5. on increasing the price of an article by 14%, it becomes 1995. what was its original price?

5.1.1

Food security and insecurity, the food system and climate change

The food system encompasses all the activities and actors in the production, transport, manufacturing, retailing, consumption, and waste of food, and their impacts on nutrition, health and well-being, and the environment (Figure 5.1).

5.1.1.1

Food security as an outcome of the food system

The activities and the actors in the food system lead to outcomes such as food security and generate impacts on the environment. As part of the environmental impacts, food systems are a considerable contributor to GHG emissions, and thus climate change (Section 5.4). In turn, climate change has complex interactions with food systems, leading to food insecurity through impacts on food availability, access, utilisation and stability (Table 5.1 and Section 5.2).

We take a food systems lens in the Special Report on Climate Change and Land (SRCCL) to recognise that demand for and supply of food are interlinked and need to be jointly assessed in order to identify the challenges of mitigation and adaptation to climate change. Outcomes cannot be disaggregated solely to, for example, agricultural production, because the demand for food shapes what is grown, where it is grown, and how much is grown. Thus, GHG emissions from agriculture result, in large part, from ‘pull’ from the demand side. Mitigation and adaptation involve modifying production, supply chain, and demand practices (through, for example, dietary choices, market incentives, and trade relationships), so as to evolve to a more sustainable and healthy food system.

According to FAO (2001a)3, food security is a situation that exists when all people, at all times, have physical, social, and economic access to sufficient, safe, and nutritious food that meets their dietary needs and food preferences for an active and healthy life. ‘All people at all times’ implies the need for equitable and stable food distribution, but it is increasingly recognised that it also covers the need for inter-generational equity, and therefore ‘sustainability’ in food production. ‘Safe and nutritious food … for a healthy life’ implies that food insecurity can occur if the diet is not nutritious, including when there is consumption of an excess of calories, or if food is not safe, meaning free from harmful substances.

A prime impact of food insecurity is malnourishment (literally ‘bad nourishment’) leading to malnutrition, which refers to deficiencies, excesses, or imbalances in a person’s intake of energy and/or nutrients. As defined by FAO et al. (2018)1410, undernourishment occurs when an individual’s habitual food consumption is insufficient to provide the amount of dietary energy required to maintain a normal, active, healthy life. In addition to undernourishment in the sense of insufficient calories (‘hunger’), undernourishment occurs in terms of nutritional deficiencies in vitamins (e.g., vitamin A) and minerals (e.g., iron, zinc, iodine), so-called ‘hidden hunger’. Hidden hunger tends to be present in countries with high levels of undernourishment (Muthayya et al. 20134), but micronutrient deficiency can occur in societies with low prevalence of undernourishment. For example, in many parts of the world teenage girls suffer from iron deficiency (Whitfield et al. 2015)5 and calcium deficiency is common in Western-style diets (Aslam and Varani 20166). Food security is related to nutrition, and conversely food insecurity is related to malnutrition. Not all malnourishment arises from food insecurity, as households may have access to healthy diets but choose to eat unhealthily, or it may arise from illness. However, in many parts of the world, poverty is linked to poor diets (FAO et al. 20187). This may be through lack of resources to produce or access food in general, or healthy food, in particular, as healthier diets are more expensive than diets rich in calories but poor in nutrition (high confidence) (see meta-analysis by Darmon and Drewnowski 20158). The relationship between poverty and poor diets may also be linked to unhealthy ‘food environments,’ with retail outlets in a locality only providing access to foods of low nutritional quality (Gamba et al. 2015) – such areas are sometimes termed ‘food deserts’ (Battersby 20129).

Whilst conceptually the definition of food security is clear, it is not straightforward to measure in a simple way that encompasses all its aspects. Although there are a range of methods to assess food insecurity, they all have some shortcomings. For example, the FAO has developed the Food Insecurity Experience Scale (FIES), a survey-based tool to measure the severity of overall households’ inability to access food. While it provides reliable estimates of the prevalence of food insecurity in a population, it does not reveal whether actual diets are adequate or not with respect to all aspects of nutrition (Section 5.1.2.1).

5.1.1.2

Effects of climate change on food security

Climate change is projected to negatively impact the four pillars of food security – availability, access, utilisation and stability – and their interactions (FAO et al. 201810) (high confidence). This chapter assesses recent work since AR5 that has strengthened understanding of how climate change affects each of these pillars across the full range of food system activities (Table 5.1 and Section 5.2).

While most studies continue to focus on availability via impacts on food production, more studies are addressing related issues of access (e.g., impacts on food prices), utilisation (e.g., impacts on nutritional quality), and stability (e.g., impacts of increasing extreme events) as they are affected by a changing climate (Bailey et al. 2015). Low-income producers and consumers are likely to be most affected because of a lack of resources to invest in adaptation and diversification measures (UNCCD 201712; Bailey et al. 201513).

5.1.2

Status of the food system, food insecurity and malnourishment

5.1.2.1

Trends in the global food system

Food is predominantly produced on land, with, on average, 83% of the 697 kg of food consumed per person per year, 93% of the 2884 kcal per day, and 80% of the 81 g of protein eaten per day coming from terrestrial production in 2013 (FAOSTAT 201814). With increases in crop yields and production (Figure 5.2), the absolute supply of food has been increasing over the last five decades. Growth in production of animal-sourced food is driving crop utilisation for livestock feed (FAOSTAT 201815; Pradhan et al. 2013a16). Global trade of crop and animal-sourced food has increased by around 5 times between 1961 and 2013 (FAOSTAT 201817). During this period, global food availability has increased from 2200 kcal/cap/day to 2884 kcal/cap/day, making a transition from a food deficit to a food surplus situation (FAOSTAT 2018; Hiç et al. 201618).

The availability of cereals, animal products, oil crops, and fruits and vegetables has mainly grown (FAOSTAT 201819), reflecting shifts towards more affluent diets. This, in general, has resulted in a decrease in prevalence of underweight and an increase in prevalence of overweight and obesity among adults (Abarca-Gómez et al. 201720). During the period 1961–2016, anthropogenic GHG emissions associated with agricultural production has grown from 3.1 GtCO2-eq yr–1 to 5.8 GtCO2-eq yr–1 (Section 5.4.2 and Chapter 2). The increase in emissions is mainly from the livestock sector (from enteric fermentation and manure left on pasture), use of synthetic fertiliser, and rice cultivation (FAOSTAT 201821).

5.1.2.2

Food insecurity status and trends

In addressing food security the dual aspects of malnutrition – under-nutrition and micro-nutrient deficiency, as well as over-consumption, overweight, and obesity – need to be considered (Figure 5.2 (g) and Table 5.2). The UN agencies’ State of Food Security and Nutrition 2018 report (FAO et al. 201822) and the Global Nutrition Report 2017 (Development Initiatives 201723) summarise the global data. The State of Food Security report’s estimate for undernourished people on a global basis is 821 million, up from 815 million the previous year and 784 million the year before that. Previous to 2014/2015 the prevalence of hunger had been declining over the last three decades. The proportion of young children (under five) who are stunted (low height-for-age), has been gradually declining, and was 22% in 2017 compared to 31% in 2012 (150.8 million, down from 165.2 million in 2012). In 2017, 50.5 million children (7.5%) under five were wasted (low weight-for-height). Since 2014, undernutrition has worsened, particularly in parts of Sub-Saharan Africa, south-eastern Asia and Western Asia, and recently Latin America. Deteriorations have been observed most notably in situations of conflict and conflict combined with droughts or floods (FAO et al. 201824).

Regarding micronutrient deficiencies known as ‘hidden hunger’, reporting suggests a prevalence of one in three people globally (FAO 2013a25; von Grebmer et al. 201426; Tulchinsky 201027) (Table 5.2). In the last decades, hidden hunger (measured through proxies targeting iron, vitamin A, and zinc deficiencies) worsened in Africa, while it mainly improved in the Asia and Pacific regions (Ruel-Bergeron et al. 201528). In 2016, 613 million women and girls aged 15 to 49 suffered from iron deficiency (Development Initiatives 201829); in 2013, 28.5% of the global population suffered from iodine deficiency; and in 2005, 33.3% of children under five and 15.3% of pregnant women suffered from vitamin A deficiency, and 17.3% of the global population suffered from zinc deficiency (HLPE 201730).

Globally, as the availability of inexpensive calories from commodity crops increases, so does per capita consumption of calorie-dense foods (Ng et al. 201431; NCD-RisC 2016a32; Abarca-Gómez et al. 2017 and Doak and Popkin 201733). As a result, in every region of the world, the prevalence of obesity (body mass index >30 kg m–2) and overweight (body mass index range between normality [18.5–24.9] and obesity) is increasing. There are now more obese adults in the world than underweight adults (Ng et al. 201434; NCD-RisC 2016a35; Abarca-Gómez et al. 201736 and Doak and Popkin 2017). In 2016, around two billion adults were overweight, including 678 million suffering from obesity (NCD-RisC 2016a37; Abarca-Gómez et al. 2017). The prevalence of overweight and obesity has been observed in all age groups.

Around 41 million children under five years and 340 million children and adolescents aged 5–19 years were suffering from overweight or obesity in 2016 (NCD-RisC 2016a38; FAO et al. 201739; WHO 201540). In many high-income countries, the rising trends in children and adolescents suffering from overweight or obesity have stagnated at high levels; however, these have accelerated in parts of Asia and have very slightly reduced in European and Central Asian lower and middle-income countries (Abarca-Gómez et al. 201741; Doak and Popkin 201742; Christmann et al. 200943).

There are associations between obesity and non-communicable diseases such as diabetes, dementia, inflammatory diseases (Saltiel and Olefsky 201744), cardiovascular disease (Ortega et al. 201645) and some cancers, for example, of the colon, kidney, and liver (Moley and Colditz 201646). There is a growing recognition of the rapid rise in overweight and obesity on a global basis and its associated health burden created through non-communicable diseases (NCD-RisC 2016a47; HLPE 201748).

Analyses reported in FAO et al. (2018) highlight the link between food insecurity, as measured by the FIES scale, and malnourishment (medium agreement, robust evidence). This varies by malnourishment measure as well as country (FAO et al. 201849). For example, there is limited evidence (low agreement but multiple studies) that food insecurity and childhood wasting (i.e., or low weight for height) are closely related, but it is very likely (high agreement, robust evidence) that childhood stunting and food insecurity are related (FAO et al. 201850). With respect to adult obesity there is robust evidence, with medium agreement, that food insecurity, arising from poverty reducing access to nutritious diets, is related to the prevalence of obesity, especially in high-income countries and adult females. An additional meta-analysis (for studies in Europe and North America) also finds a negative relationship between income and obesity, with some support for an effect of obesity causing low income (as well as vice versa) (Kim and von dem Knesebeck 201851).

As discussed in Section 5.1.1.1, different methods of assessing food insecurity can provide differential pictures. Of particular note is the spatial distribution of food insecurity, especially in higher-income countries. FAO et al. (2018)52 reports FIES estimates of severe food insecurity in Africa, Asia and Latin America of 29.8%, 6.9% and 9.8% of the population, respectively, but of 1.4% of the population (i.e., about 20 million in total; pro rata <5 million for US, <1 million for UK) in Europe and North America. However, in the USA, USDA estimates 40 million people were exposed to varying degrees of food insecurity, from mild to severe (overall prevalence about 12%) (Coleman-Jensen et al. 201853). In the UK, estimates from 2017 and 2018 indicate about 4 million adults are moderately to severely food insecure (prevalence 8%) (End Hunger UK 2018; Bates et al. 201754). The UK food bank charity, the Trussell Trust, over a year in 2017/18, distributed 1,332,952 three-day emergency food parcels to people referred to the charity as being in food crisis. Furthermore, a 2003 study in the UK (Schenker 2003)55 estimated that 40% of adults, and 15% of children admitted to hospitals were malnourished, and that 70% of undernourishment in the UK was unreported.

In total, more than half the world’s population are underweight or overweight (NCD-RisC 2017a56), so their diets do not provide the conditions for ‘an active and healthy life’. This will be more compromised under the impacts of climate change by changing the availability, access, utilisation, and stability of diets of sufficient nutritional quality as shown in Table 5.2 and discussed in detail below (Section 5.2).

5.1.3

Climate change, gender and equity

Throughout, the chapter considers many dimensions of gender and equity in regard to climate change and the food system (Box 5.1). Climate change impacts differ among diverse social groups depending on factors such as age, ethnicity, ability/disability, sexual orientation, gender, wealth, and class (high confidence) (Vincent and Cull 201457; Kaijser and Kronsell 201458). Poverty, along with socio-economic and political marginalisation, cumulatively put women, children and the elderly in a disadvantaged position in coping with the adverse impacts of the changing climate (UNDP 201359; Skoufias et al. 201160). The contextual vulnerability of women is higher due to their differentiated relative power, roles, and responsibilities at the household and community levels (Bryan and Behrman 201361; Nelson et al. 200262). They often have a higher reliance on subsistence agriculture, which will be severely impacted by climate change (Aipira et al. 201763).

Through impacts on food prices (Section 5.2.3.1) poor people’s food security is particularly threatened. Decreased yields can impact nutrient intake of the poor by decreasing supplies of highly nutritious crops and by promoting adaptive behaviours that may substitute crops that are resilient but less nutritious (Thompson et al. 201264; Lobell and Burke 201065). In Guatemala, food prices and poverty have been correlated with lower micronutrient intakes (Iannotti et al. 201266). In the developed world, poverty is more typically associated with calorically-dense but nutrient-poor diets, obesity, overweight, and other related diseases (Darmon and Drewnowski 201567).

Rural areas are especially affected by climate change (Dasgupta et al. 201468), through impacts on agriculture-related livelihoods and rural income (Mendelsohn et al. 200769) and through impacts on employment. Jessoe et al. (2018)70 using a 28-year panel on individual employment in rural Mexico, found that years with a high occurrence of heat lead to a reduction in local employment by up to 1.4% with a medium emissions scenario, particularly for wage work and non-farm labour, with impacts on food access. Without employment opportunities in areas where extreme poverty is prevalent, people may be forced to migrate, exacerbating potential for ensuing conflicts (FAO 2018a71).

Finally, climate change can affect human health in other ways that interact with food utilisation. In many parts of the world where agriculture relies still on manual labour, projections are that heat stress will reduce the hours people can work, and increase their risk (Dunne et al. 201372). For example, Takakura et al. (2017)73 estimates that under RCP8.5, the global economic loss from people working shorter hours to mitigate heat loss may be 2.4–4% of GDP. Furthermore, as discussed by Watts et al. (2018)74; people’s nutritional status interacts with other stressors and affects their susceptibility to ill health (the ‘utilisation pillar’ of food security): so food-insecure people are more likely to be adversely affected by extreme heat, for example.

In the case of food price hikes, those more vulnerable are more affected (Uraguchi 201075), especially in urban areas (Ruel et al. 201076), where livelihood impacts are particularly severe for the individuals and groups that have scarce resources or are socially isolated (Revi et al. 201477; Gasper et al. 201178) (high confidence). These people often lack power and access to resources, adequate urban services and functioning infrastructure. As climate events become more frequent and intense, this can increase the scale and depth of urban poverty (Rosenzweig et al. 2018b79). Urban floods and droughts may result in water contamination increasing the incidence of diarrhoeal illness in poor children (Bartlett 200880). In the near destruction of New Orleans by Hurricane Katrina, about 40,000 jobs were lost (Rosemberg 201081).

Differentiated impacts, vulnerability, risk perception, behaviours and coping strategies for climate change related to food security derive from cultural (gendered) norms. That is, the behaviours, tasks, and responsibilities a society defines as ‘male’ or ‘female’, and the differential gendered access to resources (Paris and Rola-Rubzen 201882; Aberman and Tirado 201483; Lebel et al. 201484; Bee 201685). In many rural areas women often grow most of the crops for domestic consumption and are primarily responsible for storing, processing, and preparing food; handling livestock; gathering food, fodder and fuelwood; managing domestic water supply; and providing most of the labour for post-harvest activities (FAO 2011a86). They are mostly impacted through increased hardship, implications for household roles, and subsequent organisational responsibilities (Boetto and McKinnon 201387; Jost et al. 201688). Water scarcity can particularly affect women because they need to spend more time and energy to collect water, where they may be more exposed to physical and sexual violence (Sommer et al. 201589; Aipira et al. 201790). They may be forced to use unsafe water in the household increasing risk of water-borne diseases (Parikh 200991). Climate change also has differentiated gendered impacts on livestock-holders’ food security (McKune et al. 201592; Ongoro and Ogara 201293; Fratkin et al. 200494) (Supplementary Material Table SM5.1).

Gender dimensions of the four pillars

Worldwide, women play a key role in food security (World Bank 201595) and the four pillars of food security have strong gender dimensions (Thompson 201896). In terms of food availability, women tend to have less access to productive resources, including land, and thus less capacity to produce food (Cross-Chapter Box 11 in Chapter 7).

In terms of food access, gendered norms in how food is divided at mealtimes may lead to smaller food portions for women and girls. Women’s intra-household inequity limits their ability to purchase food; limitations also include lack of women’s mobility impacting trips to the market and lack of decision-making within the household (Ongoro and Ogara 201297; Mason et al. 201798; Riley and Dodson 201499).

In terms of food utilisation, men, women, children and the elderly have different nutritional needs (e.g., during pregnancy or breast-feeding).

In terms of food stability, women are more likely to be disproportionately affected by price spikes (Vellakkal et al. 2015100; Arndt et al. 2016101; Hossain and Green 2011102; Darnton-Hill and Cogill 2010103; Cohen and Garrett 2010104; Kumar and Quisumbing 2013105) because when food is scarce women reduce food consumption relative to other family members, although these norms vary according to age, ethnicity, culture, region, and social position, as well as by location in rural or urban areas (Arora-Jonsson 2011106; Goh 2012107; Niehof 2016108; Ongoro and Ogara 2012109).

Integrating gender into adaptation

Women have their own capabilities to adapt to climate change. In the Pacific Islands, women hold critical knowledge on where or how to find clean water; which crops to grow in a wet or dry season; how to preserve and store food and seeds ahead of approaching storms,

floods or droughts; and how to carry their families through the recovery months. They also play a pivotal role in managing household finances and investing their savings in education, health, livelihoods, and other activities that assist their families to adapt and respond to climate effects (Aipira et al. 2017110). Decreasing women’s capacity to adapt to the impacts of climate change also decreases that of the household (Bryan and Behrman 2013111).

However, gender norms and power inequalities also shape the ability of men, women, boys, girls and the elderly to adapt to climate risks (Rossi and Lambrou 2008112). For example, women pastoralists in the Samburu district of Kenya cannot make decisions affecting their lives, limiting their adaptive capacity (Ongoro and Ogara 2012113).

