Scottish Animal Welfare Commission: welfare of isolated populations of feral animals - report

5. Welfare and ethical review

Much of the concern around the management of naturalised small ruminant populations has been focused on mortality and the way this occurs – either the numbers of deaths that occur on the St Kilda islands from natural causes, or the numbers of goats culled as part of control or eradication plans (predominantly for human goals or conservation rather than as a means to improve welfare). Although deaths of animals can be very distressing to observe, it is important that we consider the wider welfare and ethical impacts of any decision-making around management of these populations to reduce certain types of deaths.

5.1 Welfare review

In considering the welfare of the animals in different situations, we took into account three different welfare approaches. Firstly, we considered Charles Hume’s priorities for welfare⁹⁷:

• The intensity of the suffering involved

• Its usual duration

• The numbers of animals affected

• The feasibility of practical reform

Secondly, we considered Fraser’s three spheres (biological functioning, natural living, affective states⁹⁸) and thirdly we used the Five Domains Model of Animal Welfare³⁷ (as in previous SAWC reports). Fraser’s model is particularly relevant as it describes the potential conflicts between a focus on biological functioning (or mortality rates) on one hand and prioritising natural living on the other, which can be the issues when considering the welfare of feral animals. For both Fraser’s model and the Five Domains model the subjective experiences of the animals, their feelings or affective states, are important in understanding how the animal views its welfare state. In undertaking this review, we weighed the costs and benefits to the animals for the different scenarios outlined above on the basis of the Five Domains (see Appendix VI). The current situation of an unmanaged or hands-off approach is evaluated as the baseline scenario and the potential improvements or detriments of other scenarios are then compared to this baseline. A commentary of the key conclusions is presented here.

5.1.1 Unmanaged populations/hands-off approach

Currently naturalised populations of sheep and goats experience compromise due to restricted food availability seasonally; exposure to extreme winter weather; risk of injuries as a consequence of the rocky or steep terrain (although both small ruminant species are considered well adapted to these environments, it is still a risk); risk of injuries from fighting in the rut; weakness and increased disease susceptibility; increase in competition for limited resources and may experience hunger, weakness, exhaustion, and frustration. They also experience welfare enhancements through choice of diverse feeds and an ability to express dietary preferences; opportunity for exploration and investigation; selection of a variety of different shelter types; no pain or stress associated with anthropogenic management actions (e.g., castration, disbudding, shearing); an ability to engage in species-typical foraging, reproductive, play and other social interactions as well as partner choice; opportunity to express prolonged maternal care, natural weaning and complex social behaviours; positive emotional states from social interactions, control, agency and autonomy, and an ability to develop competency and overcome challenges.

Different individuals will experience things differently: some will die; some will lose body condition to the extent of compromised future survival (e.g., reduced immune capacity leading to greater impact of parasitism, weakness may lead to physical injury); some will survive until next season’s vegetation growth. Balanced against this is the opportunity for animals to experience a life where they have agency, can make choices and express social, environmental and dietary preferences.

The very young animals, less than 1 year old, and the older animals (7+ years) have much greater mortality than do prime-aged adult animals and are thus those that are most likely to experience poorest welfare. This can be separated into mortality around birth, which is the greatest risk, and mortality in the first winter for young animals. Lambs and kids dying around birth (in farmed populations, where data exist, these appear to parallel wild and naturalised populations) do so due to a combination of starvation and hypothermia, which can be particularly acute in poor weather. Neonatal small ruminants are born with limited body reserves, particularly brown fat which is activated at birth to generate heat, and these can be rapidly depleted if neonates do not stand and suck sufficient colostrum to maintain thermogenesis. As a young animal with little muscle mass, shivering is ineffective at maintaining body temperature. The biological processes of hypothermia and starvation have been well documented, but an animal’s experience of this is less well understood. In neonatal animals the period from birth until they succumb can be short, e.g., a matter of hours, as the young animal lies down and ceases to seek the udder. In human patients with hypothermia, this is similar with loss of consciousness and coma before death. For juvenile animals dying in their first winter, this may be more likely to be related to starvation and a result of not gaining sufficient weight in the first year of life. This is a slower process of gradual starvation accompanied by organ failure. Over time this is described as uncomfortable pangs that is not well located in the stomach, headache, weakness, and apathy. In monogastrics, this is accompanied by hunger contractions which may involve the whole abdomen and be sufficiently painful to disturb normal sleep⁹⁹. Whether similar sensations are experienced by ruminants is not known. In humans, starvation is associated with feelings of apathy, fatigue and constant weakness¹⁰⁰, and it may be that animals undergoing prolonged starvation also experience similar states. This may be a cause of frustration and anxiety in prey animals. Over time, emaciation ends in feelings of extreme apathy. This may also be why dead animals were often found in cleitan or caves as they may have chosen to withdraw to these areas as starvation progressed.

