Reporting standards and guidelines are the first building block for carbon footprints. This chapter introduces the landscape of such standards, including product carbon footprint standards, sectoral guidance, as well as product category rules. The chapter also discusses the PACT Pathfinder approach which seeks to integrate these different strands into a coherent whole.
Measuring Carbon Footprints of Agri‑Food Products
Eight Building Blocks
Abstract
Clear reporting standards and guidelines are essential to coordinate carbon footprint calculations across firms, sectors, and supply chain stages. Standards and guidelines can answer questions such as which activities to include, which methodologies to use, and which numbers to report and at what level of granularity. Without shared answers on these and other questions, each carbon footprint calculation may end up using different definitions or methodological choices, making the resulting numbers difficult to compare.
Over the last two decades, a landscape of standards and guidelines has emerged for measuring carbon footprints, including standards addressing food systems-specific issues. It is easiest to think of this landscape as a pyramid, as shown in Figure 4.1.
Source: OECD analysis.
In this figure, standards and guidelines shown on the left focus mainly on firm-level (or farm level) carbon footprints, while those on the right focus on product-level carbon footprints. Standards and guidelines at the bottom of the pyramid are more general (sector-agnostic) in scope, while those higher in the pyramid are increasingly specific – e.g. focusing on agriculture, or focusing on a specific sub-sector (dairy, beef, horticulture). The ultimate goal of these standards and guidelines is to support consistency in emission measurement and communication, including by ensuring consistency of calculation tools and emission factor databases, as indicated at the top of the pyramid. In the middle of Figure 4.1.
is the PACT Pathfinder initiative which explicitly aims to bridge across the different standards and guidelines, both by connecting product-level carbon footprints to firm-level reporting (for Scope 3 purposes) and by harmonising guidance across different sectors (PACT, 2023[1]).
As is clear from Figure 4.1, the two main standard setters are the Greenhouse Gas Protocol (GHG Protocol) and ISO. Both organisations have standards for firm-level (organization-level) reporting as well as product-level reporting. These standards are fairly similar. In practice, firm-level reporting commonly uses the GHG Protocol Corporate standard (which was the first of its kind when it was published in 2001) while product-level carbon footprints often use the ISO 14067 standard. An older product-level carbon footprint standard, PAS 2050, is also sometimes used.
This chapter asks to what extent existing standards provide sufficient guidance to allow a system of reliable and widespread product carbon footprints. The focus is therefore on product carbon footprint standards, although some of the other standards in Figure 4.1 will be discussed as well, when they provide relevant guidance. For example, sectoral guidance which has been developed originally for firm-level (corporate and Scope 3) emissions can be useful for product-level carbon footprints as well.
Product carbon footprints can be seen as a specialised form of life cycle assessment (LCA) (Hauschild, Rosenbaum and Olsen, 2018[2]) (Cucurachi et al., 2019[3]). The basic principles of LCA are defined in the widely used ISO 14040 and 14044 standards. The ISO 14067 standard for product carbon footprints builds on these standards and is designed to be used in conjunction with them.
In addition to the ISO 14067 standard (originally introduced in 2013, last updated in 2018, and currently undergoing revision), other product carbon footprint standards exist, notably the GHG Protocol Product Life Cycle Accounting and Reporting standard (introduced in 2011) and the PAS 2050 standard (which was the first standard on product carbon footprints, developed by the British Standards Institute in 2008 and last updated in 2011). The standards have many similarities. Whatever differences exist have limited practical relevance nowadays, as the ISO 14067 standard is now the most widely used (see Box 4.1 for a discussion of some differences). The discussion here touches on the main methodological choices and underlying principles in product carbon footprint standards, focusing mostly on the ISO 14067 standard.
A first important distinction is between an attributional and consequential approach to life cycle assessment (and hence product carbon footprints). Attributional LCA is essentially a “snapshot” at a point in time of the flows that can be ascribed to a given product or system, whereas consequential LCA asks how these would change if for example output was increased by one unit. In the context of land use change, an attributional LCA would ask whether any land use change occurred in the life cycle of the product (a concept known as direct land use change), whereas a consequential LCA would ask whether an expansion of output would, through economic and behavioral feedback and substitution effects, lead to land use change (a concept known as indirect land use change). The ISO 14067 standard can accommodate both attributional or consequential approaches, but firm-level and product-level carbon footprints typically take an attributional lens (GHG Protocol, 2022[4]); the GHG Protocol Product Standard even requires it (GHG Protocol, 2011[5]).1
A second important choice regards the functional or declared unit – i.e. the “denominator” of the LCA or carbon footprint. Environmental impacts could be expressed in terms of physical units of output, e.g. per liter of milk at the farm gate; this is referred to as a “declared unit” approach. But impacts could also be expressed in terms of the functions those products or systems fulfill, such as the nutrient content of food, e.g. environmental impacts per 100g of protein. LCAs and carbon footprints are often expressed in terms of such functional units, an approach generally favoured by the ISO 14067 standard to ensure meaningful comparisons. However, such an approach is not ideal if the goal is to transmit carbon footprint data along the supply chain. Since the supply chain involves a transfer of physical products, these should be the relevant denominator. Expressing emissions in terms of declared units also reduces the scope for confusion or incompatibility when the relevant functions of a product can be defined in different ways.2
The definition of the product system and system boundaries determines which activities (and hence impacts) are in scope of the assessment and which ones are out of scope. For example, in assessing the environmental impacts of milk, this would include a decision on whether the production of fertilisers used in growing animal feed is part of the scope or not. As the term “life cycle assessment” implies, a full LCA (and hence a full product carbon footprint) should include all relevant stages of the life cycle, from raw material extraction to the end-of-life stage (e.g. waste management, landfill); this is also known as a cradle-to-grave approach. But as noted earlier, to scale up the measurement and communication of carbon footprints in supply chains it is often more practical to focus on cradle-to-gate approaches, where each actor accounts for life-cycle emissions up to the point where the product leaves its premises. Existing standards foresee this possibility.
