2. The environmental impacts of food products: introduction to lifecycle assessment

This chapter describes lifecycle assessment (LCA), the most common method used for quantifying the environmental impacts and resource use throughout the entire life cycle of a product or service. Carbon footprinting – a subset of LCA in which only the climate impact of a product is assessed – is also discussed. The chapter looks at when and how an LCA method can be applied, the complexities and uncertainties involved in performing LCA for food products and interpreting their results, as well as the benefits and limitations of the LCA approach.

We would like to acknowledge Dr Elin Röös for generously contributing her LCA expertise to much of this chapter.

2.1 Lifecycle assessment (LCA): quantifies environmental impacts from cradle to grave of a product

2.1.1 Overview of LCA aims and process

LCA studies, following the methodology framework set by ISO14040, involve 4 phases: goal/scope definition; inventory analysis; impact assessment; and interpretation. Each phase is described in more detail in subsequent slides, but an overview follows here.

LCA: a structured way of quantifying the environmental impact and resource use

Adapted from Hellweg & Milà i Canals (2014)

 

In Phase 1, the functional unit (the product, service or system upon which the LCA is based) is defined. This could, for example, be 1kg of pork. The system boundaries are then decided; they may start with the production of the raw materials required on a farm and end with purchase of 1kg of pork at a supermarket.

In Phase 2, an inventory is made of all the inputs and outputs, resource use and emissions generated for the functional unit at each stage within the system boundaries.

In Phase 3, emissions and resource use are grouped together into specific environmental themes or impact categories, such as climate change, eutrophication, and acidification. These are referred to as mid-point indicators. Further modelling of how these impacts will affect human health, ecosystem quality and natural resource use (end-point indicators) can also be performed. Mid- and end-point indicators are discussed later in this chapter.

In Phase 4, sensitivity and uncertainty analysis are performed to determine the robustness of the results. Conclusions and recommendations are made. An example might be the recommendation of one manufacturing technique over another, or the identification of life cycle hotspots for the functional unit being studied.

One challenge within LCA, if the purpose is to compare different foods, is whether different studies use the same system boundaries, measure inputs and outputs in a consistent manner, and in how these results are interpreted. Increasing standardisation of methodologies, such as PAS 2050 (designed for carbon footprint analyses), is designed to reduce inconsistencies and allow comparison between research. This is important, especially within the food system, because of the many different production techniques and variables involved (see later in this chapter for more on this).

2.1.2 Phase 1: goal and scope definition

Phase 1: Goal and scope definition

Adapted from Hellweg & Milà i Canals (2014)

When measuring the environmental impact of a product, service or system, we must initially define the scope of the assessment. Life cycle assessment ideally considers all stages involved in the production of a product or service (sometimes termed “from cradle to grave”, or “field to fork”).

For example, the LCA scope for a loaf of bread could include producing raw materials for farming, production on the farm, harvesting and processing the crop, manufacturing the bread, delivery and storage at retail, impacts from consumption (including toasting), and transport and waste impacts all along the life cycle.

Some LCA studies deliberately limit their scope in order to focus on the impacts of a particular stage. An example might be an assessment of different production techniques for certain crops, in order to understand how different farming techniques might affect environmental outcomes.

In practice, as supply chains are often very complex, some inputs that are known to have a very small contribution to the overall environmental impact are sometimes ignored, and it is often a matter of judgement where these boundaries are set.

We also need to define what output it is whose impacts we are measuring – e.g. one kg of product, or 100 calories of product, or 1ha of farmed land and so forth. This is called the functional unit. The impacts – e.g. the quantity of carbon dioxide equivalent CO2 eq. emissions emitted in delivering this functional unit are expressed as a ratio – e.g. 0.5kg CO2 eq./kg product. The output can also be converted and expressed as CO2 eq. per average or recommended portion.

Choice of functional unit can influence results

The functional unit is the unit that is being studied. It can be an individual product (e.g. 1kg of pork or the packaging of 1 litre of soda) or a service – e.g. (to take a non-food example) delivery of a week’s supply of washed cloth nappies.

LCAs for food products normally use weight of finished food products as the functional unit (e.g. 1 kg of pork), but also use units of area (for example, the environmental impact of wheat production in terms of hectares used) or nutritional units (for example, the environmental impact of producing 100g of protein from a food product). The choice of functional unit could influence comparison between foods. For example, egg usually has a lower carbon footprint than poultry meat if the functional unit is 1 kg of product (egg or poultry meat) but if the functional unit is 1 kg of protein instead the carbon footprint is more similar between the two as eggs only contain 12% of protein compared to 20% in meat.

