ArticlePDF Available

How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope for to reduction them by 2050

An assessment of greenhouse gas emissions
from the UK food system and the scope for reduction by 2050
Prepared by: Eric Audsley, Cranfield University, UK | Matthew Brander, Ecometrica, UK | Julia Chatterton, Cranfield University, UK
Donal Murphy-Bokern, independent researcher, Germany | Catriona Webster, Ecometrica, UK | Adrian Williams, Cranfield University, UK
January 2010
Cranfield University
Bedford MK43 0AL
Murphy-Bokern Konzepte
Lindenweg 12
49393 Lohne
Kittle Yards
Edinburgh EH9 1PJ
14 January 2010
This study was conducted for WWF-UK and the Food Climate Research Network. We thank
Richard Perkins of WWF-UK and Tara Garnett of the University of Surrey (FCRN) who initiated
the work and provided valuable support. We also thank participants in the project’s conference
on 22 June 2009 who discussed preliminary results. In particular we want to acknowledge the
following for written comments: Dr Havard Prosser, National Assembly for Wales; Dorian Wynne
Davies, National Assembly for Wales; Ian Smith, the Royal Agricultural Society of England; Dr
Doug Parr, Greenpeace UK; Mike Thompson, Secretariat to the Committee on Climate Change;
Keith James, WRAP; and Dr Jeremy Wiltshire, ADAS.
The information presented here has been thoroughly researched and is believed to be accurate
and correct. However, the authors cannot be held legally responsible for any errors. There are
no warranties, expressed or implied, made with respect to the information provided. The authors
will not be liable for any direct, indirect, special, incidental or consequential damages arising out
of the use or inability to use the content of this publication.
© All rights reserved. Reproduction and dissemination of material presented here for
educational or other non-commercial purposes are authorised without any prior written
permission from the copyright holders provided the source is fully acknowledged. Reproduction
of material for resale or other commercial purposes is prohibited.
Please cite this report as follows:
Audsley, E., Brander, M., Chatterton, J., Murphy-Bokern, D., Webster, C., and Williams, A.
(2009). How low can we go? An assessment of greenhouse gas emissions from the UK food
system and the scope to reduce them by 2050. WWF-UK.
To correspond with the research team, please contact Dr Donal Murphy-Bokern
( who serves as corresponding author.
To correspond with the commissioning organisations, please contact Tara Garnett of the FCRN
( or Mark Driscoll of WWF-UK (
Foreword 2
Summary 4
Introduction 8
Study outline 13
Methodology 14
Results 37
Scenarios to achieve a 70% reduction 48
The potential of organic farming in delivering reductions 55
Effects of livestock product substitution on land use 58
Concluding discussion 64
Acronyms and abbreviations 74
Appendix 1: Terms of reference 75
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
In 2008, the Food Climate Research Network (FCRN) published a report
which estimated that
our consumption of food in the UK, from agriculture through to consumption, accounts for 19%
of all the greenhouse gas (GHG) emissions generated through the goods and services we
consume. It also argued that a reduction of up to 70% should be possible if we deployed a mix
of technological improvements and changes in consumption. The report recommended that
government should commit to reducing emissions by this amount, by 2050, and should set out a
road map for how it intends to do so, stating what proportion would be achieved through
technological and managerial improvements and what from changes in the balance of what
people eat.
This recommendation and WWF-UK’s desire to understand what approaches are needed to
reduce GHG emissions from food by 70% provided the impetus for WWF-UK and the FCRN to
join forces in commissioning a piece of work that would: first, re-examine total food chain
emissions taking into account emissions arising from agriculturally induced land use change;
and, second, investigate whether and if so how a 70% reduction in GHG emissions might be
This report, undertaken by a team of researchers from Cranfield University,
Ecometrica and Murphy-Bokern Konzepte, is the result.
We welcome it. This is an innovative piece of work. It has gone a considerable way towards
expanding our understanding of the food chain and its impacts, and of highlighting the actions
that may be needed both pre and post farm gate, both technological and behavioural, if we are
to reduce emissions. By making, as it has had to, a great many fairly major assumptions as to
both impacts (particularly with respect to land use) and as to what solutions might be possible in
the coming years, it has also underlined how much we still don’t know, and need to know.
We would like to draw attention to what we feel are the most striking aspects of this work.
A first key finding of the report is that a focus on one solution only will not lead to the reductions
that are needed. Single measures, such as the elimination of meat and dairy products from our
diet, or the decarbonisation of the supply chain, or the development of technologies to eliminate
enteric methane emissions will not by themselves cut emissions by 70%. If the UK food chain is
to make a proportionate contribution to the UK’s target of reducing its overall emissions by 80%
by 2050, then policy makers will need to put in place a combination of measures that change
not only how we produce and consume food, but also what it is we consume.
A second important finding is that the report corroborates previous estimates, by both the FCRN
and Defra
, of the contribution that food chain emissions (excluding land use change) makes to
UK GHG emissions. They all fall between 152 and 159 Mt CO
e and put the food chain’s
contribution to overall UK consumption related emissions at approximately 20%.
Third is the striking and disturbing finding of this report with respect to land use. This, to our
knowledge, is the first report that actually links changes in land use overseas to the food
consumption patterns of one country. It finds that the inclusion of CO
emissions resulting from
UK food-consumption induced land use change increases food’s footprint by 50% and
increases the contribution made by the food system to overall UK consumption related GHG
impacts to 30%.
Garnett, T. 2008. Cooking up a Storm: Food, greenhouse gas emissions and our changing climate. The Food and Climate
Research Network, Centre for Environmental Strategy, University of Surrey.
The full terms of reference for this research are provided in an appendix.
Defra. 2008. The environment in your pocket.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
The fourth striking conclusion, again one that previous studies have also drawn, is the important
contribution that meat and dairy products make to the overall footprint of the food chain.
Emissions from livestock rearing alone account for over 57% of agricultural emissions.
However, the inclusion of the land use change dimension livestock are also responsible for
more than three quarters of land use change emissions adds even more emphasis to this
Now that this report has been published, what next? We very much hope that others will use
this report as a starting point for further exploration. The report has highlighted the important
contribution played by land use change but clearly much more work needs to be done to
increase our understanding of how these impacts play out both by commodity type and by
agricultural system, as well as what we might need to do about them.
More work needs to be done to examine the trade offs and synergies with other social and
environmental goals, notably with animal welfare and biodiversity. The report suggests that a
lower-meat diet may, for example, have nutritional benefits, and it also looks at the potential
knock-on effects of reduced livestock production from an industrial perspective. In the next
phase of this work, WWF-UK and the FCRN intend to explore the broader social, ethical and
environmental implications of different mitigation scenarios more closely.
Finally, we need to do more work to make change happen. We know enough now to conclude
that the food system contributes very substantially to the problem of climate change. We also
know enough about where and how the impacts arise to start doing something about them.
Business-as-usual, and indeed even business-as-usual lite, are no longer options. We urge
decision makers, in government, the food industry and in the civil society sector to read this
report, and to start thinking urgently about what they intend to do now to create a low GHG,
sustainable food system for ourselves and for our children.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
The overall aim of this study was to develop a set of scenarios that explore how greenhouse
gas emissions from the UK food system may be reduced by 70% by the year 2050. The work is
focused on all emissions from the supply chains and systems, not just the emissions from the
UK food chain that arise in the UK. The study comprises an audit of the greenhouse gas
emissions arising from the UK food economy and an examination of the scope for substantial
reductions of these emissions.
The aim of this short and preliminary study conducted over a few months in 2009 is to stimulate
debate about the full GHG impact of the UK food chain and the scope and options for reducing
GHG emissions in line with wider climate change policy. The study is theoretical, in effect a
thought experiment based on detailed inventories of emissions and the use of life cycle
assessment (LCA) to examine the effects of measures. As far as we are aware, this is the first
study to identify systematically the proportion of global land use change attributable to
commercial agriculture linked to international trade. From this it estimates a proportion of global
land use change emissions attributable to the UK food supply chain.
In considering this report, especially the scenarios for reductions, it is important to appreciate
that we are not presenting a model or components of a model for working out the full effect of
policy choices. This report identifies the size and sources of present emissions and identifies
scenarios from these for reductions. Our scenarios set out possible directions of travel but we
emphasise that the full real-world effect of greenhouse gas mitigation strategies will depend on
the consequences of complex interactions that cannot be predicted here. Measures may open
up opportunities for synergies in specific circumstances that will be revealed in the path to a low
carbon food system giving additional benefits. Similarly, there are also risks that some
measures may trigger economic responses with unintended consequences – for example a
reduction in demand for ruminant products may cause the widespread abandonment of UK
grazing land leading to increased imports from sources closer to active land use change.
Our estimates are based on the current UK population. This is expected to increase
substantially by 2050. There will be a corresponding increase in food system emissions as the
food economy grows. But from a global perspective, this is a growth in GHG emissions that will
occur somewhere as the global population expands. By working on the basis of food system
emissions in 2005, we have avoided confusion between the effectiveness of measures and
trends in population. We also want to emphasise that our study is about the food system and
therefore does not consider other agricultural land uses – for example for biofuels. However, our
findings are applicable to the assessment of other uses of agricultural products.
Our main results are as follows:
Using a detailed inventory of emissions developed from LCA of a wide range of foods and
processes, we estimate that the supply of food and drink for the UK results in a direct emission
equivalent of 152 Mt CO
. A further 101 Mt CO
e from land use change is attributable to UK
food. Total UK consumption emissions are estimated to be about 748 Mt CO
e (excluding land
use change).
This means that direct emissions from the UK food system are about 20% of the
currently estimated consumption emissions. When our estimate of land use change emissions is
added to these, this rises to 30%.
In our work, we refer to direct emissions (excluding land use change emissions) as ‘supply
chain emissions’. Of these, about 58% arise from animal products which account for just over
30% of consumer energy intake. Two thirds of food production emissions arise in the UK, 16%
Garnett, T. 2008. Cooking up a storm. Food, greenhouse gas emissions and our changing climate. The Food and Climate
Research Network.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
arise outside Europe. Overall, about one fifth of direct UK food chain emissions occur outside
the UK. If land use change emissions are taken into account, then about a half of total food
system emissions arise outside the UK. So our results indicate that the food system in particular
presents special challenges for climate change policy focused on domestic emissions and
Taking the food chain as a whole, the supply chain emissions comprise (on a CO
basis) CO
– 102 Mt, CH
– 23 Mt, N
O 21 Mt and refrigerants – 6 Mt. Fifty-six per cent of
emissions arise from primary production (mainly farming) with CH
and N
O accounting for more
than half of these.
Land use change (mainly deforestation) driven by agricultural expansion is a hugely important
source of emissions attributable to the global food system. The UK food system is part of the
global food system contributing to the underlying forces. We estimate that global land use
change emissions account for 40% of the emissions embedded in UK consumed food and 12%
of emissions embedded in all UK consumption overall. This is based on the allocation of 2.1%
of global land use change emissions to the UK food supply chain. This estimate is based on
global average yields and land use. Managed and native grassland covers more land than
arable crops. As a result, a large proportion (around three quarters) of LUC emissions is
allocated to ruminant meat. We used alternative ways of allocating emissions which increase
allocations to crops and reduce allocations to pasture, for example by allocating according to
the economic value of crop and livestock farm outputs. This reduced emissions from beef and
sheep/goat meat production from 77 Mt CO
e to 42 Mt CO
e out of a total of 102 and 86 Mt
e respectively. So while allocation on economic value reduces the emissions attributable to
beef and sheep meat, we are confident that the broad conclusions remain across the various
allocation methods that could be used.
By assessing and attributing a proportion of land use change emissions to agricultural land use
generally, our analysis draws attention to how consumers share responsibility directly or
indirectly for the drivers behind land use change. We work on the premise that commodity
markets are highly connected. Our analysis could lead to the conclusion that transferring
consumption away from products directly linked to land use change to products from
established farmland through product certification may displace rather than reduce the
underlying pressures. This highlights the need for demand/market based approaches (e.g.
product certification and moratoria) that counter the economic forces driving land use change,
complementing ‘top-down’ government measures that seek to stop deforestation directly.
The supply chain measures we examined to achieve a 70% reduction in supply chain emissions
range from the decarbonisation of energy carriers used in food production and measures to
increase farm efficiency to technologies to reduce emissions of methane. Our results confirm
that significant reductions will involve radical structural change throughout the supply chain from
the generation of electricity through to the preparation of food. No single measure or the
combination of similar measures is capable of reducing emissions by more than about half. The
decarbonisation of the wider economy sought now by government policy by 2050 will reduce
food supply chain emissions by about 50%.
