Salmon farming - use of acoustic deterrent devices: report

5. Outcomes of evidence gathering

What effect do seals have on farmed finfish?

Direct losses

There is evidence from Scotland and other countries that direct attacks by seals can cause significant mortality in farmed fish. In 2013, Northridge and colleagues reported that 1.4 million fish were lost to seals over a 10-year period in Scottish salmon farms⁴. In response to the SAWC survey, data were presented from salmon producers organisations (Scottish Salmon Producers Organisation (SSPO) – now Salmon Scotland) that showed that up to 500,000 fish could be lost per year, with 80-92% of Scottish farms affected. Images were presented in the reports from the producer organisations (SSPO, Scottish Sea Farms) that showed injuries indicative of bite wounds, which in some cases caused the loss of substantial portions of the bodies of the fish. It was suggested that seals preferentially target the heart and liver.

Effects of seal presence on stress and subsequent disease in fish

In terms of the effect of seal presence, it was suggested that this could cause reductions in feeding, reduced growth or compromise the immune system leading to secondary disease. However, data directly linking seal presence with disease events or reductions in growth were not presented in the survey responses. It was explained that while feed consumption, growth and mortality are extensively recorded, seal presence could not be systematically monitored. Seals may also be present in the vicinity of the farm for an extended period of time rather than in a single defined 'event'. Additionally, fluctuations in growth and disease occurrence can also occur due to other factors. This means that while fluctuations in feeding and growth occur regularly and can be quantified, it is difficult to link any particular disease event or perturbation in growth or feeding to the presence of seals at particular times.

However, the opinion was strongly expressed that seal presence does cause stress to the fish, and that this has adverse effects on their health and welfare. The SSPO stated that it was universally accepted among fish farmers that seal attacks (the SSPO categorised both 'direct attacks' and 'presence' as 'attacks') cause reductions in feeding and growth. It was noted that these effects were more marked in sea pens with more seal presence/attack, which provides indirect evidence of a link. Further indirect evidence comes from observation of the video camera footage that is used to monitor feeding. It was stated that changes in fish behaviour are seen in the presence of seals and/or seal attacks. However, it should be noted that the fish do not always have sufficient space to 'escape', i.e., by moving far away from the predator, particularly when being crowded for handling, or when kept at high stocking densities in sea pens (although this is not recommended practice).

The SSPO Prescribing Vets Group presented a separate response. The members of this group are experienced fish veterinarians, who mostly work with the salmon producer companies. Their submission stated that based on their professional expertise and experience, seal attacks and presence have significant negative impacts on salmon growth, health and the incidence of disease, and that seal presence causes changes in fish behaviour. Additionally, they stated that they have witnessed instances of disease outbreak where they considered that a seal attack was the primary or an underpinning factor.

Evidence from other species

As there is no research that directly shows that salmon and trout are stressed by the presence of seals, and that stress may be responsible for down-stream effects on health and functioning, we have taken into consideration the literature relating to other species, in order to understand more generally the nature of the response of prey species to predators and to determine whether fishes are capable of showing a similar response.

The effects of the presence of predators on their prey is well known to be stressful and profound, as it is fundamentally related to survival. In addition to lethal encounters with predators, animals that are preyed upon also show changes in foraging response and increased energetic expenditure from mounting a 'fight' or 'flight' response. This is typically expressed through activation of physiological pathways, leading to increased metabolism and other changes that promote effective escape (such as increased oxygenation of tissues), and behavioural responses. Stress, even for a relatively short period of time, if severe, can lead to marked behavioural and physiological changes. For example, feeding responses were reduced for six days in two fish species towed in a net for 15 minutes and for one species (walleye pollock, Gadus chalcogrammus), the fish did not recover and the treatment resulted in 100% mortality²¹. As a significant predator of salmon, seals are undoubtedly a stressor for salmon, but the extent and impact of this on fish welfare, health and behaviour needs to be assessed, if at all possible.

