Editing IPCC:AR6/SRCCL/Chapter-5 (section)

== 5.4 Impacts of food systems on climate change ==

<span id="greenhouse-gas-emissions-from-food-systems"></span>
=== 5.4.1 Greenhouse gas emissions from food systems ===

<div id="section-5-4-1-greenhouse-gas-emissions-from-food-systems-block-1"></div>

This chapter assesses the contributions of the entire food system to greenhouse gas (GHG) emissions. Food systems emissions include CO <sub>2</sub> and non-CO <sub>2</sub> gases, specifically those generated from: (i) crop and livestock activities within the farm gate (Table 5.4, category ‘Agriculture’); (ii) land use and land-use change dynamics associated with agriculture (Table 5.4, category ‘Land Use’); and (iii) food processing, retail and consumption patterns, including upstream and downstream processes such as manufacture of chemical fertilisers and fuel (Table 5.4, category ‘Beyond Farm Gate’). The first two categories comprise emissions reported by countries in the AFOLU (agriculture, forestry, and other land use) sectors of national GHG inventories; the latter comprises emissions reported in other sectors of the inventory, as appropriate. For instance, industrial processes, energy use, and food loss and waste.

The first two components (agriculture and land use) identified above are well quantified and supported by an ample body of literature (Smith et al. 2014 <sup>[[#fn:r680|680]]</sup> ). During the period 2007–2016, global agricultural non-CO <sub>2</sub> emissions from crop and livestock activities within the farm gate were 6.2 ± 1.4 GtCO <sub>2</sub> -eq yr <sup>–1</sup> during 2007–2016, with methane (142 ± 42 MtCH <sub>4</sub> yr <sup>-1</sup> , or 4.0 ± 1.2 GtCO <sub>2</sub> -eq yr <sup>-1</sup> ) contributing in CO <sub>2</sub> -eq about twice as much as nitrous oxide (8.3 ± 2.5 MtN <sub>2</sub> O yr <sup>-1</sup> , or 2.2 ± 0.7 GtCO <sub>2</sub> -eq yr <sup>–1</sup> ) to this total (Table 2.2 in Chapter 2). Emissions from land use associated with agriculture in some regions, such as from deforestation and peatland degradation (both processes involved in preparing land for agricultural use), added another 4.9 ± 2.5 GtCO <sub>2</sub> -eq yr <sup>–1</sup> (Chapter 2) globally during the same period. These estimates are associated with uncertainties of about 30% (agriculture) and 50% (land use), as per IPCC AR5 (Smith et al. 2014 <sup>[[#fn:r681|681]]</sup> ).

Agriculture activities within the farm gate and associated land-use dynamics are therefore responsible for about 11.1 ± 2.9 GtCO <sub>2</sub> -eq yr <sup>-1</sup> , or some 20% of total anthropogenic emissions (Table 5.4), consistent with post-AR5 findings (for example, Tubiello et al. 2015). In terms of individual gases, the contributions of agriculture to total emissions by gas are significantly larger. For instance, over the period 2010–2016, methane gas emissions within the farm gate represented about half of the total CH <sub>4</sub> emitted by all sectors, while nitrous dioxide gas emissions within the farm gate represented about three-quarters of the total N <sub>2</sub> O emitted by all sectors (Tubiello 2019 <sup>[[#fn:r682|682]]</sup> ). In terms of carbon, CO <sub>2</sub> emissions from deforestation and peatland degradation linked to agriculture contributed about 10% of the CO <sub>2</sub> emitted by all sectors in 2017 (Le Quéré et al. 2018 <sup>[[#fn:r683|683]]</sup> ).