Mortality from any cause is often a significant welfare issue as it will be associated with a period of suffering which, in a farmed situation, may be alleviated by treatment or timely euthanasia, which are not available to these animals in the current state. This needs to be balanced against the additional welfare costs experienced by farmed small ruminants, such as castration, tail docking, disbudding, weaning at ages earlier than natural weaning, manipulation of social structures, gathering, handling, shearing, inversion for foot trimming, marketing, transport, holding in a lairage, stunning and slaughter.

5.1.2 Euthanasia of suffering animals

The key welfare enhancement of this strategy is reduction in suffering associated with prolonged starvation or other health issues, and a reduction in pain, hunger, weakness or exhaustion by timely euthanasia. Welfare compromises are likely to be similar to the baseline scenario and, although there may be anthropogenic disturbance in identifying and culling/euthanising the target animals, this is likely to be short lived. There is a risk that some animals may be culled that might otherwise have survived and been able to participate in future positive interactions and experienced positive emotions. Both the SSP and members of the Soay and Boreray Sheep Society told us that Soay sheep can appear to be close to death in some situations and yet survive and remain in the population for many more years. However, the number of euthanised animals that could have survived is likely to be small and overall we perceive euthanasia of suffering animals to yield a welfare benefit. In terms of Hume’s criteria, this should reduce both the severity and duration of suffering and should be practically feasible, at least for those animals which are readily accessible.

5.1.3. Support for the existing animals

We consider the welfare impacts of each of our three proposed interventions separately in this section. However, we acknowledge that it would be very unlikely that these would be effective alone and it might be relevant to consider that these must be applied with additional interactions, particularly some form of population reduction. For clarity of the welfare impacts we do still consider these as separate responses in this section.

5.1.3.1 Provision of supplementary feeding

Here we consider that providing additional feed in the winter has welfare enhancements of improving physical fitness and body mass, which would lead to better resistance to winter weather, improved immune responses to parasites, increased feelings of satiety and should increase the production of milk and colostrum in pregnant females as well as increasing birth weights of lambs/kids. Interestingly, although in domestic sheep experimental undernutrition is associated with reductions in expression of maternal care¹⁰¹, home-range quality in a natural situation is not reported to be associated with improvements in offspring survival⁵⁴. It is likely that increased nutrition will increase recruitment of juvenile females into the breeding population, as this is body mass dependent, and may also increase rates of twinning in females⁵². Thus, it may be that improved nutrition leads to more lambs/kids being born, but an equivalent absolute number of neonatal animals not surviving. Welfare compromises associated with provision of supplementary feed relate to the potential for injury and frustration from increased competition for supplementary feed and concerns that this will be a short-lived benefit as it will increase fecundity, winter survival of juveniles and hence lead to an increase in population size and density unless this is also matched to some method of population reduction. As food is likely to only be available to some animals, since there are practical constraints on providing food to some populations, this may lead to skewed social dynamics and increase population instability. Currently, for the Soay sheep there have been no large reductions in population size since 2012 and the population appears to have become more stable. Interventions that increase survival and hence increase population density beyond the carrying capacity of the area may have the paradoxical impact of increasing the likelihood of larger losses in some years. We note also that improved nutrition will likely only increase survival in the juvenile cohort: adult animals already have very high survival and older animals will still reach an endpoint, where they may not be able to compete effectively for additional food sources with younger and fitter animals.

