PSIII-4 Preliminary exploration of relationship of automated sensor data with feed intake and efficiency in lactating dairy cattle
Abstract Feed costs represent the greatest expense on a dairy farm, making feed efficiency an important trait to consider among production traits. Current tools to measure feed intake have limited application in commercial settings, due to affordability and lack of portability of technologies. Therefore, development of automated sensor-based indicator traits for feed intake could prove to be valuable. The objective of the current study was to determine if automated eartag data was associated with feed intake. Activity and inner ear temperature were collected every 19 minutes utilizing Quantified Ag eartags (n = 48 lactating cows). Ear tags were placed 5 days prior to the start of the trial, with cows ranging from 67-192 days in milk (DIM). Daily feed intake, milk weights, milk components and body weight (BW) were also recorded. Data were analyzed using PROX GLIMMIX in SAS. Dry matter intake (DMI) was modeled including fixed effects for DIM, milk weight, component composition, metabolic body weight (BW0.75), eartag activity or temperature, as well as the random effects of parity and group. To identify informative timeframes with reduced influence of environmental noise, data were analyzed over 3-day rolling windows of time. Six windows were significantly associated with dry matter intake (P ≤ 0.05) when utilizing ear tag activity. Three windows of time of ear tag temperature were found to be significantly associated with DMI (P ≤ 0.05). These findings indicate that eartag sensor data may be useful indicators of feed intake; however, days in milk and season may impact the informativeness of sensor data. Additional studies are warranted to validate the efficacy of activity and ear temperature as indicators of feed intake and determine the impact of other variables on these potential sensor indicator traits over time.