Site 1167 was located midfan to obtain as complete a section as possible without drilling too great a thickness and also to avoid shelf edge slumps. Hole 1167A was spudded in 1640 m water depth and drilled to 447.5 mbsf, achieving ~40% recovery. Because of the incomplete recovery, it was decided to cease core drilling and attempt wireline logging. This was unsuccessful, so Hole 1167B was drilled with logging-while-drilling (LWD) tools. This hole reached 261.8 mbsf before time ran out.
Site 1167 intersected predominantly clayey silty sands with dispersed rock clasts, minor beds of clays, and coarse sands. Two units were recognized by O'Brien, Cooper, Richter, et al. (2001) (Fig. F14):
a. Facies II-1 makes up most of the section. It is dark gray to reddish gray poorly sorted sandy silt, silty sand, clayey sand with dispersed granules, and pebbles and clast-poor diamicton. Foraminifers are a minor component, but diatoms, radiolarians, and sponge spicules are absent. Some benthic foraminifers are typical of the adjacent continental shelf. A few gravel beds (60% gravel and 40% matrix) are present.
b. Facies II-2 is represented by only a few beds, up to 3 m thick, of moderately sorted quartz coarse sand with granules and mud clasts.
c. Facies II-3 consists of decimeter-scale beds of dark gray clay and clay with thin (<1 mm) silt laminae and burrowed intervals. This facies has sharp upper and lower contacts with other facies. Rare sand grains are present but the facies lacks lonestones.
d. Facies II-4 comprises centimeter- to decimeter-scale beds of greenish gray to dark gray clay with dispersed sand and granules. Foraminifers and nannoplankton are common. Foraminifers are both planktonic forms (Neogloboquadrina pachyderma, sinistral) and benthic forms usually associated with mid-bathyal environments (O'Brien, Cooper, Richter, et al., 2001).
The low recovery in the hole made it impossible to estimate the number of glacial-interglacial episodes based solely on core samples. However, the LWD tools provided this information for the upper 260 m of the hole. LWD tools collect geophysical measurements of the sediment in the borehole wall using sensors in drill collars just behind the bit and are used where poor hole conditions prevent conventional logging (see the ODP Guide to Logging at www.ldeo.columbia.edu/BRG/ODP/LOGGING/TOOLS/tools.html). The configuration of LWD tools used at Site 1167 gave shallow and deep resistivity and spectral gamma readings (Fig. F15).
The combination of deep and shallow resistivity and natural gamma logs clearly shows 16 fine-grained interbeds within the 260 m logged. These clay-rich interbeds appear as low-resistivity spikes, commonly associated with small gamma peaks. Two resistivity peaks at 40 and 60 mbsf correspond to gamma lows, indicating that they are sand beds. The fine intervals represent interruptions to debris flow deposition, caused by retreat of the ice from the shelf edge. The thickest interval of low resistivity is Unit I. It has a lower gamma response than other fine intervals, probably because of a higher proportion of nonradioactive biogenic material such as diatoms. The other fine intervals are likely composed of Facies II-3 and II-4.
The thick intervals of high resistivity are mostly Facies II-1. Within these intervals, the log values are fairly uniform but blocks of slightly different resistivity suggest stacked individual Facies II-1 beds 5-20 m thick (Fig. F15).
In the absence of a sonic log or hole velocity survey, seismic data were tied to Site 1167 stratigraphy using a two-way traveltime vs. depth curve derived from core velocities corrected for rebound. Converting two-way traveltime for reflectors at the well intersection on line 149/0901 to depth in the hole produces an equivocal correlation to the hole stratigraphy with reflector depths not exactly matching major changes in lithology observed on the LWD logs. However, none of the reflector depths are more than 10 m from a significant lithologic change. The frequency content of the air gun used to collect line 149/0901 is such that 10 m is less than or equal to the limit of temporal resolution of the signal (Kallweit and Woods, 1982). This, coupled with uncertainties in the velocity model for the hole, make it difficult to determine with certainty if the reflectors really correspond to mudstone units or more subtle lithologic changes seen on the logs closer to the predicted depth. The fact that the surfaces were mapped on the basis of truncation and downlap suggests that they should not necessarily correspond to mudstones in the drill section. Therefore, each surface that intersects the hole should be considered independently.