5.1.3.2 Provision of health treatments

We consider that providing health treatments to the animals (typically parasite control and vaccination against key endemic diseases – this may only be relevant to those populations, which mix with farmed ruminants rather than truly isolated populations, where they may not be exposed to pathogens) would lead to the welfare enhancements of better physical fitness, improved food utilisation, a reduction in the duration of suffering associated with diseases and injury (as the animals will need to be gathered and can be inspected at close quarters and treated), and possibly improved gastrointestinal comfort. Evidence from previous experimental anthelminthic treatments on Soay suggest this would improve neonatal lamb survival, but may not affect juvenile growth and development. Welfare compromises associated with this approach include: increased competition for food resources and shelter if more animals survive; fear, panic, risk of injury and pain from the regular need to gather the animals; risk of social disruption; and potential impacts on antimicrobial resistance and resilience to worm burdens in populations that may already be well-adapted to the current parasite load.

Similar arguments to those presented above for provision of supplementary food are relevant, if a larger population survives because of this intervention without some form of population reduction. In addition, since this intervention relies on capture and handling of the animals, there is a greater risk of only some animals receiving health treatments and that the population becomes skewed towards those that are more readily captured. The fear, panic and injury risk of regular capture could cause some animals, e.g., on Hirta, to abandon current areas where trapping might occur, which would increase pressure on other areas, reduce the effectiveness of this option and may lead to other unintended welfare costs. Anecdotal observations of some animals becoming ‘trap-shy’ and leaving Village Bay when the summer gather happens has already been made by SSP, and this is also seen in wild sheep populations where only the bolder or tamer animals can be captured¹⁰².

5.1.3.3 Reproductive management

For this intervention the main welfare enhancement would be a reduction in juvenile males entering the rut, and hence improved survival (if this intervention is targeted at males) and potentially a reduction in juvenile ewe pregnancy (if this intervention targets females, but we note that this does not appear to be currently possible). Therefore, welfare enhancements are most likely to be focused on improving juvenile survival associated with better physical fitness and lower risks of starvation from not participating in the rut. Inducing sterility in juvenile males is unlikely to influence rates of pregnancy in juvenile females since the vast majority of matings are achieved by older males. If physical castration is used, this is associated with pain (shorter or longer term depending on the methods used⁵⁹) and the fear, panic and risk of injury associated with frequent or repeated handling to administer immunocastration (as well as some means of identifying and monitoring animals that have been enrolled in the immunocastration process and the need to repeat this probably annually). Physical castration, as well as the pain associated with the procedure, permanently excludes some males from engaging in some aspects of social behaviour, which may be positive, such as courtship. Permanent removal of animals from the breeding population would also require some anthropogenic decisions about who could breed with concomitant implications for future fitness of the population and may lead to unintended consequences in terms of welfare.

Although this intervention may hold some promise for welfare enhancement in the future, particularly short-term or reversable controls of fertility, we do not currently consider this to be practical or feasible for managing wild-living populations without much greater control of populations than is currently the case. In the future, with further developments in products or approaches, this may be an option to reconsider to limit juvenile mortality. We note also that, as with all interventions under this heading, this would require continuing activity to manage populations.

5.1.4 Population reduction

This approach requires the permanent removal of some members of the population from their current location in various ways. Where animals are culled, we assume that these are most likely to be shot, rather than caught and euthanised, and that, ideally, this will be done by competent and trained professionals capable of killing the animals quickly, so we have not considered the welfare of animals killed but only that of the remaining animals. Where animals are removed to a new location, we have considered their welfare and those of the animals left behind.

5.1.4.1 Annual culling of a proportion of the flock/herd

For the annual or semi-annual culling of a proportion of the population, we considered the welfare enhancements for the animals that are not culled to be greater access to the resources available in the environment and the potential to have improved physical fitness. If the animals have better nutrition, then they would be expected to have better body mass reserves for the winter, better resistance to parasites and a greater probability of successful breeding. The selection process should target specific groups to maximise positive effects in the remainder: aged; injured; young in poor condition. However, it is possible that sheep available for culling are not representative of the whole population and this process may remove many animals that would have survived and thrived in the environment, without removing those animals that are more vulnerable, particularly aged animals.