Participation in decision-making and politics, division of labour, resource access and control, and knowledge and skills (Nelson and Stathers 2009114) are some of the barriers to adaptation. Women’s adaptive capacity is also diminished because their work often goes unrecognised (Rao 2005115; Nelson and Stathers 2009116). Many of women’s activities are not defined as ‘economically active employment’ in national accounts (FAO 2011a117). This non-economic status of women’s activities implies that they are not included in wider discussions of priorities or interventions for climate change. Their perspectives and needs are not met; and thus, interventions, information, technologies, and tools promoted are potentially not relevant, and even can increase discrimination (Alston 2009118; Edvardsson Björnberg and Hansson 2013119; Huynh and Resurreccion 2014120).

Where gender-sensitive policies to climate change may exist, effective implementation in practice of gender equality and empowerment may not be achieved on the ground due to lack of technical capacity, financial resources and evaluation criteria, as shown in the Pacific Islands (Aipira et al. 2017121). Thus, corresponding institutional frameworks that are well-resourced, coordinated, and informed are required, along with adequate technical capacity within government agencies, NGOs and project teams, to strengthen collaboration and promote knowledge sharing (Aipira et al. 2017122).

Women’s empowerment: Synergies among adaptation, mitigation, and food security

Empowering and valuing women in their societies increases their capacity to improve food security under climate change and make substantial contributions to their own well-being, to that of their families and of their communities (Langer et al. 2015123; Ajani et al. 2013 and Alston 2014124) (high confidence). Women’s empowerment includes economic, social and institutional arrangements and may include targeting men in integrated agriculture programmes to change gender norms and improve nutrition (Kerr et al. 2016125). Empowerment through collective action and groups-based approaches in the near-term has the potential to equalise relationships on the local, national and global scale (Ringler et al. 2014126). Empowered women are crucial to creating effective synergies among adaptation, mitigation, and food security.

In Western Kenya, widows in their new role as main livelihood providers invested in sustainable innovations like rainwater harvesting systems and agroforestry (this can serve as both adaptation and mitigation), and worked together in formalised groups of collective action (Gabrielsson and Ramasar 2013127) to ensure food and water security. In Nepal, women’s empowerment had beneficial outcomes in maternal and children nutrition, reducing the negative effect of low production diversity (Malapit et al. 2015128). Integrated nutrition and agricultural programmes have increased women’s decision-making power and control over home gardens in Burkina Faso (van den Bold et al. 2015129) with positive impacts on food security.

5.1.4

Food systems in AR5, SR15, and the Paris Agreement

Food, and its relationship to the environment and climate change, has grown in prominence since the Rio Declaration in 1992, where food production is Chapter 14 of Agenda 21, to the Paris Agreement of 2015, which includes the need to ensure food security under the threat of climate change on its first page. This growing prominence of food is reflected in recent IPCC reports, including its Fifth Assessment Report (IPCC 2014a130) and the Special Report on global warming of 1.5°C (SR15) (IPCC 2018a131).

5.1.4.1

Food systems in AR5 and SR15

The IPCC Working Group (WG) II AR5 chapter on Food Security and Food Production Systems broke new ground by expanding its focus beyond the effects of climate change primarily on agricultural production (crops, livestock and aquaculture) to include a food systems approach as well as directing attention to undernourished people (Porter et al. 2014132). However, it focused primarily on food production systems due to the prevalence of studies on that topic (Porter et al. 2017133). It highlighted that a range of potential adaptation options exist across all food system activities, not just in food production, and that benefits from potential innovations in food processing, packaging, transport, storage, and trade were insufficiently researched at that time.

The IPCC WG III AR5 chapter on Agriculture, Forestry and Other Land Use (AFOLU) (Smith et al. 2014134) assessed mitigation potential considering not only the supply, but also the demand side of land uses, by consideration of changes in diets; it also included food loss and waste. AR5 focused on crop and livestock activities within the farm gate and land use and land-use change dynamics associated with agriculture. It did not take a full food system approach to emissions estimates that include processing, transport, storage, and retail.

The IPCC WG II AR5 Rural Areas chapter (Revi et al. 2014135) found that farm households in developing countries are vulnerable to climate change due to socio-economic characteristics and non-climate stressors, as well as climate risks (Dasgupta et al. 2014136). They also found that a wide range of on-farm and off-farm climate change adaptation measures are already being implemented and that the local social and cultural context played a prominent role in the success or failure of different adaptation strategies for food security, such as trade, irrigation or diversification. The IPCC WG II AR5 Urban Areas chapter found that food security for people living in cities was severely affected by climate change through reduced supplies, including urban-produced food, and impacts on infrastructure, as well as a lack of access to food. Poor urban dwellers are more vulnerable to rapid changes of food prices due to climate change.

Many climate change response options in IPCC WG II and WG III AR5 (IPCC 2014b137) address incremental adaptation or mitigation responses separately rather than being inclusive of more systemic or transformational changes in multiple food systems that are large-scale, in depth, and rapid, requiring social, technological, organisational and system responses (Rosenzweig and Solecki 2018138; Mapfumo et al. 2017139; Termeer et al. 2017140). In many cases, transformational change will require integration of resilience and mitigation across all parts of the food system including production, supply chains, social aspects, and dietary choices. Further, these transformational changes in the food system need to encompass linkages to ameliorative responses to land degradation (Chapter 4), desertification (Chapter 3), and declines in quality and quantity of water resources throughout the food-energy-water nexus (Chapter 2 and Section 5.7).

The IPCC Special Report on global warming of 1.5°C found that climate-related risks to food security are projected to increase with global warming of 1.5°C and increase further with 2°C (IPCC 2018a141).

5.1.4.2

Food systems and the Paris Agreement

To reach the temperature goal put forward in the Paris Agreement of limiting warming to well below 2°C, and pursuing efforts to limit warming to 1.5°C, representatives from 196 countries signed the United Nations Framework Convention on Climate Change (UNFCCC) Paris Agreement (UNFCCC 2015) in December 2015. The Agreement put forward a temperature target of limiting warming to well below 2°C, and pursuing efforts to limit warming to 1.5°C. Under the Paris Agreement, Parties are expected to put forward their best efforts through nationally determined contributions (NDCs) and to strengthen these efforts in the years ahead. Article 2 of the Agreement makes clear the agreement is within ‘the context of sustainable development’ and states actions should be ‘in a manner that does not threaten food production’ to ensure food security.

Many countries have included food systems in their mitigation and adaptation plans as found in their NDCs for the Paris Agreement (Rosenzweig et al. 2018a143). Richards et al. (2015) analysed 160 Party submissions and found that 103 include agricultural mitigation; of the 113 Parties that include adaptation in their NDCs, almost all (102) include agriculture among their adaptation priorities. There is much attention to conventional agricultural practices that can be climate-smart and sustainable (e.g., crop and livestock management), but less to the enabling services that can facilitate uptake (e.g., climate information services, insurance, credit). Considerable finance is needed for agricultural adaptation and mitigation by the least developed countries – in the order of 3 billion USD annually for adaptation and 2 billion USD annually for mitigation, which may be an underestimate due to a small sample size (Richards et al. 2015144). On the mitigation side, none of the largest agricultural emitters included sector-specific contributions from the agriculture sector in their NDCs, but most included agriculture in their economy-wide targets (Richards et al. 2018145).

Carbon dioxide removal (CDR). A key aspect regarding the implementation of measures to achieve the Paris Agreement goals involves measures related to carbon dioxide removal (CDR) through bioenergy (Sections 5.5 and 5.6). To reach the temperature target of limiting warming to well below 2°C, and pursuing efforts to limit warming to 1.5°C, large investments and abrupt changes in land use will be required to advance bioenergy with carbon capture and sequestration (BECCS), afforestation and reforestation (AR), and biochar technologies. Existing scenarios estimate the global area required for energy crops to help limit warming to 1.5°C in the range of 109–990 Mha, most commonly around 380–700 Mha.

Most scenarios assume very rapid deployment between 2030 and 2050, reaching rates of expansion in land use in 1.5°C scenarios exceeding 20 Mha yr-1, which are unprecedented for crops and forestry reported in the FAO database from 1961. Achieving the 1.5°C target would thus result in major competing demands for land between climate change mitigation and food production, with cascading impacts on food security.

This chapter assesses how the potential conflict for land could be alleviated by sustainable intensification to produce food with a lower land footprint (Cross-Chapter Box 6 in Section 5.6). To accomplish this, farmers would need to produce the same amount of food with lower land requirement, which depends on technology, skills, finance, and markets. Achieving this would also rely on demand-side changes including dietary choices that enable reduction of the land footprint for food production while still meeting dietary needs. Transitions required for such transformative changes in food systems are addressed in Section 5.7.

5.1.4.3

Charting the future of food security

This chapter utilises the common framework of the Representative Concentration Pathways (RCPs) and the Shared Socio-economic Pathways (SSPs) (Popp et al. 2017146; Riahi et al. 2017147 and Doelman et al. 2018148) to assess the impacts of future GHG emissions, mitigation measures, and adaptation on food security (Cross-Chapter Box 1 in Chapter 1, Sections 5.2 and 5.6).

New work utilising these scenario approaches has shown that the food system externalises costs onto human health and the environment (Springmann et al. 2018a149; Swinburn et al. 2019150; Willett et al. 2019151), leading to calls for transforming the food system to deliver better human and sustainability outcomes (Willett et al. 2019152; IAP 2018153; Development Initiatives 2018154; Lozano et al. 2018155). Such a transformation could be an important lever to address the complex interactions between climate change and food security. Through acting on mitigation and adaptation in regard to both food demand and food supply we assess the potential for improvements to both human health and the Sustainable Development Goals (Section 5.6).

This chapter builds on the food system and scenario approaches followed by AR5 and its focus on climate change and food security, but new work since AR5 has extended beyond production to how climate change interacts with the whole food system. The analysis of climate change and food insecurity has expanded beyond undernutrition to include the over-consumption of unhealthy mass-produced food high in sugar and fat, which also threatens health in different but highly damaging ways, as well as the role of dietary choices and consumption in GHG emissions. It focuses on land-based food systems, though highlighting in places the contributions of freshwater and marine production.

The chapter assesses new work on the observed and projected effects of CO2 concentrations on the nutritional quality of crops (Section 5.2.4.2) emphasising the role of extreme climate events (Section 5.2.5.1), social aspects including gender and equity (Box 5.1, and Cross-Chapter Box 11 in Chapter 7), and dietary choices (Section 5.4.6, 5.5.2). Other topics with considerable new literature include impacts on smallholder farming systems (Section 5.2.2.6), food loss and waste (Section 5.5.2.5), and urban and peri-urban agriculture (Section 5.6.5). The chapter explores the potential competing demands for land that mitigation measures to achieve temperature targets may engender, with cascading impacts on food production, food security, and farming systems (Section 5.6), and the enabling conditions for achieving mitigation and adaptation in equitable and sustainable ways (Section 5.7). Section 5.8 presents challenges to future food security, including food price spikes, migration, and conflict.

5.2.1

Climate drivers important to food security

Climate drivers relevant to food security and food systems include temperature-related, precipitation-related, and integrated metrics that combine these and other variables. These are projected to affect many aspects of the food security pillars (FAO 2018b159) (see Supplementary Material Table SM5.2, and Chapter 6 for assessment of observed and projected climate impacts). Climate drivers relevant to food production and availability may be categorised as modal climate changes (e.g., shifts in climate envelopes causing shifts in cropping varieties planted), seasonal changes (e.g., warming trends extending growing seasons), extreme events (e.g., high temperatures affecting critical growth periods, flooding/droughts), and atmospheric conditions for example, CO2 concentrations, short-lived climate pollutants (SLCPs), and dust. Water resources for food production will be affected through changing rates of precipitation and evaporation, ground water levels, and dissolved oxygen content (Cruz-Blanco et al. 2015160; Sepulcre-Canto et al. 2014161; Huntington et al. 2017162; Schmidtko et al. 2017163). Potential changes in major modes of climate variability can also have widespread impacts such as those that occurred during late 2015 to early 2016 when a strong El Niño contributed to regional shifts in precipitation in the Sahel region. Significant drought across Ethiopia resulted in widespread crop failure and more than 10 million people in Ethiopia requiring food aid (U.S. Department of State 2016164; Huntington et al. 2017165) (Figure 5.3).

Other variables that affect agricultural production, processing, and/ or transport are solar radiation, wind, humidity, and (in coastal areas) salinisation and storm surge (Mutahara et al. 2016166; Myers et al. 2017167). Extreme climate events resulting in inland and coastal flooding, can affect the ability of people to obtain and prepare food (Rao et al. 2016168; FAO et al. 2018169). For direct effects of atmospheric CO2 concentrations on crop nutrient status see Section 5.2.4.2.

5.2.1.1

Short-lived climate pollutants

The important role of short-lived climate pollutants such as ozone and black carbon is increasingly emphasised since they affect agricultural production through direct effects on crops and indirect effects on climate (Emberson et al. 2018170; Lal et al. 2017171; Burney and Ramanathan 2014172; Ghude et al. 2014173) (Chapters 2 and 4). Ozone causes damage to plants through damages to cellular metabolism that influence leaf-level physiology to whole-canopy and root-system processes and feedbacks; these impacts affect leaf-level photosynthesis senescence and carbon assimilation, as well as whole-canopy water and nutrient acquisition and ultimately crop growth and yield (Emberson et al. 2018174).

Using atmospheric chemistry and a global integrated assessment model, Chuwah et al. (2015)175 found that without a large decrease in air pollutant emissions, high ozone concentration could lead to an increase in crop damage of up to 20% in agricultural regions in 2050 compared to projections in which changes in ozone are not accounted for. Higher temperatures are associated with higher ozone concentrations; C3 crops are sensitive to ozone (e.g., soybeans, wheat, rice, oats, green beans, peppers, and some types of cottons) and C4 crops are moderately sensitive (Backlund et al. 2008176).

Methane increases surface ozone which augments warming-induced losses and some quantitative analyses now include climate, long-lived (CO2) and multiple short-lived pollutants (CH4, O3) simultaneously (Shindell et al. 2017177; Shindell 2016178). Reduction of tropospheric ozone and black carbon can avoid premature deaths from outdoor air pollution and increases annual crop yields (Shindell et al. 2012179). These actions plus methane reduction can influence climate on shorter time scales than those of carbon dioxide reduction measures. Implementing them substantially reduces the risks of crossing the 2°C threshold and contributes to achievement of the SDGs (Haines et al. 2017180; Shindell et al. 2017181).

5.2.2

Climate change impacts on food availability

Climate change impacts food availability through its effect on the production of food and its storage, processing, distribution, and exchange.

5.2.2.1

Impacts on crop production

Observed impacts. Since AR5, there have been further studies that document impacts of climate change on crop production and related variables (Supplementary Material Table SM5.3). There have also been a few studies that demonstrate a strengthening relationship between observed climate variables and crop yields that indicate future expected warming will have severe impacts on crop production (Mavromatis 2015182; Innes et al. 2015183). At the global scale, Iizumi et al. (2018)184 used a counterfactual analysis and found that climate change between 1981 and 2010 has decreased global mean yields of maize, wheat, and soybeans by 4.1, 1.8 and 4.5%, respectively, relative to preindustrial climate, even when CO2 fertilisation and agronomic adjustments are considered. Uncertainties (90% probability interval) in the yield impacts are –8.5 to +0.5% for maize, –7.5 to +4.3% for wheat, and –8.4 to –0.5% for soybeans. For rice, no significant impacts were detected. This study suggests that climate change has modulated recent yields on the global scale and led to production losses, and that adaptations to date have not been sufficient to offset the negative impacts of climate change, particularly at lower latitudes.

Dryland settlements are perceived as vulnerable to climate change with regard to food security, particularly in developing countries; such areas are known to have low capacities to cope effectively with decreasing crop yields (Shah et al. 2008185; Nellemann et al. 2009186). This is of concern because drylands constitute over 40% of the earth’s land area, and are home to 2.5 billion people (FAO et al. 2011187).

Australia

In Australia, declines in rainfall and rising daily maximum temperatures based on simulations of 50 sites caused water-limited yield potential to decline by 27% from 1990 to 2015, even though elevated atmospheric CO2 concentrations had a positive effect (Hochman et al. 2017188). In New South Wales, high-temperature episodes during the reproduction stage of crop growth were found to have negative effects on wheat yields, with combinations of low rainfall and high temperatures being the most detrimental (Innes et al. 2015189).

Asia

There are numerous studies demonstrating that climate change is affecting agriculture and food security in Asia. Several studies with remote sensing and statistical data have examined rice areas in north-eastern China, the northernmost region of rice cultivation, and found expansion over various time periods beginning in the 1980s, with most of the increase occurring after 2000 (Liu et al. 2014190; Wang et al. 2014191; Zhang et al. 2017192). Rice yield increases have also been found over a similar period (Wang et al. 2014193). Multiple factors, such as structural adjustment, scientific and technological progress, and government policies, along with regional warming (1.43°C in the past century) (Fenghua et al. 2006194) have been put forward as contributing to the observed expanded rice areas and yield in the region. Shi et al. (2013)195 indicate that there is a partial match between climate change patterns and shifts in extent and location of the rice-cropping area (2000–2010).

There have also been documented changes in winter wheat phenology in Northwest China (He 2015196). Consistent with this finding, dates of sowing and emergence of spring and winter wheat were delayed, dates of anthesis and maturity was advanced, and length of reproductive growth period was prolonged from 1981–2011 in a study looking at these crops across China (Liu et al. 2018b197). Another study looking in Northwest China demonstrated that there have been changes in the phenology and productivity of spring cotton (Huang and Ji 2015198). A counterfactual study looking at wheat growth and yield in different climate zones of China from 1981–2009 found that impacts were positive in northern China and negative in southern China (Tao et al. 2014199). Temperature increased across the zones while precipitation changes were not consistent (Tao et al. 2014200).

Similar crop yield studies focusing on India have found that warming has reduced wheat yields by 5.2% from 1981 to 2009, despite adaptation (Gupta et al. 2017201), and that maximum daytime temperatures have risen along with some night-time temperatures (Jha and Tripathi 2017202).

Agriculture in Pakistan has also been affected by climate change. From 1980 to 2014, spring maize growing periods have shifted an average of 4.6 days per decade earlier, while sowing of autumn maize has been delayed 3.0 days per decade (Abbas et al. 2017203). A similar study with sunflower showed that increases in mean temperature from 1980 to 2016 were highly correlated with shifts in sowing, emergence, anthesis, and maturity for fall and spring crops (Tariq et al. 2018204).

