Ambient noise in the marine environment has increased since the 1950s due to the rise of human activities at sea (Frisk, 2012). In response to the heightened concerns about the potential impacts of noise on marine life, many countries have reached international agreements to monitor underwater noise levels (Van der Graaf et al., 2012; OSPAR, 2017; Reeve, 2019). In Europe, the Marine Strategy Framework Directive (MSFD) requires Member States to avoid sound sources that are likely to cause significant impact in the marine environment (Tasker et al., 2010). The MSFD sets out a series of indicators to assess the environmental status of an area, which for underwater noise is the proportion and spatial distribution of days on which sound sources exceed levels likely to entail significant impact.

Human activities such as pile-driving, seismic surveys and sonars produce some of the most powerful sounds underwater (Gordon et al., 2003). These impulsive sounds are defined as acute, broadband, transient signals with a rapid onset and a rapid decay (<1 s) and are considered particularly hazardous to marine mammals (Southall et al., 2007). However, during the propagation, the acoustic characteristics of sound change at rates that vary according to the specific environmental conditions. The greatest change in the acoustic characteristics of impulsive sounds occurs within ∼ 10 km from the source (Hastie et al., 2019) and, therefore, the hazardous characteristics of these sounds will vary with distance (Southall et al., 2007). Impulsive sounds have the potential to impact cetaceans through direct injury at shorter distances (here near-field) and through behavioral disturbance at longer distances (here far-field; Erbe et al., 2018). Measures to mitigate near-field effects from injury are widely adopted (Bröker et al., 2015; Verfuss et al., 2016). However, potential far-field behavioral effects and the longer-term consequences of any short-term disturbance remain challenging to assess (Pirotta et al., 2018). To date, most of the studies on the effect of impulsive noise on wild marine mammals have focused on harbor porpoises (Phocoena phocoena) and seals (Phoca vitulina) (e.g., Russell et al., 2016; Brandt et al., 2018; Graham et al., 2019). Other species, such as bottlenose dolphins (Tursiops truncatus), may react differently to this disturbance but information on this is sparse (David, 2006; Finneran et al., 2015; Graham et al., 2017). Due to the lack of empirical data, estimates of potential behavioral effects required for the regulation of marine energy developments have a high level of uncertainty for these animals (Merchant, 2019).

This uncertainty can be especially challenging where offshore energy developments are being considered within or near to areas used by protected cetacean populations. Over the last decade, three major energy developments have been undertaken in NE Scotland, near a Special Area of Conservation (SAC) that was established in the Moray Firth to protect bottlenose dolphins. These included a 2D seismic survey for oil and gas exploration in 2011, and the installation of foundation piles for two offshore wind farms in 2017 (Beatrice Offshore Wind Farm) and 2019 (Moray East Offshore Wind Farm), all of which are known to result in high levels of impulsive underwater noise (Madsen et al., 2006; Thomsen et al., 2011). The Habitats Directive (92/43/EEC) requires regulators to ensure that the Favorable Conservation Status of the SAC is maintained. Therefore, higher levels of assessment and protection were required prior to all three of these projects. The Appropriate Assessments undertaken by the regulator, concluded that there were no likely long-term impacts on the protected bottlenose dolphin population and permissions were granted (Berr, 2007; MS-LOT, 2014; MS-LOT and Marine Scotland Science, 2014). However, significant objections were raised during some of these processes, arguing that dolphins could be displaced from the southern coast of the Moray Firth, an important area for this population (Cheney et al., 2013).

Given the sensitivities surrounding this issue and the level of uncertainty in the predictions, we used passive acoustic monitoring (PAM) to investigate whether far-field effects of impulsive noise from these offshore activities resulted in displacement of dolphins from coastal waters in the southern Moray Firth. To do so, echolocation data loggers (CPODs) were deployed to study variation in dolphin detections in areas exposed to different levels of impulsive noise. First, year-to-year variability in dolphin occurrence was investigated in each of those areas. Second, Before-After Control-Impact analyses (BACI; Underwood, 1992; Smith, 2002) were performed at two temporal scales to assess potential differences in detections within these areas. At the medium temporal scale, the BACI analysis compared years in which impulsive noise from offshore activities was present or absent. At the small temporal scale, finer-scale patterns within years with offshore activity were explored by comparing days in which impulsive noise was present or absent. Lastly, during the construction of Moray East wind farm, a more extensive PAM array was deployed in the coastal area nearest this development to investigate whether dolphins were displaced away from it and toward the coast during piling days.