Food systems emissions beyond the farm gate, such as those upstream from manufacturing of fertilisers, or downstream such as food processing, transport and retail, and food consumption, generally add to emissions from agriculture and land use, but their estimation is very uncertain due to lack of sufficient studies. The IPCC AR5 (Fischedick et al. 2014 <sup>[[#fn:r684|684]]</sup> ) provided some information on these other food system components, noting that emissions beyond the farm gate in developed countries may equal those within the farm gate, and cited one study estimating world total food system emissions to be up to 30% of total anthropogenic emissions (Garnett 2011 <sup>[[#fn:r685|685]]</sup> ). More recently, Poore and Nemecek (2018) <sup>[[#fn:r686|686]]</sup> , by looking at a database of farms and using a combination of modelling approaches across relevant processes, estimated a total contribution of food systems around 26% of total anthropogenic emissions. Total emissions from food systems may account for 21–37% of total GHG emissions (medium confidence).

Based on the available literature, a break-down of individual contributions of food systems emissions is show in Table 5.4, between those from agriculture within the farm gate (10–14%) (high confidence); emissions from land use and land-use change dynamics such as deforestation and peatland degradation, which are associated with agriculture in many regions (5–14%) (high confidence); and those from food supply chain activities past the farm gate, such as storage, processing, transport, and retail (5–10%) (limited evidence, medium agreement). Note that the corresponding lower range of emissions past the farm gate, for example, 2.6 GtCO <sub>2</sub> -eq yr <sup>–1</sup> (Table 5.4), is consistent with recent estimates made by Poore and Nemecek (2018). Contributions from food loss and waste are implicitly included in these estimates of total emissions from food systems (Section 5.5.2.5). They may account for 8–10% of total anthropogenic GHG emissions (low confidence) (FAO 2013b <sup>[[#fn:r687|687]]</sup> ).

<div id="section-5-4-1-greenhouse-gas-emissions-from-food-systems-block-2"></div>

<span id="table-5.4"></span>
<!-- START IMG -->
<!-- TABLE IMG -->
<!-- IMG TITLE -->
'''Table 5.4'''

<span id="ghg-emissions-gtco2-eq-yr1-from-the-food-system-and-their-contribution-to-total-anthropogenic-emissions.-mean-of-20072016-period."></span>
<!-- IMG CAPTION -->
'''GHG emissions (GtCO2-eq yr–1) from the food system and their contribution (%) to total anthropogenic emissions. Mean of 2007–2016 period.'''
<!-- IMG FILE -->
[[File:f551116aadc33969b6bcfc8d8bf8be84 Table-5.4.png]]

Notes: Food system emissions are estimated from a) FAOSTAT (2018), b) US EPA (2012), c) Poore and Nemecek (2018) and d) Fischedick et al. (2014) (using square root of sum of squares of standard deviations when adding uncertainty ranges; see also Chapter 2); e) rounded to nearest fifth percentile due to assessed uncertainty in estimates. Percentage shares were computed by using a total emissions value for the period 2007–2016 of nearly 52 GtCO <sub>2</sub> -eq yr <sup>–1</sup> (Chapter 2), using GWP values of the IPCC AR5 with no climate feedback (GWP-CH <sub>4</sub> =28; GWP-N <sub>2</sub> O=265).

<!-- END IMG -->
<span id="greenhouse-gas-emissions-from-croplands-and-soils"></span>
=== 5.4.2 Greenhouse gas emissions from croplands and soils ===

<div id="section-5-4-2-greenhouse-gas-emissions-from-croplands-and-soils-block-1"></div>

Since AR5, a few studies have quantified separate contributions of crops and soils on the one hand, and livestock on the other, to the total emissions from agriculture and associated land use. For instance, Carlson et al. (2017) <sup>[[#fn:r689|689]]</sup> estimated emissions from cropland to be in the range of 2–3 GtCO <sub>2</sub> -eq yr <sup>-1</sup> , including methane emissions from rice, CO <sub>2</sub> emissions from peatland cultivation, and N <sub>2</sub> O emissions from fertiliser applications. Data from FAOSTAT (2018) <sup>[[#fn:r690|690]]</sup> , recomputed to use AR5 GWP values, indicated that cropland emissions from these categories were 3.6 ± 1.2 GtCO <sub>2</sub> -eq yr <sup>–1</sup> over the period 2010–2016. Two-thirds of this were related to peatland degradation, followed by N <sub>2</sub> O emissions from synthetic fertilisers and methane emissions from paddy rice fields (Tubiello 2019) <sup>[[#fn:r691|691]]</sup> . These figures are a subset of the total emissions from agriculture and land use reported in Table 5.4. Asia, especially India, China and Indonesia accounted for roughly 50% of global emissions from croplands. Figure 5.9 shows the spatial distribution of emissions from cropland according to Carlson et al. (2017) <sup>[[#fn:r692|692]]</sup> , not including emissions related to deforestation or changes in soil carbon.