Mountain people in the Hindu-Kush Himalayan region encompassing parts of Pakistan, India, Nepal, and China, are particularly vulnerable to food insecurity related to climate change because of poor infrastructure, limited access to global markets, physical isolation, low productivity, and hazard exposure, including Glacial Lake Outburst Floods (GLOFs) (Rasul et al. 2019205; Rasul 2010206; Tiwari and Joshi 2012207; Huddleston et al. 2003208; Ward et al. 2013209; FAO 2008210; Nautiyal et al. 2007211; Din et al. 2014212). Surveys have been conducted to determine how climate-related changes have affected food security (Hussain et al. 2016213; Shrestha and Nepal 2016214) with results showing that the region is experiencing an increase in extremes, with farmers facing more frequent floods as well as prolonged droughts with ensuing negative impacts on agricultural yields and increases in food insecurity (Hussain et al. 2016215; Manzoor et al. 2013216).

South America

In another mountainous region, the Andes, inhabitants are also beginning to experience changes in the timing, severity, and patterns of the annual weather cycle. Data collected through participatory workshops, semi-structured interviews with agronomists, and qualitative fieldwork from 2012 to 2014 suggest that in Colomi, Bolivia, climate change is affecting crop yields and causing farmers to alter the timing of planting, their soil management strategies, and the use and spatial distribution of crop varieties (Saxena et al. 2016217). In Argentina, there has also been an increase in yield variability of maize and soybeans (Iizumi and Ramankutty 2016218). These changes have had important implications for the agriculture, human health, and biodiversity of the region (Saxena et al. 2016219).

Africa

In recent years, yields of staple crops such as maize, wheat, sorghum, and fruit crops, such as mangoes, have decreased across Africa, widening food insecurity gaps (Ketiem et al. 2017220). In Nigeria, there have been reports of climate change having impacts on the livelihoods of arable crop farmers (Abiona et al. 2016221; Ifeanyi-obi et al. 2016222; Onyeneke 2018223). The Sahel region of Cameroon has experienced an increasing level of malnutrition. This is partly due to the impact of climate change since harsh climatic conditions leading to extreme drought have a negative influence on agriculture (Chabejong 2016224).

Utilising farmer interviews in Abia State, Nigeria, researchers found that virtually all responders agreed that the climate was changing in their area (Ifeanyi-obi et al. 2016225). With regard to management responses, a survey of farmers from Anambra State, Nigeria, showed that farmers are adapting to climate change by utilising such techniques as mixed cropping systems, crop rotation, and fertiliser application (Onyeneke et al. 2018226). In Ebonyi State, Nigeria, Eze (2017)227 interviewed 160 women cassava farmers and found the major climate change risks in production to be severity of high temperature stress, variability in relative humidity, and flood frequency.

Europe

The impacts of climate change are varied across the continent. Moore and Lobell (2015)228 showed via counterfactual analysis that climate trends are affecting European crop yields, with long-term temperature and precipitation trends since 1989 reducing continent-wide wheat and barley yields by 2.5% and 3.8%, respectively, and having slightly increased maize and sugar beet yields. Though these aggregate affects appear small, the impacts are not evenly distributed. In cooler regions such as the United Kingdom and Ireland, the effect of increased warming has been ameliorated by an increase in rainfall. Warmer regions, such as Southern Europe, have suffered more from the warming; in Italy this effect has been amplified by a drying trend, leading to yield declines of 5% or greater.

Another study examining the impacts of recent climate trends on cereals in Greece showed that crops are clearly responding to changes in climate – and demonstrated (via statistical analysis) that significant impacts on wheat and barley production are expected at the end of the 21st century (Mavromatis 2015229). In the Czech Republic, a study documented positive long-term impacts of recent warming on yields of fruiting vegetables (cucumbers and tomatoes) from 4.9 to 12% per 1°C increase in local temperature, but decreases in yield stability of traditionally grown root vegetables in the warmest areas of the country (Potopová et al. 2017230). A study in Hungary also indicated the increasingly negative impacts of temperature on crops and indicated that a warming climate is at least partially responsible for the stagnation in crop yields since the mid-1980s in Eastern Europe (Pinke and Lövei 2017231).

In summary, climate change is already affecting food security (high confidence). Recent studies in both large-scale and smallholder farming systems document declines in crop productivity related to rising temperatures and changes in precipitation. Evidence for climate change impacts (e.g., declines and stagnation in yields, changes in sowing and harvest dates, increased infestation of pests and diseases, and declining viability of some crop varieties) is emerging from detection and attribution studies and ILK in Australia, Europe, Asia, Africa, North America, and South America (medium evidence, robust agreement).

Projected impacts

Climate change effects have been studied on a global scale following a variety of methodologies that have recently been compared (Lobell and Asseng 2017232; Zhao et al. 2017a233 and Liu et al. 2016234). Approaches to study global and local changes include global gridded crop model simulations (e.g., Deryng et al. 2014235), point-based crop model simulations (e.g., Asseng et al. 2015236), analysis of point-based observations in the field (e.g., Zhao et al. 2016237), and temperature-yield regression models (e.g., Auffhammer and Schlenker 2014238). For an evaluation of model skills see example used in AgMIP (Müller et al. 2017b239).

Results from Zhao et al. (2017a240) across different methods consistently showed negative temperature impacts on crop yield at the global scale, generally underpinned by similar impacts at country and site scales. A limitation of Zhao et al. (2017a) is that it is based on the assumption that yield responses to temperature increase are linear, while yield response differs depending on growing season temperature levels. Iizumi et al. (2017)241 showed that the projected global mean yields of maize and soybean at the end of this century do decrease monotonically with warming, whereas those of rice and wheat increase with warming but level off at about 3°C (2091–2100 relative to 1850–1900).

Empirical statistical models have been applied widely to different cropping systems, at multiple scales. Analyses using statistical models for maize and wheat tested with global climate model scenarios found that the RCP4.5 scenario reduced the size of average yield impacts, risk of major slowdowns, and exposure to critical heat extremes compared to RCP8.5 in the latter decades of the 21st century (Tebaldi and Lobell 2018242). Impacts on crops grown in the tropics are projected to be more negative than in mid – to high-latitudes as stated in AR5 and confirmed by recent studies (e.g., Levis et al. 2018243). These projected negative effects in the tropics are especially pronounced under conditions of explicit nitrogen stress (Rosenzweig et al. 2014244) (Figure 5.4).

Reyer et al. (2017b) examined biophysical impacts in five world regions under different warming scenarios: 1°C, 1.5°C, 2°C, and 4°C warming. For the Middle East and northern African region a significant correlation between crop yield decrease and temperature increase was found, regardless of whether the effects of CO2 fertilisation or adaptation measures are taken into account (Waha et al. 2017245). For Latin America and the Caribbean the relationship between temperature and crop yield changes was only significant when the effect of CO2 fertilisation is considered (Reyer et al. 2017a246).

A review of recent scientific literature found that projected yield loss for West Africa depends on the degree of wetter or drier conditions and elevated CO2 concentrations (Sultan and Gaetani 2016247). Faye et al. (2018b) in a crop modelling study with RCPs 4.5 and 8.5 found that climate change could have limited effects on peanut yield in Senegal due to the effect of elevated CO2 concentrations.

Crop productivity changes for 1.5°C and 2.0°C. The IPCC Special Report on global warming of 1.5°C found that climate-related risks to food security are projected to increase with global warming of 1.5°C and increase further with 2°C (IPCC 2018b248). These findings are based among others on Schleussner et al. (2018); Rosenzweig et al. (2018a)249; Betts et al. (2018)250, Parkes et al. (2018)251 and Faye et al. (2018a)252. The importance of assumptions about CO2 fertilisation was found to be significant by Ren et al. (2018) and Tebaldi and Lobell (2018).

AgMIP coordinated global and regional assessment (CGRA) results confirm that at the global scale, positive and negative changes are mixed in simulated wheat and maize yields, with declines in some breadbasket regions, at both 1.5°C and 2.0°C (Rosenzweig et al. 2018a253). In conjunction with price changes from the global economics models, productivity declines in the Punjab, Pakistan resulted in an increase in vulnerable households and poverty rate (Rosenzweig et al. 2018a).

Crop suitability. Another method of assessing the effects of climate change on crop yields that combined observations of current maximum-attainable yield with climate analogues also found strong reductions in attainable yields across a large fraction of current cropland by 2050 (Pugh et al. 2016254). However, the study found the projected total land area in 2050, including regions not currently used for crops, climatically suitable for a high attainable yield similar to today. This indicates that large shifts in land-use patterns and crop choice will likely be necessary to sustain production growth and keep pace with current trajectories of demand.

Fruits and vegetables. Understanding the full range of climate impacts on fruits and vegetables is important for projecting future food security, especially related to dietary diversity and healthy diets.

However, studies for vegetables are very limited (Bisbis et al. 2018255). Of the 174 studies considered in a recent review, only 14 described results of field or greenhouse experiments studying impacts of increased temperatures on yields of different root and leafy vegetables, tomatoes and legumes (Scheelbeek et al. 2018256). Bisbis et al. (2018) found similar effects for vegetables as have been found for grain crops. That is, the effect of increased CO2 on vegetables is mostly beneficial for production, but may alter internal product quality, or result in photosynthetic down-regulation. Heat stress reduces fruit set of fruiting vegetables, and speeds up development of annual vegetables, shortening their time for photoassimilation. Yield losses and impaired product quality result, thereby increasing food loss and waste. On the other hand, a longer growing season due to warmer temperatures enables a greater number of plantings and can contribute to greater annual yields. However, some vegetables, such as cauliflower and asparagus, need a period of cold accumulation to produce a harvest and warmer winters may not provide those requirements.

For vegetables growing in higher baseline temperatures (>20°C), mean yield declines caused by 4°C warming were 31.5%; for vegetables growing in cooler environments (≤20°C), yield declines caused by 4°C were much less, on the order of about 5% (Scheelbeek et al. 2018257). Rippke et al. (2016)258 found that 30–60% of the common bean growing area and 20–40% of the banana growing areas in Africa will lose viability in 2078–2098 with a global temperature increase of 2.6°C and 4°C respectively. Tripathi et al. (2016)259 found fruits and vegetable production to be highly vulnerable to climate change at their reproductive stages and also due to potential for greater disease pressure.

In summary, studies assessed find that climate change will increasingly be detrimental to crop productivity as levels of warming progress (high confidence). Impacts will vary depending on CO2 concentrations, fertility levels, and region. Productivity of major commodity crops as well as crops such as millet and sorghum yields will be affected. Studies on fruits and vegetables find similar effects to those projected for grain crops in regard to temperature and CO2 effects. Total land area climatically suitable for high attainable yield, including regions not currently used for crops, will be similar in 2050 to today.

5.2.2.2

Impacts on livestock production systems

Livestock systems are impacted by climate change mainly through increasing temperatures and precipitation variation, as well as atmospheric carbon dioxide (CO2) concentration and a combination of these factors. Temperature affects most of the critical factors of livestock production, such as water availability, animal production and reproduction, and animal health (mostly through heat stress) (Figure 5.5). Livestock diseases are mostly affected by increases in temperature and precipitation variation (Rojas-Downing et al. 2017260). Impacts of climate change on livestock productivity, particularly of mixed and extensive systems, are strongly linked to impacts on rangelands and pastures, which include the effects of increasing CO2 on their biomass and nutritional quality. This is critical considering the very large areas concerned and the number of vulnerable people affected (Steinfeld 2010261; Morton 2007262). Pasture quality and quantity are mainly affected through increases in temperature and CO2, and precipitation variation.

Among livestock systems, pastoral systems are particularly vulnerable to climate change (Dasgupta et al. 2014263) (see Section 5.2.2.6 for impacts on smallholder systems that combine livestock and crops). Industrial systems will suffer most from indirect impacts leading to rises in the costs of water, feeding, housing, transport and the destruction of infrastructure due to extreme events, as well as an increasing volatility of the price of feedstuff which increases the level of uncertainty in production (Rivera-Ferre et al. 2016b264; Lopez-i-Gelats 2014265). Mixed systems and industrial or landless livestock systems could encounter several risk factors mainly due to the variability of grain availability and cost, and low adaptability of animal genotypes (Nardone et al. 2010266).

Considering the diverse typologies of animal production, from grazing to industrial, Rivera-Ferre et al. (2016b)267 distinguished impacts of climate change on livestock between those related to extreme events and those related to more gradual changes in the average of climate-related variables. Considering vulnerabilities, they grouped

the impacts as those impacting the animal directly, such as heat and cold stress, water stress, physical damage during extremes; and others impacting their environment, such as modification in the geographical distribution of vector-borne diseases, location, quality and quantity of feed and water and destruction of livestock farming infrastructures.

With severe negative impacts due to drought and high frequency of extreme events, the average gain of productivity might be cancelled by the volatility induced by increasing variability in the weather. For instance, semi-arid and arid pasture will likely have reduced livestock productivity, while nutritional quality will be affected by CO2 fertilisation (Schmidhuber and Tubiello 2007268).

Observed impacts. Pastoralism is practiced in more than 75% of countries by between 200 and 500 million people, including nomadic communities, transhumant herders, and agropastoralists (McGahey et al. 2014269). Observed impacts in pastoral systems reported in the literature include decreasing rangelands, decreasing mobility, decreasing livestock numbers, poor animal health, overgrazing, land degradation, decreasing productivity, decreasing access to water and feed, and increasing conflicts for the access to pasture land (high confidence) (López-i-Gelats et al. 2016270; Batima et al. 2008271; Njiru 2012272; Fjelde and von Uexkull 2012273; Raleigh and Kniveton 2012274; Egeru 2016275).

Pastoral systems in different regions have been affected differently. For instance, in China changes in precipitation were a more important factor in nomadic migration than temperature (Pei and Zhang 2014276). There is some evidence that recent years have already seen an increase in grassland fires in parts of China and tropical Asia (IPCC 2012277). In Mongolia, grassland productivity has declined by 20–30% over the latter half of the 20th century, and ewe average weight reduced by 4 kg on an annual basis, or about 8% since 1980 (Batima et al. 2008278). Substantial decline in cattle herd sizes can be due to increased mortality and forced off-take (Megersa et al. 2014279). Important, but less studied, is the impact of the interaction of grazing patterns with climate change on grassland composition. Spence et al. (2014)280 showed that climate change effects on Mongolia mountain steppe could be contingent on land use.

Conflicts due to resource scarcity, as well as other socio-political factors (Benjaminsen et al. 2012281) aggravated by climate change, has differentiated impact on women. In Turkana, female-headed households have lower access to decision-making on resource use and allocation, investment and planning (Omolo 2011282), increasing their vulnerability (Cross-Chapter Box 11 in Chapter 7, Section 5.1.3).

Non-climate drivers add vulnerability of pastoral systems to climate change (McKune and Silva 2013283). For instance, during environmental disasters, livestock holders have been shown to be more vulnerable to food insecurity than their crop-producing counterparts because of limited economic access to food and unfavourable market exchange rates (Nori et al. 2005284). Sami reindeer herders in Finland showed reduced freedom of action in response to climate change due to loss of habitat, increased predation, and presence of economic and legal constraints (Tyler et al. 2007285; Pape and Löffler 2012286). In Tibet, emergency aid has provided shelters and privatised communally owned rangeland, which have increased the vulnerability of pastoralists to climate change (Yeh et al. 2014287; Næss 2013288).

Projected impacts. The impacts of climate change on global rangelands and livestock have received comparatively less attention than the impacts on crop production. Projected impacts on grazing systems include changes in herbage growth (due to changes in atmospheric CO2 concentrations and rainfall and temperature regimes) and changes in the composition of pastures and in herbage quality, as well as direct impacts on livestock (Herrero et al. 2016b289). Droughts and high temperatures in grasslands can also be a predisposing factor for fire occurrence (IPCC 2012290).

Net primary productivity, soil organic carbon, and length of growing period. There are large uncertainties related to grasslands and grazing lands (Erb et al. 2016)291, especially in regard to net primary productivity (NPP) (Fetzel et al. 2017292; Chen et al. 2018293). Boone et al. (2017) estimated that the mean global annual net primary production (NPP) in rangelands may decline by 10 gC m–2 yr–1 in 2050 under RCP8.5, but herbaceous NPP is likely to increase slightly (i.e., average of 3 gC m–2 yr–1) (Figure 5.6). Results of a similar magnitude were obtained by Havlík et al. (2015)294, using EPIC and LPJmL on a global basis. According to Rojas-Downing et al. (2017)295, an increase of 2°C is estimated to negatively impact pasture and livestock production in arid and semi-arid regions and positively impact humid temperate regions.

Boone et al. (2017)296 identified significant regional heterogeneity in responses, with large increases in annual productivity projected in northern regions (e.g., a 21% increase in productivity in the USA and Canada) and large declines in western Africa (–46% in Sub-Saharan western Africa) and Australia (–17%). Regarding the length of growing period (LGP, average number of growing days per year) Herrero et al. (2016b)297 projected reductions in lower latitudes due to changes in rainfall patterns and increases in temperatures, which indicate increasing limitations of water. They identified 35°C as a critical threshold for rangeland vegetation and heat tolerance in some livestock species.

Rangeland composition. According to Boone et al. (2017), the composition of rangelands is projected to change as well (Chapter 3). Bare ground cover is projected to increase, averaging 2.4% across rangelands, with increases projected for the eastern Great Plains, eastern Australia, parts of southern Africa, and the southern Tibetan Plateau. Herbaceous cover declines are projected in the Tibetan Plateau, the eastern Great Plains, and scattered parts of the Southern Hemisphere. Shrub cover is projected to decline in eastern Australia, parts of southern Africa, the Middle East, the Tibetan Plateau, and the eastern Great Plains. Shrub cover could also increase in much of the Arctic and some parts of Africa. In mesic and semi-arid savannas south of the Sahara, both shrub and tree cover are projected to increase, albeit at lower productivity and standing biomass. Rangelands in western and south-western parts of the Isfahan province in Iran were found to be more vulnerable to future drying–warming conditions (Saki et al. 2018298; Jaberalansar et al. 2017299).

Soil degradation and expanding woody cover suggest that climate-vegetation-soil feedbacks catalysing shifts toward less productive, possibly stable states (Ravi et al. 2010300) may threaten mesic and semi-arid savannas south of the Sahara (Chapters 3 and 4). This will also change their suitability for grazing different animal species; switches from cattle, which mainly consume herbaceous plants, to goats or camels are likely to occur as increases in shrubland occur.

Direct and indirect effects on livestock. Direct impacts of climate change in mixed and extensive production systems are linked to increased water and temperature stress on the animals potentially leading to animal morbidity, mortality and distress sales. Most livestock species have comfort zones between 10oC–30oC, and at temperatures above this animals reduce their feed intake 3–5% per additional degree of temperature (NRC 1981301). In addition to reducing animal production, higher temperatures negatively affect fertility (HLPE 2012302).