From the perspective of welfare compromise the process of culling animals could, depending on the way this is done, be accompanied by significant welfare costs to the animals of fear, panic, risk of injury or death from flight, and severe pain from surviving mis-shots or ricochets. Unless there is a very careful selection of small groups of animals, removal of individuals can disrupt social and family networks and impact on complex hierarchies and cultural knowledge. Where key adults (typically older females in matrilineal groups) are removed indiscriminately, there is a risk of the loss of cultural information, which can reduce the ability of the remaining population to access resources efficiently, potentially compromising survival. We were given anecdotal evidence of high impact culling causing marked disturbances to social behaviours, habitat use and kidding behaviour in feral goats, which led to an increase in neonatal mortality following culling. Although closed seasons (periods when it is forbidden to shoot deer) apply to the killing of female deer in Scotland¹⁰³, there are no such restrictions on killing of other ruminants living in a functionally wild state. Thus, it is possible that culling may occur for animals that are pregnant and raising dependent offspring. There is increased risk of prolonged suffering from starvation and anxiety of young animals, if their mothers are killed before they are both nutritionally and psychologically weaned, in addition to the other social costs of removing females from the wider social group.

In this scenario we considered that animals may be shot in the field, although it is also possible that animals could be herded or gathered and then selected for euthanasia. This is more commonly carried out on North Ronaldsay and was also described for the goats living in the Cheviots. In these situations there is a focus on conservation and management of the environment as much as any concern for the welfare of the animals. In these cases, although the welfare costs of being gathered remain, there is more scope to carefully select animals to remove to preserve social structures as far as possible. However, these actions are really only suitable for the removal of relatively small numbers of animals and, for example for the islands in St Kilda, the numbers of animals that would need to be removed to achieve any significant population reduction would be significant.

It is also relevant here to address the potential unintended welfare consequences of significant reductions in population density. For example, it was suggested to us that the population should be reduced to about a third of the animals on Hirta (approximately 600 animals). However, there is evidence (as presented above) that not all mortality is population-density specific, that crash years occurred when the population was far below what it currently is and there is a risk of inducing greater population instability, which may lead to further periods of ‘boom and bust’. As animals can adapt quickly to the increased food intake available (as described above for supplementary feeding), this level of heavy culling would need to be maintained to continually suppress the population. There are thus very significant risks that in some environments a process of heavy culling of animals may lead to a net increase in welfare compromise compared to the baseline scenario.

5.1.4.2 Annual trapping for removal of a proportion of the animals to be rehomed

The positive welfare enhancements for this scenario would be the same as those outlined above for removal of a portion of the population by culling for those animals not removed, such as improved access to scarce resources and a potential improvement in physical fitness. The welfare compromises would also be related to the risks of injury when fleeing capture; fear, panic and anxiety from the human disturbance and disruption of social networks and social dynamics by removal of individuals. Issues and concerns around which animals are removed, and the impacts on population fitness and population instability are also relevant here to the welfare of those animals that remain after a portion of the population has been trapped and removed.

For this scenario there are also the additional welfare issues surrounding the removal of the animals from the environment and where they are then placed. For all removed animals there are the welfare costs of fear and anxiety associated with gathering and handling in animals unused to human contact; fear, inanition, risk of pain, injury and anxiety, as well as possible separation from important social companions, if the animals are confined before removal; the risks associated with prolonged transport either at sea (for Soay sheep) or by road for other animals – these include prolonged periods of hunger and thirst, if they are not provided with food or water during the journeys, or if they are too anxious to eat or do not recognise the offered food as something to eat; risk of injuries from falling during movements of ships or trucks or fighting when crowded together; fear from close human contact and unfamiliarity (novel environments, novel sounds, novel odours, etc.), nausea, etc. These welfare impacts of transport have been well described for domestic sheep¹⁰⁴ and are likely to be exacerbated for feral animals, as they have less experience of close human contact. There may be welfare enhancements when the animals ultimately arrive at their final destination, which would be dependent on the new location meeting the animals’ physical and behavioural needs, and assuming they are able to be rehabilitated and can be supported through any disease or environmental risks.

For situations where small groups of animals can be removed, particularly in family groups, and to an appropriate location, this may bring new opportunities for positive welfare to the removed animals from better access to resources and allowing animals to exploit new locations. In these situations any negative impacts of removal and transport will likely be short term and can be overcome by the positive welfare benefits. Where larger numbers of animals are required to be removed, we have concerns about the feasibility of finding a suitable location, which will meet the welfare needs of the animals and exceed their current welfare state. The practical issues surrounding rehoming of large numbers of animals are likely to be very significant. The inability to preserve family structures, when large numbers of animals are moved, and whether this will represent a better welfare situation than the current state are also significant concerns, as well as the welfare compromises of removal from the current location.

