Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry

Transcription

1 Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry

2 Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry Dairy Australia has produced this abridged version of the Final report: Farming Carbon Footprint of the Australian Dairy Industry written by PE Australasia and the Tasmanian Institute of Agriculture (August 2012). PE Australasia is part of PE INTERNATIONAL. The full version of the Final report: Farming Carbon Footprint of the Australian Dairy Industry can be accessed by contacting the Dairy Australia Library at or Australian Milk Production Carbon Footprint at the farmgate: by the numbers 1.11 kgco 2 eq/kg FPCM: average Australian milk production carbon footprint for 2009/10 The Australian milk production carbon footprint is one of the lowest internationally The Australian milk production carbon footprint is comparable to countries with advanced dairying industries 140 farms from six dairy regions participated in farm system study 15 manufacturing facilities participated in processing system study 57% farm emissions due to enteric fermentation II

3 Contents Executive summary Overview Carbon footprint results What do these results tell us? Conclusions Methodology Life cycle model descriptions Life cycle inventory Dairy Australia s sustainability work Nomenclature References Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry 1

4 Executive summary This is a summary of the detailed farm system project report on the final results of a carbon footprint study of the Australian dairy industry. This is the first of two final project reports and focuses on the methodology and results from the farming system only. This report contains the average carbon footprint results for the production of Australian milk at the farmgate. The scope of the system covers all inputs and outputs of dairy farming, and the results are presented for the functional unit of 1 kg of fat and protein corrected milk (FPCM) produced at the farmgate. A second related report focuses on the product manufacturing/processing system, and reports the carbon footprint results for a number of products that are representative of the Australian dairy industry. An industry cross-section of primary data has been collected from 140 farms and 15 manufacturing facilities across the country. The results define the carbon impact of Australian dairy products and allow farmers and manufacturers to understand the contribution their operations make to the greenhouse gas (GHG) emissions from the nine nominated products. Results for the production of Australian milk show the average carbon footprint at the farmgate for 2009/10 was 1.11 kgco 2 eq per kg FPCM. More than half of the emissions were associated with enteric fermentation (see Figure 1). Eight per cent were associated with manure management and 10% with manure excreted onto pasture. Fertiliser and energy inputs were each associated with 8%. Purchased feed was associated with 9%. Energy 8% Purchased feed 9% Fertilisers 8% Manure excreted on pasture 10% Enteric fermentation 57% Manure management 8% Figure 1: Relative contribution (%) of different farming phases to the average Australian milk production carbon footprint at the farmgate 2

5 1. Overview Introduction Around the world, there is increasing consumer demand for more information on the environmental impacts of products, and their influence on greenhouse gas (GHG) emissions. Agriculture is a notable contributor to global emissions of GHGs, and in particular methane and nitrous oxide. National and global studies have already concluded that the agricultural sector contributes 10 to 12% of overall global GHG emissions. Of this, livestock is assumed to be responsible for the largest part, at nearly 80% of global agricultural GHG emissions. Concerns over the role of livestock and the dairy sector s contribution to climate change have led to numerous scientific studies aimed at improving scientific knowledge about the sector s emissions, including measuring its carbon footprint and conducting life cycle assessment (LCA) studies in several countries. What is a carbon footprint? A carbon footprint is a comprehensive scientific assessment of the GHG contribution of a product or service. There are specific international standards for measuring a carbon footprint, as well as specific dairy industry carbon footprinting guidelines published by the International Dairy Federation (IDF 2010). What is a life cycle assessment (LCA)? LCA offers the most comprehensive method to quantify environmental impact (including the carbon footprint) over the whole life cycle of a product. There are several international standards for LCA to ensure consistency in its measurement. Why conduct this study on the Carbon Footprint of the Australian Dairy Industry? The primary reason for this study is to: satisfy increasing demand for environmental information on products by quantitatively evaluating the environmental impacts, and in particular the carbon footprint, of major Australian dairy products. The study aims to: quantify the carbon footprint of major Australian dairy products (products are listed in Table 3), from farm to distributor s warehouse or export harbour develop weighted average Australian product carbon footprints and, if possible, differentiate these by farm practices establish an auditable monitoring system and framework for a reproducible carbon footprint reporting system that can be updated and expanded to include other environmental indicators. Other reasons for the study include: obtaining a detailed picture of life cycle inventory data for the industry, and quantifying material and energy inputs, emission releases and material outputs identifying environmental impacts of products for the entire Australian dairy industry, differentiated by farm practices, including GHG emissions and total energy use (fuel and electricity) that represents the Australian average, based on annual production figures generating a reliable basis for carbon footprinting, environmental product reporting and product labelling (e.g. environmental product declarations; carbon reduction label) supporting strategic decision-making to promote a more competitive Australian dairy industry and to build the basis for sustainable leadership of the dairy industry in Australia and internationally using the results to help assess the overall performance of the entire dairy industry in order to identify resource use, emissions hotspots and areas of focus for potential improvement along the supply chain. How is this study being reported? Two separate final project reports have been produced. One focuses on the farm system and the carbon footprint results for the production of average Australian milk. The other focuses on the product manufacturing/processing system and the carbon footprint results for nine products that are representative of the Australian dairy industry. What does this summary contain? This is a summary of the detailed final project report for the farm system only, including the methodology and results. It contains the average farmgate carbon footprint result for the production of Australian milk for the 2009/10 financial year. Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry 3