There is no published study which assesses the impact or extent of a stress response by salmon to the presence of seals. However, recently published papers indicate that fishes of other species show an acute physiological stress response in the presence of a predator. In a study of the response to practices, such as handling and slaughter in European whitefish (Coregonus lavaretus), fish were implanted with cardiac monitors²². During the experiment, an unintended event occurred where the holding cage of the group of test whitefish was held close to a cage of rainbow trout (Oncorhynchus mykiss), a natural predator of the whitefish. There was a sustained rise in heart rate that lasted at least 12 hours. The authors stated that such a response is suggestive of a major allostatic load on the fish that would have had a negative impact on other physiological processes.

In studies performed with qingbo (the Chinese barbed carp, Spinibarbus sinensis) and zebrafish (Danio rerio), the fish showed behavioural responses, increases in physiological stress responses (cortisol release), and increased metabolic rates in the presence of predator species^23,24,25). However, these papers only considered the short-term response of the fishes and not a longer-term impact of the presence of a predator on behaviour and health.

There are studies of responses of prey species to the presence of predators in mammalian species that are of relevance here. Studies of predator-prey relationships on stress responses and longer-term consequences are sparse, but some intriguing studies relevant to the current issue are available, involving terrestrial and aquatic animals, including impacts of predation on seals themselves. Cape fur seals (Arctocephalus pusillus pusillus), which are preyed upon by great white sharks (Carcharodon carcharias), have been shown to have higher chronic physiological stress responses (faecal glucocorticoid metabolites), when living in areas of high shark predation, when compared with fur seals in areas of lower predation²⁶. Wild ungulate species are also vulnerable to predation, as are domestic ungulates (sheep (Ovis aries)), cattle (Bos taurus), pigs (Sus domesticus)) and poultry species. In the presence of high predator density, wild sheep species escape to areas of their range that are perceived to be safer (something generally not available to domestic animals and farmed salmon which are confined) and then remain immobile until the threat has passed. With greater predation pressure animals maintain longer periods of reduced activity²⁷. Greater flight distances, increased vigilance and reduced time spent in maintenance behaviours (such as feeding) are also seen in several wild ungulate species when the predation risk increases. Prolonged avoidance of pastures, where predator attacks have occurred, are also reported in domestic ungulates where they have the opportunity to express these responses²⁸. Chronic or prolonged stress, such as that experienced when animals are frequently exposed to predator attack or in areas of high predator population density, can cause reduced growth rates, impair reproductive function, reduce immune responses and increase disease susceptibility²⁹. In cattle farming in Australia, for example, predation by wild dogs (dingoes (Canis familiaris)) is thought to cause stress-related impacts, including reduced weight gain and poor reproduction (lactation and delayed oestrus³⁰) even in the absence of direct attacks on individuals.

Section summary

There are reports that document incidents in which seals have attacked and killed large numbers of farmed salmon. Anecdotal evidence from producers agree that seals may cause high level of mortality. However, there are no data currently available to show that the close presence of seals causes chronic stress leading to increased susceptibility to disease, or other negative aspects of health and welfare. However, it is likely that the fish show a similar stress response to seals as a prey mammal would to the close proximity of a predator. There is also increasing evidence that fishes respond in this way, although farmed animals, including fishes, are often physically prevented from showing avoidance or distancing responses from predators, which may intensify the stress response.

Evidence for effects of ADDs on cetaceans and seals

The main welfare concerns for cetaceans and seals with respect to ADD use are temporary or permanent loss of hearing, interference with the ability of cetaceans to hunt and navigate using their own sonar, and avoidance of habitats where ADDs are deployed, which may affect foraging and reproduction. These aspects will be discussed below.

Cetacean hearing thresholds compared with ADD output

Hearing is central to vital behaviours in cetaceans, such as communication, prey location, predator detection and navigation³¹. The characteristics of hearing that are important with respect to ADD disturbance are the range and sensitivity of hearing. Cetaceans are classified into low-, mid- and high-frequency hearing groups based on their hearing ranges³². Minke whales (Balaenoptera acutorostrata) are considered to be in the low-frequency class, killer whales (Orcinus orca) in the mid-frequency class and bottlenose dolphins (Tursiops truncatus) and harbour porpoises (Phocoena phocena) in the high frequency class³³. Hearing thresholds of the relevant cetaceans and seals are shown in Table 1. For reference, reports state that the transmission frequency of ADDs range between 5 to 27 kHz¹¹ or 2-40 kHz¹² and at a source level (intensity of sound) of between 170-200 dB re 1 µPa. However, since this study was published, most ADD manufacturers have reduced the frequency transmitted by their devices. It is thought that the output of most devices is now below 10kHz, but manufacturers do not typically publish this information.

