In this section we describe the brittle structures focusing on those with lengths that are greater than core diameter. Numerous smaller-scale structures are also visible, particularly when the core surface has dried, but detailed description of these is beyond the scope of the shipboard analyses. The brittle structures were observed in cores from lithologic Subunit IIB and below in Holes 1150A and 1150B. The highest number of structures were observed in RCB cores, although some structures were also observed in XCB cores. Probably, drilling disturbances in XCB cores overprinted many structures.
For the large-scale structures, we have assessed the abundance and type of structures that occur. For abundances, we have divided fractures into two types: open and healed. Most of the brittle structures were healed and filled with fine-grained dark gray clayey gauge. The width of the infill varies from 2 mm to a few tenths of a millimeter. No secondary mineralization or recrystallization (fibrous crystal growth) were observed. Open structures, though not as common as healed structures, were also abundant. These were probably drilling induced, although they were often formed along pre-existing weakness planes. These fractures were not considered in the structural interpretation unless they were partially filled.
Within the healed fracture category, we have further assessed the abundance of faults and joints. The distinction between faults and joints is based on the visible apparent displacement. The displacement was estimated primarily from the offset of burrows, carbonate nodule-like accumulations, and dissolution and precipitation rims around manganese or iron because sedimentary layering was rarely observed. As a result, the number of faults per meter is probably underestimated and the number of joints is overestimated.
In addition, we measured the apparent dip of the large-scale structures with a conventional contact goniometer. The apparent dip is less than or equal to the real dip of any plane; therefore, dip angles that are reported here are minimum values. For curved fault traces (e.g., listric faults), the predominant apparent dip was measured. True dip measurements of faults are presented in "Downhole Measurements". Finally, we reconstructed the orientation of healed and open fractures and bedding planes relative to geographical north using paleomagnetic directions.
The observed faults generally had the following appearance: (1) relatively straight with sharp fault traces, (2) relatively straight and sharp parallel fault sets with variable spacings (0.5-10 cm), (3) relatively straight and sharp conjugate fault sets, that often had the same (usually high) dip angle, but opposite dip directions, (4) anastomosing fault traces that generally occurred as close- and wide-spaced thin branches that diverged and rejoined over distances of a few millimeters to several centimeters (Fig. F84), (5) wavy or slightly curved faults (e.g., listric faults), (6) sets of two faults at a spacing of ~3-7 mm that were slightly curved toward each other at their termination, and (7) minor or secondary straight-branching splay faults that departed at acute angles from a major fault. The apparent dip angle varies from 10° to 75°, with the majority (>45%) being steep-dipping fault traces. Locally, single isolated faults that are mostly very sharp and thin cross the core at moderate angles.
The displacement along faults generally varies from 0 to 4.5 cm, although the offset most often is a few millimeters in length. On the other hand, an offset greater than the scale of the core was observed for some rare faults (see "Downhole Measurements").
Normal faults are most commonly observed (~67% of all faults) (Figs. F84, F85). In addition, oblique faults and a few generally low-angle reverse faults were found. The strike-slip component of the oblique faults appeared to be on average 25% of the related dip-slip component.
Uncertainties in the interpretation of the sense of displacement originate from observations of (1) core surfaces that show the trace of an oblique normal fault and (2) extremely steep reverse faults that are parallel or anastomosing. For oblique normal faults, which have a low dip angle, the strike-slip component of the oblique offset may appear as reverse-fault displacement at the fault's trace. For steep reverse faults, the movement may have partially consisted of small rotation of one fault block due to complicated interference of the fault and joint pattern and related space problems.
The observed joints generally had the following appearance: (1) small and thin joints, (2) small and thin parallel joint sets, (3) small and thin conjugate planes near major faults, (4) anastomosing joints, with some closely spaced and showing partly similar features and shapes as dewatering structures described by Knipe (1986) (Fig. F86), and (5) shear joints, which occur at acute to perpendicular angles to major faults (Fig. F85). A considerable number of these joints are steeply dipping (80°-90°). Often, the dip angle changes along the length of the joint.
Most joints appear to be extensional; however, the distinction between shear and extension joints is not always clear because of abundant hybrids between both end-members. As a result, both types are referred to simply as joints.
The different kinds of structures were counted in each core from Cores 186-1150B-1R through 50R, and then the total number was divided by recovered length in meters (Fig. F87; Table T25, also available in ASCII format). The number of healed fractures (joints and faults) increases with depth and has distinct peaks, whereas the number of open fractures (moderate to steep-dipping open fractures that were formerly healed fractures) changes little downhole (Fig. F87). A clear relationship can be seen between the joint and fault numbers per meter. The number of faults, though underestimated, is generally higher than the number of joints; otherwise, the two have similar downhole variation.