A vegetarian diet (with dairy and eggs), a 66% reduction in livestock product consumption, and
the adoption of technology to reduce nitrous oxide emissions from soils and methane from
ruminants are measures that each have the potential to reduce direct supply chain emissions by
15-20%. Modifying consumption has a particularly important role to play and consumption
measures offer opportunities for reductions that could be implemented in the near future. In
addition, consumption measures align with other public policies, particularly health. A switch
from red to white meat will reduce supply chain emissions by 9% but this would increase our
reliance on imported soy meal substantially. Our analysis indicates that the effect of a reduction
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
in livestock product consumption on arable land use (which is a critical component of the link
with deforestation) will depend on how consumers compensate for lower intakes of meat, eggs
and dairy products. A switch from beef and milk to highly refined livestock product analogues
such as tofu and Quorn could actually increase the quantity of arable land needed to supply the
UK. In contrast, a broad-based switch to plant based products through simply increasing the
intake of cereals and vegetables is more sustainable. We estimate that a 50% reduction in
livestock production consumption would release about 1.6 Mha of arable land (based on the
yield of crops supplying the UK) used for livestock feed production. This would be offset by an
increase of about 1.0 Mha in arable land needed for direct crop consumption (based on UK
yields). In addition to the release of arable land, between 5 and 10 Mha of permanent grassland
would be available for extensification, other uses, or re-wilding. Such changes would open up
‘game-changing’ opportunities but there needs to be careful assessment made in the
development policy if unintended consequences are to be avoided. A contraction in the
livestock sector that might follow a significant change in consumption could trigger a collapse of
livestock production in the UK. The consequences for the emissions from the UK food chain
would then depend on developments elsewhere. Completely unregulated, such a collapse could
reinforce expansion in low cost exporting countries, even adding to forces driving land use
Our examination of measures that raise production and nitrogen use efficiency indicates that
this approach has the potential for savings that are less than consumption based measures.
This is supported by the scientific literature. However we acknowledge and set out evidence
from elsewhere that this too has an important role to play. We anticipate too that there are
potential synergies between production efficiency measures and consumption measures that
we have not been able to simulate – for example a reduction in livestock product consumption
may synergise with efforts to raise the efficiency of nitrogen use in the food system. There are
also possible synergies between efforts to raise production efficiency and the use of
technologies to reduce emissions directly. Consumption based measures would mean a
significant contraction in livestock production for UK consumption and this opens up
opportunities to restructure agriculture in a way that enhances the benefits of production
efficiency measures. In addition, from a global perspective, reductions in livestock consumption
and measures to increase production efficiency synergise with efforts to eliminate deforestation.
Improving production efficiency and reducing production emissions directly will mean embracing
new technologies. These need to be carefully applied to whole systems to raise system eco-
efficiency. Our analysis indicates there is little scope for emission reductions through the
exclusion of production technologies – for example through the widespread adoption of organic
farming. We estimate from analysis of recently published work that a complete conversion to
organic farming in the UK with corresponding changes in diet would reduce supply chain
emissions by about 5%.
Emissions from fish consumption were quantified, but expansion in fish production to replace
other livestock products was not considered owing to concerns about the sustainability of wild
fish stocks. This though has significant potential depending on the success of developing new
aquaculture systems.
Very significant change in the food system is required to achieve a 70% reduction in supply
chain emissions. The consumption and farm technology changes align with other policy
objectives, for example public health, nitrate emissions, ammonia emissions and biodiversity.
The scenarios set out here do not have definitive implications for animal welfare outcomes in
one direction or another. The reduction in animal products consumption generally as set out in
consumption measures opens up opportunities to improve welfare. However, measures to
increase production efficiency at the animal level raise questions about the welfare
consequences. This underscores the importance of whole system analyses and an emphasis
on whole system solutions rather than just on interventions at the individual animal level.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Our results also show that a 70% reduction in supply chain emissions (i.e. excluding land use
impact) may be possible without significant changes in consumption. However, if repeated
across the developed and developing world, such a high level of livestock product consumption
would require a large expansion in global agriculture and would make contraction and
convergence of emissions difficult. Per-capita UK meat consumption is more than twice the
world average, and nearly three times that of developing countries. As the global food system
becomes more resource constrained and developing countries lift themselves out of poverty,
consumption based measures will acquire relevance beyond just the UK’s greenhouse gas
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
This study examines the feasibility of achieving a significant reduction (possibly 70%) in
greenhouse gas (GHG) emissions from the UK food system by 2050. The work is consumption
based. It relates UK consumption to all direct and indirect emissions from the supply of food for
UK consumption, both in the UK and overseas. The study comprises an audit of the greenhouse
gas emissions arising from the UK food economy and an examination of the scope for
substantial reductions of these emissions in this timeframe. The overall aim was to develop a
set of scenarios that explore how greenhouse gas emissions from the UK food system may be
reduced by 70% by the year 2050.
To achieve this, two broad objectives were addressed:
1. To compile a complete inventory of all UK food consumption from domestic production and
imports, distribution and consumption, including direct greenhouse gas (GHG) emissions
related to primary and post-primary production and indirect emissions resulting from Land
Use and Land Use Change (LULUC) associated with this production.
2. To develop and assess a set of scenarios to reduce these emissions by 70% based on
measures from both production and consumption systems by 2050.
The work was prompted by the suggestion from the FCRN that the UK government should
commit to achieving a 70% or more absolute reduction in food-related GHG emissions by
Since then, the UK Climate Change Act 2008 which aims to improve carbon
management and support the transition towards a low carbon economy in the UK has been
enforced. It seeks to demonstrate strong UK leadership internationally, with a commitment to
share of responsibility for reducing global emissions globally. Targets include an 80% reduction
in UK greenhouse gas emissions through action in the UK and abroad by 2050, and reductions
in CO
emissions of at least 26% by 2020, against a 1990 baseline. The 80% target translates
into a 77% reduction in relation to 2005. This research examines in outline if and how changes
to the UK food system can make a significant contribution to this target. It also identifies the
relevance of this domestic target focused on emissions from the UK to the emissions arising
from the wider UK food system.
Worldwide, agriculture and related up-stream activities such as fertiliser manufacture plus land
use change are responsible for about a third of the world’s greenhouse gas emissions (Figure
1). In primary agricultural production, the profile and underlying causes of GHG emissions is
different to most other sectors. N
O from the nitrogen cycle dominates direct greenhouse gas
emissions from crops in terms of global warming potential, accounting for about 70% of the
GHG emission from wheat production for example. In addition, methane from livestock
production, particularly from cattle and sheep, is a potent global warming gas emission.
Methane and nitrous oxide emissions have risen in a pattern similar to CO
. Agriculture’s role in
carbon dioxide emissions arises mainly from land use change rather than fossil fuel use.
UK greenhouse gas inventories indicate that 7% of UK emissions are attributable to UK
made up of the equivalent of 51 Mt of CO
e as carbon dioxide (11%), methane
(37%) and nitrous oxide (53%). This is only a small proportion of total emissions attributable to
the food system. There are also emissions from the manufacture of farm inputs, food
processing, distribution, retailing and preparation. The manufacture of nitrogen fertilisers
Garnett, T. 2008. Cooking up a storm. Food, greenhouse gas emissions and our changing climate. The Food and Climate
Research Network.
HM Government. 2006. Climate change, the UK programme.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
(registered in GHG inventories as an industrial emission) is the most important cause of direct
emissions upstream of agriculture. About 900,000 tonnes of nitrogen as fertiliser is used in UK
agriculture each year. Assuming 80% is ammonium nitrate and 20% is urea
, the manufacture
of this fertiliser emits the equivalent of 6 Mt of carbon dioxide, the equivalent of about 1% of the
GHG emissions in the UK.
The UK is a net importer of many foods and emissions from the production of imports are not
reflected in UK inventories. Previous analyses indicate that overall, UK agriculture, fertiliser
production, and livestock agriculture in near-neighbouring countries for export to the UK is
responsible for the emission of about 62 Mt carbon dioxide per year, equivalent of 10% of
emissions attributed to the UK in inventories. Livestock products represent the majority of
imports from these nearby counties. Their production, especially of poultry and pig meat, is
similar in LCA terms to that of the UK. So drawing on UK LCA data
, it is estimated that the
production of these imported livestock commodities emits the equivalent of about 3.7 Mt carbon
dioxide on a life cycle basis up to the farm gate. Land use change in other countries is also
excluded from national emissions inventories. So it can be concluded that the role of the UK
food system in global greenhouse gas emissions is far greater than that indicated by UK
emissions attributable to UK agriculture.
Figure 1. Flow of global greenhouse gas emissions
A number of studies have made estimates of the wider emissions from the food system. The
University of Surrey based Food Climate Research Network reports that the UK food chain
Williams, A, Audsley, E and D Sandars. 2006. Determining the environmental burdens and resource use in the production of
agricultural and horticultural commodities. Defra project report IS0205.
Williams, A, Audsley, E and D Sandars. 2006. Determining the environmental burdens and resource use in the production of
agricultural and horticultural commodities. Defra project report IS0205.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
(production, processing and retail) accounts for 19% of UK consumption GHG emissions, i.e.
the equivalent of 159 Mt of carbon dioxide
The UK Cabinet Office
reports 18% with just under half attributed to UK farming and fishing.
For Western Europe as a whole, the EU Environmental Impact of Products (EIPRO) study
identified food as responsible for 20-30% for most categories of environmental burdens,
including greenhouse gas emissions. For greenhouse gas emissions, this 20-30% attributable
to food comprises 4-12% for meat, 2-4% for dairy products, and about 1% for cereal products.
So livestock products account for 6-16% of greenhouse emissions attributable to Europe. An
equivalent estimate for the world is 18%.
In addition to direct emissions from the food chain, there is also the UK’s share of indirect
emissions due to land use change, e.g. deforestation, which in total are estimated to account for
18% of global emissions. Land use change emissions attributable to the UK food economy have
not been estimated prior to this study, but even 1% (reflecting the UK population as a proportion
of the global population) of the 7,300 Mt of CO
e due to deforestation globally is very significant
(73 Mt CO
e). Overall, it is clear that the delivery of food up to the point of consumption is
significant: food is comparable to transport and domestic energy consumption in terms of its role
in personal carbon footprints.
Greenhouse gas emissions from UK agricultural production have fallen since 1990.
It is
difficult to assess trends in greenhouse gas emissions for the food economy as a whole as they
are the result of a number of counteracting and poorly understood activities – for example rising
commodity consumption is counteracted by increased production efficiency in Europe, and
increased energy efficiency in manufacturing is counteracted by increased car use in shopping.
Overall, further but modest reductions in emissions from primary production are expected up
until 2010.
Due to the intrinsic connection with biological processes causing emissions of
nitrous oxide and methane, step-changes in emissions are more difficult to achieve compared
with, for example, the electricity sector. Against this background, Defra expects UK agricultural
emissions to rise by 6.5% between 2010 and 2020 although the government’s low carbon
transition plan anticipates a 6% reduction in agricultural emissions on 2008 levels by 2050.
Life cycle assessments such as those set out in the Cranfield study
consistently reveal the
large burdens associated with the production of livestock commodities. Livestock are estimated
to account for 70% of agricultural land use worldwide (30% of the Earth’s land surface) and
more than half of the greenhouse gas emissions attributable to agriculture.
Reducing livestock
production would reduce emissions directly through reductions in methane from ruminants and
waste management, and nitrous oxide from forage and feed production. Indirect reductions
would result from reduced nitrogen related enrichment of habitats, from nitrate leaching and
ammonia emissions. The biggest effect for the environment may be through the indirect effects
Garnett, T. 2008. Cooking up a storm. Food, greenhouse gas emissions and our changing climate. The Food and Climate
Research Network.
Cabinet Office. 2008. Food matters. Towards a strategy for the 21st century. The Cabinet Office Strategy Unit, UK.
Tukker, A, Huppes, G, Guinée, J, Heijungs, R, de Koning, A, van Oers, L, Suh, S, Geerken, T, Van Holderbeke, M, Jansen, B and
P Nielsen. 2006. Environmental Impact of Products (EIPRO). Analysis of the life cycle environmental impacts related to the final
consumption of the EU-25. Main report IPTS/ESTO project.
Steinfeld, H, Gerber, P, Wassenaar, T, Castel, V, Rosales, M and C de Hann. 2006. Livestock’s long shadow. FAO.
HM Government. 2006. Climate change, the UK programme.
Defra. 2008. The UK climate change programme.
HM Government. 2009. The UK low carbon transition plan.
Williams, A, Audsley, E and D Sandars. 2006. Determining the environmental burdens and resource use in the production of
agricultural and horticultural commodities. Defra project report IS0205.
Steinfeld, H, Gerber, P, Wassenaar, T, Castel, V, Rosales, M and C de Hann. 2006. Livestock’s long shadow. FAO.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
of livestock on land use change where the production of crops for the livestock sector is a factor
driving deforestation.
An estimated 18% of global GHG emissions arise from land use change and forestry (Figure 1).
These estimates are uncertain and emission estimates range from 2,899 Mt of carbon dioxide to
8,601 Mt (20% of carbon dioxide emissions).
Deforestation is by far the largest component of
land use change emissions (Figure 2). Drawing on FAO statistics
, 58% of the deforestation is
driven by commercial agriculture. The role of agriculture as a driver can be complex with
interaction with other drivers such as road building, logging and population growth. Accepting
the uncertainty in estimates and drivers, it remains clear that land use change is connected to
agriculture and this is a significant cause of emissions attributable to the global food economy. It
is worth noting, for course, that deforestation of the UK to supply agricultural land has taken
place over millennia and much reforestation occurred in the 20th century. The associated CO
emissions from this historical deforestation have long been assimilated into the Earth’s
Figure 2. Sources of emissions from global land use change 2000
Most public debate about food and deforestation is focused in direct links between land use
change and the UK food system. Considering the dominance of the tropics in land use change
(Figure 3), this focuses attention on produce from these regions, particularly soy and beef from
South America and palm oil from South-east Asia. This approach to the problem regards
deforestation as attributable to UK food consumption when UK consumed food is grown on
recently converted land. For example, if the UK consumes palm oil and a proportion of this
demand is met by converting forest to palm oil plantations, the emissions from the conversion of
Ramankutty, N, Gibbs HK, Achard, F, Defries, R, Foley, JA and RA Houghton. 2007. Challenges to estimating carbon emissions
from tropical deforestation. Global Change Biology, 13, 51–66.