6 2. Carbon footprint results Carbon footprint of average Australian milk production at the farmgate The average carbon footprint of the 140 sampled farms for 2009/10 was calculated to be 1.11 kgco 2 eq per kg FPCM. This was the weighted mean of the annual milk production per farm. Weighting was used to normalise the results, as in some instances larger producers have a greater effect on the overall mean. The sample error or precision was +/- 8%. This means the average carbon footprint for the population was within a range of +/- 8% of the 1.11 kgco 2 eq (1.02 kgco 2 eq to 1.20 kgco 2 eq). The standard deviation of the average carbon footprint was kgco 2 eq (24%). The lower median (1.08 kgco 2 eq) and positive skewness (1.5) show that the sample was slightly biased towards a small number of large producers. Table 1: Average Australian milk production carbon footprint results at the farmgate Average carbon footprint 1.11 kgco 2 eq/kg FPCM Median 1.08 Standard deviation Skewness 1.5 Range 1.49 Minimum 0.78 Maximum 2.27 Impact of different farming phases on the average Australian milk production carbon footprint Impact of enteric fermentation Enteric fermentation was the most significant contributor to the carbon footprint, comprising 57% of all GHG emissions. Impact of manure management and manure excreted on pastures The overall contribution of emissions associated with manure was 18% (0.19 kgco 2 eq). Manure management, which includes storage and application to arable land and pastures, accounted for 8% of the carbon footprint (2% nitrous oxide and 6% methane). Emissions associated with manure excreted onto pastures during grazing accounted for 10% of the carbon footprint. This comprised nitrous oxide (9%) and methane (1%). Impact of fertiliser Fertiliser use accounted for 8% of the carbon footprint. Ninety per cent of the farms used mineral fertilisers and soil conditioners, and 15% used imported organic fertilisers. The production and supply of mineral fertilisers is an energy-intensive process and was responsible for 5% of the carbon footprint. The corresponding nitrous oxide emissions from the field during the application and breakdown of nitrogen fertilisers were associated with 3% of the carbon footprint. Purchased feed: forage 1% Purchased feed: concentrate 8% Energy 8% Fertiliser supply 5% Fertiliser application 3% Manure: methane from grazing 1% Enteric fermentation 57% Manure: nitrous oxide from grazing 9% Manure: methane from storage 6% Manure: nitrous oxide from storage/application 2% Figure 2: Relative contribution (%) of different farming phases to the average Australian milk production carbon footprint 4

7 Impact of energy Eight per cent of the carbon footprint was associated with electricity supply used for pumps and mills, and fuels for generators, tractors, other engines and auxiliary materials. Though there is significant variation between farms, the average amount of diesel consumed to produce 1 kg of milk was litres and the average amount of electricity was kwh. Impact of feed Purchased feed accounted for 9% of the carbon footprint, comprising forages (1%) and concentrates (8%). By-products such as brewers grain and fruit pulp were only used by 9% of the farms. As their environmental burden is comparatively low the overall impact on the carbon footprint is less than 1%. Impact of milk yield Figure 3 shows the correlation between the carbon footprint and the average annual milk yield. For farms with a higher milk yield, the overall farm GHG emissions were distributed over a larger amount of milk. Consequently the carbon footprint decreased. Table 2 groups the participating farms according to the average FPCM yield per cow. Most of the farms achieved an annual FPCM yield of 6,000 7,000 kg. The carbon footprint decreased with a higher FPCM yield. The highest use of grain and concentrate was in the milk yield group above 8,000 kg. The other groups had a similar concentrate input. Impact of concentrate and land use change Table 2 also shows that the amount of grain and concentrate input per kg of milk was very similar. The highest input of grain and concentrate can be found in the milk yield group above 8,000 kg. For some of the farms, the purchased grain and concentrate mix contains soy and palm kernel products. In this study we assumed that soy and palm kernel products were imported from Brazil and Malaysia and were associated with emissions from land use change (LUC; see Land use change on page 12). This assumption is realistic for palm kernel products, while for soya this is a worst-case scenario, as some soy is imported from the United States or grown in Australia and therefore has no associated impact from LUC. The countries from which the soya is sourced changes from year to year. In 2007, most Australian imports came from Brazil, while in 2009 most came from Argentina and the United States. Carbon footprint (kgco 2 eq/kg FPCM) R2 = ,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 Figure 3: Correlation of milk yield and carbon footprint Table 2: Average carbon footprint according to milk yield per cow Average annual milk yield (kg FPCM) Average annual milk yield, kg per cow < 5,000 kg 5,000 6,000 kg 6,000 7,000 kg 7,000 8,000 kg > 8,000 kg Carbon footprint kgco 2 eq per kg FPCM Number of farms kg concentrate per kg milk Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry 5