Another aspect of defining the system boundaries is the use of cut-off criteria for excluding certain processes, inputs, or outputs. For example, the ISO 14067 standard notes that emissions related to the production of capital goods (such as the emissions involved in the production of a tractor) can be excluded if this would not significantly change conclusions.3
The treatment of carbon offsets is another aspect of the definition of system boundaries. A carbon offset or carbon credit is a certificate purchased by an organisation through, for example, an emissions trading scheme or through funding an emissions reduction project unrelated to the life cycle of the product. The ISO 14067 carbon footprint standard (as well as the Greenhouse Gas Protocol’s Product Standard and PAS 2050) prohibit the inclusion of carbon offsets in the system boundary: the product carbon footprint should therefore represent actual emissions and removals which occurred during the life cycle of the product.
Production processes often involve multiple outputs, and conducting an LCA or product carbon footprint calculation thus requires allocation rules. In the context of dairy farming, for example, an allocation rule answers the question of how the total environmental impacts of the farm should be allocated between milk and beef outputs; in dairy processing, allocation rules are needed to allocate impacts between different types of dairy products (butter, skim milk powder, etc.). The ISO standards indicate a preference for avoiding allocation rules whenever possible. For example, an arable farmer with several crops might be able to identify which inputs were used for which crops, avoiding the need to use an allocation rule. Where this is not possible, the ISO 14067 standard stipulates that allocation should be done “in a way that reflects the underlying physical relationships”; and where this is not possible either, allocation should be done “in a way that reflects other relationships,” for example in proportion to economic value.4
The ISO 14067 carbon footprint standard also defines some overarching principles which should guide practitioners seeking to conduct a carbon footprint assessment using the standard. These include:
Relevance: The selection of data and methods is appropriate to the assessment of the GHG emissions and removals arising from the system under study.
Completeness: All GHG emissions and removals that provide a significant contribution to the product carbon footprint are included.
Consistency: Assumptions, methods and data are applied in the same way throughout the carbon footprint calculation.
Coherence: Methodologies, standards and guidance documents that are already recognised internationally and adopted for product categories are applied, to enhance comparability between product carbon footprints within any specific product category.
Accuracy: Quantification of the carbon footprint should be accurate, verifiable, relevant and not misleading, and bias and uncertainties are reduced as far as is practical.
Transparency: All relevant issues are addressed and documented in an open, comprehensive and understandable presentation of information. Any relevant assumptions are disclosed and methodologies and data sources used are appropriately referenced. Any estimates are clearly explained and bias is avoided.
The GHG Protocol’s Product Standard specifies accounting principles similar to these, as does PAS 2050. In all three standards, the principles as well as the more detailed requirements are designed to ensure the reliability of carbon footprint estimates, by reducing the room for both systematic and non-systematic error.
However, by construction the main product carbon footprint standards cannot cover all methodological questions which may arise in calculating product carbon footprints. Further guidance is therefore necessary to avoid methodological inconsistencies. Relevant sources are the PACT Pathfinder Framework, sectoral guidance, and product category rules, which are discussed below.
As noted earlier, the three existing product carbon footprint standards (ISO 14067:2018, the GHG Protocol Product Standard, and PAS 2050:2011) are quite similar. This is partly by construction: for example, GHG Protocol built on the existing PAS 2050 standard when developing its Product Standard, and this standard was in turn taken into account during the revision of the PAS 2050 standard (GHG Protocol, n.d.[6]; BSI, 2011[7]). ISO and GHG Protocol also collaborate to reduce the divergence between their standards. Nevertheless, some differences remain.
Interviews with practitioners reveal that the ISO 14067:2018 standard is currently the most widely used. Its popularity is partly explained by its membership of a broader family of standards such as the ISO 14040/14044 standards for LCA as well as ISO standards explaining how GHG statements can be verified and validated (ISO 14064-3), and standards detailing the competences required for teams which do the verification and validation (ISO 14066). Hence, the differences between ISO 14067 and the GHG Protocol and PAS 2050 standards may often not matter much in practice. However, as will be shown later, some calculation tools are still aligned with the older PAS 2050 standard rather than with the more recent ISO 14067 standard.
One area where standards differ is in the hierarchy of allocation rules proposed. All standards agree that allocation should be avoided where possible, by subdivision or system expansion. But beyond that, the standards diverge. While ISO 14067:2018 and GHG Protocol prioritise physical relationships over economic or other allocation methods, PAS 2050:2011 prioritises supplementary sectoral guidance followed by economic allocation. Physical allocation in PAS 2050:2011 thus is only possible if sectoral guidance for it exists.
Another area where standards differ is in their exclusion criteria. ISO 14067:2018 is not very prescriptive in this regard: activities or life cycle stages can be excluded if this is not expected to “significantly” alter the conclusions. The GHG Protocol standard is similarly flexible: exclusion is allowed if there is a data gap and if an estimation determines that the data would be “insignificant”. PAS 2050:2011, by contrast, provides more concrete guidance: exclusions are allowed for emission sources that would constitute less than 1% of total life cycle emissions as long as at least 95% of total emissions are accounted for.
Some other textual differences between the standards are unlikely to cause differences in the assessment results. As an example, ISO 14067:2018 states that offsets are “not allowed” while PAS 2050 and the GHG Protocol standard states that offsets are “not included”. The latter means that offsets cannot be counted as part of the product carbon footprint but can be reported separately as additional information.
Firms are increasingly expected to report and reduce their Scope 3 emissions, but quantifying these emissions is currently challenging partly due to a lack of harmonisation of methodologies and partly due to difficulties in sharing data across complex supply chains (OECD/BIAC/WEF, 2023[8]; OECD, 2024[9]). The Partnership for Carbon Transparency (PACT) Pathfinder initiative aims to tackle both obstacles.
PACT is hosted by the World Business Council for Sustainable Development (WBCSD) and works with stakeholders from different industries, as well as standard-setting bodies, reporting organisations, and industry initiatives. The vision of PACT was outlined earlier: if firms can receive accurate data from suppliers regarding the carbon footprint of purchased inputs on a cradle-to-gate basis, and if firms can add their own emissions, they can in turn provide accurate cradle-to-gate product carbon footprint data to their customers. However, realising this vision requires greater harmonisation of methodologies as well as interoperable technological solutions to transmit data along the supply chain.