<div id="section-5-4-2-greenhouse-gas-emissions-from-croplands-and-soils-block-2"></div>

<span id="figure-5.9"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 5.9'''

<span id="cropland-ghgs-consist-of-ch4-from-rice-cultivation-co2-n2o-and-ch4-from-peatland-draining-and-n2o-from-n-fertiliser-application.-total-emissions-from-each-grid-cell-are-concentrated-in-asia-and-are-distinct-from-patterns-of-production-intensity-carlson-et-al.-2017."></span>
<!-- IMG CAPTION -->
'''Cropland GHGs consist of CH4 from rice cultivation, CO2, N2O, and CH4 from peatland draining, and N2O from N fertiliser application. Total emissions from each grid cell are concentrated in Asia, and are distinct from patterns of production intensity (Carlson et al. 2017).'''
<!-- IMG FILE -->
[[File:d57a239ac8c5d09076adf93bcd243517 Figure-5.9-1024x446.jpg]]

Cropland GHGs consist of CH <sub>4</sub> from rice cultivation, CO <sub>2</sub> , N <sub>2</sub> O, and CH <sub>4</sub> from peatland draining, and N <sub>2</sub> O from N fertiliser application. Total emissions from each grid cell are concentrated in Asia, and are distinct from patterns of production intensity (Carlson et al. 2017).

<!-- END IMG -->
<span id="greenhouse-gas-emissions-from-livestock"></span>
=== 5.4.3 Greenhouse gas emissions from livestock ===

<div id="section-5-4-3-greenhouse-gas-emissions-from-livestock-block-1"></div>

Emissions from livestock include non-CO <sub>2</sub> gases from enteric fermentation from ruminant animals and from anaerobic fermentation in manure management processes, as well as non-CO <sub>2</sub> gases from manure deposited on pastures (Smith et al. 2014 <sup>[[#fn:r693|693]]</sup> ). Estimates after the AR5 include those from Herrero et al. (2016) <sup>[[#fn:r694|694]]</sup> , who quantified non-CO <sub>2</sub> emissions from livestock to be in the range of 2.0–3.6 GtCO <sub>2</sub> -eq yr <sup>-1</sup> , with enteric fermentation from ruminants being the main contributor. FAOSTAT (2018) <sup>[[#fn:r695|695]]</sup> estimates of these emissions, renormalized to AR5 GWP values, were 4.1 ± 1.2 GtCO <sub>2</sub> -eq yr <sup>–1</sup> over the period 2010–2016.

These estimates of livestock emissions are for those generated within the farm gate.Adding emissions from relevant land-use change, energy use, and transportation processes, FAO (2014a) and Gerber et al. (2013) estimated livestock emissions of up to 5.3 ±1.6 GtCO <sub>2</sub> -eq yr <sup>–1</sup> circa the year 2010. This data came from original papers, but was scaled to SAR global warming potential (GWP) values for methane, for comparability with previous results.

All estimates agree that cattle are the main source of global livestock emissions (65–77%). Livestock in low and middle-income countries contribute 70% of the emissions from ruminants and 53% from monogastric livestock (animals without ruminant digestion processes such as pigs and poultry), and these are expected to increase as demand for livestock products increases in these countries (Figure 5.10). In contrast to the increasing trend in absolute GHG emissions, GHG emissions intensities, defined as GHG emissions per unit produced, have declined globally and are about 60% lower today than in the 1960s. This is largely due to improved meat and milk productivity of cattle breeds (FAOSTAT 2018 <sup>[[#fn:r696|696]]</sup> ; Davis et al. 2015 <sup>[[#fn:r697|697]]</sup> ).