The data from the table above shows that the hearing ranges of seals and cetaceans largely overlap, so that there is no frequency that can be used by ADDs to target seals that is outwith that of all species of cetaceans found near Scotland.

Assessing the effects of ADD transmission on cetacean hearing, behaviour and welfare

There are a few studies that have directly measured the hearing range of cetaceans and the parts of the hearing range at which hearing damage occurs^32,33. The data from these audiometry studies have been used, together with sound pressure level and exposure times created by ADD outputs, in calculations that suggest that hearing damage may occur following exposures ranging from two minutes to several hours, depending on the specifications of the ADD deployed¹¹.

A range of impacts on cetaceans has also been observed from in-situ studies, with individuals' reactions ranging from no reaction to ADD sounds to faster swimming away from ADDs. For example, studies using ADDs designed to exclude cetaceans from construction sites have shown that porpoises and baleen whales (such as the minke whale) avoid areas in which they are deployed^9,39.

There have been a number of studies that suggest that ADDs may cause animals to leave areas of habitat. Studies from Canada have recorded animals retreating from ADD transmissions and lower numbers of cetaceans in areas where ADDs were deployed^40,41,42. A study in Scotland found that harbour porpoises tend to avoid areas where ADDs are active, but not exclusively, as animals were detected feeding within 200m of an ADD site⁴³. In order to understand the potential geographical spread of the ADD sound, one study¹⁶ created a map depicting areas of sea off the western coast of Scotland and around the Northern Isles that were likely to receive levels of ADD sound that could affect cetaceans. The resulting map showed almost total geographical coverage of some areas, but given the different duty cycles or schedules of use, this does not mean that the transmission was constant.

However, there are some methodological issues with these studies. In studies that observe the response of wild or captive animals to ADD deployment, only small numbers of animals are typically used. The studies are often of a short duration, which is not appropriate for understanding the effects of long-term use of ADDs in commercial finfish farming.

However, local environmental conditions, such as water depth, sediment type, slope of the seabed, and the complexity and topography of the seabed and coastline, can affect the propagation loss of ADD outputs and hence their effectiveness against seals or their adverse effects on cetaceans^12,43. These environmental conditions are difficult to account for in modelling studies. Critically, many of these studies were based on older models of the ADDs, which were often used for prolonged periods of transmission, and used a wider range of frequencies and a longer duty cycle (duration of transmission) than current models of ADDs. These studies also mostly used a particular brand of ADD which was not widely used in Scotland.

Additionally, if ADDs were causing the permanent exclusion of cetaceans from their home ranges, it would be expected that reductions in the numbers of cetaceans living in the areas where fish farms are located would be observed. An assessment which considered information on the abundance and distribution of the major species (produced by large-scale surveys carried out in European waters), suggests that cetacean numbers are stable in the North Sea and on the European continental shelf areas, but data are not available for the west coast where large numbers of fish farms are located⁴⁵.