8 3. What do these results tell us? Carbon dioxide from land use change 2% Nitrous oxide 18% Methane 63% Carbon dioxide 17% The average Australian dairy industry milk production carbon footprint at the farmgate for 2009/10 was calculated to be 1.11 kgco 2 eq per kg FPCM with a standard deviation of kgco 2 eq. Enteric fermentation While GHG emissions differ considerably between individual farms, emissions from enteric fermentation always dominate the carbon footprint, including 57% of the average Australian milk production carbon footprint. These emissions cannot be completely avoided as they are a product of the natural digestion process of the cow. Emissions are impacted by the number and proportion of nonmilk producing animals (calves, heifers and bulls) on the farm. A higher proportion of non-milk producing animals will lead to higher GHG emissions per litre of milk produced. Emissions are impacted by the quality of feed used on the farm. Figure 4: Contribution of single greenhouse gases to average Australian milk production carbon footprint Figure 4 shows the contribution of individual GHGs to the carbon footprint of milk. On average, 2% of these were related to LUC. 109 out of 140 farms were using palm kernel or soy products in their feed ratio. The carbon footprints of 81 farms had less than 2% associated with LUC, 21 farms had an impact of 2% to 5%, and five farms had an impact of between 5% and 26%. Impact of key assumptions Key assumptions have been made that impact the carbon footprint and have to be considered when bringing the results into context. Two key assumptions were identified within this study: 1. The factor used to allocate farm emissions to milk and meat: An Australian industry-specific allocation factor based on collected primary data is used. This factor is higher than the IDF (2010) default factor but more accurate as it is based on actual data of the analysed sample. 2. Uncertainty related to the country of origin of soy production and directly associated GHG emissions from LUC in these countries: A worst-case scenario was applied as the actual origin of soy products was unknown. Manure The prevalence of grazing on the participating farms means emissions associated with manure (0.12 kgco 2 eq per kg FPCM) were largely related to manure excreted directly onto pasture. Manure management (storage and application) was associated with 0.09 kgco 2 eq per kg FPCM, with most emitted as methane from manure storage. Manure management is an important part of farm nutrient budgeting and is closely related to fertiliser inputs. Mineral fertiliser production and application were associated with 8% of overall emissions. Feed and land use change Purchased concentrate and grain were associated with 8% of overall GHG emissions. One quarter of these were related to LUC, although for individual farms that used the relevant feed stocks the LUC contribution to GHG emissions may be more significant. As discussed under Land use change (page 12), we have assumed a worst-case scenario that all soy and palm kernel products are from countries associated with LUC impacts. An efficient way to reduce the footprint of concentrate input would be to avoid concentrates sourced from countries associated with LUC. Energy use Energy use represented 8% of overall GHG emissions, with high variation of energy use emissions between farms. Data collection process The use of a web-based data collection tool directly linked to LCA software enabled an efficient data collection and analysis. Close collaboration with agricultural experts and farmers during the development and testing of questionnaires and the LCA model contributed greatly to creating a reliable carbon footprinting tool for the dairy industry. 6

9 4. Conclusions Results: The average Australian milk production carbon footprint at the farmgate for 1 kg FPCM was 1.11 kgco 2 eq. The average Australian milk production carbon footprint at the farmgate was found to be within a narrow range of result values ( kgco 2 eq) reported by 50% of previous carbon footprint milk studies, according to the 2009 IDF global literature review. Enteric fermentation was the largest contributor to the carbon footprint at 57%, and is influenced by a number of factors including milk yield and feed conversion efficiency. The carbon footprint value is not simply a function of concentrate ratios or milk yield. Many factors influence GHG emissions in the farming phase. Hotspot analysis and reduction potentials can be best derived at an individual farm level. The impact of LUC on the average carbon footprint is low, however for individual farms that use the relevant feed sources (soy and palm kernel products) it can be significant. Inventory and analysis: Average carbon footprints for farm practices were calculated and analysed based on each farm s annual productivity per cow and concentrate requirements to produce 1 kg FPCM. The analysed sample of 140 farms achieved a precision of +/- 8% in relation to all Australian farms. This is in line with the goal of the study. Results were calculated for farms in six of the eight Australian dairy regions. It was decided to exclude data collection from farms in far north Queensland and Western Australia as it was not statistically significant. Data from a sub-sample of farms were analysed to give insight into the complexity and range of issues on different farms. How are the results being communicated? This summary provides an overview of the detailed farm system project report. The full farm system project report provides an in-depth discussion of the study s methodology and processes. The full product manufacturing/processing system report is to be released separately. Work is also underway to publish the methodology and final results of this study in peer-reviewed scientific journal articles. Customised reports are also being prepared for each farm that participated in the study. These individual reports will be confidential to the farmer and will assess individual farm carbon footprint results against the average Australian milk production results. Critical review As this study is not being used publicly for comparative purposes, in accordance with the ISO international standard it will be critically reviewed from internal and external perspectives to increase its validity and credibility. This includes: a critical internal peer review conducted by Peerasak Sanguansri of CSIRO an external peer review conducted by two independent international experts in LCA of dairy production. Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry 7

