The use of Accelerometer Technology to Predict Time Budget and Head Shaking Behaviour in Ewes Managed on Pasture
Year: 2026
Type: article
Abstract: This study aimed to evaluate a tri‐axial accelerometer attached to a collar to determine the ability to accurately predict rumination, lying, grazing, and head shaking behaviour of ewes on pasture. Collars with accelerometers recording at 1 Hz were attached to 30 Romney type ewes managed in groups of 5. The behaviour of each group was observed for 1 day between 0830 and 1630 h using 30‐s instantaneous scan observations. Head shaking behaviour was stimulated in the yards using a toy water pistol and analysed as present or not present. Grazing activity was well classified by a Random Forest algorithm using the out‐of‐bag error estimation with an accuracy of 83.4% (95% CI: 82.9–83.9%). The accuracy for ruminating activity was 82.6% (95% CI: 82.1–83.0%), and for lying activity 85.5% (95% CI: 85.2–86.0%). Grazing, ruminating, and lying activities were predicted with sensitivity values of 80.0%, 68.0%, and 79.4%, respectively. The specificity values for grazing, ruminating and lying were 86.6%, 87.6%, and 89.3%, respectively. The precision values for grazing, ruminating, and lying were 85.0%, 65.1%, and 81.7%, respectively. However, the accuracy to classify head shaking behaviour was only 52.2% with poor performance metrics potentially due to the low sampling frequency of the accelerometers in combination with a behaviour of short duration. The results show that accelerometers can monitor the behaviour of ewes on pasture; however, the sampling frequency likely needs to be higher to detect behaviours of short durations.
DOI: 10.1002/jag2.70065
Authors: Karin E. Schütz, Chanatda Somchit, Frances J. Huddart, Laura B. Hunter, Trevor Watson, Cheryl O’Connor, Bryan R. Thompson, Peter Green, Mhairi A. Sutherland
Institutions: Bioeconomy Science Institute, Beef + Lamb New Zealand
Language: English
Subfield: Small Animals
Field: Veterinary
Domain: Health Sciences
SDG: Zero hunger
Open Access status: hybrid
OpenAlex ID: W7151407385