Accounting for inter-annual <Emphasis Type="Italic">variability</Emphasis> of farm activity data for calculation of greenhouse gas emissions in dairy farming

Purpose

This study examines the inter-annual variability of production data in an organic dairy farm and its effect on the estimation of product-related greenhouse gas emissions (GHG) using a detailed material flow model. It is believed that the examination of only one production year may not adequately reflect temporal representativeness and may therefore lead to unreliable results. The current study also provides a method to deal with variability when temporal representativeness cannot be ensured.

Methods

All material flows related to milk production from six consecutive milk years in an organic dairy farm in northern Germany were analysed. The milk yield of the 75 to 91 cows varied between 5418 and 7102 kg energy corrected milk (ECM) per cow and year. GHG emissions were estimated using calculation guidelines from the International Dairy Federation (IDF) and the Intergovernmental Panel on Climate Change (IPCC). Emissions were calculated in the Flow Analysis and Resource Management (FARM) model ensuring mass balances for nitrogen and phosphorous in every subsection of the model. Based on the variability of crop yields, the number of years for representative average data was calculated as well as an uncertainty when only a limited number of years was available.

Results and discussion

Estimated GHG emissions varied between 0.88 and 1.09 kg CO₂-eq kg^?1 ECM^?1 (mean, standard deviation of the mean = 0.97 and 0.07 kg CO₂-eq kg^?1 ECM^?1). Emissions from ruminant digestion had the highest contribution (50.9 ± 2.3) percent in relation to overall product-related GHG emissions. Direct emissions from soil showed the highest coefficient of variation (36%) due to simultaneous changes in fertilization amount, crop yield and milk yield which showed no significant direct relationship. The number of years needed to be assessed for representative average yields was between 27 and 215 years for clover grass and maize silage, respectively. When performing a sensitivity analysis based on the variability of crop yields, the assessed farm showed reliable results with average data of at least 4 years.

Conclusions

Temporal representativeness should be dealt with explicitly in GHG assessments for dairy farming. If the representativeness of crop yields cannot be ensured, an uncertainty bandwidth of the results based on variability of yields can provide a basis for comparing different farms or farming systems. This approach could also be extended to other variabilities in dairy farming for more reliability of results.