Potenzial der Mid-Infrarot-Spektrometrie bei Kuhmilchprobenzur Abschätzung der Rationszusammensetzung

Milk composition of dairy animals is influenced by the composition of the ration fed. The objective of this study was to determine if the percentages and absolute amounts of hay, grass silage, pasture, maize silage and concentrate in the feed ration can be estimated using MIR spectrometry of milk. A total of 10200 milk samples from 90 dairy cows were collected, and the intakes of all ration components were measured. Using partial least squares regression (PLS), equations were developed to estimate ration compositions corresponding to each milk sample. To evaluate accuracy, the correlation between observed and estimated values (R) and ratio to performance (RPD) were used. Notable R values (for kg/for %) were observed for the ration proportion of pasture (0.85/0.87), maize silage (0.74/0.75) and concentrate intake (0.75/0.73). Estimation of groups of feedstuffs (all forages, energy-dense feedstuffs) resulted in R values of >0.8. Including the parameters milk yield and/or concentrate intake into PLS improved R values by up to 0.08. The results indicate a potential use of MIR spectra as a promising predictor for ration composition of dairy cows.