The choice of the functional unit should reflect the purpose of the LCA-study. For example, if the purpose is to compare the production of pork using different animal feed inputs, a relevant functional unit could be 1 kg of pork meat at the farm gate. For another study comparing different ways of securing protein in our diet (e.g. via meat, fish, dairy, legumes etc.), a functional unit of 1 kg of protein might be better. However, as food purchases are seldom made on the basis of the nutrient content of food, a functional unit of 1 kg is often very relevant.

The relative impacts of different products could change, depending on which functional unit is used.

2.1.3 Phase 2: assessing inputs and outputs

Phase 2: Inputs and outputs for all stages are included

Adapted from Hellweg & Milà i Canals (2014)

Data are collected for inputs and outputs at each stage within the scope of the LCA. Inputs might be energy, land and water use; outputs are the products and waste leaving the system at each stage.

Emissions are estimated for each of these, for example CO₂ emissions for the energy use to produce the chosen functional unit. The data used for this purpose may be source data (such as the measured energy used in a factory process), i.e. specific to the focus of the LCA. Alternatively the data may be more generic and drawn from existing databases which contain average impacts arising. Clearly, the more specific the data the more accurate the results. Whatever the source of the data, the data need to be adjusted to the functional unit of the LCA.

Production systems often:

  • Produce more than one product
  • Contain the use of recycled material

Allocating emissions and inputs between products is important but can be difficult

If a factory manufactures three different products, one of which is the functional unit of an LCA study, then the energy use in that factory needs to be allocated appropriately to the functional unit being studied (maybe 1/3 of energy use), rather than the total energy used by the factory as a whole. Sometimes allocation is a complex process. For example a dairy system produces not just milk but also meat (male calves, dairy cows no longer lactating). Some of the overall herd emissions need to be allocated to this meat production – but on what basis? In terms of their relative economic value? In terms of the calories or protein they supply as compared with the protein or calories supplied by the milk? Different studies adopt different approaches.

All of the data included contributes to the final LCA results. Emissions and resource use at all stages that lie within the defined scope of the LCA need to be included because they all contribute to the environmental impact of the product or service in question.

How do we deal with the problem of allocating emissions to products?

  1. Avoid allocation altogether
    If possible it is preferable to avoid having to divide emissions between co-products by, for example, expanding the system boundaries to include all products produced in the system – for example by including both the milk and the meat produced in a dairy system in the functional unit. However, if the purpose of the study is to compare milk with some other beverage emissions needs to be split between the milk and the meat.
  2. Substitution
    One way of handling this is to consider alternatives to the co-products on the global market based on the reasoning that when a co-product of a production system enters the market the production of some other product with the same function can be ‘avoided’. For example, the beef produced in the dairy system would avoid production of other beef meat e.g. from a suckler cow system or meat from pork or poultry. The emissions from the production of milk is then calculated as the total emissions from the dairy system minus the emissions from the production of the alternative ‘avoided’ product, in this case the beef.
  3. Allocate based on physical relationships (e.g. mass or energy) or economic relationships
    Another simple way of doing this is to allocate by mass or economic value. For example if an economic allocation method is used then the relative economic value of the co-products is used as the basis for allocating impacts – e.g. the relative value on the market of milk and meat. The precise impacts will vary as the market fluctuates.

2.1.4 Phase 3: assessment of impacts and allocation to indicators

Phase 3: Impacts are assessed and allocated to indicators

Adapted from Hellweg & Milà i Canals (2014)

The total emissions and resource use from all life cycle stages are grouped into environmental themes or categories (mid-point indicators) according to the impact they cause, and are converted to common impact units. Examples of impact categories are climate change, eutrophication potential, water use and land use. Effects that these impacts can have on humans and eco-systems may also be modelled. Results are then presented using what are called end-point indicators; these refer to broader issues of concern such as human health, biodiversity/ecosystem services. The end-point indicators are fewer in number than the mid-point indicators and therefore might be easier to interpret. However, the modelling of end-point effects is associated with great uncertainties.

The results can be shown for each different impact category, or presented in relation to particular stages in the product’s life. Thus, the water use for producing 1kg of bread can be shown for the whole life cycle and for each stage within that (such as farming, manufacture etc.).