5.1.4.3 Annual darting of animals for removal of a proportion of the population to be rehomed

Most of the welfare concerns and enhancements in this scenario remain the same as discussed for 5.1.4.2, with the exception that darting the animals may be better for their welfare than the use of traps and gathers to capture them. However, a comparison study of darting or use of drive nets for capturing desert bighorn sheep suggested that 23% of darted sheep died as a result of the capture process, compared to 2.2% of sheep captured by netting¹⁰⁵. In this case the animals were herded by helicopter to what was perceived to be a safe location and then darted from within 10 m. Causes of mortality were physical injuries and imprecise dart hits for 43% of mortalities and capture myopathy for 38%. A more recent review of capture of wild deer in Australia, equivalently small and relatively fragile species, also raised concerns about efficacy and welfare impacts of different techniques to capture small ruminants¹⁰⁶, including darting, oral baiting, or trapping and netting. Given these uncertainties and possible very significant negative welfare implications of darting, we do not consider that this is an appropriate method to use to reduce populations in naturalised small ruminants in Scotland.

5.1.5 Managing the populations in situ

If naturalised sheep or goat populations were managed in line with current practices for normally domesticated populations of these species, we considered the welfare enhancements to be improved food provision, improved health care and perhaps some additional protection from poor weather. This could lead to improved thermal comfort, vitality, feelings of satiety and gastrointestinal comfort. The welfare compromises we envisaged include potential loss of dietary choice and increased food competition, if animals are confined; loss of agency, autonomy and habitat selection; short-term risks of injury relating to capture and experiences of fear, panic and frustration associated with adaptation to confinement and close human contact; pain from farm management practices associated with castration or disbudding (if required); possible disruption of hierarchies and social networks; a reduced ability to choose preferred social companions; restrictions on mate choice; restrictions on the length of maternal care and social contacts and pain, fear, anxiety and loss associated with the removal of social partners at various points. There would also be welfare costs associated with reproductive management (e.g., pain of castration) and removal of animals (for slaughter for meat or other purposes), which have been discussed above.

Evaluation of the impacts of this strategy should include the likelihood of improving welfare and the requirement for significant and possibly permanent presence of humans in some locations (Hirta) to allow these interventions to be carried out. There are also welfare implications of trying to turn animals not used to human handling and confinement into ‘farm’ animals, which may not be considered an acceptable welfare compromise⁶. The evidence from farmed hill sheep, as described above, suggests that mortality rates of up to 15-25% are common in lambs to weaning, a figure that is not different to the 27% average mortality of Soay lambs, although in some cases the upper values may be much higher by year or farm. Mortality of juvenile Soay sheep is high (overall 64% of females and 70% of males do not reach their first birthday). Comparatively almost no farmed males and 50% of females may not survive to a year old in a farmed situation either, but these animals are killed by humans rather than dying of natural causes. The welfare costs of dying of cold weather, starvation and parasite load need to be weighed against dying in an abattoir after capture, transport, confinement in a lairage and, ideally, stunning before slaughter, as absolute numbers of dead animals are not a useful metric. Thereafter, the survival of adult ewes is almost identical in a farmed or naturalised situation for females until they reach around 7 years of age. Older females are typically culled in a farmed setting, whereas in the naturalised populations they may begin to have lower survivability as food competition, pregnancy, winter weather and parasite load have an impact. However, there is some evidence that Soay sheep can still be reproductively active in old age, and there are no obvious changes in mobility and behaviour until animals are very old. Older males also show similar declines in survival, typically dying at a younger age than females. There is no good comparative data in farmed sheep for rams, although it is evident that males have poorer survival than females in every age cohort. Equivalent data for goats are patchy and hard to find, but, where they exist, support a similar conclusion.

Overall, one of the key questions is an ethical one: is it better for the animals to die in a functionally wild state of starvation, winter weather and disease load, or to die at human hands following transport and removal to an abattoir, typically at a younger age in the managed state? Other interventions do not appear to achieve an overall positive welfare balance to make this approach better for welfare than the baseline ‘hands off’ state.

