Technical ReportPDF Available

Climate Footprints of Norwegian Dairy and Meat - a Synthesis

Authors:
  • CICERO Center for Climate and Environmental Research - Oslo

Abstract and Figures

This report reviews the current literature on meat, milk and dairy, with a special focus on Norway. To understand differences in reported emissions, the report explains the variation in methodological approaches such as division over co-products, functional unit selection, and system boundaries. Cattle meat, milk and dairy emissions are analyzed and compared with selected other foods that could act as a replacement, according to the various system boundaries used in the studies. Emissions from meat and dairy in Norway are compared with the Nordics and west-Europe, and other regions where relevant. Comparisons are also made between different production systems, including conventional and organic, intensive and extensive, and beef production from different types of cows. Finally, the report analyses the relative impacts of the different life cycle stages of meat and milk production and consumption. In a short section, it highlights some of the potentials for change of milk and meat impacts on the climate that emerged from the literature. Key findings summarize emissions for meat from dairy cows (around 19,5 kg CO2 equivalents per kg product), young bulls (around 19 kg CO2eq/kg), suckler cows (around 30 kg CO2eq/kg) and milk (around 1,2 CO2eq/kg). Norway’s emissions from combined meat-milk production are higher than in other Nordic and Western European countries, mainly because other countries have higher yields and lower methane emissions. Cattle meat and milk emit more than potential alternatives. Use of functional units and comparison between products depends on the stakeholders and context for comparison. In Norwegian meat and milk production, on-farm processes play by far the largest role, with around 78% of the emissions. Pre-farm stages contribute 22%. Most, around 38%, come from methane from ruminant digestion. Importantly, few if any studies present allocations over the full life cycle, which means that proportions for pre-, on—and post-farm emissions may change significantly when including all life cycle stages. Finally, the report finds no clear differences between conventional and organic meat and milk production in terms of climate impact, while intensive and extensive systems both have large mitigation potential.
Content may be subject to copyright.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy
and Meat – a Synthesis
A literature study of emissions of Norwegian dairy and meat products compared to other
relevant products and regions, commissioned by TINE AS
Bob van Oort, Robbie Andrew
July 2016
CICERO Senter for klimaforskning
P.B. 1129 Blindern, 0318 Oslo
Telefon: 22 85 87 50
Faks: 22 85 87 51
E-post: admin@cicero.uio.no
Nett: www.cicero.uio.no
CICERO Center for International Climate
and Environmental Research
P.O. Box 1129 Blindern
N-0318 Oslo, Norway
Phone: +47 22 85 87 50
Fax: +47 22 85 87 51
E-mail: admin@cicero.uio.no
Web: www.cicero.uio.no
Title: Climate Footprints of Norwegian Dairy and Meat a Synthesis
Authors: Bob van Oort and Robbie Andrew
CICERO Rapport 2016:06
Financed by: TINE and the Norwegian Research Council (SIS KliMAT)
Project (CICERO project number):: 30821 and 30790
Project manager: Bob van Oort
Quality manager: Asbjørn Aaheim and Elisabeth Lannoo
Keywords: Climate Footprints; Emissions; Meat; Milk; Dairy; Norway
Abstract: This report reviews the current literature on meat, milk and dairy, with a special focus on Norway. To
understand differences in reported emissions, the report explains the variation in methodological approaches such
as division over co-products, functional unit selection, and system boundaries. Cattle meat, milk and dairy emissions
are analyzed and compared with selected other foods that could act as a replacement, according to the various
system boundaries used in the studies. Emissions from meat and dairy in Norway are compared with the Nordics
and west-Europe, and other regions where relevant. Comparisons are also made between different production
systems, including conventional and organic, intensive and extensive, and beef production from different types of
cows. Finally, the report analyses the relative impacts of the different life cycle stages of meat and milk production
and consumption. In a short section, it highlights some of the potentials for change of milk and meat impacts on
the climate that emerged from the literature.
Key findings summarize emissions for meat from dairy cows (around 19,5 kg CO2 equivalents per kg product),
young bulls (around 19 kg CO2eq/kg), suckler cows (around 30 kg CO2eq/kg) and milk (around 1,2 CO2eq/kg).
Norway’s emissions from combined meat-milk production are higher than in other Nordic and Western European
countries, mainly because other countries have higher yields and lower methane emissions. Cattle meat and milk
emit more than potential alternatives. Use of functional units and comparison between products depends on the
stakeholders and context for comparison. In Norwegian meat and milk production, on-farm processes play by far
the largest role, with around 78% of the emissions. Pre-farm stages contribute 22%. Most, around 38%, come from
methane from ruminant digestion. Importantly, few if any studies present allocations over the full life cycle, which
means that proportions for pre-, onand post-farm emissions may change significantly when including all life cycle
stages. Finally, the report finds no clear differences between conventional and organic meat and milk production in
terms of climate impact, while intensive and extensive systems both have large mitigation potential.
Language: English
Rapporten kan bestilles fra:
CICERO Senter for klimaforskning
P.B. 1129 Blindern
0318 Oslo
Eller lastes ned fra:
http://www.cicero.uio.no
The report may be ordered from:
CICERO (Center for International Climate and
Environmental Research Oslo)
PO Box 1129 Blindern
0318 Oslo, NORWAY
Or be downloaded from:
http://www.cicero.uio.no
Contents
1 Introduction .......................................................................................................................................................... 1
1.1 REPORT AIM ................................................................................................................................................................... 1
1.2 BACKGROUND ............................................................................................................................................................... 2
1.2.1 Agricultural Emissions ................................................................................................................................................. 4
1.2.2 The Norwegian Context ................................................................................................................................................ 6
1.3 LIFE CYCLE ASSESSMENT ............................................................................................................................................ 7
1.4 CONSISTENCY AMONG ESTIMATES ........................................................................................................................... 8
1.4.1 Appropriate comparison: Functional units ..................................................................................................................... 9
1.4.2 Data Specificity ........................................................................................................................................................... 10
1.4.3 Co-products ................................................................................................................................................................. 11
1.4.4 System boundaries ....................................................................................................................................................... 12
2 Approach and Results........................................................................................................................................14
2.1 LITERATURE SEARCH, SYNTAX .................................................................................................................................. 14
2.2 GENERAL OVERVIEW ................................................................................................................................................. 15
2.2.1 Comparison of food items............................................................................................................................................. 17
2.2.2 Comparison of functional units .................................................................................................................................... 24
2.3 DAIRY ............................................................................................................................................................................ 26
2.4 BEEF............................................................................................................................................................................... 32
2.5 PRODUCTION METHODS ............................................................................................................................................ 36
2.5.1 Comparison of organic and conventional production ..................................................................................................... 36
2.5.2 Other production methods ............................................................................................................................................ 40
2.6 WHICH LIFE CYCLE STAGES OF MEAT AND DAIRY HAVE THE GREATEST IMPACT? ......................................... 42
2.7 POTENTIAL FOR CHANGE? ........................................................................................................................................ 49
3 Key points and final remarks............................................................................................................................53
4 Brief list of definitions .......................................................................................................................................57
5 Acknowledgements ............................................................................................................................................59
6 References ...........................................................................................................................................................60
6.1 CITATIONS IN THE REPORTS TEXT .......................................................................................................................... 60
6.2 SOURCES OF EMISSION DATA USED IN TABLES ...................................................................................................... 63
6.3 BIBLIOGRAPHY (NOT MENTIONED IN CITATION LIST) ......................................................................................... 68
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
1
1 Introduction
1.1 Report Aim
TINE commissioned this report to create a factual, objective basis for a better, unbiased and
critical understanding of the emissions (CO2 and equivalents of other greenhouse gases) of
Norwegian meat and dairy production.
The report provides a context for TINE’s “nutrition strategy towards 2018” which, in a separate
report, will evaluate the role of meat and dairy in a sustainable and climate friendly diet
(sustainable nutrition). The role of agriculture in terms of sustainability is increasingly relevant,
in terms of both climate change, population change, and public opinion. There are many
references to the climate impacts of agriculture, especially cattle, on climate.
A complete lifecycle analysis of Norwegian produced meat and dairy, covering the existing
variety of different production methods, (types of) energy use, and all inputs and outputs with
an effect on the climate, would give the best answers, but this was too ambitious for this report.
Thus, the report does not create new knowledge, but summarizes current and relevant
knowledge on emissions in the Norwegian dairy and meat production. There is an
understanding that different regions, production methods, and different inputs and outputs and
system boundaries related to emission numbers result in a range of different answers to the
amount of emissions related to meat and dairy production. Very few studies include a complete
life cycle of these products, and equally few studies compare production methods or products
using the same approach for Norway. Thus, we focus on findings for Norway, and compare
and complement these with similar studies in the Nordic countries (Sweden, Denmark, Finland),
and comparable countries in Europe (e.g. Netherlands, UK, Germany). Furthermore, we use
data from the rest of Europe or other at the global level to put Norwegian emissions in
perspective. We also analyze emissions in the different steps in the production-consumption
chain to assess which factors contribute to higher or lower emissions in the meat and dairy
industry. Finally, with an eye on the need to curb climate change and thus emissions, and the
potential role and consequences for the meat and dairy industry in this, we analyze emissions of
some alternative products that potentially play a role as meat or dairy substitutes.
Sustainable and climate friendly food production and consumption are also issues of increasing
focus and relevance in the scientific community. The request of TINE for this report coincides
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
2
with an increased focus at CICERO on the needs and options to decrease emissions on the
production and consumption side of the food value chain.
1.2 Background
Through natural processes, livestock generate emissions of gases that have a warming effect on
the world’s climate. While those emissions might still be regarded as ‘natural’, the enormous
scale of the industry across the globe means that those emissions contribute significantly to the
warming already seen in global temperatures. Emissions from the livestock industry as a whole
have been estimated to contribute almost 15% of total anthropogenic emissions of greenhouse
gases (GHGs; Gerber et al., 2013). Natural processes are not the only sources of greenhouse
gas emissions from agriculture, with significant use of fossil fuel, both as an energy source and
as an ingredient in fertiliser manufacture, as well as carbon emissions from land-use change,
both deforestation and draining of wetlands.
A large proportion of emissions from the livestock industry come from ruminant animals (cattle,
sheep, and others) and their management, with beef and cow milk production contributing
about 60% combined to the industry’s total global emissions (Gerber et al., 2013). In 2013, there
were about 3.7 billion ruminant livestock globally (FAOSTAT, 2016).
In Norway, the agricultural sector is responsible for about 8% of total emissions, some 4.4
MtCO2e per year. These have gone down slightly in recent years, from 4.9 Mt in 1990 to 4.4 Mt
in 2015 (preliminary estimate; Figure 1) largely as a result of reduced numbers of cattle and an
increase in the use of concentrated feed in place of fodder (Arbeidsgruppe til Landbruk og
Klima, 2016).
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
3
Figure 1: Trends of greenhouse gas emissions from the agricultural sector in Norway, 1990-2015.
These exclude emissions from production on drained wetlands, on-farm energy use, and all off-farm
emissions (Source: SSB).
However, these figures represent the agricultural ‘sector’ as defined in international accounting
terms, and thereby exclude important emissions such as those from energy use on farms (e.g.,
tractor fuel) and, most significantly, agricultural production on drained wetlands. When
wetlands are drained, the rich carbon content of their soils gradually combines with oxygen
from the air to form carbon dioxide, which escapes to the atmosphere. Despite these drained
wetlands amounting to only about 6% of Norway’s agricultural area, their slow oxidation adds
about 1.8 Mt CO2 of annual emissions. When these additional emissions are included, the total
from agriculture increases to about 6.3 Mt, or 12% of Norway’s total greenhouse gas emissions
(Arbeidsgruppe til Landbruk og Klima, 2016). The livestock sector in Norway contributes about
90% of this total (Grønlund & Harstad, 2014), while globally the proportion is lower because
of emissions from other forms of agriculture, such as rice cultivation. Table 1 presents various
contributions to agriculture’s emissions in Norway.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
4
Table 1: Sources of emissions from land-use in Norway (Source: Grønlund & Harstad,
2014)
1000 tonnes
CO2e
% total land
use
1892
30%
924
15%
604
9%
449
7%
310
5%
1785
28%
149
2%
228
4%
6340
100%
Furthermore, emissions reported officially by SSB and to the UNFCCC include only direct
emissions, i.e. those that occurred in the sector in Norway. They therefore exclude emissions
that occur upstream in the supply chain and those associated with imported goods and services.
These are sometimes called indirect emissions, resulting as they do indirectly from the activities
of the agricultural sector. Similarly direct emissions also exclude those occurring in necessary
downstream activities such as those in the food-processing sector, and in food distribution and
retail.
1.2.1 Agricultural Emissions
Emissions of greenhouse gases associated with agricultural production include both on-farm
and off-farm emissions. On-farm emissions are those that occur in the agricultural context, such
as carbon dioxide (CO2) emissions from use of tractors and other machinery, methane (CH4)
emissions from ruminant digestion (‘enteric fermentation’) and manure, and nitrous oxide
(N2O) emissions from fertiliser use and urine. Off-farm emissions are mostly CO2 and occur in
other parts of the supply chain, such as in electricity generation, fertiliser production,
transportation, refrigeration, and food processing. While CO2 is the most important greenhouse
gas globally, CH4 and N2O are significantly more important in the agricultural context.
