Aussagegenauigkeit der Milchleistungsprüfung unter Bedingungen automatischer Melkverfahren – Vergleich deutscher und kanadischer Modellansätze

Abstract. Title of the paper: Reliability of milk recording applying automatic milking - comparison of German and Canadian model approaches
Due to a high variability of milking intervals within animals rate of milk secretion and milk yield per hour in automatic milking systems (AMS) are more variable than in conventional milking systems. Further reasons are technical problems and the absence of milking persons cows with problems are to be milked. The calculation of milk yield obtained in 24 hours, only based on milking during the test day, is not precise enough. To calculate the average milk yield in a testing period as many milking as possible should be taken into consideration. Milk yield calculated according to that procedure does not correspond with milk composition of the test day.

Based on 85012 milking on one farm the amount of milking was calculated, required to minimize the variability of milk yield per hour and to obtain a high correlation of the calculated daily yield with the "real" milk yield during the sampling period. Depending on number and state of lactation it was found that between 13 and 16 milking are required to obtain a maximum of accuracy. In all classes 12 milking would result in 95% of the maximum accuracy. Since farm management and type of the AMS may affect the results additional types of AMS and more farms should be included into the evaluation.