Velocity and density of sediments of Eirik Ridge, Labrador Sea: control by porosity and mineralogy

A 767-m section of late Neogene (0-8 Ma) terrigenous sediments was cored at Ocean Drilling Program (ODP) Site 646. Continuous downhole geophysical logs, 161 laboratory measurements of core porosity and density, and 63 laboratory measurements of core velocity are used to analyze in detail the effects of porosity and mineralogy on the acoustic properties at this site. Porosity (determined from a resistivity log) agrees well with rebound-corrected laboratory measurements. Mineralogical variations (potassium feldspar, quartz plus plagioclase, calcite plus opal, and clay minerals) for the interval 206-737 mbsf were determined by matrix inversion of three logs: bound water, potassium, and uranium/ thorium ratio. These calculated mineralogical variations are similar in major features to mineral abundances from smear slides, but the wide depth spacing of smear slides and their subjective, semiquantitative mineral abundances preclude a detailed comparison. Calculated grain densities from mineralogy are consistent with laboratory measurements. A pseudodensity log from porosity and grain density is similar in character to the rebound-corrected, bulk-density measurements on cores, but about 0.1 g/cm3 lower than core measurements in the interval 340-737 mbsf. We found from our analyses that a strong synergy exists between downhole geophysical logs and core measurements of porosity and density: (1) core recovery is best at shallow depths, and logs are more reliable at greater depths; and (2) agreement between laboratory and log measurements corroborates the different assumptions made when analyzing the two data types. At Site 646, this synergy does not extend to laboratory measurements of velocity; laboratory velocities are lower than in-situ velocities, but higher than expected when rebound is considered. Observed trends of laboratory and log porosity, density, and velocity as a function of depth at Site 646 are in reasonable agreement with empirical trends. In contrast, empirical relationships of velocity to porosity do not agree well with our data. Application of Hookean elastic equations to our data is hampered by the lack of shear wave velocities and the sensitivity of the technique to small errors in porosities. Nevertheless, this theoretical approach yields a pseudovelocity log that agrees remarkably well with observed in-situ log velocities.