Velocity-porosity relationships for slope apron and accreted sediments in the Nankai Trough Seismogenic Zone Experiment, Integrated Ocean Drilling Program Expedition 315 Site C0001



In this study, we focused on the porosity and compressional wave velocity of marine sediments to examine the physical properties of the slope apron and the accreted sediments. This approach allows us to identify characteristic variations between sediments being deposited onto the active prism and those deposited on the oceanic plate and then carried into the prism during subduction. For this purpose we conducted ultrasonic compressional wave velocity measurements on the obtained core samples with pore pressure control. Site C0001 in the Nankai Trough Seismogenic Zone Experiment transect of the Integrated Ocean Drilling Program is located in the hanging wall of the midslope megasplay thrust fault in the Nankai subduction zone offshore of the Kii peninsula (SW Japan), penetrating an unconformity at ∼200 m depth between slope apron sediments and the underlying accreted sediments. We used samples from Site C0001. Compressional wave velocity from laboratory measurements ranges from ∼1.6 to ∼2.0 km/s at hydrostatic pore pressure conditions estimated from sample depth. The compressional wave velocity-porosity relationship for the slope apron sediments shows a slope almost parallel to the slope for global empirical relationships. In contrast, the velocity-porosity relationship for the accreted sediments shows a slightly steeper slope than that of the slope apron sediments at 0.55 of porosity. This higher slope in the velocity-porosity relationship is found to be characteristic of the accreted sediments. Textural analysis was also conducted to examine the relationship between microstructural texture and acoustic properties. Images from micro-X-ray CT indicated a homogeneous and well-sorted distribution of small pores both in shallow and in deeper sections. Other mechanisms such as lithology, clay fraction, and abnormal fluid pressure were found to be insufficient to explain the higher velocity for accreted sediments. The higher slope in velocity-porosity relationship for accreted sediments can be explained by weak cementation, critical porosity or differences in loading history.

1. Introduction

Sediments in subduction zones are typically unlithified, with more than 60% initial porosity [e.g., Bray and Karig, 1985]. As sediments are buried and tectonically incorporated into the prism, porosity decreases, causing an increase in the effective elastic moduli of the sediments. The physical properties of these sediments evolve in various ways in response to changes in the large-scale architecture, deformation, mass balance, and seismogenic behavior of subduction zones [e.g., Bangs and Westbook, 1991; von Huene and Scholl, 1991; Moore and Vrolijk, 1992; Erickson and Jarrard, 1998; Bilek and Lay, 1999; Moore and Saffer, 2001; Gettemy and Tobin, 2003; Saffer, 2007].

The characteristic evolution of physical properties in underthrust or accreted sediments is expected to differ from that of general basin sediments because of a more rapid increase in overburden pressure and compressive stress than what would be caused by gradual burial in a purely depositional setting [Saffer and Bekins, 2002; Tsuji et al., 2008]. The purpose of this study is to compare changes in physical properties between slope apron and accreted sediments as a function of depth, and then to explain how characteristic differences between these materials shed light on the growth and development of the accretionary wedge.

Three Integrated Ocean Drilling Program (IODP) expeditions (314, 315, and 316) were conducted at the Nankai Trough off Kumano, Kii Peninsula, Japan, in 2007 to early 2008 comprising Stage 1 of the Nankai Trough Seismogenic Zone Experiment (NanTroSEIZE). Site C0001 is located in the hanging wall of the megasplay fault, and penetrates the boundary between slope apron and accreted sediments (Figures 1 and 2). In Expedition 314, logging while drilling was performed at this site to measure physical properties, including compressional wave velocity [Kinoshita et al., 2008]. Core samples were later obtained at Site C0001 during Expedition 315; these were the first scientific core samples recovered by the drilling ship Chikyu as part of an IODP operation [Ashi et al., 2008]. Although core samples from accreted sediments were also acquired at Sites C0004 and C0002, the recovery of accreted sediments at C0001 proved to be the best out of the whole NanTroSEIZE operation thus far [Kinoshita et al., 2008; Ashi et al., 2008; Kimura et al., 2008]. The core samples provide a good opportunity to examine deposition and the evolution of the upper accretionary prism.

Figure 1.

The location of NanTroSEIZE Site C0001 (modified from Kinoshita et al. [2008]). (a) Map of Japan and bathymetry off Kii Peninsula. (b) The location of Site C0001 in a seismic profile of the Nankai accretionary prism off Kii Peninsula.

Figure 2.

(a) A detailed seismic profile of Site C0001. Coring depth is shown. (b) Stratigraphic summary of Site C0001. Depth is in meters below seafloor (mbsf).