FAO. 2007. State of the world’s forests.
Baumert, KA, Herzog, T and J Pershing. 2005. Navigating the numbers: Greenhouse gas data and international climate change
policy. World Resources Institute.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
forest land to plantation are allocated to the palm oil produced on that land. However, it is
possible that switching consumption to foods which are grown on existing agricultural land (to
reduce direct land use change) will displace the production on that land to other areas, some of
which will be converted from other land use types (causing indirect land use change). Therefore
there are direct connections to land use change, and there are indirect connections via global
commodity trading.
Figure 3. Locations of net deforestation
This study accepts that the global food system is highly connected and indirect effects must be
considered. In this, the boundary between agricultural land and other land use can be regarded
as a frontier. As the global demand for food or other agricultural products increases, global
agricultural output expands. Over the last 50 years, much of this production expansion has been
achieved through increases in yield rather than area. However, the relative growth in yields has
declined steadily and is now lower than the growth in population. This is a strong pointer
towards increased pressure on land use change.
Figure 4. The rate of growth in the world’s population is now greater than the rate of growth in
crop yields (FAOSTATata by Dr Stephan Bringezu (Wuppertal Institute, Germany))
FAO. 2005. Global forest resources assessment: progress towards sustainable forest management.
Compiled from FAOSTAT data by Dr Stephan Bringezu. Wuppertal Institute, Germany.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Based on an analysis of an inventory of greenhouse gas emissions attributable to the UK food
system on a life cycle basis (including emissions from land use and land use change), this
research developed food system scenarios integrating production and consumption mitigation
Research on mitigation necessarily examines component emissions in detail leading to
identification of individual opportunities for change and incremental progress. This work takes a
radical approach in focusing on the effects of a combination of step-changes to reduce
greenhouse gas emissions attributable to the UK food system in line with the target for the UK
as a whole. It does this by looking at combinations of step-changes in the consumption, trade,
processing and production of food.
The study comprised three phases integrated as shown in Figure 5. Phase 1 addressed the
question of the size and sources of emissions from the UK food system currently. These
comprise emissions from four categories: the production of the food commodities (primary
production), emissions from processing, distribution, retail and preparation (post farm gate
emissions), land use emissions, and land use change emissions. Phase 2 looked at the
mitigation potential of specific production and consumption measures. Phase 3 of the study
examined how these may be radically reduced over the next 40 years in line with current targets
for the UK as a whole.
UK and
regional food
system GHG
Communications and research delivery, interface with Phase 2, sup port to policy development
Data on
•Primary production
•Manufacture, distribution
•Land-use change
Analyses of commodity
flows and
Food system
System scenario
‘Socolow’s wedges
Part 1
Part 2
Part 3
Figure 5. Project overview
The research took an LCA based approach to estimate direct emissions from the food chain.
This was augmented by estimates of emissions arising from land use change to provide
estimates of all emissions attributable to the UK food system, including emissions arising from
imports, net of exports. The allocation of global land use change emissions to the UK food
economy was a particular focus.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
The foundation of our work is an inventory of emissions from the supply of food for UK
consumption. This comprises emissions from primary production (farming and fishing),
processing, distribution and retail, consumption, and land use change attributable to the UK
food system. We have based our analysis on data for UK food commodity consumption. We
were unable to find reliable data for palm oil used in food, so we have included palm oil used for
the oleochemical industry drawing on trade data. In addition, there are well developed synergies
between a range of non-food industries, pet food and human food production. There are also
about 300,000 horses in the UK, of which a good proportion will receive some concentrates that
may not have been accounted for. So while we have used data on food consumption, there may
have been some over-accounting of items in the food sector as a result of connections with non-
food uses, but we are confident that this is small compared with food.
The work conducted here is based on the 2005 population of 60.5 million. It will undoubtedly
have changed by 2050, but no one can say exactly by how much. The current UK forecast from
the ONS
is for the population to increase to 77 million, an increase of 27%. Population
forecasting is difficult because of immigration and emigration. The analyses were all calculated
on the basis of a constant population. We felt that this gives a sufficient clarity in understanding
the directions needed to achieve major reductions in emissions.
Methodology for the inventory of emissions from primary agricultural production
The work was based on a detailed analysis of commodity consumption, production and trade
data from the UK Department for Environment, Food and Rural Affairs (Defra), the United
States Department of Agriculture (USDA) and the United Nations Food and Agriculture
Organisation (FAO). Unless otherwise stated, all data on commodity flows come from the
FAOSTAT data for 2005, accessed in early 2008. This provided a full list of crop and animal
commodities and their quantities entering the UK food system for final consumption. It includes
food and drink. Table 1 presents the full list of data for food commodities entering the UK food
system in 2005.
These data were used to compile an inventory of emissions from primary production
agriculture, fishing and fish farming. They were also used for the calculation of LULUC
emissions. A separate source of data was used for the processing and distribution phases. UK
imports of major temperate arable crop and livestock commodities are dominated by supplies
from near neighbours. The data relate to primary commodities, that is, products such as olive oil
are set out in terms of tonnes of olives, beer and whisky as barley, wine as grapes etc.
Table 1. Net UK imports, production and consumption of food commodities (2005).
Consumption is human consumption only – excluding crop commodities used for animal feed.
Data on consumption are independent of data on production and imports and so do not align
arithmetically. Data is per thousand tonnes
Commodity Net Import
Almonds 27
Misc. meat 8
Anise, badian, fennel etc. 8
Milk 2013
Apples 754
Millet 17
Apricots 70
Mushrooms and truffles 131
Artichokes 1
Natural honey 27
Asparagus 7
Nutmeg, mace etc. 1
Avocados 40
Misc. nuts 23
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Bananas 702
Oats -28
Barley -1176
Misc. oilseeds 43
Green beans 6
Olives 438
Dried beans, cowpeas 123
Onions (inc. shallots) 322
Bird eggs 76
Oranges 1018
Bovine meat 260
Other melons 158
Broad and horse beans -160
Palm oil
Brassicas 49
Papayas 8
Carrots and turnips 52
Peaches and nectarines
Cashew nuts 28
Pears and quinces 238
Cassava (fresh and dried)
Peas, dry 2
Cauliflowers and broccoli 124
Peas, green 10
Misc. cereals 302
Pepper (Piper spp.) 6
Cherries 26
Pig meat 554
Chestnuts 2
Pineapples 361
Chickpeas 18
Pistachios 6
Chillies and peppers, dry 8
Plantains 16
Chillies and peppers 139
Plums and sloes 116
Cinnamon (canella) 1
Potatoes 973
Misc. citrus fruit 39
Chicken meat 317
Cocoa beans 363
Misc. pulses -133
Coconuts (incl. copra) 154
Pumpkins, squashes 36
Coffee, green 135
Rabbit meat 0
Cottonseed 10
Rape- and mustard seed
Cranberries, blueberries 5
Raspberries etc. 8
Cucumbers and gherkins 123
Rice, paddy 602
Currants, gooseberries 12
Rye -1
Dates 17
Sesame seed 14
Duck, goose, guinea fowl 4
Sheep and goat meat 34
Edible offal 64
Sorghum 6
Eggplants (aubergines) 16
Misc. spices 9
Figs 11
Spinach 8
Misc. fruit 63
Misc. starchy roots 15
Garlic 11
Strawberries 51
Ginger 13
Sugar beet -2075
Grapefruit and pomelo 170
Sugar cane 8532
Sunflower seed 382
Groundnuts 253
Sweet potatoes 20
Guavas, mangoes etc. 62
Tangerines etc. 348
Hazelnuts 9
Tea and Maté 125
Kiwi fruit 35
Tomatoes 1305
Leeks etc. 15
Turkey meat -17
Misc. leguminous veg. 0
Misc. vegetables 3188
Lemons and limes 136
Walnuts 13
Lentils 18
Watermelons 40
Lettuce and chicory 167
Wheat -1049
Linseed -34
Yams 6
Maize 1336
Soy oil***
Based on FAOSTAT trade data received in January 2008 including palm oil for non-food uses.
Includes grapes as wine.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Table 2. Commodities and countries of production included in the Defra-funded project FO0103
(“Comparative LCA”)
Commodity Alternative Country to the UK
Beef Brazil
Chicken meat Brazil
Lamb New Zealand
Strawberries Spain
Tomatoes Spain
Potatoes Israel
Apples New Zealand
Results of the research at Cranfield
were used to estimate emissions from the production of
major commodities, including the production of animal feedstuffs. The Cranfield data resource
includes 10 main commodities: bread wheat, potatoes, oilseed rape, beef, pig meat, lamb,
poultry meat, eggs, milk and tomatoes. Use of these results also avoided double counting, as
emissions relating to livestock feed production (feed wheat, barley, beans, maize, soya, forage
maize and grassland) are included in the livestock LCA figures. Further results for domestic and
overseas production (and delivery to the RDC) were obtained from the Defra-funded project
This included comparative burdens of seven food commodities (Table 2). This was
supplemented by reports from the literature
for other commodities that are not included in the
Cranfield work. Where no data were found, proxy values were used and rational adaptations
were made to the model. For example, all tree fruits, except for oranges, were assumed to be
apples. Transport adjustments were made when needed to allow for imports.
Regional consumption data was also obtained, and the UK inventory was divided into datasets
for each individual country, to enable analysis of regional differences to be considered.
Primary production is defined as all activities and emissions arising from commodity production
up to and including arrival at the regional distribution centre (RDC). For most items, this was as
raw commodities, although some processing was included for a few items and is discussed
later. Post-primary production includes activities such as processing, distribution to retail, retail
itself, cooking and waste disposal. The parallel systems in the food service sector were also
Further data were obtained from the Defra-funded project FO0404
that was led by ADAS and
assessed the applicability of PAS 2050 for agriculture and food. It provided data on apples,
onions, pineapples, tea, coffee and cocoa. The scientific literature was also searched and other
sources were identified and used. Care was needed in using other data, e.g. Carlsson-
included a value for rice which was strikingly high, but this was partly because she
used a 20 year horizon for the GWP of methane, which is a large emitting term. Converting this
to a 100-year time horizon reduced this portion of the burden about threefold (Table 3), although
the value per tonne of rice is still appreciably higher than other cereals.
Williams, A, Audsley, E and D Sandars. 2006. Determining the environmental burdens and resource use in the production of
agricultural and horticultural commodities. Defra project report IS0205, as further developed under Defra project IS0222.
Williams, AG, Pell, E, Webb, J, Tribe, E, Evans, D, Moorhouse, E and P Watkiss. 2009. Comparative life cycle assessment of
food commodities procured for UK consumption through a diversity of supply chains. Final Report to Defra on Project FO0103.
See reference list at end of this report for full list.
Defra project FO0404. Scenario building to test and inform the development of a BSI method for assessing GHG emissions from
Carlsson-Kanyama, A. 1998. Climate change and dietary choices — how can emissions of greenhouse gases from food
consumption be reduced? Food Policy, 23 (3/4), 277–293.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Due to the varied nature and detail of results, values were adjusted to fit the context of the study
so that they were comparable with other values in the inventory in terms of scope, boundary
conditions and functional units. For example, all values were adjusted to include transport up to
the Regional Distribution Centre (RDC). Data were adjusted too so that the functional unit was a
tonne of commodity production in most cases.
Furthermore, for some commodities there were no complete Life Cycle Assessment (LCA)
studies. Appropriate proxy values were chosen and adapted. For example pineapples were
used as a substitute for most other exotic fruit including bananas; oats were taken as an
average of spring and winter barley; strawberries used for other soft fruits etc. All values were
converted into Global Warming Potential (GWP) on a 100 year time horizon, kg CO
e) per tonne of commodity production, using the IPCC conversion factors.
Table 3. Global warming potentials of gases over 20, 100 and 500 year timescales
Gas, kg GWP
, kg CO
e. GWP
, kg CO
e. GWP
, kg CO
Carbon dioxide, (CO
) 1 1 1
Methane (CH
) 72 25 7.6
Nitrous oxide (N
O) 289 298 153
Total UK primary production emissions were obtained by multiplying the total consumption of
each raw commodity by its burdens per tonne of production including the transport to the RDC.
Methodology for the inventory of emissions from fishing and fish farming
Fish may be caught or farmed and vertebrates or invertebrates (shellfish). Vertebrates are
divided mainly into demersal (bottom feeders) and pelagic. Our wild fish consumption is still
dominated by demersal white fish like cod and haddock as well as tuna (pelagic). Vertebrate
fish farming is dominated by salmon and trout and invertebrates by mussels, with some
production of langoustines. Shellfish are also imported from overseas (as far away as the Far
For caught fish, the main burden is the energy used in fishing, including refrigeration.
feeding stage dominates farmed fish production.
While energy consumption and GHG
emissions are closely related to each other, it must be noted that the environmental impacts of
fishing and fish farming are more diverse and complex than these alone.
