Abstract
Harbour porpoises (Phocoena phocoena) are disturbed by underwater noise from impact piling during construction of offshore wind farms, with multiple studies reporting displacement of porpoise up to 10-20 km from piling. However, evidence of the magnitude of porpoise responses to unabated large monopile installation in the UK Southern North Sea is currently lacking. A robust understanding of displacement effects is essential to assessing and managing cumulative effects of wind farm construction and informing marine noise policy development. Here, we build on existing studies to assess porpoise responses to unabated piling of monopile foundations at the Dogger Bank C (DBC) offshore wind farm, adjacent to the Southern North Sea Special Area of Conservation (SAC) for harbour porpoise. Specifically, this study aimed to use a similar methodology to studies undertaken in the Moray Firth, Scotland, in order to compare the observed magnitude of porpoise responses to unabated piling across the two locations. An array of 27 passive acoustic monitoring (PAM) moorings was deployed across an area up to 55 km from piling locations. All moorings were equipped with an echolocation click logger (CPOD), with eight moorings also having a broadband acoustic recorder for sampling raw acoustic data, including piling noise. The monitoring period was 5.5 months long and overlapped with foundation installation activities at three adjacent offshore wind farms. Porpoise acoustic presence was extracted from CPODs, whereas broadband acoustic data were used to estimate audibility of pile driving pulses over distance to assess the suitability of the DBC PAM array to fully capture the effects of impact piling on porpoises. The probability of porpoises responding to impact piling was modelled as a function of distance from the piling location and other covariates. As far as possible, the analytical approach aligned with prior studies at wind farms in the Moray Firth, to assist in comparability with these studies. Responses to piling were assessed by calculating the proportional changes in porpoise acoustic presence during a 12-hour response period post-piling and a temporally-matched baseline period 48 hours before piling. Six piling ‘batches’, each of up to three monopiles, were deemed suitable for modelling based on temporal separation from other piling activities at DBC and spatial proximity of the piles within a batch. The presence or absence of any porpoise response to piling was fitted with a binomial distribution and a probit link function using Generalized Linear Models (GLM). Models that only had one distance covariate indicate that distance relative to the last pile in a batch provided the best fit. Adding further covariates one at a time revealed that the combination of distance to the last pile, batch piling duration, and month (as a factor) yielded the best performance. Two models were used to generate predictions of probability of response to impact piling over distance and estimate effective deterrence ranges (EDRs). The probability of response to piling decayed steeply within the first ~ 10 km from piling, then levelled out to approach an asymptote at the maximum distances recorded. The model using only distance to the last pile produced an EDR of 14.4 km whereas the best model yielded variable predictions depending on month and batch durations, with the month of May suggesting a greater probability of response compared to August. Porpoise acoustic presence exhibited a gradual decrease over the study period; therefore, it is challenging to determine whether porpoises responded less to piling as the campaign progressed or if the analytical approach limits the identification of responses during periods with low detection rates. Although this trend of decreasing detections through the summer coincided with the progression of 10 piling at DBC, it appears most likely to be attributable to natural variability in porpoise occurrence and is supported by past surveys undertaken in the area. This study provides key data specific to large, unabated monopiles and the UK-portion of the southern North Sea to inform noise management within harbour porpoise SACs, including the estimation of EDRs. The estimated EDR of 14.4 km at DBC is within the reported range of EDRs estimated for unabated piling at other offshore wind farms in the North Sea, and it is close to the lower bound of the 15-20 km range recommended for unabated piling in a recent review of data and less than the 20 km EDR currently recommended by the Joint Nature Conservation Committee (JNCC).