At Site 1108, a comprehensive downhole profile of physical properties measurements was obtained on whole cores using the MST and on discrete samples from split cores. Even though MST data are reliable only in APC and nonbiscuited XCB and RCB cores, MST measurements were nevertheless made on all 51 RCB cores from Site 1108, irrespective of core condition. Our rationale was to obtain physical properties data in the event that geophysical logging of the borehole ultimately proved to be impossible or limited, as was the case for this site. Depending on the induration of the sediment, thermal conductivity was measured from either unconsolidated whole cores or rock slices.
Bulk density, grain density, and porosity were calculated from the wet mass, dry mass, and dry volume of each sample using the moisture and density (MAD) method for soft and semilithified sediment in Cores 180-1108B-1R through 18R (0-156.3 mbsf). Below this depth, the index properties were determined through the MAD hard-rock method. All data are presented in Table T11 (also in also in ASCII format in the ASCII TABLES directory).
The MAD-determined bulk densities of the two near-seafloor samples (lithostratigraphic Unit I; see "Lithostratigraphic Unit I") are 1.44 and 1.48 g·cm-3 (Fig. F40). Lack of recovery prevented measurement of index properties from 1.43 to 63.13 mbsf (lithostratigraphic Unit II). From 63.13 to 83 mbsf (lithostratigraphic Unit III and Subunit IVA), bulk densities were significantly higher, averaging 2.03 g·cm-3. Following an interval of no recovery between 91.3 and 110.6 mbsf, bulk densities decrease to an average 1.92 g·cm-3 within Cores 180-1108B-16R and 17R (130-158.6 mbsf; Fig. F40), coinciding with the upper limits of an observed brittle fault zone (see "Structural Geology"). Below this depth, bulk densities increase slightly downhole. A second interval of lower bulk density is present between 350 and 400 mbsf (average values of 2.18 g·cm-3) and coincides with the top of another zone of brittle deformation (see "Structural Geology").
Little correlation exists between the bulk densities derived from direct measurement (i.e., MAD-derived densities) and the gamma-ray attenuation porosity evaluator (GRAPE) derived densities. This lack of correlation is likely a function of core fracturing, biscuiting, and a reduced core diameter, all of which affect the GRAPE-derived densities. Nevertheless, restoring the MAD-derived bulk densities onto the GRAPE-derived densities (Fig. F41) indicates that the maximum GRAPE-derived densities agree reasonably well with the MAD-derived densities, especially above ~200 mbsf. From 220 mbsf to the base of the borehole (485.2 mbsf), the maximum GRAPE-derived densities appear to be systematically lower than the MAD-derived densities by 0.2-0.3 g·cm-3 although the same general trends exist (e.g., the local minimum at 360-375 mbsf). We conclude that despite the large scatter within the GRAPE-derived density measurements, the maximum value either underestimates or approximates the bulk sediment density. A full compilation of GRAPE data is presented with the MST measurement data set (in also in ASCII format in the ASCII TABLES directory) on the accompanying LDEO CD-ROM.
Based on the index properties measurements, the average grain density is 2.72 g·cm-3 with minimum and maximum grain densities of 2.62 and 2.80 g·cm-3, respectively (Fig. F40). Maximum sediment grain densities exist over a depth range of 230-300 mbsf and correlate with the thick, high-frequency, volcanogenic turbidites of lithostratigraphic Subunit IVA (see "Lithostratigraphy"). Grain densities show large scatter in the interval from 100 to 170 mbsf, varying from 2.62 to 2.75 g·cm-3. Much of this scatter is attributed to variations in carbonate content (see "Organic Geochemistry"). From 170 to 270 mbsf, grain densities increase toward a maximum of 2.80 g·cm-3, after which they decrease back to the average value of 2.72 g·cm-3 (i.e., from 270 to 380 mbsf). Below 380 mbsf, the grain density is nearly constant to the bottom of Hole 1108B (485.2 mbsf). Although thin carbonate-rich layers exist within the section, for example, at 220 and 250 mbsf (see "Lithostratigraphy"), the general variation in grain density with depth does not show a relationship with carbonate content.