All developed nations report national emissions inventories annually to the United Nations
Framework Convention on Climate Change (UNFCCC). The format and structure of these
inventories is carefully designed by the Intergovernmental Panel on Climate Change (IPCC)
with consistent methodologies between all countries. Emissions fall into five ‘sectors’: Energy;
Industrial Processes and Product Use; Agriculture; Land Use, Land Use Change and Forestry
(LULUCF); and Waste. However, many readers do not understand is that this Agriculture sector
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
5
includes only those types of emissions that do not occur in other sectors, such as the methane
from enteric fermentation, but exclude all other emissions, such as energy use on the farm and
in fertiliser production, which are included in the Energy sector. When discussing total
emissions in Norway’s agriculture sector, it is therefore inappropriate to report only the total of
the IPCC Agriculture category given in the national emissions inventory. The IPCC accounting
framework is set up in this way to prevent any double counting of emissions.
While the national emissions inventory covers all greenhouse gas emissions in Norway, it
intentionally does not include emissions overseas. In contrast, a carbon footprint necessarily
includes emissions overseas, if they are generated in the product’s supply chain. In the present
context, emissions associated with soy bean production in Brazil and their transportation to
Norway should be included in the calculation of a carbon footprint for Norwegian meat and
dairy products.
Some experts argue that grazing leads to increased carbon content of soils, i.e. carbon that is
sequestered from the atmosphere via grass, leading some to suggest that increased grazing will
help in the mitigation effort against climate change. Certainly some soils under grazing do gain
carbon, but this is highly dependent on the type of soil and how long it has been grazed for.
Organic soils, as discussed above, lose large amounts of carbon following draining, while
mineral soils can gain carbon. In Norway, mineral soils are estimated to be storing additional
carbon every year, and these additions are included in SSB’s estimates, which are submitted as
the National Inventory Report to the UNFCCC. However, not all footprint analyses include
these soil carbon fluxes, which is an important aspect to be aware of during interpretation.
To add together the emissions of different greenhouse gases it is necessary to use what is called
a metric, and the most frequently used of these is Global Warming Potential (GWP), which
allows for conversion of the values of each gas emission to the equivalent warming effect of
CO2. The current standard values, as used in national reporting to the UNFCCC, are 25 for
CH4 and 298 for N2O. That is, emission of 1 kg of CH4 has the same warming potential as 25
kg of CO2, and 1 kg of N2O has the warming potential of 298 kg of CO2. Therefore emissions
of CH4 and N2O are multiplied by these factors first before all three gasses are added together
and presented in terms of the equivalent amount of CO2 that would result in the same warming,
denoted kgCO2e.
A complicating factor is that these equivalency factors are based on integrating the warming
effect over a 100-year period, and, arguably, with the 2 °C threshold potentially only 20-30 years
away (Friedlingstein et al., 2014), shorter integration periods could be more appropriate in a
policy context to reflect near-term warming. Shorter periods yield a significantly higher global
warming potential for methane, making it as much as three times more important (i.e. the factor
of 25 increases to as much as 85). Probably because of resistance to national emission accounts
suddenly changing quite substantially, this issue remains largely unaddressed, and the arbitrarily
chosen 100-year timeframe is almost always used, as a matter of convention.
Because of significant variations in production methods, climate, and other factors (e.g. Opio
et al. 2013), emissions from livestock per unit of final product vary significantly around the
world (Figure 2). In Western Europe, emissions are very low by world standards, but there are
also variations within this region, and it is important to have estimates specific to Norway.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
6
Figure 2: Greenhouse gas emissions per unit of carcass weight by category and world
region (Source: Gerber et al., 2013)
1.2.2 The Norwegian Context
It is useful to compare the environmental consequences of Norwegian agricultural production
to other countries for several reasons. The most important is perhaps to determine where
Norwegian production lies relative to ‘best practice’, and to gain understanding of what scope
there is to change Norwegian practices to reduce environmental impacts. This understanding
might lead to implemented changes, or might be used to explain to Norwegian consumers and
regulators why Norwegian production results in different environmental outcomes. A
secondary reason would be a market assessment, looking to understand the positions of
potential international competitors as a way of dealing with risks of changes in the trading
environment. Either way, comparison with other countries can lead to important lessons.
Norway’s agricultural production context is significantly different to that of many other
countries. With a short and cold growing season, prevalence of thin soils, steep and isolated
farm plots, and small proportion of arable land, Norway is relatively poorly suited to agricultural
production. Of the approximately 1 million hectares of agricultural land, 45% is suitable only
for grass production (Blandford et al. 2015). Furthermore, significant use is made of non-
agricultural land (utmarka) for grazing. This context, along with the high domestic cost
structure, means that sustaining agricultural production requires significant financial support
but also that Norway has specific production methods, input requirements, opportunities for
economies of scale, etc. Understanding this particular Norwegian context is necessary when
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
7
comparing the environmental consequences of Norwegian agricultural production to those of
other countries.
Therefore, one should be careful when choosing countries to compare with Norway.
Comparing with New Zealand production, for example, would not be appropriate, despite
claims made in Norwegian media in 2014 that conditions in the two countries were identical,
with the only difference being the complete lack of subsidies in New Zealand (Magnus, 2014).
New Zealand lies much closer to the equator than Norway, receives considerably more annual
sunshine, is without snow over much of the country in winter, and has very large contiguous
areas suitable for agriculture with good soils. The contrast in farming conditions could hardly
be starker. While we could learn some lessons from New Zealand’s methods of agricultural
production, to a large degree that country’s production is a poor point of comparison for
Norwegian production.
In contrast, countries such as Sweden, Denmark, and Switzerland have much more similar
conditions and agricultural production models to Norway and therefore serve as useful points
of comparison. We will focus on these countries in this report.
Within the Norwegian context, the issue of imported concentrated feed, such as soya beans
from Brazil, has loomed large in the Norwegian media. While soy meal made up only about
10% of concentrated feed in 2015, overall imports amounted to 45% (Landbruksdirektoratet,
2016). About 60% of a Norwegian cow’s diet is roughage (grazed or baled), so the amount of
soy in the overall diet is perhaps 3% by weight. However, soy contributed about 35% of the
protein to cows’ diets in 2015 (Volden, 2016).
As noted, emissions from agriculture in Norway have declined slightly since 1990, and this is
primarily a result of developments in the milk industry. Increased use of concentrated feed and
breeding have both led to increased milk yield per cow, resulting in turn in a decline in cow
numbers and a consequent decline in emissions. A further consequence of this is the
development of the suckling cow industry to make up for reduced meat production from the
milk industry, and this development means that beef meat in Norway is produced from these
two industries.
1.3 Life Cycle Assessment
While there are several different types of ‘carbon footprint’ found in the literature, the most
suitable and widely used method available for estimating the carbon footprint of products is
Life Cycle Assessment (LCA, sometimes Life Cycle Analysis). The core purpose of LCA is to
estimate environmental impacts associated with all stages of the production chain, use, and
disposal of a product; carbon emissions are one such impact. In the agricultural context, this
means not only estimation of impacts from on-farm activities, but also from all activities in the
supply chain. The use (i.e. consumption) and disposal phases are not always included in
assessments.
Industry has run simple in-house LCAs since the 1960s, but it was not until 1990 that the term
was coined (PE International 2013). Because results of LCAs could vary very widely based on
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
8
assumptions made and methodologies used, international standards were established, beginning
in 1993 and eventually becoming ISO 14040 and 14044, in 2006. The existence of these
standards, and the requirement that they be followed if an LCA is to be published in academic
journals, has helped to ensure greater consistency and transparency in the LCA field.
Nevertheless, the range of permissible assumptions and methodologies mean that LCA results
require careful interpretation, as we will discuss in the next section.
The LCA standards lay out a four-step procedure, although only the first three are strictly
required.
The first, Goal and Scope Definition, makes a clear statement of the purpose of the
assessment, defines the ‘functional unit’ (described below in section 1.4.1), and sets out
the system boundary: what will be included in and excluded from the assessment (see
section 1.4.4).
The second stage, Life Cycle Inventory, involves the collation of all relevant data within
the specified scope: all resources consumed and all flows of waste. All quantities are
scaled to the specified functional unit: for example how much of each emission in the
supply chain results from production of one kilogram of cheese. Software and existing
databases are very often used to help in this process.
Then follows Impact Assessment, in which inventory components are translated to
(potential) environmental impacts (e.g., via the global warming potential, discussed
above) and potentially all different impacts are combined into an overall score.
Finally an Interpretation of the results leads to discussion and conclusion, particularly
relating to the consequences, the sensitivity of the analysis to particular assumptions,
and any limitations of the study.
1.4 Consistency Among Estimates
In 2014, researchers at UiB and Bioforsk wrote an opinion piece in the newspaper Dagbladet
suggesting that Norway’s emissions from agriculture could be significantly reduced while
maintaining domestic food supply, largely by reducing consumption of red meat (Gaasland et
al. 2014). While their analysis was based on a detailed and complex model, for the purposes of
the article they presented just a few numbers to support their case, including the proportion of
Norway’s emissions coming from land use and the emissions per kilogram of meat from suckler
cows (ammekyr) and sheep. Three weeks later came a response from researchers at NMBU
challenging the figures used and conclusions given by UiB and Bioforsk, pointing to ‘official
figures’ that contradicted what had been presented (Aass & Vangen 2014). Furthermore, the
NMBU researchers argued that various factors were overlooked and, in particular, that
Norwegian cows are used for both milk and meat, so the emissions should be divided between
these two products.
At the global level, there remains widespread confusion in the media and society as to whether
the emissions from livestock agriculture amount to 15% of the global total (Gerber et al. 2013)
or 50% (Goodland & Anhang 2009).
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
9
When researchers cannot agree on the appropriate figures to use, it becomes impossible for the
public, business, or policy-makers to make informed decisions based on these figures, resulting
in both confusion and a danger that the most ‘suitable’ numbers are used, those that best fit the
goals. There are a number of reasons why data on emissions differ between different sources,
and in this section we will describe the most important of these.
According to González et al. (2011), the carbon footprint of cucumbers produced in Sweden
can vary between 0.08 and 2.6 kgCO2e/kg product, depending on whether they’re grown
outdoors in summer (low) or in fuel-oil heated glasshouses in the off season (high). Sometimes,
as in the case of Swedish cucumbers, the production method is most likely the reason for the
differences in carbon footprint estimates, and this is indeed the information that we seek from
LCAs. However, there are several other reasons why estimates can vary; we will discuss these
below.
1.4.1 Appropriate comparison: Functional units
It is conventional wisdom that one should not “compare apples with oranges”, but if the
question is how to best provide fruit for consumption while achieving various relevant policy
goals, then exactly such a comparison is required. The question then becomes whether and how
such comparisons should be performed. While it might be obvious that we should not compare
meat with shoes when considering options for nutrition, it might not be so obvious that it is
inappropriate to compare bacon with lettuce, as evidenced by widespread media attention in
late 2015 on that very subject (e.g. Withnall, 2015). Bacon and lettuce serve entirely different
purposes in the diet, and the role of lettuce is certainly not to provide calories, so any
comparison on a calorific basis is misleading at best.
One of the primary goals of LCA is to allow comparability between products that serve the
same purpose so as to identify the environmental consequences of the choice. Examples include
comparison of paint with wallpaper, re-usable nappies with disposable nappies, nuclear power
with bioenergy, and brooms with leaf-blowers.
In the case of paint, a researcher might specify the inputs required for, and environmental
consequences resulting from, production of one litre of paint. However, such a ‘functional unit’
would make comparison impossible with different paints that require different surface
preparation, or different numbers of applications, let alone comparison with wallpaper. Rather,
the researcher might choose to specify inputs and outputs for one square metre of internal wall
covered for 10 years, with an implication that all required maintenance of that wall covering is
included.
This process of defining the functional unit is critical in LCA, and different choices can lead to
significant differences in the analytical results. It is therefore necessary to identify which
properties of the products to compare: Does a drink need to be white? Does it need to be
suitable for use with breakfast cereal? Does it need to contain high levels of calcium? Does it
need to mix well with coffee? In contrast, when comparing two effectively identical products
with different production methods (e.g. conventional and organic milk, or Norwegian and Swiss
milk) then this identification of properties is less important.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
10
Because of the sensitivity of LCA results to the choice of functional unit, the case for making
this choice must be transparently made, and the LCA community has established clear
procedures for doing this (e.g. Weidema et al. 2004). Ideally, the same researchers should
perform comparative LCAs of two products at the same time, so that the functional unit, scope,
and all assumptions are the same. However, with the considerable effort required to undertake
an LCA and the enormous number of products, making comparisons based on existing,
disparate literature is often required. One should be careful to identify how comparable two
separate LCAs are when presenting their results.
For example, when comparing meat to alternative products, it seems reasonable to use a protein
basis. However, protein is not the sole reason that consumers purchase meat; one should also
consider the nutritional completeness of proteins, fat content, taste, ease of preparation,
versatility, among many other potential characteristics. While for some comparing meat with
powdered protein might seem a bridge too far, it is not necessarily clear when comparisons are
in fact reasonable.
Comparing fresh dairy milk with alternatives such as soymilk, oat milk, and rice milk simply per
litre of product ignores differences in the nutritional content, cooking properties, or cultural
reasons for consuming these products, along with (macro-) economic consequences such as the
effects on national trade balances. The fat- and protein-corrected milk (FPCM) measure partly
addresses the issue of differing nutritional contents, essentially elevating fat and protein content
as the most important factors. However, because products are generally inherently different to
some degree, it generally is not possible to choose a functional unit that makes them perfectly
comparable, and simplifications are necessary. One could compare milk to two separate
products that each provide one of milk’s services: healthy beverage and baking ingredient.
1.4.2 Data Specificity
In collating data for a life cycle assessment, averages are not only unavoidable but also entirely
necessary. Data from a specific Norwegian farm on a specific day are very unlikely to be
representative of the average Norwegian farm on an average day. Moreover, conditions change
through the course of the year, for example with different feed requirements and availability,
and from year to year with climatic, management and market variations. The international
origins of imported feed can change from year to year, the proportions of different cattle breeds
used in the industry change, the ratio of suckling cows to milk cows, the yield, the number and
size of farms, and so on. Because of such variation, carbon footprints of products must be
calculated and presented as averages. Some LCAs use data averaged over three years or more
to reduce their sensitivity to short-term variations.