5.1.6 Removal of all animals from the current location

This intervention should only be considered, if the welfare state of the animals in their current location is considered to be unacceptable and unlikely to be improved by any management intervention. For decisions made on the basis of other goals (such as conservation, avoiding environmental degradation or achieving other human-desired traits) then we consider that the 7 Principles of Ethical Wildlife Control¹⁰⁷ should first be applied to ensure that this is an appropriate course of action. Removal of all the animals in some locations (e.g., St Kilda) will impact on native wildlife, such as the St Kilda field mouse, Apodemus sylvaticus hirtensis, which utilise sheep carcases as a food source. These may be forced to switch to eating seabird chicks, with unintended conservation and welfare consequences for other species.

5.1.6.1 Removal of the whole population to another site

The welfare compromises for this approach are the same as those outlined under 5.1.4.2 for removal of a proportion of animals to another location, although the impacts on the residual population do not apply (or may only apply for a short period, if this needs to be achieved over a number of years). Thus, compromises would be largely associated with the fear, anxiety, panic and potential injuries associated with the capture of the population and welfare issues associated with transport to the new location. There may also be reduced opportunity for choice or expression of preferences (of diet and environmental resources) and risks associated with exposure to new pathogens, for which the animals may have no acquired immunity. Possible disruptions of social networks, hierarchies and family structures may also occur. Welfare enhancements are very dependent on the nature of the environment to which the animals are removed. Possible enhancements could include access to better food quality and provision, access to veterinary treatments, greater protection from winter weather, opportunities to explore new environments and reduced competition for food and other resources. These are more likely to be achieved when removing relatively small populations (e.g., some of the goat herds), where social structures can be maintained and new locations can be more readily found. If moving large populations, such as the complete removal of all sheep from the islands of St Kilda, it is impractical to expect they can be moved to a new location in their entirety, or even in functional units of families and social partners, and thus there will be some considerable social disruption. It is also worth noting that removal of the animals will not solve the issues of population density and other interventions, as described above, would need to be applied to the animals in their new location.

5.1.6.2 Culling of the entire population

The most radical intervention, and probably the least socially acceptable, would be the culling of all animals in the location. This may also be constrained by the practical issues of whether it would be feasible to do this in all the remote locations that small ruminants are found and the time scales over which this would need to be done. The welfare compromises outlined 5.1.4.1 would apply to all the culled animals, and the impacts on the animals not immediately culled would persist until the entire population had been killed. There are no welfare enhancements that would apply to this scenario, but also no unintended welfare costs for the small ruminants, thus there is greater clarity around the welfare impacts of this approach than any other situation. Once all the animals are dead, there is also certainty that there would be no further welfare impacts.

5.1.7 Section summary

In our review of the welfare implications of the different scenarios we have identified, we conclude that only one scenario would result in an overall net welfare benefit: the increased use of euthanasia for animals identified as clearly suffering. This would not impact on the number of animals likely to experience the welfare compromise of starvation, injury, or disease, but could be unambiguously shown to reduce suffering of individuals, without impacting on positive welfare enhancements for other animals. It is likely that this will be more easily achieved for more accessible populations, but approaches to facilitate this in more remote areas could be considered. We also consider that there may be some promise in the use of fertility control (particularly of females using reversible methods) that may have welfare value, but for which there is not yet a clear and justifiable approach that could be implemented for the populations we are considering. We have also identified some scenarios where the welfare impacts are clearly very significant and should not be attempted: specifically darting animals to remove from the location. For all other scenarios there are uncertainties around the welfare impacts and risk of unintended consequences that make it hard to conclude they will result in an overall improvement in welfare and significant risks that welfare could be reduced. Some of these are context specific: for small populations of animals (particularly goats) in relatively accessible locations (i.e., those not on islands) then the opportunity to move some or all of specific social groups to a new location should also have welfare benefits that may outweigh the welfare costs of the intervention. These are not considered viable approaches to larger groups of animals or where the need to undertake a long sea journey is required.

5.2 Practicalities of different approaches

Although in the preceding section we considered each scenario separately, it is clear that some of those that require care for the animals (provision of supplementary foods and health care) will lead to an increase in population size and must therefore also be paired with some form of population reduction to ensure that a new welfare problem is not inadvertently created by the intervention by increasing numbers of animals. Provision of care has practical issues in terms of feasibility and the likely need to have permanent human presence in the location. However, as this needs to be paired with some form of population control to be successful, we have focused on this as the main area of constraint for practical reasons.