Increased turbidite frequency within lithostratigraphic Subunit IVA indicates the rapid input of clastic material, as shown by the large sediment accumulation rates over the 200-380 mbsf interval (see "Sediment Accumulation Rate"). The steady increase in the percentage of relatively high density material of 2.8-3.04 Ma age between 250 and 350 mbsf (see "Biostratigraphy") possibly indicates the input of clastics from a different source area, such as the unroofing of the metamorphic core complexes of the D'Entrecasteaux Islands or the first denudation of mafic-rich ferromagnesium rocks from the forearc sequences (see "Lithostratigraphy").
Porosity is calculated
from the MAD index properties measurements used to determine bulk density and
grain density. Consequently, the porosity curve as a function of depth mirrors
that of the bulk density, with minor differences caused by variations in grain
density. Porosity varies from 38% to 49% at depths of 63-122 mbsf to 20%-30%
toward the base of Hole 1108B (Fig. F40).
The general behavior of porosity as a function of depth for Hole 1108B is
consistent with the expected negative exponential variation that characterizes
normally compacting sediments. This empirical relationship is termed Athy's Law
(Athy, 1930). Fitting a least-squares exponential curve to the observed
porosity-depth relationship, the predicted surface porosity is 48% with a
compaction decay constant of 0.0016 m-1 (Fig. F42A;
correlation coefficient of 0.68). The inverse of the decay constant (617 m) can
be physically interpreted as the depth over which porosity is reduced by a
factor of ~
(i.e., 1/e) of its initial value, that is, a compaction "decay depth."
It is clear from this figure that the depth-porosity behavior of the deeper
sediments is not consistent with the surface porosity. Further, if the surface
porosity is constrained to be 78%, then the compaction decay depth must vary
between 120 and 200 m to envelope the scatter of porosity with depth. Such
compaction constants are unrealistically low.
Based on the regression least-squares curve shown in Figure F42A, the porosity data from Hole 1108B show an obvious discontinuity between the near-surface sediments and the remainder of the data set beginning at 63.13 mbsf. Core recovery within the 14.5-63.13 mbsf zone was poor, consisting of pebbles and rubble of basalt, gneiss, granodiorites, and sandstones (see "Lithostratigraphy"). Sediments below 63.13 mbsf have porosities inconsistent with their depth, suggesting that they may have been compacted by the weight of overlying units that have since been eroded. Sediment erosion is consistent with (1) the interpretation of reflection seismic data indicating the truncation of reflectors at the seafloor (see "Miocene-Quaternary Arc and Forearc" in "Regional Setting" in the "Background and Regional Setting" chapter) and (2) the measured high sediment velocities (e.g., 2024-2713 m·s-1 at depths of 70-130 mbsf) beneath the inferred unconformity at 62.7 mbsf. Sediment erosion could also contribute to the apparent low sedimentation rate derived from biostratigraphy (see "Sediment Accumulation Rate"). To estimate the thickness of eroded sediment, we recalculated a least-squares exponential curve, using only the porosity data between 63.13 and 485.2 mbsf. For this case, the least-squares fit predicts an initial surface porosity of 44% and the compaction decay constant translates to a decay depth of 740 m (see Fig. F42B; correlation coefficient of 0.58). Extrapolation of this porosity-depth relationship to the surface, assuming an initial porosity of 73%-75% (i.e., equal to the near-surface sediments from Hole 1108B), implies that ~385 m of sediment has been removed from a section that once existed above lithostratigraphic Unit IV in Hole 1108B.