Indirect impacts to mixed and extensive systems are mostly related to the impacts on the feed base, whether pastures or crops, leading to increased variability and sometimes reductions in availability and quality of the feed for the animals (Rivera-Ferre et al. 2016b303). Reduced forage quality can increase CH4 emissions per unit of gross energy consumed. Increased risk of animal diseases is also an important impact to all production systems (Bett et al. 2017304). These depend on the geographical region, land-use type, disease characteristics, and animal susceptibility (Thornton et al. 2009305). Also important is the interaction of grazing intensity with climate change. Pfeiffer et al. (2019)306 estimated that, in a scenario of mean annual precipitation below 500 mm, increasing grazing intensity reduced rangeland productivity and increased annual grass abundance.

Pastoral systems. In Kenya, some 1.8 million extra cattle could be lost by 2030 because of increased drought frequency, the value of the lost animals and production foregone amounting to 630 million USD (Herrero et al. 2010307). Martin et al. (2014)308 assessed impacts of changing precipitation regimes to identify limits of tolerance beyond which pastoral livelihoods could not be secured and found that reduced mean annual precipitation always had negative effects as opposed to increased rainfall variability. Similarly, Martin et al. (2016)309 found that drought effects on pastoralists in High Atlas in Morocco depended on income needs and mobility options (see Section 5.2.2.6 for additional information about impacts on smallholder farmers).

In summary, observed impacts in pastoral systems include changes in pasture productivity, lower animal growth rates and productivity, damaged reproductive functions, increased pests and diseases, and loss of biodiversity (high confidence). Livestock systems are projected to be adversely affected by rising temperatures, depending on the extent of changes in pasture and feed quality, spread of diseases, and water resource availability (high confidence). Impacts will differ for different livestock systems and for different regions (high confidence). Vulnerability of pastoral systems to climate change is very high (high confidence), and mixed systems and industrial or landless livestock systems could encounter several risk factors mainly due to variability of grain availability and cost, and low adaptability of animal genotypes. Pastoral system vulnerability is exacerbated by non-climate factors (land tenure issues, sedentarisation programmes, changes in traditional institutions, invasive species, lack of markets, and conflicts) (high confidence).

5.2.2.3

Impacts on pests and diseases

Climate change is changing the dynamics of pests and diseases of both crops and livestock. The nature and magnitude of future changes is likely to depend on local agroecological and management context. This is because of the many biological and ecological mechanisms by which climate change can affect the distribution, population size, and impacts of pests and diseases on food production (Canto et al. 2009310; Gale et al. 2009311; Thomson et al. 2010312; Pangga et al. 2011313; Juroszek and von Tiedemann 2013314; Bett et al. 2017315).

These mechanisms include changes in host susceptibility due to CO2 concentration effects on crop composition and climate stresses; changes in the biology of pests and diseases or their vectors (e.g., more generational cycles, changes in selection pressure driving evolution); mismatches in timing between pests or vectors and their ‘natural enemies’; changes in survival or persistence of pests or disease pathogens (e.g., changes in crop architecture driven by CO2 fertilisation and increased temperature, providing a more favourable environment for persistence of pathogens like fungi), and changes in pest distributions as their ‘climate envelopes’ shift. Such processes may affect pathogens, and their vectors, as well as plant, invertebrate and vertebrate pests (Latham et al. 2015316).

Furthermore, changes in diseases and their management, as well as changing habitat suitability for pests and diseases in the matrix surrounding agricultural fields, have the ability to reduce or exacerbate impacts (Bebber 2015317). For example, changes in water storage and irrigation to adapt to rainfall variation have the potential to enhance disease vector populations and disease occurrence (Bett et al. 2017318).

There is robust evidence that pests and diseases have already responded to climate change (Bebber et al. 2013319), and many studies have now built predictive models based on current incidence of pests, diseases or vectors that indicate how they may respond in future (e.g., Caminade et al. 2015320; Kim et al. 2015321; Kim and Cho 2016322; Samy and Peterson 2016323; Yan et al. 2017324). Warren et al. (2018)325 estimate that about 50% of insects, which are often pests or disease vectors, will change ranges by about 50% by 2100 under current GHG emissions trajectories. These changes will lead to crop losses due to changes in insect pests (Deutsch et al. 2018) and weed pressure (Ziska et al. 2018326), and thus affect pest and disease management at the farm level (Waryszak et al. 2018327). For example, Samy and Peterson (2016)328 modelled bluetongue virus (BTV), which is spread by biting Culicodes midges, finding that the distribution of BTV is likely to be extended, particularly in Central Africa, the USA, and Western Russia.

There is some evidence (medium confidence) that exposure will, on average, increase (Bebber and Gurr 2015329; Yan et al. 2017330), although there are a few examples where changing stresses may limit the range of a vector. There is also a general expectation that perturbations may increase the likelihood of pest and disease outbreaks by disturbing processes that may currently be at some quasi-equilibrium (Canto et al. 2009331; Thomson et al. 2010332; Pangga et al. 2011333). However, in some places, and for some diseases, risks may decrease as well as increase (e.g., drying out may reduce the ability of fungi to survive) (Kim et al. 2015334; Skelsey and Newton 2015335), or tsetse fly’s range may decrease (Terblanche et al. 2008336; Thornton et al. 2009337).

Pests, diseases, and vectors for both crop and livestock diseases are likely to be altered by climate change (high confidence). Such changes are likely to depend on specifics of the local context, including management, but perturbed agroecosystems are more likely, on theoretical grounds, to be subject to pest and disease outbreaks (low confidence). Whilst specific changes in pest and disease pressure will vary with geography, farming system, pest/pathogen – increasing in some situations decreasing in others – there is robust evidence, with high agreement, that pest and disease pressures are likely to change; such uncertainty requires robust strategies for pest and disease mitigation.

5.2.2.4

Impacts on pollinators

Pollinators play a key role on food security globally (Garibaldi et al. 2016338). Pollinator-dependent crops contribute up to 35% of global crop production volume and are important contributors to healthy human diets and nutrition (IPBES 2016339). On a global basis, some 1500 crops require pollination (typically by insects, birds and bats) (Klein et al. 2007340). Their importance to nutritional security is therefore perhaps under-rated by valuation methodologies, which, nonetheless, include estimates of the global value of pollination services at over 225 billion USD2010 (Hanley et al. 2015341). As with other ecosystem processes affected by climate change (e.g., changes in pests and diseases), how complex systems respond is highly context dependent. Thus, predicting the effects of climate on pollination services is difficult (Tylianakis et al. 2008342; Schweiger et al. 2010343) and uncertain, although there is limited evidence that impacts are occurring already (Section 5.2.2.4), and medium evidence that there will be an effect.

Pollination services arise from a mutualistic interaction between an animal and a plant – which can be disrupted by climate’s impacts on one or the other or both (Memmott et al. 2007344). Disruption can occur through changes in species’ ranges or by changes in timing of growth stages (Settele et al. 2016345). For example, if plant development responds to different cues (e.g., day length) from insects (e.g., temperature), the emergence of insects may not match the flowering times of the plants, causing a reduction in pollination. Climate change will affect pollinator ranges depending on species, life-history, dispersal ability and location. Warren et al. (2018)346 estimate that under a 3.2°C warming scenario, the existing range of about 49% of insects will be reduced by half by 2100, suggesting either significant range changes (if dispersal occurs) or extinctions (if it does not). However, in principle, ecosystem changes caused by invasions, in some cases, could compensate for the decoupling generated between native pollinators and pollinated species (Schweiger et al. 2010347).

Other impacts include changes in distribution and virulence of pathogens affecting pollinators, such as the fungus Nosema cerana, which can develop at a higher temperature range than the less-virulent Nosema apis; increased mortality of pollinators due to higher frequency of extreme weather events; food shortage for pollinators due to reduction of flowering length and intensity; and aggravation of other threats, such as habitat loss and fragmentation (González-Varo et al. 2013348; Goulson et al. 2015349; Le Conte and Navajas 2008350; Menzel et al. 2006351; Walther et al. 2009352; IPBES, 2016353). The increase in atmospheric CO2 is also reducing the protein content of pollen, with potential impact on pollination population biology (Ziska et al. 2016354).

In summary, as with other complex agroecosystem processes affected by climate change (e.g., changes in pests and diseases), how pollination services respond will be highly context dependent. Thus, predicting the effects of climate on pollination services is difficult and uncertain, although there is medium evidence that there will be an effect.

5.2.2.5

Impacts on aquaculture

This report focuses on land-based aquaculture; for assessment of impacts on marine fisheries both natural and farmed see the IPCC Special Report on the ocean and cryosphere in a changing climate (SROCC).

Aquaculture will be affected by both direct and indirect climate change drivers, both in the short and the long-term. Barange et al. (2018)355 provides some examples of short-term loss of production or infrastructure due to extreme events such as floods, increased risk of diseases, toxic algae and parasites; and decreased productivity due to suboptimal farming conditions. Long-term impacts may include scarcity of wild seed, limited access to freshwater for farming due to reduced precipitation, limited access to feeds from marine and terrestrial sources, decreased productivity due to suboptimal farming conditions, eutrophication and other perturbations.

FAO (2014a)357 assessed the vulnerability of aquaculture stakeholders to non-climate change drivers, which add to climate change hazards. Vulnerability arises from discrimination in access to inputs and decision-making; conflicts; infrastructure damage; and dependence on global markets and international pressures. Other non-climate drivers identified by McClanahan et al. (2015)1425 include: declining fishery resources; a North–South divide in investment; changing consumption patterns; increasing reliance on fishery resources for coastal communities; and inescapable poverty traps created by low net resource productivity and few alternatives. In areas where vulnerability to climate change is heightened, increased exposure to climate change variables and impacts is likely to exacerbate current inequalities in the societies concerned, penalising further already disadvantaged groups such as migrant fishers (e.g., Lake Chad) or women (e.g., employees in Chile’s processing industry) (FAO 2014a)1426.

In many countries the projected declines co-occur across both marine fisheries and agricultural crops (Blanchard et al. 2017358), both of which will impact the aquaculture and livestock sectors (Supplementary Material Figure SM5.1). Countries with low Human Development Index, trade opportunities and aquaculture technologies are likely to face greater challenges. These cross-sectoral impacts point to the need for a more holistic account of the inter-connected vulnerabilities of food systems to climate and global change.

5.2.2.6

Impacts on smallholder farming systems

New work has developed farming system approaches that take into account both biophysical and economic processes affected by climate change and multiple activities. Farm households in the developing world often rely on a complex mix of crops, livestock, aquaculture, and non-agricultural activities for their livelihoods (Rosenzweig and Hillel 2015359; Antle et al. 2015360). Across the world, smallholder farmers are considered to be disproportionately vulnerable to climate change because changes in temperature, rainfall and the frequency or intensity of extreme weather events directly affect their crop and animal productivity as well as their household’s food security, income and well-being (Vignola et al. 2015361; Harvey et al. 2014b362). For example, smallholder farmers in the Philippines, whose survival and livelihood largely depend on the environment, constantly face risks and bear the impacts of the changing climate (Peria et al. 2016363).

Smallholder farming systems have been recognised as highly vulnerable to climate change (Morton, 2007364) because they are highly dependent on agriculture and livestock for their livelihood (high confidence) (Dasgupta et al. 2014365). In Zimbabwe, farmers were found vulnerable due to their marginal location, low levels of technology, and lack of other essential farming resources. Farmers observed high frequency and severity of drought; excessive precipitation; drying of rivers, dams and wells; and changes in timing and pattern of seasons as evidence of climate change, and indicated that prolonged wet, hot, and dry weather conditions resulted in crop damage, death of livestock, soil erosion, bush fires, poor plant germination, pests, lower incomes, and deterioration of infrastructure (Mutekwa 2009366).

In Madagascar, Harvey et al. (2014b)367 surveyed 600 small farmers and found that chronic food insecurity, physical isolation and lack of access to formal safety nets increased Malagasy farmers’ vulnerability to any shocks to their agricultural system, particularly extreme events. In Chitwan, Nepal, occurrence of extreme events and increased variability in temperature has increased the vulnerability of crops to biotic and abiotic stresses and altered the timing of agricultural operations; thereby affecting crop production (Paudel et al. 2014368). In Lesotho, a study on subsistence farming found that food crops were the most vulnerable to weather, followed by soil and livestock. Climate variables of major concern were hail, drought and dry spells which reduced crop yields. In the Peruvian Altiplano, Sietz et al. (2012)369 evaluated smallholders’ vulnerability to weather extremes with regard to food security and found that resource scarcity (livestock, land area), diversification of activities (lack of alternative income, education deprivation) and income restrictions (harvest failure risk) shaped the vulnerability of smallholders. See Section 5.2.2.2 for observed impacts on smallholder pastoral systems.

Projected impacts. By including regional economic models, integrated methods take into account the potential for yield declines to raise prices and thus livelihoods (up to a certain point) in some climate change scenarios. Regional economic models of farming systems can be used to examine the potential for switching to other crops and livestock, as well as the role that non-farm income can play in adaptation (Valdivia et al. 2015370; Antle et al. 2015371). On the other hand, lost income for smallholders from climate change-related declines (for example, in coffee production), can decrease their food security (Hannah et al. 2017372).

Farming system methods developed by AgMIP (Rosenzweig et al. 2013373) have been used in regional integrated assessments in Sub-Saharan Africa (Kihara et al. 2015374), West Africa (Adiku et al. 2015375); East Africa (Rao et al. 2015376), South Africa (Beletse et al. 2015377), Zimbabwe (Masikati et al. 2015378), South Asia (McDermid et al. 2015379), Pakistan (Ahmad et al. 2015380), the Indo-Gangetic Basin (Subash et al. 2015381), Tamil Nadu (Ponnusamy et al. 2015) and Sri Lanka (Zubair et al. 2015382). The assessments found that climate change adds pressure to smallholder farmers across Sub-Saharan Africa and South Asia, with winners and losers within each area studied. Temperatures are expected to increase in all locations, and rainfall decreases are projected for the western portion of West Africa and southern Africa, while increases in rainfall are projected for eastern West Africa and all studied regions of South Asia. The studies project that climate change will lead to yield decreases in most study regions except South India and areas in central Kenya, as detrimental temperature effects overcome the positive effects of CO2.

These studies use AgMIP representative agricultural pathways (RAPs) as a way to involve stakeholders in regional planning and climate resilience (Valdivia et al. 2015)1427. RAPs are consistent with and complement the RCP/SSP approaches for use in agricultural model intercomparisons, improvement, and impact assessments.

New methods have been developed for improving analysis of climate change impacts and adaptation options for the livestock component of smallholder farming systems in Zimbabwe (Descheemaeker et al. 2018383). These methods utilised disaggregated climate scenarios, as well as differentiating farms with larger stocking rates compared to less densely stocked farms. By disaggregating climate scenarios, impacts, and smallholder farmer attributes, such assessments can more effectively inform decision-making towards climate change adaptation.

In Central Asia, a study using the bio-economic farm model (BEFM) found large differences in projected climate change impact ranging from positive income gains in large-scale commercial farms in contrast to negative impacts in small-scale farms (Bobojonov and Aw-Hassan 2014384). Negative impacts may be exacerbated if irrigation water availability declines due to climate change and increased water demand in upstream regions. In Iran, changes in rainfall and water endowments are projected to significantly impact crop yield and water requirements, as well as income and welfare of farm families (Karimi et al. 2018385).

Climate change impacts on food, feed and cash crops other than cereals, often grown in smallholder systems or family farms are less often studied, although impacts can be substantial. For example, areas suitable for growing coffee are expected to decrease by 21% in Ethiopia with global warming of 2.4°C (Moat et al. 2017) and more than 90% in Nicaragua (Läderach et al. 2017386) with 2.2°C local temperature increase.

Climate change can modify the relationship between crops and livestock in the landscape, affecting mixed crop-livestock systems in many places. Where crop production will become marginal, livestock may provide an alternative to cropping. Such transitions could occur in up to 3% of the total area of Africa, largely as a result of increases in the probability of season failure in the drier mixed crop–livestock systems of the continent (Thornton et al. 2014387).

In Mexico, subsistence agriculture is expected to be the most vulnerable to climate change, due to its intermittent production and reliance on maize and beans (Monterroso et al. 2014388). Overall, a decrease in suitability and yield is expected in Mexico and Central America for beans, coffee, maize, plantain and rice (Donatti et al. 2018389). Municipalities with a high proportional area under subsistence crops in Central America tend to have less resources to promote innovation and action for adaptation (Bouroncle et al. 2017390).

In summary, smallholder farmers are especially vulnerable to climate change because their livelihoods often depend primarily on agriculture. Further, smallholder farmers often suffer from chronic food insecurity (high confidence). Climate change is projected to exacerbate risks of pests and diseases and extreme weather events in smallholder farming systems.

5.2.3

Climate change impacts on access

Access to food involves the ability to obtain food, including the ability to purchase food at affordable prices.

5.2.3.1

Impacts on prices and risk of hunger

A protocol-based analysis based on AgMIP methods tested a combination of RCPs and SSPs to provide a range of projections for prices, risk of hunger, and land-use change (Hasegawa et al. 2018391) (Figure 5.7 and Supplementary Material Table SM5.4.). Previous studies have found that decreased agricultural productivity will depress agricultural supply, leading to price increases. Despite different economic models with various representations of the global food system (Valin et al. 2014392; Robinson et al. 2014393; Nelson et al. 2013394; Schmitz et al. 2014395), as well as having represented the SSPs in different ways, for example, technological change, land-use policies, and sustainable diets (Stehfest et al. 2019396; Hasegawa et al. 2018397), the ensemble of participating models projected a 1–29% cereal price increase in 2050 across SSPs 1, 2 and 3 due to climate change (RCP 6.0). This would impact consumers globally through higher food prices, though regional effects will vary. The median cereal price increase was 7%, given current projections of demand. In all cases (across SSPs and global economic models), prices are projected to increase for rice and coarse grains, with only one instance of a price decline (–1%) observed for wheat in SSP1, with price increases projected in all other cases. Animal-sourced foods (ASFs) are also projected to see price increases (1%), but the range of projected price changes are about half those of cereals, highlighting that the climate impacts on ASFs will be felt indirectly, through the cost and availability of feed, and that there is significant scope for feed substitution within the livestock sector.