Apart from resource
use and emissions to the environment, there are major problems about fish stocks, wastage
from the returns to sea of undersized fish etc. Most UK fish farming includes fish meal from wild
caught fish, so that expansion of domestic production is limited by the availability of the wild fish
supply. For these reasons, we did not include any scenarios about increasing fish consumption.
Forster, P and Ramaswamy, V. 2007. Changes in atmospheric constituents and in radiative forcing, in: IPCC AR4 WG1, Report
climate change 2007, The physical science basis. (overall web
Tyedmers, PH, Watson, R and D Pauly. 2005. Fueling global fishing fleets. Ambio, 34 (8), 635–38.
Thrane, M. 2006. Energy consumption in the Danish fishery. Identification of key factors. Journal of Industrial Ecology, 8 (1–2),
Ellingsen, H and Aanondsen, A. 2006. Environmental impacts of wild caught cod and farmed salmon - A comparison with chicken.
International Journal of Life Cycle Assessment, 11, 60–65.
Papatryphon, E, Petit, J, Van der Werf, HMG and SJ Kaushik. 2007. Life cycle assessment of trout farming in France: A farm level
approach. Proceedings 5th international conference on LCA in foods, 25–26 April 2007, 71–77, Gothenburg, Sweden.
Cappell, R, Wright, S and F Nimmo. 2007. Sustainable production and consumption of fish and shellfish. Environmental impact
analysis. Final report to Defra from Royal Haskoning. Project code 9S6182.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Fish production and consumption are not as well defined in data as other foods are.
simplified the data on consumption to allow us to use what data there are from fishing and fish
farming. Data resources are dominated by results from Scandinavian research. Part of the
problem is the yield is not always clear whether weights refer to gross weight or net weight after
filleting or removing shells etc. The best LCA studies include all stages through to the retail, but
these do not cover all fish types. The simplified consumption data set that we derived is
provided in Table 4. It should be noted that these data relate to fish as purchased by
consumers. Weight loss in the fish supply chain is high and so the quantities of commodity fish
used are higher than the quantities shown here.
Table 4. Simplified data for UK fish consumption and specific emissions of GHG
Fish type Gross wt,
Wastage rates
Net wt,
kt Gross Net
emissions kt
Salmon 163 25% 122 3.0 366
Trout 9 25% 6 4.5 29
Imports – long
Tuna 98 30% 73 1.9 2.6 194
Shellfish 111 25% 83 5.7 7.6 633
UK and imports
from EU
Wet fish (by
difference) as cod 276 25% 207 6.6 1370
Shellfish 37 25% 28 5.1 6.9 190
Total fish 692 519 5.4 2781
Data sources:
Methodology for the inventory of emissions from processing, distribution, retailing and
The RDC (Regional Distribution Centre) is a nominal boundary in our reporting. We adopted this
boundary because data sources vary in detail about end points and because of the imprecision
in judging where primary production ends and processing begins. In some cases, the
manufacturing, processing and packaging has been included in the pre-RDC side, e.g. liquid
Ellingsen, H, Olaussen, JO and IB Utne. 2009. Environmental analysis of the Norwegian fishery
Papatryphon, E, Petit, J, Van der Werf, HMG and SJ Kaushik. 2007. Life cycle assessment of trout farming in France: A farm level
approach. Proceedings 5th international conference on LCA in foods, 25–26 April 2007, 71–77. Gothenburg, Sweden.
Hospido, A and Tyedmers, P. 2005. Life cycle environmental impacts of Spanish tuna fisheries. Fisheries Research, 76, 174–186.
Baruthio, A, Aubin, J, Mungkung, R, Lazard, J and HM Van der Werf. 2008. Environmental assessment of Filipino fish/prawn
polyculture using Life Cycle Assessment. 6th International Conference on LCA in the Agri-Food Sector, 12–14 November 2008,
Ziegler, F, Nilsson, P, Mattsson, B, and Y Wahher. 2003. Life Cycle Assessment of frozen cod fillets including fishery-specific
environmental impacts. Int J LCA, 8 (1), 39–47.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
milk. Milling wheat, however, is processed into bread, biscuits etc. and this is all included in the
post RDC data. Secondary processing of meat etc. into sausages, pies, pizzas etc. is included
post RDC.
We expected that it would not be possible to examine the post-farm gate part of the food system
along commodity lines but that the combination of the pre-farm gate along commodity lines and
the post farm gate analyses of food types (e.g. bakery and fresh, preserved or frozen) would
deliver an adequate basis for scenario building and assessment. Food product consumption
data were obtained from Defra’s Family Food Datasets.
These include values for home and
eating out consumption. These data were supplemented by calculating energy use in
distribution, purchasing, processing, refrigeration and cooking using the models of Mila i
In addition, wastage rates were taken from the Family Food Survey, together with the
original source in WRAP’s food waste study.
The main data inputs for each activity were taken
mainly from Mila i Canals, which includes much from a Swedish study.
Detailed manufacturing
energy was mainly taken from data compiled by Carlson-Kanyama and Faist
, with milk
processing data taken from Foster et al.
and other specific processes from Hanssen et al.
, Braschkat et al.
; Koroneos et al.
; Cordella et al
; Hospido et al.
; Berlin
, and
Hospido et al.
The food and drink consumption data do not conveniently account for all
commodities produced and some simplifications were needed in estimating the processing
energies and associated GHG emissions. For example, soft drink production was based on that
of sparkling bottled water
with additional sugar and some extra processing energy.
The individual values were summed and cross-checked against top level data for energy use
and refrigerant leakage for manufacturing, domestic food related energy consumption, service
Milà i Canals, L, Muñoz, I, McLaren, S and M Brandão. 2007. LCA methodology and modelling considerations for vegetable
production and consumption. CES Working Paper 02/07, University of Surrey. ISSN: 1464-8083. This paper is a result of the Rural
Economy and Land Use (RELU) programme funded project RES-224-25-0044 (
WRAP. 2008. The food we waste.
Sonesson, U, Janestad, H and B Raaholt. 2003. Energy for preparation and storing of food – Models for calculation of energy use
for cooking and cold storage in households. SIK-Rapport, 709, 1–56. Gothenburg, Sweden, SIK.
Carlsson-Kanyama, A and Faist, M. 2000. Energy use in the food sector: A data survey. AFN report 291, Swedish Environmental
Protection Agency, Stockholm, Sweden.
Foster, C, Green, K, Bleda, M, Dewick, P, Evans, B, Flynn A and J Mylan. 2006. Environmental impacts of food production and
consumption: A report to the Department of Environment, Food and Rural Affairs. Manchester Business School. Defra, London.
Foster, C, Audsley, E, Williams, AG, Webster, S, Dewick, P and K Green. 2007. The environmental, social and economic impacts
associated with liquid milk consumption in the UK and its production. A review of literature and evidence. Report to Defra under
project EVO 2067 for the Milk Roadmap Team.
Hanssen, OJ, Rukke, E-O, Saugen, B, Kolstad, J, Hafrom, P, von Krogh, L, Raadal, HL, Rønning, A and KS Wigum. 2007. The
environmental effectiveness of the beverage sector in Norway in a factor 10 perspective. Int J LCA, 12 (4), 257–265.
Jungbluth, N. 2005. Comparison of the environmental impact of drinking water vs. bottled mineral water. Manuscript for the
SGWA information bulletin and GWA (Gas Water Sewage). Commissioned by Swiss Gas and Water Association (SVGW). ESU
services, Uster, Switzerland.
Braschkat, J, Patyk, A, Quirin, M and GA Reinhardt. 2003. Life cycle analysis of bread production – a comparison of eight different
options. 4th International Conference: Life Cycle Assessment in the Agri-food sector. 6–8 October, 9–16. Horsens, Denmark.
Koroneos, C, Roumbas, G, Gabari, Z, Papagiannidou, E and N Moussiopoulos. 2005. Life cycle assessment of beer production in
Greece. Journal of Cleaner Production, 13, 433–439.
Cordella, M, Tugnoli, A, Spadoni, G, Santarelli, F and T Zangrando. 2008. LCA of an Italian Lager Beer. Int J LCA, 13 (2), 133–
Hospido, A, Moreira, MT and G Feijoo. 2005. Environmental analysis of beer production. Int. J. Agricultural Resources
Governance and Ecology, 4, 2.
Berlin, J. 2002. Environmental life cycle assessment (LCA) of Swedish semi-hard cheese. International Dairy Journal, 12, 939–
Hospido, A, Vazquez, ME, Cuevas, A, Feijoo, G and MT Moreira. 2006 Environmental assessment of canned tuna manufacture
with a life-cycle perspective. Resources, Conservation and Recycling, 47, 56–72.
Jungbluth, N. 2005. Comparison of the environmental impact of drinking water vs.bottled mineral water. Manuscript for the SGWA
information bulletin and GWA (Gas Water Sewage). Commissioned by Swiss Gas and Water Association (SVGW). ESU services,
Uster, Switzerland.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
sector and retail.
, Utley and Shorrock
; James et al
The cross-checking
suggested that the sum of individual cooking energies was about half that surveyed, suggesting
a substantial inefficiency in cooking activities. Refrigerant emissions from road transport and
retail were taken from a recent study by Brunel University
plus corporate social responsibility
reports from supermarkets and food processors (e.g. Co-op, Sainsbury’s, Morrisons, ASDA,
M&S, Waitrose, Tesco, United Biscuits, Unilever, Northern Foods, Weetabix, SAB Miller,
Premier Foods, Adnams, Brake Bros). These varied widely in value, with one from Tesco being
particularly useful on refrigerant leakage.
One area for which we could find no data was
refrigerants from shipping. These, like some large industrial facilities may well be based on low
GHG refrigerants anyway.
Alcoholic drinks
One area of consumption data in which the Family Food Survey data clearly under-reported
consumption was alcoholic drinks, which are dominated by beer, wine and cider. Sources
included the Office of National Statistics (ONS) PRODCOM reports – PRODucts of the
European COMmunity, which is a European Union (EU)-wide scheme
, the British Beer &
Pub Association, UK Quarterly Beer Barometer, the ONS survey on drinking and health, and the
Revenue & Customs reporting on alcohol “clearances”, presumably after duty has been paid
and the Wine and Spirit Trade Association.
These sources indicated a difficulty in obtaining
reliable statistics in this area. This difficulty is widely acknowledged and actually caused the part
of ONS responsible for the beer PRODCOM report to test the reliability of their survey data.
Food services
The data on impacts of eating out and obtaining food and drink from the service sector are
much more uncertain than those for domestic consumption. There is work under way for WRAP
and Defra on quantifying these environmental impacts, but results are not available (the Defra-
funded study was only due to start in the summer 2009). There are some top level indications of
energy consumption in the BERR data, but these are incomplete or may overlap functions, e.g.
general hotel operation plus cooking. In the service sector, practices and serving environments
will vary considerably (e.g. chip shop to haute cuisine restaurant). Additional energy is often
used for heating plates, keeping prepared food hot (or cold) and ambience. Wastage rates can
be very high, but are not quantified. After due consideration, it was decided to assume that all
DECC. 2008 Energy Consumption in the UK, Industrial data tables, 2008 update.
BERR. 2009a. Carbon dioxide emissions estimates and fuel used in electricity generation 1990 to 2007.
BERR. 2009b. Energy Consumption in the UK, service sector data tables, 2008 update.
Utley, JI and Shorrock, LD. 2006. Domestic energy fact file. Report for Defra by BRE.
James, SJ, Swain, MJ, Brown, T, Evans, JA, Tassou, SA, Ge, YT, Eames, I, Missenden, J, Maidment, G. and D Baglee. 2009.
Improving the energy efficiency of food refrigeration operations. Presented at The Institute of Refrigeration, 5 February 2009.
LACORS. 2007. UK implementation of fluorinated greenhouse gases and ozone-depleting substances regulations. Market
intelligence and risk-based Implementation model. From:
Tassou, S, Hadawey, A, Ge, Y and D Marriot. 2009. Presentation on “Greenhouse gas impacts of food retailing”. Defra-funded
project FO0405.
Tesco. 2009. Measuring our carbon footprint.
ONS. 2007a. Product sales and trade. PRA 15960, Beer. 2007.
ONS. 2007b. Product sales and trade. PRA 15940. Cider & Other Fruit Wines. 2007.
ONS. 2007c. Product sales and trade. PRA 15930. Wines. 2007.
Revenue & Customs. 2009.
WSTA. 2005. Wine and data sheet – December 2005.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
cooking, cooling and “presentation” energies were twice that of domestic food preparation and
that the wastage rate for all food and drink was fixed at 30%.
Our analyses include consideration of the use of plant-based livestock product analogues and
other direct ‘like-for-like’ substitutes. Direct substitutes to animal products were estimated on the
basis of soya milk replacing dairy milk, margarine replacing butter and soya cheese replacing
dairy cheese. No LCA studies have apparently been performed on soya milk or cheese, so
estimates were made on the basis of the mixtures of soya meal, soya oil and sugar needed to
produce the gross compositions cited on product labels and/or the composition tables provided
by the Food Standards Agency
, together with some processing energy. Meat substitution was
by replacement of the dry weight of all meats by the dry weight of a mixture of alternatives.