Effect of ADD transmissions on seal welfare

ADDs are considered to work by either causing auditory pain in the seal, such that the animal avoids the area where pain is experienced (known as conditioned place avoidance), or by creating aversive but non-painful acoustic stimuli that alter the behaviour of the animal to avoid exposure (known as a startle response). Newer types of ADD evoke a startle response. Seals may suffer acute and chronic exposure to the sounds from ADDs, which may have negative effects on their welfare. With a wide hearing range of 0.5-40 kHz³⁴, it is unclear how temporary or permanent impairment, or even loss of hearing may affect seals, but impacts may include reduced dynamic range (sound intensity), frequency discrimination and passive listening space. These effects in turn could impact detection of predators and prey, and communication with other seals, such as competition between males for females, and vocalising in the breeding season to attract mates⁴⁴, as well as increased energetic costs caused by moving greater distances⁴⁶. Even temporary impairment of hearing could have slight cumulative effects that become permanent injuries. It could be argued that permanent loss of hearing caused by ADDs could result in more seal depredation from sea pens as the seals will become "immune" to the deterrent effects of the ADDs and may become more dependent on an assured supply of food. In some studies, use of ADDs has resulted in increased losses to seals through a "dinner bell" effect of the ADDs, which attracts the seals when switched on¹¹.

It has been estimated that over short time periods, permanent damage to hearing may occur in harbour seals at a distance of only 7m from ADDs, but permanent damage may also occur at distances up to 60m when the exposure occurs over longer periods of months and years¹¹. Exposure times that cause hearing impairment vary between different brands of ADD, ranging from 3 minutes to 57 hours and 51 minutes¹¹.

Studies in the wild suggest variable responses of the seals to ADD deployment, with some seals avoiding sites with ADDs, sometimes for only a short period, and others where ADDs do not appear to have much impact on seal behaviour. Harbour seals may stop foraging and move away when they are within 1 km of ADDs, which are being used at levels of 134.6 dB re 1 μPa (RMS)⁴⁶. Seals have also been reported lifting their heads out of the water or even hauling or leaping out of the water when close to ADDs to avoid the sound, potentially reducing foraging time^46,47.

Another study was carried out in which seven harbour seals were tagged to examine their potential exposure to ADDs and potential for auditory impairment, if they were present. All seals were potentially exposed to ADDs at 51 of 56 sites, where mean ambient noise levels were exceeded. Temporary auditory impairment was expected in one of the seven seals (14.3%) across 1.7% of waters per 24 hours, which over time could impact significantly on the local seal population⁴⁴. Another study used FaunaGuard Seal modules, which operate at 0.2-200 kHz for 3-10 seconds (142 dB re 1 μPa) and found that there was no permanent hearing impairment, if the seals were more than 100-200 metres away⁴⁸.

In humans, persistent low-level noise exposure may increase stress hormone levels, blood pressure and heart rate, leading to hypertension, arrhythmia, dyslipidemia, increased blood viscosity and blood glucose, and the activation of blood clotting factors, consequently increasing the risk of cerebrocardiovascular diseases such as stroke, ischaemic heart disease, acute myocardial infarction, heart failure, and arterial hypertension⁴⁹. Therefore, chronic noise from ADDs may have similar effects on the health of seals in close proximity to finfish farms. As well as the direct effects of ADDs on seal behaviour, stress responses (such as a startle response), and possible hearing damage, effective use of ADDs may also exclude seals from feeding, resting and breeding habitats. Depending on the availability of suitable alternative habitats, this may affect hunger and feeding motivation (with a potential impact on increasing attacks on farmed fish), social interactions, reproductive success, and disease resistance as outcomes of chronic stress.

Although ADDs are normally aversive to harbour seals, captive seals habituated quickly when exposed to sounds at 146 dB, when they were fed, which suggests that seals are willing to overcome discomfort, if they are hungry, so that motivation may be an important factor in the effectiveness of ADDs¹⁵. However, hearing impairment would not be expected at this source level³³. Habituation times vary greatly between studies from a few days to up to several years¹¹.

Section summary

Although the evidence comes from a number of different sources and methodologies, it indicates that the sound transmitted from ADDs has the potential to cause hearing damage in cetaceans when exposed to the sound over a period of time. Exclusion from home ranges and feeding areas is more difficult to determine given the paucity of data. There is also evidence for the potential for hearing to be damaged in seals, and other impacts on seal welfare. The change in use of ADDs over time, and the variety of different brands with different specifications in current use also means that it is difficult to determine what the current situation is, compared to these published studies.

How effective are ADDs? Are there any viable alternatives?