CPODs were successfully deployed from 2009 to 2019 and provided more than 10,000 days of CPOD data. Gaps in the dataset occurred due to a combination of device failure and logistical constraints. Data from 2012 were removed from these analyses because only one CPOD was recovered that year from the Impact Area (Supplementary Figure 1).

The analysis of 10 years of PAM data highlighted that dolphins used Moray Firth inshore areas regularly, albeit the extent of use varied from year to year without any consistent relationship to the impulsive noise generated by offshore activities. Nevertheless, at a fine temporal scale within offshore activity years, there was a significant change in dolphin occurrence depending upon the presence or absence of impulsive noise on different days.

The results of this study suggest that the impulsive noise generated by offshore activities did not cause any dolphin displacement from the southern coast of the Moray Firth. The southern coast is the closest area to the offshore activities within this bottlenose dolphin population’s range (Arso Civil et al., 2019): the seismic survey took place 20–30 km away, and piling at Beatrice 50–70 km and Moray East 40–70 km, from the southern coast. Predicted maximum received noise levels were 139 dB and 128 dB re 1 μPa²s during the seismic survey and during piling at Beatrice wind farm, respectively (unweighted single pulse SEL; Thompson et al., 2013; Graham et al., 2019) and 141 dB re μPa²s in the worst-case scenario for Moray East wind farm (MORL, 2016). Our analyses showed that dolphins continued using the southern coast of the Moray Firth during the seismic survey and impact pile-driving. These results are in line with recent studies that found that displacement of marine mammals from impulsive noise sources only occurs at shorter distances. For instance, displacement of harbor porpoises and baleen whales, which are more sensitive to noise (Southall et al., 2019b), has been reported up to 20 km away from impulsive noise sources (Dähne et al., 2013; Dunlop et al., 2018; Graham et al., 2019; Southall et al., 2019a; Sarnocińska et al., 2020). There are no similar studies for bottlenose dolphins but, since they are less sensitive than baleen whales and harbor porpoises to noise, shorter ranges of displacement would be expected.

Contrary to expectations, the BACI analysis at the smaller temporal scale, showed an increase in dolphin detections on the southern Moray Firth coast on days with impulsive noise. Furthermore, this increase was consistent between all three offshore projects. The short-term increase in dolphin detections observed only at the smallest temporal scale is in line with previous studies that found that behavioral alterations due to noise exposure last less than a day (Thompson et al., 2013; Van Beest et al., 2018). Our findings are consistent with those of Bailey et al. (2010b) who, based on criteria in Southall et al. (2007) suggested that dolphins in this study area might exhibit modifications in behavior at around 50 km away from a piling vessel. Acute noises can modify marine mammal group size and group behavior (Visser et al., 2016; Curé et al., 2021), which in turn can cause changes in vocalizations (Henderson et al., 2012). Therefore, one possible explanation for the observed increase in dolphin detections is that noise modified bottlenose dolphin group sizes or group behavior. Although whistle vocalization rates have been linked to group sizes (Quick and Janik, 2008), echolocation click rates cannot be directly linked to the size of bottlenose dolphin groups (Nuuttila et al., 2013). Since our study was based on echolocation data loggers, we cannot test whether the increase in noise levels modified dolphin group sizes. Moreover, since CPOD detections and the number of individuals are not directly linked (Nuuttila et al., 2013), it does not follow that the increase in detections observed here is indicative of changes in group size. Changes in marine mammal vocalizations due to distant (>20 km) anthropogenic noise sources have been observed previously (Risch et al., 2012; Blackwell et al., 2013; Shannon et al., 2016). Therefore, another explanation of the increase in detections is that dolphins might have changed their vocalization rate (Blackwell et al., 2015) or the amplitude of their calls (Holt et al., 2009; Parks et al., 2011) in response to the impulsive noise generated by the offshore activities. In an experiment with captive bottlenose dolphins, the playback of pile driving noise resulted in an increase in the number of clicks produced by these animals (Branstetter et al., 2018). Caution is required when extrapolating information from trained to wild animals, but an increase in the click rate could explain the observed increase in detections by our CPODs during the impulsive noise events. Results obtained from the fine scale spatial analysis during the construction of Moray East wind farm also support this hypothesis. No displacement from the outer toward the inner array was observed during piling days, while a subtle but consistent increase in dolphin detections was observed on both arrays. Research is needed to further investigate potential changes in the acoustic characteristics of bottlenose dolphin clicks during far-field impulsive noise events. Passive acoustic devices that collect information on marine mammal click characteristics could be deployed during future offshore developments to explore in more detail whether these activities can be linked to differences in dolphin click rates or click amplitude.