We focused mainly on changes in porosity and compressional wave velocity. Compressional wave velocity is one of the primary expressions of elastic bulk and shear moduli, and porosity is a major control on velocity [Wyllie et al., 1956; Mavko et al., 1998]. Many studies have examined these properties in various sediments. The relationships between compressional wave velocity and porosity have also been reported [e.g., Raymer et al., 1980; Marion et al., 1992; Hyndman et al., 1993; Erickson and Jarrard, 1998; Gettemy and Tobin, 2003; Hoffman and Tobin, 2004; Spinelli et al., 2007]. Comparison of relationships between compressional velocity and porosity for various depositional histories presented in this study will identify characteristics of accreted sediments expressed by quantitative physical properties.

In this study, we present results of porosity-depth curve for Site C0001 and compressional velocities from laboratory tests, and show the differences in porosity-velocity relationship between slope apron sediments and accreted sediments.

Secondary Electron Microscopy (SEM) and micro-X-ray CT scan analyses were conducted to examine whether variability can be linked to differences in texture.

2. Geologic Overview

The Nankai Trough lies on the Pacific side of SW Japan (Figure 1). Seismic studies have revealed the general features of the associated accretionary prism [e.g., Park et al., 2002; Moore et al., 2007]. The “megasplay” fault is a characteristic structure which branches from the decollement at seismogenic depth and cuts through the prism to the surface. This fault could potentially be activated as a tsunamigenic fault during a major seismic event [Moore et al., 2007].

Site C0001 is located in the hanging wall of the megasplay fault [Kinoshita et al., 2008; Ashi et al., 2008] (Figures 1 and 2). Coring was conducted from the seafloor to 458.0 m below seafloor (mbsf) at Site C0001 during Expedition 315. Two lithologic units were identified (Figure 2) [Ashi et al., 2008]. Unit I extends from the seafloor to 207.17 mbsf and is Quaternary to late Pliocene in age (Figure 2b). The lithology of Unit I consists mainly of silty clay and clayey silt with minor thin interbeds and irregular patches of sand, sandy silt, silt, and volcanic ash. The siliciclastic interbeds show sharp bases, faint plane-parallel laminae, normal size grading, and diffuse tops. Such textures are typical of fine-grained turbidites. Fine sand layers interbedded with thin mud are observed in the bottom of Unit I from 196.76 to 207.17 mbsf (Figure 2b). The sand grains in these layers consist mostly of detrital quartz and feldspar with abundant sedimentary and low-grade metasedimentary rock fragments. Unit I is defined as the slope apron sediments, and is identifiable in the seismic profile by parallel reflectors of interbeds (Figure 2a).

The boundary between Units I and II is an unconformity, identified as a strong reflector at the bottom of the slope apron sediments in the seismic profile (Figure 2a). Unit II ranges from 207.17 mbsf to the bottom of the coring depth at 457.8 mbsf and is late Pliocene to late Miocene in age. The dominant lithology of Unit II is bioturbated mud. Biogenic materials are much more rare in Unit II than in Unit I. The abundance of interbeds of silt or volcanic ash is also lower in Unit II (Figure 2b). Unit II is interpreted as the upper accretionary prism on the basis of the seismic profile and age [Ashi et al., 2008].

The bedding of Unit I and Unit II dips gently about ∼0° to ∼20° throughout with local exceptions showing higher dip [Kinoshita et al., 2009].

Mineral composition and clay fraction within these formations have been examined by X-ray diffraction analysis [Kinoshita et al., 2009]. The main components of the sediments are quartz, plagioclase, calcite and clay minerals. The fractions of quartz and plagioclase remain constant at about 20 wt% in Unit I. The calcite and clay contents are inversely correlated, with calcite content decreasing from ∼30 to ∼0 wt% and clay content linearly increasing from ∼30 to ∼50 wt% in Unit I. In Unit II, the clay content abruptly increases to ∼65 wt% and plagioclase decreases to ∼10% at the boundary [Kinoshita et al., 2009]. The fractions are constant with depth in Unit II.

3. Porosity

Smectite is relatively abundant in the sediments in Unit II because total clay content increases abruptly from ∼45 wt% to ∼65 wt% at the boundary as described above. The percentage of smectite relative to other clay minerals remains almost constant at around 35% [Guo et al., 2009]. Therefore, the ratio of smectite content to bulk sediments increases from ∼10% to ∼20% in Unit I and is ∼25% in Unit II. The presence of large amounts of smectite can cause an overestimation of interstitial water content in sediments [Brown and Ransom, 1996]. This means that porosity derived from initial shipboard measurements may be an overestimation of in situ conditions. In this study, we correct these measurements using the method suggested by Brown and Ransom [1996] which can calibrate out the porosity overestimation resulting from smectite content. The amount of smectite was estimated by Guo et al. [2009] as described.