To achieve greater harmonisation of methodologies for product-level carbon footprints, PACT has developed the PACT Pathfinder Framework (PACT, 2023[1]).
The PACT Pathfinder Framework first sets out a hierarchy of approaches:
When product-specific guidelines (so-called product category rules, see below) already exist, firms should prioritise these, as long as they meet certain quality criteria. In particular, product category rules should only be used if they are developed in accordance with ISO standards; if they were developed using a multistakeholder process and independent peer review; if they are applicable to the geography where the product is produced or sold; and if the product category rules are reviewed at least every five years to ensure they are up to date.
If product-specific guidelines do not exist, firms should use sector-specific rules built on recognised standards, in conjunction with the guidance in the Pathfinder Framework.
If sector-specific rules do not exist, firms should fall back on cross-sectoral standards such as the ISO 14067 carbon footprint standard, in conjunction with the guidance in the Pathfinder Framework.
Next, the Pathfinder Framework provides guidance on the scope and boundaries of calculations. This guidance explains, for example, the use of a cradle-to-gate approach based on a declared unit rather than a functional unit, as discussed above.
The Pathfinder Framework then provides more detailed guidance on how to calculate product carbon footprints. Firms should include all “attributable processes”, i.e. all processes associated with services, materials, or energy flows that become, make, or carry a product throughout its life cycle. Firms can exclude a process if this would likely represent less than 1% of the total, and if the sum of excluded processes is less than 5% of the total.
For each process, firms should calculate emissions as: Activity data (amount of activity) x Emission factor (kg GHG per unit of activity) x Global Warming Potential (kg CO2-equivalent per kg of GHG). Activity data can include a firm’s material inputs (e.g. purchased fertiliser or feed) expressed in physical quantities; energy inputs (e.g. purchased electricity); or its own production processes. To the maximum extent possible, firms should use primary activity data. For emission factors, the Pathfinder Framework similarly prioritises primary data. For purchased inputs, primary data would be obtained from suppliers; if this is not available, firms can use emission factors from secondary databases. For a firm’s own activities primary data would mean, for example, direct on-site measurement. In many contexts this is currently not feasible at scale; in that case, secondary emission factors can be used, as long as these come from high-quality databases (listed in the guidance). The Pathfinder Framework specifies that where emissions are calculated using a model that takes primary data as input (as will often be the case in agriculture), the resulting emissions estimate would also be considered primary data. For Global Warming Potentials, the Pathfinder Framework aligns with other standards in referring to the latest information provided by the Intergovernmental Panel on Climate Change (IPCC).
The Pathfinder Framework provides a decision tree on how to allocate emissions in multi-output processes:
Try to avoid allocation. What looks like a multi-output process may in fact consist of single-output processes; if such “process subdivision” is possible, it should be applied.
If this is not possible, use the allocation rules outlined in product category rules or sector-specific guidance, if these meet the requirements of the Pathfinder Framework.
If such guidance is not available, but if there is a dominant, identifiable substitute product, apply “system expansion”. This is a procedure where the carbon footprint of the product being studied is calculated by taking the total carbon footprint of the multi-output process and subtracting the carbon footprint of substitutes for the co-products (i.e. the other outputs).5
When the above is not possible, the Pathfinder Framework asks what the ratio is of the economic value of the co-products. 6 If this ratio is greater than five, then one co-product can be considered the main driver of the process, and economic allocation can be used – that is, the emissions are allocated proportionally to the economic value (e.g. revenues) associated with the different products.
If the ratio is equal to or lower than five, the Pathfinder Framework asks whether there exists an underlying physical relationship between the co-products. If so, a physical allocation method should be used.
If no physical relationship exists, allocation can be done using economic allocation or alternative approaches.
The Pathfinder Framework also contains specific guidance on how to account for emissions from land use change and emissions and removals from “land management” (i.e. agriculture and forestry). However, the Pathfinder Framework notes that this guidance will be updated to reflect the final GHG Protocol Land Sector and Removals Guidance, discussed below.
Finally, the Pathfinder Framework also contains guidance on preferred data sources, as well as requirements regarding assurance and verification and on minimum required data elements to be exchanged alongside product-level carbon footprints. This will be discussed below in the context of facilitating data flows across the supply chain.
As noted above, the Pathfinder Framework gives priority to product category rules and sectoral guidance, as long as these meet certain quality safeguards. Product category rules are discussed in more detail below; this section discusses sectoral guidance. The Pathfinder Framework prioritises sectoral guidance which is built on cross-sectoral standards such as ISO or the GHG Protocol. For food systems, the relevant guidance here includes the GHG Protocol’s Agriculture Guidance and its forthcoming Land Sector and Removals Guidance.7 These are developed to facilitate implementation of the ‘core’ GHG Protocol standards for Corporate and Scope 3 reporting.
4.3.1. GHG Protocol’s Agricultural Guidance
The Agricultural Guidance (GHG Protocol, 2014[10]) provides guidance on questions which may arise when trying to report GHG fluxes (emissions and removals) from agricultural activities. For example, when a farm’s livestock is grazing on land owned by a third party, the Agricultural Guidance clarifies how emissions should be allocated between the owner of the livestock and the owner of the land. The Guidance also discusses common challenges and solutions for the collection of activity data.
The Agricultural Guidance may be subject to change given the forthcoming Land Sector and Removals Guidance (discussed below), and may even be replaced by it. At the time of writing, however, the Land Sector and Removals Guidance was not yet officially published, and hence the Agricultural Guidance remains relevant.
An important element of the Agricultural Guidance is its description of how changes in carbon stocks (in biomass, dead organic matter, soil organic matter, and harvested products) should be accounted for, and how firms should report their GHG fluxes (i.e. emissions and removals). These are summarised in Table 4.1. The table takes the perspective of an agricultural producer, i.e. Scope 1 here refers to on-farm emissions.
