Prediction of calving in dairy cows using a tail-mounted tri-axial accelerometer: A pilot study

Dystocia and stillborn calves are recurring problems for dairy farms, leading to high costs. By implementing a real-time monitoring system, problems could be reduced by supporting farmers to detect the onset of parturition and to intervene in case of dystocia. The hypothesis of this study was that the onset of parturition is detectable through analyses of the movement pattern of the clam's tail, recorded by a tri-axial accelerometer that was fixed on the upper part of the cow's tail. Cows (n = 5, approx. 1 week before calving) were housed in single straw-bedded boxes. Animal behaviour was video-recorded (24 h/d) and evaluated by encoding the events 'frequency and duration of tail raising 5 h pre-partum', 'rupture of the amniotic sac' and 'expulsion of the calf'. In parallel, the accelerometer data collected from two days before calving to parturition were analysed. We developed an algorithm to detect the tail raising and created a decision function based on the frequency and duration of the tail raising. Exceeding the threshold led to a birth alarm. In each of the evaluated calvings, the alarm was triggered a short time before the expulsion of the calves, at 33, 32, 121, 6 and 71 min for cows 1 to 5, respectively. These preliminary results indicate that an accelerometer to detect tail movements may be useful to predict parturition. Further research is required to refine the algorithm and the decision function, to analyse predictability of dystocia and to develop a real-time alarm system under field conditions. (C) 2017 IAgrE. Published by Elsevier Ltd. All rights reserved.