<div id="section-5-4-3-greenhouse-gas-emissions-from-livestock-block-2"></div>

<span id="figure-5.10"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 5.10'''

<span id="global-ghg-emissions-from-livestock-for-19952005-adapted-from-herrero-et-al.-2016a."></span>
<!-- IMG CAPTION -->
'''Global GHG emissions from livestock for 1995–2005 (adapted from Herrero et al. 2016a).'''
<!-- IMG FILE -->
[[File:91828479746a640ab3c39bab4e05a70c Figure-5.10-1024x469.jpg]]

Global GHG emissions from livestock for 1995–2005 (adapted from Herrero et al. 2016a <sup>[[#fn:r1440|1440]]</sup> ).

<!-- END IMG -->
<div id="section-5-4-3-greenhouse-gas-emissions-from-livestock-block-3"></div>

Still, products like red meat remain the most inefficient in terms of emissions per kg of protein produced in comparison to milk, pork, eggs and all crop products (IPCC 2014b <sup>[[#fn:r698|698]]</sup> ). Yet, the functional unit used in these measurements is highly relevant and may produce different results (Salou et al. 2017 <sup>[[#fn:r699|699]]</sup> ). For instance, metrics based on products tend to rate intensive livestock systems as efficient, while metrics based on area or resources used tend to rate extensive systems as efficient (Garnett 2011 <sup>[[#fn:r700|700]]</sup> ). In ruminant dairy systems, less intensified farms show higher emissions if expressed by product, and lower emissions if expressed by Utilizable Agricultural Land (Gutiérrez-Peña et al. 2019 <sup>[[#fn:r701|701]]</sup> ; Salvador et al. 2017 <sup>[[#fn:r702|702]]</sup> ; Salou et al. 2017 <sup>[[#fn:r703|703]]</sup> ).

Furthermore, if other variables are used in the analysis of GHG emissions of different ruminant production systems, such as human-edible grains used to feed animals instead of crop waste and pastures of marginal lands, or carbon sequestration in pasture systems in degraded lands, then the GHG emissions of extensive systems are reduced. Reductions of 26% and 43% have been shown in small ruminants, such as sheep and goats (Gutiérrez-Peña et al. 2019 <sup>[[#fn:r704|704]]</sup> ; Salvador et al. 2017 <sup>[[#fn:r705|705]]</sup> ; Batalla et al. 2015 <sup>[[#fn:r706|706]]</sup> and Petersen et al. 2013 <sup>[[#fn:r707|707]]</sup> ). In this regard, depending on what the main challenge is in different regions (for example, undernourishment, over-consumption, natural resources degradation), different metrics could be used as reference. Other metrics that consider nutrient density have been proposed because they provide potential for addressing both mitigation and health targets (Doran-Browne et al. 2015 <sup>[[#fn:r708|708]]</sup> ).

Uncertainty in worldwide livestock population numbers remains the main source of variation in total emissions of the livestock sector, while at the animal level, feed intake, diet regime, and nutritional composition are the main sources of variation through their impacts on enteric fermentation and manure N excretion.

Increases in economies of scale linked to increased efficiencies and decreased emission intensities may lead to more emissions, rather than less, an observed dynamic referred to by economists as a ‘rebound effect’. This is because increased efficiency allows production processes to be performed using fewer resources and often at lower cost. This in turn influences consumer behaviour and product use, increasing demand and leading to increased production. In this way, the expected gains from new technologies that increase the efficiency of resource use may be reduced (for example, increase in the total production of livestock despite increased efficiency of production due to increased demand for meat sold at lower prices). Thus, in order for the livestock sector to provide a contribution to GHG mitigation, reduction in emissions intensities need to be accompanied by appropriate governance and incentive mechanisms to avoid rebound effects, such as limits on total production.

