Animal Behavior and Well-Being 1

摘要

Formerly, cattle behavior studies were performed by direct visualobservations or video analysis, which require considerable resourcesand time. Recently, accelerometers have been successfully used forremote monitoring of behavior in dairy calves. However, this validationwas made under temperate weather conditions, and to the authorsunderstanding no similar data exist for the tropics where heat stressaffects cattle behavior. Thus, the y-axis values recorded by HOBOPendant G accelerometers tied to the right rear leg (perpendicular to thefloor) in confined Holstein calves (n = 5; 53 ± 20 d old) in Puerto Ricowere evaluated as indicators of lying (L) and standing (S). Data wererecorded at 1-s intervals from 0700 to 0900h (n = 23,408) on February7, 2019. The L and S events were visually recorded as gold standardsfor validation. Transitional movements between L and S were excluded.To determine the best way to classify the obtained data, its overall meanand standard deviation (SD) during the visual L and S events were calculatedand used to create a series of respective accelerometer categories(L or S) in the ranges between 0.4, 0.6, 0.8, 1, 1.2, 1.4, 1.6, 1.8, 2, 2.2,2.4 or 2.6 SD bellow to above the mean. Proc FREQ (SAS) was usedto determine the sensitivity (likelihood that a visually recorded eventwas also recorded by the accelerometer), visual-accelerometer misclassifications(L-S or S-L) and non-classified observations in each createdcategory. All categories successfully identified over 99% of the recordedL and S events, with less than 0.04 and 0.16% of L-S and S-L misclassifications,respectively. However, the amount of non-classificationsincreased as the category range decreased (0.47 to 55% from 2.6 to 0.4SD, respectively). The category that maximized the sensitivity for theL (99.89%) and S (99.98%) events, while maintaining low levels ofthe L-S (0.02%) and S-L (0.11%) misclassifications and non-classifiedvalues (1.23%) was the 2 SD bellow and above the mean. The evaluatedaccelerometers may be useful tools for remote monitoring of L and Sbehavior in dairy calves in the tropics.