LCA does not just measure GHGs

LCA can be used to measure many different environmental impacts, one of which is the global warming potential or greenhouse gas emissions (GHG). Other impacts include:

  • Ozone layer depletion
  • Acidification
  • Eutrophication
  • Photochemical smog
  • Ecotoxicity
  • Land use
  • Water use
Holistic assessment  avoids burden shifting,  whereby improving one  metric might adversely  affect another.

Measuring more than one factor in this way means we are able to see wider environmental impacts, and can avoid unintended negative consequences from reductions in one specific impact (sometimes called “burden shifting”).

For example, if an intervention is designed to reduce global warming impact, but as a consequence this increases land-use requirements, then it is important to know this.

These impacts can be further allocated to general impact categories (related to actual physical impacts) concerning human health, ecosystem quality, climate change and resource use.

Environmental impacts can be presented as Mid- & end-point indicators

Mid-point indicators represent impacts as specific environmental themes.

End-point indicators are categories of impact contributed to by mid-point indicators, representing impacts on broader issues of concern such as human health and ecosystem quality. For example depletion of the ozone layer and increases in photochemical smog (presented as mid-point indicators) can both contribute negatively to human health; high land-use and water requirements place a burden on resources.

Modelling end-points thus allows LCA studies to communicate the real impacts on the environment in a less abstract way. However not all LCA studies specifically take this extra step due to the difficulties in modelling end-point effects and the increased uncertainties associated with this step.

2.1.5 Phase 4: Interpretation of results

The final phase is the interpretation of the results.

Adapted from Hellweg & Milà i Canals (2014)

This can include drawing conclusions about the environmental impact of the product being studied, and identifying hotspots both in relation to life cycle stages (such as the stages that have the highest environmental impacts), impact categorisations (which environmental impacts are most of concern as a result of producing this product), or both (which stages contribute the most to the key environmental impacts).

Results are also analysed and tested to see how robust they are using sensitivity analysis and/or uncertainty analysis (i.e. use of different input data, comparison with model data, choice of system boundaries, functional units, allocation principles etc.).

Environmental hotspots can be defined, e.g. life cycle stages with high impact, and conclusions are drawn about the overall impact of the functional unit being assessed.

Recommendations can also be made to reduce environmental impacts, based on the hotspots identified, and may be placed in the context of wider environmental challenges. For example, the GHG of a product can be put in the context of wider climate change concerns.

LCA allows us to identify environmental hotspots

Source: Basset-Mens & van der Werf (2005) 

 

This example illustrates the relative impacts on eutrophication, climate change, acidification, terrestrial ecotoxicity, energy use, land use and pesticide use from different pig farming systems. It is clear that the production of feed is the hotspot in all impact categories except for acidification.

References: 

General references

  • Dr Elin Röös, LCA researcher, personal communication

Further reading:

2.1.1

2.1.2

Further reading:

2.1.3

Further reading:

2.1.4

Further reading:

2.1.5

2.2 Lifecycle assessment of food products is a complex undertaking.

All LCA research faces common challenges when defining the scope of the life cycle to be measured, what functional unit to use, how to deal with changes over time and problems associated with allocation of shared inputs and outputs. However, the agriculture stage presents unique complexities for LCA, as defining the system boundaries is particularly difficult.

LCA Challenges

Source: FCRN and Dr Elin Röös.

For example, how should the emissions from deforestation of an area be allocated to the products that are grown on this land when arguably all agricultural production contributes to deforestation: even production on existing agricultural land, by preventing another crop being grown could be seen as responsible for ‘driving’ that production onto forest land?

  • There may be multiple food and non food drivers of deforestation (e.g. both cattle ranching and logging)?
  • Additionally, direct emissions are usually not easily measured; consider for examples the difference between measuring the emissions from a factory chimney and measuring the emissions from hectares of agricultural land. Therefore, emissions usually have to be modelled rather than measured, a step which is associated with great uncertainties. Furthermore, emissions from agriculture vary widely depending on geographical location, production method, temperature, soil and rainfall patterns, and more. Generic data are therefore often used, so reducing the accuracy of the findings. To add to the complexity, soils can act as greenhouse gas sources (emissions) or sinks (absorbing emissions) and both land-use and agricultural practices can change over time.

In addition, quantifying emissions from animals (and their manure) is also difficult because of a large number of variables involved including breed differences, quantifying methane emissions from enteric fermentation, different feeds and production systems, different impacts of different manure treatments etc.