Because of the effort required to collect data, and consequent cost, sometimes data from
previous studies are used. Data or information from one domain (e.g. Danish dairy farms) might
be transferred are transferred for use in another domain (e.g. Norwegian dairy farms). While
such transferring saves expense and time, one should carefully documented and identify the
similarities and differences between the two domains to prevent biasing the assessment.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
11
The variation in conditions of the supply chain also necessitates an assessment of both
uncertainty and sensitivity. Uncertainty means ‘how sure are we of the result?’ while sensitivity
means ‘how much would the result change if a particular data point were to change?’ Sensitivity
is very important for two reasons. Firstly, it gives some indication of how the result might vary
in future. Secondly, it points to ‘hot spots’ in the supply chain, changes to which would lead to
significant changes to the footprint. For example, it might turn out that the amount of
supplementary feed fed to livestock has a large effect on the carbon footprint, or perhaps the
age at slaughter of milking cows. Knowing to which parts of the supply chain the result is
sensitive is therefore very valuable.
1.4.3 Co-products
When there are two or more products of a production process, the environmental impacts of
the process must be divided between those products, e.g. milk and meat from the dairy industry.
Yet how much of the carbon emissions occurring in the supply chain to the farm gate come
from the meat production and how much from milk?
There are three standard methods to resolve this:
Physical Allocation makes assumptions about how the inputs used in the farm
physically end up in the milk and meat. An example might be to use the nitrogen
content of milk and meat to divide the nitrogen fertiliser impacts.
Economic Allocation allocates all environmental consequences based on the economic
value of the products: if the process produces 2 kroner of milk and 1 krone of meat,
then two-thirds of the impacts are allocated to the milk and one-third to the meat. The
argument for economic allocation is that economic demand drives production.
System Expansion (also Substitution and Avoided Burden) isolates one of the co-
products by subtracting the environmental consequences of the most likely alternative
means of producing the other co-products (Weidema, 2000).
We note in passing that there are two further approaches to dealing with co-products. One is
to simply ignore one of the co-products and allocate all impacts only to the other. The second
is to leave the two co-products combined and report the environmental impacts associated with
two products at the same time (e.g. the joint production of milk and meat from the dairy
industry, Blandford et al., 2015). Neither of these is particularly useful.
Physical Allocation and Economic Allocation are termed attributional methods, describing the
present state but giving a poor indication of the consequences of a change. According to Plevin
et al. (2014), attributional approaches give misleading advice to decision-makers. The reason for
this is that these studies look at averages, not at margins, i.e. that any change in the scale of
production merely results in a linear scaling of impacts. In addition, both methods are
normative: arbitrarily supporting a particular worldview. The System Expansion method, on the
other hand, is consequential: it indicates what would happen when changing from one
production method to another. While this gives more appropriate guidance to decision-makers,
it comes at the expense of relying on specific scenarios: the results are valid only for the change
specified in the scenario precisely because marginal factors change with quantity produced,
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
12
although sensitivity analyses can go some way to mitigating this problem. Unfortunately, while
the information provided is more accurate, the System Expansion method is considerably more
complex. Almost all studies in the literature use attributional methods. While such studies can
be highly informative, they are not ideal for describing how environmental impacts would
change in switching from one production (method) to another.
1.4.4 System boundaries
While the ideal goal of an LCA study is to assess the entire lifecycle of a product, from ‘cradle
to grave’, this is not always feasible. Figure 3 depicts the common stages of the life cycle of a
product. Sensitivity to post-retail stages can be very large and outside of the control of the
producer. For example, whether a consumer drives to the supermarket in a 20-year-old car or
takes public transportation can have a huge effect on the total footprint of their food
consumption. Disposal stages can also be beyond the control of the producer, with significantly
different consequences dependent on the decisions of consumers to compost food waste, send
it to incineration, or to biogas production, and similarly on the options provided by
municipalities for waste collection and disposal. Indeed, how much food the consumer wastes
can increase their dietary footprint by more than 50%. The impact of food wastage also depends
on the type of food and its GHG impact up to consumption (or waste). For cereal for example,
the driver seems to be mostly the wastage volume, whereas for meat, the driver is the carbon
intensity of the commodity. The FAO (2013) reports that at the global level, products of animal
origin account altogether for about 33 percent of total carbon footprint, whereas their
contribution to food wastage volumes is only 15 percent.
However, important lessons can be learned from evaluating the entire life cycle of the product,
and producers do have some influence over the post-retail stages. In a seminal study, Procter &
Gamble analysed the entire life cycle of laundry detergent, finding that more than 80% of energy
use occurred in the consumer stage, mostly in heating water (Saouter & van Hoof, 2002). This
led to the development of cold-water detergents, with potentially significant consequences for
global energy consumption. Manufacturers also have some control over impacts of the disposal
stage of their products by designing with repair and recycling in mind. Consumer-stage food
waste can be greatly reduced by wrapping products in plastic film, with the cucumber being a
clear example, lasting up to three times as long when wrapped in plastic, greatly reducing waste,
and directly translating into reduced production and environmental impacts (Aldrige & Miller,
2012). Therefore, it can be beneficial to include post-retail stages in life-cycle assessments.
When different parts of the supply chain are included in an LCA, different terms are used to
describe the assessment (see also figure 3):
‘Cradle to grave’ is used when the full life-cycle is included in the system boundary,
‘cradle to gate’ describes the supply chain only up to production at the factory or farm,
‘cradle to plate’ or ‘field to fork’ (‘jord til bord’) describes the process specifically for
food products to the point of actual consumption (and therefore should include
purchasing, transportation home, storage in the home, and preparation).
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
13
In this report, we refer to several studies that used the ‘whole farm’ system boundary. These are
LCAs essentially the same ‘cradle to farm gate(Crosson et al., 2010), as they generally include
GHG emissions from all processes up until the point the primary product is sold from the farm.
Emissions from production of external farm inputs (e.g. concentrate feeds and fertilisers) are
also typically included in the analysis. However, to avoid misinterpretations, we have as much
as possible kept the same wording of the original papers and use ‘whole farm’ alongside ‘cradle
to farm gate’ and other system boundaries.
Figure 3: Graphical representation indicating different system boundaries and which parts
of the supply chain they include.
System boundaries extend not only along the supply chain, but also describe the depth of
analysis at each stage. An LCA is produced by creating an inventory of each input in the supply
chain and assessing their cumulative impacts. However, supply chains are always complex, with
inputs such as use of services often assumed to introduce negligible environmental impact
compared to physical processes. In the early 2000s, it became clear that the assumption that
many contributions to the life-cycle impacts were small wrong, with up to 50% of life-cycle
impacts being ‘truncated’ in this way (Lenzen, 2001). As a result, LCAs now typically combine
supply-chain-specific inventory analysis and databases that include the life-cycle impacts of
generic (i.e. averaged) services and other inputs that were previously considered negligible.
Other life-cycle impacts that may or may not be included in an LCA system boundary are: land-
use change emissions, soil carbon fluxes, consequential effects for food production elsewhere
(particularly important for bioenergy LCAs), pesticide manufacture and use, and more.
Raw
materials Transport Farm
production Processing Distribution Storage
& Retail Transport Storage &
Consumption Disposal
Cradle Farm Gate Shelf Plate Grave
‘Whole farm’ Retail Gate
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
14
2 Approach and Results
2.1 Literature search, syntax
To cover the available literature, we followed a number of different approaches. Firstly, we
performed a systematic search using the following (table 2) search terms in the ORIA
(www.oria.no) and Google Scholar (www.scholar.google.com) databases.
Table 2: Search syntax used in the database searches
Emissions AND
Products AND
Production method
AND
Location
Emission* OR
footprint OR LCA
Agriculture OR food
OR dairy OR milk
OR beef
Production OR
ecologic*
Norway OR Nordic OR
Scandinavi* OR Sweden OR
Switzerland
Utslipp OR *avtrykk
OR livssyklus*
Jordbruk OR mat
OR meieri OR melk
OR kjøtt
drifts* OR
økologisk*
Norge OR Nordisk OR
Skandinavi* OR Sverige OR
Sveits
As an example of this approach, Google Scholar initial results (80.400 hits) were further limited
by using a cut-off date from 2000 to 2016 (20.500 hits). Narrowing the syntax to just including
Norway and making LCA a necessary inclusion (Emission* OR footprint AND LCA AND
food OR dairy OR milk OR beef AND Norway) within the 2000-2016 range, the number of
hits were further reduced to 5.390. Narrowing the search even further to articles published
between 2000-2016 containing all of the words “Emission * AND LCA AND food AND
Norway”, the exact phrase “lifecycle analysis”, and at least one of the words “dairy milk beef
production ecologic” anywhere in the article yielded 138 results. We scanned these results for
relevance and included them in the attached bibliography.
While systematic, the search for e.g. (Emission* OR footprint OR LCA) AND (Agriculture OR
food OR dairy OR milk OR beef) AND (Production OR ecologic*) AND (Norway OR Nordic
OR Scandinavi* OR Sweden OR Switzerland) in oria.no gave 15 hits, while the Norwegian
search gave no hits. This indicates that the available literature is limited, or that the key words
used are not delivering the desired results. To account for this potential gap, we included other
approaches to cover the available literature and sources to (Norwegian meat- and dairy)
emissions data included trawling through the reference lists of available and newly identified
literature, and communications and literature and other data exchange with TINE and experts
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
15
at e.g. the Norwegian University of Lifesciences (NMBU). A final list of relevant material is
included in the bibliography.
2.2 General overview
The sections in this general overview will present some comparisons of emissions of products,
distinguishing between system boundaries, countries and functional units. The details of
Norwegian emissions for each food category (meat, dairy) or production systems will be
highlighted and analyzed more in depth in the consecutive sections (2.3 and onward).
The total number of references including emission numbers for products in Norway is relatively
low, with 21 references, some of which are indirect references (referred to in another report or
article). As source of our references, we only use research articles or reports that are considered
to present objective data. Thus, any reports from sources that could have an interest in
representing the data subjectively are omitted. Likewise, our search and sources does not include
newspaper articles and websites and similar, with the exception of illustrating a point or
discussion in the media.
Otherwise, the search resulted in a bibliography of 168 articles or reports which were considered
relevant to the topic, 118 of which were used to extract emission numbers for different products
and countries, and 21 of these included emission numbers to Norway (for various products).
Table 3 shows how many emission numbers related to Norwegian food items (covering
different products, often using both average, upper and lower ranges for the same products)
from each study. The number of emission data for different Norwegian products added up to
135. By far the most studies relate to meat (52), while 24 studies concerned dairy products. Fish
and other food/drinks were covered by respectively 28 and 27 studies, and eggs by 4 studies.
Some important and recent reference works on Norwegian or Nordic emissions in the
agricultural sector were also consulted. These include e.g. Arbeidsgruppe til Landbruk og Klima
(2016), Andersen Nesse (2015), or Landbruks- og matdepartementet (2016). These and many
other works are extensively used in the discussion, but as these are reviews of research and
emission data already presented in other reports as is this report these reference works are
not listed in table 3. A typical example of this is the often-quoted emission data from Bonesmo
et al. (2013) who used the HOLOS model adapted for Norwegian dairy and beef production
system.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
16
Table 3: Studies including emissions for Norway, with counts for each study indicating the
number of emission values used from that reference and which system boundary the study
used.
Reference
cradle to
retail gate
cradle to
farm gate
whole farm
model
(blank)
Blonk et al. (2009)
1
Bonesmo et al. (2013)
12
Ellingsen et al. (2009)
1
FHL (2009)
2
Findus (2008)
3
Grønlund (2015)
5
Grønlund and Harstad (2014)
4
Grønlund and Mittenzwei (2016)
5
Hille et al. (2012)
64
Leip et al. (2010)
1
Mittenzwei (2015)
6
Nymoen and Hille (2010)
3
Pelletier et al. (2009)
1
Refsgaard et al. (2011)
7
Roer et al. (2013)
6
Silvenius and Grönroos (2003)
1
Storlien and Harstad (2015)
2
Svanes et al. (2011)
1
Ziegler and Valentinsson (2008)
1
Ziegler et al. (2013)
2
Åby et al. (2015)
7
Most studies and results presented could not be compared directly. There are differences in
methodology, as allocation between meat and milk and sometimes the system boundaries or the
factors they include are different. The scale of measurement may differ, with some studies based
on one or a few farms, others on farm modelling and yet others on national averages. These
latter have a tendency to show higher results, partly since more flows are covered than in the
other two types but mainly due to changed weighting factors for methane and nitrous oxide
introduced in 2007, which make results from older studies slightly lower than results from newer
studies (Sonesson et al. 2010).
Notarnicola et al. (2013) mention that the most commonly considered system boundary is the
cradle to farm-gate because of the lack of sufficiently detailed information in the cradle to retail
or consumer supply chains. Those studies including post-farm processes usually simplify the
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
17
input/output flows related to the agricultural phase. In addition, if a comparative LCA is
undertaken, and it is known that a particular part of the system is identical between the two or
more processes being compared, sometimes that part is omitted entirely. For example, in
comparing organic and conventional chicken production, an assumption might readily be made
that all post-retail phases of the system are identical and therefore do not need to be enumerated.
Thus, final product emission numbers may differ and be incomplete for many reasons.