These were 20% (by protein content) of the textured fungal food Quorn, 20% (by protein
content) of tofu and 60% (by protein content) of a mixture of pulses (soya, chickpea, kidney
beans, dried peas, green beans and green peas) consumed directly. The substitution reduces
protein intake and energy from meat by 33% and 45% respectively (note this is not a change in
the whole diet, just this part). All other aspects of diet were assumed to remain the same. Other
approaches are possible, but this provided a convenient substitution for intake. An earlier
approach of substituting on the basis of the same energy and protein would have led to an
untenably high dry matter intake of the alternatives.
There were no complete LCA reports on Quorn, but Nonhebel and Raats
calculated energy
use and material flows in Quorn production. We derived the GWP from this source. The main
microbial energy substrate in Quorn production is molasses, but the large increase in production
that would be needed to support a meat-free diet would mean that the amount of molasses
currently available as a by-product would be greatly exceeded. Much is currently used in animal
feed. So, a main effect would be growing more sugar from domestic beet or overseas cane.
We initially calculated a value for tofu production based on the gross composition and an
estimate of manufacturing energy, although other studies subsequently came to light.
study by Muroyama et al.
is more detailed and process based than that of Håkansson et al.
which seems to give a very high value, but much is based on the cost of energy and an
estimated conversion factor. We cannot say if one is undoubtedly more reliable than the other,
but the results of Muroyama et al. seem more plausible (and were much closer to ours) and
were subsequently used.
Egg substitution is very speculative and is based on a hypothetical alternative derived from soya
Vitamin B
, iron and calcium dietary requirements were taken from Salmon
and related to the
animal-based and vegetable-based alternatives to estimate supplementation requirements.
Vitamin B
production was assumed to be the same as the synthetic production of the amino
acid lysine and existing inventory values were used for iron and calcium. It is worth noting
immediately that the quantities of B
needed are very small, because the amounts in any
foodstuff are only a few µg per 100 g, compared with several g of fat, protein or carbohydrate,
Food Standards Agency. 2002. McCance and Widdowson's The composition of foods. 6th Summary Edition. Cambridge, Royal
Society of Chemistry.
Nonhebel, S and Raats, J. 2007. Environmental impacts of meat substitutes: comparison between Quorn and pork. Proceedings
5th international conference on LCA in foods, 25–26 April, 73–75. Gothenburg, Sweden.
Muroyama, K, Hayashi, T, Ooguchi, M and J Hayashi. 2003. Evaluation of environmental impact for tofu production on the basis of
cumulative CO
emission unit. Environmental Science, 16 (1), 25–32.
Håkansson, S, Gavrilita, P and X Bengoa. 2005. Comparative Life Cycle Assessment pork vs tofu. Life Cycle Assessment,
1N1800, Group 5 Stockholm.
Salmon, J. 1991. Dietary reference values: A guide. HMSO.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
so that the unit burdens of producing B
would need to be extraordinarily high to have any
substantial effect on the overall impacts of a supplemented food.
GHG emissions from the ‘Regional Distribution Centre (RDC) to retail
The energy used to deliver food from the RDC to retail stores and during retail itself (including
refrigeration, heating, lighting and ventilation) varies according to storage temperatures and
throughput, as well as distance. Additional emissions of GHG also occur from mobile chillers
and those used in retail outlets as well as land-filling wasted food. The GHG emissions of
different foods were estimated from Tassou et al.
and based on the storage temperatures in
the RDC and retail stores (Table 5). The landfill emissions assume a relatively low wastage rate
from RDC to retail of 1% over all food types. This is an area still being researched in Defra and
WRAP funded studies so is an arbitrary estimate based on informed opinion. The same
wastage rate was assumed in service sector supply chain.
Table 5. Estimates of GHG emission for different food types depending on the temperature of
storage and delivery, kg CO
e/kg. The letters have these meanings, with the first applying to the
RDC and the second to retail: A = Ambient, R = Refrigerated, F = Frozen, M = Milk (fresh)
Source of emissions AA RA RR FF MM
Electricity 0.001 0.008 0.50 0.61 0.036
Refrigerants 0.000 0.000 0.59 0.38 0.044
Road fuel & Oil 0.052 0.018 0.026 0.017 0.016
Landfill 0.012 0.012 0.015 0.012 0.012
Total 0.065 0.038 1.1 1.0 0.11
Shopping transport energy
Much food is currently bought by using cars or buses, which use petrol or diesel. The energy
used in shopping came from Pretty et al. (2005).
They calculated that the average shopping
basket weighs 28kg and the mean distance travelled is 6.4km. Assuming a set of ways of
travelling to shops, the GHG emitted per kg is 0.034kg CO
e/[kg shopping] (Table 6).
Table 6. Energy used and GHG emitted during the average shopping trip (based on Pretty et al.
Transport modes Proportion Fuel, litres/km Occupancy rate MJ/kg kg CO
Car 59% 0.081 1 0.80 0.057
Bus 8% 0.40 30 0.13 0.009
Walking 30%
Cycling 3%
Total 100%
Weighted mean 0.48 0.034
Cold storage in homes and food service sector
Once food and drink are taken home or delivered to a food service sector outlet, some is stored
in refrigerators or deep freezes. In the service sector, open top devices are also used (e.g. for
salad bars) and drinks may be stored in cellars, behind-bar cabinets and trays and served
through chilled pipes. Fridges and freezers are typically the most power consuming item in the
home as the top level BRE data shows. The energy use in domestic households was estimated
Tassou, S, Hadawey, A, Ge, Y and D Marriot. 2009. Presentation on “Greenhouse gas impacts of food retailing”. Defra-funded
project FO0405
Pretty, JN, Ball, AS, Lang, T and JIL Morison. 2005. Farm costs and food miles: An assessment of the full cost of the UK weekly
food basket. Food Policy, 30, 1–19.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
from typical appliance energy usage and the throughput of food and drink in households. It was
then allocated to each refrigerated or frozen item by weight (Table 7). Energy use in the service
sector was derived from BERR’s top level data. It was applied at the sectoral level and the best
estimate of the value per kg food or drink is also given.
Table 7. GHG emissions data for cold storage in home and the food service sector used in this
Domestic Service sector
Chilled Frozen Refrigerated and frozen
e/kg product 3.5 kgCO
e/kg product
24 kgCO
e per person per year
Cooking energy
The cooking energy data originate from the formulae of Sonesson et al.
(with later
interpretations from Carlsson-Kanyama & Faist
). Sonesson et al.’s formulae seem to be based
on best practice. Work at Campden BRI has shown that cooking energies can vary widely for
the same food type by using different equipment and cooking methods (e.g. stir fry, roast,
boiling or frying. The variation of individuals is also considerable, e.g. observe the amount water
boiled to make one cup of coffee, let alone the effects of portion size on cooking energy (in
which smaller portions are more energy intensive in most cases). We started by quantifying a
range of foods using most likely methods and applying Sonesson et al.’s formulae, which
generated a range of cooking intensities with a maximum of 10 MJ/kg. Using an equal mixture
of electricity and gas as energy carriers (i.e. 5 MJ delivered electrical energy and 5 MJ delivered
net energy from natural gas) causes the emission of 1.3kg CO
e/kg, which was reduced to 0.09
with low CO
energy supplies. The food types given in the Family Food Survey were ranked by
expert opinion. Small allocations were included for take-away items that were consumed in the
home to allow for some re-heating. This bottom-up modelling approach was found to
underestimate the energy used in cooking from the BRE top-down survey data
by about 50%
Given that the scaling applied was relatively coarse and without being able to obtain more
detailed activity data, all individual values were then doubled and were used in the subsequent
Wasted food management
It was assumed that most food waste currently goes to landfill with very limited energy recovery,
0.49kg CO
e/kg waste. The improved method is based on data from the Holsworthy centralised
anaerobic digester in Devon, in which food wastes and manure are co-digested. The results of
Cumby et al
were analysed and used to calculate a net credit from electricity generation
0.031kg CO
e/kg waste. This allows for the extra fuel of collection etc.
Enteric and sewage emissions
These were omitted from the study owing to lack of resources and the expectation that the
effects of dietary change would have relatively small effect on these. Furthermore, while the
change in available energy mixture would have some effects on reducing the impact of
Sonesson, U, Janestad, H and B Raaholt. 2003. Energy for preparation and storing of food – Models for calculation of energy use
for cooking and cold storage in households. SIK-Rapport, 709, 1–56. Gothenburg, Sweden, SIK.
Carlsson-Kanyama, A and Faist, M. 2000. Energy use in the food sector: a data survey. AFN report 291, Swedish Environmental
Protection Agency, Stockholm, Sweden.
Utley, JI and Shorrock, LD. 2006. Domestic energy fact file (2006). Report for Defra by BRE.
Cumby, TR, Sandars, DL and E Nigro. 2004. Physical assessment of the environmental impacts of centralised anaerobic
digestion – Defra-funded project CC0240.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
wastewater and sewage sludge management, the overall range of possibilities really deserves a
separate study in its own right. Also, the focus of the study was on production, distribution and
The ‘top-down’ method of calculating land use change emissions attributable to
agricultural production
This approach involves estimating total observed land use change (LUC) emissions caused by
commercial food production, and allocating that total “pool” of emissions to different food-types
consumed in the UK based on their global average land-area requirements per unit of
production. It should be noted that this approach does not divide emissions into emissions
arising from LUC directly connected to crop consumed (direct emissions) and indirect emissions
arising from the effect of land use for consumed crops displacing other crops to agricultural land
obtained by LUC (indirect emissions). This is based on a methodology published by
Central to the approach is the consideration that agricultural commodity markets
are global and interconnected, and all demand for agricultural land contributes to commodity
and land prices, and therefore contributes to land use change. The steps are set out in Table 8.
There are a number of advantages to this approach. Firstly, the emissions allocated to different
food-types will not sum to a figure which is greater than actual observed LUC emissions. This is
important to maintain the integrity of a consumption-based emissions accounting approach (i.e.
total emissions allocated should not exceed total emissions, also known as the “100% rule”).
Secondly, food-types which have high land use requirements (e.g. beef) are allocated higher
LUC emissions, and switching to food-types with lower land use requirements will show a
reduction in LUC emissions. In addition, measures such as a reduction in total food
consumption will show a reduction in LUC emissions. Thirdly, the method recognises that all
demand for agricultural land contributes to LUC pressures (either directly or indirectly), and
therefore all demand for agricultural land (via the consumption of agricultural commodities)
should be allocated a share of LUC emissions.
Table 8. Steps in conducting the ‘top-down’ method to estimate land use change greenhouse
gas emissions attributable to UK food consumption
Step 1. Estimate total LUC emissions per year (GtCO
Step 2. Estimate the proportion of total LUC caused by commercial agriculture, including ranching (% of
Step 3. Divide LUC emissions attributable to agriculture (derived from Steps 1 and 2) by total
commercial agricultural land area to derive LUC emissions per hectare (tCO
Step 4. Calculate land requirement for each food commodity consumed (hectares/tonne of commodity)
Step 5. Multiply LUC emissions factor (from Step 3) by commodity land requirement (from Step 4) to
derive LUC emissions per tonne of commodity (tCO
Step 6. Multiply LUC factor per tonne of commodity (from Step 5) by total quantity of each commodity
consumed in the UK (tCO
Step 7. Sum the LUC emissions calculated for each commodity (from Step 6) to derive total LUC
emissions associated with UK food consumption
One of the disadvantages of this method is it does not pick out the possible differences between
food-types which happen to have the same land-area requirements per unit of output. For
example, if palm oil and rape seed oil had similar land-area requirements per unit of output then
they would be allocated the same LUC emissions (although the actual total (direct and indirect)
LUC impacts may be different – e.g. palm oil may cause higher total emissions than rapeseed
oil). This limitation in the accounting method may have the perverse effect of directing
consumption towards commodities that have higher LUC impacts. One possible solution to this
issue is to implement a decision-rule when considering mitigation options, e.g. if switching from
Tipper, R, Hutchison, C and M Brander. 2009. A practical approach for policies to address GHG emissions from indirect land use
change associated with biofuels. Ecometrica, UK.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
a high land-requirement food to a low land-requirement food, the low land-requirement food
should not be associated with direct LUC.
Details of the use of the ‘top-down’ method as applied to this study
Global average yields for crops and livestock land requirements were used in the analysis,
rather than the yields of crops and the land requirements of livestock directly consumed by the
UK. This approach was adopted to reflect the integration of world commodity markets. The UK’s
demand for commodities contributes to world prices generally, rather than to prices for
commodities with a specific land requirement, and therefore world average yields are
considered appropriate. This approach also avoids the possibility of “playing” the accounting
system by consuming commodities from higher yield regions, and leaving lower yield production
for others to consume (with total emissions remaining the same).
Quantifying land use for animal production presented a special challenge, particularly for
grassland. We used estimates of total arable crop use in livestock production in 2002
estimate arable crop use in livestock production in 2005 by adjusting the 2002 figures to
account for changes in livestock production between 2002 and 2005. Total livestock production
in 2005 and associated permanent grassland was screened to identify the world’s ‘commercial’
livestock production and associated pasture. This was done in order to exclude large areas of
extensively grazed pasture which are not connected to global commodity markets. FAO country
level livestock production, import, export and land use data sets were synchronised with each
other to allow screening using all parameters. A country was defined as having a commercial
livestock industry connected to world trade if its exports or imports were greater than 0.5% of
world imports or exports and production was greater than 0.5% of world production in 2005. We
examined several screens against countries most would regard as connected and not
connected to world trade and this screen proved most efficient against these sense checks.
