Conservatism in the Underwater Noise Modelling

  1. Continuing on from the information presented in paragraphs 454 to 461, a number of conservative assumptions were adopted in the underwater noise model. These measures of conservatism are summarised in this section and highlight that both PTS and TTS onset ranges predicted using the SELcum threshold are likely to lead to overestimates in the ranges and therefore should be interpreted with caution.
  2. The underwater noise modelling assumed that the maximum hammer energy would be reached and maintained at all locations, whereas this is unlikely to be the case based on examples from other offshore wind farms, e.g. Beatrice Offshore Wind Farm, where the mean actual hammer energy averages were considerably lower than the maximum assessed in the Environmental Statement and only six out of 86 asset locations reached maximum hammer energy (Beatrice Offshore Wind Farm Limited, 2018).
  3. Additionally, the piling procedure simulated in the model does not allow for short pauses in piling (e.g. for realignment) and therefore the modelled SELcum is likely to be an overestimate since, in reality, these pauses will reduce the noise exposure that animals experience whilst moving away.
  4. The underwater noise modelling assessment also assumed that animals swim directly away from the noise source at constant and conservative average speeds based on published values. Whilst this buffers the uncertainty with respect to the directionality of their movement, it may lead to overestimates of the potential range of effect as animals are likely to exceed these speeds. For example, Otani et al. (2000) reported horizontal speed for harbour porpoise can be significantly faster than vertical speed and cite a maximum speed of 4.3 m/s (compared to 1.5 m/s used in the underwater noise model).
  5. The underwater noise model accounts for the SELcum metric as an equal-energy rule, where exposures of equal-energy are assumed to produce the same noise-induced threshold shift regardless of how the energy is distributed over time. Since for intermittent noise (such as piling) the quiet periods between noise exposures will allow some recovery of hearing compared to continuous noise, the equal-energy rule is likely to overestimate the extent of impact. Additionally, modelling of concurrent piling assumed piling will exactly coincide and strike piles simultaneously, whereas in reality this is highly unlikely and could lead to overestimates in the injury and/or disturbance ranges.
  6. The impulsive noise is likely to undergo transition into non-impulsive noise at distance from the noise source due to a combination of factors (e.g. dispersion of the waveform, multiple reflections from sea surface and seafloor, and molecular absorption of high frequency energy). The empirical evidence suggest that such shifts in impulsivity could occur within 10 km from the noise source (Hastie et al., 2019). However, since the precise range at which this transition occurs is unknown, the underwater noise model adopted the impulsive thresholds at all ranges. This is likely to lead to an overly precautionary estimate of injury ranges at larger distances (tens of kilometres) from the noise source.

                        Sensitivities of marine mammals to underwater noise

                        Harbour porpoise
                        Injury
  1. Scientific understanding of the biological effects of threshold shifts is limited to the results of controlled exposure studies on small numbers of captive animals (reviewed in Finneran (2015)) where TTS are experimentally induced (given it is unethical to induce PTS in animals) and thresholds for PTS extrapolated using TTS growth rates. Kastelein et al. (2013) demonstrated that hearing impairment as a result of exposure to piling noise is likely to occur where the source frequencies overlap the range of peak sensitivity for the receptor species, rather than across the whole frequency hearing spectrum. The study demonstrated that for simulated piling noise (broadband spectrum), harbour porpoise hearing around 125 kHz (the key frequency for echolocation) was not affected. Rather, a measurable threshold shift in hearing was induced at frequencies of 4 kHz to 8 kHz, noting the magnitude of the hearing shift was relatively small (2.3 dB to 3.6 dB at 4 kHz to 8 kHz) due to the lower received SELs at these frequencies. This was due to most of the energy from the simulated piling occurring in lower frequencies (Kastelein et al., 2013). Kastelein et al. (2017) confirmed sensitivity declined sharply above 125 kHz in a following study.
  2. The duty cycle of fatiguing noises is also likely to affect the magnitude of a hearing shift, (e.g. hearing may recover to some extent during inter-pulse intervals (Kastelein et al., 2014)). Other studies reported that whilst a threshold shift can accumulate across multiple exposures, the resulting shift will be less than the shift from a single, continuous exposure with the same total SEL (Finneran, 2015).
