Reducing dairy ammonia emissions: A Swedish case study

Ammonia Reduction is of growing importance throughout the world Ammonia in the atmosphere can cause acid rain 90% of ammonia emissions in Sweden originate from agriculture activities

A project at the research farm, Brogarden, at the Swedish University of Agriculture Sciences in Skara has shown that it is possible to reduce on farm level emission of ammonia considerably.

By applying known techniques and routines, ammonia released from the dairy farm has been reduced by as much as 70 %. The areas that have been studied are feeding, housing, manure handling, manure storage and spreading.

Ammonia in the atmosphere can cause acid rain and be harmful to both aquatic environments and plants. The Swedish Government has set up 15 environmental goals. One of these states that the ammonia releases to the atmosphere should be reduced by 15 % from the level of 1995 by? the year 2010. Ninety percent of all ammonia emissions in Sweden originate from agricultural activities with 70 % of this from milk production, Major sources of ammonia release on dairies include; ventilation from the buildings, manure storage and application, and from grazing animals. Storage and spreading of manure is subject to several regulations, such as storage capacity and spreading time. Storage for slurry should also have a crust in order to minimize gas releases. Spreading techniques using low emission systems are promoted and subject to financial support.

Rebuilding of the cowshed

The herd at Brogarden was housed in a cowshed built in 1918, equipped with long stalls for tied-up cows. After the first winter season (Sept-99 – April –00) the old stalling, manure handling system, gutters and floors were demolished and new short stalls, new milking, ventilation and manure handling equipment were installed. A new mixing pit for slurry and a manure store with canvas cover was built. The new short-stalls were 180 cm long and 120 cm wide. The rearmost 32 cm consisted of 4 aluminium bars with rubber covering, so called rubber slats, and the rest of the stalls were equipped with 20 mm thick rubber mats. The old computer controlled feed cars for roughage and concentrate were kept, but a new feed mill and a mixing unit for cereals was installed.

Special efforts were made to build the manure removal system in such a way that ammonia emission would be minimised. The gutter was sloped towards an urine channel covered with –perforated plates for efficient urine separation. The urine channel was equipped with an auger for mechanical cleaning. The manure was removed by a hydraulic operated scraper system. Urine and feaces were not mixed in the cross channel but taken out of the shed in separate systems and then mixed in the pit outside the building. The cows’ drinking water was running in plastic pipes placed, just under the surface, in the concrete floor of the manure gutter. The idea was to cool the surface and thereby reduce the ammonia emission. The temperature of the drinking water was increased by some 5-6 °C, which most likely was also positive for the cows.

The winter period before rebuilding (Sept. 1999 to April 2000) is referred to as the reference period and the next three winter periods as test periods. The projectwill finish in September 2003.

Figure 1. New short stalls equipped with rubber slats


The herd at Brogården consisted of 42 cows of the Swedish Red and White breed with an average production of 10,200 kg ECM (energy-corrected milk), with 4.2% fat, 3.4% protein and 4.5 mmol/l urea in the milk. Calves were taken to a neighbouring farm and brought back as pregnant heifers one month before calving. Feed rations were formulated individually for each cow according to her monthly milk yield and body-condition score. During the reference period (October 1999 – April 2000) the feed ration contained grass/clover silage, hay, small grains (70% oats and 30% wheat) and protein concentrate. During the first winter season after the reconstruction of the barn, the protein level in the feed ration was decreased, on average, 10 gram crude protein per kg of dry matter and the protein balance in the rumen was adjusted to be near 0. This was achieved by adding dried sugar-beet pulp to diets of high-yielding cows and increasing the proportion of small grains (40% barley, 40% wheat and 20% oats) in the diet. This ration was also fed during the second winter period. Table 1 shows milk production, total feed intake, dietary composition and nitrogen efficiency for cows yielding >35 kg ECM/day during the first three winter periods of the project. (Nadeau et al, 2003)

Table 1. Milk production, feed ration and nitrogen efficiency of high-producing dairy cows (>35 kg ECM).


Reference period



ECM1, kg/day




Milk fat, kg/day

1.7  1.6 1.7

Milk protein, kg/day

1.2 1.2 1.3

Milk urea, kg/day  

5.1 5.2 4.4

Total feed, kg DM/day

21.2 22.9 21.6

Forage, kg DM/day  

8.6 7.9 8.3

Concentrate, kg/day

12.6 15.1 13.3

Forage, % of total feed

41  35 38

Metabolic energy, MJ/kg DM

21.6 12.3 12.3

Crude protein, g/kg DM

170  160 165


7.7 8.2 7.7
PBV3, g/day 362 0   366

RDP4, g/kg DM  

109 95 113

RUP5, g/kg DM  

61 64 52

NDF6, g/kg DM  

375 360 373

Starch , g/kg DM  

172 190 187

Nitrogen efficiency7, %  

32 33 34

1 ECM= Energy-Corrected Milk.
2 AAT = Amino acids absorbed in small intestine.
3 PBV = Protein Balance in the Rumen.
4 RDP = Rumen-Degradable Protein.
5 RUP = Rumen-Undegradable Protein.
6 NDF = Neutral Detergent Fibre.
7 Calculated as N in milk / N in feed.