Agricultural production on native wild grassland is not connected to land use change. Our
screen had the merit of excluding most of the world’s native grassland – e.g. the Savannahs of
Africa and the native grasslands of Mongolia.
The totals for arable crops used in livestock production were allocated to world livestock
production using the rates of feed use for livestock products as identified in the Cranfield
From this, the inputs of the major feed commodities per tonne of output were identified.
The land area required was calculated as for crops for direct human consumption using average
global yields.
The allocation of pasture was done in a slightly different way. The starting point was the
assumption that commercial pasture use is dominated by cattle for milk and beef, and sheep
and goats for meat. The screen described above was used to identify the area of permanent
pasture and the corresponding meat and milk production connected to world trade. The
cultivated pasture (i.e. pasture sown on arable or potentially arable land) was added to this
resulting in an estimate of the total grassland area used to support commercial livestock
production connected to world trade. Due to the practice of multiple or combined grazing it was
not possible to calculate land requirements for specific commodity types, e.g. bovine meat,
sheep meat and milk. Therefore the use of pasture land was treated as a single process and the
associated emissions were allocated by economic output value. To avoid using values
influenced by local subsidies and markets, representative world prices were derived from
average producer prices in Australia and New Zealand in 2005. Total emissions from cultivated
pasture were allocated between beef, sheep/goat meat and milk. All the emissions from
permanent pasture on land not suitable for arable crops were allocated to the meats. The work
Steinfeld, H, Gerber, P, Wassenaar, T, Castel, V, Rosales, M and C de Hann. 2006. Livestock’s long shadow. FAO.
Williams, A, Audsley, E and D Sandars. 2006. Determining the environmental burdens and resource use in the production of
agricultural and horticultural commodities. Defra project report IS0205.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
therefore considers dairying as based on cultivated pasture, based on an assumption that
commercial dairying worldwide is generally conducted on cultivated pasture while commercial
permanent grassland is generally used for meat production. The analysis is based on the IPCC
estimate of land use change emissions
and data on the primary drivers of land use change
(Table 9). This identified the total global land use change emissions arising from agriculture.
Table 9. Basis for identifying the proportion of deforestation attributable to commercial
agriculture as a basis for partitioning land use change emissions
Annual average forest loss between 2000 and 2005 - Africa: 4 million ha
Annual average forest loss between 2000 and 2005 - Asia/Pacific: 3.7 million ha
Annual average forest loss between 2000 and 2005 - Latin America: 4.4 million ha
Total 12.1 million ha
% of deforestation due to large scale agriculture - Africa: 12%
% of deforestation due to large scale agriculture - Asia/Pacific: 29%
% of deforestation due to large scale agriculture - Latin America: 47%
% of deforestation due to small scale permanent agriculture – Africa: 59%
% of deforestation due to small scale permanent agriculture - Asia/Pacific: 13%
% of deforestation due to small scale permanent agriculture - Latin America: 13%
Proportion of total LUC emissions attributable to commercial agriculture: 58.1%
The estimates of LUC emissions associated with UK food consumption resulting from this
methodology should be interpreted with care, especially when considering mitigation options.
The method is based on an attributional approach which allocates LUC emissions based on the
average land area requirements of the foods consumed in the UK. Attributional LCA (ALCA) is
useful for allocating "responsibility" for emissions, based as closely as possible on the causal
relationship between the emissions and the entity to which they are allocated. It is also the
appropriate approach for consumption-based carbon accounting as it avoids double-counting
emissions. However, it does not capture all the complexities and consequences of specific
mitigation actions or policies.
In order to quantify the full GHG consequences of an action, consequential LCA (CLCA) is
required. CLCA looks at marginal changes arising from actions and quantifies all the
consequences which flow from this. The attributional approach is therefore useful for estimating
the size of LUC emissions attributable to UK food consumption, and it can indicate possible
mitigation options, but it does not accurately quantify the actual emissions reductions achieved
by different mitigation options. For example, attributional LCA may show that in the current
agricultural system, beef has more embedded emissions than poultry meat. The attributional
approach is essentially a system of accounting emissions and attributing them to commodities
as currently produced and consumed. However, it does not say what the full consequences of a
significant shift from beef to poultry would be. For example, a reduction in beef consumption
may increase reliance on male calves from the dairy herd reducing the burdens from beef
production. It should be noted that this limitation with attributional analysis arises for most
emissions sources across the economy. For example, a grid average emissions factor is used
when allocating emissions from electricity consumption (within an ALCA). However, when
quantifying the actual emissions reductions from reducing electricity consumption the grid
IPCC. 2007. Climate change 2007. Synthesis Report.
FAO. 2007. State of the world’s forests.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
margin should be used, and other consequences from the action should also be taken into
account. The relationship between the attributional figures for LUC and the emissions
reductions achieved by specific mitigation options is likely to be less close than for other
emissions sources, given the complexity of the causal interactions between demand for a food
commodity and LUC (particularly indirect land use change). Attributional figures help to indicate
possible mitigation options, such as switching from foods which have high land area
requirements to those that have lower land area requirements. However, such options should
be investigated in greater detail using consequential analysis, in order to accurately assess the
emissions reductions achieved.
Uncertainties arising from the ‘top-down’ method
Estimates of land use change emissions have high uncertainty
, and perhaps the highest
uncertainty of any emissions source. There is therefore high uncertainty associated with the
estimate of total LUC emissions used in this study (the 8.5 GtCO
e figure derived from the
IPCC’s Fourth Assessment Report
). There is also high uncertainty associated with the
estimate of the proportion of total LUC emissions attributable to commercial agriculture, which is
based on the FAO’s State of the World’s Forests Report 2009.
Land use change is driven by
the interaction of numerous proximate and underlying causes, and attributing a proportion to a
single cause will be approximate.
A further source of uncertainty in the calculations relates to the allocation of emissions
associated with pasture use. Data were not available for the average pasture land area
requirements for livestock commodities and therefore the LUC emissions associated with the
use of pasture land were allocated between beef, sheep and goat meat, and milk products on
the basis of economic value (and other underlying assumptions). There are a number of further
steps in the methodology which could be performed in different ways, for example the allocation
of emissions could be undertaken on the basis of economic value rather than land area
requirement per unit of commodity. This approach would reduce the allocation of LUC
emissions to high land requirement commodities such as beef and sheep meat, but the LUC
emissions associated with these commodities would remain relatively high due to their high
economic value.
Differentiated emissions value for pasture land/credit for increased carbon sequestration in
pasture land
The method for estimating LUC emissions attributable to UK food consumption uses a single
emissions factor for agricultural land, i.e. 1.43 tCO
e/hectare of agricultural land used. There is
a case for using different emissions factors for pasture land and cropland, as grassland
generally has higher carbon stocks than cropland. In order to apply this approach average
carbon stock figures for pasture and cropland are required. The average baseline carbon stock
of land converted to agriculture and a method for calculating emissions factors to reflect the
relative contribution of pasture or cropland to LUC emissions would also be required. It is also
necessary to ensure that when the calculated emissions factors are multiplied by the total area
of each land use type, the total emissions figure equals the total LUC emissions associated with
agriculture (to avoid over or under allocating emissions). Further complexities may arise if the
categories of pasture land and cropland are considered too broad, and differences within these
categories are accounted for, such as the variation in carbon stocks depending on crop type,
Ramankutty, N, Gibbs, HK, Achard, F, Defries, R, Foley, JA and RA Houghton. 2007. Challenges to estimating carbon emissions
from tropical deforestation. Global Change Biology, 13, 51–66.
Barker T, Bashmakov, I, Bernstein, L, Bogner, JE, Bosch, P, Dave, R, Davidson, O, Fisher, BS, Gupta, S, Halsnæs, K, Heij, BJ,
Kahn Ribeiro, S, Kobayashi, S, Levine, MD, Martino, DL, Masera, O, Metz, B, Meyer, L,. Nabuurs, G-J, Najam, A, Nakicenovic, N,
Rogner, H-H, Roy, J, Sathaye, J, Schock, R, Shukla, P, Sims, REH, Smith, P, Tirpak, DA, Urge-Vorsatz, D and D Zhou. 2007:
Technical summary. In: Climate change 2007: Mitigation. contribution of Working Group III to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change [B Metz, OR Davidson, PR Bosch, R Dave and LA Meyer (eds)], Cambridge University
Press, Cambridge, United Kingdom and New York, NY, USA.
FAO. 2009. State of the world’s forests report
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
land management practices, and soil type.
The possibility of using different emissions
factors for different land use types should be the subject of further research. An alternative
approach may be to apply a single emissions factor for agricultural land, but to introduce a
“credit” or derogation for commodities which are from agricultural systems which can be shown
to avoid direct and indirect land use change. For example, if commodities are produced on
marginal or degraded land which would not have been used for any other purpose, they may
have neutral or even positive effects on direct carbon stocks (e.g. on the degraded land), and
will not displace other agricultural activities (and therefore avoid indirect land use change).
Land use emissions, soil carbon changes
The world’s soils are estimated to contain 1,500 Gt of organic carbon which is roughly twice
that in the atmosphere.
Oxidation of soil organic matter accounts for a natural flux of about 75
Gt per year through which carbon entering the soil from plants is returned to the atmosphere.
The UK has a net emission of 2 Mt CO
from land according to the UK GHG inventory.
Grassland absorbs 8 Mt, and crop land releases 15 Mt. Losses from arable soils include the
oxidation of fenland peat, which is an irreversible loss. The uptake of carbon by grassland
includes increased storage in temporary grassland which is partly offset by emissions from the
arable phase land in these mixed-farming rotations. Climate change rather than land use is
implicated in long-term reductions.
This has not been formally introduced into the UK GHG
inventory, but the possibility is being considered.
In the UK context, these fluxes from soil are reversible and not intrinsically linked to agriculture
on stable soils such as those in northern Europe. However, we recognise that expansion of
agricultural land on a global scale, especially the expansion of arable land, would increase soil
carbon losses.
Regional emissions
The study examined differences in consumption between the UK regions and the implications
for emissions. The consumption of commodities by the English, Welsh, Scottish and Northern
Irish was derived from the Family Food Survey consumption data. The commodity contents of
all product categories from milk to various types of ready meals and food service products were
estimated from the commodity composition of each. The total commodity values consumed
were then summed and compared with the pre-RDC FAO data. Agreement was reasonably
good (70% to 110% for most commodities). The FAO data were then scaled by the population
of each part of the UK and the per capita consumption of commodities was obtained.
Mitigation measures
The main aim of the study was to consider potential scenarios for reducing human-induced
GHG emissions attributable to the UK food system by 70% by 2050. To examine reductions in
the region of 70%, scenarios require several mitigation measures to be implemented together.
Wang, Z, Han, X and L Li. 2008. Effects of grassland conversion to croplands on soil organic carbon in the temperate Inner
Mongolia. Journal of Environmental Management 86, 529–534.
Kim, H, Kim, S and B Dale. 2009. Biofuels, land use change, and greenhouse gas emissions: Some unexplored variables.
Environmental Science and Technology, 43, 961–967.
Schlesinger, WH and Andrews, JA. 2000. Soil respiration and the global carbon cycle. Biogeochemistry 48, 7–20.
NAEI. 2005. UK emissions of air pollutants 1970 to 2005.
Bellamy, PH, Loveland, PJ, Bradley, RI, Lark, RM and GJD Kirk. 2005. Carbon losses from all soils across England and Wales
1978–2003. Nature, 437, 245–248.
Thomson, AM. 2008. Inventory and projections of UK emissions by sources and removals by sinks due to land use, land use
change and forestry. Annual Report, July. Defra Contract GA01088, CEH No. C03116.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
The research first examined the effect of individual mitigation measures on the emissions
inventory. The first stage was to be a free-thinking listing of all possible measures (e.g. change
from red meat to white meat; reduce GHG emissions form livestock waste). We identified 7
consumption and 23 production measures (Table 10).
Production measures
A series of production measures was drawn up and a model developed to test the overall
impact of these when applied to all commodities. The majority of GWP values for these
measures were derived using the Cranfield model and values for other commodities scaled or
inferred from proxy values.
Zero electricity from fossil fuels
This measure assumed that all electricity could be produced from non-fossil fuel based sources.
Electricity burdens within the Cranfield model were adjusted to reflect this and other commodity
values were scaled in proportion to the reduction achieved.
Zero enteric emissions
This highly speculative measure assumed the development of technology or feed to completely
remove or perfectly capture enteric emissions from ruminants. The Cranfield model is structured
such that it was possible to set enteric methane emission factors for beef, dairy cattle and sheep
to zero. Using data from the comparative LCA study
, GWP values for commonly imported
livestock commodities such as Brazilian beef and New Zealand lamb were scaled appropriately
based on results for UK livestock.
O release inhibitor with fertiliser
This measure assumed that fertiliser could be produced such that N
O emissions from soils
could be completely prevented. To simulate this, the IPCC emission factor EF1 (emission factor
for N
O emissions from N inputs
was set to zero. This assumes that a nitrification and
denitrification inhibitor can stop N
O emissions from synthetic N fertiliser.
Anaerobic digestion (AD) of manure (no stored manure emissions)
This was applied to manure from all non-grazing stock and it was assumed that the emissions of
methane from the point of manure capture were zero. The benefits of anaerobic digestion were
quantified as credits from removing methane emission from managed manure and credits for
generating electricity. The electricity generated was taken from Parsons
and the benefits are
summarised in Table 11.