  3. In order to reduce exposure to noise, cetaceans are able to undertake some self-mitigation measures (e.g. the animal can change the orientation of its head so that noise levels reaching the ears are reduced), or it can suppress hearing sensitivity by one or more neurophysiological auditory response control mechanisms in the middle ear, inner ear, and/or central nervous system. Kastelein et al. (2020) highlighted the lack of reproducibility of TTS in a harbour porpoise after it was exposed to repeated airgun noises, and suggested self-mitigation may lead to the discrepancies.
  4. It is important to highlight that extrapolating the results from captive bred studies to how animals may respond in the natural environment should be treated with caution as there are discrepancies between experimental and natural environmental conditions. In addition, the small number of test subjects does not account for intraspecific differences (i.e. differences between individuals) or interspecific differences (i.e. extrapolating to other species) in response. However, based on the latest scientific evidence, PTS is a permanent and irreversible hearing impairment. It is therefore anticipated that harbour porpoise is sensitive to this effect as the loss of hearing would affect key life functions (such as mating and maternal fitness, communication, foraging, predator detection) and could lead to a change in an animal’s health (chronic) or vital rates (acute) (Erbe et al., 2018). In addition to studies conducted in controlled environments, there is also evidence on noise-induced hearing loss, based on inner ear analysis in a free-ranging harbour porpoise (Morell et al., 2021). Considering the above, a potential consequence of a disruption in key life functions is that the health of impacted animals would deteriorate and potentially lead to reduced birth rate in females and mortality of individuals (Costa, 2012).
                        Behavioural disturbance
  1. As a small cetacean species, harbour porpoise is vulnerable to heat loss through radiation and conduction. They have a high metabolic requirement, with a need to forage frequently to lay down sufficient fat reserves for insulation. Kastelein et al. (1997) found in a study of six, non-lactating, harbour porpoise that they require between 4% and 9.5% of their body weight in fish per day. In the wild, porpoises forage almost continuously day and night to achieve their required calorific intake (Wisniewska et al., 2016), meaning they are vulnerable to starvation if foraging is interrupted.
  2. It is well documented that there is variance in behavioural responses to increased underwater noise and it is context specific. Factors such as the activity state of the receiving animal, the nature and novelty of the noise (i.e. previous exposure history), and the spatial relation between noise source and receiving animal are important in determining the likelihood of a behavioural response and therefore their sensitivity (Ellison et al., 2012). Empirical evidence from monitoring at offshore wind farms during construction suggests that pile driving is unlikely to lead to 100% avoidance of all individuals exposed, and that there will be a proportional decrease in avoidance at greater distances from the pile driving source (Brandt et al., 2011). Graham et al. (2019) demonstrated this dose-response at Horns Rev Offshore Wind Farm, where 100% avoidance occurred in harbour porpoises at up to 4.8 km from the piles, whilst at greater distances (10 km plus) the proportion of animals displaced reduced to < 50%). More recently Graham et al. (2019) studied responses of harbour porpoise to piling at the Beatrice Offshore Wind Farm, and suggested that harbour porpoise may adapt to increased noise disturbance over the course of the piling phase, thereby showing a degree of tolerance and behavioural adaptation. Graham et al. (2019) also demonstrated that the probability of occurrence of harbour porpoise (measured as porpoise positive minutes) increased exponentially moving further away from the noise source. Similarly, a study of seven offshore wind farms constructed in the German Bight demonstrated that detections of harbour porpoise declined several hours before the start of pling within the vicinity (up to 2 km) of the construction site and were reduced for about one to two hours post-piling (Brandt et al., 2018). At the maximum effect distances (from 17 km out to approximately 33 km) avoidance only occurred during the hours of piling. Brandt et al. (2018) found harbour porpoise detections during piling resulted in a measurable response at noise levels exceeding 143 dB re 1 µPa2s and at lower received levels (i.e. at greater distances from the source) there was little evident decline in porpoise detections. These studies demonstrate the dose-response relationship between received noise levels and declines in porpoise detections although noting that the extent to which responses could occur will be context specific such that, particularly at lower received levels (i.e. 130 to 140 dB re 1 µPa2s), detectable responses may not be apparent from region to region.
