Estimating Sound Exposure Levels Due to a Broadband Source over Large Areas of Shallow Sea

D. S. Manul’chev, Andrey G. Tyshchenko, Mikhail Fershalov, Pavel S. PetrovJournal of Marine Science and Engineering 2022 · DOI

TL;DR

This paper presents a novel method for accurately estimating sound exposure levels (SEL) over large shallow sea areas, crucial for environmental monitoring. By reconstructing a broadband source's effective spectrum from a single hydrophone measurement and employing advanced mode parabolic equations, the authors demonstrate a computationally efficient way to map acoustic noise distribution.

What Problem Does This Paper Address?

Monitoring anthropogenic noise pollution in marine environments, especially over vast shallow sea areas, is a significant challenge. Traditional methods relying on dense sensor networks are impractical and costly. Furthermore, accurately modeling the propagation of broadband acoustic signals, like those from seismic surveys, is computationally intensive. Existing 3D parabolic equation (PE) models, while accurate, are too slow for real-time monitoring or large-scale simulations. This paper addresses the need for a computationally efficient yet accurate method to estimate sound exposure levels (SEL) over large areas, enabling better assessment of noise impacts on marine fauna and timely mitigation efforts.

Methods

The authors propose a three-step procedure:

  1. Effective Source Spectrum Reconstruction: They develop a technique to reconstruct the effective spectrum of a broadband source (like an airgun array) from a single reference hydrophone measurement. This bypasses the need to know the source's exact characteristics beforehand.
  2. Multi-frequency Acoustic Field Simulation: Using the reconstructed source spectrum, they simulate the acoustic field across the area of interest. This is achieved through Mode Parabolic Equations (MPE), a class of models known for their computational efficiency compared to full 3D PE models. Specifically, they compare two MPE approaches:
  3. Narrow-Angle Mode Parabolic Equation with Mode Coupling (MPE-MI): This method accounts for interactions between different acoustic modes but has limitations in handling wide horizontal apertures.
  4. Adiabatic Wide-Angle Mode Parabolic Equation (PDMPE): This approach can handle wider horizontal angles, crucial for scenarios with significant horizontal refraction, but does not explicitly model mode coupling.
  5. SEL Distribution Computation: The Sound Exposure Level (SEL) distribution is then calculated by integrating the simulated acoustic field over the source's frequency range for a grid of points across the area.

Datasets

The study utilizes experimental data collected during a seismic survey in the shallow part of the shelf Sakhalin Island, Russia. Two autonomous underwater acoustic recorders were deployed to capture seismic pulses. The data from one recorder (P1) was used to reconstruct the source spectrum, while data from the second recorder (P2) was used for validating the simulation results by comparing the predicted pulse waveforms with the observed ones. The paper also refers to existing geoacoustic models and measurements for the Sakhalin shelf area.

Key Insights and Ideas

  • Single-hydrophone source reconstruction is feasible: The ability to accurately estimate a broadband source's spectrum from a single measurement significantly simplifies the setup for large-area acoustic monitoring.
  • MPE methods offer a good balance: Mode Parabolic Equations provide a computationally efficient alternative to full 3D PE models, making large-scale simulations practical.
  • Wide-angle modeling is crucial for shallow seas: The study highlights that for shallow-water environments with sloping bottoms, horizontal refraction is a significant factor. Wide-angle models (like PDMPE) are essential for accurately capturing these effects, outperforming narrow-angle models even when mode coupling is considered.
  • SEL is a practical metric for impact assessment: The paper demonstrates how SEL can be effectively computed and mapped, providing a valuable tool for assessing the potential impact of noise on marine life.

Key Results

  • The proposed method successfully reproduces the pulse waveform at a validation point (P2) with reasonable accuracy in both time and frequency domains.
  • The wide-angle PDMPE model showed significantly more accurate results than the narrow-angle MPE-MI model, especially in scenarios with horizontal refraction. The difference in SEL predictions could reach up to 8 dB in certain areas.
  • While the overall SEL was reproduced with sufficient accuracy for practical applications, discrepancies were observed in the lower frequency range (10–50 Hz) of the spectra, suggesting potential limitations in the geoacoustic model.

Limitations and Future Work

  • Geoacoustic Model Accuracy: The authors acknowledge that the accuracy of SEL distribution modeling is limited by the knowledge of the sea bottom's structure. Discrepancies in the lower frequency spectra suggest that a more detailed geoacoustic model is needed.
  • Future Work: The authors propose incorporating a geoacoustic inversion algorithm into their software. This would allow for the reconstruction of bottom layer structures based on acoustic measurement data, further improving SEL distribution modeling. They also note that while MPE-based techniques offer a good balance of accuracy and computational cost, further work is needed to fully integrate mode interaction into wide-angle models. The potential of Gaussian beams for 3D modeling is also mentioned as an alternative to consider.

Summary updated

Abstract

3D sound propagation modeling in the context of acoustic noise monitoring problems is considered. A technique of effective source spectrum reconstruction from a reference single-hydrophone measurement is discussed, and the procedure of simulation of sound exposure level (SEL) distribution over a large sea area is described. The proposed technique is also used for the modeling of pulse signal waveforms at other receiver locations, and results of a direct comparison with the pulses observed in the experimental data is presented.
BibTeX
@article{manulchev2022estimating,
  author = {D. S. Manul’chev and Andrey G. Tyshchenko and Mikhail Fershalov and Pavel S. Petrov},
  title = {{Estimating Sound Exposure Levels Due to a Broadband Source over Large Areas of Shallow Sea}},
  journal = {Journal of Marine Science and Engineering},
  year = {2022},
  doi = {10.3390/jmse10010082},
}
Figure