Pilot whale stranding: acoustic analysis report

6. Discussion

The stranding of 55 pilot whales on a beach at Tolsta, Isle of Lewis was one of the largest seen in Scotland at the time. While the work undertaken by SMASS to understand the pathology of the stranded animals is critical, there are also questions about the acoustic environment that the animals were exposed to prior to the stranding, and the potential influence that these may have had on the animals. Here, we have reviewed data from underwater sound recorders, at sites in the region and close vicinity of the stranding, and investigated potential sources of sounds.

Analysis of the data suggests a large amount of cetacean activity in the location of the Tolsta acoustic mooring in the week prior to the pilot whale stranding. However, while 1,330 unique detection events from the broadband data seems a large number for such a short time period (10^th to 16^th July 2023), this analysis is not able to differentiate between species, individual animals and groups of animals, nor can it determine if the same individual made repeated calls. This means that the number of individuals or groups present was almost certainly substantially lower than the number of unique detection events. However, it does indicate that the area is used regularly by cetaceans, and other studies have shown a significant correlation between metrics of acoustic detection and density of cetaceans such as harbour porpoise (Brookes et al. 2013, Williamson et al. 2016). Furthermore, vocalisations occurring toward the end of a ten minute recording window may have continued into the next recording window but are treated as separate encounters given the 20 minute gap between duty cycled recordings. Several of the detected whistles fall into the range of pilot whale vocalisation frequencies. Further investigation to determine the number of pilot whale detections may be useful but would require engagement of an expert on pilot whale acoustics, given the lack of available pilot whale automated detectors. This report relies on a qualitative assessment of the presence of cetaceans.

Impulsive sounds were recorded twice at the Tolsta site in the early hours of the day of the stranding. A series of pronounced, impulsive sounds at ~12 kHz were detected at 01:00 UTC on 16^th July (Figure 11) in the period leading up to the discovery of the stranded pilot whales later that morning. The pulsed 12 kHz signals are consistent with a vessel echosounder, with a three pulse sequence identified potentially representing direct arrival at the recorder and echoes from the seabed and sea surface. Given the time of arrival of the pulses, it is hypothesised that the first was the direct arrival of the sound at the recorder, while the second (with a short interval from the first) was the echo from the seabed, which the recorder was moored in close proximity to. The third pulse, with a longer interval, may be the echo from the sea surface.

At 05:00 UTC on 16^th July, impulsive sounds at ~50 kHz were detected by both the F-POD (Figure 8) and the broadband recorder (Figure 12) at the Tolsta site. The sounds coincided with detection of ship noise in the recordings, which increases the likelihood of this sound source being a vessel based echosounder. However, there was no evidence in the AIS or VMS data of vessels being present in the area at the time of either of the impulsive noise events, suggesting that the source may have been a small fishing vessel, leisure craft, or other vessel that is not required to operate AIS or VMS. F-POD data showed the presence of porpoises and delphinids between bouts of this sonar activity during 16^th July (Figure 8).

Pilot whales have been shown to make behavioural changes such as more frequent changes in direction (Quick et al. 2017) in response to some impulsive sounds, including scientific sonar (EK60). It would be useful to understand the area of sea over which the sounds detected by the recorders were audible to allow an interpretation of the likelihood of them being a factor in the stranding, but unfortunately, this is not possible for several reasons. Firstly, the location where the sound was being produced is unknown, meaning it cannot be back propagated to understand the source level (how loud the sound was at its source), information which is required to understand the range at which the sound could be heard. Secondly, in both cases, the sound amplitude diminished throughout the recording, indicating that either the sound level at source was decreasing or the sound source was moving away from the recorder. If the sound source was moving, it is not known what path it took through the area (due to the lack of records in the AIS or VMS data) and this would affect the area of sea that was impacted by the sound.

Long-finned pilot whales’ peak hearing sensitivity is approximately 40 kHz with highly sensitive hearing between 4 and 64 kHz (Pacini et al. 2010). While absolute sound levels were not quantifiable, the relative strength of the impulsive sounds described above for 16^th July at 01:00 UTC and 05:00 UTC were above ambient sound levels and so have the potential for pilot whales to have heard them. However, since most transducers direct sound energy downwards in a narrow beam, it is unlikely that they would ensonify a wide area. Given that we do not know the location of the sound sources, nor of the pilot whales prior to the stranding, we cannot determine whether the pilot whales were exposed to these sounds. Additionally, given the uncertainty around the time of the stranding, it cannot be determined whether the recorded sounds occurred before or after the stranding event. As a result, the potential influence of these sounds in the stranding is unknown. Any future inferences on potential sources of the sound and the resultant influence on the pilot whales’ behaviour and stranding must therefore be caveated and contextualised with results of pathological analysis on the individuals and any other available data.

Analysis of the broadband acoustic data from the Garenin site, located west of the Isle of Lewis, identified regular presence of continuous anthropogenic noise, most likely shipping noise. There were very few cetacean detections, but other biological sounds, likely from fish such as haddock, were recorded. Although there were some geophysical surveys being undertaken in the wider area at the time, no impulsive sounds were identified within the broadband data. It is likely that most of the equipment being used for the geophysical surveys was operating at frequencies higher than the broadband devices were capable of recording. Such frequencies are also beyond the hearing ranges of cetaceans, and also attenuate rapidly, so are unlikely to have ensonified large areas of sea. Additionally, the sub bottom profiler that was operating at frequencies around 100 kHz appears to have a narrow beam width (~2°); frequency attenuation and the narrow beam width may have prevented it from propagating over larger distances.

The information provided by the MOD does not suggest any obvious activity that may have been a factor in the strandings. Previous mass strandings linked with military activity around the world have primarily been associated with the use of active sonar (D’Amico et al. 2009; Dolman et al. 2011; Parsons et al. 2008), or explosives (Brownlow et al. 2015), neither of which have been reported as having been undertaken within the area in the week before the stranding.