CONCLUSION

Permeability measurements show values of ~2–6 x 10–19 m2 when the effective pressure is ~2.5 MPa and the drained compressibility of the sample is ~1.3 x 10–8 Pa–1. If these values are representative of permeability and compressibility of fault wallrock under in situ conditions, diffusion of pore pressure in the wallrock would affect a characteristic length of 20–50 m in ~100–1000 yr. Considering that the fault zone itself is 20–30 m wide and that the diffusion distance varies with the square root of time, equilibration of pore pressure between the fault zone should be achieved in <100,000 yr. This implies that a higher pore pressure in the fault zone than in the surrounding sediments can be maintained only during transient events. These transient events may relate to slip events and to décollement propagation episodes (Bourlange et al., 2003).

Failure tests show that permeability increases as a result of the formation of a faultlike slip plane in the sample when confining stress is low. Sample 190-1173A-55X-5 (vertical) was fractured at 0.2-MPa confining effective stress, and this induced an increase of permeability by one order of magnitude. This fractured sample was then used in the triaxial cell for permeability measurements under increasing confining pressure. During this second phase, permeability decreased sharply between 1.23 and 1.57 MPa, whereas the sample experienced only a small (0.01) void ratio change (see Table T2; Fig. F2). This sharp permeability decrease likely corresponds to fracture closure. This suggests that once hydraulically conductive fractures have been formed, they may remain hydraulically conductive up to a relatively high confining stress (1.2 MPa). This permeability cycle presents similarity with transient increases in expelled water flow observed during sample loading (Byrne et al., 1993). These observations were interpreted as a transient increase in permeability associated with dilation and shear zone formation in the sample.

Friction coefficients determined in this study (0.37–0.40) are in the range expected for the clay-rich lithology and are not exceptionally low. This suggests a high fluid pressure is required inside the fault zone for decoupling at the décollement level. If a Coulomb wedge model is used (Dahlen, 1984), sliding along the décollement is allowed only when the excess pore pressure ratio, b*, at the décollement is >0.6:

b* = (Pf – Ph)/(Pt – Ph),

where

Pf = interstitial fluid pressure,
Ph = hydrostatic fluid pressure, and
Pt = lithostatic pressure.

This value is significantly higher than the 0.47 value estimated from compaction curves (Screaton et al., 2002) at ODP Site 808. This suggests a higher pore pressure in the fault zone during episodes of sliding.

NEXT