The thick gouge interval in the basement section at Hole 976B (750-785 mbsf) seems to consist of metric scale, discrete bands of cataclastic rocks with anastomosing geometry. The band of cataclastic rock surround slices of wall rock with minor or no evidence of brittle deformation (Fig. 4), a geometry common to brittle fault zones (e.g., Koestler and Ehrmann, 1991; Sibson, 1996); therefore, we interpret this gouge interval as corresponding to a significant fault zone within the basement. The fault-gouge interval coincides with a major change in basement lithology, which is richer in gneissic rocks downhole (Fig. 3A). This observation, together with the rock types recovered and the brittle structures encountered, suggests that this interval marks a major fault in the basement section. According to the seismic images, we suggest that the fault zone may correspond to a less than 40° westward-dipping brittle fault (Fig. 5). The footwall of this fault is characterized by an increasing downward dip of the reference surfaces (below 710 mbsf the magnetic foliation dips more than 70°). The sense of shear in this fault could not be determined by any of the available drilling data. Formation MicroScanner data at Hole 976B on shallow-dip planes from fractures or foliations below the fault zone show a north-northeast to north-northwest variable trend with a general dip (<30°) towards the west (de Larouzière et al., Chap. 24, this volume).
We consider the majority of the breccia intervals to be cataclasites (sensu Twiss and Moores, 1992), as clear cataclastic structures have been recognized in most breccia samples. Structures include grain disaggregation and fracturing (Pl. 3, fig. 1, fig. 6; Pl. 4, fig. 3), and cataclastic flow, foliation, or lineation (Pl. 1, fig. 3; Pl. 3, fig. 1, fig. 5, fig. 6; Pl. 4, fig. 1, fig. 2, fig. 3, fig. 4, fig. 5). Cataclasites are known to form by frictional sliding and fracture processes in different tectonic settings: wrench tectonics (Blenkinsop and Sibson, 1992; Tanaka, 1992), thrusting or nappe-stacking (House and Gray, 1982), and extensional faulting (Malavieille, 1993).
A growing array of geological and geophysical data suggests that the grain-scale deformation mechanism in cataclasites, and the state of stress in active faults are largely controlled by high pore-fluid pressures in the fault zone (e.g., Carter et al., 1990; Wintsch et al., 1995). These pressures are caused by either fluids migrating to and through the faults during their movement or by fluids trapped in fine-grained clay gouge zones (Chester et al., 1985; Byerlee, 1990; Evans, 1990). The presence of fluids favors cataclastic deformation and hydraulic fracture by reducing the effective confining pressure, and cataclastic flows develop through successive cycles of softening-hardening of fault zones (Babaie et al., 1991; Wintsch et al., 1995).
Extensive internal microfracturing (splintered grains) observed in some metamorphic porphyroclasts of the reported samples may have developed by rapid decompression during exhumation of the metamorphic basement produced by extensional faulting. Further cataclastic structures postdate the intragranular microfractures (Pl. 4, fig. 1, fig. 2, fig. 3).
There is strong evidence for superficial extension in breccias and metamorphic rocks, thereby creating open spaces within displaced host rocks. In the uppermost basement cores from Holes 976B and 976E (Pl. 1, fig. 2, fig. 3, fig. 4), fractures have sediment-fill derived from above and are therefore considered neptunian dikes and sills (e.g., Bernoulli and Jenkyns, 1974). Neptunian dikes prove that the dilation and brittle tensional fracturing of the basement at Site 976 occurred in a submarine environment (i.e., by the late Serravallian; Shipboard Scientific Party, 1996). In Hole 976E, sediment injection into open fractures from below are denoted by convex-up laminations in sediment fill, thus suggesting strong fluid circulation between a connected open-fracture closed system in the basement high (Pl. 1, fig. 5).
Sediment injections into open fractures, and pervasive dolomitization, together with the distinctive geochemical composition in cataclastic zones, indicate fluid-assisted processes for brittle deformation at Site 976. Coarse-grained saddle dolomite suggests dolomitization by fluids at high temperature, comparable to those in which dolomitization resulted in experimental studies, that is, as low as 100°C (Gaines, 1980; Sibley et al., 1994). Our data reveal that the brecciated basement at Site 976 was an important pathway for fluids, probably including heat advention associated with the tectonic regime. Hydrothermal solutions are potential dolomitizing fluids, at least for replacive nonplanar and saddle dolomite. However, whether fluids correspond to younger or older sea water, somewhat modified by host-rock interaction, or just to hydrothermal brines, they deserve to be investigated by further geochemistry analyses, which is beyond the scope of this paper.