The distinctly low density of the proto-décollement zone influences the structural evolution both seaward and landward of the deformation front. Options that explain this low density include high radiolarian porosity, low grain density, and overpressuring.
Because the low-density interval at Site 1044 correlates with the upper half of the radiolarian claystone lithology (Fig. F5), we suspected that the cause of the low density was a high proportion of porous radiolarian tests. However, analysis of a series of samples from Site 672 penetrating the proto-décollement zone demonstrated that the radiolarian tests as open pore volumes comprise only 4% to 5% of the proto-décollement section (Wallace, Chap. 1, this volume). The density shift at the top of the proto-décollement zone is -0.12 g/cm3. Replacing the sediment matrix with 5% porosity because of radiolarians will only shift the density by 0.035 g/cm3 or ~30% of the observed decrease in density at the top of the proto-décollement zone.
Changes in grain density may also account for part of the density decrease in the proto-décollement zone. The generally high smectite and total clay contents decrease in the downhole transition to the proto-décollement zone (Fig. F5). According to the X-ray diffraction studies (Tribble, 1990), the decrease in smectite and total clays is accounted for by an increase in quartz and feldspar, a shift consistent with the presence of dispersed volcanic ash in this section. Therefore, the low density of the proto-décollement zone cannot be accounted for by an increase in smectite with its internally bound water. However, dispersed volcanic glass (variable but anomalously low grain density) and opal in radiolarians, sponge spicules, and diatoms (grain density of 2.1 g/cm3) may contribute to the observed density decrease. Because the amount of volcanic glass and opal is undetermined, this idea cannot be explicitly tested.
Finally, the low density and underconsolidation (Taylor and Leonard, 1990) of the proto-décollement zone could be explained by overpressuring. However, a modeling study indicates that the proto-décollement zone should have consolidated given the duration of time since its deposition (~20 m.y.) and permeabilities of the overlying section (Screaton and Ge, in press). Thus, the maintenance of low density caused by overpressuring is arguable.
In summary, no single factor can adequately explain the origin of the low-density proto-décollement zone. However, the combined effects of radiolarian porosity, low grain density, and overpressuring may combine to account for the observed density minima.
The proto-décollement zone was originally defined at Site 672 and correlated to the décollement zone at Site 671 based on lithology, age, and the small-scale faulting and extensional mud-filled veins in lower Miocene radiolarian claystones (Shipboard Scientific Party, 1988b). Shipboard scientists related the deformation at Site 672 to seaward propagation of thrusting from the deformation front, even though this hypothesis seemed mechanically unrealistic.
The LWD data indicate that the zone of faulting and veining lies in the zone of low density (Fig. F5). A careful analysis of the 3-D seismic data indicates that small-displacement normal faults sole out at the level of the proto-décollement zone (Fig. F7) (Teas, 1998). This later observation suggests that the proto-décollement zone is inherently weak, predicting the development of the décollement zone in this interval.
Given that the section is relatively uniform claystone above, through, and below the décollement zone, its development in the radiolarian mudstone would not be obviously predicted from lithology alone. The proto-décollement zone (with listric normal faulting) and the décollement zone (with thrusting) are correlated with the low-density radiolarian mudstone (Figs. F5, F6, F8). This low-density interval is most obvious seaward of the deformation front but persists as a residual of two low-density spikes, even at Site 948, farthest landward (Figs. F4, F6). Siliceous sediments are as strong or stronger than clays, so the radiolarian lithology offers no strength advantage (Fossum and Fredrich, 1998). However, the lower density and higher porosity is generally correlated with lower strength (Hoshino et al., 1972; Vernik et al., 1993). Therefore, the low density and high porosity of the radiolarian mudstone may account for the localization of failure in this lithology.
The normal faults are common in the section seaward of the deformation front, often sole out in the proto-décollement zone, and arguably disturb the seafloor <800 m from the deformation front (Fig. F7) (Teas, 1998). The latter observation suggests that the normal faulting is young and the compressional state of stress does not propagate far from the limits of the accretionary prism. Such extension may be caused by the flexure of the Tiburon Rise as it approaches the subduction zone. Moreover, incipient thrusting is not known more than several hundred meters seaward of the main frontal thrust (Fig. F7). Thus, any proto-thrust zone is virtually undeveloped at this margin in contrast to Cascadia (Cochrane et al., 1994) or southwest Japan (Morgan and Karig, 1995). The low taper angle of the northern Barbados accretionary prism suggests minimal coupling along the décollement zone (Davis, 1984), which is consistent with the absence of a proto-thrust zone. Simulations of fluid flow from the consolidating accretionary prism and incoming sediment section suggest elevated fluid pressures seaward of the deformation front, which could facilitate faulting of any type (Stauffer and Bekins, in press).
The high heat flow seaward of the deformation front at Site 672 and in surface measurements nearby was explained by fluid flow either along the proto-décollement zone or in sandy turbidites at greater depth (Fisher and Hounslow, 1990a). Anomalously fresh pore water at the level of the sandy turbidites argues for active flow in this interval (Shipboard Scientific Party, 1988b). However, the value of the proto-décollement zone as a conduit transmitting fluid from beneath the accretionary prism has been questioned because (1) the pore-water anomalies at Site 672 are not considered a strong indicator of fluid flow through the proto-décollement zone (Shipboard Scientific Party, 1997), and (2) there is no heat-flow anomaly at Site 543, which is closer to the deformation front but lacks the sandy turbidite section at depth (Tokunaga, in press). The statistical correlation and volumetric strain analysis by Tokunaga (in press) indicates that more rapid consolidation of the mudstones surrounding the sandy turbidites beneath the accretionary prism could account for the fluid sources necessary to produce the heat-flow anomaly at Site 672.