Abstract
• Offshore wind farm construction requires regulatory assessment of the numbers of protected marine mammals that may experience auditory injury or disturbance from impulsive piling noise. However, there is high uncertainty within available assessment frameworks, and limited opportunities to validate outputs.
• We present data on underwater noise levels and porpoise behavioural responses during construction at Ocean Winds’ Moray West wind farm in 2023. Specifically, we compare measurements of 1) underwater noise and 2) harbour porpoise behavioural responses with predictions made during regulatory assessments.
• Broadband acoustic recorders were used to measure noise levels at 750 m and 2,000 m during impact piling of thirteen 9.5 -10 m monopiles, with additional opportunist recordings at up to 15 km. Harbour porpoise responses were assessed between 1 km and 33.4 km from each monopile using an array of 60 echolocation detectors (CPODs) as reported in Benhemma-Le Gall et al. (2024). Analysis of changes in echolocation click detections provided a proxy for behavioural responses during the installation of seven monopiles, where detections during 24 hours after piling were compared with matched periods two days before piling started.
• Measured maximum received noise levels were typically within 1 dB (range -2.4 dB to + 3 dB) of values predicted from acoustic modelling at two focal sites used in earlier regulatory assessments. INSPIRE v5.2, an acoustic modelling tool used in many current UK assessments, was used to conduct blind retrospective modelling of spatial variation in received noise levels from all piling locations where field measurements were available. There was a strong correlation between measured and INSPIRE modelled received noise levels, but with a tendency to over-predict received levels within ~7 km of source and under-predict at greater distances. Overall, 68% of the measurements were lower than model predictions. Where measured values exceeded INSPIRE predictions, all were within 3 dB.
• Using recommended dose-response curves from Graham et al. (2017), modelled predictions of porpoise responses within the areas covered by our PAM array suggested that there would be a > 75% chance of disturbance throughout its 35 km range. In contrast, observed changes in echolocation detections indicated that the probability of disturbance was < 50% at distances beyond 5 km from piling. Monitoring data indicated that <100 porpoises would be disturbed by each piling event, compared to worst-case predictions of > 4,500 in regulatory assessments.
• These results provide confidence in current approaches to acoustic modelling but suggest that existing approaches to estimating disturbance to harbour porpoise from impact piling in UK waters are likely to be unrealistically conservative. This could constrain efforts both to assess impacts on protected populations and to identify mitigation measures that are proportionate to realistic levels of risk.