  3. Building on earlier work presented in Southall et al. (2007) and the mounting literature in this area, Southall et al. (2021) introduced a concept of behavioural response severity spectrum with progressive severity of possible responses within three response categories: survival (e.g. resting, navigation, defence), feeding (e.g. search, consumption, energetics), and reproduction (e.g. mating, parenting). For example, at the point of the spectrum rated seven to nine (where sensitivity is highest) displacement is likely to occur resulting in movement of animals to areas with an increased risk of predation and/or with sub-optimal feeding grounds. A failure of vocal mechanisms to compensate for noise can result in interruption of key reproductive behaviour including mating and socialising, causing a reduction in an individual’s fitness leading to potential breeding failure and impact on survival rates.
  4. There are limitations of the single step-threshold approach for strong disturbance and mild disturbance as it does not account for inter-, or intraspecific variance or context-based variance. However, according to Southall et al. (2021) harbour porpoise within the area modelled as ‘strong disturbance’ would be most sensitive to behavioural effects and therefore may have a response score of seven or above. Mild disturbance (score four to six) could lead to effects such as changes in swimming speed and direction, minor disruptions in communication, interruptions in foraging, or disruption of parental attendance/nursing behaviour (Southall et al., 2021). Therefore, at the lower end of the behavioural response spectrum, the potential severity of effects is reduced and whilst there may be some detectable responses that could result in effects on the short-term health of animals, these are less likely to impact on the survival rate of the animal.
  5. Although harbour porpoise may be able to avoid the disturbed area and forage elsewhere, there may be a potential effect on reproductive success of some individuals. As aforementioned, it is anticipated that there would be some adaptability to the elevated noise levels from piling and therefore survival rates are not likely to be affected. The assessment is highly conservative due to uncertainties associated with the effects of behavioural disturbance on vital rates of harbour porpoise, as it assumes the same level of sensitivity for both strong and mild disturbance, noting that for the latter the sensitivity is likely to be lower.
                        Bottlenose dolphin
                        Injury
  1. Individual dolphins experiencing PTS would suffer a biological effect that could impact the animal’s health and vital rates (Erbe et al., 2018). Bottlenose dolphin are classed as HF cetaceans (Southall et al., 2019). As described for harbour porpoise, there are frequency-specific differences in the onset and growth of a noise-induced threshold shift in relation to the characteristics of the noise source and hearing sensitivity of the receiving species. For example, exposure of two captive bottlenose dolphins to an impulsive noise source between 3 kHz and 80 kHz found that there was increased susceptibility to auditory fatigue between frequencies of 10 to 30 kHz (Finneran et al., 2013).
                        Behavioural disturbance
  1. Bottlenose dolphin are thought to be less vulnerable to the effects of disturbance than harbour porpoise; with larger body sizes – and lower metabolic rates – the necessity to forage frequently is lower in comparison. Bottlenose dolphin is largely coastally distributed in relation to the Array marine mammal study area and are more abundant during spring and summer compared to autumn and winter months (Paxton et al., 2016).
  2. Limited information is available regarding the specific sensitivities of bottlenose dolphin as most studies have concentrated on harbour porpoise. A study of the response of bottlenose dolphin to piling noise during harbour construction works at the Nigg Energy Park in the Cromarty Firth (north-east Scotland) found that there was a measurable (albeit weak) response to impact and vibration piling with animals reducing the amount of time they spent in the vicinity of the construction works (Graham et al., 2017). Another study investigating dolphin detections in the Moray Firth during impact piling at the Moray East and Beatrice Offshore Wind Farms found surprising results at small temporal scales with an increase in dolphin detections on the southern Moray coast on days with impulsive noise compared to days without (Fernandez-Betelu et al., 2021). Predicted maximum received levels in coastal areas were 128 dB re. 1 µPa2s and 141 dB re. 1 µPa2s during piling at Beatrice Offshore Wind Farm Ltd and Moray Offshore Wind Farm Ltd, respectively (Fernandez-Betelu et al., 2021). The authors of this study warn that caution must be exercised in interpreting these results as increased click changes do not necessarily equate to larger groups sizes but may be due to a modification in behaviour (e.g. an increase in vocalisations during piling). The results, however, do suggest that impulsive noise generated during piling at the offshore wind farms did not cause any displacement of bottlenose dolphins from their population range. Notably, the received levels during piling at Moray Offshore Wind Farm are higher than those predicted for the outer isopleths (130 dB and 135 dB re. 1 µPa2s) that overlap with the Coastal East Scotland MU during piling at the Array, suggesting that disturbance at these lower noise levels is unlikely to lead to displacement effects.
