Measuring Carbon Footprints of Agri‑Food Products

2. Background: Four findings about GHG emissions in food systems

Abstract

This chapter highlights four main findings about GHG emissions in food systems. First, agricultural production and land use change account for a significant share of the GHG emissions in food systems. Second, food products differ strongly in terms of their carbon footprint. Third, there is large heterogeneity among producers of the same product. Finally, a fourth finding is that many options exist to reduce GHG emissions from food production. The chapter discusses implications for carbon footprint measurement of agri-food products.

Four findings from the scientific literature to date provide important background for carbon footprint measurement in food systems.

First, agricultural production and land use change account for a significant share of the GHG emissions in food systems. Figure 2.1, using data from Tubiello et al. (2021[1]), shows that at the global level these two stages account for most of the GHG emissions.1 For high-income countries, the role of other stages of the supply chains becomes more important: in Europe and North America, other stages of the supply chain accounted for more than half of domestic food systems emissions in 2019. However, even in these regions, GHG emissions from agricultural production are a significant source of emissions (41% in Europe, 38% in North America).2 One implication of these findings is that the carbon footprint of a food product as found in a supermarket or restaurant depends heavily on emissions which occurred upstream in the supply chain. In other words, a life-cycle view is essential. Another implication is that any methodology for calculating carbon footprints should take into account the potential impact of emissions from land use change, at least for those commodities where that impact is likely to be significant.

A second finding in the literature is that food products differ strongly in terms of their carbon footprint. Synthesizing 570 studies covering 40 products, nearly 40 000 farms, and 119 countries, Poore and Nemecek (2018[2]) showed large differences between food products in terms of carbon footprints (as well as other environmental impacts). On average, the carbon footprint of food products is higher for animal-based foods than for plant-based foods; within the animal-based foods, carbon footprints are on average higher for ruminant products (beef, lamb, cheese) (Figure 2.2).3

However, the same study also shows that there is large heterogeneity among producers of the same product, the third key finding regarding food systems emissions. Figure 2.3 shows this heterogeneity at the global level for selected protein-rich products. The grey bar in the chart shows the range between the 10% best and 10% worst producers globally in terms of carbon footprint, indicating large variation around the median values. Poore and Nemecek (2018[2]) note that their data also shows a large range for wheat, maize, and rice, even within major growing areas (the Australian wheat belt, the US corn belt, and the Yangtze river basin). These differences may be due to different farm management practices and techniques, variations in local climate and soil conditions, and interactions between these.

Finally, a fourth finding is that many options exist to reduce GHG emissions from food production, especially when the full supply chain is considered. For example, the production of nitrogen fertiliser currently relies on the use of natural gas, making it an emissions-intensive production process. It is possible to replace this with a production process based on renewable energy, which would allow for a significant reduction in the carbon footprint of nitrogen fertiliser production. On the farm, a wide range of farm management techniques and existing and future technological options can help reduce emissions. These include for example inputs such as feed additives to reduce methane emissions from enteric fermentation, or enhanced efficiency fertiliser to reduce nitrous oxide emissions; as well as changes in production practices (e.g. to increase soil carbon sequestration). Downstream supply chain actors similarly have many options to reduce emissions, from lower-emission vehicles for road transport to reducing the leakage of refrigerants. Across the food supply chain, reducing food loss and waste would similarly reduce emissions per unit of product delivered to the final consumer.

Taken together, these findings suggest that three levers can be used to reduce emissions of food systems (Deconinck, Jansen and Barisone, 2023[3]):

• Shifting to products with a lower average carbon footprint, e.g. from animal-based products to plant-based products. This requires information on the average carbon footprint of a product category.

• Within each product category, shifting to suppliers with a lower carbon footprint. At the farm stage, this could mean, for example, shifting from higher-emitting dairy producers to lower-emitting ones; but the same logic applies to other stages of the supply chain (e.g. shifting to fertiliser producers with a lower carbon footprint). Such shifts require information on supplier-specific carbon footprints.

• Incentivising producers everywhere to adopt techniques (e.g. farm management practices or technological solutions) which reduce their emissions. This requires that producers can access information on which techniques can reduce carbon footprints, not just in general but in their specific business. It also means that when producers are purchasing inputs with lower emissions (e.g. nitrogen fertiliser produced using renewable energy), this should be reflected in the estimated carbon footprint of their products. Again, this applies to the farm stage as well as to other stages of the food supply chain.

Carbon footprints in food systems should ideally be reliable enough to enable all three of these levers. The importance of agriculture in total GHG emissions, as well as the heterogeneity of emission intensities among farmers, argues for using primary data. Calculation methods should also be able to take into account emission reductions through changing techniques (such as farm management practices or new technological solutions) – and carbon footprint estimates should be updated regularly to capture such changes over time. There is a potential tension here between ensuring that methods are able to capture context-specific factors and avoiding trade barriers arising from divergent approaches in different countries.

References

[4] Crippa, M. et al. (2021), “Food systems are responsible for a third of global anthropogenic GHG emissions”, Nature Food, Vol. 2/3, pp. 198-209, https://doi.org/10.1038/s43016-021-00225-9.

[3] Deconinck, K., M. Jansen and C. Barisone (2023), “Fast and furious: the rise of environmental impact reporting in food systems”, European Review of Agricultural Economics, Vol. 50/4, pp. 1310-1337, https://doi.org/10.1093/erae/jbad018.

[6] Deconinck, K. and L. Toyama (2022), “Environmental impacts along food supply chains: Methods, findings, and evidence gaps”, OECD Food, Agriculture and Fisheries Papers, No. 185, OECD Publishing, Paris, https://doi.org/10.1787/48232173-en.

[5] Garsous, G. (2021), “Developing consumption-based emissions indicators from Agriculture, Forestry and Land-use (AFOLU) activities”, OECD Food, Agriculture and Fisheries Papers, No. 171, OECD Publishing, Paris, https://doi.org/10.1787/b2b24307-en.

[2] Poore, J. and T. Nemecek (2018), “Reducing food’s environmental impacts through producers and consumers”, Science, Vol. 360/6392, pp. 987-992, https://doi.org/10.1126/science.aaq0216.

[1] Tubiello, F. et al. (2021), Pre- and post-production processes along supply chains increasingly dominate GHG emissions from agri-food systems globally and in most countries, Copernicus GmbH, https://doi.org/10.5194/essd-2021-389.

Notes

← 1.  See also Crippa et al. (2021[4]) for a similar analysis.

← 2.  Moreover, these estimates look only at emissions which take place within national boundaries. Given international trade in agri-food commodities, it is possible that the share of agricultural production and land use change in consumption-based emissions would be larger in these regions once trade is taken into account. Better product carbon footprint data could help improve estimates of consumption-based emissions. On consumption-based emission estimates, see Garsous (2021[5]) and Deconinck and Toyama (2022[6]).

← 3.  The analysis of Poore and Nemecek (2018[2]) results in a greater share of total emissions accounted for by land use change and agricultural production, at around 81% versus 70% in the Tubiello et al. (2021[1]) data. This is partly explained by the use of a different method (“bottom-up” extrapolation from detailed life cycle assessments in the case of Poore and Nemecek; “downscaling” from global cross-sectoral estimates in the case of Tubiello et al) and partly by a different time period, as the data in Tubiello et al. (2021[1]) refers to 2019 while the estimates of Poore and Nemecek (2018[2]) are a synthesis of numerous studies which took place prior to 2018. (The relative contribution of land use change has been falling over time).