Feed efficiency and surplus feed intake among Norwegian Red dairy cows

Feed in livestock production is a major cost driver and closely linked to GHG emissions from the sector. Hence, there is a great deal of interest to improve feed efficiency of livestock and promote economic and environmental sustainability of livestock production. We estimated feed efficiency of 48 early- to mid-lactation Norwegian Red dairy cows based on intake and performance data collected for 54 d using a stochastic feed requirement frontier (SFRF). The SFRF predicted the minimum feed intake each cow required to support its production and maintenance needs, accounting for parity, lactation stage, body reserve mobilization, unobserved cow-specific heterogeneity, and statistical noise. Feed efficiency of each cow was then estimated in 2 ways: first by computing a feed efficiency score as a ratio of the prediction from the SFRF to the observed feed intake of each cow, and second by computing surplus feed intake (SFI) of each cow as a difference between observed and predicted feed intake. Furthermore, it was shown that feed efficiency can be decomposed into time-invariant and time-variant parts. Results showed that the average feed efficiency score was 89.7%, and it was time-invariant for the period covered during the study. The score implies 10.3% of the observed feed intake was SFI. In absolute terms, the SFI amounted to 13.3 kg of DM/wk on average. However, cows in the least feed efficient quartile had SFI of almost 5 times higher than cows in the most feed efficient quartile. Furthermore, cow-specific feed intake variation accounted for only 53.9% of the total feed intake variation. The remaining 46.1% of the feed intake variation was due to statistical noise. Out of the cow-specific feed intake variation, 82.1% was due to differences in feed efficiency, whereas the remaining 17.9% was due to other sources of unobserved cow-specific heterogeneity. When implemented with lifetime data, the feed efficiency decomposition approach presented in this study enables understanding of the temporal trajectory of feed efficiency within and across lactations. Furthermore, it allows identification of feed efficiency that is repeatable over lifetime, facilitating the inclusion of feed efficiency in breeding programs.