DISCUSSION AND CONCLUSIONS

The motivation for analyzing grain-size distributions in such a comprehensive manner was to determine whether or not the character of the overlying sediment column exerts any influence on the transfer of hydrothermal fluids, either into or out of the underlying igneous basement. Because fluid migration is sensitive to the physical properties of sediments, one indirect way to address the link from lithology to hydrology is to determine whether or not physical properties change as a function of grain size. For the most part, our study quantified two obvious relations: hemipelagic and/or turbidite mud is finer grained than turbidite sand and silt, and physical properties within the stratigraphic column change in response to both initial sediment texture and depth of burial.

Shipboard measurements showed that a large range exists in the physical properties of sediments, particularly within the upper 100-150 m of the stratigraphic successions. Porosity values, for example, range from 80% to 30% (Fig. 5A). Much of this scatter can be attributed to the interlayering of several lithologies within lithostratigraphic Subunits IA and IB. When data from the coarser samples are segregated from the porosity values of mud samples, two compaction trends emerge (Fig. 5B). All but two porosity values for sandy samples fall between 55% and 35%, but there is no systematic change in sand porosity over a depth range of 0-120 mbsf. Conversely, mud porosity near the seafloor is greater than 70%; values drop to ~40% at depths below 500 mbsf. Data from the mud samples fit a compaction curve (of the form n = az^b) to depths of ~200 mbsf. Below 200 mbsf, a linear compaction trend provides a better fit to the data (Fig. 5A). Separation between the sand-layer data and the mud compaction trend is pronounced within the upper 40 m of the sediment column (Fig. 5C). At depths greater than 50 mbsf, overlap begins to occur between the mud compaction gradient and the porosity field for sand (Fig. 5B).

Giambalvo et al. (2000) carefully examined samples that were collected for consolidation tests and were able to discriminate between hemipelagic and turbidite muds. Initial porosity values (i.e., prior to consolidation tests) for the hemipelagic specimens are significantly higher than initial porosity values for turbidite muds. Although their grain-size characteristics are similar, the hemipelagic muds contain more foraminifers, and scanning electron microscopy (SEM) showed that their grain fabrics are random to subvertical. Random grain orientations probably result from deposition as fecal pellets and/or flocculated aggregates. In contrast, the turbidite muds contain few, if any, foraminifers, and their grain fabrics show systematic bed-parallel alignment of phyllosilicates (Giambalvo et al., 2000). X-ray diffraction analyses did not reveal differences in mineralogy between the two types of mud (Underwood and Hoke, Chap. 5, this volume). Interbedding between these two types of mud, with similar texture and mineralogy but contrasting initial grain fabrics, helps explain why there is so much scatter in mud porosity values within the upper 150 m of sediment (Fig. 5B).

One of the more intriguing phenomena to address during Leg 168 involves the transformation from a hydrologically open basement to a sealed basement. The early-stage sedimentary carapace of highly porous and permeable mud is gradually transformed to a thicker section of sediment that contains a more highly compacted and relatively impermeable seal at its base. With the exception of Site 1027, which includes a basal unit of basaltic sills, breccia, and carbonate-rich mud, the lithology resting above igneous basement is hemipelagic mud (Unit II). Once this fine-grained material compacts sufficiently, the basement becomes sealed, but exactly when this happens remains uncertain. At Sites 1030 and 1031, upflow of fluids through the sediment cover was inferred from pore-water profiles of conservative elements, and the estimated rate of upflow is ~2 mm/yr (Shipboard Scientific Party, 1997a). Evidently, the overburden at Sites 1030 and 1031 is too thin (<45 m) to collapse the pore fabric of the basal mud unit.

Giambalvo et al. (2000) showed that sediments from the Site 1030/1031 seepage localities are overconsolidated; underconsolidated conditions might be expected if fluid pressures were significantly greater than hydrostatic. Instead, fluid overpressures appear to be 5 kPa at the basement/sediment interface (Giambalvo et al., 2000). Because of their random grain fabrics, porosities for hemipelagic muds from the upflow sites are consistently higher when compared to undifferentiated mud (turbidite and hemipelagic) from comparable depths (<42 mbsf) at sites of no flow (Fig. 5D). The average difference in porosity between the two groups of shallow mud samples is ~7%. Those contrasts in porosity translate into 10x differences in permeability; in addition, modeling indicates that the hemipelagic mud of Unit II could sustain geochemically detectable flow (>0.1 mm/yr) up to burial depths of 150 m, assuming an overpressure of 5 kPa (Giambalvo et al., 2000). Thus, the overlying turbidite section (Subunits IA and IB) is entirely responsible for increasing the lithostatic load enough to compact the basal hemipelagic mud, but the textural characteristics of the turbidites are probably not important in the process of sealing the basement.