The Agricultural Guidance states that fluxes should be reported for each subcategory in Table 4.1. Importantly, regarding CO2 fluxes, the Guidance requires that only CO2 emissions from land use change are reported under Scope 1 emissions, while other CO2 fluxes (emissions or removals) due to land use management, as well as CO2 sequestration due to land use change, and CO2 emissions from biofuel combustion, should be reported under a separate category for “Biogenic carbon”. The Guidance does not require firms to report separately on different non-mechanical sources (e.g. enteric fermentation, manure management).
Note: This table illustrates the requirements and minimum, best practice recommendations for disaggregating agricultural GHG flux data in inventories. Please note that the proposed Land Sector and Removals Guidance would include important changes to these requirements, as discussed below.
Source: GHG Protocol (2014[10]).
4.3.2. GHG Protocol’s Land Sector and Removals Guidance
GHG Protocol is currently also preparing a Land Sector and Removals Guidance. A Draft for Pilot Testing and Review was published in September 2022 (GHG Protocol, 2022[4]) and, following feedback from pilot testers and stakeholders, is currently being refined. The Guidance would apply to all firms which have “land sector” activities (e.g. agriculture, forestry) in its operations or in its value chain, and would make Scope 3 reporting a requirement for these firms. In addition, the Guidance would also apply to firms reporting removals (including technology-based removals), and to firms that buy or sell carbon credits from land sector or removal activities. The Guidance would notably introduce clear guidelines on when and how removals can be reported (including removals through, for example, soil carbon sequestration).
The Draft Guidance proposes three new principles in addition to the principles of relevance, accuracy, completeness, consistency, and transparency listed in the ‘core’ GHG Protocol standards. These are:
Conservativeness: Use conservative assumptions, values, and procedures when uncertainty is high. Conservative values and assumptions are those that are more likely to overestimate GHG emissions and underestimate removals.
Permanence: Ensure mechanisms are in place to monitor the continued storage of reported removals, account for reversals, and report emissions from associated carbon pools.
Comparability: Where relevant, firms should apply common methodologies, data sources, assumptions, and reporting formats such that the reported GHG inventories from multiple firms can be compared.
The discussion here focuses on those aspects of the Land Sector and Removals Guidance most relevant to food systems.8 As in the Agricultural Guidance, the Draft Guidance requires that CO2 emissions from land use change should be reported, but it expands this requirement to also cover methane and nitrous oxide emissions due to land use change (e.g. from burning vegetation or peatland drainage, or from the mineralisation of nitrogen in soil due to losses of soil carbon).
Moreover, the Draft Guidance goes beyond the Agricultural Guidance in requiring that net biogenic CO2 emissions from land management (e.g. loss of soil carbon due to farm management practices) need to be reported in the relevant scope, rather than in a separate “Biogenic carbon” category as is the case in the Agricultural Guidance.
Net biogenic CO2 removals from land management (e.g. soil carbon sequestration due to the use of cover crops) as well as from land use change (e.g. reforestation) could optionally be reported under the relevant scope, but only if a range of additional criteria are met:
The calculation of net land carbon stock changes includes at a minimum any changes in carbon stock due to biomass, dead organic matter, and soil carbon.
There is ongoing storage monitoring documented in a land management plan or monitoring plan so that carbon remains stored and any losses can be detected.
There is traceability: when net removals occur in the firm’s supply chain, these can only be reported as Scope 3 removals if there is physical traceability to the land where carbon is stored or potentially to the first point of collection or processing facility – this requirement was still subject to discussion in the Draft Guidance given the difficulty of achieving traceability in supply chains.
There is primary data: Firms should only include net removals if it can be accounted for using primary data.
There is limited uncertainty: Firms should only include net removals if the estimated increase in the land carbon stock is statistically significant based on uncertainty estimates.
Moreover, firms would be required to report any losses of land carbon stocks either as emissions or reversals. This would also apply if firms lose the ability to monitor land carbon stocks associated with previously reported removals.
The above criteria apply to removals due to land management as well as those due to land use change, although the latter is not always clear from the current text of the Guidance.9 The Guidance foresees that firms could also optionally report gross biogenic land CO2 removals and gross emissions separately.
The Land Sector and Removals Guidance would also introduce requirements for firms to report an additional “land tracking metric” such as indirect land use change emissions, carbon opportunity costs, and/or land occupation, for Scopes 1, 2, and 3.
Table 4.2 summarises the reporting requirements most relevant to food systems.
Note: Table only shows categories relevant to food supply chains. This is a simplified representation; please refer to the full Guidance for details.
Source: GHG Protocol (2022[4]).
The Draft Guidance also includes an extensive discussion on how best to calculate emissions in each of these categories.
Given the importance of CO2 removals, it is useful to compare different standards in this regard (Box 4.2).
A unique feature of the land sector (agriculture and forestry and other land management activities) is that the biogenic carbon cycle removes CO2 from the atmosphere and transfers it to storage in biogenic carbon pools (above- and belowground biomass, dead organic matter, and soil organic matter). Changes in land use (e.g. from cropland to grassland) and changes in land management practices (e.g. use of cover crops) can increase these removals.
The ISO 14067 product carbon footprint standard proposes the following treatment:
Emissions and removals from direct land use change shall be included in the product carbon footprint, and shall be documented separately in the report
Emissions and removals from land use (land management) should be included (a weaker requirement), and if included shall be documented separately
Emissions and removals resulting from indirect land use change should be considered for inclusion, and if included shall be documented separately
The ISO 14067 standard thus treats land-related emissions and removals symmetrically. This is different from the Land Sector and Removals Guidance, which requires reporting of emissions but makes reporting of removals optional and subject to additional criteria.
The ISO 14067 standard also requires that both emissions and removals of the above categories should be documented separately in the carbon footprint study report, whereas the Land Sector and Removals Guidance requires this only for the removals.
The Pathfinder Framework proposes that emissions from direct land use change shall be included, as well as emissions and removals from land management. Emissions from indirect land use change should be reported separately but not included in the carbon footprint. The Pathfinder Framework does not explicitly cover removals from direct land use change. Importantly, the Pathfinder Framework at the moment does not explicitly impose the same criteria for removals as the Land Sector and Removals Guidance, although the relevant section of the Pathfinder Framework will be revisited once the Guidance is finalised. Table 4.3 summarises these requirements.