Surface pp-2 is predicted to be at 68 mbsf at Site 1167, just below a major shift in gamma ray and resistivity logs at 62 mbsf (Fig. F15). This level also features a major discontinuity in magnetic susceptibility values. A mudstone horizon is present at 59 mbsf. It seems likely that pp-2 is the seismic expression of the lithologic change seen in the geophysical logs. Surface pp-3 at 133 mbsf is ~12 m below the nearest overlying mudstone and 8 m above the next below (Fig. F15). It does correspond quite closely to a step in the gamma log that may represent a change to slightly cleaner, thicker units within this fairly uniform interval. Surface pp-4 at 226 mbsf is just below major changes in geophysical logs and sediment composition including magnetic properties (Shipboard Scientific Party, 2001), grain size, and clay mineralogy (Forsberg et al., 2001). Surfaces pp-5 and pp-7 do not intersect the logged part of the section. Surface pp-5 probably intersects the hole at 265 mbsf, whereas pp-7 is close to total depth.
The combination of sediment facies in the hole and the seismic facies indicate that the Prydz Channel Fan is dominated by debris flows with thin interbedded facies. Unit I records a period of hemipelagic deposition, as indicated by the fine-grained sediments and diatoms and sponge spicules. The lonestones are likely ice-rafted detritus, and normally graded sand beds are turbidites. Unit I in Hole 1167A is identical to the sediments recovered from across the fan in shallow cores (O'Brien et al., 1995; Golding, 2000).
For Facies II-1, poor sorting, abundant floating clasts, reworked shelf benthic foraminifers, and lack of visible grading indicate deposition as debris flows derived from ice grounded at the shelf break. On seismic lines, the debris flows show as reflection-poor intervals that may be mounded and, in places, show chaotic mounded reflectors. The downslope terminations of some of the youngest flows can still be seen as down-to-basin steps in the lower fan surface. The spacing of seismic reflectors is consistent with the scale of unit thickness of Facies II-1 at Site 1167 as indicated by the logging data. The GI gun used has a signal with a frequency content that allows the resolution of interfaces 10 m apart, which is similar to the minimum thickness of Facies II-1 packages (Fig. F15). Lenticular units with chaotic fill represent slumps formed by failure of fan sediment rather than primary debris flows derived directly from the grounding zone.
Facies II-2 probably represents turbidites generated by mass movement on the fan surface or by sinking of dense, turbid plumes (Hunter et al., 1996). Turbidity currents are also suggested by presence of the channel-levee units visible on seismic lines. The rarity of channel-levees is consistent with the rarity of Facies II-2 at Site 1167. Facies II-3 and II-4 represent interruptions to debris flow sedimentation. Laminated Facies II-3 is probably bottom-current deposits, whereas Facies II-4 is hemipelagic sediment deposited with little current activity.
At the fan top on the shelf, the high amplitude of the pp-12 reflector and several prefan shelf reflectors suggest high impedance contrasts that may result from sustained erosion and/or overcompaction associated with either a hiatus or relatively severe subglacial erosion (Cochrane et al., 1995; Cochrane and Cooper, 1991; Solheim et al., 1991). The change from higher aggradation to stronger progradation suggests a reduction in the volume of subglacial till deposited on the shelf compared to the volume of material delivered to the shelf edge at the glacier sole. This could indicate an increase in basal shear stress or a reduction in the strength of the subglacial till. Both such effects are consistent with the development of a fast-flowing ice stream (Alley et al., 1989). The contrasting persistence of vertical aggradation post-pp-12 times shown in Four Ladies Bank might result from deposition by slower-moving ice that did not transport much debris to the shelf edge (Boulton, 1990).