Of particular interest is the increased nitrogen efficiency from 32 to 34% for cows yielding >35 kg ECM per day. Milk yield increased by 1.5 kg ECM/day and total feed intake by 0.4 kg DM/day from 1999-00 to 2001-02 resulting in a 6% decrease in nitrogen contents of urine and feaces for high producing cows. Averaged over all lactating cows, including dry cows, nitrogen efficiency increased from 27 to 30%. Consequently, increasing milk yield, while controlling nitrogen intake, improved the nitrogen efficiency.

Ammonia emission from the building


A number of parameters were measured in the stanchion barn during the reference period. A data logger recorded the ammonia and carbon dioxide concentrations in the air, ventilation rate, outside and inside air temperatures and humidity. In general, the indoor climate was poor with a low ventilation rate and there were occasional high concentrations of ammonia and air leakage into the barn through a dung culvert. However, cow health and reproductive performance were generally good and were comparable with other Swedish herds.

Reference period

The ammonia release during the reference period 1999-2000 is presented in Figure 2.

The results of the measurements of ammonia emission and ventilation parameters are summarized in Table 2. The average ammonia emission from the cowshed was 24 g per cow per day. During the same period, the average ammonia concentration in the cowshed was 7.9 ppm.

The variation in ammonia emission during a day is illustrated by Figure 4. Two maxima of ammonia emission are seen, around 8 a.m. and the other around 5 p.m. (Gustafsson et al, 2003).

Figure 2. Release of ammonia during the reference period 1999-2000

Table 2. Results of the measurements of ammonia emission, ammonia concentration, ventilation rate, carbon dioxide concentration, inside and outside temperature during the reference period 1999-2000. Values of 110 daily averages

Parameter Average  Standard deviation Max - Min

Ammonia emission, g/cow and day  

24 4.6 37 - 13

Ammonia concentration, ppm  

7.9 3.2 13.9 - 2.1

Ventilation rate, m3/cow and h  

152 29 218 - 87

Carbon dioxide concentration, ppm  

2103 424 3592 - 1435

Inside air temperature, oC  

16.5 0.7 19.3 - 14.9

Outside air temperature, oC  

1.3 3.7 9.0 -

Figure 3. Ammonia emission during one day ( February 18, 2000) of the reference period.

Demonstration period

The ammonia release during the demonstration period 2000-2001 is presented in Figure 4. The rebuilding of the barn resulted in a lower emission. The total emission from the cows was reduced by about 20 %, from 24 g/cow/day to 19 g/cow/day. This is the result of an increased ventilation rate and low evacuation of air from the manure culvert and the urine drainage system preventing air leakage in combination with efficient urine separation and cooling of the manure gutter. The indoor climate improved considerably; the average ammonia concentration was lowered from 8 ppm before the re-design of the barn to 3 ppm.

Average values of ammonia release and concentrations, ventilation rate, carbon dioxide, inside and outside temperatures are presented in Table 3.

Figure 4. Release of ammonia during the demonstration period 2000-2001

Table 3. Results of the measurements of ammonia emission, ammonia concentration, ventilation rate, carbon dioxide concentration, inside and outside temperature during the demonstration period 2000-2001. Values of 56 daily averages

Parameter Average  Standard deviation Max - Min
Ammonia emission, g/cow and day 18 3.9 28 - 10

Ammonia concentration, ppm  

3.2 0.5 4.3 - 2.1

Ventilation rate, m3/cow and h  

346 70 477 - 200

Carbon dioxide concentration, ppm  

1537 347 2595 - 1012

Inside air temperature, oC  

13.5 2.3 17.7 - 6.5

Outside air temperature, oC  

1.3 3.7 9.0 -


The fans for exhaust air from the building were all placed in a separate room, from which the air was led into a channel and passed through a biofilter. The filter consisted of chopped wood and straw and had an irrigation system for keeping it wet. It measured 140 square meters and the thickness was 0.5 meter, giving a contact time between air and material of about 15 sec. The main part of the ammonia in the air was dissolved in the water as nitrate and mainly converted to gaseous nitrogen by the bacteria in the bed. Measurements of the outgoing air showed only traces of ammonia left after passing through the filter. The total nitrogen losses from the project are presented with and without the biofilter, see Figure 7.

