Sensor validation to record rumination and analysis of behavioral parameters of dairy Gyr heifers in feedlot systems

The use of sensors for livestock monitoring has grown in the last decade, motivated not only by the search for improvements in animal production and health, but also by the better use of physical, financial, and human resources in the management of livestock. Since precision livestock farming depends on technological innovation, including the use of sensors, it is essential to validate their applicability in confined animals. Therefore, the objective was to evaluate the functioning of the SCR Heatime® HR System sensor in heifers and analyze the behavioral differences of these animals in two feedlot systems (tie-stall and loose-housing). Eleven Gyr heifers of 10 months of age and an average body weight of 179 ± 26 kg were used. Animals were equipped with the SCR Heatime® HR System 7 days before the evaluations for adaptation, with the behavioral and system studies carried out during two periods each of 5 days, for 8 h a day. Five heifers were housed in tie-stall bedding and six were housed in loose-housing. After the evaluations for a period, animals have been switched systems in the following period. While the animals had the sensors attached to their bodies, behavioral evaluations were also performed visually, every 3 min by two evaluators per system. Behavior activities were characterized as standing, lying, idle standing, idle lying, ruminating standing, ruminating lying, feed, and water intake. The regression for the rumination activity recorded between the sensor and the visual assessment was significant only for loose-housing (P = 0.0002), but the Pearson correlation between both was negative and low (r =  - 0.25; P = 0.0002). The sensor overestimated rumination by 27.3% in loose-housing (28 vs. 22 min/2 h) and 38.5% in tie-stall (36 vs. 26 min/2 h). Thus, when considering the different feedlot systems, validation was not possible due to the overestimation of time spent on rumination activity determined by the sensor. In the behavioral comparison between the systems by visual assessment, there was a significant difference (P < 0.05) for the behaviors of motor activity, idle lying, total idle time, and feed intake between the animals in the two housing systems. As for idle standing, ruminating standing, total rumination time, and water intake, there was no difference (P > 0.05). However, the time spent ruminating lying was greater from 10:00 to 12:00 h (43 ± 3 min/2 h) in the tie-stall, whereas in the loose-housing the animals distribute more uniformly their rumination from 08:00 to 16:00 h (26 min/2 h on average). Therefore, the use of the SCR Heatime® HR System sensor is not recommended for Gyr heifers housed in tie-stall and loose-housing systems. Regarding the daytime range of rumination and the accuracy improvements of similar sensors, it is recommended to develop them in both feedlot systems using the periods of less rumination in early morning in tie-stall and late afternoon in loose-housing.