The sequence of events that led to the geometry seen in Prydz Channel must have been as follows:
The pattern of foreset truncation followed by topset preservation suggests cycles of erosion followed by increased sediment accommodation on the shelf. On nonglaciated margins, such cycles would be the result of sea level cycles superimposed on subsidence (Posamentier et al., 1988). In the case of Prydz Bay, tectonic subsidence is probably low compared to the rates of glacial erosion and deposition; however, some isostatic subsidence of the shelf edge induced by shelf progradation is possible (Boulton, 1990; ten Brink et al., 1995).
Consideration needs to be given to the controls on subglacial deposition. Glacial erosion and deposition are controlled by a range of factors (Boulton, 1990). High ice velocities produce high basal shear stress, which favors erosion. High vertical effective pressure, which is the weight of ice thickness minus water pressure at the bed, also produces high basal shear stress. Till rheology also influences erosion and deposition (Boulton and Hindmarsh, 1987; Murray, 1997). Till deposition is favored by lower velocities, lower vertical effective pressure, and rapid basal melting that delivers sediment to the glacier sole. Basal freezing is thought to entrain sediment in the basal ice (Alley et al., 1997).
For a major ice mass grounded at the shelf edge, vertical effective pressure is likely to vary with sea level change though the glacial cycle as well as ice volume (Boulton, 1990). The pattern of truncated foresets separated by a topset that passes into a foreset can be explained by the response of the ice stream to a sea level cycle. Falling sea level increases the height of ice above buoyancy and increases the depth of the deforming bed, eroding the sediments beneath. As sea level begins to rise, the process reverses, depositing a subglacial topset bed as the ice decouples from the deforming bed (Murray, 1997) overlain by grounding zone deposits and proximal glaciomarine facies if the grounding zone retreats from the shelf (Boulton, 1990). This retreat phase produces the foreset-topset unit. The preservation of topset beds in Prydz Channel from successive erosion episodes could result from relatively minor erosion beneath the outer part of the glacier during each major advance. The slight seaward dip of topsets suggests that preservation was enhanced by subsidence of the outer shelf sediment wedge caused by differential compaction and loading of the crust (ten Brink et al., 1995).
Although the tie between the drill site and seismic section is uncertain, consideration of the likely processes on the fan surface might explain why the mapped surfaces do not correspond to the obvious fine-grained units. It is thought that trough mouth fans grow primarily during episodes when the ice is grounded at the shelf edge (Boulton, 1990; Vorren and Laberg, 1997). The erosion surfaces mapped within the fan probably formed during periods of relative sediment starvation when the ice had retreated. During such periods, the fan surface could be reworked by contour currents or have a hemipelagic drape deposited on it, forming mudstone intervals. Boulton (1990) points out that debris deposited near the top of the fan may be reworked during such "interglacial" periods. This is also suggested by the surficial slump scars near the top of the Prydz Channel Fan (Fig. F4). Therefore, the erosion surfaces mapped within the fan could conceivably be overlain by thin debris flows as well as by hemipelagic intervals. Such reworked debris flows might contain less clay and so appear to have lower gamma and resistivity values than debris flows beneath. This may be the case for surfaces pp-3 and pp-4, which correspond to subtle changes in logging values (Fig. F15). Such debris flows may also contain foraminifers with "interglacial" isotope values.
The isopach maps show that the Prydz Channel started as a fairly broad feature that deposited the more sheetlike interval pp-12 to pp-7 but deposited a strongly lobate fan from pp-7 to pp-2. This suggests preferential downcutting and sediment transport along the channel axis through that time. The last stage of sedimentation (pp-2 to present) produced a more drapelike deposit with debris flows and, finally, hemipelagic sediments deposited evenly across the fan.