Review of the velocity and density measurements made during Leg 199 reveals consistent patterns in the variability of the properties and factors that influence them. Velocity tends to display greater variability than density. Overall, the agreement between discrete sample wet bulk density and corrected GRA density is very good. The match between PWL velocity and velocities determined with the insertion and contact probes is weaker. The coring process disrupts the sediment fabric to various degrees, potentially resulting in volumetric changes that affect the bulk density and, consequently, the velocity. The fabric disruption also affects the rigidity of the sediment, which may explain the greater variability displayed by the velocity.
Simple differences in lithology, as determined visually, explain much of the differences in the properties of the sediments. Radiolarian ooze and radiolarite are distinctive in their properties. These sediments display uniformly low bulk densities, even at burial depths greater than 200 m. The siliceous sediments also are characterized by unusually high velocity for their density. There are several possible explanations for this anomaly. One explanation is that the shape of the radiolarians and their interlocking spines creates a stiff sediment framework. Another possibility is that low density results in part from high intraparticle porosity and the relatively large size of radiolarians and low interparticle porosity contribute to greater rigidity of the sediment structure. This latter explanation has been proposed to explain unusually high velocity in foraminifer-rich sediments (Hamilton et al., 1982; Bachman, 1984). As a result of either interlocking spines or low interparticle porosity, increased sediment rigidity produces a higher shear modulus and higher than expected velocities in the siliceous sediments. The pelagic clays, as a group, are characterized by low velocity and low bulk density and lack consistent trends with depth or composition. The properties of the nannofossil ooze are more regular in the relationship between porosity and velocity and the increase in velocity with depth. With increasing burial depth, the transformation to chalk results in a significant velocity increase.
Regression analyses indicate that sediment composition has a significant influence on wet bulk density. Regressions of wet bulk density with weight percent CaCO3, LAS mineralogy, and bulk sediment geochemistry all demonstrate a positive correlation between bulk density and calcium carbonate content. The association between bulk density and weight percent CaCO3 is well established to the extent that GRA density has been used as a proxy for weight percent CaCO3 (Herbert and Mayer, 1991; Mayer, 1991; Hagelberg et al., 1995). The high correlation coefficient for the Leg 199 sediments, however, is somewhat deceptive in that the best-fit line is anchored by concentrations of data at approximately zero CaCO3 and between 75 and 95 wt% CaCO3 with a fall off of data from the trend line between these end points (Fig. F7). Results of the regressions of wet bulk density with LAS mineralogy and bulk geochemistry data explain a greater amount of variance in bulk density than the regression with percent CaCO3 alone. Higher R2 values in these analyses reflect the greater number of variables in the regression models and the larger sample populations. As with the use of the GRA density as a carbonate proxy, Vanden Berg and Jarrad (2004) used a regression of the GRA density with the LAS data to convert the GRA density into a mineralogy proxy.
In contrast to the excellent correlation between the bulk density and the various measures of sediment composition, correlation of the velocity with weight percent CaCO3, LAS mineralogy, and bulk sediment geochemistry is weak to nonexistent. Despite the lack of quantitative relationships between velocity and sediment composition, there are general patterns of differences in velocity among the principal lithologies.
It is difficult to assess the success in estimating in situ values of density and velocity. The agreement between laboratory and logging densities is good, after the laboratory composite sections have been shifted and compressed to match the GRA density with the HLDT density (Fig F9). However, characterizing the elastic rebound by the difference in the logging and laboratory densities produces ambiguous results without a clear relationship between density reduction and overburden pressure or depth (Fig. F11). The lack of clear trends in the mechanical rebound in part results from shallow penetration at Sites 1218 and 1219 and the relatively short overlapping intervals of laboratory and logging densities. The lack of logging velocities limits the evaluation of the in situ velocities estimated from the laboratory data. The extent to which density is primarily responsible for impedance differences may be advantageous in generating synthetic seismograms from density profiles alone. Matching reflectors on the synthetic seismograms with those of the seismic profiles may be the ultimate test of the accuracy of the estimates of in situ density and velocity.