2.2.1 Comparison of food items
As discussed in section 1.4, “emission values” depend on many factors, including the system
boundaries used, if land use and land-use change or waste are included, how the emissions are
distributed over the different co-products of an animal, which type of production system was
used, and the unit in which the emissions are expressed. Variation in these and more factors
makes inter-comparison of emission data at this level near impossible. Of all data collected,
“cradle-to-retail-gate” and “cradle-to-farm-gate” were the most used system boundaries (see
table 4), but even between these, methods and inclusion or exclusion of certain factors (such as
waste or land use/change) could differ and hence the comparability between emissions
numbers.
Nevertheless, the results give us a general idea of how emissions of different food items
compare based on general knowledge of emissions in land-use, of differences between
ruminants and mono-gastric animal, differences in transportation distance, and in waste. The
following sections will analyse these differences in more detail, and highlight some data and
studies with multiple comparisons with the same methodology.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
18
Table 4: Count of the type of system boundary used or indicated in the collected studies
from Norway to global level. The overview is not comprehensive, as system boundaries
are not always indicated or registered for each study in the database
Norway
Nordic
Europe
Global
Total
cradle to grave
36
1
37
whole LCA (excluding
waste)
16
4
20
cradle to retail gate
66
15
147
275
503
cradle to farm gate
23
19
107
49
198
whole farm model
41
41
Near all studies find that there is a large difference in carbon footprint between beef on the one
hand, and pork and chicken meat on the other, regardless of where in the world production
takes place (Norden, 2014). Table 5 shows an approximation of how different food items relate
to each other, comparing emissions (per weight) of the collected data on different foods across
a selection of system boundaries. Indeed, the table shows great differences in emissions between
the main product groups, i.e. meat, dairy, eggs, fish, other foods and vegetarian. Also within
each category (e.g. meat) there can be large differences, especially for beef, various cow meat
1
and mutton (sheep - and goat), which have much higher emissions per kg product than pork or
chicken. Lamb and sheep meat emit slightly more than beef, largely because beef’s emissions
per kilogram are reduced with some emissions allocated to milk. The clear division to make here
is that ruminant livestock produce substantially higher emissions than other livestock.
Refsgaard et al. (2011) argues that the environmental impact from animal and vegetable
products often differs by a factor of 10. Our results also show large variations in impact between
animal and vegetable products. The differences hinge on whether we compare vegetable
products with dairy (milk has about 2,5-3,5 times higher emissions than wheat) or meat (meat
from dairy cows has about 21-29 times higher emissions than wheat), and which system
boundary is used. The factor is again different when comparing nutrient value instead of weight
(table 8), but it should be born in mind that such a comparison may be meaningless.
One notable study (or rather: news coverage of a study) underlining the point of meaningless
comparisons denied that vegetables have lower emission than meat: The “lettuce versus bacon”
news story (e.g., Nosowitz, 2015) seems to make a baffling and contradictory claim: It is possible
that adjusting our diets from meat-heavy to produce-heavy could actually result in an increase
in greenhouse gas emissions. However, the article was based on a study that calculated that in
an unlikely, extreme modelling situation, one diet could be devised where lettuce could be worse
1
See the list of definitions at the end of this report.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
19
than pork meat. However, there are many critics to both the study and the presentation of it in
the media. A main take-home message is that it is an invalid and extremely unlikely comparison,
since we are never going to scale lettuce consumption up to the point where we obtain all our
(replacement for meat) calories from it.
Table 5: Emissions (average of kg CO2eq/kg product) for selected food products collected
in this report, including studies from many countries. The table differentiates between
different system boundaries used in the studies, and averages emissions per food
category. Meat is generally calculated per carcass weight, and milk as fat and protein
extracted weight
Category/product
cradle to
grave
whole
LCA (ex.
waste)
cradle to
retail
gate
cradle to
farm gate
whole
farm
model
unknown
Meat
5,63
13,71
15,79
20,51
14,55
11,79
Beef
28,35
39,03
29,60
14,24
24,69
Dairy cows
18,40
15,33
21,40
18,00
Suckler cow
29,67
28,15
28,55
Veal/Young bulls
19,48
16,83
Sheep, Lam, Goat
22,12
41,57
27,64
Pork
3,83
8,39
5,51
5,36
2,58
4,99
Chicken
6,23
4,41
4,13
3,46
1,32
3,07
Dairy
4,50
3,71
5,38
1,29
0,97
1,93
Milk, cattle
1,41
1,14
3,21
1,23
0,97
1,09
Milk, small
ruminants
6,39
3,00
Yoghurt
1,24
Cream
5,22
Ice cream
2,60
Butter
9,50
20,32
Margarine
1,50
Cheese
6,80
8,93
9,48
Eggs
4,44
2,10
3,39
3,40
0,93
3,00
Fish
2,97
1,76
6,29
3,00
Cod
4,47
2,70
5,67
4,00
Herring
1,47
0,89
1,10
1,40
Mackerel
0,95
3,18
Pangasius
3,00
Salmon
3,25
4,22
3,20
Shrimp
22,90
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
20
Category/product
cradle to
grave
whole
LCA (ex.
waste)
cradle to
retail
gate
cradle to
farm gate
whole
farm
model
unknown
Other food/drink
0,82
2,08
0,71
Wheat
0,87
0,53
0,67
Potatoes
0,43
0,20
Pulses
1,20
Rice
4,00
4,00
Cabbage and roots
0,31
Tomatoes
2,04
Apples
0,30
Strawberries
0,26
Vegetarian
1,00
0,40
2,48
1,58
Vegetarian burger
2,60
7,30
Tofu
2,00
Soy milk
1,00
0,40
0,74
Oat milk
0,42
Davis et al. (2010), who compared meals with varying protein sources (similar content of
protein, fat and energy), showed that a meal with a pea burger is associated with significantly
less GHG’s compared to a pork chop meal. However, this study highlighted the need for
efficient processing of products with vegetable protein such as veggie burgers, since these
products are often sold frozen due to small stock units, which can result in high-energy costs
for freezing and frozen storage. Of the other alternatives, especially some types of fish or
seafood (e.g. shrimps) have much higher emissions, due to the catching methods.
A second highlight in the table 5 is the great differences between emission numbers using
different system boundaries. Generally, the more “steps” from cradle to grave are included in
the analysis, the higher the emission for a product. This does not become immediately obvious
from the category averages (which may or may not include all products for each boundary
analyses), but comparing for a product across the different system boundaries one can see for
e.g. beef, cattle milk, or pork, that “cradle to retail” gives higher emissions than “cradle to farm
gate” or for the “whole farm model”.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
21
Table 6: Overview of emissions (average of kg CO2eq/kg product) for selected food
products comparing within the system boundary “cradle to farm gate” between different
countries and products.
Category/product
Norway
Nordic
West-Europe
Global
Meat
15,95
9,66
20,41
25,22
Beef
22,00
24,00
26,34
32,83
Dairy cows
16,06
18,95
11,27
Suckler cow
34,00
27,50
Sheep, Lam, Goat
57,00
21,50
Pork
4,50
4,48
6,12
3,49
Chicken
2,73
2,54
4,77
2,06
Dairy
1,53
1,05
1,17
1,10
Milk, cattle
1,53
1,05
1,17
1,10
Eggs
3,93
1,70
Fish
3,30
6,70
4,65
16,27
Cod
3,60
6,70
Herring
1,10
Pangasius
3,00
Salmon
3,23
8,20
Shrimp
22,90
Other food/drink
2,61
2,00
0,80
Wheat
0,53
Pulses
2,00
0,80
Rice
4,00
Vegetarian
2,48
Vegetarian burger
2,60
Tofu
2,00
Comparing between countries in table 6, using only one much used system boundary (cradle to
farm gate), suggests that Norway has lower emissions for a number of products such as beef
and meat from dairy cows, or fish, when compared to the Nordics, west-Europe or global
numbers. Some other products, such as meat from suckler cows or milk seem to end up higher
in Norway compared to the other regions, but to analyse the specific reason for this these
differences must be analysed and sometimes using reports and details from only single articles
to account for differences in system boundaries etc. Beef from South America for example has
a significantly higher climate impact than European beef due to high CO2 emissions from LUC
as well as high CH4 emissions due to low animal productivity. Other potential meat alternatives
such as fish, pulses or vegetarian also have much lower emissions than beef or cow meat per kg
product, but the difference with pork or chicken is much smaller. Only a few studies exist with
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
22
enough data to do inter-comparisons of emissions across production systems or regions. These
studies and comparisons feature in section 2.5 in this report where we focus on different
production methods, especially conventional versus organic systems and intensive versus
extensive systems
2
.
A comparison between products just for Norway (table 7), differentiating studies with different
system boundaries, shows similar findings. Cow meat generally has higher emissions (per kg
product) than other types of meat, with suckler cows having highest emissions, followed by beef
and then by veal and dairy cows. Pork meat and chicken meat register much lower emissions (5
to 10 times lower) in comparison with various cow meat. Fish has about 5 to 10 times lower
emissions, except for lobster due to the intensive fishing method. Vegetables and fruit also have
much lower emissions when comparing per kg product, with up to a factor of 100 difference
when comparing strawberries with veal. Cheese and butter are relative intensive dairy products
and have higher emissions than just milk.
The details for why certain emissions are higher than other will be discussed further down in
this report. The relative emissions for these food items stem from a number of selected studies,
but are comparable to many studies. Norden (2014) has similar findings for fish, finds that
vegetables in general are associated with fairly low GHG emissions and have generally lower
life-cycle GHG emissions than animal products. Grain products, e.g. wheat flour, typically have
emissions of around 0.5 kg CO2-eq per kg, while potatoes and other root vegetables such as
carrots are particularly efficient in cultivation, since the yield is high per ha, resulting in low
GHG emissions per kg product. GHG emissions from greenhouse products, such as tomatoes,
are very sensitive to the source of heating of the greenhouse. Substituting fossil fuels with
biofuels will thus have a significant impact on the product’s emissions. Generally, vegetables
grown in open air have lower emissions than products grown in greenhouses using fossil fuels,
but the report states - transport of such products can be of importance for vegetables
imported to the Nordic countries. As example they bring the well-known Spanish tomatoes vs
imported tomatoes example: transport emissions represent almost half of the Spanish tomatoes
total emissions, resulting in a slightly higher impact than (Swedish) tomatoes cultivated in
greenhouse with bio-fuels but significantly lower CF than tomatoes grown in greenhouse using
fossil fuels.
2
See brief list of definitions at the end of this report.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
23
Table 7: Emissions (average of kg CO2eq/kg product) for food products collected in this
report, for Norway (based on 21 available studies). The table differentiates between
different system boundaries used in the underlying studies, and averages emissions per
food category
Category/product
cradle to retail gate
cradle to farm gate
whole farm model
unknown
Beef
22,00
14,24
Dairy cows
16,06
21,40
Suckler cow
34,00
28,15
Veal/Young bulls
22,00
16,83
Sheep, Lam, Goat
18,70
27,64
Pork
4,95
4,50
2,58
Chicken
3,30
2,73
1,32
Milk, cattle
1,32
1,53
0,97
Butter
15,07
Margarine
1,50
Cheese
9,90
Cod
2,70
3,60
4,27
Herring
0,89
1,20
Lobster
86,20
Mackerel
0,95
Salmon
3,25
3,23
Saithe
2,60
Bread
0,94
Wheat
0,87
0,53
Oats
0,75
Potatoes
0,43
Rice
4,00
4,00
Cabbage and
roots
0,32
Apples
0,30
Strawberries
0,22
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
24
2.2.2 Comparison of functional units
In a final general comparison of emissions between different products and product groups, it is
useful to look at the functional unit or different ways of expression of emissions. As explained
in section 1.4.1, one should not “compare apples and oranges”, but depending on the needs,
this may indeed be exactly what is required. Thus, a comparison of emissions between products
based on weight (should I eat 500 gram of meat today, or should I replace that with 500 gram
of fish?) depends on whether food items are in fact potential substitutes for each other or if
they serve very different purposes in a diet. The emissions then also depend on the
requirements; is it valid to compare emissions of the amount of food or should you look
instead how fish could replace the energy, proteins or other nutrients that are provided by meat?
Finally it would be relevant to ask if the replacement would fit with the other items on the
plate for the dinner planned that day, or the quality or financial aspects when purchasing or
comparing fish versus meat.
Table 8 shows the emissions of some comparable food items (edibles, including meat, fish and
vegetarian options in the upper section of the table, and dairy or drinkables in the lower section
of the table. It is clear that meat has more proteins per kg than most other products (except
cheese). It is also clear that while meat has a high-energy content, fish and several vegetarian
substitutes (but not tofu or pulses) are higher in energy. In general, the CO2-emission per kg
food is much higher for the animal products than for the plant products, although the
differences decrease especially between meat and milk when the energy content of food is
considered. The emission from cattle meat is from about 11 to 23 CO2-eq per kcal, from milk
is around 2.5 CO2-eq per kcal, while production of wheat only contributes with from around
0.2 kg CO2-eq per kcal. The emissions however are highest for meat regardless of functional
unit.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
25
Table 8: Overview of emissions related to a number of selected food products, comparing
emissions per kg product, per kg protein, and on an energy basis (kcal) for comparable
system boundaries used in the studies in the database: cradle-to-farm-gate for edibles,
and whole LCA without waste for dairy products. The red gradation indicates in which
edible group and items the highest emissions are, while the green gradation indicates
likewise for dairy products.
