Mesoscopic analysis of split-core features were described shipboard (Kimura, Silver, Blum, et al., 1997). Key horizons were identified and samples were extracted for postcruise analysis. Material was impregnated with low-viscosity epoxy resin and ground down into slides for optical microscopic examination. Oriented sediment chips were slowly air-dried before being mounted on metallic stubs and sputter-coated with gold for scanning electronic examination operating in secondary mode.
Whole-round samples were taken at key intervals from both the reference site and the two prism toe sites. Permeability tests were conducted on cylindrical samples using the constant-rate-of-flow permeameter (Olson and Daniel, 1981) at various values of effective stress before and after shear (Fig. F2). Precise effective stress conditions were achieved by fluctuating the confining pressure and/or the back pressure. The back pressure required to eliminate any air held within the sample was consistently greater than 300 kPa in magnitude. Samples were placed in a flexible, impermeable membrane and housed in a triaxial cell. Fluid was infused into the sample at a constant, predetermined rate, creating a hydraulic gradient and promoting fluid flow along the axis of the cylinder. The associated differential pressure was measured using a differential pressure transducer accurate to ±1 kPa. Permeability is calculated using Darcy's law at steady state, occurring when a constant volume of fluid entering and exiting the sample is coupled with a steady differential pressure. Flow rate was controlled such that differential pressure across the sample did not exceed 100 kPa, equating to a flow rate of approximately 0.5 mL/min for these low-permeability materials. Samples were deformed axially at a constant rate of strain (10-5 to 10-8 s-1), and axial stress was continuously monitored to beyond failure (5%-30% axial strain depending on the strength of the individual sample). Where the initial permeabilities were high enough (>10-17 m2), fluid was infused through the sample simultaneously with deformation. Comparing the infused fluid volume with the volume of fluid exiting the sample provided information on volumetric changes induced by shear. For example, an increase in sample volume was identified when the volume of permeant entering the sample exceeded the volume of fluid exiting the sample. This volumetric behavior, when coupled with the differential response, allowed calculation of the active permeability change during deformation (the "dynamic permeability" of Stephenson et al., 1994). Where initial permeabilities were low (<10-17 m2), the sample was sheared undrained by closing the base tap of the triaxial cell, hence preventing any pore water drainage to occur. In this instance, no fluid was infused into the sample, and the differential pressure fluctuations could be directly related to deformation-induced fluid pressure changes.
After cessation of shear, the static permeability was measured at different values of effective stress. Comparison of the permeability behavior after shear with that measured under the same conditions before deformation allowed direct assessment of the influence deformation microfabrics exerted on sediment permeability.