Note: See full text of the carbon footprint standards for additional context and guidance.
The ISO 14040/14044 standards for life-cycle assessment and the ISO 14067 standard for product carbon footprints provide important general guidance, but in calculating a product carbon footprint for a specific product, many additional questions and complexities may arise. Without additional guidance, two analysts could make different methodological choices leading to incomparable results. To prevent this lack of comparability, additional product category rules (PCR) can be developed, to provide common answers to common methodological questions in a specific product category.
The importance of these additional rules is recognised in the more general standards: the ISO 14067 standard states that if relevant PCRs exist, these should be used as long as they meet some quality criteria (one of which is that the PCR should have been developed in line with the ISO 14027 standard for the development of PCRs, or a relevant sector-specific international standard that is in line with the ISO 14044 standard for LCA). As noted above, the Pathfinder Framework similarly prioritises the use of PCRs, as long as certain quality criteria are met.
The landscape of product category rules is somewhat fragmented, as anyone can in principle develop a PCR. For food systems, a few PCRs are of particular importance, however.
A first set of PCRs are those developed as part of the EU Product Environmental Footprint (PEF). This initiative aims to set both general guidance and product-specific PEF Category Rules (PEFCR) for life-cycle assessment in the European Union (covering not just carbon footprints but 16 environmental impact categories). Because the goal is to standardise LCA calculations as much as possible, PEFCR guidance tends to be highly prescriptive. For example, the PEFCR for dairy details the specific methods, datasets and default factors to be used in calculating LCAs for five dairy product categories (liquid milk, butter, cheese, fermented milk products, dairy ingredients). The PEFCR requires the use of specific PEF datasets unless primary data is available. In addition to the PEFCR for dairy, PEFCRs exist for beer, animal feed, pet food, and pasta, with work underway on a PEFCR for marine fish.10 However, practitioners suggest that existing PEFCRs may in some cases introduce new inconsistencies (Foundation Earth, 2023[11]).
A second group of PCRs are carbon footprint standards developed by sector organisations. A prime example here is the International Dairy Federation’s Global Carbon Footprint standard for the dairy sector, first published in 2010 and most recently updated in 2022 (see below for a discussion). Other examples include the carbon footprint standard of the Global Roundtable for Sustainable Beef (GRSB, 2022[12]), and the HortiFootprint category rules, which were developed as a precursor for a PEFCR for horticultural products (Helmes et al., 2020[13]).
Finally, while not technically a product category rule, the various guidance documents produced by the Livestock Environmental Assessment and Performance (LEAP) Partnership (a multistakeholder initiative hosted by FAO) play an important role as “fallback option” in cases where a PCR is not available.11 LEAP guidance takes a life-cycle assessment approach, and is available for large ruminants, small ruminants, poultry, pigs, animal feed, and feed additives. While LEAP guidance documents are less prescriptive than PCRs, they nonetheless provide an important methodological foundation and are cross-referenced in, for example, the IDF and GRSB standards.
4.4.1. An example: IDF guidance on carbon footprints for the dairy sector
The International Dairy Federation (IDF) released its updated “Global Carbon Footprint standard for the dairy sector” in September 2022 (IDF, 2022[14]). 12 A first edition was published in 2010 and subsequently revised in 2015 and 2022. This guidance document can be used for both organisation-level and product-level reporting across the dairy life cycle (i.e. including dairy farming as well as processing).
IDF notes that guidance was necessary to avoid confusing and contradictory messages due to carbon footprint estimates based on differing methodologies and data inputs. For example, more than 4 800 peer-reviewed studies have investigated carbon footprints from dairy, but comparing these studies is difficult because of inconsistent system boundaries, allocation rules, or emission factors.
The IDF guidance is designed to be consistent with existing international standards and guidance documents. IDF distinguishes three sets of relevant standards:
General carbon footprint standards and guidelines, including the ISO standards on LCA (ISO 14040/140444) and carbon footprints (ISO 14067), the GHG Protocol standards as well as the Agricultural Guidance and the proposed Land Sector and Removals Guidance, the general guidance for the EU Product Environmental Footprint (PEF), and the PAS 2050 carbon footprint standard.
Dairy-specific guidelines, including the EU PEF category rules (PEFCR) for dairy, the FAO LEAP guidelines on large ruminants, and the dairy-specific product category rules for Environmental Product Declarations (EPD).
Guidance on specific aspects of the carbon footprint, such as information on Global Warming Potentials from IPCC reports, guidance by the GHG Protocol on accounting for removals, and guidance on carbon sequestration from the C-sequ initiative.
IDF discusses how its own guidance aligns with these.
Among other aspects, IDF provides guidance on goal, scope and boundaries, on the choice of emission factors, on allocation issues, and on accounting for land use change and carbon sequestration.
Regarding the scope, IDF explicitly aims to provide guidance for product carbon footprints using different possible scopes: cradle-to-farm gate (covering agricultural inputs and dairy farming emissions), cradle-to-factory gate (which adds emissions from the milk collection and dairy processing stages), cradle-to-purchase (which adds emissions from distribution and retail), and cradle-to-grave (which adds emissions related to use and end-of-life). IDF notes that different scopes will be relevant for different goals. Correspondingly, IDF recommends that carbon footprints should use different functional units (the “denominator” of the carbon footprint). For example, when the life cycle is covered up until the end-of-life stage, the relevant denominator would be the quantity consumed rather than the quantity purchased (to account for food waste). The choice of functional unit is discussed further below.
Regarding the boundaries, IDF provides a detailed overview of the various activities and inputs which should be included. IDF also clarifies that if a farm generates carbon credits/offsets (e.g. by carbon sequestration) and these credits are sold to a different sector, the farm can no longer include the reduction in its carbon footprint as this would lead to double-counting and double-claiming. The IDF standard makes an exception in cases where the carbon credit is maintained within the same value chain (so-called “insets”, as opposed to “offsets” traded between different value chains). This particular guidance deviates from the current GHG Protocol guidance, which generally maintains a strict separation between inventory accounting (based on actual emissions) and accounting for credits (see, for example, GHG Protocol (2004[15]), (GHG Protocol, 2011[5]), (GHG Protocol, 2022[4])). However, GHG Protocol is currently studying whether existing guidance needs to be updated in this regard (GHG Protocol, 2023[16]).