                        Grey seal
                        Injury
  1. In comparison to cetaceans, seals are less dependent on hearing for foraging, but may rely on noise for communication and predator avoidance (e.g. Deecke et al., 2002). Seals can detect swimming fish with their vibrissae (Schulte-Pelkum et al., 2007) but, in certain conditions, they may also listen to noises produced by vocalising fish in order to hunt for prey. Consequently, the ecological consequences of a noise-induced threshold shift in seals may be a reduction in fitness, reproductive output and longevity (Kastelein et al., 2018b). A study by Hastie et al. (2015a) reported that, based on calculations of SEL of tagged harbour seals during the construction of the Lincs Offshore Wind Farm (Greater Wash, UK), at least half of the tagged seals would have received noise levels from pile driving that exceeded auditory injury thresholds for pinnipeds (PTS). Nevertheless, population estimates indicated that the relevant population trend was increasing and therefore (whilst there are many other ecological factors that will influence the population health) this indicated that predicted levels of PTS did not affect a sufficient numbers of individuals to cause a decrease in the population trajectory (Hastie et al., 2015b). Hastie et al. (2015a) did note that the paucity of data on effects of noise on seal hearing means the exposure criteria used are intentionally conservative and therefore predicted numbers of individuals likely to be affected by PTS would also have been highly conservative.
  2. Reichmuth et al. (2019) reported the first confirmed case of PTS following a known acoustic exposure event in a seal. The study evaluated the underwater hearing sensitivity of a trained harbour seal before and immediately following exposure to 4.1 kHz tonal fatiguing stimulus (SPLrms was increased from 117 to 182 dB re 1 μPa). Rather than the expected pattern of TTS onset and growth, an abrupt threshold shift of >47 dB (i.e. the difference between the pre-exposure and post-exposure hearing thresholds in dB) was observed half an octave above the exposure frequency. Hearing at 4.1 kHz recovered within 48 hours, however, there was a PTS of at least 8 dB at 5.8 kHz, and hearing loss was evident for more than ten years.
  3. Despite the uncertainty in the ecological effects of PTS on seals, seals rely on hearing much less than cetaceans and therefore would exhibit some tolerance (i.e. the effect is unlikely to cause a change in either reproduction or survival rates). In addition, it has been proposed that seals may be able to self-mitigate (i.e. reduce their hearing sensitivity in the presence of loud noises in order to reduce their perceived SPL) (Kastelein et al., 2018b). Although this evidence suggests a lower sensitivity of pinnipeds to PTS, a precautionary approach has been taken within this assessment due to the potential for uncertainties surrounding their lower sensitivity to PTS.
                        Behavioural disturbance
  1. Mild disturbance has the potential to disturb seals, however this constitutes only slight changes in behaviour, such as changes in swimming speed or direction, and is unlikely to result in population-level effects. Although there are likely to be alternative foraging sites, barrier effects could either prevent seals from travelling to forage from haul-out sites or force seals to travel greater distances than is usual. Strong disturbance could result in displacement of seals from an area.
  2. Hastie et al. (2021) measured the relative influence of perceived risk of a noise (silence, pile driving, and a tidal turbine) and prey patch quality (low density versus high density), in grey seal in an experimental pool environment. The study found foraging success was highest under relatively silent conditions. When noise from tidal turbines and pile driving was introduced, foraging success remained similar to silent conditions when subjected to prey density was high. When exposed to tidal turbine and pile driving noise, the foraging success at low-density prey conditions was significant reduced. Therefore, avoidance rates were dependent on prey densities as well as the perceived risk from the anthropogenic noise and therefore it can be anticipated such decisions are consistent with a risk/profit balancing approach.