The porosity-depth curve for Site C0001 is shown in Figure 3. The corrected porosity is systematically smaller than onboard porosity. The differences between them are up to ∼15% in the deeper portion.

Figure 3.

Porosity versus depth for Sites C0001. Red dots and blue dots indicate porosities measured by onboard procedure and corrected by Brown and Ransom's [1996] method, respectively.

Porosity ranges from ∼0.8 to ∼0.45 at depth. A small recovery in porosity is observed at the unconformity (207.17 mbsf), even after the effect of smectite has been removed (Figure 3).

Because the differences in smectite content between Unit I and Unit II are relatively large, the porosity corrected by Brown and Ransom's [1996] method was applied for samples in the ultrasonic laboratory tests (Table 1) to calibrate out the effect on porosity measurements. Because porosity was measured on board using “community samples” located adjacent to the whole round samples used in this study, the porosity for samples in our ultrasonic tests is expected to match that of the closest community sample.

Table 1. Tested Samples, Porosity, and Compressional Wave Velocity
SampleDepth (mbsf)Uncorrected Porosity (%)Corrected Porosity (%)Effective Pressure (KPa)Velocity at Hydrostatic Pressure (km/s)

4. Ultrasonic Velocity Measurements

Compressional wave velocity was measured under conditions of controlled confining and pore fluid pressures. The design of the experiment is similar to that employed in other studies of saturated marine sediments [e.g., Tobin et al., 1994; Tobin and Moore, 1997; Gettemy and Tobin, 2003].

During each measurement, constant pore fluid pressure of 500 kPa was applied and confining pressure was adjusted to control the effective pressure. The maximum effective pressure for each experiment was based on a calculated in situ value derived from the measured bulk density of sediments integrated over the depth of recovery and the hydrostatic pore fluid pressure at that depth. 1 MHz Lead Zirconate Titanate (PZT) compressional wave transducers were used as a source-receiver pair. Uncertainty in velocity measurements is up to ∼2% depending mainly on differences between two displacement transducers and on error in arrival time picking. This is particularly evident at lower effective pressures.

Cylindrical sediment plugs were subsampled from each saturated whole round drill core (∼7–8 cm diameter by ∼15–20 cm length). Each plug is 3.8 cm diameter by ∼4–6 cm length. The whole round cores were sampled from intact sections identified by CT scan images on board to avoid the effects of drilling-induced damage or natural cracks. The velocity measurements were conducted parallel to the core axis. Four samples from Unit I and five samples from Unit II were tested (Table 1). All samples are silty clay to clayey silt, except for I-3 (114.3 mbsf), which contains a relatively larger amount of visible coarse grains than the others.

5. Results

The results of the compressional wave velocity measurements are presented in Figure 4 and in Table 1. The compressional wave velocity increases with effective pressure (Figure 4). The velocities for Unit I and Unit II range from ∼1.63 to ∼1.76 km/s, and from ∼1.74 to ∼1.99 km/s, respectively. The velocity for the samples just below the boundary (II-1 and II-2) is slightly slower than that above the boundary (Figure 4 and Table 1). The velocity–effective pressure relationship in general exhibits a greater positive slope angle for Unit I samples, but a lower slope throughout Unit II (Figure 4). The slope for I-3 stands out, with a much greater slope. This may be because this sample contains a relatively larger amount of coarse grains, as described above.

Figure 4.

Measured effective pressure versus compressional wave velocity using whole-round core samples. For clarity, I-3 and II-2 are shown as a gray line and a dashed line, respectively. The black circles and black squares represent the velocities for Unit I and Unit II, respectively, at calculated in situ effective stress assuming normal hydrostatic fluid pressure conditions.

The compressional wave velocity-porosity relationships are shown in Figure 5. The velocities at hydrostatic pressure are used in Figure 5 to represent an in situ value. The slope for Unit I in this relationship shows lower angle than that for Unit II. The higher angle of slope for Unit II is well observed in the relationship. For reference, the empirical relationships reported by previous work are presented in Figure 5. The curves labeled “normal compaction” and “high compaction” are those that Erickson and Jarrard [1998] proposed as a global empirical relationship between compressional wave velocity and porosity for clastic sediments. Two additional empirical relationships are shown in Figure 5 that were derived for the Shikoku Basin sediments oceanward of the deformation front off Shikoku (Ocean Drilling Program Site 1173) [Hoffman and Tobin, 2004] and from the sediments at the outermost toe of the Nankai accretionary prism (ODP site 808) [Hyndman et al., 1993] (Figure 5). Whereas the slope for Unit I in the velocity-porosity relationship is almost parallel to the empirical curves, the slope of Unit II is apparently much higher (Figure 5) than all of these foregoing relationships.