At the medium temporal scale, comparing dolphin occurrence to baseline years, the results were not consistent between offshore activities. During the seismic survey, an increase in dolphin detections was observed on the southern coast. During the construction of Beatrice wind farm, a decrease was detected in the same area and during the construction of Moray East wind farm, no significant difference was detected. In line with previous studies, dolphin detections overall were higher in the Reference Area of the inner Moray Firth, compared with the Impact Inner area on the southern coast, but the extent of use varied between years and through the season (Thompson et al., 2015; Fernandez-Betelu et al., 2019). The difference in detections at the medium temporal scale therefore seems likely to be related to the natural inter-annual variation in occurrence that was also reflected in our analyses here (Figure 2). Although potentially modified by anthropogenic factors, bottlenose dolphin occurrence is largely influenced by a range of natural drivers, such as prey abundance and oceanographic processes (Heithaus and Dill, 2002; Benjamins et al., 2015). Our results suggest that the effect of far-field noise is less important than these other natural drivers in affecting the occurrence of dolphins in this study area.

We used a BACI design to investigate the effect of impulsive noise on the coastal occurrence of dolphins. This methodology is based on the comparison of data from two similar areas (Impact and Control) in situations where only one of them is affected by a disturbance. However, finding control sites with similar ecological characteristics to impact sites, while being distant enough to be unaffected by the disturbance, is not always feasible (Underwood, 1992, 1994). The sites defined here as reference sites were chosen because they were located at greater distances from the developments but were still within the bottlenose dolphin population’s range. However, a complete lack of disturbance cannot be assured there, hence the term Reference instead of Control Area. For instance, during the construction of Moray East wind farm an increase in detections was observed at both Reference and Impact Areas during piling days, potentially indicating that noise levels might also have affected dolphins at reference sites. Nevertheless, these 10 years of passive acoustic monitoring data provided a robust dataset to investigate the effects of three offshore projects, while overcoming some of the limitations of the BACI design. The consistent results at the smallest temporal scale strengthened our hypothesis that a change in vocal behavior may have caused the observed increase in dolphin detections. While the contrasting results at the medium temporal scale from multiple offshore activities prevented us from suggesting misleading effects that might have been inferred from work conducted only during a single project. Overall, our analyses showed that long-term datasets such as this provide opportunities to detect relatively subtle differences in bottlenose dolphin behavior linked to the far-field impulsive noise generated by offshore activities.

All data are available from the Dryad Digital Repository. All datasets and R code used for the analyses are available at: https://doi.org/10.5061/dryad.6hdr7sr17.

Ethical review and approval was not required for the animal study because this was a non-invasive, acoustic observational study of bottlenose dolphin responses to offshore seismic surveys and pile-driving. The authors had no control or influence over the duration or scheduling of offshore activities. No animals were captured or tagged during this study and no research or animal ethical assessments were required. Bottlenose dolphin responses to impulsive noise were determined using remote passive acoustic devices on seabed moorings licensed for scientific use by Marine Scotland, and consented by the Crown Estate. Moorings were deployed and recovered using vessels with appropriate certification, accreditation and endorsements.

OFB, IG, and PT designed the study based on data collection that had been led by IG, KB, and PT. TB managed data collection. IG and BC processed the raw data. All analyses were designed and conducted by OFB and supervised by IG. OFB led the writing of the manuscript. All authors reviewed and contributed critically to the drafts.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.