The potential of virtual fencing for cattle and goats with regard to learning behaviour and animal welfare
Doctoral thesis
Date of Examination:2025-09-19
Date of issue:2025-10-16
Advisor:Prof. Dr. Johannes Isselstein
Referee:Prof. Dr. Johannes Isselstein
Referee:Prof. Dr. Imke Traulsen
Files in this item
Name:Ediss_Wilms_ohne CV.pdf
Size:2.67Mb
Format:PDF
Description:Dissertation
Abstract
English
The modern fencing technology ‘virtual fencing’ is becoming increasingly popular worldwide because it promises to make farmers’ work easier. Instead of actual physical fences, the virtual fence lines are invisible to the animals and are indicated by acoustic signals emitted by collars that the animals must wear. The fence lines can be modified remotely by the farmers at any time via an app, making the system flexible. This requires a stable mobile phone connection and a GPS signal, which also allows the farmers to locate and monitor their animals. Similar to electric fences, an electric pulse acts as a punishment if the animal does not turn away in response to the acoustic signal. To ensure that the animal remains within the virtual fence and does not receive permanent electric pulses, it must learn to associate the signals and react appropriately to the acoustic signal. Although many studies have proven that cattle, sheep, and goats are generally capable of this learning process, concerns still exist regarding the welfare of virtually fenced animals. The present thesis addresses these issues from various perspectives. In Chapter one, a systematic review was conducted, analysing the results of virtual fencing studies with cattle regarding the animals’ learning process and welfare. It was demonstrated that cattle learn to avoid receiving electric pulses as the calculated ratio of the number of electric pulses to acoustic signals decreased significantly over time. Also, the welfare parameters examined revealed no significant differences between the virtually and physically fenced cattle that could be reliably attributed to the virtual fencing system. Still, due to the small number of studies at that time, further investigations were recommended. Chapter two describes a virtual fencing experiment with a focus on the learning behaviour and welfare of goats. In a cross-over design, twenty ‘Blobe’ goats with offspring were either virtually or physically fenced for 12 days and underwent a virtual fencing training protocol by Hamidi et al. (2022). Observations of the goats’ basic behaviours and faecal cortisol metabolite measures as stress indicator were conducted to evaluate welfare. Except for higher values of standing vigilant in virtually fenced groups, no differences were found between the fencing systems. In addition, the development of the numbers of acoustic signals and electric pulses indicated successful learning at herd level. Overall, the results were promising, but further studies are needed to find out whether virtual fencing can negatively affect the welfare of goats. Based on the concerns on the welfare of virtually fenced animals and the legal hurdles of virtual fencing in many European countries, Chapter three deals with the potential of acoustic signals emitted by virtual fencing collars. In a first experiment, different acoustic signals and vibrations were tested to find a potential replacement for the electric pulse. However, none of the signals tried were efficient enough to change the behaviour of grazing cattle sufficiently and in the long-term. The second experiment was based on the idea that the movement behaviour of cattle on pasture could be influenced to improve grazing management. Although various training approaches were tried, cattle could not be taught to adjust their direction of travel in response to an acoustic signal while walking to a feed source. Finally, it was recommended to test different signals and a different training protocol for the first and second experiment, respectively. Overall, the results of the three chapters in combination with the insights from the General Discussion demonstrate that at least cattle can be virtually fenced, as they are generally capable of the necessary learning process and no clear disadvantage for their welfare has been proven in the studies conducted so far. For this animal species, legal changes should be considered where virtual fencing or the use of electricity on animals is prohibited, and official regulations are introduced in countries where virtual fencing is commercially available in order to ensure animal welfare. At the same time, research into a potential replacement for the electric pulse should be continued to have a current-free alternative for individual cases. Since the current research for goats is still limited, further studies are needed, especially to assess the welfare of virtually fenced goats. Finally, virtual fencing brings many benefits for farmers, animals, and the environment, and has the potential to replace physical fences in the future under certain conditions.
Keywords: Virtual Fencing; Animal welfare; Learning behaviour; cattle; Precision livestock farming; goats
