The association between transition management and modelled milk yield in multiparous dairy cows

Abstract

The transition period remains one of the major challenges for dairy cows. Although farms have adopted automated technological means to assist them in breeding and feeding management, automated tools to tailor the transition management are scarce. Previous work has developed a methodology to calculate the milk yield residuals in the transition period (MRT) and investigated the potential of this feature to monitor multiparous dairy cows in early lactation. The MRT is obtained by subtracting a prediction of the expected production from the realised production at the first test day of the lactation. The current study investigates the associations of MRT compiled at farm level with farm management practices. Therefore, transition management practices were gathered through a survey on 45 Flemish and Dutch farms. After data aggregation and cleaning, data from 33 farms were retained and analysed using a partial least squares regression model. This model has the mean MRT of the individual farms as dependent variable. The independent variables included the survey answers combined with general farm production metrics. In cross-validation, the final model had an R2 of 0.69. The regression coefficients of this model mainly revealed that the mean MRT was positively associated with the herd mean 305-day milk yield and negatively associated with the feeding of grassland products in the dry period. While this study reveals valuable insights into the relationship between MRT and farm management practices, limitations such as sample size and unbalanced data distributions warrant further investigation with larger datasets and longitudinal approaches.