2.3 Herbivores behavior monitoring
We selected sheep (Ovis aries; one of the most extensively kept livestock breeds in the grassland of Mongolia Plateau) attached with sensors to monitor the strength of actives (Fig. 2d). The accelerations of sheep were monitored using accelerometer loggers from Druid Technology Co., Ltd (Chengdu, China; https://www.ecotopiago.com). The collars were programmed to record three-axial acceleration at a frequency of 25 Hz for 1.5s every minute (i.e., 37 data points per axis and 111 data points per ACC burst) during the grazing period between the end of July and September. We conducted a correlation analysis between the frequency tri-axial of acceleration obtained for 1.5 seconds and those obtained by the device automatically calculating the values for every 6 seconds. The results showed a significant correlation, indicating that the 1.5-second frequency can represent the animal’s behavior within that minute (Supplementary Fig. 2).
The data obtained from the study were stored in a device containing an nRF52840 SoC (system on a chip) that boasts a 64 MHz microprocessor, 1 MB Flash, and 256 kB RAM. The SoC features BLE (Bluetooth Low Energy), which facilitates data transfer between the logger and personal cell phones over a range of 70 m. The device also incorporates a 3G module for data communication via the mobile network, as well as sensors for capturing temperature and light intensity in addition to the three-axial ACC sensor. A solar panel was fitted onto the logger to recharge a 200 mAh rechargeable lithium battery.
The acceleration data initially collected in millivolts was converted into gravitational acceleration units (g), which represent the rate of change of velocity concerning gravity, using tag-specific calibration values. Subsequently, each of the three signals was subject to individual smoothing through running means over a period of 1.5 seconds. Following this, the corresponding smoothed data were subtracted from the unsmoothed data for each axis. The resulting sum of all three axes yielded a value for the Overall Dynamic Body Acceleration (ODBA) experienced by the animal given by
ODBA = |Dynamic acceleration (x)| + |Dynamic acceleration (y)| + |Dynamic acceleration (z)|