10 5. Methodology The global dairy industry has developed industry-specific carbon footprinting guidelines through the IDF (2010). These guidelines provide a consistent methodology to conduct carbon footprint assessments for the dairy sector. The IDF guideline (2010) is the underlying methodology adopted for this study. Dairy production systems The overall dairy production system can be subdivided into two separate systems (see Figure 5). These are farm, which includes on-farm production of milk; and processing plant, which includes the processing/ manufacturing of milk and other ingredients into dairy products. Functional units There are two functional units used in the overall study; one for milk at the farmgate, and one for processed dairy products. The functional unit for the farm system is: 1 kg of fat and protein corrected milk (FPCM) at the farmgate. FPCM is calculated by multiplying milk production by the ratio of the energy content of a specific farm s milk to the energy content of standard milk with 4% fat and 3.3% true protein content. The dairy products investigated in the separate final report on product manufacturing/processing are listed in Table 3. These are not discussed further in this report. Farm system boundary The study required a boundary to be established to define inputs to the farm system. These boundaries were: The foreground system, which includes: all on-farm processes related to dairy production such as animal husbandry, on-farm feed production, pasture management, manure management and fuel combustion. The background system, which includes all relevant inputs to the foreground system including: off-site feed production production of synthetic fertilisers, fuels and electricity Table 3: Dairy products investigated in the product manufacturing/processing study Category Product Quantity Packaging Fresh products Full cream milk 2 L HDPE bottle Flavoured milk, chocolate 600 ml 600 ml Tetra Brik Fruit yogurt 200 g 200 g plastic cup Milk powder FCMP and SMP 25 kg Multi-wall kraft paper sack Butter Bulk butter 25 kg Plastic film Retail butter 250 g Parchment wrap Cheese Bulk cheddar 20 kg Plastic bag Retail cheddar 1 kg Plastic wrap Whey Whey protein concentrate WPC kg Multi-wall kraft paper sack Inputs Inputs Farm Milk kg FPCM Processing plant kg product Outputs Outputs Figure 5: The two dairy production systems 8

11 transport of fodder, fertilisers, chemicals and fuels to the farm water supply. Transportation of raw milk to the processing plant is allocated to the processed dairy products system. Sampling of farms A sample of 140 farms was investigated. The level of precision in relation to all Australian farms was calculated to be +/-8%, in line with the goal of the study to achieve a precision rate of less than 10% for all Australian farms (Table 4). The level of precision or sampling error determines the closeness with which the sample predicts where the true values in the population lie. The difference between the sample and the real population is the sampling error. Table 4: Sample size and level of precision achieved Total number of Australian farms (Source: van Buuren 2010) Farm sample size (n) Precision achieved (+/-%) 7, % Allocation In LCA studies, it is important to allocate environmental impacts between the main product and by-products of an agricultural system. In this study, allocation on the farm is required at the following stages: Production of feed. Production of milk at the farm. Allocation procedures followed IDF (2010) guidelines based on the principles outlined below. Many feed ingredients are co-products from a production system generating more than one product, and therefore the environmental burden should be distributed between the co-products. Allocations were performed at the following stages of feed production: Soybean meal (co-product of soy oil and soy hulls production, produced from soy beans). Canola seed meal (co-product of canola seed oil production, produced from canola seed). Palm kernel extract (co-product of palm kernel oil production, produced from palm kernels, which is a co-product to palm oil, produced from oil palm). Wet and dry distillers grain (co-product of beer fermentation, produced from grain). For the dairy farm system where the main focus is on milk production, the meat generated from surplus calves and cull dairy cows is also a co-product that needs consideration. It is therefore necessary to determine the: total emissions and allocate them between milk and meat production systems sum of live weight of all animals sold, including bull calves and culled mature animals produced at the farm sum of all milk (fat and protein corrected milk) produced at the farm. The allocation factor (AF) is then calculated for each individual farm based on the following equation developed by the IDF (2010): AF = * R R = kg livestock/kg FPCM The equation is based on a farm level survey on 531 farms in the United States and estimates the quantity of feed energy required to produce milk versus beef. This study used available Australian data to calculate the energy required to produce meat and milk. The average result for the allocation factor was 89.84%. This means that nearly 90% of emissions are attributed to milk production (see Table 5). Table 5: Allocation factor between milk and livestock at farmgate in Australia Average allocation factor 89.84% Standard deviation Median 89.86% Skewness Range 32.12% Minimum 67.48% Maximum 99.60% Impact assessment methodology and impact categories considered This study focuses on the carbon footprint, however the data necessary to calculate additional impact categories, such as eutrophication potential and primary energy demand, were also collected and can be used in future investigations. Data on resource consumption, such as primary energy demand, water and land use, were also collected for potential future analysis. Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry 9