However, the scope of LCA usually encompasses far more than the agricultural stages of the food supply chain, and all stages can present difficulties with respect to scope. For example, food products are processed, often refrigerated, transported and stored for different times. Food can be prepared in different ways (e.g. a carrot can eaten raw, microwaved, boiled or roasted). The impacts associated with all these stages need to be understood and quantified.

As such, there is a considerable degree of uncertainty when measuring the environmental impacts of food.

References: 
  • Dr Elin Röös, LCA researcher, personal communication

Further reading:

2.3 Carbon footprinting is based on the LCA approach but focuses only on greenhouse gas emissions.

2.3.1 Defining terminologies.

GHG = greenhouse gases

IPCC provide a list of gases that contribute to global warming. The most important for food production are carbon dioxide, methane, nitrous oxide and refrigerant gases

Expressed in terms of 20, 100 or 500 years:

Methane = 34 x more potent than CO2 (over 100 year time-period)
Nitrous oxide = 298 x more potent than CO2 (over 100 year time-period)
Refrigerant gases = thousands x more potent than CO2


Most discussion about GHGs outside the food and agricultural sector focuses on carbon dioxide (CO2). However, there are other greenhouse gases too, such as methane and nitrous oxide and the food system is a particularly significant source of these.

The Kyoto Protocol includes 6 main GHGs : carbon-dioxide (CO₂), methane (CH4), nitrous oxide (N2O), hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), and sulphur hexafluoride (SF6)

Different GHGs have different potency – that is, the strength of their contribution to radiative forcing (a measure of global warming) differs.

Measured over a time-period of 100 years, the global warming potential of methane is 34 times more potent than carbon dioxide; nitrous oxide is nearly 300 times more so.

The total global warming impact can be expressed in aggregated form, known as carbon dioxide equivalent (CO2 eq.). This is often also referred to as the carbon footprint (CFP), global warming potential (GWP), or simply greenhouse gases (GHG). Throughout these chapters, the term GHG is used. There are also other climate change metrics such as the Global Temperature Potential (GTP).

2.3.2 GHGs arising from the food system.

The most important GHGs arising from the food system are methane (CH4), nitrous oxide (N2O) and carbon dioxide (CO2).

Agricultural emissions are dominated by methane and nitrous oxide, although land-use change for agriculture is a major source of carbon dioxide emissions.

Post-production stages (such as food processing, transport, storage, refrigeration) give rise to notable carbon dioxide emissions, due to more significant energy and fossil-fuel use.


The food system is different from many of the sectors and activities (such as construction, heavy industry and transport) because agricultural emissions are dominated by methane and nitrous oxide. In industrial processes carbon dioxide arising from energy use tends to dominate.

This said, agriculturally induced land use change (for example deforestation, or conversion of peatlands to crop production) leads to very significant releases of carbon dioxide.

References: 

General references

  • Dr Elin Röös, LCA researcher, personal communication

Further reading:

  • Weidema, B.P., Thrane, M., Christensen, P., Schmidt, J., and Løkke, S. (2008) Carbon Footprint. Journal of Industrial Ecology 12 (1), 3–6

2.3.1

  • Myhre, G., Shindell, D., Bréon, F.-M., Collins, W., Fuglestvedt, J., Huang, J., Koch, D., Lamarque, J.-F., Lee, D., Mendoza, B., Nakajima, T., Robock, A., Stephens, G., Takemura, T. and Zhang, H. (2013). Anthropogenic and Natural Radiative Forcing. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
     

2.3.2

Food system related GHG emissions are covered in detail in Chapter 3.

2.4 The value and limitations of lifecycle assessment.

2.4.1 Values and limitations of LCA.

  • LCA enables a more complete and holistic understanding of the impact of food at all stages in its production.
  • It allows analysts to identify stages, or practices that generate particularly high negative impacts and potential solutions to be identified.
  • It also reveals the trade offs that may exist across areas of environmental concern, and that difficult decisions may need to be made.
  • Inconsistency in methodological choices (e.g. of functional units and system boundaries) make comparisons across studies difficult. There is however increasing focus on encouraging standardisation of methodologies.
  • An LCA may only give an indication of the environmental impact associated with the average or industry standard production of a product – it may not give an indication of the best case outcome of a particular food product (i.e. novel production systems with vastly reduced environmental impacts may exist for a particular product, but this may or may not be taken into account in any given LCA study).
  • LCA has historically focused on environmental, not social impacts although attempts are being made to develop a methodology for social LCAs.
  • LCA is a useful but incomplete tool for measuring of sustainability.
  • LCA can show impacts but it doesn’t necessarily tell us ‘what to do.’

