Crossing grazing calendars and bulk milk analysis : is on-field validation for grass-diet in cattle possible ?

Managing an animal system based on pastures is a time-consuming activity that exhibits many benefits.
Assessing the extent to which farms optimize pasture usage and prioritize grass feeding presents
challenges. This assessment holds significance for various labelling requirements mandating a
proportion of grass in the diet and serves to incentivize favourable management practices.
Consequently, a verification method based on milk composition seems relevant. This study seeks to
validate such hypothesis by cross-referencing 75 tank-milk analysis with 526 records from grazing
calendars and supplementary feed data. Employing hierarchical clustering, distinct management
practice groups were identified, with Groups 2 and 4 displaying the most discernible differences,
particularly in their emphasis on full grazing and grass-based diets. Subsequent analysis revealed
significant variations in milk composition among these groups, prompting the development of a
predictive model based on Partial Least Squares-Discriminant Analysis (PLS-DA) from key milk traits
predicted by milk mid-infrared spectrometry, like milk pH, Mg, K, Ca, P, C18:1 trans, C14:1 and C16. The
study yielded interesting results, with an accuracy to distinguish Group 2 from Group 4 of 0.92. Using
this PLSDA indicator would be interesting as a next step to potentially determine the proportion of
grass in diet and even the proportion of grazed grass. This could not only help in the verification for
labelling, but also become an indicator for farmers to see if their cattle best valorise the pastures.