The PWS1 and PWS2 insertion probe system was used to measure the transverse (i.e., perpendicular to the core axis) and longitudinal (i.e., along the core axis) P-wave velocities in unconsolidated sediments. Sediment velocities obtained from the upper several meters of Core 180-1108B-1R (0-8.6 mbsf) yielded values between 1518 and 1527 m·s-1, as expected for unconsolidated surficial sediments (Fig. F43). The PWS3 contact probe system was used to measure the P-wave velocity of ~10-cm3 cube samples of semilithified and lithified sediments cut from the RCB cores. Velocities were measured in the x and y (transverse) and z (longitudinal) directions (Fig. F43). The P-wave logger (PWL) measured velocities from MST measurements on unsplit cores do not correlate well with those measured using the PWS1, PWS2, and PWS3 system and show extreme scatter, possibly caused by drilling disturbance. Because of the poor quality of the PWL-measured velocities, only PWS1-, PWS2-, and PWS3-derived velocities are discussed further. Below ~150 mbsf, longitudinal velocities increase from an average of 2250 to ~3000 m·s-1 at 380 mbsf. Curiously, velocities abruptly decrease to an average of ~2600 m·s-1 below 380 mbsf. High velocities, particularly those in excess of 3000 m·s-1, correlate strongly with the existence of carbonate and siliceous cements (e.g., 220 and 380 mbsf). Triaxial seismic velocity measurements indicate that transverse and longitudinal velocities typically vary by less than 20%. The bias is toward lower longitudinal velocities compared to transverse velocities (i.e., across bedding, positive anisotropies; Fig. F44). A compilation of the PWS1, PWS2, and PWS3 data is located in Table T12 and in also in ASCII format in the ASCII TABLES directory.
As described in "Thermal Conductivity" (in "Physical Properties" in the "Explanatory Notes" chapter), methods for measuring thermal conductivity depended on the degree of induration of the sediment cored. The needle probe full-space method was applicable only to Core 180-1108B-1R. Between Cores 180-1108B-2R and 14R, measurements were either not possible because of lack of recovery or unsuccessful because of the disrupted nature of the cores. From Core 180-1108B-15R to the bottom of Hole 1108B, thermal conductivity was determined using the half-space method. Because the measurements were time consuming, the number of measurements per core was reduced from one per section to one every two sections after Core 180-1108B-20R. In addition, the number of measurement repetitions on each sample was reduced from four to two in cases where results from the first two repeat measurements varied by less than 1%. The reported value is an average of the repeat measurements. Refer to Table T13 (also in also in ASCII format in the ASCII TABLES directory) for thermal conductivity data.
In the upper several meters of sediment (Core 180-1108B-1R; 0-8.6 mbsf), measured thermal conductivities were 0.8-0.9 W·m-1·ºC-1 (Fig. F45). As core conditions became conducive to preparing half-space samples for thermal conductivity measurement (i.e., deeper than 121.64 mbsf), thermal conductivities routinely exceeded 1.0 W·m-1·ºC-1. The thermal conductivities show an overall increase downhole to ~322 mbsf, where values reach 1.7 W·m-1·ºC-1. Below this peak, thermal conductivities remain nearly constant or decrease slightly with depth. In addition to the large-scale trends, the thermal conductivity data show variations that are probably related to lithostratigraphic changes within the turbidite sequences.
Magnetic susceptibility reflects changes in magnetic mineralogy and, as a result, is widely used as a proxy for lithostratigraphic variations. The quality of these data are degraded in XCB and RCB sections where the core may be undersized with respect to the liner diameter and/or disturbed. Nevertheless, the general downhole trends can be useful for stratigraphic correlations. The MST and AMST susceptibility data are shown in Figure F31 and the full magnetic susceptibility meter (MSM) data set can be found as part of the MST compilation (in also in ASCII format in the ASCII TABLES directory) on the accompanying LDEO CD-ROM. The AMST susceptibilities are termed point values because a smaller volume is measured compared with the MST, which uses the full uncut cores (see "Paleomagnetism"). This volume difference also explains the variation in the susceptibility amplitude between the two measurement procedures. The close correlation between the two verifies that both the spatial variation and amplitude are being successfully measured.
By comparing the magnetic susceptibility with the remanent magnetization intensity (see "Paleomagnetism"), it is possible to determine if the minerals carrying the magnetic remanence are also responsible for the susceptibility. Whereas magnetite and hematite are commonly the carriers of magnetic remanence, it is usually the clay minerals and clay content that dominate the susceptibility. The ferro-magnesium-rich clays are smectite and, less frequently, chlorite. To evaluate this association for Site 1108, the magnetic susceptibility, natural gamma ray (NGR), and remanent intensity were compared (Fig. F46). No clear relationship was found between these parameters. For example, the remanent intensity maximum at 200 mbsf is not reflected in either the NGR, which increases from 190 to 210 mbsf, or the susceptibility, which tends to be relatively constant over the 190 to 215 mbsf interval. Some minor correlations do exist, such as the high susceptibility and remanence peaks at ~218 mbsf. From Figure F46 we conclude that, in general, the minerals controlling the magnetic susceptibility are different from those controlling the remanent magnetic intensity. Further, the lack of correlation between NGR and the magnetic susceptibility suggests that the clays and sands containing radioactive material are distinct from those units rich in ferromagnesium minerals.
