Behavioural response of dairy cows to different shock energies during virtual fencing training

Project details

Status: Completed

Project background

As virtual fencing technology becomes increasingly adopted in dairy systems, there is a need to understand how different electrical stimulus energies influence cow behaviour and welfare.

This project aimed to determine whether varying shock energy levels affect behavioural responses, learning outcomes, and grazing recovery following virtual fencing training, and to identify stimulus levels that support both effective livestock management and positive animal welfare outcomes.

Halter devices typically deliver a shock energy of 0.1 J. A small proportion of animals, however, require a higher shock level (up to 0.45 J) for effective management with virtual fencing technology. This project aims to better understand cattle behavioural responses to shocks delivered across that range.

Project activities

Rather than deliver a fixed shock energy to all cows, the Halter virtual fencing system customises shock energy at the individual cow level.

Cows are trained using a low shock energy of 0.1 J, which is increased after training is complete only for cows that fail to respond to the cues, or later become nonresponsive.

Forty early-lactation dairy cows were studied in 5 groups of 8 animals each. Cows were selected from the larger herd of ~500 cows on the basis that they were over 30 days in milk and had not been treated for lameness, mastitis, or metritis in the preceding fortnight. Within each group, cows were randomly allocated to receive one of 4 shock energies (0.1, 0.2, 0.3, or 0.45 J).

Behavioral responses to shock were recorded and classified using a 6-point reaction score: (1) stop, pause, or no response with no steps taken; (2) fewer than 3 steps; (3) more than 3 steps at a normal walking pace; (4) more than 3 steps at a fast walking pace and or head shake; (5) move away at a trotting pace; and (6) jump, vocalisation, or 360-degree turn. It was categorised as low to moderate being 1-3 and moderate to intense reaction 4-6.

The researcher remained blinded to shock energy allocation during behavioral observations and statistical analyses, with treatment identities revealed only after all analyses were finalised.

Project outcomes

This study found that most cow responses (84%) were mild and classified as low-level behavioural reactions.

Shock energy within the tested range did not significantly affect reaction score, the number of shocks received, or the time taken for cows to resume grazing after a stimulus. A key outcome was the identification of substantial individual variation between cows.

Individual animals explained 46% of the variation in behavioural responses and 34% of the variation in the time taken to resume grazing, indicating that cow-specific characteristics were more influential than shock energy level in determining responses.

The findings support the use of virtual fencing systems that deliver the lowest effective electrical stimulus for each animal, helping to maintain effective livestock management while supporting animal welfare outcomes.