High-frequency acoustic remote sensing of seafloor characteristics

This dissertation describes the development of a sediment classification method which compares bottom returns measured by a calibrated, moderate beam width (10°–20°), vertically oriented (0°–15°) monostatic sonar, with an echo envelope model based on high frequency (10–100 kHz) incoherent backscatter theory and sediment properties such as: mean grain size, strength and exponent of an interface roughness spectrum exhibiting power law statistics, and volume scattering coefficient.; An average echo envelope matching procedure is described where: first, sediment type (sand or fines) is established by iterating on the reflection coefficient to match the peak echo amplitude and to establish a general fit with generic values for the remaining geoacoustic parameters; then, a three parameter global optimization is performed using a combination of simulated annealing and downhill simplex searches over the allowable range of interface roughness spectral strength, sediment volume scattering coefficient, and a constrained range of reflection and bottom absorption coefficients correlated with mean grain size. Bottom echo envelopes collected at 33 kHz and 93 kHz, over substrates ranging from sand to clay, yield solutions for grain size and geoacoustic properties which are consistent with ground truth measurements.; Analyses of the estimated geoacoustic parameters for different combinations of sediment type, frequency, and transducer orientation, reveal that moderate frequencies (33 kHz) and orientations normal with the interface are best suited for this application. The ability to distinguish sands from fine-grain sediments is demonstrated based on acoustic estimation of mean grain size alone. The creation of feature vectors from estimates of mean grain size and interface roughness spectral strength shows promise for intraclass separation of silt and clay.; Estimates of spectral strength for sand substrates are relatively immune to measured echo variability, whereas estimates of mean grain size are moderately affected. The opposite trend is observed for fine-grain substrates, with spectral strength estimates varying significantly. Correlation of mean grain size and spectral strength is observed in the estimated solutions, and is especially large for sand substrates. This trend is consistent with what is observed in nature, where coarser sediments exhibit more energy in the roughness spectrum than fine-grain sediments.