Comparing the grain-size distribution, magnetic susceptibility, and remanent intensity on an expanded scale allows us to (1) investigate the cause of magnetic susceptibility variations, and (2) determine if a relationship exists with grain size, an observation that was made on various ODP cruises (e.g., Leg 169; Fouquet, Zierenberg, Miller, et al., 1998). In Figure F47A, the general decrease in magnetic susceptibility from 217.5 to 218.5 mbsf and many of the individual peaks appear to correlate with the location of fine to coarse sands, whereas the silts and clays tend to be associated with relatively lower values of magnetic susceptibility. Little relationship exists between the remanent magnetic intensity and magnetic susceptibility. Given that we should not expect any relationship between susceptibility and grain size, per se, Figure F47A implies that the sand fraction of the turbidite sequences most likely contains magnetic material derived from a different source region than the intervening clays. However, it is clear from Figure F47B that the sands do not always induce high magnetic susceptibility. For example, whereas the sands at 255-256.5 mbsf and 256.5-256.8 mbsf have relatively high magnetic susceptibility, the sands at 257.8-257.9 mbsf and 258-259 mbsf show no such correlation. It would seem that both the clays and the sands can contain magnetic material: smectite, chlorite, and pyritic clays, and magnetite within the sands. The mixing between magnetic and nonmagnetic clays and sands implies sediment input from multiple provenances. Magnetic and nonmagnetic carrying provinces are likely represented by the New Guinea mainland and associated islands (e.g., D'Entrecasteaux Islands) and the carbonate banks of the Trobriand platform and the Egum reef, respectively.
The NGR count was recorded on the MST. A full compilation of NGR values is presented with the MST measurement data set in ASCII format on the accompanying LDEO CD-ROM. Clay minerals, being charged particles, tend to attract and bond with K, U, and Th atoms so that an increasing NGR count typically correlates with increasing clay/shale content. In contrast, sand-prone and carbonate units usually tend to be characterized by low NGR counts. These relationships can be used to define the location of shale-prone and sand-prone formations down the borehole. However, this relationship will begin to break down in poorly sorted sequences, such as volcaniclastic units and/or with peculiarities in the source region mineralogy.
The NGR data are rather noisy even though the sampling period was intentionally set high (at 28 s) to increase precision. Nevertheless, the NGR profile shows a number of general trends (Fig. F46). From 62 to 160 mbsf and from 170 to 200 mbsf, there are two segments of downward-decreasing NGR count separated by an abrupt increase in the NGR count between 155 and 165 mbsf. This same location has been subjected to brittle deformation (see "Structural Geology"). From 200 to 250 mbsf the NGR count generally increases, followed by a segment from 250 to 340 mbsf in which the NGR count at first decreases and then increases. There is a decrease in the NGR count at 350-360 mbsf followed by an increasing segment to 380 mbsf. From 385 to 400 mbsf, there is a rapid decrease in the NGR count after which the average trend steadily decreases to the bottom of the hole (485.2 mbsf). The NGR count minimum at 180-200 mbsf and maximum at 370-380 mbsf correlate with clays at the base of lithostratigraphic Subunit IVB (see "Lithostratigraphy"), and the sands and conglomerates of lithostratigraphic Subunits IVA and IVC, respectively (Figs. F48A, F48B). However, the clay units between 160 and 170 mbsf are also characterized by a high NGR count (Fig. F48A). This implies that both radioactive mineral-bearing sands and clays exist in the section. Consequently, in contrast to the typical relationship of NGR count and grain size, a decreasing NGR count with depth may indicate a coarsening upward sequence and an increasing NGR count with depth may indicate a fining-upward sequence at Site 1108. Note that other factors can also affect NGR count, such as variations in the calcium carbonate content and specifics of the clay mineralogy.