Regarding emission factors and calculation methodologies, IDF requires that methodologies be consistent with the IPCC (2019[17]) refinement to the Guidelines for National GHG Inventories, or more recent versions should these become available. In particular, in choosing between Tier 1, Tier 2, and Tier 3 approaches (discussed in more detail below), the IDF standard requires that the highest-possible Tier method must be used, and recommends that at least a Tier 2 approach should be used.
Allocation rules are relevant at several stages of the dairy value chain. At the input stage, animal feed production often generates co-products, for example when oilseeds are crushed, resulting in protein meal and vegetable oil. The farm stage produces milk and meat (from surplus calves and culled dairy cows), as well as manure. The processing stage produces a variety of dairy products (e.g. liquid milk, butter, cheese). IDF provides the following guidance:
For feed, economic allocation is recommended – that is, the allocation takes place on the basis of the relative economic value of the co-products.
For farm level production, biophysical allocation is recommended between milk and meat. In particular, IDF proposes that where emissions cannot be attributed unambiguously to either milk or meat production, the “milk share” of emissions should reflect the share of net energy for lactation in total net energy requirements.
For manure, IDF recommends that manure should be considered a “residue” of dairy production, so that a cut-off approach can be used whereby no emissions are allocated to manure. However, IDF notes that where relevant, manure may need to be considered a co-product (in which case economic allocation should be used), or a waste product (in which case emissions from the treatment of manure, potentially including those occurring outside the dairy farm, should be included and allocated between meat and milk products).
For dairy products, guidelines recommend mass-based allocation based on the dry weight of milk solids (fat, protein and lactose) or, less preferred, total dry matter (milk solids as well as minerals).
Regarding land use change, IDF recommends that the GHG emissions arising from changes in carbon stocks (soil carbon and above- and below-ground biomass) due to direct land use change should be included in the carbon footprint. This applies not only to land use change on the dairy farm but also purchased inputs, notably animal feed. All land use change occurring in a 20-year period before the reference year of the carbon footprint assessment should be included. Moreover, IDF recommends that indirect land use change should be included as a sensitivity analysis (to be reported separately).
IDF notes that carbon sequestration can have a significant impact on the carbon footprint of dairy, but that there is currently no consensus on how to quantify and account for it. IDF nevertheless recommends including carbon sequestration but reporting it separately. IDF has in parallel developed the “C-sequ” guidelines for calculating carbon sequestration in cattle production systems (IDF, 2022[18]).
While the IDF standard brings a welcome degree of harmonisation to product carbon footprint calculations in the dairy sector, its definition of the functional unit (i.e. the denominator of the carbon footprint) currently makes it less suitable for facilitating economy-wide emissions accounting. IDF prescribes that cradle-to-farm gate carbon footprints of milk should be calculated not per kg of product, but per kg of fat-and-protein corrected milk (FPCM), i.e. liquid milk with 4% fat and 3.3% protein. The IDF standard argues that this “assures objective comparison between farms with different breeds or feed regimes” (IDF, 2022[14]). However, this practice obscures the actual carbon footprint of the purchased products from the point of view of a buyer, and makes it more complicated to transmit data along the supply chain. For this reason it would be preferable to express emissions by default per kg of actual product, in line with the PACT Pathfinder guidance (PACT, 2023[1]).
IDF also notes several open issues which could be addressed in future revisions of the standard. These include the allocation of manure off-farm. The current version of the standard does not assign any production emissions to manure, so that from the point of view of crop producers, manure is currently produced “emissions-free” (although crop producers of course need to account for emissions from manure application and field use in their own carbon footprint calculations). Other issues include how to account for transfers of animals between farms, or how to account for the use of feed additives and other mitigation technologies. As IDF notes, “Whilst there is not currently enough technical information available to provide a detailed calculation method, it is desirable that we make provision for the inclusion of these technologies as more evidence on their performance becomes available” (IDF, 2022[14]). IDF suggests that a mitigation technology may be included in carbon footprint calculations once it is accepted to be included in a national GHG emissions inventory, as this signals that evidence on emissions reductions is well-substantiated and internationally accepted. IDF itself has initiated the MiLCA project (in collaboration with the Global Research Alliance on Agricultural Greenhouse Gases) to develop a protocol for including mitigation actions in agricultural life cycle assessment (GRA, 2022[19]).
Another major issue is whether carbon sequestration should be included in the carbon footprint assessment, and if so, how. As noted, the current version of the IDF standard proposes to include carbon sequestration as it can be an important mitigation option, but it requires that it be reported separately. IDF notes that in future there should be more guidance in this area. In particular, it is likely that future versions of the IDF standard will also be able to incorporate the GHG Protocol’s Land Sector and Removals Guidance in this regard.
As the previous discussion shows, there exists a well-developed landscape of carbon footprint reporting standards covering both firm-level (organisation-level) and product-level reporting. Product-level carbon footprint standards are furthermore based on the widely used Life Cycle Assessment methodology. At the level of cross-sectoral guidance, multiple standards co-exist, notably the ISO and GHG Protocol standards as well as the older PAS 2050 standard. Yet, these standards are quite similar.
On top of these cross-sectoral standards, additional sector- and product-specific guidance has been created. For firm-level reporting (including Scope 3 reporting), GHG Protocol provides important additional guidance relevant for food systems. This includes its Agricultural Guidance and its forthcoming Land Sector and Removals Guidance, which is expected to be highly influential. The analysis here shows that there are some inconsistencies between the older Agricultural Guidance and the draft Land Sector and Removals Guidance regarding the treatment of CO2 emissions and removals from land management, and CO2 removals from land use change, but this would be irrelevant if the Land Sector and Removals Guidance replaces the older Agricultural Guidance.
For product carbon footprints, additional product category rules and sectoral guidance can be used. These include category rules developed as part of the EU Product Environmental Footprint initiative, as well as sectoral guidance developed for, by example, dairy, beef, and horticulture. The FAO LEAP project also provides methodological guidance for LCA of livestock and feed, which can be used as a fallback in the absence of more detailed guidance.