  3. Seal behaviour during offshore wind farm installation has been studied based on empirical data (Russell et al., 2016). Movements of tagged harbour seal during piling at the Lincs Offshore Wind Farm in the Greater Wash showed significant avoidance of the offshore wind farm by harbour seal (Russell et al., 2016). Within this study, seal abundance significantly reduced from the piling activity over a distance of up to 25 km and there was a 19% to 23% decrease in usage within this range. Nevertheless, displacement was limited to pile driving activity only, and harbour seal returned rapidly to baseline levels of activity within two hours of cessation of the piling (Russell et al., 2016). More recently, a study by Whyte et al. (2020) used tracking data from 24 harbour seal to estimate the potential effects of pile driving noise on this species. Predicted cumulative sound exposure levels (SELcum) experienced by each seal were compared to different auditory weighting functions and thresholds for TTS and PTS. The study used predictions of seal density during pile driving made by Russell et al. (2016) compared to distance from the wind farm and predicted single-strike sound exposure levels (SELss) by multiple approaches. Predicted seal density significantly decreased within 25 km or SELss (averaged across depths and pile installations) above 145 dB re 1 μPa2s. Predictions of seal density, and changes in seal density, during piling were given in Table V in Whyte et al. (2020), averaged across all water depths and piling events.
  4. Diverse reactions of tracked grey seal to pile driving during construction of the Luchterduinen and Gemini wind farms was reported in Aarts et al. (2018). Reactions ranged from altered surfacing and diving behaviour, changes in swimming direction, or coming to a halt. In some cases, however, no apparent changes in diving behaviour or movement were in Aarts et al. (2018). Similar to the conclusions drawn by Hastie et al. (2021), the study at the Luchterduinen and Gemini wind farms indicated animals were balancing risk (disturbance) with profit (prey). Approximately half of the tracked grey seal were absent from the pile driving area altogether, but this may be because animals were drawn to other more profitable areas as opposed to active avoidance of the noise, although a small sample size (n = 36 animals) means that no firm conclusions could be reached. It was notable that, in some cases, grey seal exposed to pile driving at distances shorter than 30 km returned to the same area on subsequent trips suggesting that the incentive to go to the area was stronger than potential deterrence effect of underwater noise from pile driving in some animals.
  5. Changes in behaviour and subsequent barrier effects have the potential to affect the ability of phocid seals to accumulate the energy reserves prior to both reproduction and lactation (Sparling et al., 2006). Female seals increase their foraging effort (including increased diving behaviour) before the breeding season, maximising energy allocation to reproduction. Especially during the third trimester of pregnancy, grey seal accumulate reserves of subcutaneous blubber which they use to synthesise milk during lactation (Hall et al., 2009). Therefore, grey seal foraging at-sea may be most vulnerable in this period, as maternal energy storage is extremely important to offspring survival and female fitness (Ailsa J et al., 2001, Mellish et al., 1999). Potential exclusion from foraging grounds during this time could affect reproduction rates and probability of survival.
  6. Phocid seals may also be vulnerable to disturbance during the lactation period, depending on the breeding strategy of a particular species. The lactation period for grey seal is shorter than for harbour seal, lasting around 17 days (Sparling et al., 2006) with females remaining mostly on shore, fasting. Furthermore, as grey seal females do not forage often during lactation, it is expected that they may exhibit some tolerance to disturbance as they would not spend as much time at-sea, where they can be affected by underwater noise. Following lactation however female grey seal return to the water and must forage extensively to build up lost energy reserves. Consequences of disturbance may include reduced fecundity, reduced fitness, and reduced reproductive success. Although grey seal may be able to avoid the disturbed area and forage elsewhere, there may be an energetic cost to having to move greater distances to find food, and therefore there may be a potential effect on reproductive success of some individuals.