Mjölkkobete i norr: gårdsstudier och betesanalyser

A study was conducted on seven dairy farms in the county of Västerbotten and two in Västernorrland in northern Sweden. The farmers were interviewed by phone about their grazing strategies. The Västerbotten farms were visited at three occasions, in June, July and August during grazing seasons 2021 and 2022. Updates were made on production and grazing and registrations were made in the paddock where the cows were to graze next. A drone with mounted sensors taking RGB and multispectral photos was flown over the paddock. Thereafter GPS coordinates for twelve points per paddock were noted and registrations were made, including spectral measurements with a Yara-N sensor, and registrations of grass height and density with a simple rising plate meter, and also the highest stem with a ruler. A square of 0.5×0.5 m was cut at each registration point.
The cut material was taken to the lab for botanical analysis, weighing and dry matter analysis. Dried and milled samples were sent for chemical analysis of nutrient content. Samples were also scanned with a hyperspectral camera to see the potential of the technique as an alternative to wet chemistry analysis. To assess the digestibility and predict methane production, analyses with in vitro-gas production were conducted.
The interviews showed that there is a large variation in grazing management and strategies. For example, the grazing paddocks were only mown once per season at two farms, while at three other farms they were mown after each grazing instant. At three farms, the same seed mix was used both for grazing and harvesting fields, while the other farms used a special seed mix for the grazing paddocks. The area for grazing, including grazing of regrowth of harvested fields, varied from 0.12 to 0.50 hectares per dairy cow.
Data from the dairy on milk delivery did not show any decrease in production at the start of the grazing season at the farms in the study. However, calculations of milk delivery in summer compared to winter showed a somewhat lower production in summer.
The botanical analysis showed a large variation in species composition between registration points, paddocks and sampling times. For example, the clover content varied between 0 and 80 % between single registration points. The clover content increased during the season. The percentage of dead material varied between 1 and 41 % in samples merged per farm and sampling occasion.
The nutrient content varied between farms and sampling occasions. There was a tendency that the dry matter and NDF contents were higher and the crude protein and metabolizable energy contents were lower in July, compared to June and August. The field measurements made with Yara N sensor showed relatively good correlation between model calculated data from sensor readings and the lab analysed results. The same also goes for estimations made with the hyperspectral camera.
The highest values of digestibility and thereby also for methane production were observed in June. That was also the month when the percentage of grass was highest. The July samples had the lowest values for digestibility and methane. The preliminary estimations from the hyperspectral sensor showed promising correlations for digestibility and gas production. The calculations of the cows’ daily need to graze, based on indoor feed rations, showed a large variation around the average 5 kg DM. They also showed that the pasture was not always fully used.
The economical calculations showed that the cost for grazed pasture varied between 0.53–1.34 SEK per kg DM. The largest part was labour, but the variation was large between farms, depending on differences in amount grazed, need of machine work and use of artificial fertilizers.