What are virtual fencing and virtual herding and do they affect animal welfare?

Virtual fencing works by using a GPS-enabled collar, which is fitted to the animal’s neck. The collar uses sound (audio ‘warning’ cues) and electric shocks to control animal behaviour. As an animal approaches a pre-determined virtual (invisible) barrier which has been set using an app, the collar emits warning sounds. If the animal continues towards the virtual barrier they receive an aversive electric shock [3,4]. The virtual fencing system requires that animals respond appropriately to the sound cue (i.e., stop or move in a different direction) to avoid receiving an electric shock.

Virtual herding is the use of virtual fencing technology to move animals from one location to another using an additional cue such as vibration. The vibration cue signals to the animal that they should move, while sound cues provide directional guidance during herding. If the animal does not respond in the desired way to the vibration or sound cues, they receive an electric shock. Virtual herding adds complexity to what animals need to learn in order to avoid shocks, compared with static virtual fencing for containment [5]. There is very little research published on the welfare impacts of virtual herding. One study found that during the training period, 70 per cent of shocks were experienced during herding rather than virtual fencing containment, despite herding presumably taking place over a much smaller time frame than containment [5]. This is likely due to the cows being herded (moved) having to learn what vibration as well sound cues mean and respond appropriately to both, whereas in a virtual ‘fencing’ situation, they have to learn the meaning of just one cue. Further complexity is added because animals have to keep moving (i.e., responding to the vibration or audio cues) until they reach the desired location (e.g., through a gate and into another paddock or to the dairy for milking). Currently only one commercially available virtual fencing system is capable of active virtual herding.

Virtual fencing has the potential to improve pasture and environmental management and reduce labour and costs associated with fixed fences in some situations. From an animal welfare perspective, if physical fences are removed, this could reduce wildlife fence entanglements and facilitate easier wildlife movement. There is also the potential for virtual fences to improve human and animal safety during emergencies, for example by allowing fences to be disabled remotely or by guiding animals to safer areas. 

However, the use of electric shocks to train animals raises significant ethical and animal welfare concerns. For example, is it acceptable to introduce new technology that uses electric shocks to control animals, particularly when best practice animal training principles caution against their use? What are the potential welfare implications to animals who don’t understand how to avoid shocks or who learn slowly? Does wearing the collars impact animals’ mental/emotional state? And how are animals impacted when the technology fails or when mistakes are made? 

While the intensity of the shock delivered by virtual fencing collars is typically less than the shock that electric fences are set to deliver [6], it is unknown whether these are directly comparable in terms of how the animal experiences the shock. Electric fences are visible and the animal can move away from the source of the shock, while virtual boundaries are invisible, only identifiable from sound cues, and animals are unable to move away or escape from the collars. For both electric physical fencing and virtual fencing to be effective, the electric shocks must be sufficiently aversive (unpleasant) that animals actively try to avoid them. One study found that sheep showed more avoidance behaviours towards an electrical stimulus than they did to a barking dog, and that behaviours in response to the electrical stimulus were more extreme, such as rearing, jumping or falling [7]. Reported reactions to virtual fencing electric shocks in cattle include running, jumping, head throwing and full body twitches [8], however other authors have reported that running, lunging and head shaking were not observed during virtual fencing use [9]. 

Some experiments have found that cattle confined behind a virtual fence spend less time lying down than those confined with electrified tape [6, 10], which could indicate that virtually fenced cows are less relaxed or comfortable. However, other small-scale studies have found no differences in lying time between virtually fenced and physically fenced cattle [9,11]; more research in this area is needed. 

During the initial phase of training, animals do not know what the audio cue means and therefore cannot avoid receiving the electric shock(s) [4]. This could lead to anxiety and psychological stress during the learning period [4]. Learning is successful when an animal is able to understand the virtual fencing cues and control how they interact with the technology, which generally means learning how to avoid electric shocks. Studies have reported that cattle generally learn to respond to the audio cue alone after one to six electric shocks or within hours to days, but high individual variation has been observed in animals’ rates of learning [1,3,5,6,12]. 

