Peter H. Dahl

Normalized vertical reactive intensity (Qzn⋆), computed from coherent combinations of complex pressure and vertical velocity measured from a vector acoustic sensor, preserves the local interference dynamics of dispersed modes at the receiver independent of source location or propagation distance. Measurements of impulsive sources deployed during the 2022 Seabed Characterization Experiment recorded by vector acoustic sensors at three locations are used for sediment classification and geoacoustic inversion. In addition to classifying thick vs thin mud, inversion results from the thickest portion of the New England Mud Patch suggest that in May 2022 the mud sound speed at the top of the sediment is faster at 1472 [1464, 1480] m s−1 (brackets indicate 95% confidence bounds) and the gradient is smaller at 3.46 [1.03, 5.59] s−1 than inversion results from the same location in March 2017, consistent with the observed changes in oceanography. Inversions of data collected on and off the mud patch are also consistent with previously observed east–west increase in surficial mud sound speed.

Observations of tonal emissions over frequencies ∼ 9.5–36 Hz from a merchant vessel along a 31.5-km track are used to study seabed properties on the New England Mud Patch, an area distinguished by a low sound speed mud layer of thickness ∼ 10 m commencing at the water-sediment interface. The data are forward-modeled using adiabatic normal modes based on a range-varying geoacoustic model for the upper sediments, reproducing key effects such as changing mode interference patterns. For the deep sediments, a proposed linear sound speed gradient governs an increasing compressional speed with depth. Bayesian inversion yields frequency-dependent estimates of the compressional wave attenuation within the deep sediments using data from the lower frequencies for which it is possible to disambiguate the effect of attenuation from the upper mud layer. At ∼16 Hz, the highest frequency meeting this condition, this estimate is 0.17 ±0.03 dB/λ, and these inversion results are used to infer corresponding values for the higher frequencies. A compressional wave attenuation of 0.05 dB/λ in the mud layer is applied across all frequencies. Elastic effects are included in the geoacoustic candidate models used in the inversion; removing them increases compressional wave attenuation 10%–40% depending on frequency.

This article presents a spatial environmental inversion scheme using broadband impulse signals with deep learning (DL) to model a single spatially-varying sediment layer over a fixed basement. The method is applied to data from the Seabed Characterization Experiment 2022 (SBCEX22) in the New England Mud-Patch (NEMP). Signal Underwater Sound (SUS) explosive charges generated impulsive signals recorded by a distributed array of bottom-moored hydrophones. The inversion scheme is first validated on a range-dependent synthetic test set simulating SBCEX22 conditions, then applied to experimental data to predict the lateral spatial structure of sediment sound speed and its ratio with the interfacial water sound speed. Traditional geoacoustic inversion requires significant computational resources. Here, a neural network enables rapid single-signal inversion, allowing the processing of 1836 signals along 722 tracks. The method is applied to both synthetic and experimental data. Results from experimental data suggest an increase in both absolute compressional sound speed and sound speed ratio from southwest to northeast in the NEMP, consistent with published coring surveys and geoacoustic inversion results. This approach demonstrates the potential of DL for efficient spatial geoacoustic inversion in shallow water environments.

This study expands previous work examining the effects on fishes from exposure to a single 4.5 kg underwater explosive detonation. Experiments were done in the field, with fish in cages at different distances from the source. Although our earlier work reported high acoustic dosage levels (e.g. based on peak pressure) correlating with severe injuries, dosage levels that result in moderate, or mild injuries were not clearly established. Thus, in this study, caged Pacific mackerel (Scomber japonicus) were placed at targeted ranges of 150–800 m from the source. All procedures were the same as in the earlier study except that animals were left at depth for ∼3 hours post-exposure to determine immediate effects on survival. Fish were then retrieved and assessed for physical damage. The only statistically significant tissue injuries were swim bladder bruising and in a reduction in inner ear sensory hair cell density that lessened with distance from the source. Still, results must be taken with caution since they may vary with different source levels, water depths, location of the fish in the water column, and by species.