The alternative approaches for managing population size would be pharmaceutical control of female fertility, or removal of animals to an alternative location or lethal control. We have discounted control of male fertility here as it is unlikely to be practically successful as even a few males left fertile could prevent any meaningful reduction in population size. For small ruminants, with a promiscuous mating strategy and high male mobility between female hefts, population control is likely to rely heavily on control of female fertility. In zoo populations it is possible to limit female fertility through pharmaceutical means, but this is a different scenario to large-scale population management and thus there needs to be additional research and potentially technological developments before this may be practically possible. Evidence that there are not long-term impacts of contraception for the female or any subsequent offspring is also required. There are some current trials in more accessible populations (e.g., feral goats in Llandudno, Wales) that may yield useful results. However, some of the vaccine products used in international studies are not available in the UK.

As discussed⁶ and in part above, the feasibility of the other options for population management is dependent on context. Although capture and movement of small populations of relatively accessible goats in the Southern Uplands seems feasible, and has already been practised in the Cheviots, this is quite a different prospect for island populations. Capture, confinement and shipping from remote islands is likely to be impractical, incur significant risks to human and animal safety from the more remote islands (Soay and Boreray are barely accessible at all) and be prohibitively costly, even without the associated welfare costs of attempted capture, disruption of social relationships, transportation and forced human contact. Although difficult and challenging, and likely to incur very significant issues with social acceptability, culling of all animals is more practical than attempting to remove animals from remote locations. However, it may not be possible to remove all animals, and certainly not within a short space of time, and this is likely to require continuing efforts over a protracted period time.

Although not the focus of the review, we also note the loss of cultural, heritage, genetic and scientific values of the animals, if entire populations are removed from areas of historical association.

5.3 Ethical review

In this section we consider separately our ethical responsibilities to the populations under examination from the welfare impacts of any intervention. A number of different ethical considerations have been raised in this work, some of which have been helpfully evaluated already by Gambourg et al.⁶ in their review of what is owed to feral populations. These include the significant suffering associated with starvation; the historical reasons why animals are in these situations (essentially because humans placed them there), which suggest some aspects of human responsibility; the condition of wildness or naturalness often embodied in these populations; negative environmental damage caused by the animals (this relates most to the goat populations, although possible impacts of Soay sheep on seabird populations has also been suggested); conservation value through the unique genetic resources represented by heritage populations, and the scientific value of studying these populations. These authors⁶ argue that there are different values and concerns at stake when considering feral populations, not all of which relate directly to the welfare of the animals themselves. Here we consider a number of ethical issues that seem relevant to us and were raised in our investigations or were specifically requested that we consider.

An important feature of our investigation of the different species has been that the animals have found themselves in their current locations due to human activity. Neither species of small ruminant evolved on the British Isles and both were brought here around the same time in prehistory. This suggests two relevant considerations: firstly, that it seems anomalous to refer to one species (goats) as an invasive, alien species and where the dominant activity is eradication, whereas sheep are treated as domestic or feral animals to be protected. The welfare needs of both species are not affected by these differing designations, but often what is permitted in their treatment is. Secondly, both species are in their current locations due to human activity. Even if the animals were placed in their current locations by populations of humans who are no longer alive, a relevant ethical question is whether we still retain some responsibility for their welfare.

The nature and extent of our responsibility to feral animals, and indeed to other species that have also been brought to these islands, but are now considered as ‘wild’ (e.g., rabbits) is complex and may be related to both our categorisation and views of these species, their biological adaptability to the environment and a wider understanding of our impact on animals more generally, including truly wild populations. For example, most populations of wild animals are impacted by our activities, whether that is historical eradication of large predators (which impact on population dynamics of large herbivores), introduction of new ones (e.g. domestic dogs), fencing, reintroductions, culling or other impacts. In our previous SAWC report¹⁰⁸ on animal welfare risk assessment guidance for wild animal translocations we considered the relevance of human responsibility to those populations. We considered there that this should be evaluated based on the ability of the animals to live independently in their new environment, and human intervention may reduce over time so as to be minimal or absent. It should be noted that even at the lowest level of human responsibility and intervention, we still considered there may be a need for euthanasia of injured or diseased individuals and for welfare risk assessments to be conducted. Even where animals may not be directly managed there is a responsibility on the landowners to demonstrate stewardship, and this should include consideration of the welfare of the animals living there, as well as any other goals such as conservation. As we discuss above, there is an argument that this is a greater responsibility for feral animals or those moved by human intervention to the current location than for truly wild animals which have evolved in situ.