Figure 5.

Porosity versus compressional wave velocity. Black circles and black squares show the data from the laboratory measurements for Unit I and Unit II, respectively, at hydrostatic conditions. Estimated error in velocity is 2%. Least squares line fits for Unit I and Unit II are represented as dashed lines. Empirical relationships reported by previous works are shown. High and normal compaction curves are shown as gray lines. The empirical relationships observed at the toe of the accretionary prism [Hyndman et al., 1993] and in the Shikoku basin from Site 1173 [Hoffman and Tobin, 2004] are shown as black and broken lines, respectively.

6. Textural Characteristics of Sediments

Textural observations were conducted on the test samples. Each core was freeze-dried to preserve the textures present in what are basically unlithified sediments.

In the scanning electron microscope (SEM) observation of the broken surface, sharp-edged clay minerals are well preserved even in accreted sediments and also in slope apron sediments (Figure 6). Spinelli et al. [2007] suggest that cemented sediments will show textures with vague clay mineral boundaries. If the vague clays are evidence for cementation in unlithified sediments, the presence of sharp-edged clay minerals in accreted sediments identified in this study indicates that cementation had not occurred. Ujiie et al. [2003] suggested another type of cementation taking place in early stages of diagenesis. This would be caused by authigenic clay aggregates binding particles, and is found in sediments of equivalent depth for the decollement zone at Site 1173 and in decollement zone itself at Site 1174, ODP Leg 190. The clay aggregates clearly bind sediments particles to retain porosity. However, in the sediments in Site C0001 such clay aggregates were not identified, suggesting that this alternative type of lithification had not taken place either.

Figure 6.

Secondary electron microscope (SEM) images of (a) slope apron sediments (I-1, 66.57 mbsf, and porosity 59.8%) and (b) accreted sediments (II-5, 447.83 mbsf, and porosity 45.3%). The observed surface was a fracture opened by hand. Sharp edged clay minerals are well preserved in both sediments.

Micro-X-ray CT observation was conducted to observe pore geometry. The micro-X-ray CT system used for the observation was a TESCO Corporation, HMX225-Actis. The distance from the source of the X-ray beam to the sample was about 25 mm. The size of the pixels on the image is about 5 microns.

The pore textures for slope apron sediments and accreted sediments are presented in Figure 7. Pores are identified as black spots in each image. The smallest features observed on the images are 20 microns diameter, which may correspond to the spatial resolution of the scanner. Focusing on larger size of pore at least 20 μm, pores are rounded and subrounded for both sediments (Figures 7b and 7d). While pore size for accreted sediments is well sorted, representing a homogenous texture, that for slope apron sediments shows more heterogeneity. Larger pores (up to ∼150 μm in diameter) are rarely observed, and only in the shallower samples (Figure 7a).

Figure 7.

Micro-X-ray CT images for (a and b) slope apron sediments and (c and d) accreted sediments. Enlarged images show rounded or subrounded pore geometry (Figures 7b and 7d).

7. Discussion

The elevated dependence of velocity on porosity in the accretionary prism over that of the slope apron at Site C0001 may be a fundamental characteristic of the accreted sediments. Compressional wave velocity-porosity relationships have been reported in many previous studies of sediments from sedimentary basin and accretionary prisms [e.g., Hyndman et al., 1993; Erickson and Jarrard, 1998; Gettemy and Tobin, 2003; Hoffman and Tobin, 2004; Spinelli et al., 2007]. These studies suggested that the velocity-porosity relationship is affected mainly by lithology, shale (clay) fraction, cementation, pore shape (pore texture) and loading history.

The decrease in calcite and plagioclase and increase in clay fraction in Unit II should make the compressional wave velocity slower at constant porosity because the bulk modulus of plagioclase and calcite grains is over 3 times larger than that of the clay minerals [Mavko et al., 1998]; increasing clay content in clastic sediments is typically accompanied by velocity decrease [e.g., Han and Nur, 1986]. Therefore, the differences in mineral composition between Units I and II as described above cannot be used to explain the higher slope in the velocity-porosity relationship for Unit II.

