The transition from aggregated lactation yields to single test day records in breeding value estimation procedures for milk production traits allows a more detailed consideration of environmental factors. In this study the definition and significance of environmental effects on milk, fat and protein yield were analysed with multiple trait repeatability test day model. As criteria for the validation of different models for the fixed effects, the residual mean squared error and the correlation between true and estimated breeding values (r_HI) were taken. A new method was developed to calculate r_HI depending on the used operational values in the estimation procedure and the true variance of the effect under consideration. The unfavourable herd structure in Southern Germany leads to problems in genetic evaluations because the number of observations per herd test day is small. Therefore, one main part of the study consisted in optimization of the contemporary group definition, with special regard to the interaction between parity and herd test day. The results showed higher r_HI, if that interaction was excluded from the model. This was especially valid for the breeding values of cows from smaller herds, but also for sires and led to the recommendation to exclude this interaction from the model in the routine evaluation. The analyses for the effect of pregnancy have shown, that an interaction between milk production traits and calving interval exists. Therefore, this factor is inappropriate to correct for the pregnancy effect in a test day model. The linear, quadratic and cubic covariable 'days carried calf' was found to be best to correct yield traits for the pregnancy effect. The investigation of further environmental factors, like stage of lactation, year and month of calving, calving age and region led to a proposal for the statistical test day model, that will be used in developing a routine genetic evaluation in Southern Germany.
«
The transition from aggregated lactation yields to single test day records in breeding value estimation procedures for milk production traits allows a more detailed consideration of environmental factors. In this study the definition and significance of environmental effects on milk, fat and protein yield were analysed with multiple trait repeatability test day model. As criteria for the validation of different models for the fixed effects, the residual mean squared error and the correlation bet...
»