Variation in estimates of N <sub>2</sub> O emissions are due to differing (i) climate regimes, (ii) soil types, and (iii) N transformation pathways (Charles et al. 2017 and Fitton et al. 2017). It was recently suggested that N <sub>2</sub> O soil emissions linked to livestock through manure applications could be 20–40% lower than previously estimated in some regions. For instance, in Sub-Saharan Africa and Eastern Europe (Gerber et al. 2016 <sup>[[#fn:r709|709]]</sup> ) and from smallholder systems in East Africa (Pelster et al. 2017 <sup>[[#fn:r710|710]]</sup> ). Herrero et al. (2016a) estimated global livestock enteric methane to range from 1.6–2.7 Gt CO <sub>2</sub> -eq, depending on assumptions of body weight and animal diet.

<span id="greenhouse-gas-emissions-from-aquaculture"></span>
=== 5.4.4 Greenhouse gas emissions from aquaculture ===

<div id="section-5-4-4-greenhouse-gas-emissions-from-aquaculture-block-1"></div>

Emissions from aquaculture and fisheries may represent some 10% of total agriculture emissions, or about 0.58 GtCO <sub>2</sub> -eq yr <sup>–1</sup> (Barange et al. 2018 <sup>[[#fn:r711|711]]</sup> ), with two-thirds being non-CO <sub>2</sub> emissions from aquaculture (Hu et al. 2013 <sup>[[#fn:r712|712]]</sup> ; Yang et al. 2015 <sup>[[#fn:r713|713]]</sup> ) and the rest due to fuel use in fishing vessels. They were not included in Table 5.4 under agriculture emissions, as these estimates are not included in national GHG inventories and global numbers are small as well as uncertain.

Methodologies to measure aquaculture emissions are still being developed (Vasanth et al. 2016 <sup>[[#fn:r714|714]]</sup> ). N <sub>2</sub> O emissions from aquaculture are partly linked to fertiliser use for feed as well as aquatic plant growth, and depend on the temperature of water as well as on fish production (Paudel et al. 2015 <sup>[[#fn:r715|715]]</sup> ). Hu et al. (2012) <sup>[[#fn:r716|716]]</sup> estimated the global N <sub>2</sub> O emissions from aquaculture in 2009 to be 0.028 GtCO <sub>2</sub> -eq yr <sup>-1</sup> , but could increase to 0.114 GtCO <sub>2</sub> -eq yr <sup>–1</sup> (that is 5.72% of anthropogenic N <sub>2</sub> O–N emissions) by 2030 for an estimated 7.10% annual growth rate of the aquaculture industry. Numbers estimated by Williams and Crutzen (2010) <sup>[[#fn:r717|717]]</sup> were around 0.036 GtCO <sub>2</sub> -eq yr <sup>-1</sup> , and suggested that this may rise to more than 0.179 GtCO <sub>2</sub> -eq yr <sup>–1</sup> within 20 years for an estimated annual growth of 8.7%. Barange et al. (2018) <sup>[[#fn:r718|718]]</sup> assessed the contribution of aquaculture to climate change as 0.38 GtCO <sub>2</sub> -eq yr <sup>–1</sup> in 2010, around 7% of those from agriculture.

CO <sub>2</sub> emissions coming from the processing and transport of feed for fish raised in aquaculture, and also the emissions associated with the manufacturing of floating cultivation devices (e.g., rafts or floating fish-farms), connecting or mooring devices, artificial fishing banks or reefs, and feeding devices (as well as their energy consumption) may be considered within the emissions from the food system. Indeed, most of the GHG emissions from aquaculture are associated with the production of raw feed materials and secondarily, with the transport of raw materials to mills and finished feed to farms (Barange et al. 2018 <sup>[[#fn:r719|719]]</sup> ).

