Abstract Monitoring behavior changes of livestock may help producers expedite intervention and treatment in both intensive and extensive systems. Sensor technologies can identify behavioral trends, offering the potential to detect abnormal behavior associated with health status and animal welfare changes. The aim of recent studies was to investigate behavior deviations using tri-axial accelerometers and global positioning systems (GPS). Accelerometers were effective tools in determining behavior changes due to the onset of illnesses and parturition. Eight Brahman-cross heifers were fitted with accelerometers to study water consumption behavior. Two of the heifers were diagnosed with Bovine Ephemeral Fever, which symptoms include lameness and stiffness. Movement Intensity (MI), a metric calculated from accelerometer axis data, was lower (P ≤ 0.01) during the day of disease diagnosis compared to 48 h prior. Ten ewes were each fitted with HerdDogg biometric accelerometers, with five ewes also fitted with accelerometers to investigate behavior before and after illness caused by the unintentional feeding of mold-contaminated feedstuffs. Activity levels were lower for days 1 to 4 after feeding (P < 0.05) than day -4 to 0 before feeding. Thirteen Debouillet ewes, with accelerometer ear tags, were used to determine the ability of random forest machine learning techniques to accurately identify and predict parturition-related behavior. Prediction models were able to predict seven mutually-exclusive behaviors and active vs. inactive behavior with 66.7% and 87.2% accuracy, respectively. The use of GPS tracking can help identify proper pasture rotation timing. Thirty-two and twenty-nine cattle were fitted with GPS collars in central Arizona to evaluate the temporal changes in association patterns on two herds of Corriente cattle managed under two different stocking densities: high stocking density (0.417 animal ha-1) and low stocking density (0.123 animals ha-1), respectively. Half-weight index (HWI) value was calculated for each pair of GPS-tracked cattle to determine the time cattle spent within 75 m and 500 m of others. Cattle within both pastures exhibited low mean association values (HWI < 0.25) at both spatial scales at the beginning of the study with increasing levels of independence throughout the study. Cattle dispersed throughout the study (P < 0.01) and traveled farther from water (P < 0.01), likely in search of forages. The combination of GPS and accelerometers can improve detection of welfare issues. Seventeen cattle were tracked with both sensors and sixteen cattle tracked with only GPS across two years to determine the potential of detecting water system failures. Movement intensity was greater (P = 0.03) and cattle remained closer to water (P = 0.01) on the days of simulated water failure compared to control days. Real-time sensors have the potential to provide caretakers notification of changes in livestock behavior that can occur due to disease, parturition, or infrastructure failure.
Read full abstract