Declining food availability caused by climate change is likely to lead to increasing food cost impacting consumers globally through higher prices and reduced purchasing power, with low-income consumers particularly at risk from higher food prices (Nelson et al. 2010398; Springmann et al. 2016a399 and Nelson et al. 2018). Higher prices depress consumer demand, which in turn will not only reduce energy intake (calories) globally (Hasegawa et al. 2015400; Nelson et al. 2010401; Springmann et al. 2016a402 and Hasegawa et al. 2018403), but will also likely lead to less healthy diets with lower availability of key micronutrients (Nelson et al. 2018404) and increase diet-related mortality in lower and middle-income countries (Springmann et al. 2016a405). These changes will slow progress towards the eradication of malnutrition in all its forms.

The extent that reduced energy intake leads to a heightened risk of hunger varies by global economic model. However, all models project an increase in the risk of hunger, with the median projection of an increase in the population at risk of insufficient energy intake by 6, 14, and 12% in 2050 for SSPs 1, 2 and 3 respectively compared to a no climate change reference scenario. This median percentage increase would be the equivalent of 8, 24 and 80 million (full range 1–183 million) additional people at risk of hunger due to climate change (Hasegawa et al. 2018406).

5.2.3.2

Impacts on land use

Climate change is likely to lead to changes in land use globally (Nelson et al. 2014407; Schmitz et al. 2014408 and Wiebe et al. 2015409). Hasegawa et al. (2018)410 found that declining agricultural productivity broadly leads to the need for additional cropland, with 7 of 8 models projecting increasing cropland and the median increase by 2050 projected across all models of 2% compared to a no climate change reference (Figure 5.7). Not all regions will respond to climate impacts equally, with more uncertainty on regional land-use change across the model ensemble than the global totals might suggest. For example, the median land-use change for Latin America is an increase of cropland by 3%, but the range across the model ensemble is significant, with three models projecting declines in cropland (–25 to –1%) compared to the five models projecting cropland increase (0–5%). For further discussion on land-use change and food security see Section 5.6.

5.2.4

Climate change impacts on food utilisation

Food utilisation involves nutrient composition of food, its preparation, and overall state of health. Food safety and quality affects food utilisation.

5.2.4.1

Impacts on food safety and human health

Climate change can influence food safety through changing the population dynamics of contaminating organisms due to, for example, changes in temperature and precipitation patterns, humidity, increased frequency and intensity of extreme weather events, and changes in contaminant transport pathways. Changes in food and farming systems, for example, intensification to maintain supply under climate change, may also increase vulnerabilities as the climate changes (Tirado et al. 2010411).

Climate-related changes in the biology of contaminating organisms include changing the activity of mycotoxin-producing fungi, changing the activity of microorganisms in aquatic food chains that cause disease (e.g., dinoflagellates, bacteria like Vibrio), and increasingly heavy rainfall and floods causing contamination of pastures with enteric microbes (like Salmonella) that can enter the human food chain. Degradation and spoilage of products in storage and transport can also be affected by changing humidity and temperature outside of cold chains, notably from microbial decay but also from potential changes in the population dynamics of stored product pests (e.g., mites, beetles, moths) (Moses et al. 2015412).

Mycotoxin-producing fungi occur in specific conditions of temperature and humidity, so climate change will affect their range, increasing risks in some areas (such as mid-temperate latitudes) and reducing them in others (e.g., the tropics) (Paterson and Lima 2010413). There is robust evidence from process-based models of particular species (Aspergillus/Aflatoxin B1, Fusarium/deoxynivalenol), which include projections of future climate that show that aflatoxin contamination of maize in Southern Europe will increase significantly (Battilani et al. 2016414), and deoxynivalenol contamination of wheat in Northwestern Europe will increase by up to three times current levels (van der Fels-Klerx et al. 2012b, a)1429

Whilst downscaled climate models make any specific projection for a given geography uncertain (Van der Fels-Klerx et al. 2013)1430, experimental evidence on the small scale suggests that the combination of rising CO2 levels, affecting physiological processes in photosynthetic organisms, and temperature changes, can be significantly greater than temperature alone (Medina et al. 2014415). Risks related to aflatoxins are likely to change, but detailed projections are difficult because they depend on local conditions (Vaughan et al. 2016416).

Foodborne pathogens in the terrestrial environment typically come from enteric contamination (from humans or animals), and can be spread by wind (blowing contaminated soil) or flooding – the incidence of both of which are likely to increase with climate change (Hellberg and Chu 2016417). Furthermore, water stored for irrigation, which may be increased in some regions as an adaptation strategy, can become an important route for the spread of pathogens (as well as other pollutants). Contaminated water and diarrheal diseases are acute threats to food security (Bond et al. 2018418). Whilst there is little direct evidence (in terms of modelled projections) the results of a range of reviews, as well as expert groups, suggest that risks from foodborne pathogens are likely to increase through multiple mechanisms (Tirado et al. 2010419; van der Spiegel et al. 2012420; Liu et al. 2013421; Kirezieva et al. 2015422; Hellberg and Chu 2016423).

An additional route to climate change impacts on human health can arise from the changing biology of plants altering human exposure levels. This may include climate changing how crops sequester heavy metals (Rajkumar et al. 2013424), or how they respond to changing pest pressure (e.g., cassava produces hydrogen cyanide as a defence against herbivore attack).

All of these factors will lead to regional differences regarding food safety impacts (Paterson and Lima 2011425). For instance, in Europe it is expected that most important food safety-related impacts will be mycotoxins formed on plant products in the field or during storage; residues of pesticides in plant products affected by changes in pest pressure; trace elements and/or heavy metals in plant products depending on changes in abundance and availability in soils; polycyclic aromatic hydrocarbons in foods following changes in long-range atmospheric transport and deposition; and presence of pathogenic bacteria in foods following more frequent extreme weather, such as flooding and heat waves (Miraglia et al. 2009426).

In summary, there is medium evidence, with high agreement that food utilisation via changes in food safety (and potentially food access from food loss) will be impacted by climate change, mostly by increasing risks, but there is low confidence, exactly how they may change for any given place.

5.2.4.2

Impacts on food quality

There are two main routes by which food quality may change. First, the direct effects of climate change on plant and animal biology, such as through changing temperatures changing the basic metabolism of plants. Secondly, by increasing carbon dioxide’s effect on biology through CO2 fertilisation.

Direct effects on plant and animal biology. Climate affects a range of biological processes, including the metabolic rate in plants and ectothermic animals. Changing these processes can change growth rates, and therefore yields, but can also cause organisms to change relative investments in growth vs reproduction, and therefore change the nutrients assimilated. This may decrease protein and mineral nutrient concentrations, as well as alter lipid composition (DaMatta et al. 2010427). For example, apples in Japan have been exposed to higher temperatures over 3–4 decades and have responded by blooming earlier. This has led to changes in acidity, firmness, and water content, reducing quality (Sugiura et al. 2013428). In other fruit, such as grapes, warming-induced changes in sugar composition affect both colour and aroma (Mira de Orduña 2010429). Changing heat stress in poultry can affect yield as well as meat quality (by altering fat deposition and chemical constituents), shell quality of eggs, and immune systems (Lara and Rostagno 2013430).

Effects of rising CO2 concentrations. Climate change is being driven by rising concentrations of carbon dioxide and other GHG’s in the atmosphere. As plants use CO2 in photosynthesis to form sugar, rising CO2 levels, all things being equal, enhances the process unless limited by water or nitrogen availability. This is known as ‘CO2 fertilisation’. Furthermore, increasing CO2 allows stomata to partially close during gas exchange, reducing water loss through transpiration. These two factors affect the metabolism of plants, and, as with changing temperatures, affects plant growth rates, yields and their nutritional quality. Studies of these effects include meta-analyses, modelling, and small-scale experiments (Franzaring et al. 2013431; Mishra and Agrawal 2014432; Myers et al. 2014433; Ishigooka et al. 2017434; Zhu et al. 2018435; Loladze 2014 and Yu et al. 2014436).

With regard to nutrient quality, a meta-analysis from seven Free-Air Carbon dioxide Enrichment (FACE), (with elevated atmospheric CO2 concentration of 546–586 ppm) experiments (Myers et al. 2014), found that wheat grains had 9.3% lower zinc (CI 5.9–12.7%), 5.1% lower iron (CI 3.7–6.5%) and 6.3% lower protein (CI 5.2–7.5%), and rice grains had 7.8% lower protein content (CI 6.8–8.9%). Changes in nutrient concentration in field pea, soybean and C4 crops such as sorghum and maize were small or insignificant. Zhu et al. (2018)437 report a meta-analysis of FACE trials on a range of rice cultivars. They show that protein declines by an average of 10% under elevated CO2, iron and zinc decline by 8% and 5% respectively. Furthermore, a range of vitamins show large declines across all rice cultivars, including B1 (–17%), B2 (–17%), B5 (–13%) and B9 (–30%), whereas vitamin E increased. As rice underpins the diets of many of the world’s poorest people in low-income countries, especially in Asia, Zhu et al. (2018) estimate that these changes under high CO2 may affect the nutrient status of about 600 million people.

Decreases in protein concentration with elevated CO2 are related to reduced nitrogen concentration possibly caused by nitrogen uptake not keeping up with biomass growth, an effect called ‘carbohydrate dilution’ or ‘growth dilution’, and by inhibition of photorespiration which can provide much of the energy used for assimilating nitrate into proteins (Bahrami et al. 2017438). Other mechanisms have also been postulated (Feng et al. 2015439; Bloom et al. 2014440; Taub and Wang 2008441). Together, the impacts on protein availability may take as many as 150 million people into protein deficiency by 2050 (Medek et al. 2017442). Legume and vegetable yields increased with elevated CO2 concentration of 250 ppm above ambient by 22% (CI 11.6–32.5%), with a stronger effect on leafy vegetables than on legumes and no impact for changes in iron, vitamin C or flavonoid concentration (Scheelbeek et al. 2018443).

Increasing concentrations of atmospheric CO2 lower the content of zinc and other nutrients in important food crops. Dietary deficiencies of zinc and iron are a substantial global public health problem (Myers et al. 2014444). An estimated two billion people suffer these deficiencies (FAO 2013a445), causing a loss of 63million life-years annually (Myers et al. 2014446). Most of these people depend on C3 grain legumes as their primary dietary source of zinc and iron. Zinc deficiency is currently responsible for large burdens of disease globally, and the populations who are at highest risk of zinc deficiency receive most of their dietary zinc from crops (Myers et al. 2015447). The total number of people estimated to be placed at new risk of zinc deficiency by 2050 is 138 million. The people likely to be most affected live in Africa and South Asia, with nearly 48 million residing in India alone. Differences between cultivars of a single crop suggest that breeding for decreased sensitivity to atmospheric CO2 concentration could partly address these new challenges to global health (Myers et al. 2014448).

In summary, while increased CO2 is projected to be beneficial for crop productivity at lower temperature increases, it is projected to lower nutritional quality (e.g., less protein, zinc, and iron) (high confidence).

5.2.5

Climate change impacts on food stability

Food stability is related to people’s ability to access and use food in a steady way, so that there are not intervening periods of hunger. Increasing extreme events associated with climate change can disrupt food stability (see Section 5.8.1 for assessment of food price spikes).

5.2.5.1

Impacts of extreme events

FAO et al. (2018)449 conducted an analysis of the prevalence of undernourishment (PoU) and found that in 2017, the average of the PoU was 15.4% for all countries exposed to climate extremes (Supplementary Material Figure SM5.2). At the same time, the PoU was 20% for countries that additionally show high vulnerability of agriculture production/yields to climate variability, or 22.4% for countries with high PoU vulnerability to severe drought. When there is both high vulnerability of agriculture production/yields and high PoU sensitivity to severe drought, the PoU is 9.8 points higher (25.2%). These vulnerabilities were found to be higher when countries had a high dependence on agriculture as measured by the number of people employed in the sector. Bangkok experienced severe flooding in 2011–2012 with large-scale disruption of the national food supply chains since they were centrally organised in the capital city (Allen et al. 2017450).

The IPCC projects that frequency, duration, and intensity of some extreme events will increase in the coming decades (IPCC 2018a451, 2012452). To test these effects on food security, Tigchelaar et al. (2018)453 showed rising instability in global grain trade and international grain prices, affecting especially the about 800 million people living in extreme poverty who are most vulnerable to food price spikes (Section 5.8.1). They used global datasets of maize production and climate variability combined with future temperature projections to quantify how yield variability will change in the world’s major maize-producing and exporting countries under 2°C and 4°C of global warming.

Tesfaye et al. (2017)454 projected that the extent of heat-stressed areas in South Asia could increase by up to 12% in 2030 and 21% in 2050 relative to the baseline (1950–2000). Another recent study found that drier regions are projected to dry earlier, more severely and to a greater extent than humid regions, with the population of Sub-Saharan Africa most vulnerable (Lickley and Solomon 2018455).

5.2.5.2

Food aid

Food aid plays an important role in providing food security and saving lives after climate disasters. In 2015, 14.5 million people were assisted through disaster-risk reduction, climate change and/ or resilience building activities (WFP 2018456). However, there is no agreement on how to better use emergency food aid, since it can come with unintended consequences for individuals, groups, regions, and countries (Barrett 2006457). These may include negative dependency of food recipients (Lentz et al. 2005458) or price increases, among others.

Some authors state that tied food aid provided as ‘in kind’ by the donor country hampers local food production (Clay 2006459), although others found no evidence of this (Ferrière and Suwa-Eisenmann 2015460). Untied cash aid can be used to buy food locally or in neighbouring countries, which is cheaper and can contribute to improving the livelihoods of local farmers (Clay 2006461).

Ahlgren et al. (2014)462 found that food aid dependence of Marshall Islands due to climate change impacts can result in poor health outcomes due to the poor nutritional quality of food aid, which may result in future increases of chronic diseases. In this regard, Mary et al. (2018)463 showed that nutrition-sensitive aid can reduce the prevalence of undernourishment.

In summary, based on AR5 and SR15 assessments that the likelihood of extreme weather events will increase, (e.g., increases in heatwaves, droughts, inland flooding, and coastal flooding due to rising sea levels, depending on region) in both frequency and magnitude, decreases in food stability and thus increases in food insecurity will likely rise as well (medium evidence, high agreement).

This chapter assesses the contributions of the entire food system to greenhouse gas (GHG) emissions. Food systems emissions include CO2 and non-CO2 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. 2014680). During the period 2007–2016, global agricultural non-CO2 emissions from crop and livestock activities within the farm gate were 6.2 ± 1.4 GtCO2-eq yr–1 during 2007–2016, with methane (142 ± 42 MtCH4 yr-1, or 4.0 ± 1.2 GtCO2-eq yr-1) contributing in CO2-eq about twice as much as nitrous oxide (8.3 ± 2.5 MtN2O yr-1, or 2.2 ± 0.7 GtCO2-eq yr–1) 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 GtCO2-eq yr–1 (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. 2014681).

Agriculture activities within the farm gate and associated land-use dynamics are therefore responsible for about 11.1 ± 2.9 GtCO2-eq yr-1, 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 CH4 emitted by all sectors, while nitrous dioxide gas emissions within the farm gate represented about three-quarters of the total N2O emitted by all sectors (Tubiello 2019682). In terms of carbon, CO2 emissions from deforestation and peatland degradation linked to agriculture contributed about 10% of the CO2 emitted by all sectors in 2017 (Le Quéré et al. 2018683).

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. 2014684) 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 2011685). More recently, Poore and Nemecek (2018)686, 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 GtCO2-eq yr–1 (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 2013b687).

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)689 estimated emissions from cropland to be in the range of 2–3 GtCO2-eq yr-1, including methane emissions from rice, CO2 emissions from peatland cultivation, and N2O emissions from fertiliser applications. Data from FAOSTAT (2018)690, recomputed to use AR5 GWP values, indicated that cropland emissions from these categories were 3.6 ± 1.2 GtCO2-eq yr–1 over the period 2010–2016. Two-thirds of this were related to peatland degradation, followed by N2O emissions from synthetic fertilisers and methane emissions from paddy rice fields (Tubiello 2019)691. 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)692, not including emissions related to deforestation or changes in soil carbon.

Emissions from livestock include non-CO2 gases from enteric fermentation from ruminant animals and from anaerobic fermentation in manure management processes, as well as non-CO2 gases from manure deposited on pastures (Smith et al. 2014693). Estimates after the AR5 include those from Herrero et al. (2016)694, who quantified non-CO2 emissions from livestock to be in the range of 2.0–3.6 GtCO2-eq yr-1, with enteric fermentation from ruminants being the main contributor. FAOSTAT (2018)695 estimates of these emissions, renormalized to AR5 GWP values, were 4.1 ± 1.2 GtCO2-eq yr–1 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 GtCO2-eq yr–1 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 2018696; Davis et al. 2015697).

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 2014b698). Yet, the functional unit used in these measurements is highly relevant and may produce different results (Salou et al. 2017699). 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 2011700). 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. 2019701; Salvador et al. 2017702; Salou et al. 2017703).

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. 2019704; Salvador et al. 2017705; Batalla et al. 2015706 and Petersen et al. 2013707). 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. 2015708).

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 N2O 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 N2O 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. 2016709) and from smallholder systems in East Africa (Pelster et al. 2017710). Herrero et al. (2016a) estimated global livestock enteric methane to range from 1.6–2.7 Gt CO2-eq, depending on assumptions of body weight and animal diet.

Emissions from aquaculture and fisheries may represent some 10% of total agriculture emissions, or about 0.58 GtCO2-eq yr–1 (Barange et al. 2018711), with two-thirds being non-CO2 emissions from aquaculture (Hu et al. 2013712; Yang et al. 2015713) 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. 2016714). N2O 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. 2015715). Hu et al. (2012)716 estimated the global N2O emissions from aquaculture in 2009 to be 0.028 GtCO2-eq yr-1, but could increase to 0.114 GtCO2-eq yr–1 (that is 5.72% of anthropogenic N2O–N emissions) by 2030 for an estimated 7.10% annual growth rate of the aquaculture industry. Numbers estimated by Williams and Crutzen (2010)717 were around 0.036 GtCO2-eq yr-1, and suggested that this may rise to more than 0.179 GtCO2-eq yr–1 within 20 years for an estimated annual growth of 8.7%. Barange et al. (2018)718 assessed the contribution of aquaculture to climate change as 0.38 GtCO2-eq yr–1 in 2010, around 7% of those from agriculture.

CO2 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. 2018719).

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. 2014720) 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 2012721). The proportion of upstream/downstream emissions fall significantly for less-intensive and more-localised production systems (Mottet et al. 2017a722).

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 2007723). 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. 2012724). 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. 2015725).

A study conducted by Wakeland et al. (2012)726 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. 2008727); processing of sugar and oils also requires large amounts of energy.