<span id="greenhouse-gas-emissions-from-inputs-processing-storage-and-transport"></span>
=== 5.4.5 5.4.5 Greenhouse gas emissions from inputs, processing, storage and transport ===

<div id="section-5-4-5-greenhouse-gas-emissions-from-inputs-processing-storage-and-transport-block-1"></div>

Apart from emissions from agricultural activities within the farm gate, food systems also generate emissions from the pre- and post-production stages in the form of input manufacturing (fertilisers, pesticides, feed production) and processing, storage, refrigeration, retail, waste disposal, food service, and transport. The total contribution of these combined activities outside the farm gate is not well documented. Based on information reported in the AR5 (Fischedick et al. 2014 <sup>[[#fn:r720|720]]</sup> ) and Poore  and  Nemecek  (2018) , we estimate their total contribution to be roughly 5-10% of total anthropogenic emissions (Table 5.4). There is no post-AR5 assessment at the global level in terms of absolute emissions. Rather, several studies have recently investigated how the combined emissions within and outside the farm gate are embedded in food products and thus associated with specific dietary choices (see next section). Below important components of food systems emissions beyond the farm gate are discussed based on recent literature.

Refrigerated trucks, trailers, shipping containers, warehouses, and retail displays that are vital parts of food supply chains all require energy and are direct sources of GHG emissions. Upstream emissions in terms of feed and fertiliser manufacture and downstream emissions (transport, refrigeration) in intensive livestock production (dairy, beef, pork) can account for up to 24–32% of total livestock emissions, with the higher fractions corresponding to commodities produced by monogastric animals (Weiss and Leip 2012 <sup>[[#fn:r721|721]]</sup> ). The proportion of upstream/downstream emissions fall significantly for less-intensive and more-localised production systems (Mottet et al. 2017a <sup>[[#fn:r722|722]]</sup> ).

'''Transport and processing''' . Recent globalisation of agriculture has promoted industrial agriculture and encouraged value-added processing and more distant transport of agricultural commodities, all leading to increased GHG emissions. Although often GHG-intensive, food transportation plays an important role in food chains: it delivers food from producers to consumers at various distances, particularly to feed people in food-shortage zones from food-surplus zones. (Section 5.5.2.6 for assessment of local food production.)

To some extent, processing is necessary in order to make food supplies more stable, safe, long-lived, and in some cases, nutritious (FAO 2007 <sup>[[#fn:r723|723]]</sup> ). Agricultural production within the farm gate may contribute 80–86% of total food-related emissions in many countries, with emissions from other processes such as processing and transport being small (Vermeulen et al. 2012 <sup>[[#fn:r724|724]]</sup> ). However, in net food-importing countries where consumption of processed food is common, emissions from other parts of the food lifecycle generated in other locations are much higher (Green et al. 2015 <sup>[[#fn:r725|725]]</sup> ).

A study conducted by Wakeland et al. (2012) <sup>[[#fn:r726|726]]</sup> in the USA found that the transportation-related carbon footprint varies from a few percent to more than half of the total carbon footprint associated with food production, distribution, and storage. Most of the GHGs emitted from food processing are a result of the use of electricity, natural gas, coal, diesel, gasoline or other energy sources. Cookers, boilers, and furnaces emit carbon dioxide, and wastewater emits methane and nitrous oxide. The most energy-intensive processing is wet milling of maize, which requires 15% of total USA food industry energy (Bernstein et al. 2008 <sup>[[#fn:r727|727]]</sup> ); processing of sugar and oils also requires large amounts of energy.

<span id="greenhouse-gas-emissions-associated-with-different-diets"></span>
=== 5.4.6 Greenhouse gas emissions associated with different diets ===

<div id="section-5-4-6-greenhouse-gas-emissions-associated-with-different-diets-block-1"></div>