The calculated effective vertical stress increases linearly with depth, from ~5 MPa at 700 mbsf to 10 MPa at 1180 mbsf (Fig. F64). The fracture pattern is useful for constraining the timing and mechanisms of the deformation and the properties of the affected rocks.
For clayey sediments, the curve of porosity against burial depth usually shows a very characteristic flat shape at the beginning, representative of a large decrease in porosity, which means that the primarily high water-bearing sediments are very sensitive to pressure increase (Fig. F62). The mechanical behavior changes in conjunction with this compaction and with loss of water and porosity. When the brittle regime is reached, the observed changes with depth should include (1) a steepening of fracture planes, (2) a decrease in anastomosing and complexity of fracture branching, and (3) a decrease in number of faults with an increase in the amount of displacement on the rarer major faults.
As can be seen in Figure F87A, the number of healed fractures per meter increases on average with depth (basic minimum values increase), contrary to the above predictions. However, the change in porosity with depth is rather minor and probably cannot account for great changes in rock properties (see "Physical Properties"). Also, because of the observation of dewatering features, it is unlikely that the sediments are brittle over the entire length of the study interval. A more likely interpretation is that the increase of the number of healed fractures per meter results from overprinting by the two fault zones at 900-950 and 1030-1070 mbsf. Alternatively, the observed increase in clay content with depth may have permitted more healed fractures to form and thus account for compaction-related volume change (e.g., Guiraud and Seguret, 1987).
The porosity also has an important influence on the dip and the shape of the fracture traces. Fractures tend to become steeper and sharper with decreasing porosity. In clay-rich sediments, the process of fracturing starts with local rotation of clay particles to shear parallel orientation before a fracture plane evolves. In finer grained sediments, these zones are thinner than in coarser grained sediments. Also, continuing movement leads to an intensification of this fabric and consequently to a thinning of the fault gauge (Maltman, 1987). Furthermore, at smaller porosity, fractures tend to be sharper and fewer anastomosing faults occur. However, the observed fractures do not change shape or type with increasing depth. Again, multiple fracturing events, where more recent fractures overprint older ones, and downhole variations in lithology could be factors. The overprinting explanation is supported by FMS and resistivity data from Hole 1150B, which indicate that steeper faults tend to be active and flatter faults tend to be inactive (see "Downhole Measurements").
There are several assumptions that can be made concerning the regional stress field. The steep joints visible through the whole examined depth interval indicate a three-dimensional stress field with a dominant vertical effective principal stress and low confining horizontal effective principal stresses (at least one low value of stress in one horizontal direction). When confining pressure increases, hybrid shear-extension fractures form, becoming progressively flatter and evolving a shear component on the plane of failure. The abundant very small branching and anastomosing fissile microcracks and cracks, which could not be taken into account for statistics and are only mentioned in the VCDs, are thought to have formed in the first stages of brittle failure (Hobbs et al., 1976; Maltman, 1987; Knipe, 1986). With continued stress, they could unite to form through-going fault planes, if this is mechanically easier than creating new fractures. This evidence also accounts for the scoured and wavy shapes of a large part of the fractures.
Taking into account the obvious displacement and steepness of the joints and faults, it follows that a combined stress field is present. Some randomly distributed faults show unusually high amounts of displacement (several centimeters) not typical for simple compaction. Additionally, the majority of normal faults are moderate to steep dipping, which requires on the one side a low water content and on the other side a low confining pressure (especially for the subvertical joint fractures). Borehole logging results that show an elliptical diameter of the Holes 1150A and 1150B and the stereoplots of the FMS-detected faults indicate a combined stress field. The observed fracture pattern mirrors a superposition of a common sedimentary basin stress field and the Japan Trench-related forearc stress field. That means a predominant vertical maximum stress caused by an overload of the sedimentary column and an extension component with the lowest horizontal stress directed east-west.
We reconstructed orientations of healed and open fractures and bedding planes using characteristic magnetic directions obtained from cores in Hole 1150B. Dip azimuths for these surfaces were reoriented into geographic coordinates (i.e., relative to true north) using stable magnetic declinations. The method of restoring the azimuthal orientation of cores by paleomagnetic directions has been applied to identify an original structural attitude in many past DSDP and ODP studies. Fractures, which are commonly developed in the slope region of subduction zones, have been reoriented and utilized to indicate the stress field (e.g., Niitsuma, 1986). In Hole 1150B we recovered a fractured sequence in cores (see "Lithostratigraphy") that have fairly constant magnetic directions, making these suitable for reorientation (Fig. F88).