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)728 updated Nelson et al. (2016)729, 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)730, and one on life-cycle assessments (Poore and Nemecek 2018731). Poore and Nemecek (2018)732 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. 2018733). For example, in the USA 4% of food sold (by weight) is beef, which accounts for 36% of food-related emissions (Heller and Keoleian 2015734). 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)735 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 2018a736). Given that higher-income countries typically have higher emissions per capita, results are particularly applicable in such places.

However, Springmann et al. (2018a)737 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).

5.5.1

Supply-side mitigation options

5.5.1.1

Greenhouse gas mitigation in croplands and soils

The mitigation potential of agricultural soils, cropland and grazing land management has been the subject of much research and was thoroughly summarised in the AR5 (Smith et al. 2014779) (see also Chapter 2, Section 2.5.1 and Chapter 6, Section 6.3.1). Key mitigation pathways are related to practices reducing nitrous oxide emissions from fertiliser applications, reducing methane emissions from paddy rice, reducing both gases through livestock manure management and applications, and sequestering carbon or reducing its losses, with practices for improving grassland and cropland management identified as the largest mitigation opportunities. Better monitoring reporting and verification (MRV) systems are currently needed for reducing uncertainties and better quantifying the actual mitigation outcomes of these activities.

New work since AR5 has focused on identifying pathways for the reductions of GHG emissions from agriculture to help meet Paris Agreement goals (Paustian et al. 2016800 and Wollenberg et al. 2016801). Altieri and Nicholls (2017)802 have characterised mitigation potentials from traditional agriculture. Zomer et al. (2017)803 have updated previous estimates of global carbon sequestration potential in cropland soils. Mayer et al. (2018)804 converted soil carbon sequestration potential through agricultural land management into avoided temperature reductions. Fujisaki et al. (2018) identify drivers to increase soil organic carbon in tropical soils. For discussion of integrated practices such as sustainable intensification, conservation agriculture and agroecology, see Section 5.6.4.

Paustian et al. (2016)805 developed a decision-tree for facilitating implementation of mitigation practices on cropland and described the features of key practices. They observed that most individual mitigation practices will have a small effect per unit of land, and hence they need to be combined and applied at large scales for their impact to be significant. Examples included aggregation of cropland practices(for example, organic amendments, improved crop rotations and nutrient management and reduced tillage) and grazing land practices (e.g., grazing management, nutrient and fire management and species introduction) that could increase net soil carbon stocks while reducing emissions of N2O and CH4.

However, it is well-known that the portion of projected mitigation from soil carbon stock increase (about 90% of the total technical potential) is impermanent. It would be effective for only 20–30 years due to saturation of the soil capacity to sequester carbon, whereas non-CO2 emission reductions could continue indefinitely. ‘Technical potential’ is the maximum amount of GHG mitigation achievable through technology diffusion.

Biochar application and management towards enhanced root systems are mitigation options that have been highlighted in recent literature (Dooley and Stabinsky 2018806; Hawken 2017807; Paustian et al. 2016808; Woolf et al. 2010809 and Lenton 2010810).

5.5.1.2

Greenhouse gas mitigation in livestock systems

The technical options for mitigating GHG emissions in the livestock sector have been the subject of recent reviews (Mottet et al. 2017b811; Hristov et al. 2013a,b; Smithers 2015812; Herrero et al. 2016a813; Rivera-Ferre et al. 2016b814) (Figure 5.11). They can be classified as either targeting reductions in enteric methane; reductions in nitrous oxide through manure management; sequestering carbon in pastures; implementation of best animal husbandry and management practices, which would have an effect on most GHG; and land-use practices that also help sequester carbon. Excluding land-use practices, these options have a technical mitigation potential ranging 0.2–2.4 GtCO2-eq yr–1 (Herrero et al. 2016a815; FAO 2007816) (Chapters 2 and 6.)

The opportunities for carbon sequestration in grasslands and rangelands may be significant (Conant 2010817), for instance, through changes in grazing intensity or manure recycling aimed at maintaining grassland productivity (Hirata et al. 2013818). Recent studies have questioned the economic potential of such practices in regard to whether they could be implement at scale for economic gain (Garnett et al. 2017819; Herrero et al. 2016a820 and Henderson et al. 2015821). For instance, Henderson et al. (2015)822 found economic potentials below 200 MtCO2-eq yr-1. Carbon sequestration can occur in situations where grasslands are highly degraded (Garnett 2016823). Carbon sequestration linked to livestock management could thus be considered as a co-benefit of well-managed grasslands, as well as a mitigation practice.

Different production systems will require different strategies, including the assessment of impacts on food security, and this has been the subject of significant research (e.g., Rivera-Ferre et al. 2016b1447). Livestock systems are heterogeneous in terms of their agroecological orientation (arid, humid or temperate/highland locations), livestock species (cattle, sheep, goats, pigs, poultry and others), structure (grazing only, mixed-crop-livestock systems, industrial systems, feedlots and others), level of intensification, and resource endowment (Robinson 2011824).

The implementation of strategies presented in Figure 5.11 builds on this differentiation, providing more depth compared to the previous AR5 analysis. Manure management strategies are more applicable in confined systems, where manure can be easily collected, such as in pigs and poultry systems or in smallholder mixed crop-livestock systems. More intensive systems, with strong market orientation, such as dairy in the US, can implement a range of sophisticated practices like feed additives and vaccines, while many market-oriented dairy systems in tropical regions can improve feed digestibility by improving forage quality and adding larger quantities of concentrate to the rations. Many of these strategies can be implemented as packages in different systems, thus maximising the synergies between different options (Mottet et al. 2017b825).

See the Supplementary Material Section SM5.5 for a detailed description of livestock mitigation strategies; synergies and trade-offs with other mitigation and adaptation options are discussed in Section 5.6.

5.5.1.3

Greenhouse gas mitigation in agroforestry

Agroforestry can curb GHG emissions of CO2, CH4, and N2O in agricultural systems in both developed and developing countries (see Glossary for definition) (see Chapter 2, Section 2.5.1 and Figure 2.24). Soil carbon sequestration, together with biological N fixation, improved land health and underlying ecosystem services may be enhanced through agricultural lands management practices used by large-scale and smallholder farmers, such as incorporation of trees within farms or in hedges (manure addition, green manures, cover crops, etc.), whilst promoting greater soil organic matter and nutrients (and thus soil organic carbon) content and improve soil structure (Mbow et al. 2014b826) (Table 5.5). The tree cover increases the microbial activity of the soil and increases the productivity of the grass under cover. CO2 emissions are furthermore lessened indirectly, through lower rates of erosion due to better soil structure and more plant cover in diversified farming systems than in monocultures. There is great potential for increasing above-ground and soil carbon stocks, reducing soil erosion and degradation, and mitigating GHG emissions.

These practices can improve food security through increases in productivity and stability since they contribute to increased soil quality and water-holding capacity. Agroforestry provides economic, ecological, and social stability through diversification of species and products. On the other hand, trade-offs are possible when cropland is taken out of production mainly as a mitigation strategy.

Meta-analyses have been done on carbon budgets in agroforestry systems (Zomer et al. 2016827; Chatterjee et al. 2018828). In a review of 42 studies, (Ramachandran Nair et al. 2009829) estimated carbon sequestration potentials of differing agroforestry systems. These include sequestration rates ranging from 954 (semi-arid); to 1431 (temperate); 2238 (sub-humid) and 3670 tCO2 km–2 yr–1 (humid). The global technical potential for agroforestry is 0.1–5.7 Gt CO2e yr–1 (Griscom et al. 2017830; Zomer et al. 2016831; Dickie et al. 2014832) (Chapter 2, Section 2.5.1). Agroforestry-based carbon sequestration can be used to offset N2O and CO2 emissions from soils and increase methane sink strength compared to annual cropping systems (Rosenstock et al. 2014833).

Agroforestry systems with perennial crops, such as coffee and cacao, may be more important carbon sinks than those that combine trees with annual crops. Brandt et al. (2018)834 showed that farms in semi-arid regions (300–600 mm precipitation) were increasing in tree cover due to natural regeneration and that the increased application of agroforestry systems were supporting production and reducing GHG emissions.

5.5.1.4

Integrated approaches to crop and livestock mitigation

Livestock mitigation in a circular economy. Novel technologies for increasing the integration of components in the food system are being devised to reduce GHG emissions. These include strategies that help decoupling livestock from land use. Work by van Zanten et al. (20181451)shows that 7–23 g of animal protein per capita per day could be produced without livestock competing for vital arable land. This would imply a contraction of the land area utilised by the livestock sector, but also a more efficient use of resources, and would lead to land sparing and overall emissions reductions.

Pikaar et al. (2018)836 demonstrated the technical feasibility of producing microbial protein as a feedstuff from sewage that could replace use of feed crops such as soybean. The technical potential of this novel practice could replace 10–19% of the feed protein required, and would reduce cropland demand and associated emissions by 6–7%. These practices are, however, not economically feasible nor easily upscalable in most systems. Nonetheless, significant progress in Japan and South Korea in the reduction and use of food waste to increase efficiencies in livestock food chains has been achieved, indicating a possible pathway to progress elsewhere (FAO 2017837; zu Ermgassen et al. 2016838). Better understanding of biomass and food and feed wastes, value chains, and identification of mechanisms for reducing the transport and processing costs of these materials is required to facilitate larger-scale implementation.

Waste streams into energy. Waste streams from manure and food waste can be used for energy generation and thus reduction in overall GHG emissions in terms of recovered methane (for instance through anaerobic digestion) production (De Clercq et al. 2016839) or for the production of microbial protein (Pikaar et al. 2018840). Second-generation biorefineries, once the underlying technology is improved, may enable the generation of hydro-carbon from agricultural residues, grass, and woody biomass in ways that do not compete with food and can generate, along with biofuel, high-value products such as plastics (Nguyen et al. 2017841). Second-generation energy biomass from residues may constitute a complementary income source for farmers that can increase their incentive to produce. Technologies include CHP (combined heat and power) or gas turbines, and fuel types such as biodiesel, biopyrolysis (i.e., high temperature chemical transformation of organic material in the absence of oxygen), torrefaction of biomass, production of cellulosic bioethanol and of bioalcohols produced by other means than fermentation, and the production of methane by anaerobic fermentation. (Nguyen et al. 2017842).

Technology for reducing fossil fuel inputs. Besides biomass and bioenergy, other forms of renewable energy substitution for fossil fuels (e.g., wind, solar, geothermal, hydro) are already being applied on farms throughout the supply chain. Energy efficiency measures are being developed for refrigeration, conservation tillage, precision farming (e.g., fertiliser and chemical application and precision irrigation).

Novel technologies. Measures that can reduce livestock emissions given continued research and development include methane and nitrification inhibitors, methane vaccines, targeted breeding of lower-emitting animals, and genetically modified grasses with higher sugar content. New strategies to reduce methanogenesis include supplementing animal diets with antimethanogenic agents (e.g., 3-NOP, algae, chemical inhibitors such as chloroform) or supplementing with electron acceptors (e.g., nitrate) or dietary lipids. These could potentially contribute, once economically feasible at scale, to significant reductions of methane emissions from ruminant livestock. A well-tested compound is 3-nitrooxypropanol (3-NOP), which was shown to decrease methane by up to 40% when incorporated in diets for ruminants (Hristov et al. 2015843).

Whilst these strategies may become very effective at reducing methane, they can be expensive and also impact on animal performance and/or welfare (Llonch et al. 2017844). The use of novel fertilisers and/or plant species that secrete biological nitrification inhibitors also have the potential to significantly reduce N2O emissions from agricultural soils (Subbarao et al. 2009845; Rose et al. 2018846).

Economic mitigation potentials of crop and livestock sectors. Despite the large technical mitigation potential of the agriculture sector in terms of crop and livestock activities, its economic potential is relatively small in the short term (2030) and at modest carbon prices (less than 20 USD tC–1). For crop and soil management practices, it is estimated that 1.0–1.5 GtCO2-eq yr–1 could be a feasible mitigation target at a carbon price of 20 USD tC–1 (Frank et al. 2018, 2017; Griscom et al. 2016847; Smith et al. 2013848; Wollenberg et al. 2016849). For the livestock sector, these estimates range from 0.12–0.25 GtCO2-eq yr–1 at similar carbon prices (Herrero et al. 2016c850; Henderson et al. 2017851).

But care is needed in comparing crop and livestock economic mitigation potentials due to differing assumptions.

Frank et al. (2018)852 recently estimated that the economic mitigation potential of non-CO2 emissions from agriculture and livestock to 2030 could be up to four times higher than indicated in the AR5, if structural options such as switching livestock species from ruminants to monogastrics, or allowing for flexibility to relocate production to more efficient regions were implemented, at the same time as the technical options such as those described above. At higher carbon prices (i.e., at about 100 USD tC–1), they found a mitigation potential of supply-side measures of 2.6 GtCO2-eq yr-1.

In this scenario, technical options would account for 38% of the abatement, while another 38% would be obtained through structural changes, and a further 24% would be obtained through shifts in consumption caused by food price increases. Key to the achievement of this mitigation potential lay in the livestock sector, as reductions in livestock consumption, structural changes and implementation of technologies in the sector had some of the highest impacts. Regions with the highest mitigation potentials were Latin America, China and Sub-Saharan Africa. The large-scale implementability of such proposed sweeping changes in livestock types and production systems is likely very limited as well as constrained by long-established socio-economic, traditional and cultural habits, requiring significant incentives to generate change.

In summary, supply-side practices can contribute to climate change mitigation by reducing crop and livestock emissions, sequestering carbon in soils and biomass, and by decreasing emissions intensity within sustainable production systems (high confidence). The AR5 estimated the total economic mitigation potential of crop and livestock activities as 1.5–4.0 GtCO2-eq yr–1 by 2030 at prices ranging from 20–100 USD tCO2-eq (high confidence). Options with large potential for GHG mitigation in cropping systems include soil carbon sequestration (at decreasing rates over time), reductions in N2O emissions from fertilisers, reductions in CH4 emissions from paddy rice, and bridging of yield gaps. Options with large potential for mitigation in livestock systems include better grazing land management, with increased net primary production and soil carbon stocks, improved manure management, and higher-quality feed. Reductions in GHG emissions intensity (emissions per unit product) from livestock can support reductions in absolute emissions, provided appropriate governance structures to limit total production are implemented at the same time (medium confidence).

5.5.1.5

Greenhouse gas mitigation in aquaculture

Barange et al. (2018)853 provide a synthesis of effective options for GHG emissions reduction in aquaculture, including reduction of emissions from production of feed material, replacement of fish-based feed ingredients with crop-based ingredients; reduction of emissions from feed mill energy use, improvement of feed conversion rates, improvement of input use efficiency, shift of energy supply (from high-carbon fossil fuels to low-carbon fossil fuels or renewables), and improvement of fish health. Conversion of 25% of total aquaculture area to integrated aquaculture-agriculture ponds (greening aquaculture) has the potential to sequester 95.4 million tonnes of carbon per year (Ahmed et al. 2017854).

Proposed mitigation in aquaculture includes avoided deforestation. By halting annual mangrove deforestation in Indonesia, associated total emissions would be reduced by 10–31% of estimated annual emissions from the land-use sector at present (Murdiyarso et al. 2015855). Globally, 25% mangrove regeneration could sequester 0.54–0.65 million tonnes of carbon per year (Ahmed et al. 2017856) of which 0.17–0.21 million tonnes could be through integrated or organic shrimp culture (Ahmed et al. 2018857).

5.5.1.6

Cellular agriculture

The technology for growing muscle tissue in culture from animal stem cells to produce meat, for example, ‘cultured’, ‘synthetic’, ‘in vitro’ or ‘hydroponic’ meat could, in theory, be constructed with different characteristics and be produced faster and more efficiently than traditional meat (Kadim et al. 2015858). Cultured meat (CM) is part of so-called cellular agriculture, which includes production of milk, egg white and leather from industrial cell cultivation (Stephens et al. 2018859). CM is produced from muscle cells extracted from living animals, isolation of adult skeletal muscle stem cells (myosatellite cells), placement in a culture medium which allow their differentiation into myoblasts and then, through another medium, generation of myocytes which coalesce into myotubes and grow into strands in a stirred-tank bioreactor (Mattick et al. 2015860).

Current technology enables the creation of beef hamburgers, nuggets, steak chips or similar products from meat of other animals, including wild species, although production currently is far from being economically feasible. Nonetheless, by allowing bioengineering from the manipulation of the stem cells and nutritive culture, CM allows for reduction of harmful fatty acids, with advantages such as reduced GHG emissions, mostly indirectly through reduced land use (Bhat et al. 2015861; Kumar et al. 2017b862).

Tuomisto and de Mattos (2011)863 made optimistic technological assumptions, relying on cyanobacteria hydrolysate nutrient source, and produced the lowest estimates on energy and land use. Tuomisto and de Mattos (2011)864 conducted a lifecycle assessment that indicates that cultured meat could have less than 60% of energy use and 1% of land use of beef production and it would have lower GHG emissions than pork and poultry as well. Newer estimates (Alexander et al. 2017865; Mattick et al. 2015866) indicate a trade-off between industrial energy consumption and agricultural land requirements of conventional and cultured meat and possibly higher GWP than pork or poultry due to higher energy use. The change in proportion of CO2 versus CH4 could have important implications in climate change projections and, depending on decarbonisation of the energy sources and climate change targets, cultured meat may be even more detrimental than exclusive beef production (Lynch and Pierrehumbert 2019867).

Overall, as argued by Stephens et al. (2018)868, cultured meat is an ‘as-yet undefined ontological object’ and, although marketing targets people who appreciate meat but are concerned with animal welfare and environmental impacts, its market is largely unknown (Bhat et al. 2015 and Slade 2018869). In this context it will face the competition of imitation meat (meat analogues from vegetal protein) and insect-derived products, which have been evaluated as more environmentally friendly (Alexander et al. 2017870) and it may be considered as being an option for a limited resource world, rather than a mainstream solution. Besides, as the commercial production process is still largely undefined, its actual contribution to climate change mitigation and food security is largely uncertain and challenges are not negligible. Finally, it is important to understand the systemic nature of these challenges and evaluate their social impacts on rural populations due to transforming animal agriculture into an industrialised activity and its possible rebound effects on food security, which are still understudied in the literature.

Studies are needed to improve quantification of mitigation options for supply chain activities.

5.5.2

Demand-side mitigation options

5.5.2.1

Mitigation potential of different diets

A systematic review found that higher consumption of animal-based foods was associated with higher estimated environmental impact, whereas increased consumption of plant-based foods was associated with an estimated lower environmental impact (Nelson et al. 2016871). Assessment of individual foods within these broader categories showed that meat – especially ruminant meat (beef and lamb) – was consistently identified as the single food with the greatest impact on the environment, on a global basis, most often in terms of GHG emissions and/or land use.