Manure storage

During the reference year, the manure was stored as farmyard manure, the solids on a concrete pad outside the barn and the urine in a pit, covered with a floating layer of Leca pebbles. The losses from the storage for solid manure and urine were 6.5 and 5.0% respectively, calculated on total incoming nitrogen to the store. Losses of ammonia were strongly related to the ambient air temperature.

Figure 5. Losses from manure pad and urine pit.

Corresponding losses from the canvas-covered slurry storage were 0.2%. The new storage had capacity for almost one year of production and it was mainly built under the ground surface to provide low temperatures. The new slurry handling system with a canvas-covered storage reduced the ammonia losses during storage by about 95%, from 4.1 to 0.2 kg NH3-N per cow and year.

Figure 6. Losses from the roof-covered slurry store.

Manure spreading

Solid manure was spread in a traditional way with a broad cast spreader, partly in early spring, prior to planting season, and partly in the autumn after harvesting. Urine was spread by means of a spreader equipped with a boom with trailing hoses, placing the urine in bands on the ground, on grassland in the spring and summer.

Out of the total nitrogen spread, 12% was lost from the solid manure and 30% from the urine corresponding to a total loss of 10.4 kg NH3-N per cow and year.

Both band spreading and shallow injection were used for spreading of slurry in the first growing season after rebuilding of the barn. Handling the manure as slurry also meant a new strategy for spreading; about 60% of the total amount of slurry was band spread into growing wheat in the spring. The remaining 40% was injected on grassland in late autumn. The average losses were 5-6% of total nitrogen applied, which corresponds to 6.1 kg NH3-N per cow and year.

Total losses

During the first year of measurements after rebuilding (October 2000 – April 2001), the losses of nitrogen were reduced by 377 kg compared with the corresponding period during the reference year, corresponding to a reduction of nitrogen losses by 70%, provided all ammonia is absorbed in the biofilter. If the biofilter is excluded, the total reduction of nitrogen during the observation period was 231 kg, corresponding to a reduction of 44%. (Sannö et al, 2003).

Figure 7. Ammonia losses from Brogården prior to and after rebuilding.

The main reasons for the reduction were changes in the manure handling system inside the building, at storage and at spreading, as well as the biofilter. The losses per ton of milk produced have decreased from 2.30 to 1.35 kg NH3-N without biofilter and from 2.30 to 0.71 kg NH3-N with biofilter.


Financial support for the project was provided by LIFE-Environment within EU, Swedish Farmers´ Foundation for Agricultural Research, Arla Foods AB, Svenska Lantmännen AB, the Swedish Dairy Association, DeLaval AB, Svenska Foder AB, JTI - the Swedish Institute of Agricultural and Environmental Engineering, the departments of Agricultural Biosystems and Technology, Agricultural Research Skara, and Animal Environment and Health at the Swedish University of Agricultural Sciences, and Skara community.


G. Gustafsson, K-H. Jeppsson, J. Hultgren, J-O Sannö. 2003. Techniques to reduce ammonia release from cowshed with tied dairy cattle. Submitted to the International Symposiums on “Gaseous and odour emission from Animal Production Facilities in Bygholm, Denmark, June 1-4. 2003.

Monteny, G.J., Kant, P.P.H., 1997. Ammonia emission and possibilities for its reduction in dairy cow houses: A review of Dutch developments. In: Voermans, J.A.M., Monteny, G.J. (Eds.), Ammonia and odour emissions from animal production facilities. NVTL, Rosmalen, The Netherlands, Vol. 1, pp. 355-364.

Sannö, J-O. Nadeau, E. Karlsson, S. 2003. Reduction of ammonia emission in milk production - a practical study. Submitted to the International Symposiums on - Gaseous and odour emission from Animal Production Facilities in Bygholm, Denmark, June 1-4. 2003.

Sannö, J-O., Hultgren, J., Gustafsson, G., Jeppsson, K-H., Nadeau, E., Karlsson, S., Cederberg, C., Bergström, S., Henriksson, M. 2002. Minskning av ammoniakavgången från mjölkproducerande besättningar -Delrapport av projektet LIFE Ammoniak. SCB, 2003. Utsläpp av ammoniak till luft i Sverige 2001 (Emission of ammonia to air in Sweden in 2001). Statistiska meddelanden, MI 37 SM 0201.

Related Links:

Shabtai Bittman, Grant Kowalenko, Naveen Patni and Laurens van Vliet