deeper in the basin, with free gas trapped below a widespread bottom simulating reflector (BSR) [Doan et al., 2011]. In addition, the style of faulting and landward tilting in Kumano Basin is typical near the seaward edge of a forearc basin [e.g., Susilohadi et al., 2005; Contardo et al., 2008]. In the Hawke Bay forearc of New Zealand, for example, an impressive anticline at the basin's seaward edge incorporates the forearc deposits, and tectonic deformation within the interior of the basin has created numerous smaller-scale structural closures [Paquet et al., 2011]. Those types of structures might be good targets for petroleum exploration. What sets Kumano Basin apart as an exploration target is the deep bathymetric “notch” along the KBEFZ [Martin et al., 2010], which breaches the prospective trap and seal. The prospects for accumulation and preservation of hydrocarbons would be more attractive without that zone of fluid venting.

6 Conclusions

[38] Uplift of the Nankai accretionary prism and landward tilting prompted Kumano Basin's birth, but sedimentation thereafter was modulated by three co-mingled factors: episodic expansion and landward migration of accommodation space during uplift along the megasplay; Pleistocene uplift and erosion of sediment sources on land; and incision of a robust network of gullies and canyons on the upper slope, which must have intensified during the eustatic lowstands associated with Northern Hemisphere glaciation. Thus, as with any such example, the sediments of Kumano forearc basin record a history of interwoven tectonic, climatic, eustatic influences.

[39] CRS consolidation tests show that mudstones from the Kumano Basin have compression indices that range from 0.39 to 0.78 (average = 0.55). Cc values in unit III are higher than Cc values in unit II, and we attribute those differences to higher contents of total clay minerals. All of the mudstones that we tested are overconsolidated (OCR > 1). Values of in situ permeability range from 2.6 × 10–17 to 2.5 × 10–18 m2 and show no systematic changes with burial depth or lithologic attributes. We suggest that efficient lateral drainage through landward-dipping turbidite sand beds keeps pore fluid pressures within the mudstone interbeds at (or close to) hydrostatic values. In addition, one or more of the following may have facilitated the apparent overconsolidation in lower parts of the basin: authigenic clay (or silica) cementation, realignment of microfabric, and tectonic consolidation. Cementation is more likely within unit III, where porosity remains anomalously high and nearly unchanged with depth.

[40] Judging from our lab tests and numerical models, fluids (hydrocarbon gases and interstitial water) probably discharge where sand beds crop out along the “notch” at the seaward edge of the basin. The basin's distal edge, therefore, is not very promising for hydrocarbon traps and seals. Farther landward, on the other hand, we see several on-lap surfaces that formed during landward migration of the basin's depocenter. Locally, those on-laps might seal the fluids transmitted along inclined turbidite beds (Figure 8), but for those seals to be effective, the on-laps must juxtapose sandy intervals against thick intervals of low-permeability mudstone. In the final analysis, although some elements of the petroleum system might be favorable, the overall prospects for Kumano Basin are not very promising.

Acknowledgments

[41] This research used samples and data provided by the Integrated Ocean Drilling Program. We thank the captains, drilling crew, laboratory technicians, and fellow shipboard scientists for their assistance in sampling during IODP Expedition 315. We also thank Pierre Henry and two anonymous reviewers for providing constructive suggestions that improved the paper. Financial support was provided by grants from the National Science Foundation (OCE-0751819 and OCE-0752114) and the Consortium for Ocean Leadership (T315A58, T315B58, and T315C58).

Ancillary