Can diurnal variation in velocity of grazing pregnant Merino ewes be used to identify lambing?

Abstract

The development of remote monitoring of animal behaviour will have many implications. One major issue with these technologies is that they generate large datasets. Therefore, methods are required that can reduce these datasets to more manageable sizes and therefore help to improve decision making for livestock management and welfare. It is known that pregnant ewes change their behaviour around lambing, with one observed change being that the ewe drops behind the main flock as lambing approaches. A benefit from knowing when the ewe is about to lamb, or has lambed, could be a reduction in lamb mortality at this critical time. Dobos et al (2010) showed that Bayesian change point analysis could be used to identify the onset of lambing using mean daily speed as a metric. However, there may be other methods that could be used to identify lambing. To test if changes in diurnal variation with velocity as a metric can be used to identify lambing, data from an investigation on shelter use by pregnant grazing Merino ewes (Taylor et al. 2011) was summarised for this analysis. The mean hourly velocities (m/s) calculated from GPS locations taken at 10 minute intervals from five grazing pregnant Merino ewes seven days before lambing, at lambing, three and seven days after lambing (period) were analysed using a mixed model. Variation in hourly velocity between ewes in all periods was high. No significance difference (P>0.05) was found in diurnal variation between periods. Mean hourly velocity peaked at 5h and at 16h for ewes 7d before, 3d after and 7d after lambing. At lambing mean hourly velocity was reduced, with two short peaks at 5h and 9h and two higher peaks at 12h and 20h. Because of the large variation in individual ewe velocity in all periods, further research is required to determine if these changes in peak velocity times correlate with changes in behaviour at lambing.