The PACT Pathfinder Framework provides a bridge between firm-level and Scope 3 reporting on the one hand and product carbon footprints on the other. Its aim is to provide a clear hierarchy of which standards to use. It prioritises well-developed product category rules, followed by sector-specific rules, followed by cross-sectoral standards such as the ISO and GHG Protocol standards. It also provides supplemental guidance to ensure consistency.
Despite these efforts of harmonisation, some areas of ambiguity remain. The treatment of CO2 emissions and removals from land management and land use change does not appear to be fully streamlined yet between the GHG Protocol Agriculture Guidance, the draft Land Sector and Removals Guidance, the ISO 14067 standard, and the PACT Pathfinder Framework.
In addition, some issues that appear settled in the current standards framework may need to be reviewed over time. One is the question of indirect land use change (ILUC). Existing standards take an “attributional” approach to carbon footprints, which asks whether any land use change occurred in the life cycle of the product. However, as markets are connected, growing demand for commodity A in one region might displace its production of commodity B to a different region, where it might cause land use change. The GHG Protocol draft Land Sector and Removals Guidance and the ISO 14067 standard recognise the importance of this issue, but do not currently require inclusion of ILUC in the ‘regular’ carbon footprint calculation. To provide correct incentives, it may however be desirable to include ILUC, perhaps by providing reference tables with estimates of ILUC effects for major commodities.
Another question concerns allocation rules. Where the same production process creates several co-products (as is often the case in food systems), there is a question of how to allocate emissions across the different products. Existing standards provide guidance on allocation rules. As noted earlier, all standards advocate for avoiding allocation as much as possible. Beyond that, however, the preferred approach differs. Product category rules currently suggest using economic allocation for animal feed (e.g. between protein meal and vegetable oil), biophysical allocation for dairy cows (between milk and meat), and a mass-based allocation for dairy processing (between, for example, butter, skim milk powder). There appears to be a lack of scientific research on how these different allocation rules could impact economic behaviour, and hence on how allocation rules should be designed to provide the correct incentives.
Finally, it must be noted that standards will require continuous updating as scientific insights and techniques evolve. For example, the IDF guidance for carbon footprints in the dairy sector notes a lack of consensus on how to quantify soil carbon sequestration and hence recommends reporting it separately for the time being. Moreover, as standards depend on each other, a modification in a cross-sectoral standard or important guidance document should be reflected in more specialised standards built on top of them. This is one example of the need for continuous improvement, discussed in more detail below.
References
[7] BSI (2011), The Guide to PAS 2050:2011. How to carbon footprint your products, identify hotspots and reduce emissions in your supply chain, British Standards Institution.
[3] Cucurachi, S. et al. (2019), “Life Cycle Assessment of Food Systems”, One Earth, Vol. 1/3, pp. 292-297, https://doi.org/10.1016/j.oneear.2019.10.014.
[11] Foundation Earth (2023), LCA Methodology for Environmental Food Labelling - Beta Version 1.0, Foundation Earth.
[16] GHG Protocol (2023), Survey on Need and Scope for Updates or Additional Guidance: Market-based Accounting Approaches Survey Memo, GHG Protocol, https://ghgprotocol.org/sites/default/files/Market-based%20accounting%20Survey%20Memo.pdf.
[4] GHG Protocol (2022), GHG Protocol Land Sector and Removals Guidance (Draft for Pilot Testing and Review, September 2022), https://ghgprotocol.org/land-sector-and-removals-guidance#supporting-documents.
[10] GHG Protocol (2014), GHG Protocol Agricultural Guidance, GHG Protocol, https://ghgprotocol.org/agriculture-guidance.
[5] GHG Protocol (2011), Product Life Cycle Accounting and Reporting Standard, GHG Protocol, https://ghgprotocol.org/sites/default/files/standards/Product-Life-Cycle-Accounting-Reporting-Standard_041613.pdf.
[15] GHG Protocol (2004), A Corporate Accounting and Reporting Standard - Revised Edition, https://ghgprotocol.org/sites/default/files/standards/ghg-protocol-revised.pdf.
[6] GHG Protocol (n.d.), Quantifying the greenhouse gas emissions of products - PAS 2050 & the GHG Protocol Product Standard: A short guide to their purpose, similarities and differences, https://ghgprotocol.org/sites/default/files/2022-12/GHG%20Protocol%20PAS%202050%20Factsheet.pdf.
[19] GRA (2022), RfP – Protocol for including mitigation actions in agricultural LCA assessments (MiLCA), https://globalresearchalliance.org/n/rfp-milca/.
[12] GRSB (2022), Beef Carbon Footprint Guideline, Global Roundtable for Sustainable Beef, https://grsbeef.org/grsb-beef-carbon-footprint-guideline/.
[2] Hauschild, M., R. Rosenbaum and S. Olsen (eds.) (2018), Life Cycle Assessment, Springer International Publishing, Cham, https://doi.org/10.1007/978-3-319-56475-3.
[13] Helmes, R. et al. (2020), Hortifootprint Category Rules : Towards a PEFCR for horticultural products, Wageningen Economic Research, Wageningen, https://doi.org/10.18174/526452.
[18] IDF (2022), C-sequ: Life cycle assessment guidelines for calculating carbon sequestration in cattle production systems, https://cdn.shopify.com/s/files/1/0603/5167/6609/files/Bulletin_of_IDF_B519_CSequ_Life_cycle_assessment_guidelines_for_calculating_carbon_sequestration_in_cattle_production_systems_CAT_d27849b4-519c-4951-8a6c-108db7cba6ea.pdf?v=1663055790.
[14] IDF (2022), The IDF Global Carbon Footprint Standard for the Dairy Sector, International Dairy Federation, https://shop.fil-idf.org/products/the-idf-global-carbon-footprint-standard-for-the-dairy-sector.
[20] Innovation Center for U.S. Dairy (2019), Scope 1 & 2 GHG Inventory Guidance, https://ghgprotocol.org/sites/default/files/2023-03/Guidance_Handbook_2019_FINAL.pdf.
