Evaluation of an activity and rumination monitor in dairy cattle grazing two types of forages

The aim of the present study was to evaluate the accuracy of a newer version of an activity- and rumination-monitoring system by comparison against direct visual observations, for the following three different types of behaviour: grazing, resting (described as lying or standing idle) and ruminating for cows grazing either annual ryegrass or chicory-based swards. Eight non-lactating Holstein–Friesian cows were fitted with the sensor tags, and grazed on annual ryegrass pasture for a target consumption of 10 kg DM ryegrass/cow.day for 7 days. The experiment was then repeated with cattle offered a similar allowance of chicory. Observations were conducted by two trained observers in two observation periods each day, to capture the above described behaviours. In each period, electronic behavioural measurements were recorded continuously by the sensors, while visual observations were also continuous (during observation periods), and the two datasets were matched. On average, each cow was visually observed for 87.2 min/day. For each behavioural state (at 1-min intervals, n = 6963), probability of agreement, sensitivity, specificity and positive predicted value were determined for grazing as 98%, 98.3%, 97.3% and 98.9% respectively, for resting as 80%, 77.5%, 99.1% and 92.9% and for ruminating as 87%, 86.9%, 98.4% and 90.68%. Concordance correlation coefficient (CCC) and Pearson correlations (r) were used to investigate the relationships between visual observations and data generated from the tags. Different behaviours were analysed separately. Significant correlations were found for the three behaviours (grazing: CCC = 0.99, r = 0.99; resting: CCC = 0.95, r = 0.97; ruminating: CCC = 0.80, r = 0.80), with no differences detected between the two forages. We conclude that, under the conditions of the present study, the activity- and rumination-monitoring system tag measured grazing, resting and ruminating behaviours with high accuracy on the basis of comparison to visual observations.