Abstract
Offshore piling for wind farm construction has attracted significant attention due to the high noise emission levels associated with hydraulic hammering. While the impact of underwater noise on marine biology is well-documented, the propagation of airborne piling noise over the sea has remained understudied. This research utilizes detailed numerical calculations, specifically the Green's Function Parabolic Equation (GFPE) method, to estimate noise levels up to 10 km from the source. By combining measured noise emission levels from pinpile installations with idealized vertical sound speed profiles and effective impedances for non-flat sea surfaces (Sea States 2, 3, and 4), the study evaluates the competing effects of atmospheric refraction and surface scattering. The results indicate that scattering by a rough sea surface significantly exceeds the refractive effects of downwind propagation. Consequently, the presence of wind—even when blowing toward a receiver—actually increases the attenuation of piling sound, identifying a perfectly flat sea surface as the worst-case scenario for noise propagation.
Introduction
The global transition toward renewable energy has accelerated the construction of offshore wind farms. A critical phase of this construction is "piling," where massive hydraulic hammers drive steel foundations into the seabed. While environmental impact assessments frequently address the acoustic threat to marine mammals, the airborne component of this noise is equally relevant for coastal communities and offshore personnel.
Predicting airborne noise over the sea is complex due to the unique physics of the marine boundary layer. Sound waves are subject to vertical wind and temperature gradients that cause refraction, as well as the stochastic nature of the sea surface which causes scattering. This study aims to quantify these effects and determine the primary drivers of long-distance noise propagation in offshore environments.
Methodology
The researchers implemented a multi-scale numerical framework to model the acoustic environment accurately:
1. Source Characterization
Initial acoustic data were collected from a floating vessel located 280 meters from a piling operation at Thornton Bank, off the Belgian coast. These measurements provided the baseline source levels for the numerical models.
2. Micro-scale Modeling (FDTD)
To account for the complex geometry of sea waves, the 2D Finite-Difference Time-Domain (FDTD) method was employed. This allowed the team to simulate how sound pulses interact with specific wave shapes corresponding to Sea States 2, 3, and 4 (based on the Pierson-Moskowitz spectrum). From these simulations, "effective impedances" were derived, allowing the rough sea surface to be represented as a simplified flat surface with specific acoustic properties in subsequent models.
3. Long-distance Propagation (GFPE)
For distances up to 10 km, the Green’s Function Parabolic Equation (GFPE) method was used. The GFPE is highly efficient for long-range propagation and can integrate atmospheric refraction effects caused by wind and temperature gradients.
4. Atmospheric Profiling
The study utilized Monin-Obukhov similarity theory (MOST) to generate realistic vertical wind speed and temperature profiles. These profiles are essential for calculating the downward refraction typically found in downwind conditions.
Key Insights and Results
The Dominance of Scattering
The study identifies a fundamental "tug-of-war" between two physical phenomena: downwind refraction and rough sea scattering. Downwind refraction typically bends sound waves back toward the surface, which would normally increase noise levels at a distance. However, the research found that the scattering effect of even a mildly rough sea (Sea States 2–4) is significantly more powerful than the refractive effect.
The "Worst-Case" Scenario
Counter-intuitively, the study concludes that a perfectly flat, mirror-like sea surface represents the worst-case scenario for noise propagation. Without waves to scatter the acoustic energy, the rigid water surface acts as a near-perfect reflector, allowing noise to travel much further with minimal attenuation. Even when wind is blowing toward a receiver, the waves generated by that wind provide a net benefit by reducing noise levels by 5 to 15 dBA compared to a hypothetical calm sea.
Acoustic Metrics and Audibility
- Source Power: The total A-weighted peak source level during piling was estimated at approximately 145 dBA.
- Frequency Filtering: The atmosphere and sea surface effectively act as a low-pass filter. Over a 10 km distance, high-frequency components are almost entirely absorbed or scattered, leaving only low-frequency "thuds" audible to the receiver.
- Atmospheric Stability: The choice of atmospheric stability (neutral vs. unstable) can result in a variance of up to 25 dBA in predicted noise levels at the 10 km mark, highlighting the importance of accurate meteorological data in impact assessments.
Discussion and Limitations
While this study provides a robust framework for offshore noise prediction, several limitations remain. The model assumes a point source for the piling impact, which may simplify the complex vibrations of a 48-meter steel pile. Additionally, the "effective impedance" approach primarily accounts for the coherent part of the sound field, potentially neglecting energy transferred to an incoherent field.
Furthermore, the study focused on Sea States 2–4, as piling operations are generally suspended during higher sea states for safety reasons. Future research could benefit from incorporating measured vertical profiles for humidity and temperature to further refine the accuracy of the GFPE calculations at varying altitudes.
Conclusion
This research demonstrates that environmental factors—specifically sea surface roughness—play a decisive role in the attenuation of airborne piling noise. The finding that wind-induced waves mitigate noise propagation, even in downwind conditions, provides a new perspective for environmental impact assessments in the offshore wind sector. For coastal noise management, the calmest sea states should be treated as the primary concern for long-distance sound transmission.
Summary updated