[21] Innovation Center for U.S. Dairy (2019), Scope 3 GHG Inventory Guidance, https://ghgprotocol.org/sites/default/files/2023-03/Scope_3_Handbook_2019_FINAL%20%2811%29_0.pdf.
[17] IPCC (2019), 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories.
[22] Kan, D. and M. Vieira (2020), Life cycle analysis of horticultural products: Memo on capital goods modelling, Wageningen University.
[9] OECD (2024), “Towards more accurate, timely, and granular product-level carbon intensity metrics: A scoping note”, Inclusive Forum on Carbon Mitigation Approaches Papers, No. 1, OECD Publishing, Paris, https://doi.org/10.1787/4de3422f-en.
[8] OECD/BIAC/WEF (2023), Emissions Measurement in Supply Chains: Business Realities and Challenges, World Economic Forum, https://www3.weforum.org/docs/WEF_Emissions_Measurement_in_Supply_Chains_2023.pdf.
[1] PACT (2023), PACT Pathfinder Framework: Guidance for the Accounting and Exchange of Product Life Cycle Emissions, Version 2.0, https://www.carbon-transparency.com/media/b13h4b25/pathfinder-framework_report_final.pdf.
[23] WRAP (2022), Scope 3 GHG Measurement & Reporting Protocols: Sector Guidance for Food & Drink Businesses, WRAP.
← 1. Most LCA databases (discussed in detail further below) are attributional. The “Big Climate Database” created by 2.-0 LCA Consultants and Concito is an exception. See https://lca-net.com/blog/a-bigger-and-better-climate-database-of-540-food-items/ (accessed 16 October 2024).
← 2. The ISO 14067 standard gives the example of one tonne of steel: because it can be transformed into many different products, it is difficult to pin down a single functional unit, making it more appropriate to express emissions per tonne of steel (a declared unit).
← 3. The treatment of capital goods is ambiguous in existing standards. In contrast with ISO 14067, the ISO 14040 standard (on which the ISO 14067 standard is based) states that capital goods should be taken into account (Kan and Vieira, 2020[22]). PAS 2050 excludes capital goods with a lifespan of over one year, unless supplementary requirements dictate otherwise. The GHG Protocol Product Standard states that capital goods are “non-attributable” processes, i.e. they are ‘not directly connected to the studied product during its lifecycle because they do not become the product, make the product, or directly carry the product through its lifecycle’ and are hence not required to be included. For agri-food products, capital goods may in fact contribute a significant amount of emissions, particularly in horticulture, which means their inclusion may be necessary to provide a reliable assessment of carbon footprints. For this reason, the Hortifootprint Category Rules (Helmes et al., 2020[13]), which provide supplementary guidance for the horticulture sector, do require the inclusion of capital goods such as greenhouses.
← 4. The recommendation to use an allocation method that reflects underlying physical relationships should not be misinterpreted as favouring a mass-based approach (i.e. allocation based on the relative mass of different co-products). In some cases, allocation by mass could reflect underlying physical relationships (this might for example be the case in allocating emissions of a plane or truck used for transporting different types of goods), but when there is no actual physical relationship, mass-based allocation is not necessarily appropriate.
← 5. One downside of the “system expansion” methodology is that it no longer guarantees that carbon footprints of co-products will add up to total emissions of the multi-product process. Imagine a process with three outputs A, B, and C and with total emissions of 60 tonnes of CO2 equivalent. Imagine that there exist substitute products A’, B’, and C’ where for each substitute the emissions are 30 tonnes of CO2 equivalent. In that case, calculating the carbon footprint of A by subtracting the carbon footprint of the substitutes B’ and C’ would result in a carbon footprint of zero – and the same would be true when the procedure is applied to B and C. This could be prevented by stipulating that system expansion can only be used for one product (e.g. A), with carbon footprints for the remaining co-products set equal to their substitutes. But in that case, the resulting carbon footprints are still somewhat arbitrarily dependent on which product is chosen as the ‘main’ product, and on the emissions of unrelated processes for producing B’ and C’.
← 6. There is an ambiguity here in the Pathfinder Framework, as the examples in the document seem to suggest that what matters for this calculation is the co-products’ price per unit rather than their share in total revenues (as would seem more logical).
← 7. In addition, there exist other sectoral guidance documents as well. One example is the WRAP Scope 3 Sector Guidance for Food & Drink Businesses (WRAP, 2022[23]). Another example is guidance for the US dairy industry which was explicitly recognised by the GHG Protocol as being in conformance with the requirements of the GHG Protocol standards. See Innovation Center for US Dairy (2019[20]) (for Scope 1 and 2) and Innovation Center for US Dairy (2019[21]) (for Scope 3).
← 8. The Guidance also contains provisions for how to account, for example, the carbon stored in long-lasting products (such as wood products) and how to account for technological carbon removals, which will typically be less relevant in food supply chains.
← 9. Section 7.1.1 in Chapter 7 (which deals with land use change) explains that net CO2 removals due to land use change are accounted for as land management net CO2 removals and covered under Chapter 8 (which deals with land management), although Chapter 8 does not explicitly mention removals due to land use change.
← 10. In addition to these official PEFCRs, a number of “shadow” PEFCRs exist, developed by industry actors without involvement of the European Commission. These include common wheat flour, food fermentation and soybean, the food and drink sector, fresh products, fruits and vegetables, green coffee, poultry meat, red meat, rice, soft drinks, and vegetable oils and protein meal products.
← 11. See https://www.fao.org/partnerships/leap/overview/the-partnership/en/ (accessed 23 February 2024).
← 12. IDF refers to its guidance as a standard. While there is no commonly accepted nomenclature, the term “standard” is typically used for documents which provide a framework with broader applicability (e.g. ISO standards, or the GHG Protocol Corporate Standard), whereas the terms “guidelines” and ‘guidance’ are typically used for additional advice, recommendations or clarifications (e.g. the GHG Protocol’s Agricultural Guidance, the PACT Pathfinder Guidance). For consistency, the IDF document is therefore described here as “guidance”.