12 Global warming potential (GWP) The current Intergovernmental Panel on Climate Change (IPCC) global warming potential (GWP) factors have been applied to convert non-carbon dioxide emissions into carbon dioxide equivalent emissions (CO 2 eq). The two most common factors are 1 kg methane = 25 kgco 2 eq and 1 kg of nitrous oxide = 298 kgco 2 eq (IPCC 2007). Cut-off criteria Ninety-five per cent of the likely life cycle emissions are covered by the system being studied. Where processes are excluded from the analysis, a justification is provided. The cut-off rules are consistent with the goal and scope of the study as well as the IDF (2010) methodology. Data requirements, assumptions and limitations Data collection followed accounting principles as defined in relevant guidance documents. The main assumptions and limitations, such as time period of collection, technology, completeness, accuracy, precision, geographical coverage, consistency, transparency and reproducibility are defined below. Time period Inventory data collected and results represent the 2009/10 financial year. Completeness, accuracy and precision Details on the number of farms sampled and the level of precision are outlined under Sampling of farms on page 9. Geographical coverage This study covers annual production volumes for six of the eight Australian dairy regions. Consistency The methodology and data are consistent across all areas of investigation. All deviations from this principle have been documented and reasons given. Transparency and data collection A customised, transparent and fully auditable web-based software system was developed to ensure data quality was maintained during collection and modelling. This system featured customised questionnaires and enabled farmers to submit their data securely by remote access. This system allowed easy entry of data on an annual basis, and included a system for quality checking, data entry tracking and instant validity checking. The system also tracked data modifications. Dairy company field officers worked with farmers to collect and submit the farm data directly from farms. Reproducibility, updateability and documentation There is a need to ensure that data can be easily reproduced and updated by an independent practitioner. All data sources have been documented in detail in order to ensure reproducibility. Assumptions All assumptions made in this study have been documented. Sub-sample study A study of a sub-sample of farms was conducted to give insight into the complexity and divergence of issues on different farms. This study is discussed in the detailed farm project report. Note: for a more detailed discussion on the methodology of the study, consult the full farm project report. Technology The current technologies and agricultural methods used in Australia on the farm, for transportation and at manufacturing sites are the basis for the study. 10

13 6. Life cycle model descriptions Data sources All impacts related to activities on the farms have been calculated based on primary data. Impacts associated with fuel and electricity inputs are included using National Greenhouse Accounts Factors (NGA, DCCEE, 2010) data. Impacts related to farm inputs are either from the GaBi LCA database 1 or literature data. PE Australasia has undertaken a review of previous studies and analysis of existing life cycle inventories to establish the current status of data relevant to the study and to identify gaps. Data from existing studies have been used to validate data collected in this study. Input materials and farm operations Feed Animal feed is partly produced on-farm and partly purchased. It can be divided into concentrate feed and forages. Concentrate feed can be individual ingredients (e.g. cereals or pulses) or completely balanced supplements (e.g. pelletised). Seventy-four per cent of forages were produced on-farm, and can be pasture, hay or silages. GHG emissions related to on-farm feed production are accounted for in the emissions from fertilisers and energy. GHG emissions associated with the supply of forages, concentrates and other feed imported to the farm are accounted through extensive data sets provided by GaBi database (PE: GaBi 5). A detailed list of feedstock considered is included in the full farm project report. The feed uptake via grazing is estimated by calculating the difference between the available energy from feedstock reported purchased, and energy requirement of the relevant livestock class. Fertiliser GHG emissions associated with the supply of mineral fertilisers are taken from the GaBi database (PE: GaBi 5). Organic fertilisers imported to the farm are treated as waste products; for example, burdens are not allocated to the dairy production system. Methodology for calculating nitrous oxide emissions associated with all fertiliser application is described under Agricultural soils on this page. 1 PE Australasia has developed the GaBi database, which is an in-house, extensive data set of various processes providing more than 2,000 cradle-to-gate material data sets, 8,000 intermediary chemical process models, and thousands of LCA projects from quality-controlled industry projects. The GaBi database is the most comprehensive, up-to-date life cycle inventory database globally. It is available, well known and consistent with the European Commission s European Reference Life Cycle Database. Australian data for energy are included in the GaBi database. All datasets included on the GaBi databases are consistent (e.g. same boundaries, allocation rules, modelling principles) and are clearly documented. Energy State-specific emission factors are used to account for electricity and natural gas supplied to farms (DCCEE 2010). National averages are considered for the supply and combustion of diesel, gasoline and LPG as listed in DCCEE (2010). Unrecorded diesel consumption of field operations (associated with operations carried out by external contractors) is estimated according to USDA (2000) and Hanna (2005) (See the full farm project report). Water GHG emissions associated with on-site pumping of water are assumed to be captured in energy use. Data for water supplied from town supply are taken from the GaBi database (PE: GaBi 5). Agricultural soils Nitrous oxide emissions from managed soils occur naturally as intermediate gaseous emissions in the reaction sequence of nitrification and denitrification. Nitrous oxide emissions are calculated following the IPCC (2006) tiered approach using DCCEE emission factors. This tiered approach accounts for direct emissions (Tier 1), and indirect emissions (Tier 2) from leaching, runoff and atmospheric deposition of reactive nitrogen. Data on nitrogen application rates of mineral and organic fertilisers are based on primary data. Animal husbandry Methane enteric fermentation Enteric fermentation of cattle is the digestive process during which carbohydrates are broken down in an anaerobic microbial fermentation by micro-organisms in the rumen of the animals. Methane is created as a by-product and mostly emitted through the mouth of the cow. Methane emissions are calculated following the DCCEE Tier 2 approach and are dependent on the characteristics of the animal feed used on a particular farm, as well as the amount of feed taken in. Feed intake is calculated according to DCCEE (4A.1a). The decisive feed quality parameter is the mean feed digestibility. Mean digestibility is calculated from the actual feed intake of milkers and other stock classes. Default digestibility values for different forages and concentrates are deposited in the model (see detailed full farm project report for a complete list). Methane manure management According to the IPCC (2006), methane emissions from manure management depend mainly on the amount of volatile solids excreted, the specific manure management system used on the farm, and the region-specific methane conversion rate, which is determined based on average yearly temperature. Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry 11

