JASCO Applied Sciences conducted a sound source characterisation study of the Hywind Scotland floating offshore wind farm, involving in situ acoustic recording over three winter months (October 2020 to January 2021) at the Hywind site and at a Control site 14 km away. JASCO field engineers conducted the deployment and retrieval of the recording instrumentation from the Hywind support vessel, MCS SWATH 1 (out of Peterhead) under the oversight of the Hywind Operations Room.
The recording instrument at the Hywind site employed a four-hydrophone tetrahedral array to provide bearing discrimination between sounds from different directions. The location of the recorder was selected to enable the acoustic isolation of one Hywind structure such that a noise signature from that unit could be extracted without contamination from the four other floating turbines in the farm. One further purpose of the directional array was to allow for an analysis of the location of transient noises from the Hywind mooring system. In a previous study of the Hywind prototype off Norway in 2011, occasional high amplitude mooring transients were detected but their precise origin was unknown. The recording program was completed successfully with recovery of a continuously recorded, 24-bit, acoustic data set (10 Hz to 32 kHz) from each site, a total of approximately 6.6 TB data.
Analysis of the recorded data was conducted to assess the spectral content of the sound signature from the Hywind structures. Continuous tonal noise, associated with rotating rotor and generator components below 500 Hz, was clearly evident and showed corelation with wind speed. Temporal variability in similar frequency tones was suggestive of different concurrent signature for individual Hywind turbines arriving simultaneously. The other key feature of the overall Hywind noise was the presence of frequent broadband transient sounds with a median duration of 1.5 s. These transients were audibly associated with strain and friction in the mooring system and showed a strong positive corelation in occurrence with wave height. Directional analysis of transient noise from three of the HYWIDND turbines indicates that the mooring noise is predominantly being generated in mooring components close to the floating spar and not from components farther down each mooring cable.
A quantitative analysis of the impulsiveness of the soundscape at Hywind was undertaken using an impulse detector as well as studying the distribution of the per-minute kurtosis. The SEL of each detected impulse was summed, which showed that the impulsive SEL was generally 6 dB below the daily total SEL. The mean duration of the impulses was on the order of 1.5 s, longer than the 1.0 s typically used to identify impulses for the purposes of assessing the effects of sound on hearing. The soundscape at Hywind had a greater kurtosis than at the Control site; however, the kurtosis was not high enough to be considered impulsive. Based on these three measures, it is recommended that the non-impulsive temporary threshold shift (TTS) sound exposure level (SEL) thresholds be applied to the wind farm sounds.
The total noise levels (tonal and transient) from HS1were extracted and back propagated to derive decidecade band source levels for a single Hywind Scotland system at five winds speeds between 5 and 25 kn. Unexpectedly, the total noise level from HS1 was higher in 5 kn of wind than at 10 kn of wind. The resulting median broadband source levels ranged from 162.5 to 167.2 dB re 1 µPa²m² with the maximum 95th percentile at 25 kn of 172 dB re 1 µPa²m². The HS1 source levels were used to model a basic noise footprint for the entire five turbine windfarm. The modelling shows that the distance to the averaged background SPL level (110 dB re1µPa) from the centre of the OWF (i.e., where the radiating noise decays to approximately the broadband ambient level) at the quietest state in 10 kn of wind, was approximately 4 km and in 25 kn of wind it was 13 km. A high-frequency cetacean (porpoise) would need to remain within 50 m of a turbine for 24 h before there would be a risk of temporary hearing threshold shift (15 kn wind).