Tuesday, August 4, 2015

Use and effectivity of sensor systems


Precision dairy farming is on the rise. For instance, in the Netherlands 19 % of the dairy farmers are milking with a milking robot. I am travelling home from the 3rdNorth American conference on Precision Dairy Farming, held in Rochester, MN (just as 2 years ago). I heard there that automatic milking is now really gaining momentum. One of the major brand of milk robots told me they were working over hours to produce all new systems that are to be installed in North America.

We had an interesting meeting in Rochester, mostly aimed at dairy farmers and extension workers. Marcia Endres (University of Minnesota) took the initiative for the meeting and did the main part of the organization together with Jeffrey Bewley (University of Kentucky). That meant that besides a number of academic presentations there were quite a number of more practical presentations. Farmers that presented their use of sensor systems and automation. The striking similarity between the farmers, who all seemed to be quite successful in the application of precision dairy farming systems, was that they all spent time behind the computer. They liked and worked with the information they could gain from the system. In order to successfully use precision dairy technology, you need to learn to use the system and spend daily time behind the screen. Precision dairy farming is more than only replacing labour with computers. Besides, you need to know your cows as well. You have to learn to link the information on the screen to the real life of the cows.

Next year, we are organizing the first truly International Conference on Precision Dairy farming in Leeuwarden, the Netherlands. I am already looking forward to it.

At the Rochester meeting I had the honour to be invited to give the opening presentation and was asked to provide insight in the situation in the Netherlands with regard to precision dairy farming. The slides of my presentation can be found on slideshare. Besides developments in of the Dutch dairy sector, the use of automatic milking systems and the Smart Dairy Farming project, results of two recently published papers of Wilma Steeneveld were used.

In the first paper, the use of sensor systems, both on farms with an automatic milking system as on farms with a conventional milking systems was investigated, alongside with reasons that farmers had to invest in these systems and the use of these systems. The study was done in co-operation with Accon-AVM, a large agricultural accountancy firm. Questionnaires were sent out to 1,672 farmers. In total, 512 responded. Of these 512, 212 (41 %) had a sensor system on their farm. Because we can expect a response bias (you are less tempted to respond to a questionnaire about sensor systems if do not have such a system), the proportion of Dutch dairy farmers with sensor systems is most likely lower than 41 %. It became apparent that especially in the last couple of years, farmers deliberately invested in sensor systems for estrus detection and rumination measurements. But it became also clear that quite some farmers have sensors on their farm that they did not deliberately choose for. They were a part of the new milking system. Consequently, these systems are not very well used.

In the second paper, the effect of sensor systems on milk production, days to first service and somatic cell count of the herd was evaluated. Data on production, health and reproduction over the years 2003-2013 was available for 414 of the 512 farms that responded, divided over farms with sensors in combination with an automatic milking system (n=103), with sensor systems in combination with a conventional milking system (n=49) and farms without (n=262) sensor systems. Having sensor systems was associated with a higher average production per cow on farms with an automatic milking system, and with a lower average production per cow on farms with a conventional milking system in the years after investment. Farms with an automatic milking system had on average 12,000 cells/ml higher somatic cell count after adoption of the sensor system whereas farms with a conventional milking system had on average 10,000 cells/ml lower somatic cell count.
Other developments in the Netherlands that are more than worth mentioning is the Smart Dairy Farming project. Besides work on precision dairy farming applications for reproduction, young stock rearing and metabolism, the most interesting and intriguing work is aimed at sharing of data. There is an enormous amount of data collected on dairy farms. The dairy farmer owns these raw data, but in practise they can only be used by the equipment collecting the data. There will be such more value in these data when others, after permission of the farmer, can use those data. This might give chain perspectives, such as the dairy industry being able to ensure the qualty of the production process as well as the feeding company being able to evaluate the effect of changes in the composition of composite feed. But also developers of decision support systems could be able to use data from multiple sources and thus make better applications. Interesting and challenging developments!!