A key concept in the translocation of wild animals is the ability of the animal to live independently in the environment. This could be seen to be relevant also to feral animals, where the need for human responsibility for welfare may be related to the biological adaptations of the animal to the current environment. For example, moving an animal to a location where it does not possess suitable environmental adaptations to survive (e.g., moving an animal adapted to a particular nutritional diet to an environment without those foods, or a flightless bird to an environment with predators) would require continual human intervention to ensure good welfare and it is unlikely that that responsibility could diminish over time. However, if animals are moved to a similar environment to where they evolved, and where they are able to adapt and flourish, then the need for human intervention is likely to be less and may be considered to be negligible over time. Biologically, feral sheep and goats in the British Isles behave similarly to wild progenitors and appear to be able to adapt and utilise the environment in a way that suggests they are able to live independently. Their survival and longevity reflect those of wild populations, and (for the comparison of unmanaged mortality where the data are present) appear to have similar survival to domestic animals of the same species. Therefore, there does not seem to be a biological limitation to their ability to survive without human intervention in most environments. However, a relevant consideration is that some of these populations are kept on islands and are unable to disperse or seek new areas if the environmental conditions are harsh. It is relevant to note that all populations are constrained in some way – either by social behavioural constraints that prevent populations spreading (such as hefting), by intraspecific competition (e.g., feral goats on Rum), by geographical features (e.g., islands, mountains, rivers etc) or by human-made features such as roads or housing. Nonetheless, we consider that movement of some animals to island habitats, which they cannot leave, does present an ethical need to have greater consideration for their welfare than in other environments.

A key component of our consideration of the welfare of animals in this report is to evaluate the wider aspects of welfare, which go beyond the prevention of suffering. Suffering in wild and feral animals is an ethical concern, and we accept that dying from starvation is severe and unpleasant. From a biological functioning view of animal welfare this may appear to be the greatest concern and a need to prevent suffering to be the over-riding issue. This is a particular ethical view that prioritises the alleviation of suffering as the most important aspect of welfare. However, a wider view of animal welfare also needs to consider the more positive aspects of an animal’s life, where freedom, choice and agency are also valued by animals, and thus a balance of preserving these important attributes of the life of an animal needs to be considered alongside a desire to prevent suffering. This also includes an ethical consideration of the risk of future harms (such as impacts on future adaptability) by attempts to reduce the suffering of the current population. We were cognisant of the risks of imposing an anthropocentric preference on the animals whilst not giving full weight to animal agency through the decisions that the animals themselves might make. Undertaking this cost-benefit analysis of a range of positive and negative welfare impacts (a number of which we have noted above) is hampered by the lack of real scientific data to assess this and therefore an ethical view on the potential risks and likelihoods of welfare harms compared to welfare benefits remains challenging.

From a welfare point of view, the welfare needs of a species, and our ethical responsibilities, do not differ because we might consider or value that animal differently, even though the legal position might differ. Thus, we find it unhelpful to apply labels to animals (pest for example), which are often interpreted to allow welfare harms to some groups of animals to be permitted and which would not be allowed for other species. In this case, ethically, we do not agree that culling of goats should be permitted under conditions which would not be allowed for wild deer (during the breeding season for example) simply because goats are considered invasive and non-native. From a welfare perspective, the risks of harms to dependent young are identical and thus a similar consideration should be applied.

A final ethical consideration is the accessibility of some populations under consideration and the inaccessibility of others. In considering the different scenarios we presented above, it is clear that some interventions can be more readily applied to some populations than others, and indeed that the level of welfare concerns expressed are somewhat population dependent. We accept that it may be problematic to apply some proposed welfare interventions to all populations considered, and that the welfare implications of attempting to carry out some interventions may differ. This is a pragmatic position that reflects the risks to human safety as well as the inability to mitigate all welfare hazards, alongside the desire to alleviate some suffering where this is possible to do so.