50% yield increase
This measure assumed that with no increase in fertiliser application rates or change to land
requirements it would be possible to increase crop yields by 50%.
Zero N
O from nitrate fertiliser production
This is a specific emission from one stage in fertiliser manufacture that relates only to nitrate
production and is associated with N
O emissions that can be abated already to some degree.
Defra project FO0103. Comparative life-cycle assessment of food commodities procured for UK consumption through a diversity
of supply chains.
IPCC. 2006. IPCC Guidelines for national greenhouse gas inventories.
Parsons, DJ. 1984. A survey of literature relevant to the economics of anaerobic digestion of farm animal waste. Divisional Note
DN. 1225, National Institute of Agricultural Engineering, Silsoe, UK.
Parsons, DJ. 1986. The economics of the treatment of dairy-cow slurry by anaerobic-digestion. Journal of Agricultural Engineering
Research, 35 (4), 259–276.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Improved Feed Conversion Ratio
Without reference to method, this measure assumed that it would be possible to improve the
feed conversion ratio (FCR) of livestock by 25% over the next 40 years, that is, the ratio of mass
of all food eaten to body mass gain over a specified period of time. If body mass gain is greater,
or food consumption reduced this reduces the ratio. Thus the Cranfield model was used and the
FCR reduced by 25% for pig meat, poultry and eggs. For beef, dairy cattle and lamb, the
efficiency of fattening was increased by 25% to give an equivalent effect.
Table10. Details of mitigation measures
Zero fossil fuels (electricity and other energy
Very low carbon fuels (1% of standard) – including diesel
No enteric methane emissions from
No enteric methane emissions from ruminants
O inhibitor with fertiliser (no N
O from soils) No N
O from fertiliser applied to soils
Anaerobic digestion (AD) of manure (no
stored manure emissions)
No methane from manure and all used in AD to produce bio-energy
50% yield increase Crops having 50% increase in yield with associated increase in inputs
Zero N
O from nitrate fertiliser production No N
O from fertiliser production through perfect filtration
25% improvement in feed conversion
Improved feed conversion ratio (FCR) for finishing meat animals and in egg
and milk production
N use efficiency in crop production increased
by 50%
Reduce losses of nitrogen by denitrification, volatilisation or leaching by 50%
Livestock production based on by-products
(grass still used for ruminants)
Concentrates produced only using by-products plus beans and wheat where
Minimum tillage (where possible) Reduce tillage energy to levels of minimum tillage for all crops
Organic production Commodity production using organic methods rather than the non-organic
assumed elsewhere
Energy, processing, distribution, retail and preparation (post RDC)
Low carbon energy for cooking Reduced emissions from cooking by using very low carbon fuels (7% of
standard), but the same amount of process energy in the home and service
Low carbon energy for supply chain chilling Reduced emissions from refrigeration and freezing by using very low carbon
fuels (1.25% of standard), but the same amount of process energy.
Refrigerant emissions still the same
50% saving in energy inputs into processing Assumed more efficient food industry using 50% of energy in embedded
materials and process energy
Low GWP potential refrigerants Low GWP potential refrigerants used in transport and retail cold shelves
Low carbon transport in processing and
Very low carbon fuels (1.25% of standard) used in the RDC and retail stores
Energy recovery from food waste using AD No reduction in waste arising, but better management with energy recovery
Low energy use in consumer transport 10% of current energy used by shoppers and in transport to service outlets
95% reduction in GWP of packaging Reduced GWP from packaging in the supply chain (5% of standard), e.g.
much lower wastage, less material &/or fuel efficient recycling
75% reduction in GWP from shopping bags 25% of current GWP by more re-use of shopping bags from retail and take-
away service sector outlets etc.
Low GWP home refrigerants Low GWP potential refrigerants in homes and service sector outlets
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
No meat Meat is replaced by fungal protein, tofu and pulses
66% reduction in livestock products Livestock products are reduced and other food increased by 29%
50% reduction in livestock products Livestock products are reduced and other food increased by 21%
Red to white meat Red meat is replaced by white meat with an increase in vegetables (NB there
is still some shortage of vitamins, but these have small burdens of production)
No dairy milk Dairy milk and products are replaced by soy based milk products
No rice Rice is replaced by wheat and potatoes
No eggs Eggs are replaced by “soy synthetic egg”
All avoidable food waste avoided Unavoidable waste (WRAP definition) still to landfill etc. but less production
Table 11. Manure dry matter (DM) outputs and GHG emissions credits from using anaerobic
Manure DM
output per t
kWh/t manure
GHG Credit,
kg CO
Credit from
stopping CH
emissions, kg
e /t commodity
Total GHG
Credit t CO
e /t
Pig 1.9 196 373 250 398 0.65
Poultry 0.7 420 273 183 8 0.19
Beef 8.0 155 1,243 831 1,135 2.00
Milk 0.4 155 55 36 47 0.08
Eggs 1.1 420 448 299 17 0.32
N use efficiency in crop production increased by 50%
This assumed that N losses by denitrification, volatilisation and leaching were all reduced, thus
requiring lower N supplies for the same yield.
Livestock from by-products
The concentrates were re-formulated to use by-products (rapemeal, brewer's grains and
wheatfeed) as protein sources replacing imported soya and maize by-products. The aim was to
maintain the same metabolisable energy and digestible crude protein content. Where necessary
UK feed peas/beans were increased to balance the diet. Although by-products are fully used at
present, increasing amounts will become available with increased biodiesel and bioethanol
production and similarly in a scenario with refined cereal based products replacing meat.
Minimum tillage
This assumes reduced energy consumption for cultivation, equivalent to the energy required for
minimum tillage. The Cranfield model allows most crop commodities to be modelled at this
reduced energy input level, and burdens for other crops were scaled according to the
proportional reduction in appropriate proxy crops.
Zero fossil fuels
Further to the zero electricity from fossil fuels measure, this assumes that all other energy
requirements could be produced from renewable sources. To model this it was assumed that all
pre-RDC carbon dioxide (CO
) emissions arose from combustion, and thus were subtracted
from GWP values for each commodity. This was applied using both the Cranfield model and
various studies from the literature search which produced sources that gave a breakdown of
GWP into component gases. With post RDC cooking, a slightly less effective change was
assumed, given that some gas or solid fuels would always be needed.
100% organic production
This measure assumed that all commodities would be produced using organic production. This
presents some difficulties because the production of all commodities currently consumed is
unlikely to be possible using an all organic scenario (e.g. less poultry and pig production seems
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
inevitable, while beef and sheep would increase). Estimating possible production levels is not
without difficulty and a recent study for the Soil Association by a team from Reading
illustrated this well, with extensive discussion of their findings in the FCRN. They
did not have the resources to model all land use and production thoroughly but used statistical
data from the Farm Business Survey on yields and farm types around England and Wales to
estimate production from yields or farm types. These produced quite disparate results for good
reasons, e.g. wheat production going down by about 35% or 65%. Crops like oilseed rape and
sugar beet are rarely (if at all) grown in the UK, because there is currently apparently no UK
organic market. We would still need oil and sugar in our diets (although not necessarily as much
as we have now) and these would need to be sourced from somewhere. It is inconceivable that
a market would not develop and that some domestic sugar and oil would be produced, although
overseas production might dominate an open market. Another aspect of this is what that wider
context is. In a 100% organic world, global land use would be very different and the ability to
import would change too, so adding further to speculation. The range of commodities actually
consumed would be determined by market forces (it is reasonable to assume). So, with barley
production falling by about 50%
, the amounts of poultry and pig products would be in direct
competition with barley for malting. Factors like this add to the complexity of any forecasting.
We, like the Reading University team, did not have the resources to model an all organic future
as well as could be wished for. The main comparison simply considers the substitution of
current consumption by the same amount of commodities produced organically. This is unlikely,
but it provides some quantification of the differences in GHG emissions between the production
systems. While some commodities have been analysed with the Cranfield model, it does not
address all commodities, especially those produced overseas nor any fruit or field vegetables
(except potatoes). There are some other LCA studies that study organic production, but the
picture is incomplete. Where there were gaps in the data (e.g. fruits) missing values were
assumed to be no different from non-organic production.
Some explorations of alternative production scenarios based on Jones & Crane were also
explored, but they are limited in what they can offer.
It should be noted that post-farm gate, it can only be assumed that distribution and cooking are
essentially the same. Critical comment is inevitable and to avoid re-runs of well worn arguments
about the Cranfield model, we present results from four independently conducted studies on
one the biggest single terms – milk. All the results are broadly similar, without any systematic
differences between non-organic and organic milk production (Table 12).
The systems approach of the Cranfield LCA model enabled the main commodities to be
modelled with the N
O-N portion of the GWP attributable to fertiliser manufacture removed and
other commodity values scaled from this.
Jones, P and Crane, R. 2009. England and Wales under organic agriculture: how much food could be produced? CAS Report
18, Centre for Agricultural Strategy, University of Reading.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Table 12. Comparisons of GHG emissions from milk production by organic and non-organic
production to the farm gate per m
Study Non-organic Organic
Williams et al., 2006
1.10 1.20
Cederberg & Mattsson (2000)
1.10 0.95
Thomassen et al. (2008) (on farm)
0.70 0.90
Wiltshire et al. (2009)
1.2 (high yield), 1.4 (low yield) 1.30
The results of the organic scenarios stand apart somewhat from the main body of results and
are in a separate section. Note that the implications for LUC emissions which are not included
are enormous.
Post RDC measures
The same amount of process energy is used for cooking, but with very low carbon electricity
and some gas. This would reduce emissions to 7% of standard.
Chilling energy
The same amount of process energy is used for refrigeration and freezing in homes and service
sector outlets, but using very low carbon electricity, i.e. 1.25% of standard emissions.
Refrigerant emissions are assumed not to change.
Food processing
A more efficient food processing industry is assumed that reduces external energy input to 50%
of current levels using a combination of embedded energy in materials and more efficient use of
process energy.
Distribution chain refrigerants
The current generation of chiller units used in transport and retail cold shelves mainly use high
GWP potential refrigerants. This assumes that low GWP ones can be used. In general, larger
static plants already tend to use lower GWP refrigerants and/or leak less.
Distribution chain fuels
This assumes that in the RDC and retail outlets, very low carbon fuels electricity (1.25% of
standard) is used.
All current food waste to AD
It is assumed that the current level of food waste arising is maintained, but it is managed better.
Instead of going to landfill, food waste is co-digested with energy recovery as electricity.
Distribution chain delivery energy
Very low carbon fuels (1.25% of standard) are used for transport between the RDC and retail
stores and food service outlets.
Williams, A, Audsley, E and D Sandars. 2006. Determining the environmental burdens and resource use in the production of
agricultural and horticultural commodities. Defra project report IS0205, as further developed under Defra project IS0222.
Cederberg, C and Mattsson, B. 2000: Life cycle assessment of milk production — a comparison of conventional and organic
farming. Journal of Cleaner Production, 8(1), 49–60.
Thomassen, MA, van Calker, KJ, Smits, MCJ, Iepema, GL and IJM de Boer. 2008. Life cycle assessment of conventional and
organic milk production in the Netherlands. Agricultural Systems 96, 95–107.
Wiltshire, J, Tucker, G, Williams, AG, Foster, C, Wynn, S, Thorn, R and D Chadwick. 2009. Scenario building to test and inform
the development of a BSI method for assessing GHG emissions from food. Defra research report FO0404.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Shopping transport
It is assumed that 10% of current energy (and hence GHG emissions) is used by shoppers and
in transport to service outlets.
Reduced GWP from packaging in the supply chain (5% of standard), e.g. much lower wastage,
less material and/or fuel efficient recycling
Shopping bags
More re-use of shopping bags from retail and take-away service sector outlets etc. is assumed,
so resulting in GHG emissions falling to 25% of the current level.
Refrigerants (end users)
Current units use high GWP refrigerants in homes and service sector outlets, although leak
much less than mobile or retail units. It is assumed that low leakage and low GWP potential
refrigerants are used.
Consumption measures
There is now consensus that consumption based mitigation will have changes to livestock
consumption as a major element. Specifying relevant changes in diet presents a special
challenge. We adopted an open approach to developing consumption (diet) measures which
are not constrained by fixed approaches such as ‘vegetarian’ or ‘vegan’, or by a percentage of
the population adopting such approaches. The important thing for the research question is
commodity consumption at the population level rather than the proportion of the population
adopting a particular diet. It is not sufficient to change one dietary component, for example to
reduce beef consumption. The change must be made to the entire diet to reflect increases in
some components in response to decreases in others.
Measures examined include the direct substitution of livestock products using plant based
alternatives. However, we also consider more comprehensive whole diet changes that are not
anchored by efforts to match nutrient profiles, for example protein intake. To guide us, we
examined consumption profiles across the world – first by examining the FAO country profiles of
agriculture and diet
seeking examples of consumption patterns at the population level that
show how food systems may change in relation to diet. Throughout most of the world, significant
deviations from the commodity consumption characteristic of the UK are associated with greater
incidence of under-nutrition, so most countries with low livestock product intakes cannot be
used as examples. Japan is an example of a country with a combination of lower calorie intake,
and lower intakes of animal products (54% that of the UK). The consumption of meat and
especially dairy products is low. However fish consumption is high and individual calorie intake
is low. Turkey was identified as a country with low incidences of under-nutrition with calorie
intakes similar to the UK even though the intake of animal products is only 36% that of the UK.