Product
kg CO2eq
/kg
gr.protein
/100 gr
kg
CO2(eq.)/kg
protein
Kcal
/kg
gr. CO2eq
/kcal
Edibles: cradle to farm gate
Meat:
Beef
29,60
21
170,19
1440
15,28
Dairy cows
15,33
21
66,30
1440
11,15
Suckler cow
29,67
21
160,23
1440
23,61
Sheep, Lam,
Goat
41,57
20
238,00
2210
18,81
Pork
5,36
19
28,19
2230
2,02
Chicken
3,46
19
16,78
1970
1,39
Fish:
Salmon
4,22
20
21,20
2240
1,44
Mackerel
3,18
19
16,00
1870
1,70
Cod
5,67
18
28,33
810
4,44
Herring
1,10
17
5,00
2930
0,38
Eggs:
3,40
12
26,60
1420
2,39
Vegetarian:
Wheat
0,53
12
4,34
3355
0,16
Rice
4,00
8
52,63
3515
1,14
Tofu
2,00
8
17,00
770
2,60
Vegetarian burger
2,60
7
16,00
1920
0,57
Pulses
1,20
2
5,33
1140
1,75
Dairy: whole LCA excluding waste
Milk, cattle
1,14
3
33,93
463
2,46
Soy milk
0,40
4
10,00
410
0,98
Cheese
6,80
27
25,18
3510
1,94
Yoghurt
1,24
4
28,84
685
1,81
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
26
Looking at meat only, cow meat again has higher emissions regardless of functional unit, except
for mutton. Within the category cow-meat, dairy cows are most “climate friendly” especially in
terms of protein, although these numbers are in relation to what the “Norwegian Food
Composition Table 2016” reports as protein and energy content and this source does not
distinguish between beef and dairy or suckler cows. For the dairy products, cheese is an outlier
about both protein and energy content. When comparing emissions for milk and soymilk we
see that milk has higher emissions both per weight, per protein content and per energy content.
Functional units can have a great say in the different ways of allocation of emissions, as
Gonzalez-Garcia et al. (2013a, in Notarnicola et al 2015) note when discussing the effect of
different allocation methods among milk, cream and butter on the total life cycle results: in
addition to a mass allocation approach, the authors performed a sensitivity analysis in which
economic and protein-based allocations were applied to the system. The results showed that
economic allocation improved the environmental performance of milk production by 34 %,
whereas protein-based allocation worsened the results by up to 5 %. Gonzales Garcia et al.
(2013b) analysed the effect of different allocation approaches and found that mass allocation
improved the impact of cheese more than the economical one, because the economic value of
whey per unit of mass is lower than that of cheese.
2.3 Dairy
In this section, we will analyse in depth what the available literature finds regarding dairy
emissions of production in Norway, compared to other regions, and considering different
production systems (conventional, organic, etc.). As earlier described in the report, emissions
are very dependent on the system boundaries. This means we can only compare countries and
production systems when also the system boundaries are taken into consideration. Even then,
there will be differences in what the analysis includes and excludes, but the available literature
and differences between studies still makes the results valid.
Table 9 shows, not surprisingly, that milk has the lowest emissions as compared to “milk
derivatives” butter, cheese, cream and yoghurt. Especially cheese production is emission
intensive compared to milk. Also not surprisingly, there is a trend towards the more inclusive
the system boundary, the higher the emissions, though this finding is not consistent, and there
are large variations between the different system boundary emissions. Also the ranges are large
at times: For Norway, the average emission for milk across system boundaries and production
methods is 1.15 kg CO2eq/kg. The range (0,50-1,92) is larger than the emission factor itself,
indicating that there is great variability in the emissions, due to many factors: system boundary,
production method, and between farms (with different soils, number of animals, yield per
animal, energy use, etc.).
Because dairy cows need to be milked regularly, distances to the milking parlour are usually
short. This means intensive grazing takes place nearby the farm, or grazers are kept indoors
permanently. Therefore, Nijdam et al. (2012) argue, livestock management systems of dairy
farms generally do not vary greatly, with values between 1 and 1.5 kg CO2-eq/kg milk (12
studies). Weiske et al. (2006) give an average of 1.4 kg CO2-eq/kg for milk for the EU-15. In a
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
27
study by the FAO (2010), an average of 1.3 kg CO2-eq/kg is calculated for Western Europe.
The differences can be traced back to soil condition and consequent N2O emissions (De Vries
and De Boer, 2010), feed composition and race (related to yield) (Vergé et al., 2007), intensity
of farming (mainly related to yield and diet) and manure management (Haas et al., 2001;
Phetteplace et al., 2001; Weiske et al., 2006).
Table 9: Overview of emissions related to milk and dairy, comparing Norway with other
regions, and distinguishing between different production methods and system boundaries.
Emissions are given as average and range in kg CO2eq/kg product.
Product / System boundary
/ Production method
Norway
Nordic
West-
Europe
Europe
Global
Milk, cattle
1,15
0,50-1,92
1,06
0,87-1,24
1,33
0,95-1,70
1,40
1,30-1,50
3,75
1,00-10,80
cradle to grave
Conventional
1,66
1,23-2,4
whole LCA (excluding
waste)
Conventional
1,14
1,09-1,24
cradle to retail gate
1,32
0,84-1,92
0,94
0,87-1,00
1,59
1,19-1,70
3,86
1,00-10,80
Conventional
1,32
0,84-1,92
0,94
0,87-1,00
1,55
1,19-1,70
3,72
1,00-10,00
Organic, grass based
1,67
1,60-1,70
4,01
1,50-10,80
cradle to farm gate
1,53
1,47-1,59
1,05
1,00-1,10
1,17
0,95-1,50
1,40
1,30-1,50
1,10
1,09-1,10
Conventional
1,53
1,47-1,59
1,05
1,00-1,10
1,18
0,95-1,50
1,40
1,30-1,50
1,10
1,09-1,10
Mixed
1,10
Organic
1,19
Organic, grass based
1,05
whole farm model
0,97
0,50-1,36
Conventional
0,92
0,50-1,30
Organic, grass based
1,07
0,82-1,36
Yoghurt
1,24
whole LCA (excluding
waste)
Conventional
1,24
Cream
5,22
2,96-6,12
cradle to grave
Conventional
5,22
2,96-6,12
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
28
Product / System boundary
/ Production method
Norway
Nordic
West-
Europe
Europe
Global
Butter
15,07
8,80-22,40
21,55
9,50-27,60
cradle to grave
Conventional
9,50
cradle to retail gate
Conventional
15,07
8,80-22,40
25,57
23,50-27,60
Margarine
1,50
1,05-1,95
cradle to retail gate
Conventional
1,50
1,05-1,95
Cheese
9,90
6,30-14,40
6,80
3,44-9,23
7,97
6,80-9,00
whole LCA (excluding
waste)
Conventional
6,80
3,44-9,23
cradle to retail gate
Conventional
9,90
6,30-14,40
7,97
6,80-9,00
Overall, the picture that emerges is that Norway has higher milk production emissions
compared to other countries in the Nordic region (Sweden, Finland, Denmark), but lower
emissions than (Western) Europe and globally. This is true when one looks at cradle-to-retail
boundaries. However, when comparing cradle-to-farm-gate analyses, Norway has highest milk
emissions across the compared regions. The variety or range between emission data from
Norwegian studies is the largest compared across the regions, indicating that the greatest variety
is within Norway, and not between Norway and other regions. However in the cradle-to-farm-
gate studies, the lowest Norwegian emissions for milk (1.47) are higher or near the highest
emissions for other regions (1.10-1.50) suggesting that Norwegian milk has higher emissions
than elsewhere at least when comparing conventional production. Butter production on the
other hand seems to be less emission intensive in Norway than in west Europe, while cheese
production is more emission intensive both compared to west Europe and other Nordic
countries.
Table 10 compares dairy products within single studies and confirms the results of table 9: milk-
derived products in general have larger emissions than milk itself, and especially cheese has
relatively high emissions.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
29
Table 10: This table shows emissions of a variety of dairy products compared in two
studies/reports, one in Denmark and one in the UK. Within each of the studies, the same
production system (conventional), system boundaries (whole LCA excluding waste in the
one, cradle to grave in the other) and methodology are used, making emissions numbers
comparable within each study for different products. Reference: Werner et al. 2014
(Denmark) and Tesco 2012 (UK).
Product
kg CO2eq/kg
Denmark, whole LCA (ex. waste)
Milk, mini milk 0,50% fat
1,09
Milk, skim milk 0,30% fat
1,09
Milk, butter milk 0,50% fat
1,24
Milk, yoghurt 0,50% fat
1,24
Cheese, 20+ 17% fat
8,47
Cheese, 30+ 31% fat
9,23
Cheese, smoked
6,05
Cheese, cottage 20+ 4% fat
3,44
Ice cream
2,80
UK, cradle to gave
Semi-Skimmed Milk
1,41
Skimmed Milk
1,23
Whole Milk
1,58
Tesco Fresh Single Cream
2,96
Tesco Fresh Double Cream
6,12
Tesco Fresh Extra Thick Double Cream
6,12
Tesco Whipped Cream
5,00
Tesco Fresh Whipping Cream
5,10
Creamfields Cream
6,00
Tesco English Salted/Unsalted Butter
9,50
Comparing the Danish study (table 10) with Norway (table 9) for “whole LCA” is not possible,
but judging from the other system boundaries for we see that Denmark likely has lower
emissions than Norway. The English study (table 10) is similarly not comparable to Norway,
but - again judging from the other system boundaries - it appears that Norwegian and English
product emissions are fairly similar.
A final comparison of emissions of dairy products in the Nordic countries (including Norway)
lists the different studies and system boundaries used in these.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
30
Table 11: Overview of emissions of dairy products in the Nordic countries (including
Norway) from the different studies, listing product, country, system boundaries and
reference.
Product
kg
CO2eq/kg
Location
Main system boundary
Reference
Milk, cattle
0,99
Denmark
cradle to farm gate
Kristensen et al. (2011)
Milk, cattle
1,09
Denmark
whole LCA (excluding
waste)
Werner et al. (2014)
Milk, cattle
1,09
Denmark
whole LCA (excluding
waste)
Werner et al. (2014)
Milk, cattle
1,24
Denmark
whole LCA (excluding
waste)
Werner et al. (2014)
Milk, cattle
1,02
Norway
whole farm model
Bonesmo et al. (2013)
Milk, cattle
0,82
Norway
whole farm model
Bonesmo et al. (2013)
Milk, cattle
1,36
Norway
whole farm model
Bonesmo et al. (2013)
Milk, cattle
1,20
Norway
cradle to retail gate
Hille et al. (2012)
Milk, cattle
0,84
Norway
cradle to retail gate
Hille et al. (2012)
Milk, cattle
1,92
Norway
cradle to retail gate
Hille et al. (2012)
Milk, cattle
1,17
Norway
whole farm model
Mittenzwei (2015)
Milk, cattle
1,47
Norway
cradle to farm gate
Roer et al. (2013)
Milk, cattle
1,59
Norway
cradle to farm gate
Roer et al. (2013)
Milk, cattle
1,54
Norway
cradle to farm gate
Roer et al. (2013)
Milk, cattle
0,64
Norway
whole farm model
Storlien and Harstad
(2015)
Milk, cattle
0,50
Norway
whole farm model
Storlien and Harstad
(2015)
Milk, cattle
1,00
Norway
whole farm model
Åby et al. (2015)
Milk, cattle
0,90
Norway
whole farm model
Åby et al. (2015)
Milk, cattle
1,30
Norway
whole farm model
Åby et al. (2015)
Milk, cattle
0,87
Sweden
cradle to retail gate
Cederberg and Flysjö
(2004b)
Milk, cattle
1,00
Sweden
cradle to farm gate
Cederberg and Flysjö
(2004b)
Milk, cattle
1,10
Sweden
cradle to farm gate
Cederberg and Flysjö
(2004a)
Milk, cattle
1,05
Sweden
cradle to farm gate
Cederberg and Stadig
(2003)
Milk, cattle
1,00
Sweden
cradle to retail gate
de Vries and de Boer
(2010)
Milk, cattle
0,99
Sweden
cradle to retail gate
Smedman et al. (2010)
Butter
14,00
Norway
cradle to retail gate
Hille et al. (2012)
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
31
Product
kg
CO2eq/kg
Location
Main system boundary
Reference
Butter
8,80
Norway
cradle to retail gate
Hille et al. (2012)
Butter
22,40
Norway
cradle to retail gate
Hille et al. (2012)
Margarine
1,50
Norway
cradle to retail gate
Hille et al. (2012)
Margarine
1,05
Norway
cradle to retail gate
Hille et al. (2012)
Margarine
1,95
Norway
cradle to retail gate
Hille et al. (2012)
Cheese
11,30
Denmark
Hille et al. (2012)
Cheese
8,47
Denmark
whole LCA (excluding
waste)
Werner et al. (2014)
Cheese
9,23
Denmark
whole LCA (excluding
waste)
Werner et al. (2014)
Cheese
6,05
Denmark
whole LCA (excluding
waste)
Werner et al. (2014)
Cheese
3,44
Denmark
whole LCA (excluding
waste)
Werner et al. (2014)
Cheese
9,00
Norway
cradle to retail gate
Hille et al. (2012)
Cheese
6,30
Norway
cradle to retail gate
Hille et al. (2012)
Cheese
14,40
Norway
cradle to retail gate
Hille et al. (2012)
Cheese
8,80
Sweden
Hille et al. (2012)
The relatively low emissions of milk production in (Norway and) the Nordics (see tables 9 and
11) compared to Europe and at the global level, is discussed in several studies (e.g. Norden,
2014). These low emissions are especially due to the high animal productivity and high feed
efficiency in Europe. Several studies (table 11) find that emissions from milk in Norway, Sweden
and Denmark have a carbon footprint at the farm-gate slightly above 1 kg CO2-eq per kg milk
(in fact, 1.11 kg CO2eq/kg on average of presented studies), not including emissions from LUC.
