Title

Estimating dry matter intake of transition dairy cows through multiple on-cow accelerometer sensors

Document Type

Abstract

Publication Date

2021

Publisher

American Dairy Science Association

Journal

Jounral of Dairy Science

Volume

104

Issue

Suppl. 1

Pages

109

Language

en.

Keywords

accelerometer, intake, sensor technology

Abstract

Decreased dry matter intake (DMI) around calving is attributed to postpartal health disorders in dairy cows. Therefore, the objective of this study was to evaluate the feasibility of using 3D accelerometer sensors to estimate individual intakes of transition dairy cows. Eighteen multiparous Holstein cows housed in bedded pack pens during close-up were fitted with 3 sensors (Onset; Pocasset, MA) that record acceleration in the 3-axis (i.e., x, y, and z), one placed on the lateral side of the left hind leg and 2 attached to a halter directly superpose over the jaw and nose. After calving, cows were moved to a freestall barn bedded with straw. Cows were assigned 2 groups, a calibration group (A; n = 9) and a validation group (B; n = 9). Accelerations and individual intakes were collected from −7 to 7 d relative to parturition. Sensors were set to record accelerations at 1 min intervals. Acceleration models highly associated with DMI determined in a previous study (Carpinelli et al., 2019; J Dairy Sci, 102:11483) were used to cross-reference accelerometer data and DMI in group A. Six additional variables were derived from jaw and nose accelerations by measuring the change in acceleration between 2 consecutive time points (Lag-time). The REG procedure of SAS was used in group A to generate an intercept (B0) and slope (B1). Then, these were used in group B to derive DMI from previous acceleration models (DMIA) and compared this against the actual DMI using the CORR and MIXED procedures of SAS. DMIA was closest (P = 0.58) to the actual DMI using the models LagNoseZ+LagJawZ (14.0 vs 13.6 kg/d ± 0.5 kg/d) and JawX+JawZ+JawY+LagJawY (13.8 vs 13.6 kg/d ± 0.5 kg/d). Moreover, the LagNoseZ+LagJawZ model was able to produce a DMIA that exhibited a rapid decrease (P = 0.03) commonly observed around calving in actual DMI, while JawX+JawZ+JawY+LagJawY model did not (P > 0.05). Overall, DMI and DMIA were not correlated (P ≥ 0.05; r ≤ 0.20) in all tested models. Peripartal DMI is highly vari- able and challenging to estimate with sensor-based methods. However, the similar DMI and DMIA observed in this study is encouraging and demonstrates a great potential for this approach to estimate DMI around calving, which could help in the future to flag cows at risk of developing a postpartal disease.

This document is currently not available here.

Share

COinS