By observing the interrelationships of different structures, their internal characteristics (e.g., porosity contrast of deformed zones and adjacent undeformed sediment), and permeability behavior during laboratory deformation, it is possible to constrain the timing, the role of fluid pressure, and principal stress orientation within the prism, the décollement, and the underthrust unit.
Kink bands display both extensional and compressional offsets and may be both gravitational and tectonic in origin. There is no marked porosity difference within the kink bands and adjacent, undeformed sediment (as revealed by SEM observation) and displacements are small (1-2 mm), indicative of very minor finite strains along discrete discontinuities (Fig. F13).
Deformation bands offset kink bands and contain collapsed porosity. The close association with kink bands and relative crosscutting relationships suggests they formed by progressive strain and grain reorientation from deflected kink-band boundaries, and subsequent densification within the band. The bimodal orientation is feasibly a result of timing and extent of fluid pressure variations during shear (Fig. F13). Such volumetric behavior typifies normally consolidated sediment, and deformation bands oriented obliquely to the primary (consolidation) fabric may form in response to translation of horizontal principal stresses in the sedimentary pile. Laboratory tests (in particular Samples 170-1040B-2H-3, 136-150 cm, and 170-1040C-27R-1, 96-111 cm) consolidated during shear, accompanied by decreasing permeability both during and after shear. If this behavior is representative of natural processes, it can be inferred that such fabrics within the prism would not play a significant role in bulk prism hydrology. However, given that little overpressure exists within the prism, the fluid expulsion necessitated by porosity collapse of the deformation bands has been sufficiently transmitted out of the prism at some point in the structural history. Bedding-parallel deformation bands contain intensely aligned particles, constrained to extremely narrow horizons and are more representative of slip-under conditions of reduced effective stress. The geometries displayed here share similar characteristics to experimental structures of sheared overconsolidated sediment (Skempton, 1966). Typically, a dilating shear zone creates a zone of weakness that allows strain softening to occur within a deformation band, allowing large shear strains to develop within narrow discontinuities. The accompanying hydrological response may be a drop in fluid pressure (Bolton et al., 1999) or a concomitant increase in permeability (Stephenson et al., 1994; Bolton et al., 1998). Transient dilation during shear induces only small increases in permeability before ultimate porosity collapse yields a more permanent decrease in permeability both parallel and perpendicular to the shear zone (Brown et al., 1994). Although the actively deforming sediment may have contributed to sediment dewatering during formation of the shear zones, ultimate porosity collapse reduces the flow capacity. Shear deformation within the prism is therefore unlikely to significantly affect the bulk hydrology.
The dewatering and deformational history of the décollement appears less systematic than the prism. Microstructure associations suggest the variable presence of high-porosity zones and low-porosity zones (Fig. F14). Lower porosity regions are both clay rich and silt rich where cementation has occurred. Deformation appears constrained to clay-rich domains, and although porosity has collapsed, calcite and rhodocrosite precipitates within the shear zones suggest focused flow at some point in the structural history. Geochemical observations show that the upper, more brittle portion of the décollement is the principal focus of flow.
Given that the layer-parallel fracture networks transect extremely indurated and, presumably, intensely overconsolidated sediment, we can extrapolate our deformation experiments to processes occurring within the décollement. It has been documented (Fig. F12A) that localized, brittle failure can result in both fluid pressure decrease and a permeability increase of nearly four orders of magnitude with fluctuating fluid pressure (Fig. F12C). The resultant fracture flow is manifest as intervals of fluid pressure build up and shear, until failure and increased interconnectivity of the fractures allows the release of pressurized fluids. At low fluid pressures, flow is matrix dominated and of low permeability, switching to fracture flow and high permeability as the fluid pressure is raised. The mechanism relies upon connectivity between areas of different hydraulic potential that allow fluid circulation into more permeable horizons.
The presence of authigenic minerals within the lower, more ductile part of the décollement indicates that some fluid flow has also occurred within this interval, although the lack of discrete fabrics suggests that this flow is likely to be matrix dominated and therefore would not display the same variation of permeability with effective stress.
Although less numerous, deformation bands with morphologies comparable to those observed within the prism are present in the underthrust section. They maintain an appearance similar to the spaced foliation (Fig. F15) found near scaly fabric zones in the Barbados accretionary prism (Labaume et al., 1997), and to those formed in laboratory experiments (Maltman, pp. 426-429 in Borradaile et al., 1982). The bands are regularly spaced and constrained to certain intervals and have geometries similar to mud-filled veins found at the Peru margin (Brothers et al., 1996).
Other structures, normal faults in particular, are observed throughout the entire underthrust unit. Their prevalence in the upper Unit U1, which has undergone the most significant compaction, suggests that differential compaction is one operative mechanism of unit thinning. The fact that many of the fault surfaces are planar suggests that they have formed solely in response to loading by the overlying prism, rather than in response to flexure of the oceanic plate, where subsequent loading would result in compacted, curviplanar fault surfaces. We are precluded from possible interpretation of the consolidation behavior of the underthrust sediments from laboratory testing, as stresses achievable by the experimental apparatus were not of sufficient magnitude to exceed the maximum in situ consolidation stress.
The lower chalk unit displayed notably high permeabilities, even given the extremely indurated nature of the sediment. Figure F11 indicates that both the initial permeability and postshear permeability are of similar values, even though axial deformation resulted in discrete faults that bore a close resemblance to the normal faults seen in section. It seems unlikely that faulting within Unit U3 would impact significantly on bulk dewatering.