Figure 5.12 shows the technical mitigation potentials of some scenarios of alternative diets examined in the literature. Stehfest et al. (2009)872 were among the first to examine these questions. They found that under the most extreme scenario, where no animal products are consumed at all, adequate food production in 2050 could be achieved on less land than is currently used, allowing considerable forest regeneration, and reducing land-based GHG emissions to one third of the reference ‘business-as-usual’ case for 2050, a reduction of 7.8 GtCO2-eq yr-1. Springmann et al. (2016b)873 recently estimated similar emissions reduction potential of 8 GtCO2-eq yr–1 from a vegan diet without animal-sourced foods. This defines the upper bound of the technical mitigation potential of demand side measures.

Herrero et al. (2016a)874 reviewed available options, with a specific focus on livestock products, assessing technical mitigation potential across a range of scenarios, including ‘no animal products’, ‘no meat’, ‘no ruminant meat’, and ‘healthy diet’ (reduced meat consumption). With regard to ‘credible low-meat diets’, where reduction in animal protein intake was compensated by higher intake of pulses, emissions reductions by 2050 could be in the4.3–6.4GtCO2-eqyr-1 , compared to a business-as-usual scenario.

Of this technical potential, 1–2 GtCO2-eq yr–1 come from reductions of mostly non-CO2 GHG within the farm gate, while the remainder was linked to carbon sequestration on agricultural lands no longer needed for livestock production. When the transition to a low-meat diet reduces the agricultural area required, land is abandoned, and the re-growing vegetation can take up carbon until a new equilibrium is reached. This is known as the land-sparing effect.

Other studies have found similar results for potential mitigation linked to diets. For instance, Smith et al. (2013) analysed a dietary change scenario that assumed a convergence towards a global daily per-capita calorie intake of 2800 kcal per person per day (11.7 MJ per person per day), paired with a relatively low level of animal product supply, estimated technical mitigation potential in the range 0.7–7.3 GtCO2-eq yr–1 for additional variants including low or high-yielding bioenergy, 4.6GtCO2-eqyr–1 if spare land is afforested.

Bajželj et al. (2014)875 developed different scenarios of farm systems change, waste management, and dietary change on GHG emissions coupled to land use. Their dietary scenarios were based on target kilocalorie consumption levels and reductions in animal product consumption. Their scenarios were ‘healthy diet’; healthy diet with 2500 kcal per person per day in 2050; corresponding to technical mitigation potentials in the range 5.8 and 6.4 GtCO2-eq yr-1.

Hedenus et al. (2014) explored further dietary variants based on the type of livestock product. ‘climate carnivore’, in which 75% of the baseline-consumption of ruminant meat and dairy was replaced

by pork and poultry meat, and ‘flexitarian’, in which 75% of the baseline-consumption of meat and dairy was replaced by pulses and cereal products. Their estimates of technical mitigation potentials by 2050 ranged 3.4–5.2 GtCO2-eq yr-1, the high end achieved under the flexitarian diet. Finally, Tilman and Clark (2014) used stylised diets as variants that included ‘peseatarian’, ‘Mediterranean’, ‘vegetarian’, compared to a reference diet, and estimated technical mitigation potentials within the farm gate of 1.2–2.3 GtCO2-eq yr-1, with additional mitigation from carbon sequestration on spared land ranging 1.8–2.4 GtCO2-eq yr-1.

Studies have defined dietary mitigation potential as, for example, 20 kg per person per week CO2-eq for Mediterranean diet, versus 13 kg per person per week CO2-eq for vegan (Castañé and Antón 2017876). Rosi et al. (2017)877 developed seven-day diets in Italy for about 150 people defined as omnivore 4.0 ± 1.0; ovo-lacto-veggie 2.6 ± 0.6; and vegan 2.3 ± 0.5 kg CO2-eq per capita per day.

Importantly, many more studies that compute the economic and calorie costs of these scenarios are needed. Herrero et al. (2016a)878 estimated that once considerations of economic and calorie costs of their diet-based solutions were included, the technical range of 4.3–6.4 GtCO2-eq yr–1 in 2050 was reduced to 1.8–3.4 GtCO2-eq yr–1 when implementing a GHG tax ranging from 20–100 USD tCO2. While caloric costs where low below 20 USD tCO2, they ranged from 27–190 kcal per person per day under the higher economic potential, thus indicating possible negative trade-offs with food security.

In summary, demand-side changes in food choices and consumption can help to achieve global GHG mitigation targets (high confidence). Low-carbon diets on average tend to be healthier and have smaller land footprints. By 2050, technical mitigation potential of dietary changes range from 2.7–6.4 GtCO2-eq yr–1 for assessed diets (high confidence). At the same time, the economic potential of such solutions is lower, ranging from 1.8–3.4 GtCO2-eq yr–1 at prices of 20–100 USD tCO2, with caloric costs up to 190 kcal per person per day. The feasibility of how to create economically viable transitions to more sustainable and healthy diets that also respect food security requirements needs to be addressed in future research.

5.5.2.2

Role of dietary preferences

Food preference is an inherently cultural dimension that can ease or hinder transformations to food systems that contribute to climate change mitigation. Consumer choice and dietary preferences are guided by social, cultural, environmental, and traditional factors as well as economic growth. The food consumed by a given group conveys cultural significance about social hierarchy, social systems and human-environment relationships (Herforth and Ahmed 2015).

As suggested by Springmann et al. (2018a), per capita dietary emissions will translate into different realised diets, according to regional contexts including cultural and gendered norms (e.g., among some groups, eating meat is perceived as more masculine (Ruby and Heine 2011). In some cases, women and men have different preferences in terms of food, with women reporting eating healthier food (Imamura et al. 2015879; Kiefer et al. 2005880; Fagerli and Wandel 1999881): these studies found that men tend to eat more meat, while women eat more vegetables, fruits and dairy products (Kanter and Caballero 2012882).

Food preferences can change over time, with the nutrition transition from traditional diets to high-meat, high-sugar, high-saturated fat diets being a clear example of significant changes occurring in a short period of time. Meat consumption per capita consistently responds to income with a saturating trend at high income levels (Sans and Combris 2015883; Vranken et al. 2014884). Some emerging economies have rapidly increased demand for beef, leading to pressure on natural resources (Bowles et al. 2019885). In another example, by reducing beef consumption between 2005 and 2014, Americans avoided approximately 271 million metric tonnes of emissions (CO2-eq) (NRDC 2017886). Attending farmers markets or buying directly from local producers has been shown to change worldviews (Kerton and Sinclair 2010887), and food habits towards healthier diets (Pascucci et al. 2011888) can be advanced through active learning (Milestad et al. 2010889).

Regarding the options to reduce meat intake in developed countries, research shows that there is an apparent sympathy of consumers for meat reduction due to environmental impacts (Dagevos and Voordouw 2013890), which has not been exploited. Social factors that influence reducing meat consumption in New Zealand include the need for better education or information dispersal regarding perceived barriers to producing meat-reduced/less meals; ensuring there is sensory or aesthetic appeal; and placing emphasis on human health or nutritional benefits (Tucker 2018891).

Different and complementary strategies can be used in parallel for different consumer’s profiles to facilitate step-by-step changes in the amounts and the sources of protein consumed. In the Netherlands, a nationwide sample of 1083 consumers were used to study their dietary choices toward smaller portions of meat, smaller portions using meat raised in a more sustainable manner, smaller portions and eating more vegetable protein, and meatless meals with or without meat substitutes. Results showed that strategies to change meat eating frequencies and meat portion sizes appeared to overlap and that these strategies can be applied to address consumers in terms of their own preferences (de Boer et al. 2014892).

5.5.2.4

Insect-based diets

Edible insects are, in general, rich in protein, fat, and energy and can be a significant source of vitamins and minerals (Rumpold and Schlüter 2015906). Approximately 1900 insect species are eaten worldwide, mainly in developing countries (van Huis 2013907). The development of safe rearing and effective processing methods are mandatory for utilisation of insects in food and feed. Some insect species can be grown on organic side streams, reducing environmental contamination and transforming waste into high-protein feed. Insects are principally considered as meat substitutes, but worldwide meat substitute consumption is still very low, principally due to differences in food culture, and will require transition phases such as powdered forms (Megido et al. 2016908 and Smetana et al. 2015909). Wider consumer acceptability will relate to pricing, perceived environmental benefits, and the development of tasty insect-derived protein products (van Huis et al. 2015910; van Huis 2013911). Clearly, increasing the share of insect-derived protein has the potential to reduce GHG emissions otherwise associated with livestock production. However, no study to date has quantified such potential.

5.5.2.5

Food loss and waste, food security, and land use

Food loss and waste impacts food security by reducing global and local food availability, limiting food access due to an increase in food prices and a decrease of producer income, affecting future food production due to the unstainable use of natural resources (HLPE 2014912). Food loss is defined as the reduction of edible food during production, postharvest, and processing, whereas food discarded by consumers is considered as food waste (FAO 2011b913). Combined food loss and waste amount to 25–30% of total food produced (medium confidence). During 2010–2016, global food loss and waste equalled 8–10% of total GHG emissions (medium confidence); and cost about 1 trillion USD per year (low confidence) (FAO 2014b914).

A large share of produced food is lost in developing countries due to poor infrastructure, while a large share of produced food is wasted in developed countries (Godfray et al. 2010915). Changing consumer behaviour to reduce per capita over-consumption offers substantial potential to improve food security by avoiding related health burdens (Alexander et al. 2017916; Smith 2013917) and reduce emissions associated with the extra food (Godfray et al. 2010918). In 2007, around 20% of the food produced went to waste in Europe and North America, while around 30% of the food produced was lost in Sub-Saharan Africa (FAO 2011b919). During the last 50 years, the global food loss and waste increased from around 540 Mt in 1961 to 1630 Mt in 2011 (Porter et al. 2016920).

In 2011, food loss and waste resulted in about 8–10% of total anthropogenic GHG emissions. The mitigation potential of reduced food loss and waste from a full life-cycle perspective, for example, considering both food supply chain activities and land-use change, was estimated as 4.4 GtCO2-eq yr–1 (FAO 2015a, 2013b921). At a global scale, loss and waste of milk, poultry meat, pig meat, sheep meat, and potatoes are associated with 3% of the global agricultural N2O emissions (more than 200 Gg N2O-N yr–1 or 0.06 GtCO2-eq yr–1) in 2009 (Reay et al. 2012922). For the USA, 35% of energy use, 34% of blue water use, 34% of GHG emissions, 31% of land use, and 35% of fertiliser use related to an individual’s food-related resource consumption were accounted for as food waste and loss in 2010 (Birney et al. 2017923).

Similar to food waste, over-consumption (defined as food consumption in excess of nutrient requirements), leads to GHG emissions (Alexander et al. 2017924). In Australia for example, over-consumption accounts for about 33% GHGs associated with food (Hadjikakou 2017925). In addition to GHG emissions, over-consumption can also lead to severe health conditions such as obesity or diabetes. Over-eating was found to be at least as large a contributor to food system losses (Alexander et al. 2017926). Similarly, food system losses associated with consuming resource-intensive animal-based products instead of nutritionally comparable plant-based alternatives are defined as ‘opportunity food losses’. These were estimated to be 96, 90, 75, 50, and 40% for beef, pork, dairy, poultry, and eggs, respectively, in the USA (Shepon et al. 2018927).

Avoiding food loss and waste will contribute to reducing emissions from the agriculture sector. By 2050, agricultural GHG emissions associated with production of food that might be wasted may increase to 1.9–2.5 GtCO2-eq yr–1 (Hiç et al. 2016928). When land-use change for agriculture expansion is also considered, halving food loss and waste reduces the global need for cropland area by around 14% and GHG emissions from agriculture and land-use change by 22–28% (4.5 GtCO2-eq yr–1) compared to the baseline scenarios by 2050 (Bajželj et al. 2014929). The GHG emissions mitigation potential of food loss and waste reduction would further increase when lifecycle analysis accounts for emissions throughout food loss and waste through all food system activities.

Reducing food loss and waste to zero might not be feasible. Therefore, appropriate options for the prevention and management of food waste can be deployed to reduce food loss and waste and to minimise its environmental consequences. Papargyropoulou et al. (2014930) proposed the Three Rs (i.e., reduction, recovery and recycle) options to prevent and manage food loss and waste. A wide range of approaches across the food supply chain is available to reduce food loss and waste, consisting of technical and non-technical solutions (Lipinski et al. 2013931). However, technical solutions (e.g., improved harvesting techniques, on-farm storage, infrastructure, packaging to keep food fresher for longer, etc.) include additional costs (Rosegrant et al. 2015932) and may have impacts on local environments (FAO 2018b933). Additionally, all parts of food supply chains need to become efficient to achieve the full reduction potential of food loss and waste (Lipinski et al. 2013934).

Together with technical solutions, approaches (i.e., non-technical solutions) to changes in behaviours and attitudes of a wide range of stakeholders across the food system will play an important role in reducing food loss and waste. Food loss and waste can be recovered by distributing food surplus to groups affected by food poverty or converting food waste to animal feed (Vandermeersch et al. 2014935). Unavoidable food waste can also be recycled to produce energy based on biological, thermal and thermochemical technologies (Pham et al. 2015936). Additionally, strategies for reducing food loss and waste also need to consider gender dynamics with participation of females throughout the food supply chain (FAO 2018f937).

In summary, reduction of food loss and waste can be considered as a climate change mitigation measure that provides synergies with food security and land use (robust evidence, medium agreement). Reducing food loss and waste reduces agricultural GHG emissions and the need for agricultural expansion for producing excess food. Technical options for reduction of food loss and waste include improved harvesting techniques, on-farm storage, infrastructure, and packaging. However, the beneficial effects of reducing food loss and waste will vary between producers and consumers, and across regions. Causes of food loss (e.g., lack of refrigeration) and waste (e.g., behaviour) differ substantially in developed and developing countries (robust evidence, medium agreement). Additionally, food loss and waste cannot be avoided completely.

5.5.2.6

Shortening supply chains

Encouraging consumption of locally produced food and enhancing efficiency of food processing and transportation can, in some cases, minimise food loss, contribute to food security, and reduce GHG emissions associated with energy consumption and food loss. For example, Michalský and Hooda (2015)938, through a quantitative assessment of GHG emissions of selected fruits and vegetables in the UK, reported that increased local production offers considerable emissions savings. They also highlighted that when imports are necessary, importing from Europe instead of the Global South can contribute to considerable GHG emissions savings. Similar results were found by Audsley et al. (2010)939, with exceptions for some foods, such as tomatoes, peppers or sheep and goat meat. Similarly, a study in India shows that long and fragmented supply chains, which lead to disrupted price signals, unequal power relations perverse incentives and long transport time, could be a key barrier to reducing post-harvest losses (CIPHET 2007940).

In other cases, environmental benefits associated with local food can be offset by inefficient production systems with high emission intensity and resource needs, such as water, due to local conditions. For example, vegetables produced in open fields can have much lower GHG emissions than locally produced vegetables from heated greenhouses (Theurl et al. 2014941). Whether locally grown food has a lower carbon footprint depends on the on-farm emissions intensity as well as the transport emissions. In some cases, imported food may have a lower carbon footprint than locally grown food because some distant countries can produce food at much lower emissions intensity. For example, Avetisyan et al. (2014)942 reported that regional variation of emission intensities associated with production of ruminant products have large implications for emissions associated with local food. They showed that consumption of local livestock products can reduce emissions due to short supply chains in countries with low emission intensities; however, this might not be the case in countries with high emission intensities.

In addition to improving emission intensity, efficient distribution systems for local food are needed for lowering carbon footprints (Newman et al. 2013943). Emissions associated with food transport depend on the mode of transport, for example, emissions are lower for rail rather than truck (Brodt et al. 2013944). Tobarra et al. (2018)945 reported that emissions saving from local food may vary across seasons and regions of import. They highlighted that, in Spain, local production of fruits and vegetables can reduce emissions associated with imports from Africa but imports from France and Portugal can save emissions in comparison to production in Spain. Additionally, local production of seasonal products in Spain reduces emissions, while imports of out-of-season products can save emissions rather than producing them locally.

In summary, consuming locally grown foods can reduce GHG emissions, if they are grown efficiently (high confidence). The emissions reduction potential varies by region and season. Whether food with shorter supply chains has a lower carbon footprint depends on both the on-farm emissions intensity as well as the transport emissions. In some cases, imported food may have a lower carbon footprint because some distant agricultural regions can produce food at lower emissions intensities.

5.6.1

Land-based carbon dioxide removal (CDR) and bioenergy

5.6.2

Mitigation, food prices, and food security

Food prices are the result of supply, demand and trade relations. Earlier studies (e.g., Nelson et al. 20091008) showed that recent climate impacts that reduced crop productivity led to higher prices and increased trade of commodities between regions, with asymmetric impacts on producers and consumers. In terms of published scenario analyses, the most affected regions tend to be Sub-Saharan Africa and parts of Asia, but there is significant heterogeneity in results between countries. Relocation of production to less affected areas buffers these impacts to a certain extent, and offers potential for improvements in food production technologies (Hasegawa et al. 20181009; van Meijl et al. 20171010; Wiebe et al. 20151011; Lotze-Campen et al. 20141012; Valin et al. 20141013; Robinson et al. 20141014).

A newer, less studied impact of climate change on prices and their impacts on food security is the level of land-based mitigation necessary to stabilise global temperature. Hasegawa et al. (2018)1015, using an ensemble of seven global economic models across a range of GHG emissions pathways and socioeconomic trajectories, suggested that the level of mitigation effort needed to reduce emissions can have a more significant impact on prices than the climate impacts themselves on reduced crop yields (Figure 5.14). This occurs because in the models, taxing GHG emissions leads to higher crop and livestock prices, while land-based mitigation leads to less land availability for food production, potentially lower food supply, and therefore food price increases.

Price increases in turn lead to reduced consumption, especially by vulnerable groups, or to shifts towards cheaper food, which are often less nutritious. This leads to significant increases in the number of malnourished people. Frank et al. (2017)1016 and Fujimori et al. (2017)1017 arrived at the same conclusions for the 1.5°C mitigation scenario using the IAM Globiom and ensembles of AgMIP global economic models. While the magnitude of the response differs between models, the results are consistent between them. In contrast, a study based on five global agroeconomic models highlights that the global food prices may not increase much when the required land for bioenergy is accessible on the margin of current cropland, or the feedstock does not have a direct completion with agricultural land (Lotze-Campen et al. 20141018).

