Time-of-Flight Measurements of Acoustic Wave Speed in a Sandy Sediment at 0.6–20 kHz

There is considerable interest within the underwater acoustics community as to whether a fluid model or a poroelastic (Biot) model provides a more accurate representation of sandy sediments. One key metric used to determine this is the acoustic wave speed in the seabed, since the Biot model predicts a sound speed that is frequency dependent whereas the traditional fluid model assumes a sound speed that is constant with frequency. Results obtained during the 1999 Sediment Acoustics Experiment (SAX99) showed some evidence of sound-speed dispersion [IEEE J. Ocean. Eng., vol. 27, no. 3, pp. 413-428, 2002]. The results were consistent with Biot model predictions that employed inputs based on geophysical measurements made at the site. However, only a limited data set was obtained at frequencies from 1 to 10 kHz where the model exhibited its greatest sound-speed variation. Furthermore, these were relative-rather than absolute-measurements of sound-speed dispersion. During the SAX04 sea trial, conducted in autumn 2004 about a kilometer from the location of the SAX99 site, acoustic data were collected on receivers buried in the seabed using a pair of transmitters located within the seabed and a third located in the water column directly above the buried receivers. This source geometry enabled direct time-of-flight (TOF) measurements of acoustic wave speed along all three Cartesian axes. The results are normalized by the acoustic wave speed in the overlying water. Horizontal measurements yielded absolute dispersion estimates but the vertical data were limited to relative estimates due to uncertainty in the depths of the receivers. Results show dispersion within the error limits of the measurement with normalized sediment sound speed increasing from 1.05 at 600 Hz to 1.13 at 20 kHz. The frequency dependence of the measured sound-speed ratios reported on in this paper is in agreement with a simplified poroelastic model [J. Acoust. Soc. Amer., vol. 110, no. 5, pp. 2276-2281, 2001] evaluated using physical parameters measured nearby during SAX99, but the measured sound-speed ratios are about 3% lower than the model predicts; however, some of the vibracores taken at the SAX04 site indicate the presence of small mud inclusions at about 1-m depth, and model results using the oases seismoacoustic model indicate that the lower sound speeds are consistent with the presence of a thin muddy layer. In addition, sound speed along the vertical axis showed substantially greater variability with frequency than did the measurements along the horizontal axes. Results obtained from a simple numerical model indicate that the greater variability in the vertical direction can be explained by interference from reflected arrivals from a low-speed reflector at approximately 1-m depth. Using the porosity β as a free parameter, a best fit of the poroelastic model to the data is obtained for β = 0.425 . Although this is higher than the value of β = 0.385 measured in the sandy sediment during SAX99, heuristic arguments based on the self-consistent model results and the vibracores are presented to support the hypothesis that localized muddy inclusions at the experimental site increased the average porosity over the horizontal propagation paths and resulted in the lower sound-speed ratios.