Total apparent (i.e. all commodity entering the food system including food that is wasted) daily
calorie intake is 3,340 per day compared with 3,440 in the UK. Of key importance is that this is
achieved with daily calorie intake in livestock products being just 385 (12%) compared with
1056 (31%) in the UK. These significant reductions in livestock production consumption are
dominated by very significant reductions in meat intake with more moderate reductions in dairy
and especially egg consumption. Increased cereal intake compensates in terms of calories. This
pattern of commodity consumption was used as a template for identifying realistic dietary
measures involving significant reductions in livestock products. This was used to formulate
measures resulting in a 50% and 66% reduction in the consumption of livestock products. The
66% reduction option was chosen because the pattern of commodity use in Turkey can be used
FAO. 2004. Country profiles. Statistical Yearbook.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
to identify possible consequences. The 50% reduction measure was derived from this to
examine the consequences of halving livestock product calorie intake.
This gives the following changes:
UK diet with a 50% reduction in livestock product intake:
% consumption compared to current diet
Ruminant meat 30
Milk 60
Butter 60
Cheese 60
Eggs 90
Demersal fish 27
Poultry 40
Pig meat 40
Other animal fats 30
This gives a total livestock product calorie intake of 526 compared with 1,056. This is
compensated by increasing cereal, fruit, pulse, potato, vegetable and vegetable oil consumption
by 21%.
UK diet with a c. 66% reduction in livestock product intake compared with current diet:
% consumption compared to current diet
Ruminant meat 20
Milk 50
Butter 33
Cheese 20
Eggs 66
Demersal fish 9
Poultry 33
Pig meat 20
Other animal fats 20
Offal 33
This gives a total livestock product daily calorie intake of 359 compared with 1,056,
compensated for by a 29% increase in crop production consumption.
Scenario generation
Scenarios examined the effects of combinations of production and consumption measures. This
presented a complex challenge in trying to simulate significant interactions between measures,
particularly with respect to the nitrogen cycle. In delivering insight into the scope for reductions,
we opted for intermediate scenarios comprising combinations of measures around particular
themes. These are as follows:
Non-mobile energy reducing GWP from the fuel input to non-mobile equipment that
typically use electricity or gas, such as ventilation and cooking. Typically this would
comprise use of renewable energy for electricity or nuclear power, with a shift from gas to
electricity in food preparation;
Mobile energy” – reducing GWP from the fuel input to mobile equipment that typically use
diesel and also GWP from fertiliser production from gas. Typically this would involve
replacing diesel with hydrogen or electric engines in vehicles and a new method of fertiliser
production using electricity not gas;
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Direct GHG emissions” – directly reducing direct emissions of GHGs to the atmosphere:
refrigerants, methane, nitrous oxide. Typically this would be non GHG refrigerant gas and
techniques for reducing methane emitted by ruminants;
Production efficiency reducing GWP by reducing waste, increasing food conversion
efficiency and crop yields, and reducing the energy required in the production processes of
Consumption” – changing consumption;
Conservation” – recycling and avoiding wasteful use.
Combining measures
The general procedure was applied when quantifying the emissions reductions in themes (i.e. a
set of measures). The theme was determined, e.g. non-mobile energy. The primary measures
associated with it were identified. The potential for emissions reduction with the primary
measures were estimated using expert judgement. This included an assessment of technical
feasibility, cost and societal acceptance. These were made over time up to 2100 and quantified
on the basis of the percentage of the total possible reduction achieved over time. The
secondary measures were similarly quantified, but with a lower rate of uptake to reflect the
higher importance of the primary measures. Interactions between measures were carefully
scrutinised to avoid any double counting. Thus if the consumption of livestock products was
reduced, then the potential for savings from methane emissions was also reduced.
The scenarios indicate potential tracks which would result in an eventual 70% saving in
emissions. There are an infinity of possible combinations of themes which can be constructed to
achieve 70%, but equally there is no single theme which can.
Our consumption based scenario focuses on livestock products. In addition to reducing
emissions directly, less meat consumption and production could mean reduced emissions of
GHG from arable land as more land would be available for crops for human consumption which
could then be grown with less fertiliser-N giving further reductions in N
O emissions. However,
complete removal of livestock products is an extreme option which is not realistic and presents
very significant nutritional challenges. So, consumption options other than vegetarianism or
veganism were considered in developing the consumption based scenario. The role of meat,
dairy, eggs and fish, out-of-season and refrigerated products was examined. This included for
example examination of the effect of replacing one type of meat with another. A simple scenario
analysis indicated that the substitution of beef and lamb through increasing poultry and pigmeat
consumption would lead to a reduction in the direct GHG emissions from primary production of
about 6 Mt CO
e. However, such a simple analysis based only on our existing LCA results is
inadequate in estimating the full effects of such a change. To more fully quantify rigorously the
potential impacts of such a change, the emissions from changing land use, e.g. tilling
grasslands to produce cereals for pig and poultry feeds, need to be estimated as well as the
effects of increased soy consumption. In addition, long-term changes to N inputs also need to
be taken into account and a proper net GHG budget prepared. For example, while CO
emissions from soil will increase following conversion of grassland to arable, the availability of N
from soil organic matter will lead to reduced emissions of N
O from N fertiliser application.
The land resource based food chain was one approach used to configure a scenario that
viewed livestock as a means to utilise resources not suitable or needed for the production of
plant products. Ruminants are fed only on the grass grown on the land not suitable for crops,
while no crops are grown solely for consumption by pigs and poultry. In this scenario, land
currently used directly or indirectly for livestock farming could be freed up for other purposes,
such as carbon sequestration. This is a complex scenario requiring detailed study to elicit an
accurate assessment of potential reduction in GHG emissions.
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Current emissions from primary production – up to the Regional Distribution Centre
LCI values for commodities to the RDC used in this study. The descriptions are those used by
the FAO. Note that feed crops such as feed wheat are not shown. These have already been
used in the calculation of the LCI of animal products. Results for all commodities are provided in
Table 13. These are condensed to commodity categories in Table 14.
Table 13. Greenhouse gas emissions (CO
e/kg) from the production of commodities in the UK,
the rest of Europe (RoE) and the rest of the world (RoW) for direct UK consumption
kg CO
e/kg commodity
Almonds 0.88
Anise, badian, fennel etc. 1.41
Apples 0.32 0.43 0.88
Apricots 0.43
Artichokes 0.48
Asparagus 1.94 2.22 2.39
Avocados 0.43 0.88
Bananas 1.33
Barley 3.24 3.35
Beans (incl. cowpeas), dry 0.61
Beans, green 1.55 10.70
Beef 12.14 12.26 32.00
Cabbages, other brassicas 0.22 0.48 0.64
Carrots and turnips 0.35 0.46
Cashew nuts 1.06
Cauliflowers and broccoli 1.94 2.22 2.39
Misc. cereals 0.37 0.49
Cherries 0.32 0.43 0.88
Chestnuts 0.43
Chickpeas 0.77 0.80
Chicken meat 2.84 2.95 2.60
Chillies and peppers, dry 1.30
Chillies and peppers, green 5.88 3.12
Cinnamon (canella) 0.87
Citrus fruit, misc. 0.51
Cocoa beans 0.74
Coconuts (incl. copra) 1.78
Coffee, green 8.10
Cranberries, blueberries 1.39
Cucumbers and gherkins 3.79 1.30
Currants and gooseberries 0.84
Dates 0.32 0.88
Eggplants (aubergines) 1.30
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Eggs 2.94 3.04
Figs 0.43
Misc. fruit 0.43 0.88
Garlic 0.57 0.68
Ginger 0.88
Grapefruit and pomelo 0.51 0.70
Grapes 0.42 0.75
Grapes as wine 0.65 1.08
Groundnuts 0.65
Guavas, mangoes etc. 1.78
Hazelnuts 0.43 0.88
Kiwi fruit 0.43 0.88
Misc. leguminous veg. 1.55
Lemons and limes 0.51
Lentils 1.06
Lettuce and chicory 1.15 1.00 10.00
Maize 0.45
Milk, whole, fresh 1.19
Millet 0.47
Mushrooms and truffles 1.00 1.11
Natural honey 1.00 1.00 1.00
Nutmeg, mace and cardamoms 0.87
Misc. nuts 0.88
Oats 0.38 0.12
Misc. oilseeds 2.20
Olives 3.66
Onions (inc. shallots) 0.37 0.48
Oranges 0.51
Other melons (incl. cantaloupes) 1.55 1.74
Palm nuts-kernels (nut equiv.)/Oil 2.23
Papayas 0.88
Peaches and nectarines 0.43 0.88
Pears and quinces 0.32 0.43 0.88
Peas, dry 0.51 0.62 0.15
Peas, green 0.29 0.40
Pepper (Piper spp.) 0.87
Pig meat 4.45 4.56
Pineapples 1.78
Pistachios 0.88
Plantains 1.33
Plums and sloes 0.32 0.43 0.88
Potatoes 0.26 0.51
Pumpkins, squash and gourds 2.22
Rapeseed and mustard seed 2.09
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Raspberries and other berries 0.84 0.95 1.41
Rice, paddy 3.50
Rye 0.38 0.49
Sesame seed 1.05
Sheep and goat meat 14.61 12.00
Sorghum 0.47
Soy oil 0.77 0.80
Spices 0.87
Spinach 2.22
Strawberries 0.84 1.06 1.39
Sugar beet 0.10
Sugar cane and misc. sugar crops 0.09
Sunflower seed 2.20
Tangerines, mandarins etc. 0.51
Tea and Maté 0.87
Tomatoes 3.79 1.30
Turkey meat 3.76 3.87
Walnuts 0.88
Watermelons 1.33 1.33
Wheat – Milling 0.52 0.63 0.66
Yams 0.88
1 One composite number for fish
2 Values for a few commodities, such as milk, actually extend to retail
Table 14. Greenhouse gas emissions from the primary production of food for consumption in the
UK – up to the RDC
Food category kt CO
Red meat 19,400
Milk 17,200
White meat 10,900
Cereals, including for brewing and distilling 9,750
Vegetables & legumes 5,380
Oil-based crops 4,060
Salad crops 3,580
Fish 2,780
Grapes & wine 2,610
Temperate & Mediterranean fruit 2,220
Rice 1,860
Exotic fruit 1,780
Eggs 1,650
Sugar 1,200
Beverages 1,180
Nuts 254
Misc. including spices 79
Total 85,883
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Gas emitted (% of total GWP from primary production)
Carbon dioxide CO
Nitrous oxide N
O (24%)
Methane CH
Livestock product components account for 61% of direct primary production emissions while
serving about one third of calorie intake.
Table 15. Location of GHG emissions from the primary production of commodities for UK
UK RoE RoW Total
GHG for all commodities (kt CO
e) 56,400 15,500 13,600 85,500
Proportion from regions 66% 18% 16% 100%
Current emissions from processing, distribution, retail and food preparation (post
regional distribution centre)
A summary of the inventory of emissions from the processing, distribution, retail and preparation
of food for UK consumption is provided in Table 16.
Table 16. Greenhouse gas emissions from processing, distribution and retail for consumption in
the UK – after the regional distribution centre
consumption, kt
e / year
Eating out, kt
e / year
Total, kt CO
/ year
Cooking 11,100 4,410 15,510
Manufacturing 12,200 2,720 14,920
Food storage energy 11,200 2,170 13,370
Refrigerants 4,630 1,270 5,900
Electricity 4,530 1,090 5,620
Landfill of food waste 2,550 928 3,478
Washing-up 1,970 257 2,227
Road fuel & oil 1,380 271 1,651
Travel to outlet 1,330 113 1,443
Packaging 719 136 855
Landfill 488 155 643
Carrier bags and take-away containers 391 51 441
Food storage refrigerants 61 180 241
Total 52,549 13,751 66,300
Gas emitted as % of total GWP
Carbon dioxide CO
Nitrous oxide N
O 0%
Methane CH
Refrigerants 9%
How low can we go? An assessment of greenhouse gas emissions from the UK food system and the scope to reduce them
by 2050
Land use change emissions
Land use change emissions - background data
Table 17. Summary of global land use and LUC data used
Total world agricultural area (for comparison) 4,946 Mha
Total world arable and permanent crop area 1,244 Mha
Total pasture area connected to world trade 2,232 Mha*
Total agricultural land area – excluding non-commercial pasture 3.475 Mha*
Total world area used for commercially traded livestock (pasture and crops) 2,710 Mha*
Total world area used for directly consumed crops 765 Mha
Total UK land requirements for directly consumed crops (food only) 7.469 Mha
UK land requirement for directly consumed crops as % of total food crop land 0.98%
UK population as % of world population 0.9%
Total LUC emissions attributable to commercial agriculture 5 GtCO
% of LUC emissions attributable to UK food consumption 2.1%
* See Table 19.
Screened as set out in methods – production from countries accounting for more than 0.5% of world trade
AND production
Land use change emissions attributable to crops for direct human consumption
Table 18. Arable land and crop (directly consumed by humans) commodity consumption data in
emission calculations and the associated estimated LUC emissions
per tonne of
Emissions per
tonne of food
of food