14 The amount of volatile solids is calculated based on stock class-specific excretion rates. Excretion rates are calculated following the DCCEE Tier 2 approach (4.B.1). Methane conversion rates are taken from IPCC (2006). The Australian manure management systems that were considered in the analysis and methane conversion rates are outlined in the detailed full farm project report. The amount of urine and faeces going to each system is estimated based on the percentage of time each stock class is held in each system. If a solid-liquid separation system for liquid manure is used, it is assumed following Birchell (2008) that 20% of volatile solids are separated and stored as solid manure. Nitrous oxide manure management Excreted dung and urine causes nitrous oxide emissions during the breakdown of ammonia and other forms of nitrogen (nitrification and denitrification) on grazing land, and during manure storage and application of manure on arable land and pastures. Nitrous oxide emissions from manure management depend on the amount of nitrogen excreted as faeces and urine and the type of manure management. The DCCEE (4.B.1) Tier 2 algorithms and emission factors are used for calculations. The amount of nitrogen excreted is calculated based on the herd s protein intake through feed and the protein required for growth, reproduction and milk production. Crude protein intake is calculated by taking the actual feed intake of milkers and other stock classes. Values for default crude protein contents of forages and concentrates are taken from Dairy Australia (2010). The amount of urine and faeces entering each manure management system is estimated based on the percentage of time each stock class is held in each system. Land use change Land use change (LUC) is a challenging and complex area of the LCA process. According to the IDF (2010), GHG emissions arising from direct LUC, i.e. transformation from non-agricultural to agricultural land, shall be included in the LCA of dairy products. The assessment of the impact of LUC should include all direct LUC occurring on or after 1 January Five per cent of the total emissions arising from LUC should be included in the GHG emissions of these products in each year over the 20 years following LUC. All agricultural area in Australia was assumed to have been under cultivation before Thus the assumption is that the dairy sector is not responsible for direct LUC emissions. However, some of the imported feedstock is cultivated in areas where LUC occurs. This was assumed to be the case for palm kernel meal and soybean products. The emission factors considered in the LCA are given in Table 6. Table 6: Emission factors for LUC for affected feedstock Product Emission factor Source Palm kernel 1.02 kgco 2 eq/kg product Carlton 2011 Soybean meal 8.66 kgco 2 eq/kg product PE: GaBi 5 Soybean cake (Brazil) 7.69 kgco 2 eq/kg product FAO

