Skip to Main navigation Skip to Left navigation Skip to Main content Skip to Footer

University of Minnesota Extension

Extension is almost done building a new website! Please take a sneak peek or read about our redesign process.

Extension > Agriculture > Dairy Extension > Precision dairy > An update on precision dairy farming—2016

Print Icon Email Icon Share Icon

An update on precision dairy farming—2016

Marcia Endres

What is going on in the area of Precision Dairy? About 350 people attended the 1st International Precision Dairy Farming Conference in Leeuwarden, Netherlands June 21 to 23 to learn more about it. The program included a day of farm tours where attendees could see some of the technologies at work on dairy farms.

Many of the presentations and posters focused on individual cow sensor technologies that measure various parameters on the cows such as rumination, feeding behavior, lying time, activity, rumen pH, cow location in the pen, etc. One of the farms on our tour was using a sensor that measures eating behavior and it can also locate cows in the pen within 25 inches of accuracy. The presenter showed us on the tablet display where each of the ~200 cows in the group were located at the time and we could see the dots moving as a cow went from place to place at the feedbunk or she moved back to the freestall resting area. This sure makes it easier to find a cow that needs to be bred, treated, fetched, etc.

There seemed to be an increase on the number of studies that investigated the use of measurements in milk such as progesterone, enzymes, color, conductivity, somatic cell count, milk fat, milk protein, milk lactose, etc. as another option for improving health and reproductive performance of dairy cows. These options do not require cows to wear a sensor. Another option discussed by a few speakers was the use of video technology to measure things like body condition or lameness as cows exit the parlor or milking robot box.

There is growing interest in finding out how to best utilize all these technologies for early disease detection and going beyond heat detection. The challenge becomes developing appropriate algorithms that are based on a true gold standard. How many false positives or false negatives might we detect? What is more important to consider - whether we have a few extra cows to check that turn out to be healthy (more labor) or whether we miss some cows that are sick? We have seen some positive results on disease detection on farm and research studies, but more work is yet needed to optimize the use of these tools for improving animal health. As we have more and more cows housed in groups and labor becomes more difficult to find, these technologies will help improve animal welfare, especially as better algorithms are developed.

Another major area of discussion was automatic milking systems (AMS). Topics included improvements in AMS data collection, factors that influence milk production per AMS, daily data integration to provide the producer with action lists, labor efficiency, barn layouts, use of AMS in larger herds, etc. A Dutch study found that 2 or 3 robots per pen resulted in less time spent checking cows than 5 AMS per pen. In the larger herd study from Finland, most farms with more than 500 cows still used 2 AMS per pen, but one farm had 8 AMS all in one group – the time spent fetching and finding cows was much greater. They could benefit from using a cow location sensor! The same study found that larger herds (average size 644 cows) fetched less cows per AMS (3.0 vs. 4.8), had more milkings per day (2.85 vs. 2.7), and more milk per cow (81.2 vs 70.1 lb per day) than smaller herds (average size of 128 cows). When milking cows exclusively with AMS, the larger herds had a special group with 1 or 2 AMS used specifically for fresh cows with less cows per group and more comfortable housing. More time was spent with this group of cows. Most large herds had a separation area for cows at the exit of the AMS or commitment pen and this area was used for breeding, dry-off, herd health, etc. Most of the farms were not using these facilities very effectively. Fetching cows that do not come to the AMS voluntarily is a major labor component of AMS dairy farms. Larger herds fetched an average of 3 cows per robot, whereas smaller herds fetched an average of 4.8 cows per robot.

A study from Israel demonstrated the use of slow moving fences to improve attendance to the AMS and reduce number of cows fetched daily. There was an increase of 47% in the number of milkings per day and a 17% increase in milk yield when the fences were used. In terms of barn layout, a Dutch study reported that important factors that positively influenced AMS performance indicators included presence of split-entry waiting area to the AMS, right-angled alignment of the AMS to the feed bunk, and having a straw separation area for special needs cows. However, more of the variation was explained by factors not related to barn layout such as feed management.

These are just some examples of the research results presented at the 1st International Precision Dairy Farming Conference. The 2nd International Precision Dairy Conference will be held in Rochester, MN, in June 2019. The next U.S. Precision Dairy Conference will be held in Lexington, KY, May 31 to June 1, 2017. Mark your calendars!


July, 2016

  • © Regents of the University of Minnesota. All rights reserved.
  • The University of Minnesota is an equal opportunity educator and employer. Privacy