A 5-year ONR Departmental Research Initiative (DRI) on the interaction of high-frequency sound with the seafloor began in October 1997. The DRI addresses high-frequency sound penetration into, propagation within, and scattering from the shallow-water seafloor at a basic research (6.1) level. This overview summarizes the acoustics issues associated with the DRI, outlines the environmental characterization required to address the acoustic issues in a shallow-water experiment, and concludes with a brief description of some aspects of the first experiment. Program descriptions for investigators associated with the DRI follow in separate sections.
One motivating issue for the DRI is the detection and classification of objects, such as mines, buried in sediments. The character and relative strength of the returns from the buried object and from the sediment itself will determine whether detection and classification are possible. An improved understanding of the coupling of sound into sediments, of the propagation and attenuation within the sediment, and of the scattering from the sediment interface and from interior inhomogeneities should lead to improved models for predicting when buried objects can be detected and classified.
Basic research is particularly needed for sandy sediments that have sound speeds greater than that of the water above. Simulations in which the sediments are modeled as fluids with flat interfaces predict a critical grazing angle, often in the 20°–30° range, below which no appreciable acoustic energy will penetrate into the sediment. There is substantial evidence [1-4] that penetration occurs below this critical grazing angle, implying that the fluid-sediment, flat-interface model is inadequate. It has also been shown that using a visco-elastic model to account for shear effects does not significantly change the predictions for the compressional wave within sand sediments [5]. Thus, it is an important scientific issue to fully understand the factors that contribute to the acoustic penetration at subcritical grazing angles.
Important scientific issues also remain regarding acoustic penetration into sandy sediments at angles above the critical angle, where the acoustic coupling will be substantially higher. In silty or muddy sediments the critical angle may be very small or even nonexistent, implying relatively high penetration levels in the sediment over essentially all grazing angles. For acoustic penetration above the critical angle, loss of spatial or temporal coherence due to forward scattering from the rough sediment interface or volume inhomogeneities is important for imaging (and possibly other means of classifying) buried objects. At very low grazing angles (but still above critical), propagation effects due to vertical gradients could be especially important, as is an accurate knowledge of attenuation.
Basic research is also needed to better understand backscattering and general bistatic scattering from sediments at high frequencies. Questions remain about the dominant scattering mechanisms versus frequency (interface roughness, volume inhomogeneities, bubbles, sediment grains, and other discrete scatterers). The importance of fine-scale stratification and sound speed gradients for modeling bottom interaction at high frequencies needs to be better understood, and the importance of multiple scattering also needs to be clarified. In the context of detecting and classifying buried objects, accurately modeling the interference due to seafloor reverberation is important. A better understanding of high-frequency acoustic propagation in sediments should improve high-frequency acoustic modeling and help improve the performance of high-frequency sonars in shallow water.
Three mechanisms have been considered as possible contributors to subcritical penetration: (1) the porous nature of the sediment, which could lead to a second, slow compressional wave [1] with a wave speed less than the speed of sound in water, in which case there would be no critical angle for that wave; (2) roughness of the water/sediment interface, which could diffract or refract energy into the sediment [6, 7]; (3) volume inhomogeneities within the sediment, which could scatter the evanescent wave (propagating along the water/sediment interface) into the sediment. Experimental data acquired both in the field and in the laboratory have been interpreted using a porous (Biot) model for the sediment which can predict a slow compressional wave [1]. However, it has been shown that the results of experiments carried out to date could also be explained as a result of roughness at the sediment/water interface [6, 7]. Furthermore, recent modeling results [8, 9] show sediment volume inhomogeneities within about a wavelength of the water/sediment interface could also cause significant subcritical penetration.
Evidence is accumulating on the importance of seafloor roughness for subcritical acoustic penetration. However, seafloor characterization has been lacking in acoustic measurements, and none of the experimental analyses quantitatively relates the strength and properties of the penetrating field to detailed measurements of the surface roughness. Consequently, it still is not known if other mechanisms also contribute. The measurement program being developed under the DRI includes extensive environmental measurements which should allow a quantitative evaluation of the contributions of all three mechanisms mentioned above.
While acoustic penetration into sediments at subcritical grazing angles is an important issue, the DRI program will also address the broader range of technical issues associated with scattering from sediments, propagation and attenuation within sediments, and scattering from buried objects. In the context of these issues, the effects of the temporal evolution of the seafloor are also of interest.
Interface roughness and volume inhomogeneities are known to be important scattering mechanisms for sediments. Questions remain about the relative size of the contributions of these two mechanisms, the spatial variability of such scattering, and the temporal dependence due to biological reworking of the sediment. For volume scattering, discrete scatterers, such as shells, may also be important. How such discrete scattering should be treated in the context of stochastic modeling remains an issue.