Underwater acoustic communication is heavily influenced by intersymbol interference caused by the delay spread of multipaths. In this article, communication sequences transmitted from a drifting source were received by a fixed acoustic vector receiver system consisting of an accelerometer-based vector sensor and a pressure sensor, which can measure the three-directional components of vector quantity and pressure at a point. The underwater acoustic communication experiment was conducted in water approximately 30 m deep off the south coast of Geoje Island, South Korea, in May 2017 during the Korea Reverberation Experiment. Acceleration signals received by the vector sensor were converted to pressure-equivalent particle velocities, which were then used as input for a four-channel communication system together with acoustic pressure. These four channels have multipaths with different amplitudes but the same delay times, providing directional diversity that differs from the spatial diversity provided by hydrophone arrays. To improve the communication performance obtained from directional diversity, the Multichannel Combined Bidirectional Block-based Time Reversal Technique was used, which combines bidirectional equalization with time-reversal diversity and block-based time reversal that was robust against time-varying channels. Communication performance was compared with the outcomes produced by several other time reversal techniques. The results show that the Multichannel Combined Bidirectional Block-based Time Reversal Technique using a vector sensor achieved superior performance under the environmental conditions considered in this article.

Vector acoustic measurements of underwater noise made continually over a 24-h period within a small bay on the west side of Puget Sound, Washington State are discussed. The bay contains the Washington State ferry landing at Kingston with measurement site approximately 1 km distant from the landing, at water depth 42 m. Periodic, nearby ferry traffic, and episodic large-vessel shipping activity punctuate the observations which will be compared with those made in deeper waters. Potential and kinetic energies in key shipping decidecimal bands are equal within calibration uncertainty; a practical result towards the inference of kinematic properties from pressure-only measurements within inland waters complementing similar results obtained in off shore waters. An important exception exists when the noise field is dominated by a vessel passing directly over the sensor, as was demonstrated by the deployment vessel moving at speed 7.5 knots. Observed near the seafloor, broadband noise emissions from a vessel passing directly above exhibit frequency bands where potential acoustic energy is greater than kinetic energy (and vice versa in neighboring frequency bands). This phenomenon is also examined with a model based on multiple interfering reflections from the seabed, sea surface, and buried sediment layers.

For acoustic measurements made within the water column, it is well understood that observability of the effects of seabed shear rigidity varies widely depending, e.g., on source type, frequency range and measurement geometry. Measurements of underwater noise from a passing ship made with a vector sensor are discussed in this context. Data are expressed in a bounded, non-dimensional form known as circularity that provides a signal independent of the ship noise-source spectrum. The system was deployed 1.45 m above the seabed in waters of depth 77 m on the New England Mud Patch as part of the May 2022 Sediment Characterization Experiment. Broadband noise measurements (15i–500 Hz) of the ship are made during which the ship range varies from 10 to 1.5 km at closest point of approach. Observations are compared to an interpretative model based on the normal modes and a geoacoustic model consistent with studies conducted at this location; the defining characteristic being a low-speed mud layer constituting the upper ∼10 m of the seabed that transitions with increasing sand content. Emphasis in this presentation is on parameterization of the deeper seabed structure involving compressional sound speed gradients and sediment shear rigidity.

Observed near the seabed, broadband noise emissions from a source or vessel passing directly above the sensor exhibit frequency bands where potential acoustic energy is greater than kinetic energy while the opposite occurs in neighboring frequency bands. The condition where the dynamic and kinematic energy forms differ in this manner is characteristic to interference involving steep angles or near-normal incidence reflection from the seafloor. Recent experiments conducted on the New England Mud Patch (NEMP) and the neighboring (towards offshore) shelf break waters demonstrate the sensitivity to seabed structure, which evolves from the soft mud layers at the NEMP into the harder sediments that are exposed on the continental slope. Results are expressed as a ratio of kinetic to potential energy as a function of frequency, with resulting data series yielding information on seabed properties. Variance components as function of frequency are also isolated based on a singular value decomposition of a matrix constructed from the data series (known as singular spectrum analysis). The components are associated with different sub- bottom layering features. A model for kinetic and potential energy combined with candidate geoacoustic models is used to interpret both the data series and components.