Global Positioning System (GPS)-based virtual fences offer the potential to improve the management of grazing animals. Prototype collar devices utilising patented virtual fencing algorithms were placed on six Angus heifers in a 6.15 hectare paddock. After a “no fence” period, sequential, shifting virtual fences restricted the animals to 40%, 60%, and 80% of the paddock area widthways and 50% lengthways across 22 days. Audio cues signaled the virtual boundary, and were paired with electrical stimuli if the animals continued forward into the boundary. Within approximately 48 h, the cattle learned the 40% fence and were henceforth restricted to the subsequent inclusion zones a minimum of 96.70% (±standard error 0.01%) of the time. Over time, the animals increasingly stayed within the inclusion zones using audio cues alone, and on average, approached the new fence within 4.25 h. The animals were thus attentive to the audio cue, not the fence location. The time spent standing and lying and the number of steps were similar between inclusion zones (all p ≥ 0.42). More lying bouts occurred at the 80% and lengthways inclusion zones relative to “no fence” (p = 0.04). Further research should test different cattle groups in variable paddock settings and measure physiological welfare responses to the virtual fencing stimuli.

The eShepherd^® virtual fencing system being commercialized for cattle has the potential to exclude cattle from environmentally sensitive areas. Animals are given audio cues to indicate a fence line via a neckband device. An electrical pulse is administered if the animal continues moving forward following an audio cue. A commercial trial was conducted in South Australia to assess whether virtual fencing technology could exclude 20 cattle from an area of regenerating saplings; across 44 days; using a contoured fence line. The results showed that the cattle were able to rapidly learn the virtual fencing cues; responding appropriately to the audio cue for 74.5% of 4378 audio signals; and were excluded from the regenerating area for 99.8% of the trial period with the more complex fence line (contoured; not straight) in place. IceQube R’s^® measuring lying time and bouts showed no consistent increasing or decreasing pattern of change. At the trial conclusion; the feed available in the protected zone was double the quantity and quality of the grazed zone. Technical issues occurred with some of the pre-commercial prototype devices; but those versions are now obsolete. This study observed a single group of cattle in one paddock; further testing of the virtual technology is warranted.

Grazing cattle can both negatively and positively impact riparian zones, dependent on controlled grazing management. Virtual fencing technology, using collar devices that operate via GPS can provide audio cues and electrical stimuli to temporarily exclude cattle from specified areas as desired. An early experimental prototype automated virtual fencing system was tested in excluding ten cattle from a riparian zone in Australia. Animals were given free access to an 11.33-hectare area for three weeks, excluded from river access by a virtual fence for ten days (2.86-hectare inclusion zone), followed by free access again for six days. Animals were almost exclusively contained by the virtual fence. All animals received audio cues and electrical stimuli with daily fence interactions, but there was high individual variation with some animals first approaching the fence more often than others. Overall, there was an approximately 25% probability that animals would receive an electrical stimulus following an audio cue. Individual associative learning may have been socially-facilitated by the group’s behaviour. Following fence deactivation, all animals re-entered the previously excluded area. Further research with more groups and longer periods of exclusion using updated collar devices would determine the scope of virtual fencing technology for cattle grazing control.

This series of experiments was undertaken to develop a low animal impact, inexpensive and simple method of continuous logging of body temperature in free-ranging adult cattle for periods up to 7 days. Two different types of logger (iButton and Vemco minilogger) and three different methods of attachment (surgical implant, anal girth strap and anal tail attachment) for the iButton were investigated. The technique based on the iButton temperature logger, secured to the tail head was the most effective. This technique required minimal animal handling and provided reliable and accurate measurement of temperature in freely behaving cattle.

This study evaluated the relationship between rectal temperature (T_REC, °C) and vaginal temperature (T_VAG, °C) in grazing Bos taurus heifers, to develop an understanding of the reliability of these measures as estimates of core body temperature. Nineteen Angus heifers (BW = 232.2 ± 6.91 kg) were implanted with intra-rectal and intra-vaginal data loggers. Rectal temperature and T_VAG were simultaneously recorded at 20 s intervals over 18.5 h. Heifers were housed as a singular cohort on grazing pastures for the duration of the study. A strong linear relationship (R² = 0.72, p < 0.0001) between the measurement sites was identified. The mean difference between T_REC and T_VAG was small, in which T_VAG was on average 0.22 ± 0.01 °C lower than T_REC. Individual twenty second T_REC and T_VAG data were used to determine the pooled mean T_REC and T_VAG and then to highlight the within measure variation over time. The coefficient of variation was, on average, lower (p < 0.001) for T_VAG (0.38%) than T_REC (0.44%), indicating that T_VAG exhibited less variation. Overall, the results from the current study suggest that a strong relationship exists between T_REC and T_VAG, and that T_VAG may be a more reliable estimate of core body temperature than T_REC in grazing Bos taurus heifers.

Understanding social interactions in livestock groups could improve management practices, but this can be difficult and time-consuming using traditional methods of live observations and video recordings. Sensor technologies and machine learning techniques could provide insight not previously possible. In this study, based on the animals’ location information acquired by a new cooperative wireless localisation system, unsupervised machine learning approaches were performed to identify the social structure of a small group of cattle yearlings ($n=10$) and the social behaviour of an individual. The paper first defined the affinity between an animal pair based on the ranks of their distance. Unsupervised clustering algorithms were then performed, including K-means clustering and agglomerative hierarchical clustering. In particular, K-means clustering was applied based on logical and physical distance. By comparing the clustering result based on logical distance and physical distance, the leader animals and the influence of an individual in a herd of cattle were identified, which provides valuable information for studying the behaviour of animal herds. Improvements in device robustness and replication of this work would confirm the practical application of this technology and analysis methodologies.