Few portable devices that record grazing time have been tested on goats. The Lifecorder Plus is a commercial device, based on a single-axis accelerometer, that calculates and stores in memory the average activity level (range: 0–9) every 2 min. Previous studies determined that activity levels greater than or equal to 0.5 were specific to grazing, while those less than 0.5 could be considered as other activities. The aims of the present study were to determine the threshold value that minimised the prediction error and to validate the ability of the Lifecorder Plus device to record goat grazing time at vegetative temperate pasture under rotational grazing management. Twenty goats (5 in 2015 and 15 in 2017) were fitted with the Lifecorder Plus on their neck and had access to pasture for 8−12 h/d. The device was evaluated by comparing 2-min visual observations of trained observers to the activity level recorded at the same time by the device, resulting in a total of 187 h of observations. Two analysis methods were used to assess the device’s accuracy: a confusion matrix and the mean squared prediction error. Lifecorder Plus recorded grazing time with 93 % accuracy, 95 % precision, 97 % sensitivity and 73 % specificity. Relative prediction errors calculated at the hour and day scales were low, averaging 0.09 and 0.05 of actual grazing time, respectively. Overall, there was no confusion between grazing and other goat activities. Lifecorder Plus overestimated grazing time by 3%, probably due to other activities of goats (lack of specificity). Placing an additional Lifecorder Plus on a leg may help identify non-grazing activities (e.g. long-duration walking, running) and thus reduce this slight overestimation. No bias due to sward height was detected. In conclusion, for rotationally grazed vegetative temperate pastures, Lifecorder Plus detects grazing activities of goats with high accuracy and precision and is suitable for studying variations in grazing time, particularly at the day scale in small paddocks.

很少有记录放牧时间的便携式设备在山羊上进行过测试。Lifecorder Plus 是一种基于单轴加速度计的商业设备，每 2 分钟计算一次平均活动水平（范围：0-9）并将其存储在内存中。先前的研究确定，大于或等于 0.5 的活动水平特定于放牧，而低于 0.5 的活动水平可被视为其他活动。本研究的目的是确定最小化预测误差的阈值，并验证 Lifecorder Plus 设备在轮牧管理下记录植被温带牧场山羊放牧时间的能力。20 只山羊（2015 年有 5 只，2017 年有 15 只）在他们的脖子上安装了 Lifecorder Plus，并且可以进入牧场 8-12 小时/天。通过将受过训练的观察者的 2 分钟视觉观察与设备同时记录的活动水平进行比较来评估该设备，从而产生总共 187 小时的观察结果。使用两种分析方法来评估设备的准确性：混淆矩阵和均方预测误差。Lifecorder Plus 以 93% 的准确度、95% 的精确度、97% 的灵敏度和 73% 的特异性记录放牧时间。在小时和天尺度计算的相对预测误差很低，分别为实际放牧时间的平均 0.09 和 0.05。总体而言，放牧和其他山羊活动之间没有混淆。Lifecorder Plus 高估了 3% 的放牧时间，可能是由于山羊的其他活动（缺乏特异性）。在腿上放置一个额外的 Lifecorder Plus 可能有助于识别非放牧活动（例如 长时间步行、跑步），从而减少这种轻微的高估。没有检测到由于草皮高度造成的偏差。总之，对于旋转放牧的温带植物性牧场，Lifecorder Plus 能够以高精度和精确度检测山羊的放牧活动，适用于研究放牧时间的变化，特别是在小围场的日间尺度上。