2.4.2 LCAs have historically focussed on environmental impacts.

Currently, the majority of LCA studies focus on environmental impacts and do not concern themselves with social impacts.

As such, the impacts on employment, health, well-being, community benefits and human rights are not measured or represented (although the use of end point indicators gestures in this direction).

Introducing social and economic issues adds complexity. For example GHG emissions arising from a certain food producing activity may be high, but may also support many people’s livelihoods, have positive or negative health consequences, or represent value to communities in other ways.

LCA methodologies and standards are beginning to evolve in an attempt to incorporate measurements of social impacts into assessment.

2.4.3 Can LCAs determine whether something is sustainable?

  • No, because LCA does not consider the number of users or needs and the broader value or importance of any given product or activity.
  • LCA quantifies impacts but does not give “sustainability” thresholds.
  • LCA cannot address broader societal questions about the impacts of consumption and consumerism in society in general.

However

  • LCA results are indispensable as an aid to making assessments about sustainability.

The LCA measurement of a product is useful when comparing different products with a similar function, but in the context of sustainability this needs to be estimated in terms of how much of the product is used in absolute terms, what the trends are in its usage, what value it has for society and other social and ethical aspects of its production or use.

For example, if the GHG of a certain food product is very high per kg, but consumption levels are very low, then its overall impact might be lower than a product with lower GHG per kg but which is consumed in large quantities. Similarly, if a food product has a high nutritional value then this should be considered within the perspective of society’s needs. Certain animal production systems may have higher GHG impacts than others, but the welfare of the animals raised may be better than those reared in lower GHG-impact systems.

References: 

General references

  • Dr Elin Röös, LCA researcher, personal communication

2.4.1

2.4.2

For more on Social Life Cycle Assessment, see: UNEP/Setac Life Cycle Initiative (2009) Guidelines for Social Life Cycle Assessment of Products. vol. 15

2.4.3

For a more detailed analysis of different perspectives on the usefulness of LCA, see Garnett, T. (2014) Three Perspectives on Sustainable Food Security: Efficiency, Demand Restraint, Food System Transformation. What Role for Life Cycle Assessment? Journal of Cleaner Production 73, 10–18 (in press article available open access here)

  • Vergnaud, A.C., Norat, T., Romaguera, D., Mouw, T., May, A.M., Travier, N., Luan, J., Wareham, N., Slimani, N., Rinaldi, S., Couto, E., Clavel-Chapelon, F., Boutron-Ruault, M.C., Cottet, V., Palli, D., Agnoli, C., Panico, S., Tumino, R., Vineis, P., Agudo, A., Rodriguez, L., Sanchez, M.J., Amiano, P., Barricarte, A., Huerta, J.M., Key, T.J., Spencer, E.A., Bueno-de-Mesquita, B., Büchner, F.L., Orfanos, P., Naska, A., Trichopoulou, A., Rohrmann, S., Hermann, S., Boeing, H., Buijsse, B., Johansson, I., Hellstrom, V., Manjer, J., Wirfält, E., Jakobsen, M.U., Overvad, K., Tjonneland, A., Halkjaer, J., Lund, E., Braaten, T., Engeset, D., Odysseos, A., Riboli, E., and Peeters, P.H.M. (2010). Meat Consumption and Prospective Weight Change in Participants of the EPIC-PANACEA Study. American Journal of Clinical Nutrition 92 (2), 398–407
  • Godfray, H.C.J. and Garnett, T. (2014). Food Security and Sustainable Intensification. Philosophical Transactions of the Royal Society B: Biological Sciences 369 (1639)

2.5 Conclusions.

  • LCA is the most common method used for quantifying the environmental impacts and resource use throughout the entire life cycle of a product or service.
  • It comprises four stages: goal and scope definition, allocations of inputs and outputs to stages of the lifecycle, assessment of impacts, and the interpretation of results.
  • LCA of agricultural products is difficult and uncertain.
  • Carbon footprinting is based on the LCA approach but focuses only on greenhouse gas emissions.
  • LCAs have historically focussed on environmental impacts, but are increasingly incorporating societal considerations.
  • LCA is an extremely useful tool but cannot by itself indicate the sustainability of a product or process