Accelerometer derived rumination monitoring detects changes in behaviour around parturition

Monitoring cattle for parturition is a labour-intensive task that is often not feasible in extensive environments. On-animal sensors, however, provide a means of remotely monitoring animals, without the labour requirements of traditional observation. This paper explores the use of a rumination algorithm with accelerometer ear tags to observe the changes in rumination around calving. Calving cattle (n = 17) were housed in an intensive calving barn and movement data was collected using tri-axial accelerometer ear tags. A heuristic approach, whereby data was manually visualised to identify ruminating and non-ruminating periods, was used to train a rumination model tailored to each individual animal. Rumination time, as calculated using the rumination model, decreased on 0d and 1d, with a decline in rumination commencing 6 h prior to parturition. These results explore the use of an individualised rumination model for monitoring changes in rumination relative to parturition. Further studies applying this model in various production systems could lead to the development of a universally applicable model for integration with a commercially available sensor system. These commercial tags could subsequently be used to improve the management of cattle around the parturition period, leading to improved productivity and welfare.