We focused our study on continuous pieces of core that were 20 cm long or longer. Pieces of these lengths are required to determine magnetic directions accurately and to ensure the direction is constant over the interval (Fig. F88). Planes of fractures were measured in three dimensions using all surfaces of the working half in the core-face coordinates system (Shipboard Scientific Party, 1995). In this coordinates system the plane of the split-core surface is called the apparent horizontal plane. A 360° orientation net is used with pseudo-north (0°) in the upcore direction. Dip azimuths (perpendicular to the strike) and dip angles were then rotated into the ODP core reference scheme by a 90° counterclockwise rotation about a horizontal east-west axis. We sketched positions of fractures to record which piece was measured. When the piece had a stable magnetization, it was reoriented into geographic coordinates using the paleomagnetic declination. The example in Figure F89 demonstrates the effectiveness of paleomagnetic reconstruction, which is evident from the improvement in clustering of healed fracture planes once they are placed in geographic coordinates. In Hole 1150B we did not correct the azimuths for the deviation of the hole from vertical because the hole deviations were negligible (<3.5°).
We made measurements of 369 healed fractures, 290 of which could be reoriented (Table T26, also available in ASCII format). Because the pattern of the horizontal frequency of measurements is similar to that of the occurrence of healed fractures (Fig. F90A) (see "Lithostratigraphy"), the results should be representative of healed fractures in this hole.
Histograms of all dip azimuths and dip angles are shown in Figures F90B and F90C. The two clusters of reoriented dip azimuths are nearly antiparallel and clearly bimodal with clusters near 90° and 270°. On the other hand, the distribution of dip angles shows an unimodal pattern with most angles ranging between 45° and 80°. The mode of the directions shows a similar pattern with the fracture analysis from the FMS logging data (see "Downhole Measurements").
The downhole variation of the dip azimuth can be subdivided into domains (Fig. F91):
Plotting and contouring the poles to the fracture planes on stereoplots (Fig. F92) illustrates the differences between domains. In the upper domain poles cluster into separate west and east groups with a greater number in the west. The middle domain shows similar bimodal clusters, although the clusters are less concentrated than in the upper domain. Additionally, minor clusters are recognized in the diagonal direction. In the lower domain the poles are more scattered than in the upper domain, and the clusters have a northwest-southeast trend.
The plots indicate that eastward- and westward-dipping fracture planes exist throughout the sedimentary column, with a bias for eastward-dipping planes in the upper and middle domains. In the lower domain, southeast and northwest-dipping planes dominate. The middle domain appears to be a mixture of both the upper and lower domains. These domain boundaries generally coincide with frequency of fractures (see "Lithostratigraphy"). An increasing frequency of fractures generally occurs in the middle domain. This indicates that the fracture pattern in the middle is a mixture of two pattern systems of east-west and southeast-northwest fracturing planes. Horizontal and vertical P-wave velocities have larger dispersion in the interval between 920 and 1010 mbsf, which could be related to the higher number of healed fractures and possibly the more variable fracture plane orientations that occur in this interval.
Two dominant fracture orientations, east-west and northwest-southeast, have apparently formed in different stress fields. The orientation of healed fractures in the upper domain strongly suggests an east-west extensional stress field. This agrees with borehole breakout information that indicates north-south compression and east-west extension. The bias toward eastward-dipping healed fractures might suggest an asymmetric sense of the extension.
However, the fracture direction in the lower domain indicates a northwest-southeast extensional field although the interval is accompanied by other minor directions. An analysis of conjugate healed faults with paleomagnetic orientation revealed an extensional direction of 117° at Site 584 (Niitsuma, 1986). The direction is consistent with the motion of the Eurasian plate relative to the Pacific plate (~120°). Niitsuma considered that structures were formed during uplift of the middle slope related to accretionary processes in the trench. The direction derived from the lower domain is consistent with the direction at Site 584, possibly caused by the same mechanism as Site 584.
Sixty-six open fractures through the hole were reoriented in the same manner as the healed fractures (Table T27, also available in ASCII format). Poles to the fracture planes (Fig. F93) are divided into the same three domains as the healed fractures. The open fractures have a high density of poles in the south for the middle domain and in the southeast and northeast for the lower domain. The open fractures may have resulted from release of the north-south compressional stress in the hole. The concentration of poles to the west in the upper domain may indicate that many of the open fractures originated from healed fractures that were cracked during drilling or handling of cores.
We also obtained several orientation data of bedding planes through the hole (Fig. F94). After reorientation, bedding planes have a gentle dip with a mean bedding plane dipping slightly, but insignificantly, to the east.