Effect of cow traffic system on cow performance and AMS capacity

Abstract: Robotic milking in Automatic Milking systems (AMS) is proposed to reduce manual labour and at the same time increase milk yield by increasing milking frequency. In order to increase milking frequency, it is essential to have well-functioning cow traffic. Investing in an AMS is a great capital investment for the farmer, thus it is of major importance to ensure maximal AMS capacity. This study investigated the effect of the traffic systems Feed First™ and Free cow traffic with and without waiting area (WA) on cow performance and AMS capacity. The effects were estimated with multivariable linear regression models, accounting for other potentially influencing factors. The farmers’ satisfaction with the traffic systems was investigated as well as how often and how many cows that were fetched to the milking unit and how much time the farmers spent on fetching cows to the milking unit. Data was collected from 165 dairy farms in Denmark and the Netherlands for the period 1st July 2011 to 1st July 2012. The data was retrieved from the management systems on each farm as well as from a telephone based quantitative interview with the farmers. Cows in Feed First traffic produced on average 0.6 kg less milk per day compared to cows in Free cow traffic with WA and 0.7 kg less milk per day compared to cows in Free cow traffic without WA (p<0.001). Cows in Feed First traffic produced on average 0.1 kg less milk per milking compared to cows in Free cow traffic with WA (p<0.05). On average, cows in Feed first traffic visited the milking unit 0.1 less times per day compared to cows in Free cow traffic with and without WA (p<0.001). Descriptive data showed that average milking duration per robot and day was 16.6 h, 17.1 h and 16.7 h in Feed First, Free cow traffic with WA and Free cow traffic without WA respectively. Average amount of milk per robot day was 1529.8 kg, 1583.7 kg and 1550.6 kg for Feed First, Free cow traffic with WA and Free cow traffic without WA respectively. Feed First traffic resulted in on average 138.1 milkings per robot and day whereas Free cow traffic with and without WA resulted in 142.2 and 140.4 milkings per robot and day respectively. The estimates from the statistical model, i.e. adjusted for the potential effects of other variables in the model, showed that Feed First traffic resulted in 5.6 fewer visits per robot and day compared to Free cow traffic without WA (p<0.05). Average number of cows per robot was 55 in Feed First traffic and Free cow traffic with WA and 54.1 in Free cow traffic without WA (p<0.001). Farmers with Feed First traffic reported that they fetched cows on average 1.5 times per VMS and day compared to 2.5 in Free cow traffic with WA (p<0.001) and 2.3 in Free cow traffic without WA (p<0.05). Farmers in Feed First traffic also reported that they fetched on average 16.6 cows per VMS and day compared to 24.0 in Free cow traffic with WA (p<0.05). In terms of cow performance, Free cow traffic seems to be more favourable than Feed First traffic. Weather the differences are due to actual differences in traffic systems or management is hard to determine with the information available. Several management factors, such as feeding strategy, determine the success of cow traffic. These factors might be more important in order to obtain successful cow traffic than type of traffic system. More research, taking these factors into account, is needed to investigate the question further and to confirm the results in this study.