Temporal variability in automatically-collected liveweight data affects estimation of dry matter intake of dairy cows

Abstract

Daily liveweight monitoring can be done using an automated weighing system. However, liveweight of dairy cows is extremely variable between and within individual cows and days due to various factors. Large variation in daily liveweight changes can be seen when measured at different times of the day, and that may affect the estimation of daily energy required by cows, and directly affect the estimation of dry matter intake. The objectives of the present study were to quantify the variation in cow’s liveweight automatically recorded in the morning and afternoon, and subsequently demonstrate the differences in estimates of dry matter intake. Morning and afternoon liveweight data from 64 primiparous and multiparous lactating Holstein-Friesian cows were used in this study. Two analyses were performed, without (raw) and with smoothing liveweight data using a spline function to minimize measurement errors and reduce the variability in liveweight over time. Dry matter intake for individual animals was estimated using back-calculation of the energy requirement from liveweight, liveweight change, milk production and milk composition. Results showed that, on average, cows were significantly heavier in the afternoon (Mean ± SD = 614.3 ± 78.20 kg) compared to morning (593.2 ± 76.08 kg; P < 0.01). In terms of variability, the residual standard deviation indicated that the afternoon liveweight had more fluctuations (P < 0.01) over the time compared to those recorded in the morning. The estimation of both individual liveweight change and dry matter intake using morning and afternoon liveweight in their raw form showed unrealistic values ranging from −110.0 to +96.0 and −337.57 to + 399.74 kg/day, respectively. The smoothing of automatically-collected liveweight data substantially reduced the variability of liveweight, liveweight change and dry matter intake. The estimated dry matter intake ranged between 1 and 5% of liveweight/day when calculated using adjusted liveweight values. Based on the residual standard deviation for estimated dry matter intake, we concluded that a single liveweight measurement per cow in the morning is adequate to adjust the automatically-collected liveweight to reduce variability in the estimated intake.