The Biot model for the sediment potentially plays an important role in modeling high-frequency sediment acoustics. First, it leads to the possibility of the slow wave mechanism for penetration. Second, even if further experimentation shows that the slow wave is not a significant contributor to penetration, the possibility exists that the Biot sediment model and the fluid sediment model will produce noticeably different predictions for the strength of the penetrating field for the normal (or fast) wave due to scattering from interface roughness or volume inhomogeneities. Further efforts to incorporate the Biot sediment model into simulations should show if Biot/fluid model differences are important in this context. The model used for the sediment (Biot or fluid) may also be an important issue in predicting backscattering or general bistatic scattering back into the water. Thus, it is important to understand when the simpler fluid model is adequate and when a Biot sediment model is necessary for accurate high-frequency modeling.
Some aspects of sediment acoustics can only be modeled (aside from empirically) by going beyond the relatively simple description contained in the fluid model: these include attenuation versus frequency and the corresponding velocity dispersion. It would be of great benefit to test predictions of the Biot model (and other models) with acoustic measurements coupled with a broad range of sediment characterization, so that the sediment acoustic models can be highly constrained. One potential complication of such a comparison is that in the Biot model, volume inhomogeneities might lead to coupling into rapidly attenuating slow waves which may appear simply as increased fast wave attenuation. Thus, knowledge of volume inhomogeneities may be important in this context.
Essentially all models of sediment acoustics treat the granular sediment as an effective medium. At very high frequencies, the onset of scattering from sand grains will become important, and a continuum model for the sediment begins to break down. The frequency at which this breakdown begins to be noticeable, however, has not been established.
At grazing angles well above the critical angle, a substantial fraction of the sound incident on the seafloor will penetrate the water/sediment interface. The ability to use such fields to image buried objects, however, will be limited by forward scattering from the rough interface and from volume inhomogeneities within the sediment. Such scattering will degrade the spatial coherence of the acoustic field as well as distort and elongate the waveform of the propagating pulse. A need exists to better understand the magnitude of these effects and their frequency dependence. In particular, the ability to model these effects reliably from knowledge of sediment structure needs to be developed.
Scattering from buried objects is a complicated topic, since scattering from both the object and the overlying sediment needs to be predicted accurately in order to determine the expected signal-to-noise ratio. Coupling of sound energy into and back out of the sediment plays a crucial role in determining the strength of the return from the object. For incident fields below the critical angle, this coupling can be quite weak, preventing the object from being observed (e.g., see Ref. [10], where the coupling was based on relatively low levels of random roughness). This makes accurate modeling of such acoustic coupling all the more important. It also raises the question of how much this coupling might be increased if in addition to multiscale random roughness, a well-defined ripple component were present. This is now an important issue, since the tentative experiment site (to be discussed shortly) presently contains significant ripples remaining from the 1998 hurricane season.
Effects of sediment relaxation after object burial are also important. In some geometries scattering from disturbed sediment may betray the presence of an artificially buried object, even if the object itself cannot be observed. Thus, the temporal change in acoustic scattering as disturbed sediment relaxes in both space and time is of considerable interest.
In order to address the many acoustic issues just summarized at a basic research level, very extensive and very detailed environmental characterization is necessary. Only an outline of the characterization needs to be given here. Further details are contained in the investigator reports.
The basic concept is that the sediment roughness and internal structures must be known on sub-wavelength scales (i.e., to about 1 cm), and the temporal dependence needs to be monitored. In addition, the sediment properties required to define the parameters used in the Biot sediment model need to be determined.
To reduce the complexity of environmental characterization that could be encountered in complex shallow-water sites, the goal has been to select an experimental site that is relatively benign. An ideal site would have a homogeneous, silicate sand sediment with no layering, mud inclusions, or bubbles in the top meter of the seabed and with a minimum of biological activity. The sound speed in the sediment should also be high enough to produce a critical angle in the 20°–30° range. Also, the homogeneity of the seafloor should extend laterally over a region about 0.5 km by 0.5 km so that measurements taken at nearby locations can be legitimately compared. Thus, the site survey used as part of the site selection process must include examination of at least these properties of the sediment.
Within the sediment, vertical profiles in all basic properties are needed: compressional and shear speeds, compressional and shear attenuations, density, porosity, permeability, and grain size. Direct measurements of the grain bulk modulus and water/sediment reflection coefficients are useful for further constraining Biot parameter values. This needs to be supplemented with microscopic analysis of grain and pore structure to round out the determination of Biot model parameters. This microscopic analysis would also provide input to permeability models to supplement direct measurements of this difficult to measure quantity.