15 7. Life cycle inventory Data quality, review and verification Incoming data were checked on a weekly basis. For every farm the following checks were conducted: Consistency of milk yield and herd size. Validation of protein and fat content of milk. Check of GWP factors. Consistency of herd size heifers and herd size milkers. Plausibility of concentrate ratios. Whenever a data issue was discovered, field officers were asked to revise or confirm the values in question. The checks were then repeated to validate that the issue was resolved. 148 farms provided completed questionnaires. Manual data correction was needed for six farms. Data from eight farms were removed as the data quality was considered too inconsistent to be included in the study. As a result, 140 farms were analysed in the study. Table 7 gives an overview of the farm data. Farm size ranged from 40 to 1,795 ha, with an average farm size of 261 ha. The standard deviation shows that the majority of the farms lie within a range of 216 ha. As the median and the mode are below the average farm size and the skewness is positive, a small number of large farms tend to bias the average. The average herd size and total milk sold show a similar tendency. The average annual milk yield is evenly distributed around the average (median and mode lie close to the average and skewness is close to zero). DGAS approach, and differences between the DGAS and LCA approaches to estimating the GHG emissions of dairy farms PE Australasia and the Tasmanian Institute of Agriculture comprehensively compared the approach applied in this study (within this comparison it is referred to as the LCA approach) and the Dairy Greenhouse Gas Abatement Strategies Calculator (DGAS) approach. The comparison aimed to review the LCA model and also give a detailed overview of the differences between the two approaches. For further details see the full farm project report. The DGAS calculator aims to quantify the carbon footprint of milk production as per National Greenhouse Gas Inventory (NGGI) methodology and the Kyoto Protocol. DGAS can also inform methodology development for the Australian Carbon Farming Initiative, as there is no difference between the calculations supporting DGAS and the calculations supporting the NGGI and so meets requirements for Kyoto-compliant carbon credits. The main differences between the DGAS and LCA approaches to quantify on-farm activity emissions are: 1. the multiplication factors used for methane and nitrous oxide. DGAS uses 21 (methane) and 310 (nitrous oxide), compared to 25 (methane) and 289 (nitrous oxide) in this study 2. the way emissions from upstream activities (production of farm inputs) are considered. Table 7: Overview of farm data Farm size (ha) Herd size milkers Total milk sold (kg/year) Average milk yield (kg/year) Average ,681,127 6,248 Median ,503,196 6,221 Mode ,870,705 6,371 Standard deviation ,245,749 1,439 Skewness Range 1, ,478,749 6,574 Minimum ,727 3,090 Maximum 1,795 1,000 6,714,476 9,665 Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry 13

16 8. Dairy Australia s sustainability work Dairy farmers and processors have a strong track record of working to improve financial, environmental and social outcomes for the health of the environment, workforce and the broader community. They understand that sustainability and industry profitability are interdependent. The large number of sustainability initiatives the industry currently undertakes demonstrates this. Dairy Australia invests in a number of key programs and initiatives aimed at helping farmers and processors with GHG emission reduction strategies. These include: Smarter Energy Use on Australian Dairy Farms This project will provide farmers with information and technical support to improve farm energy efficiency including funding for 900 energy audits. Funds are provided by the Department of Climate Change and Energy Efficiency and the Department of Agriculture, Fisheries and Forestry. Research into strategies to reduce greenhouse gas emissions whilst increasing productivity Specific projects include: evaluating the methane mitigation potential of a range of feed supplements and pasture forages improving nitrogen use efficiency on dairy farms for productivity and sustainability gains potential for Smart-N technology to reduce nitrogen fertiliser use. More information on practices to reduce greenhouse gas emission can be found in the Greenhouse Gas Emissions module of DairySAT at 14

17 Nomenclature Term AF CH 4 CO 2 CO 2 eq CSIRO DCCEE DGAS EP FCMP FPCM GHG GWP ha HDPE IPCC kg LCA LUC N 2 O NGA SMP WPC Explanation Allocation Factor Methane Carbon Dioxide Carbon Dioxide Equivalent Commonwealth Scientific and Industrial Research Organisation Australian Government Department of Climate Change and Energy Efficiency Greenhouse Gas Abatement Strategies Calculator Eutrophication Potential Full Cream Milk Power Fat and Protein Corrected Milk Greenhouse Gas Global Warming Potential Hectare High-Density Polyethylene Intergovernmental Panel on Climate Change Kilogram Life Cycle Assessment Land Use Change Nitrous Oxide National Greenhouse Accounts Factors Skimmed Milk Powder Whey Protein Concentrate Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry 15