There is now extensive literature on the relationship between food products and emissions, although the focus of the studies has been on high-income countries. Godfray et al. (2018) <sup>[[#fn:r728|728]]</sup> updated Nelson et al. (2016) <sup>[[#fn:r729|729]]</sup> , a previous systematic review of the literature on environmental impacts associated with food, and concluded that higher consumption of animal-based foods was associated with higher estimated environmental impacts, whereas increased consumption of plant-based foods was associated with estimated lower environmental impact. Assessment of individual foods within these broader categories showed that meat – sometimes specified as ruminant meat (mainly beef) – was consistently identified as the single food with the greatest impact on the environment, most often in terms of GHG emissions and/or land use per unit commodity. Similar hierarchies, linked to well-known energy losses along trophic chains, from roots to beef were found in another recent review focussing exclusively on GHG emissions (Clune et al. 2017) <sup>[[#fn:r730|730]]</sup> , and one on life-cycle assessments (Poore and Nemecek 2018 <sup>[[#fn:r731|731]]</sup> ). Poore and Nemecek (2018) <sup>[[#fn:r732|732]]</sup> amassed an extensive database that specifies both the hierarchy of emissions intensities and the variance with the production context (for example, by country and farming system).

The emissions intensities of red meat mean that its production has a disproportionate impact on total emissions (Godfray et al. 2018 <sup>[[#fn:r733|733]]</sup> ). For example, in the USA 4% of food sold (by weight) is beef, which accounts for 36% of food-related emissions (Heller and Keoleian 2015 <sup>[[#fn:r734|734]]</sup> ). Food-related emissions are therefore very sensitive to the amount and type of meat consumed. However, 100 g of beef has twice as much protein as the equivalent in cooked weight of beans, for example, and 2.5 times more iron. One can ingest only about 2.5 kg of food per day and not all food items are as dense in nutrition.

There is therefore ''robust evidence with high agreement'' that the mixture of foods eaten can have a highly significant impact on per capita carbon emissions, driven particularly through the amount of (especially grain-fed) livestock and products.

Given the rising costs of malnutrition in all its forms, a legitimate question is often asked: would a diet that promotes health through good nutrition also be one that mitigates GHG emissions? Whilst sustainable diets need not necessarily provide more nutrition, there is certainly significant overlap between those that are healthier (e.g., via eating more plant-based material and less livestock-based material), and eating the appropriate level of calories. In their systematic review, Nelson et al. (2016) <sup>[[#fn:r735|735]]</sup> conclude that, in general, a dietary pattern that is higher in plant-based foods, such as vegetables, fruits, whole grains, legumes, nuts, and seeds, and lower in animal-based foods is more health-promoting and is associated with lesser environmental impact (GHG emissions and energy, land, and water use) than is the current average ‘meat-based’ diet.

Recent FAO projections of food and agriculture to 2050 under alternative scenarios characterised by different degrees of sustainability, provide global-scale evidence that rebalancing diets is key to increasing the overall sustainability of food and agricultural systems world-wide. A 15% reduction of animal products in the diets of high-income countries by 2050 would contribute to containing the need to expand agricultural output due to upward global demographic trends. Not only would GHG emissions and the pressure on land and water be significantly reduced but the potential for low-income countries to increase the intake of animal-based food, with beneficial nutritional outcomes, could be enhanced (FAO 2018a <sup>[[#fn:r736|736]]</sup> ). Given that higher-income countries typically have higher emissions per capita, results are particularly applicable in such places.

However, Springmann et al. (2018a) <sup>[[#fn:r737|737]]</sup> found that there are locally applicable upper bounds to the footprint of diets around the world, and for lower-income countries undergoing a nutrition transition, adopting ‘Westernised’ consumption patterns (over-consumption, large amounts of livestock produce, sugar and fat), even if in culturally applicable local contexts, would increase emissions. The global mitigation potential of healthy but low-emissions diets is discussed in detail in Section 5.5.2.1.

In summary, food system emissions are growing globally due to increasing population, income, and demand for animal-sourced products ( ''high confidence'' ). Diets are changing on average toward greater consumption of animal-based foods, vegetable oils and sugar/sweeteners ( ''high confidence'' ) (see also Chapter 2), with GHG emissions increasing due to greater amounts of animal-based products in diets ( ''robust evidence, medium agreement'' ).

<span id="mitigation-options-challenges-and-opportunities"></span>