Increased motivation to cross the virtual boundary can increase the risk of receiving electric shocks and thus negatively impact welfare. Increased motivation to cross the boundary can occur when there are resources outside the boundary that animals want to access, for example feed, especially when feed within the boundary is limited [13], or cows walking to their uncollared calves who have wandered outside of the virtual boundary [14,15]. 

Other welfare concerns associated with virtual fencing relate to the risks of equipment malfunction [1], irritation and/or ulceration of the skin due to poor collar fit, and deliberate or accidental misuse.

While scientific research on virtual fencing has been underway for a number of years, there are still gaps in the knowledge about how virtual fencing systems impact animal welfare. Research is needed on: 

• the impacts of collar use on animals’ mental/emotional state
• the long-term animal welfare impacts
• the effects of different collar and software designs on animals
• welfare impacts of virtual herding compared to low stress handling techniques
• impacts of collars on cows with uncollared calves at foot, and
• validation of the health and behaviour alerts that are built into some collars.  

In addition, many currently published virtual fencing studies have small sample sizes and are conducted under controlled conditions – more research is needed on bigger sample sizes and under commercial conditions. Independent research is also needed to substantiate manufacturer claims about the animal welfare benefits of virtual fencing technology. 

State and territory animal welfare legislation determines where and what type of electronic devices can or cannot be used to contain livestock. Virtual fencing collars are now permitted for commercial use on cattle in all states and territories except the Australian Capital Territory, but some jurisdictions have conditions governing their use. For example, in New South Wales, users of virtual fencing technology must undergo training before use and must only activate virtual fencing collars if they are present on the same property as the cattle wearing the collars. In Victoria and South Australia, conditions of use include requiring manufacturers to obtain government approval to market virtual fencing systems, users and cattle being trained on system use, and cattle being at least six months old when wearing virtual fence collars. In Western Australia, only one brand of virtual fencing collar is permitted for use on cattle. The Northern Territory allows virtual fencing systems to be used on farmed animals when used in accordance with manufacturer’s instructions. Tasmania and Queensland do not currently have any specific restrictions on the use of virtual fencing systems for farmed animals.

New technology introduced for use in relation to animals must be subject to an ethical assessment that emphasises the importance of animal welfare before implementation. The use of technology must not replace the requirement for good animal husbandry and regular inspections of animals by humans. Any new animal technology with the potential to negatively affect animal welfare must be subjected to independent validation and animal welfare science research to ensure appropriately informed decision making. New technology must be regularly monitored for adverse animal welfare impacts and updated or removed from market where risks to animal welfare cannot be mitigated. Where new technology poses a risk to animal welfare, regulations must be developed which minimise this risk. 

The RSPCA supports the use of non-aversive methods (those that don’t require unpleasant stimuli, for example, plain wire non-electrified physical fencing) to fence and herd farm animals. Virtual fencing technology has the potential to cause injury, pain, and distress, for example due to poorly fitting collars, the application of electric shocks, the expectation of electric shocks, and individual learning differences. Where virtual fencing technology is used on farmed animals, it must: 

• be underpinned by animal learning theory
• be the subject of an extensive body of animal welfare research which indicates that negative animal welfare impacts of the technology are minimal in the contexts in which it will be used
• have built-in, effective safeguards to protect animal welfare
• be effectively regulated, including mandatory standards around the design, user training, use, and monitoring of the technology
• have regulatory oversight which is resourced sufficiently to allow appropriate enforcement
• include as a requirement of sale and use the recording and reporting of the animal welfare impacts of the technology. 

Until such time as there is extensive animal welfare research to indicate that the risk to animal welfare is minimal in the following contexts, virtual fencing technology must not be used:

• on species other than cattle
• on growing animals who are not closely monitored
• on deaf animals
• to actively herd animals
• to separate individuals or groups of animals out from a larger herd
• on cows with calves at foot
• without regular monitoring of collared animals
• without the provision of adequate food, water, shade and shelter within the containment area.