18 References The following references were used in the full farm project report and, where relevant, this summary. Birchell 2008 BOM 2010 Carlton 2011 Cederburg 2004 DA 2010 Birchall, S. Dillon, C. Dillon, W. Wriley, R: Effluent and Manure Management Database for the Australian Dairy Industry. Dairy Australia. Average annual & monthly maximum, minimum, & mean temperature. The carbon costs or palm kernel expeller and its contribution to the dairy carbon footprint in New Zealand. PDF available to download at: Life cycle inventory of 23 farms in south-western Sweden. SIK Report The Swedish Institute for Food and Biotechnology. DA Fact Sheet 5: Feed value varies in different feeds. Nutritional values of common feed. Look up tables available under Nutritional values of common feed at: DairyCo Greenhouse gas emissions on British dairy farms. DairyCo, Kenilworth. UK. February (2012) aspx. DCCEE 2009 DCCEE 2010 DPI 2009 FAO 2010 Flysjö 2012 Hagemann 2011 Hanna 2005 IDF 2009 IDF 2010 IKP 2003 IPCC 1995 IPCC 2000 IPCC 2006 IPCC 2007 National Inventory Report 2007 Volume 1. The Australian Government Submission to the UN Framework Convention on Climate Change May (Department of Climate Change and Energy Efficiency: Canberra, Australia); National Greenhouse Accounts (NGA) Factors. (Department of Climate Change and Energy Efficiency: Canberra, Australia); Nutritive Value of Feeds (Database). NSW Government, Primary Industries Agriculture. content_src=%2bdxjspwh0dhalm0elmkylmkz3d3dplmfncmljlm5zdy5nb3yuyx UlMkZ0b29scyUyRmZlcyUyRnJlbW90ZS5odG1sJTNGYWN0aW9uJTNEc2VhcmNoJmFsbD0x. Greenhouse Gas Emissions from the Dairy Sector: A Life Cycle Assessment. Food and Agriculture Organization, Rome, Italy. Flysjö A, Cederberg C, Henriksson M, Ledgard S. The interaction between milk and beef production and emissions from land use change Critical considerations in life cycle assessment and carbon footprint studies of milk. Journal of Cleaner Production (2011), doi: / j.jclepro Hagemann et al. Benchmarking of greenhouse gas emissions of bovine milk production systems for 38 countries. Animal Science and Technology (2011) Fuel required for field operations File A3-27 Iowa State University available at agdm/crops/pdf/a3-27.pdf. Environmental / ecological impact of the dairy sector: literature review on dairy products for an inventory of key issues. Bulletin 436/2009 of the International Dairy Federation: Brussels, Belgium. A common carbon footprint approach for dairy- The IDF guide to standard lifecycle assessment methodology for the dairy sector. International Dairy Federation: Brussels, Belgium. Documents/ A-common-carbon-footprint-approach-for-dairy.pdf. Institut für Kunststoffprüfung und Kunststoffkunde der Universität Stuttgart, Abteilung Ganzheitliche Bilanzierung, Climate Change 1995: The Science of Climate Change (Eds JT Houghton, LG Meira Filho, BA Callander, N Harris, A Kattenberg, K Maskell) pp 588 (Cambridge Press: Cambridge, UK). Good Practice Guidance and Uncertainty Management in National Greenhouse Gas Inventories. Available at IPCC Guidelines for National Greenhouse Gas Inventories. Prepared by the National Greenhouse Gas Inventories Programme. Eggleston, H.S., Buendia, L., Miwa, K., Ngara, T., Tanabe, K. (Eds.). IGES, Japan. Climate Change 2007: The Physical Science Basis (Ends S Solomon, D Qin, M Manning, Z Chen, M Marquis, KB Averyt, M Tignor, HL Miller) pp 996 (Cambridge Press: Cambridge, UK and New York, NY, USA). 16

19 ISO ISO Environmental Management Life Cycle Assessment Principles and Framework, ISO ISO Environmental Management Life Cycle Assessment Requirements and guidelines, ISO Israel 2009 Kristensen 2011 Lundie 2009 Murray 2009 NGA 2010 PAS 2050 PE Australasia 2010 ISO DIS Draft International Standard. International Organisation for Standardization: Environmental Management Carbon Footprint Part 1: Quantification. Israel, G. D., (2009), Determining Sample Size, University of Florida, PD00600.pdf. Kristensen et al. Effect of production system and farming strategy on greenhouse gas emissions from commercial dairy farms in a life cycle approach. Livestock Science 140 (2011) Lundie S., Schulz M., Peters G., Nebel B., Ledgard S. Carbon footprint measurement methodology report. Fonterra Co-Operative Group Limited. Jan Murray, A. Theoretical calculations of energy usage and carbon emissions from food processing plants, p. 4/6/ National Greenhouse Accounts (NGA) Factors. (2010) Department of Climate Change and Energy Efficiency, Canberra, Australia; Publicly Available Specification, Specification for the assessment of life cycle greenhouse gas emissions of goods and services, (2008). BSI, British Standard Institute, London. PE Australasia et al. Goal and Scope Report: Carbon Footprint of the Australian Dairy Industry. Confidential Report for Dairy Australia. PE Australasia December PE: GaBi 5 Software-System and Databases for Life Cycle Engineering. Copyright TM. Stuttgart, Echterdingen Sevenster 2008 SoFi Thoma 2010 Thomassen 2008 USDA 2000 Sevenster, M. de Jong, F A sustainable dairy sector Global, regional and life cycle facts and figures on greenhouse-gas emissions. CE Delft. September Software-System for Corporate Sustainability. (2012) Copyright TM. PE-International, Stuttgart, Echterdingen. Thoma el al Greenhouse Gas Emissions of Fluid Milk in the U.S., University of Arkansas. Presented at LCA Food th International Conference in Bari, Italy. Thomassen, M. et al. (2008): Attributional and consequential LCA of milk production. The International Journal of Life Cycle Assessment 13(4): Fuel Required for field operations, Machine Management Series. Iowa State University, University Extension available at References/1Guides/Equip/ISUPm709FuelRequiredFieldOperations.pdf. van Buuren 2010 Pers comms. Neil van Buuren, Dairy Australia. 22/7/2010. Summary of the Final report: Farming Carbon Footprint of the Australian Dairy Industry 17

20 Dairy Australia Limited ACN Level 5, IBM Centre 60 City Road, Southbank VIC 3006 Australia T F E enquiries@dairyaustralia.com.au 18