The Norden study concludes that “adding these emissions (the FAO estimates close to 0.1 kg
CO2 per kg milk from LUC for European milk) as well as post-farm emissions suggests that
milk production from Nordic countries lies in the lower range of European milk production
and thus worldwide”. Indeed, in their FAO report, Gerber et al. (2013) find that industrialized
regions in the world exhibit the lowest emissions for milk, ranging between 1,6-1,7 kg CO2-
eq/kg (FPCM, as expressed for milk emissions in most studies), which is higher than the
Nordics or Norway. Emissions in developing countries on the other hand emissions for milk
range between 2-9 kg CO2-eq/kg FPCM, the latter being milk emissions for sub-Saharan Africa.
In the following sections of this report, we will first analyse emissions in Norwegian meat
production versus other regions, considering different system boundaries. Then we will
compare emissions of dairy and meat production for different production methods, and discuss
which life cycle stages of meat and dairy have the greatest climate impact.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
32
2.4 Beef
This section deals more in-depth with the emissions from meat production, comparing Norway
with other countries and regions in the world. Table 12 shows that there are great differences
in emissions between different sorts of meat production, at the global all the way down to the
Norwegian scale. To understand the differences, it is necessary to compare within system
boundaries and production methods. Looking then at the whole farm (marked in yellow), with
conventional farming, we see that beef has the lowest emissions (14.24), together with veal
(15.40), closely followed again by dairy cows (19.09) and finally suckler cows (28.15)
3
. In
Norway, beef comes from dairy cows, suckler cows and veal. From the analysed studies it was
not clear which of these “beef” referred to, so we have kept this category. Looking however at
the cradle-to-farm-gate boundary with conventional production (orange marking), we see that
dairy cows have the lowest emissions (17.33), followed by beef (22.00) and again suckler cows
having most emissions (34.00). Comparing these relative distributions for the other regions, e.g.
west Europe, Europe or globally (green marking), we see a similar picture to the latter, with
dairy cows having lowest emissions followed by beef and finally suckler cows with the highest
emissions.
This latter pattern, with dairy cow emissions lowest and beef and suckler cow emissions higher
is indeed consistent with what several studies write about the comparison of emissions between
these different meat production systems, which at the same time explains a major difference
between European (and Norwegian) beef production as compared to other parts in the world.
E.g. the FAO in Opio et al (2013) describes that emissions for beef in much of the industrialized
world (Western Europe, North America and Oceania) is lower than the global average mainly
because these regions have a high efficiency in production and high feed digestibility.
3
See brief list of definitions at the end of this report.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
33
Table 12: Overview of emissions related to cattle meat, comparing Norway with other
regions, and distinguishing between different production methods and system boundaries.
Emissions are given as average and range in kg CO2eq/kg product.
Product / System boundary
/ Production method
Norway
Nordic
West-
Europe
Europe
Global
Beef
16,83
13,70-22,00
25,33
20,00-28,00
25,28
9,00-129,00
30,25
26,00-39,00
39,95
9,90-103,00
whole LCA (excluding
waste)
Conventional
27,99
28,70
cradle to retail gate
21,22
17,30-24,10
44,59
17,40-103,00
Conventional
20,43
17,30-24,10
45,22
18,90-103,00
Organic, grass based
22,00
20,40-23,90
43,92
17,40- 93,40
cradle to farm gate
22,00
24,00
20,00-28,00
26,34
9,00-129,00
30,77
26,00-39,00
32,83
9,90-80,00
Conventional
22,00
24,00
20,00-28,00
22,38
9,00-42,00
30,77
26,00-39,00
35,05
9,90-80,00
Free-range
21,80
Mixed
71,60
14,20-129,00
14,00
Organic
19,05
18,20-19,90
21,30
12,00-34,90
whole farm model
Conventional
14,24
13,70-14,79
Dairy cows
19,49
11,00-37,46
18,95
15,60-22,30
11,27
9,00-15,80
15,95
12,00-19,90
18,40
cradle to retail gate
Conventional
18,40
cradle to farm gate
16,06
11,00-18,40
18,95
15,60-22,30
11,27
9,00-15,80
15,95
12,00-19,90
Conventional
17,33
15,00-18,40
11,27
9,00-15,80
15,95
12,00-19,90
Organic, grass based
11,00
18,95
15,60-22,30
whole farm model
21,40
12,00-37,46
Conventional
19,09
15,40-25,00
Organic, grass based
24,28
12,00-37,46
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
34
Product / System boundary
/ Production method
Norway
Nordic
West-
Europe
Europe
Global
Suckler cow
30,10
25,00-34,00
27,50
25,00-30,00
cradle to farm gate
Conventional
34,00
27,50
25,00-30,00
whole farm model
Conventional
28,15
25,00-31,30
Veal/Young bulls
19,04
11,75-32,00
16,97
7,40-28,00
cradle to retail gate
Conventional
22,00
14,00-32,00
16,97
7,40-28,00
whole farm model
16,83
11,75-22,90
Conventional
15,40
Organic, grass based
17,30
11,75-22,90
On average, European beef has the lowest carbon footprint in the world, because much (80%)
of its beef comes from the dairy sector (slaughtered dairy cows, bull dairy calves), and the region
has a generally high animal productivity (Opio et al., 2013; Gerber et al., 2013). Indeed, in
Norway, Ulleberg (forthcoming) reports that around 75% of beef production comes from dairy
farms. Such a combination of both beef and milk production reduces the emissions as these are
distributed over more products (both milk and beef), though the ultimate allocation depends
on the productivity, or how much milk and meat dairy cows produce. Nijdam et al. (2012) argue
that the environmental impact of the beef from culled dairy cows is lower than that from beef
cattle mainly due to the relative efficient co-production of meat and milk in intensive systems.
The meat production from the dairy sector is also a consequence of the need to sustain milk
production through production of calves in order to keep cows lactating.
Cows reared for both milk and meat live longer (and thus produce more methane and other
emissions) than cows reared solely for meat. Studies such as the “UK GHG inventory report
1990-2012” report that beef cows produced about half the amount of methane compared to
dairy cows, which suggests that dairy cows would have about double total emissions than beef
cows. This is clearly not reflected in table 12, because it is the productivity of an animal (and
thus the distribution over milk and/or meat) that ultimately determines the emissions that a
beef or dairy cow ends up with.
Although the data in table 12 may look comparable within the system boundaries used, it is
important to notice that not all studies use the same emission factors even when using the same
or similar system boundary. A typical factor that is often omitted is land use and land-use change
related emissions, which can add a significant portion to the final emissions for beef (or milk).
For example, for beef production in Latin America pastures may be expanded into forested
areas. Consequently, land-use change is a major driver of emissions in the region, representing
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
35
approximately one-third of the footprint (Opio et al., 2013), equivalent to 24 kg CO2-eq/kg CW
an estimate with a high level of uncertainty.
Finally, between studies and even within studies calculated emissions may vary greatly. Table 13
highlights this point showing the sources of the emissions data used in this report, and the
variation within single studies even for specific sources of meat.
Table 13: Overview of emissions of cattle meat products in the Nordic countries (including
Norway) from the different studies, listing product, country, system boundaries and
reference.
Product, short
kg
CO2eq/kg
Location
System boundary
Reference
Beef
27,99
Denmark
whole LCA (excl.
waste)
Werner et al. (2014)
Beef
13,7
Norway
whole farm model
Grønlund and Mittenzwei (2016)
Beef
22
Norway
cradle to farm gate
Leip et al (2010)
Beef
14,79
Norway
whole farm model
Mittenzwei (2015)
Beef
28
Sweden
cradle to farm gate
Cederberg et al. (2009b)
Beef
20
Sweden
cradle to farm gate
Cederberg et al. (2009b)
Beef
32
Sweden
LMD (2016)
Beef
20
Sweden
LMD (2016)
Beef
23
Sweden
LMD (2016)
Beef
39
Sweden
LMD (2016)
Beef
29
Sweden
LMD (2016)
Beef
22
Sweden
LMD (2016)
Beef
29
Sweden
LMD (2016)
Beef
40
Sweden
LMD (2016)
Dairy cows
21,67
Norway
whole farm model
Bonesmo et al. (2013)
Dairy cows
12
Norway
whole farm model
Bonesmo et al. (2013)
Dairy cows
37,46
Norway
whole farm model
Bonesmo et al. (2013)
Dairy cows
15,4
Norway
whole farm model
Grønlund (2015)
Dairy cows
26
Norway
whole farm model
Grønlund and Harstad (2014)
Dairy cows
21,06
Norway
whole farm model
Grønlund and Mittenzwei (2016)
Dairy cows
15
Norway
cradle to farm gate
Refsgaard et al. (2011)
Dairy cows
11
Norway
cradle to farm gate
Refsgaard et al. (2011)
Dairy cows
17,7
Norway
cradle to farm gate
Roer et al. (2013)
Dairy cows
18,4
Norway
cradle to farm gate
Roer et al. (2013)
Dairy cows
18,2
Norway
cradle to farm gate
Roer et al. (2013)
Dairy cows
18
Norway
whole farm model
Åby et al. (2015).
Dairy cows
16
Norway
whole farm model
Åby et al. (2015).
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
36
Product, short
kg
CO2eq/kg
Location
System boundary
Reference
Dairy cows
25
Norway
whole farm model
Åby et al. (2015).
Dairy cows
18
Sweden
Cederberg and Darelius (2000)
Dairy cows
22,3
Sweden
cradle to farm gate
Cederberg and Stadig (2003)
Dairy cows
15,6
Sweden
cradle to farm gate
Cederberg and Stadig (2003)
Suckler cow
31,3
Norway
whole farm model
Grønlund (2015)
Suckler cow
34
Norway
cradle to farm gate
Refsgaard et al. (2011)
Suckler cow
25
Norway
whole farm model
Åby et al. (2015).
Veal/Young
bulls
17,25
Norway
whole farm model
Bonesmo et al. (2013)
Veal/Young
bulls
11,75
Norway
whole farm model
Bonesmo et al. (2013)
Veal/Young
bulls
22,9
Norway
whole farm model
Bonesmo et al. (2013)
Veal/Young
bulls
15,4
Norway
whole farm model
Grønlund and Mittenzwei (2016)
Veal/Young
bulls
20
Norway
cradle to retail gate
Hille et al. (2012)
Veal/Young
bulls
14
Norway
cradle to retail gate
Hille et al. (2012)
Veal/Young
bulls
32
Norway
cradle to retail gate
Hille et al. (2012)
2.5 Production methods
2.5.1 Comparison of organic and conventional production
There is an ongoing debate about the merits of ecologic or organic farming methods versus
conventional methods. The debate originally revolved around the assumed differences in
impacts on the environment, e.g. “organic agriculture which is often seen by the public as
producing food free of chemicals and being more environmentally friendly as compared to
poorly managed conventional farms” (e.g. Trewavas, 2004). While the environmental
friendliness and management practices are part of a wider discussion, the focus in this section
is specifically on the climatic impact of production systems.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
37
Table 14: Literature emission data for Norwegian dairy and cattle meat under different
production methods and system boundaries compared to other regions. Emissions are
given as average and range in kg CO2eq/kg product. Red markings illustrate lower
emissions for conventional methods, and green markings illustrate lower emissions for
organic production. Yellow indicates complete overlap of ranges.
Region
cradle to farm gate
cradle to retail gate
Product
Conventional
Organic
Conventional
Organic
Norway
Beef
16,83
13,70-22,00
Dairy cows
18,31
15,00-25,00
21,63
11,00-37,46
Suckler cow
30,10
25,00-34,00
Veal/Young bulls
15,40
Not available
17,30
11,75-22,90
22,00
14,00-32,00
Milk, cattle
1,12
0,50-1,59
1,07
0,67-1,36
1,32
0,84-1,92
Nordic
Beef
24,00
20,00-28,00
Dairy cows
18,95
15,60-22,30
Milk, cattle
1,02
0,75-1,32
0,99
0,67-1,29
0,94
0,87-1,00
West-Europe
Beef
22,38
9,00-42,00
19,05
18,20-19,90
20,43
17,30-24,10
22,00
20,40-23,90
Dairy cows
11,27
9,00-15,80
Suckler cow
27,50
25,00-30,00
Veal/Young bulls
16,97
7,40-28,00
Milk, cattle
1,18
0,95-1,50
1,19
Not available
1,55
1,19-1,70
1,67
1,60-1,70
Various studies have compared the environmental impacts of conventional, integrated and
organic farming (e.g. Refsgaard et al 2011). Trewavas (2004) mentions that there are economic
and environmental considerations for organic production, which uses less energy, and preserves
biodiversity and soils better. Indeed, the study finds that “organic farming practices generally
have positive impacts on the environment per unit of area, but not necessarily per product unit.
The variation between farms and systems however is very wide, and the only significant
differences between organic and conventional systems found in the study were soil organic
matter content, nitrogen leaching, nitrous oxide emissions per unit of field area, and land use
(all higher in organic production), and energy use (lower in organic production).
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
38
Norway is well behind Sweden and Denmark in consumption of ecological products, but
increased ecological production and consumption is a political target: 15 percent of the
production and consumption of food (both national produce and import) should be ecologic
by 2020 (Solemdal and Friss Pederssen, 2014). The reason for this is especially environmental,
and not necessarily climatic. There are somewhat different signals in the popularity of both
production and sales of ecological food, including meat and dairy. Solemdal and Friss Pederssen
(2014) find that most ecological produce is sold in Oslo and Akershus, where its popularity is
increasing an increase with 16% in supermarkets in 2013, and with 14% through other
(informal) channels. SLF (2013) reports that especially the number of ecologically fed cattle has
been increasing over the years. As a result, the report finds that ecological production of cow
milk has increased slightly from 3,4 to 3,5 percent of the total milk production, in spite of a
decrease in producers of ecological milk. Paradoxically, the ecological production of cattle meat
has been decreasing. The reasons for these may reside in price differences and subsidies.
