The objective of this study was to identify housing and management factors associated with productivity on automatic milking system (AMS) dairy farms measured as daily milk yield/AMS and daily milk yield/cow. Management, housing, and lameness prevalence data were collected from 33 AMS farms in Minnesota and Wisconsin during a farm visit. All farms in the study used free-flow cow traffic. Mixed model analysis of cross-sectional data showed that farms with automatic feed push-up via a robot produced more milk per AMS/day and per cow/day than farms where feed was pushed up manually. New versus retrofitted facility, freestall surface, manure removal system, and the number of AMS units/pen were not associated with daily milk yield per AMS or per cow. Cow comfort index (calculated as number of cows lying down in stalls divided by total number of cows touching a stall) was positively associated with daily milk yield/cow. Prevalence of lameness and severe lameness, number of cows per full-time employee, depth of the area in front of the AMS milking station, and length of the exit lane from the AMS milking station were not associated with daily milk yield per AMS or per cow. Multivariable mixed model analysis of longitudinal AMS software data collected daily over approximately an 18-mo period from 32 of the farms found a positive association between daily milk yield/AMS and average age of the cows, cow milking frequency, cow milking speed, number of cows/AMS, and daily amount of concentrate feed offered/cow in the AMS. Factors negatively associated with daily milk yield/AMS were number of failed and refused cow visits to the AMS, treatment time (the time spent preparing the udder before milking and applying a teat disinfectant after milking), and amount of residual concentrate feed/cow. Similar results were also found for daily milk yield on a per cow basis; however, as it would be expected, average days in milk of the herd were also negatively associated with daily milk yield/cow. These findings indicate that several management and cow factors must be managed well to optimize AMS productivity.
Bibliographical noteFunding Information:
We thank all the dairy producers who participated in the study. We also thank Luis Espejo (St. Augustine, FL) for help with statistical analysis and Lucas Salfer (Dassel, MN) for collection of on-farm observations, cow locomotion scores, and producer questionnaire responses. In addition, we thank Lely Industries N.V. (Maassluis, the Netherlands) for technical help with software data collection and Kelly Froelich and Michael Schmitt (both University of Minnesota) for help with software data entry. We also thank David W. Kammel (University of Wisconsin, Biological Systems Engineering Department, Madison) for his input on housing data collection and help with enrollment of Wisconsin AMS farms. Justin Siewert was partially supported by a Department of Animal Science Fellowship (University of Minnesota) and the John Brandt Memorial Scholarship (Land O'Lakes Inc., St. Paul, MN). This project was partially supported by Hatch Funds from the USDA National Institute of Food and Agriculture.
- automatic milking
- milk yield
- robotic milking