An abrupt change in clay fraction at the boundary between Unit I and Unit II can cause a change in porosity without a change in effective pressure. Marion et al. [1992] examined the effect of clay fraction on porosity and velocity. According to this study, minimum porosity is reached at ∼20% clay content for lower effective pressures from ∼0–10 MPa and higher porosities are observed as clay fraction is increased. Although the effect of smectite on porosity was calibrated and removed, a small increase in porosity is observed at the boundary between Unit I and Unit II. The abrupt increase in porosity may be due to an abrupt increase in clay content. The small decrease in velocity at the boundary is reasonably explained by the increase in porosity.

Erickson and Jarrard [1998] suggested that a higher clay fraction could produce the variable angle of higher slope in the velocity-porosity relationship depending on clay fraction for porosities less than 0.4. This is corresponds to a critical porosity suggested by Nur et al. [1998]. In this study, the effect of clay fraction might be negligible because the porosity was always larger than 0.4. The idea of critical porosity, however, is consistent with the change in slope angle in velocity-porosity relationship as will be discussed later.

No cement was observed on the SEM images, but it is still possible that some form of weak cementation is occurring at the contacts between clay grains and is not imaged at the resolution of SEM images. Because cementation significantly increases velocity without a large porosity change [e.g., Dvorkin and Nur, 1996], the higher slope of velocity-porosity relationship could be explained by cementation if cementation increases with increasing depth and compaction.

With the wavelength of the 1 MHz ultrasonic wave being in the 1.5 to 2.0 mm range for the measured velocity, only the larger pores (∼400 to ∼500 microns as 1/4 of wavelength) observed in the shallowest sample in Unit I may influence raypaths or cause scattering. The main scale of heterogeneity observed at 10–50 microns will be averaged at the wavelength scale. The pore size distribution, however, is observed to differ between them. The pore size distribution in slope apron sediments is more heterogeneous than that of the accreted sediments. This may influence the porosity-velocity relationship because larger pores or heterogeneous porous domain can influence the tortuosity of raypaths for transmitted elastic energy more than smaller pores. Further study will require mercury intrusion porosimetry to identify the pore size distribution more precisely or textural analysis for heterogeneous distribution of pore geometry.

Loading history is an inconclusive factor in terms of the relationship between velocity and porosity. Loading history for slope apron sediments should differ from that for accreted sediments. While apparent sedimentation rates for slope apron sediments are reported to be ∼140 to ∼180 m/Ma, those for accreted sediments (below the unconformity) range 80–160 m/Ma [Kinoshita et al., 2009]. Their loading history may differ in the amount of horizontal stress and strain sustained during tectonic deformation. It is not clear, however, what the quantitative relationship would be, and this represents an open question.

The idea of critical porosity is consistent with the higher slope in velocity-porosity relationship. The idea is that a change from fluid-supported suspension to frame-supported sediments (or rocks) occurs at a specific “critical” porosity. Nur et al. [1998] have reported critical porosities for different kinds of materials. Most natural rocks and sintered glass beads show critical porosities less than 0.4. This is inconsistent with this study showing higher critical porosity around 0.55 (Figure 5). Porous materials such as pumice or chalks can exhibit a higher critical porosity [Nur et al., 1998]. Because cemented grains are corresponds to the porous grains, critical porosity tends to increase when cementation occurs at grain boundaries. Furthermore, critical porosity depends on pore size distribution and is higher in cleaner and better sorted sandstone [Avseth et al., 2001]. Samples in this study are silty clay rather than sand, but it is possible that critical porosity in these materials is also dependant on pore size distribution and texture.

This is the first report of the compressional wave velocity-porosity relationship for the accreted sediments in the hanging wall thrust sheet of the megasplay fault in the NanTroSEIZE transect. In this study, we observed a characteristic compressional wave velocity-porosity relationship in the accreted sediments which shows a substantially higher slope than that predicted by global empirical relationships and that previously observed around the toe of the Nankai accretionary prism except for the cemented section in Site 1173.

Results at C0001 suggest that accreted material is affected by diagenetic cementation, albeit weaker than that observed at Site 1173 on the Muroto transect [Spinelli et al., 2007]. Our observations found no major difference in texture between slope apron and accreted units. Several processes other than cementation or sorting of pore spaces may contribute to the observed contrast in porosity-velocity relationships, such as the critical porosity effect [Nur et al., 1998], or differences in stress path history.


We thank Brian Hess, John Fournelle, Lee Powell, Lee Putman, and Neal Lord for their help with sample treatments and laboratory measurements. We also appreciate O. Tadai for assistance in micro-X-ray CT analysis. This research used samples provided by the Integrated Ocean Drilling Program (IODP). This manuscript is much improved by constructive reviews from Associated Editor T. Tsuji and an anonymous reviewer. The work of Y. Hashimoto is supported in part by the Japan Society for the Promotion of Science (JSPS) Postdoctoral Fellowship for Research Abroad.