Surprisingly, the production of ecological milk increases more than the actual sales (Stette
Høyberg, 2016), and this difference is increasing. One potential explanation for this is that
organically produced milk is increasingly mixed with conventionally produced milk prior to
sales. Some supermarkets (e.g. Rema 1000 - in Solemdal and Friss Pederssen, 2014) on the other
hand report that delivery/purchase of ecological food to supermarkets is a problem, except for
milk and one reason for the fluctuations in demand can be the price differences between
conventional and ecological products: especially for meat the difference can be high.
However, what are the emissions and climatic differences between organic and ecologic
production methods? Because of their lower impact on the environment, ecologic products are
intuitively expected to have a lower impact on climate, with lower emissions than conventional
production systems. However, in contrast to other environmental impacts, the GHG emission
differences are much less clear (Trewavas, 2004; Refsgaard et al., 2011). Table 14 sums up
findings from previous tables on emissions for meat and milk under different production
systems and system boundaries. The data are inconclusive, and point to no or only small
differences between production methods. While emissions for meat production from
Norwegian dairy cows and veal seem to be lower in a conventional production system than in
an organic production system, this difference is smaller and non-significant for veal. For milk,
the organic system seems to have lower climatic impact, but the range for conventional
production completely overlaps with the narrower range for organic production, and the small
difference in the average emissions is therefore not significant. For the Nordics, only a
comparison for milk is possible. The results are similar to Norway with a lower climatic impact
for organic production, but here the lowest organic emissions are slightly lower than for
conventional production. On the west-European scale however, most results are inconclusive.
If only considering the averages, conventional milk production gives lower emissions than
organic, while the results for meat production (beef) depend on the method (or study) used.
However, ranges are overlapping for all west-European products and studies, making the results
very inconclusive. Overall, it seems that in Norway conventional production is better for meat,
while organic production may be slightly better for milk production.
Comparing these findings with other studies, we find contrasting or diverging results. For milk,
Kristensen et al. (2011) however found in their study in Denmark that emissions were larger in
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
39
the organic system (1.27) compared to conventional (1.20). This is due to higher methane and
nitrous oxide emissions and lower milk production per animal found Thomassen et al. (2008),
who reported the same difference. Tuomisto et al. (2012) report that only Cederberg and
Mattsson (2000) and one of the cases in Olesen et al. (2006) found lower GHG emissions from
organic milk production. Refsgaard et al. (2011) on the other hand report on about 20 Swedish
case studies where emissions were higher for conventional milk and lower for organic milk
(about 0.98 kg CO2-eq per kg for conventional and 0.95 for organic). Other studies in the
Nordics or west Europe conclude that there was no difference between organic and
conventional production systems in terms of GHG emission per kg milk (Cederberg and Flysjö,
2004b; Thomassen et al., 2008; Trewavas, 2004; Tuomisto et al., 2012). For beef, both Refsgaard
et al. (2011) and Tuomisto et al. (2012) find that organic beef had lower emissions due especially
to lower emissions from industrial inputs (referring to Casey and Holden, 2006) contrary to
the findings of Table 15 for Norway.
Hille et al. (2012) report on a number of comparisons in GHG emissions between organic and
conventional productions systems. While for plant foods (but not vegetable), a majority of
comparative LCAs seem to suggest that organic products have lower carbon footprints than
conventional products, for milk and meat they too find that the results are split. In the case of
milk, most studies only show small differences between organic and conventional products,
while for meat the results diverge with some studies indicating that GHG emissions from
organic production were significantly higher and others the opposite. Refsgaard et al. (2011)
mention the importance of considering the total emissions versus the “per kg product”: milk
and beef meat generally have lower emissions in an organic production system than from the
conventional system considering overall average numbers for different types of model farms.
“The average emissions for conventionally products are from 30% to 70% higher than for the
organically products with the lowest difference for beef. There is however variation in the CO2-
emission for each of the analysed products depending on type of production system. The CO2-
emission from beef meat produced in combination with milk is only half the CO2-emission
when produced from suckler cows where the emission is around 34 kg CO2-eq per kg beef
meat.”
Some of the lack of differences can be explained by methane emissions from enteric
fermentation being higher per unit product in organic than in conventional systems, while
emissions from production of feed tend to be lower. Mondelaar et al. (2009) (in Hille et al. 2012)
pointed out that the avoidance of artificial fertilisers and pesticides in organic production, along
with less use of feed concentrates (kraftôr), had a downward influence on GHG emissions in
organic production (because of the decrease land-use effects). However, higher methane
emissions from ruminants due to a smaller fraction of concentrates in their feed (digestion
effect) and more fuel consumption for mechanical weed control were among factors with an
opposite effect, although Refsgaard et al. (1998) and others (see Hille et al. 2012) found no
differences in diesel consumption between the two systems. Hille et al. point out that yields and
fuel consumption also have an important influence on the ultimate emissions higher with low
yields.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
40
2.5.2 Other production methods
Production methods of course extend beyond the comparison of organic versus conventional.
There are many farm-level production system differences that have a large influence on the
ultimate emissions for meat or milk. Bergslid et al. (2016) mention in this context various case
studies that have partly contradicting conclusions depending on the actual management
intensity, farming context, climate and soil conditions, crops, etc. There are many different
management systems and combinations of these in cattle farming example include the duel
use of dairy cows for milk and meat, meat production from suckling cows or from intensively
fattening of oxen indoors or extensively kept and fed castrated oxen in pastures. Notarnicola et
al. (2013) note that the amount of GHG emissions per kg beef (or milk) depends on these
different cattle farming systems can have very different climate impacts.
Leip et al. (2010) compare emissions from beef on a European scale, and find that differences
in systems but also climate and other factors can be as large ranging from 6,49 kg CO2eq per
kg meat in the Italian region “Abruzzo” to 51,16 kg in the Finish region “Laensi-Suomi” (mainly
due to high emissions from organic soils). Importantly, with regard to which production system
gives lowest emissions, they find that there may be various solutions: the best performing
countries are not necessarily characterized by similar production systems, and be as diverse as
Austria and the Netherlands. While the Netherlands save emissions especially with low methane
and N2O rates indicating an efficient and industrialized production structure, Austria
outbalances the higher methane emissions by lower emissions from land use and land use
change (LULUC) indicating high self-sufficiency in feed production and a high share of grass
in the diet. However, both countries are characterized by high meat yields, while emissions in
Norway are relatively high in part due to low meat (and milk) yields and thus a less efficient
production structure. For meat, intensive maize systems show the lowest, and extensive systems
(such as in Norway) the highest emissions.
The type and quality of feed has a large influence on methane emissions: it has already been
mentioned that concentrates lower the methane emissions, while Leip et al. (2010) find a relation
between high methane emissions with animals spending much time on pastures. The FAO
(www.fao.org/gleam) also indicates that feed quality is closely correlated with enteric emissions:
Poorly digestible rations, i.e. highly fibrous ingredients, yield higher enteric methane emissions,
while Grøndahl (2010) finds that cows fed ryegrass had the lowest methane emission (25 g/kg
dry matter intake) and red clover had the highest emission (51 g/kg dry matter intake).
N2O emissions increase with the share of solid systems or manure fallen on pastures. Different
manure management systems can lead to different emission levels, and in general terms,
methane emissions are higher when manure is stored and treated in liquid systems (lagoons or
ponds), while dry manure management systems such as drylot or solid systems tend to increase
nitrous oxide emissions. Finally, high CO2 emissions (electricity, transportation) indicate a
strong dependency on feed imports and, in general, feed crops.
For milk, Leip et al. report a variability of emissions in Europe ranging from 0.41 kg CO2-eq
per kg of milk in the Italian region “Abruzzo” to 3.03 kg in the Greek region “Kriti”. To some
degree such differences can be attributed to lower milk yields, such as in Norway. However, if
feed concentrates (which give higher milk yields) are imported from overseas, they again are
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
41
often accompanied by higher emissions from land use change, as in the case of the Netherlands,
which is a typical example of an intensive system creating very low methane emissions and NO2
emissions, but “overcompensates” these by land use and land use change emissions. Overall,
the authors find that intensive maize and extensive grassland systems produce the lowest total
emissions while free ranging subsistence and climate-constrained systems emit more.
Nijdam et al. (2012) reviewed 15 LCA studies on beef production in a variety of cattle farming,
finding that the production of 1 kg of extensively farmed beef results in three to four times as
many greenhouse gas emissions as the equivalent amount of intensively farmed beef. According
to these authors, the differences in feed transformation efficiency are higher in intensive
systems; Intensive production systems result in higher total production levels and higher feed
efficiency (based on higher quality feeds) in intensive production systems results in lower GHG
emissions per kg of product compared to extensive systems (Nijdam et al. 2012). Peters et al.
(2010) compared grass-fed with the feedlot systems in Australia, similarly finding lower total
GHG emissions for the latter; the additional effort in producing and transporting feeds was
effectively offset by the increased efficiency of meat production in feedlots.
An advantage of intensive farming it seems is that technological advances to reduce GHG
emissions are often more easily implemented because the animals are housed indoors in
confined areas and there are more opportunities for handling both manure and gas emissions.
Additionally, the high costs of implementing new technologies in an intensive high-input/high-
output system can be justified, whereas a similar increase in costs will turn a low-input/low-
output system into an unprofitable enterprise (Notarnicola et al., 2013). In extensive grazing
systems on the other hand, the sequestration of GHG may balance the generally higher GHG
emissions. Again, conclusions are not clear-cut: the Norden report (2014) finds that especially
beef production by specialised beef breeds is generating large GHG emissions relative to the
amount of human edible food produced. This is particularly due to suckler cows, which
consume large amounts of feed but are producing only one calf per year and no milk for human
consumption. In contrast, dual-purpose breeds (or combi-cows), which produce both milk and
beef, are producing more human food for the same GHG emissions.
Kristensen et al. (2011) found that the production system effects on meat and milk are highly
dependent on the allocation method (between milk and meat) used. In their model, they find
that an average of 15% of total farm GHG emissions was allocated to meat. However,
depending on the method, the amounts allocated to meat range from 13% for economic value,
18% for protein mass, 23% for system expansion and up to 26% for biological allocation. The
allocation method highly influences the GHG emission per kg meat (in Kristensen: 3,41 to 7,33
kg CO2-eq. per kg meat), while the effect on the GHG emission per kg milk was lower (0,90 to
1,10 kg CO2-eq. per kg energy corrected milk). After allocation there was no significant effect
of production system on GHG emission per kg milk.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
42
2.6 Which life cycle stages of meat and dairy have the greatest
impact?
Earlier in this report, in section 1.2.1, we described the status of Norwegian agricultural
emissions. These amount to about 8% of total emissions, we stated, around 4.4 Mt CO2e per
year. This however was based on a fairly loose accounting, and if we include farming in wetland
in these numbers, the total for Norwegian agriculture goes up to about 6.3 Mt, or 12% of
Norway’s total greenhouse gas emissions (Arbeidsgruppe til Landbruk og Klima, 2016). As
stated, the livestock sector in Norway contributes about 90% of this total (Grønlund & Harstad,
2014). At the global level, agriculture contributes about 50% of the global CH4 emissions
(mostly livestock) and about 60% of the global N2O emissions (Crosson et al., 2010).
But how do milk and dairy account, how are their emissions distributed over the different production related
factors, the on- and off-farm processes?
According to Leip et al. (2010), the main emission sources vary between animals, production
systems, countries and climates. Some of the impacts, related to methane, food quality, manure
handling, are already described in section 2.5.2. Analyzing European emissions of the
agricultural sector, Leip et al. find that for beef around 39.6% of its total CO2eq is emitted as
methane, 26% as N2O and 34.4% as CO2. Methane stems primarily from ruminant digestion,
N2O stems from fertiliser use and urine, while 16.5% of the CO2 emissions come from the use
of energy and 17.9% from land use and land use change (although these latter are highly
uncertain numbers, with a wide range). For milk, the distribution of gases is similar, with 36.7%
emitted as methane, 21.3% as N2O and 42% as CO2, from which 17.7% stem from energy use
and 24.3% from land use and land use change. These percentages do include pre-farm processes
such as land-use change, but do not include the after-farm processes, and are thus somewhat
misrepresenting for the overall sources.
Reporting on a UK study by Garnett (2008), Hille et al. (2012) include some of these pre-and
post-farm processes and find that 45 % of the carbon footprint of food consumption in the
UK was allocated to primary production, 5 % to upstream processes, 21 % to processing, 15 %
to distribution and 14 % to trade, including restaurants etc. But even these percentages are not
inclusive of all processes, as the contributions of storage and preparation of food in homes and
of the waste stage have been left out of the total, although Garnett also estimated these - taken
together amounting to about that of trade. Also the upstream processes were not complete, as
only fertiliser production was counted. Other studies such as Weber and Matthews (2008) find
that only 4 % of the carbon footprint of food in the US was due to distribution, and 5 % to
trade. When counting all transport, they found this part contributed 16 % to the life cycle carbon
footprint.
Food waste is rarely included in the accounting, but this has been gaining attention over the
past few years, especially at the consumer level. Werner et al. (2014), in a Danish study, estimate
that around one third of all food produced is not consumed, and the largest share in
industrialized countries of this food waste occurs at consumer level. This would add significantly
to the overall emission of food. However, the amount of waste (at consumer level) differs
substantially for different food groups, with bread and cereals (emissions increasing with around
34% due to waste) and fruits and vegetables (increasing with 23-25% due to waste) topping the
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
43
list. Emissions of beef (and pork and chicken) increase with about 12% due to waste (going
from about 28 to 31.5 CO2eq/kg product). For milk and cheese emissions increase with 7 and
8% respectively (from 1.09 to 1.17 for milk, and from e.g. 9.23 to 9.93 for cheese with 31% fat).