Measurements of the seafloor roughness spectrum will be required for use in data-model comparisons. The ability to monitor both long-term and short-term changes in roughness is also important. Similarly, measurements of volume inhomogeneities (sound speed and density variability) are needed on centimeter scales. The effects of the biological environment on changing these stochastic properties of the sediment need to be understood and monitored.
The first experiment under the DRI is scheduled for October-November 1999 near Panama City, Florida. A tentative experimental site has been selected about 2 miles offshore in water about 18 m deep. A relatively shallow depth is necessary since divers will be working on or near the bottom throughout the experiment. The exact site has yet to be determined, but the tentative site is near 30o 7.2'N and 85o 47.5'W. The experimental activity at the chosen site will extend from 4 October 1999 to about 10 November 1999. Two ships, the R/V Seward Johnson and the R/V Pelican, will be moored at the experiment site during SAX99.
Two factors led to locating the experiment near Panama City. First, it was felt that this region contained the sandy sediment desired. Second, a location near the Coastal Systems Station (CSS) would enable 6.2 programs at CSS to participate in SAX99 and benefit from the extensive environmental characterization. Two CSS programs will participate in SAX99, as described in the investigator summaries.
The final site selection will be based on multiple inputs: the results of a set of vibracores taken near the tentative site in mid-March, a preliminary site survey in late April, a more extensive site survey in July, and a final inspection at the beginning of SAX99. (See the investigator reports by Fleischer and Sawyer and by Richardson, Briggs, et al. for further details on these site surveys.)
If an intense tropical storm or hurricane should move through the region of the tentative site this summer, heavy rains and associated runoff could deposit patches of mud and silt on top of the sand, making the site less suitable for the proposed measurements. The first choice for a backup site is on the near-shore continental shelf between Panama City and Destin, Florida, about 100 km to the northwest.
The acoustic and environmental measurements are described in the investigator reports that follow.
The ONR program managers for the DRI are J. Simmen [lead], Code 321OA (Ocean Acoustics); J. Kravitz, Code 322GG (Marine Geology & Geophysics); and J. Eckman and R. Tipper, Code 322BC (Biological/Chemical Oceanography). The coordinating scientist is Eric Thorsos, APL-UW, (206) 543-1369, -6785 (fax), eit@apl.washington.edu.
1. N. P. Chotiros, "Biot model of sound propagation in water-saturated sand,'' J. Acoust. Soc. Am. 97, 199-214, 1995.
2. J. L. Lopes, "Observations of anomalous acoustic penetration into sediment at shallow grazing angles,'' J. Acoust. Soc. Am. 99, 2473-2474, 1996.
3. H. J. Simpson and Brian H. Houston, "Analysis of laboratory measurements of sound propagating into an unconsolidated water-saturated porous media,'' J. Acoust. Soc. Am. 103, 3095-3096, 1998; Proceedings of the 16th International Congress on Acoustics and the 135th Meeting Acoustical Society of America, Vol. IV, 3033-3034.
4. A. Maguer, E. Bovio, W. L. Fox, E. Pouliquen, and H. Schmidt, "Mechanisms for subcritical penetration into a sandy bottom: Experiment and modeling results,'' SR 287, SACLANT Undersea Research Centre, La Spezia, Italy, 1998 (submitted to J. Acoust. So c. Am.).
5. A. N. Ivakin and D. R. Jackson, "Effects of shear elasticity on sea bed scattering: Numerical examples,'' J. Acoust. Soc. Am. 103, 346-354, 1998.
6. E. I. Thorsos, D. R. Jackson, J. E. Moe, and K. L. Williams, "Modeling of subcritical penetration into sediments due to interface roughness,'' Proceedings of the High Frequency Acoustics in Shallow Water Conference, Lerici, Italy, 563-569, July 1997
7. E. I. Thorsos, D. R. Jackson, and K. L. Williams, "Modeling of subcritical penetration into sediments due to interface roughness,'' submitted to J. Acoust. Soc. Am.
8. K. L. Williams and D. R. Jackson, "A model for bistatic scattering into ocean sediments for frequencies from 10-100 kHz,'' APL-UW TR 9505, Applied Physics Laboratory, University of Washington, February, 1995.
9. D. Tang (APL-UW), presented at the DRI Acoustics Workshop, November 1998, Applied Physics Laboratory, University of Washington, Seattle, WA.
10. R. Lim, K. L. Williams, and E. I. Thorsos, "Acoustic scattering by a 3-dimensional elastic object near a rough surface,'' submitted to J. Acoust. Soc. Am.
| Top of Page | Return to Contents Page |