There are varying reports on the effectiveness of ADDs. Some observational studies in fisheries have shown that ADDs are effective in deterring seals from salmon netting sites (i.e. not aquaculture sites^47,50) over a long period of time, while others have found that seals eventually return to the site. Experimental studies using live seals in a test pool have shown that the seals habituate to the sound or are able to overcome it when food was offered¹⁵. A survey of managers of marine salmon sites in Scotland and found that only 23% of them thought that the ADDs were very effective, but 50% thought that they were moderately effective⁵¹. However, it was noted that their use had increased markedly in the preceding decades. In another survey of Scottish fish farmers, there was little consensus found among fish farmers about the efficacy of ADDs, but most thought that they are effective at least some of the time⁴³. A major study in Scotland⁵ analysed the number of months in which seal predation was recorded as a function of whether the ADD was turned on or not, and also compared predation at sites where ADD use was not permitted with sites where it could be used. The results were equivocal. Overall, reports of seal predation were higher on farms using ADDs, but this likely reflects the fact that farms with seal predation issues are more likely to use ADDs. However, the small number of farms that were not permitted to use ADDs had higher predation rates than farms allowed to use ADDs, suggesting some efficacy of the devices. Responses to the SAWC survey from industry stated that ADDs formed 'part of a wider predator management strategy' and that ADDs' effectiveness depended on a number of factors such as farm location and seal predation pressure.

Other main means of deterring seals are frequent removal of mortalities from the bottom of nets, seal blinds and double-netting, maintaining good net tension and the use of extra-strength, high-density netting⁷. This netting requires higher levels of maintenance to prevent algal growth reducing water flow, but appears to have had some success in some locations. Research into electrifying the netting of the cages or the use of electrified or repellent-tasting baited 'dummy' fish are other alternatives⁷. The use of higher handrails or 'top-nets' (over the exposed water at the top of the net) were also suggested. The use of larger pen sizes, or lower stocking densities that allow the fish more room to escape or retreat were also suggested. However, many of these measures will deter seal attacks, but not necessarily seal presence.

Section summary

Uncertainty about the effectiveness of ADDs, coupled with apparent reluctance to give up their use, suggests that they may work in some cases for some of the time. Demand for their use suggests that seal predation is a constant problem for fish farmers, but also that there are not any sufficiently effective alternative solutions currently available

Developments in seal deterrent methods

When ADDs were first installed to deter seals, they tended to be turned on continuously, but more modern devices have the ability to be used only when seal presence occurs or is suspected⁷.The use of technological solutions to detect seals in the vicinity of the nets to trigger ADD transmission has been recommended⁷ and is being used by at least one ADD producer, and others may follow this route. There is potential for ADDs to be able to detect only those seals that are planning to attack (such as swimming directly towards a sea pen) and these ADDs may also be able to detect the presence or absence of cetaceans before deployment. Similarly, if the technology could distinguish between cetaceans and seals, this would allow further targeting of ADD deployment and avoid unnecessary impacts on cetaceans. However, further development and testing of the technology are required.

Further developments of a range of non-lethal control options are essential, to avoid a continuing "arms race" between seals and fish farmers in the development and use of non-excluding technologies. To resolve the issue of seals entering the pens over the sides of the pen, alternative designs could be developed. Sea pens with higher side walls that prevent seal ingress have been designed and used in areas of Tasmania, for example. Although this design requires a much bigger pen diameter than currently used in Scotland, and may pose operational difficulties, it may provide a viable solution to this problem.

More research is needed into the viability of alternatives, such as methods that startle seals. Using ADDs that elicited the startle reflex in grey seals led to a sensitisation (i.e., enhancement) of avoidance responses and resulted in avoidance of food locations near the source of the startling sounds⁵². Where ADD startle devices have been used for seals, the numbers of salmon predated by seals fell by 91%, while the numbers of harbour seals fell by 91% up to 250m away and there was no habituation. However, the studies were short term, and carried out over periods of 19 months and two months respectively^53,54.

Section summary

There are a number of promising developments that could reduce the use of or replace traditional 'constant transmission' types of ADDs. These include devices that evoke a startle response or technology that detects seal presence and/or the presence of cetaceans.