Fish, for comparison, has a 12% increase due to waste, similar to that of meat, while soy drink
emission increase is comparable to milk. Thus, reducing food losses is another improvement
option. Assuming a product loss of 20%, it is found that if meat and dairy product loss is
reduced to 17.5%, the climate effect of milk decreases by 1.75%. Sevenster and de Jong (2008)
highlight that while the most important stage in the milk life cycle up to the farm is enteric
fermentation, followed by feed production, for the total milk life cycle, electricity use due to
household storage is also significant. The IMPRO study (Sevenster and de Jong, 2008) has
calculated that changing the energy efficiency of refrigerators in households could reduce the
climate effect of milk by 1%.
The FAO (www.fao.org/gleam) similarly mentions that energy consumption occurs along the
entire supply chain. Production of fertilisers and the use of machinery for crop management,
harvesting, processing and transport of feed crops generate GHG emissions. Energy is also
spent on animal production site for ventilation, illumination, milking, cooling, etc. Finally,
livestock products are processed, packed and transported to retail points, which involves further
energy use.
The FAO presents disaggregated information on emissions from the four main processes:
enteric fermentation (about 40% of total emissions), manure management (about 10%), feed
production (about 47%) and energy consumption (both on-farm and post-farm: 5%). Figure 4
shows the distribution of emissions along the livestock supply chain at the global level. These
emissions reflect to a great extend the distribution of the emissions of dairy and meat.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
44
Figure 4: Global emissions by source. Relative contribution of main sources of emissions
from global livestock supply chains. Source: www.fao.org/gleam.
For meat and dairy in Norway in particular, Bonesmo et al. (2013) listed the emissions for 30
different farms in Norway, with minimum and maximum emissions for different factors,
ranging from fermentation, soil carbon change, energy use, manure handling, etc. (see figure 5).
They find that there is great variation between farms.
For milk, the maximum emission is 1,7 times higher than the minimum, and while most
emissions are related to fermentation, the greatest variation is found in the N2O emissions from
soil (between 2-39% in young bulls; 8,5-38% in cows and heifers; and 11-40% of the total
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
45
emissions for milk) underlining the importance of correct use of fertiliser, i.e. that purchase
fertiliser should complement livestock fertiliser (Storlien and Harstad, 2015). The difference in
soil carbon change was the next largest variable factor (down to -22% and up to 9,6% in cows
and heifers, while indirect energy use in the production of fertiliser also played a role in the farm
differences. Importantly the enteric fermentation of the animals was not a major variable
between the farms.
Figure 5: Mean, minimum, and maximum values (in percentage) of GHG emission
intensities, expressed as kg CO2eq/kg fat and protein corrected milk (FPCM) and kg
CO2eq/kg carcass weight, for culled cows/heifers and for young bulls based on data from
30 Norwegian dairy farms in 2008. Values less than 0 indicate removal from the
atmosphere (i.e., soil C gain = carbon uptake). Adapted from table 4 in Bonesmo et al.
2013.
-40 -20 020 40 60 80 100 120 140 160 180 200 220
max
min
AVG
Milk
-40 -20 020 40 60 80 100 120 140 160 180 200 220
max
min
AVG
Meat - cows and heifers
-40 -20 020 40 60 80 100 120 140 160 180 200 220
max
min
AVG
% of emissions
Meat - young bulls
Enteric CH4 Manure CH4 N2O
Soil N2O Soil C change
Off-farm barley, CO2eq Off-farm soya, CO2eq
Indirect energy, CO2eq Direct energy, CO2eq
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
46
Gerber et al. (2010) provide a breakdown of emission sources and specify that 93% of total
GHG emissions from milk production globally occur up to the ‘farm gate’. Besides emissions
on the farm, there are many other crucial emissions further down the production and
consumption chain. Sevenster and de Jong (2008) find that post-farm emissions add another
10-20% to cradle-to-farm gate emissions, still excluding household energy use such as cooling,
but including product loss. Assuming an amount of product loss varying between 5 and 20%,
Sevenster and de Jong calculate that product losses are responsible for 57% of the post farm
emissions. Almost 41% of the post farm emissions is due to milk processing (including cheese
and milk powder production). Svenskmjolk finds that post farm emissions are equivalent to
approximately 10% of the emissions up to the farm gate.
The most important post farm stages are market/consumer (36%, mostly fossil fuel use due to
the consumer driving to the shop) and packaging (29%). These estimates for post farm
emissions are under-estimates since storage of milk in the household are left out of the analyses,
while the IMPRO study has shown that household storage has a substantial impact (Sevenster
and de Jong, 2008). Notarnicola et al. (2015) identified some hotspots for dairy products other
than milk. The production of powdered and concentrated milk needed for yogurt production
for example is the main hotspot for the dairy factory phase, mainly due to the high-energy
consumption required for their production processes. Also the production of packaging
materials and energy requirements contributes significantly to yogurt emissions. Finally, the
distribution phase, consumption at the household and final disposal showed a low contribution.
Although the production of milk is the main environmental concern of cheese production,
several authors (references in Notarnicola et al.) focused on the environmental impact of cheese
manufacturing plants, and found that fossil fuel both for energy production and for transport
plays a major role here.
Refsgaard et al. (2011) also compare some of the allocation differences under different
production systems: the largest main contributor to CO2-emission from milk is the direct
emission from husbandry production, i.e. the CH4 and N2O from digestion and manure
contributing with 54% of the total emission at farm gate from conventional production and
71% from organic production. Kristensen et al. (2011) likewise find in their Danish study
comparing emissions under different production systems that organic production has higher
(98%) on-farm emissions than the conventional system (88%).
For Norwegian meat production, Bonesmo et al. (2013) find that the main culprits responsible for
GHG emissions per kg of carcass weight were, in order of relevance: soil’s nitrous oxide
emissions, indirect energy use, soil C loss and enteric methane. Figure 5 shows that there is a
difference between culled (dairy) cows and heifers (DC+H) on the one hand, and young bulls
(YB) on the other hand. Enteric fermentation has the largest emissions in both groups around
38,5-39,5% of the emissions. The variation in YB however is smaller (24-46%) than in DC+H
(23-71). The second biggest source is soil N2O (18% YB 20% DC+H) closely followed by
manure, with about 17-18%, and again more variation DC+H than in YB. Off-farm soy and
barley feed production make up around 16-18%, and direct and indirect energy use 4 to 6%
each, for both YB and DC+H. Finally, the test farms had a net soil carbon uptake (around 3%)
which could be as much as 22% in DC+H and 9% in YB.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
47
Assumptions about electricity: Determining the emissions associated with one unit of electricity
input requires an assumption about the generation mix used to produce that one consumed
unit. While Norway’s generation mix is almost entirely hydropower, through the course of the
year the country both imports and exports electricity to neighbouring countries, such that the
physical electricity at the socket is no longer almost entirely derived from hydropower.
Moreover, because of financial arrangements, certificates of origin are sold by Norwegian power
companies to other countries, such that from an accounting perspective the electricity mix in
Norway contains significant coal generation input. The combination of both physical and
financial trade of electricity across borders adds significant complexity to estimation of
emissions associated with electricity inputs in an LCA. The consequences of these assumptions
are demonstrated in Table 15.
Table 15: Breakdown of GHG emissions (in percent) caused by food consumption in
Norway in 2006. Source: Hille et al. (2012).
Process stage
Assuming Norwegian mix of
electricity*
Assuming European mix of
electricity*
Production of capital goods and
inputs to agriculture and fisheries
17
15
Primary production
57
51
Food processing
9
14
Transport, downstream of primary
production
15
13
Trade in food
2
8
Total
100
100
*Refers to electricity used for processes occurring in Norway
Based on the Norwegian generation mix, almost all of which is hydropower, electricity
contributes very low emissions to the total, and this is most obvious in food processing and the
trade stage of the supply chain, as seen in Table 15. These two stages are high users of electricity
inputs, and because of Norway’s very low emissions per unit of electricity, they contribute very
low emissions. In contrast, when the European electricity-generation mix is assumed, these two
stages have much higher emissions, and become much more significant overall. The trade stage
is no longer negligible (was 2%, now 8%) and emissions in the food processing stage increase
from 9% of the total to 14%. Consequentially, the proportional contributions of all other stages
go down because they have low electricity inputs. This shows the sensitivity to the assumptions
used in the analysis.
Farm size: An interesting point to make here is that preliminary findings from ongoing
(unpublished) research shows that diesel use may be higher on larger farms than on smaller ones
in Norway. This has implications for the emission allocations to direct energy use on the farm,
and will also play through in the relative allocations of emissions to other on-farm and off-farm
processes. There are more differences between small and large livestock holdings than that:
smaller farms are reported to have cattle out for longer times of the year than larger farms,
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
48
which has implications for the amount of feed for these animals that comes from pasture grazing
and thus affects emissions (see e.g. Landbruks- og matdepartementet 2008).
Table 16 summarizes the emissions and allocations to different life-cycle stages for milk and
meat based on some individual papers cited in this section 2.6. The table distinguishes estimates
for the three main stages pre-farm, on-farm and post-farm. Across the studies on cattle (cattle,
milk and meat), the on-farm processes play by far the largest role; in Norway, with around 78%
of the emissions. As the table shows, and many studies have reported previously, for the on-
farm emissions enteric fermentation is the greatest factor, with about 38-40%. Pre-farm stages
contribute about 22%, while fertiliser, manure and pre-farm inputs and indirect energy use play
about an equal large role with between 17 to 22%. Finally, on-farm energy use and soil carbon
storage are only small, with 5% and -4% respectively.
Table 16: Estimates of allocation of total emissions (in percentage of the emissions per kg product)
of meat and milk broken down to different life-cycle stages.
TESCO
2012
Thoma et
al. 2013
Bonesmo et al. 2013
Hille et al.
2012
FAO
Gleam
UK
USA
Norway
Global
Milk
Milk
Milk
Meat,
DC+H
Meat, YB
Food
Cattle
LUC
73
70
6
Inputs, Indirect
energy
22
22
22
5
14
Pre-farm gate ↑
22
22
22
5
20
Enteric CH4
38
38
40
39
Manure CH4+ N2O
18
18
17
26
Fertiliser N2O
21
21
18
8
Direct energy
5
5
4
2
Soil C
-3
-4
-3
LUC
3
Farm production ↑
78
78
76
45
77
Transport
4
Processing
9
7
21
Packaging
3
Distribution
3
5
15
Retail
10
6
14
Consumption
3
5
Recycle and Waste
2
Post-farm gate** ↑
27
30
50
3
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
49
As the table shows, not all studies cover all stages. The allocation and resulting figures for each
stage clearly depends on which factors are included. Therefore, the data for Bonesmo et al.
(2013), which covered until farm-gate, are an overrepresentation of the actual allocations: The
proportion for pre-farm and on-farm emissions would go down if they had added post-farm
emissions. Another major point is that land use change, which in Norway is about 30% of the
agricultural emissions. Nearly all of Norway’s agricultural land is used for livestock in one way
or another, but this often-omitted part in the lifecycle of meat and dairy in Norway plays a major
role. Also, there appears to be considerable inconsistency between the various estimates of soil
carbon fluxes, so where these data have been included their proportional role may vary greatly,
also then affecting the proportional allocation of emissions to the other life cycle processes.
Comparing the allocation of Norwegian emissions related to cattle to the more general
emissions from food (Hille et al. 2012) also highlights some interesting points. Firstly, on farm
emissions are much lower for food in general compared with livestock on-farm emissions. This
makes sense, since there are no methane and nitrous oxide emissions related to non-livestock,
which bring the overall food related emissions for on farm production down. A second point
of interest is the low pre-farm emissions: these again are lower because the much lower
proportion of emissions related to feed imports (for livestock) and related land-use change.
Finally, the proportion of post farm operations are much higher as a result of the lower pre-
and on-farm processes.
Interestingly, at the global level FAO estimate that post-farm activities contribute only 3% to
total emissions, much lower than the contributions from other sources in this table. The most
likely explanation for this is that the developed-country sources are not representative of the
global situation: in developed countries there is considerable energy use in food processing and
wholesale/retail trade, while in most other parts of the world these stages are much more basic
and therefore require very little energy input.
Finally, the table and this discussion about partial life cycle studies highlights the point that to
really understand allocations properly, there is a need for more detailed whole life-cycle LCAs,
for different products and at farm level.
2.7 Potential for change?
Cattle production has been identified as one of the causes of climate change because of cows
emitting methane, which is a GHG with a warming potential at least 25 times that of CO2.
Consequently, a number of strategies that could reduce methane emissions have been the focus
of research. This report did not extensively search for potential to improve (decrease) emissions
from milk and meat production in Norway. However, we came across several studies that
indicated some options, and comparisons between Norwegian/Nordic and other production
systems, and we use this section to bring some important issues to the attention.
In general, life-cycle GHG emissions of vegetable foods are more sensitive to alternative energy
use and efficiencies and transport modes in the supply chain than animal food’s climate impact,
since emissions of methane and nitrous oxide are so significant in milk and meat supply chains
(Norden, 2014). Thus, for milk and meat, emission cuts potentials must be sought elsewhere.
CICERO Report 2016:06
Climate Footprints of Norwegian Dairy and Meat a Synthesis
50
Production: Overall, Norwegian meat and dairy production seems to have a relatively high
environmental impact, compared to other European countries (Roer et al., 2013). In both milk
and meat production, field emissions from forage production and direct emissions from the
animals contribute significantly to the environmental burdens as assessed </