These studies highlight the need for careful design of emissions mitigation policies in upcoming decades – for example, targeted schemes encouraging more productive and resilient agricultural production systems and the importance of incorporating complementary policies (such as safety-net programmes for poverty alleviation) that compensate or counteract the impacts of climate change mitigation policies on vulnerable regions (Hasegawa et al. 20181019). Fujimori et al. (2018) showed how an inclusive policy design can avoid adverse side effects on food security through international aid, bioenergy taxes, or domestic reallocation of income. These strategies can shield impoverished and vulnerable people from the additional risk of hunger that would be caused by the economic effects of policies narrowly focussing on climate objectives only.

In summary, food security will be threatened through increasing numbers of malnourished people if land-based mitigation raises prices, unless other policy mechanisms reduce its impact (high confidence). Inclusive policy design can avoid adverse side effects on food security by shielding vulnerable people from the additional risk of hunger that would be caused by the economic effects of policies narrowly focusing on climate objectives (medium confidence).

5.6.3

Environmental and health effects of adopting healthy and sustainable diets

Two key questions arise from the potentially significant mitigation potential of dietary change: (i) Are ‘low-GHG emission diets’ likely to be beneficial for health? and (ii) Would changing diets at scale provide substantial benefits? In short, what are the likely synergies and trade-offs between low-GHG emissions diets and food security, health, and climate change? See Supplementary Material Section SM5.6 for further discussion.

Are ‘low GHG emission diets’ healthy? Consistent evidence indicates 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 associated with lower environmental impact (GHG emissions and energy, land and water use) than either the current global average diets (Swinburn et al. 20191021; Willett et al. 20191022; Springmann et al. 2016b1023), or the current average USA diet (Nelson et al. 20161024). Another study (Van Mierlo et al. 20171025) showed that nutritionally-equivalent diets can substitute plant-based foods for meat and provide reductions in GHG emissions.

There are several studies that estimate health adequacy and sustainability and conclude that healthy sustainable diets are possible. These include global studies (e.g., Willett et al. 20191026; Swinburn et al. 20191027), as well as localised studies (e.g., Van Dooren et al. 2014). For example, halving consumption of meat, dairy products and eggs in the European Union would achieve a 40% reduction in ammonia emissions, 25–40% reduction in non-CO2 GHG emissions (primarily from agriculture) and 23% per capita less use of cropland for food production, with dietary changes lowering health risks (Westhoek et al. 20141028). In China, diets were designed that could meet dietary guidelines while creating significant reductions in GHG emissions (between 5% and 28%, depending on scenario) (Song et al. 20171029). Changing diets can also reduce non-dietary related health issues caused by emissions of air pollutants. For example, specific changes in diets were assessed for their potential to mitigate PM 2.5 in China (Zhao et al. 2017b1030).

Some studies are starting to estimate both health and environmental benefits from dietary shifts. For example, Farchi et al. (2017) estimate health (colorectal cancer, cardiovascular disease) and GHG outcomes of ‘Mediterranean’ diets in Italy, and found the potential to reduce deaths from colorectal cancer of 7–10% and CVD from 9–10%, as well as potential savings of up to 263 CO2-eq per person per year. In the USA, Hallström et al. (2017) found that adoption of healthier diets (consistent with dietary guidelines, and reducing amounts of red and processed meats) could reduce relative risk of coronary heart disease, colorectal cancer, and type 2 diabetes by 20–45%, USA healthcare costs by 77–93 billion USD per year, and direct GHG emissions by 222–826 kg CO2-eq per person per year (69–84 kg from the healthcare system, 153–742 kg from the food system). Broadly similar conclusions were found for the Netherlands (Biesbroek et al. 20141031); and the UK (Friel et al. 20091032 and Milner et al. 20151033).

Whilst for any given disease, there are a range of factors, including diet, that can affect it, and evidence is stronger for some diseases than others, a recent review found that an overall trend toward increased cancer risk was associated with unhealthy dietary patterns, suggesting that diet-related choices could significantly affect the risk of cancer (Grosso et al. 20171034). Tilman and Clark (2014)1035 found significant benefits in terms of reductions in relative risk of key diseases: type 2 diabetes, cancer, coronary mortality and all causes of mortality (Figure 5.15).

5.6.3.1

Can dietary shifts provide significant benefits?

Many studies now indicate that dietary shifts can significantly reduce GHG emissions. For instance, several studies highlight that if current dietary trends are maintained, this could lead to emissions from agriculture of approximately 20 GtCO2-eq yr–1 by 2050, creating significant mitigation potential (Pradhan et al. 2013b1036; Bajželj et al. 20141037; Hedenus et al. 20141038; Bryngelsson et al. 20171039). Additionally in the USA, a shift in consumption towards a broadly healthier diet, combined with meeting the USDA and Environmental Protection Agency’s 2030 food loss and waste reduction goals, could increase per capita food-related energy use by 12%, decrease blue water consumption by 4%, decrease green water use by 23%, decrease GHG emissions from food production by 11%, decrease GHG emissions from landfills by 20%, decrease land use by 32%, and increase fertiliser use by 12% (Birney et al. 20171040). This study, however, does not account for all potential routes to emissions, ignoring, for example, fertiliser use in feed production. Similar studies have been conducted, for China (Li et al. 20161041), where adoption of healthier diets and technology improvements have the potential to reduce food systems GHG emissions by >40% relative to those in 2010; and India (Green et al. 20171042; Vetter et al. 20171043), where alternative diet scenarios can affect emissions from the food system by –20 to +15%.

Springmann et al.(2018a) modelled the role of technology, waste reduction and dietary change in living within planetary boundaries (Rockström et al. 20091044), with the climate change boundary being a 66% chance of limiting warming to less than 2°C. They found that all are necessary for the achievement of a sustainable food system. Their principal conclusion is that only by adopting a ‘flexitarian diet’, as a global average, would climate change be limited to under two degrees. Their definition of a flexitarian diet is fruits and vegetables, plant-based proteins, modest amounts of animal-based proteins, and limited amounts of red meat, refined sugar, saturated fats, and starchy foods.

Healthy and sustainable diets address both health and environmental concerns (Springmann et al. 2018b1045). There is high agreement that there are significant opportunities to achieve both objectives simultaneously. Contrasting results of marginal GHG emissions, that is, variations in emissions as a result of variation in one or more dietary components, are found when comparing low to high emissions in self-selected diets (diets freely chosen by consumers). Vieux et al. (2013)1046 found self-selected healthier diets with higher amounts of plant-based food products did not result in lower emissions, while (Rose et al. 2019)1047 found that the lowest emission diets analysed were lower in meat but higher in oil, refined grains and added sugar. Vieux et al. (2018)1048 concluded that setting nutritional goals with no consideration for the environment may increase GHG emissions.

Tukker et al. (2011)1049 also found a slight increase in emissions by shifting diets towards the European dietary guidelines, even with lower meat consumption. Heller and Keoleian (2015)1050 found a 12% increase in GHG emissons when shifting to iso-caloric diets, defined as diets with the same caloric intake of diets currently consumed, following the USA guidelines and a 1% decrease in GHG emissions when adjusting caloric intake to recommended levels for moderate activity. There is scarce information on the marginal GHG emissions that would be associated with following dietary guidelines in developing countries.

Some studies have found a modest mitigation potential of diet shifts when economic and biophysical systems effects are taken into account in association with current dietary guidelines. Tukker et al. (2011)1051, considering economic rebound effects of diet shifts (i.e., part of the gains would be lost due to increased use at lower prices), found maximum changes in emissions of the EU food system of 8% (less than 2% of total EU emissions) when reducing meat consumption by 40 to 58%. Using an economic optimisation model for studying carbon taxation in food but with adjustments of agricultural production systems and commodity markets in Europe, Zech and Schneider (2019) found a reduction of 0.41% in GHG emissions at a tax level of 50 USD per tCO2-eq. They estimate a leakage of 43% of the GHG emissions reduced by domestic consumption, (i.e., although reducing emissions due to reducing consumption, around 43% of the emissions would not be reduced because part of the production would be directed to exports).

Studying optimised beef production systems intensification technologies in a scenario of no grasslands area expansion de Oliveira Silva et al. (2016)1052 found marginal GHG emissions to be negligible in response to beef demand in the Brazilian Cerrado. This was because reducing productivity would lead to increased emission intensities, cancelling out the effect of reduced consumption.

In summary, there is significant potential mitigation (high confidence) arising from the adoption of diets in line with dietary recommendations made on the basis of health. These are broadly similar across most countries. These are typically capped at the number of calories and higher in plant-based foods, such as vegetables, fruits, whole grains, legumes, nuts and seeds, and lower in animal-sourced foods, fats and sugar. Such diets have the potential to be both more sustainable and healthier than alternative diets (but healthy diets are not necessarily sustainable and vice versa). The extent to which the mitigation potential of dietary choices can be realised requires both climate change and health being considered together. Socio-economic (prices, rebound effects), political, and cultural contexts would require significant consideration to enable this mitigation potential to be realised.

5.6.4

Sustainable integrated agricultural systems

A range of integrated agricultural systems are being tested to evaluate synergies between mitigation and adaptation and lead to low-carbon and climate-resilient pathways for sustainable food security and ecosystem health (robust evidence, medium agreement). Integration refers to the use of practices that enhance an agroecosystem’s mitigation, resilience, and sustainability functions. These systems follow holistic approaches with the objective of achieving biophysical, socio-cultural, and economic benefits from land management systems (Sanz et al. 20171053). These integrated systems may include agroecology (FAO et al. 20181054; Altieri et al. 20151055), climate smart agriculture (FAO 2011c1056; Lipper et al. 20141057; Aggarwal et al. 20181058), conservation agriculture (Aryal et al. 20161059; Sapkota et al. 20151060), and sustainable intensification (FAO 2011d1061; Godfray 20151062), amongst others.

Many of these systems are complementary in some of their practices, although they tend to be based on different narratives (Wezel et al. 20151063; Lampkin et al. 20151064; Pimbert 20151065). They have been tested in various production systems around the world (Dinesh et al. 20171066; Jat et al. 20161067; Sapkota et al. 20151068 and Neufeldt et al. 20131069). Many technical innovations, for example, precision nutrient management (Sapkota et al. 20141070) and precision water management (Jat et al. 20151071), can lead to both adaptation and mitigation outcomes and even synergies; although negative adaptation and mitigation outcomes (i.e., trade-offs) are often overlooked. Adaptation potential of ecologically intensive systems includes crop diversification, maintaining local genetic diversity, animal integration, soil organic management, water conservation and harvesting the role of microbial assemblages (Section 5.3). Technical innovations may encompass not only inputs reduction, but complete redesign of agricultural systems (Altieri et al. 20171072) and how knowledge is generated (Levidow et al. 20141073), including social and political transformations.

5.6.4.1

Agroecology

Agroecology (see Glossary) (Francis et al. 20031074; Gliessman and Engles 20141075; Gliessman 20181076), provides knowledge for their design and management, including social, economic, political, and cultural dimensions (Dumont et al. 20161077). It started with a focus at the farm level but has expanded to include the range of food system activities (Benkeblia 20181078). Agroecology builds systems resilience through knowledge-intensive practices relying on traditional farming systems and co-generation of new insights and information with stakeholders through participatory action research (Menéndez et al. 20131079). It provides a multidimensional view of food systems within ecosystems, building on ILK and co-evolving with the experiences of local people, available natural resources, access to these resources, and ability to share and pass on knowledge among communities and generations, emphasising the inter-relatedness of all agroecosystem components and the complex dynamics of ecological processes (Vandermeer 19951080).

At the farm level, agroecological practices recycle biomass and regenerate soil biotic activities. They strive to attain balance in nutrient flows to secure favourable soil and plant growth conditions, minimise loss of water and nutrients, and improve use of solar radiation. Practices include efficient microclimate management, soil cover, appropriate planting time and genetic diversity. They seek to promote ecological processes and services such as nutrient cycling, balanced predator/prey interactions, competition, symbiosis, and successional changes. The overall goal is to benefit human and non-human communities in the ecological sphere, with fewer negative environmental or social impacts and fewer external inputs (Vandermeer et al. 19981081; Altieri et al. 19981082). From a food system focus, agroecology provides management options in terms of commercialisation and consumption through the promotion of short food chains and healthy diets (Pimbert and Lemke 20181083; Loconto et al. 20181084).

Agroecology has been proposed as a key set of practices in building climate resilience (FAO et al. 20181085; Altieri et al. 20151086). These can enhance on-farm diversity (of genes, species, and ecosystems) through a landscape approach (FAO 2018g1087). Outcomes include soil conservation and restoration and thus soil carbon sequestration, reduction of the use of mineral and chemical fertilisers, watershed protection, promotion of local food systems, waste reduction, and fair access to healthy food through nutritious and diversified diets (Pimbert and Lemke 20181088; Kremen et al. 20121089; Goh 20111090; Gliessman and Engles 20141091).

A principle in agroecology is to contribute to food production by smallholder farmers (Altieri 20021092). Since climatic events can severely impact smallholder farmers, there is a need to better understand the heterogeneity of small-scale agriculture in order to consider the diversity of strategies that traditional farmers have used and still use to deal with climatic variability. In Africa, many smallholder farmers cope with and even prepare for climate extremes, minimising crop failure through a series of agroecological practices (e.g., biodiversification, soil management, and water harvesting) (Mbow et al. 2014a1093). Resilience to extreme climate events is also linked to on-farm biodiversity, a typical feature of traditional farming systems (Altieri and Nicholls 20171094).

Critiques of agroecology refer to its explicit exclusion of modern biotechnology (Kershen 20131095) and the assumption that smallholder farmers are a uniform unit with no heterogeneity in power (and thus gender) relationships (Neira and Montiel 20131096; Siliprandi and Zuluaga Sánchez 20141097).

5.6.4.2

Climate-smart agriculture

‘Climate-smart agriculture’ (CSA) is an approach developed to tackle current food security and climate change challenges in a joint and synergistic fashion (Lipper et al. 20141098; Aggarwal et al. 20181099; FAO 2013c1100). CSA is designed to be a pathway towards development and food security built on three pillars: increasing productivity and incomes, enhancing resilience of livelihoods and ecosystems and reducing, and removing GHG emissions from the atmosphere (FAO 2013c1101). Climate-smart agricultural systems are integrated approaches to the closely linked challenges of food security, development, and climate change adaptation/mitigation to enable countries to identify options with maximum benefits and those where trade-offs need management.

Many agricultural practices and technologies already provide proven benefits to farmers’ food security, resilience and productivity (Dhanush and Vermeulen 20161102). In many cases, these can be implemented by changing the suites of management practices. For example, enhancing soil organic matter to improve the water-holding capacity of agricultural landscapes also sequesters carbon. In annual cropping systems, changes from conventional tillage practices to minimum tillage can convert the system from one that either provides adaptation or mitigation benefits or neither to one that provides both adaptation and mitigation benefits (Sapkota et al. 2017a1103; Harvey et al. 2014a1104).

Increasing food production by using more fertilisers in agricultural fields could maintain crop yield in the face of climate change, but may result in greater overall GHG emissions. But increasing or maintaining the same level of yield by increasing nutrient-use- efficiency through adoption of better fertiliser management practices could contribute to both food security and climate change mitigation (Sapkota et al. 2017a1105).

Mixed farming systems integrating crops, livestock, fisheries and agroforestry could maintain crop yield in the face of climate change, help the system to adapt to climatic risk, and minimise GHG emissions by increasingly improving the nutrient flow in the system (Mbow et al. 2014a1106; Newaj et al. 20161107; Bioversity International 20161108). Such systems can help diversify production and/or incomes and support efficient and timely use of inputs, thus contributing to increased resilience, but they require local seed and input systems and extension services. Recent whole farm modelling exercises have shown the economic and environmental (reduced GH emissions, reduced land use) benefits of integrated crop-livestock systems (Gil et al. 20181109) compared different soy-livestock systems across multiple economic and environmental indicators, including climate resilience. However, it is important to note that potential benefits are very context specific.

Although climate-smart agriculture involves a holistic approach, some argue that it narrowly focuses on technical aspects at the production level (Taylor 20181110; Newell and Taylor 20181111). Studying barriers to the adoption and diffusion of technological innovations for climate-smart agriculture in Europe, Long et al. (2016)1112 found that there was incompatibility between existing policies and climate-smart agriculture objectives, including barriers to the adoption of technological innovations.

Climate-smart agricultural systems recognise that the implementation of the potential options will be shaped by specific country contexts and capacities, as well as enabled by access to better information, aligned policies, coordinated institutional arrangements and flexible incentives and financing mechanisms (Aggarwal et al. 20181113). Attention to underlying socio-economic factors that affect adoption of practices and access to technologies is crucial for enhancing biophysical processes, increasing productivity, and reducing GHG emissions at scale. The Government of India, for example, has started a programme of climate resilient villages (CRV) as a learning platform to design, implement, evaluate and promote various climate-smart agricultural interventions, with the goal of ensuring enabling mechanisms at the community level (Srinivasa Rao et al. 20161114).

5.6.4.3

Conservation agriculture

Conservation agriculture (CA) is based on the principles of minimum soil disturbance and permanent soil cover, combined with appropriate crop rotation (Jat et al. 20141115; FAO 2011e1116). CA has been shown to respond with positive benefits to smallholder farmers under both economic and environmental pressures (Sapkota et al. 2017a, 20151117). This agricultural production system uses a body of soil and residues management practices that control erosion (Blanco Sepúlveda1118 and Aguilar Carrillo 2016) and at the same time improve soil quality, by increasing organic matter content and improving porosity, structural stability, infiltration and water retention (Sapkota et al. 2017a, 20151119 and Govaerts et al. 2009).

Intensive agriculture during the second half of the 20th century led to soil degradation and loss of natural resources and contributed to climate change. Sustainable soil management practices can address both food security and climate change challenges faced by these agricultural systems. For example, sequestration of soil organic carbon (SOC) is an important strategy to improve soil quality and to mitigation of climate change (Lal 20041120). CA has been reported to increase farm productivity by reducing costs of production (Aryal et al. 20151121; Sapkota et al. 20151122; Indoria et al. 20171123) as well as to reduce GHG emission (Pratibha et al. 20161124).

Conservation agriculture brings favourable changes in soil properties that affect the delivery of nature’s contribution to people (NCPs) or ecosystem services, including climate regulation through carbon sequestration and GHG emissions (Palm et al. 20131125; Sapkota et al. 2017a1126). However, by analysing datasets for soil carbon in the tropics, Powlson et al. (2014, 2016) argued that the rate of SOC increase and resulting GHG mitigation in CA systems, from zero-tillage in particular, has been overstated (Chapter 2).

However, there is unanimous agreement that the gain in SOC and its contribution to GHG mitigation by CA in any given soil is largely determined by the quantity of organic matter returned to the soil (Giller et al. 20091127; Virto et al. 20111128; Sapkota et al. 2017b1129). Thus, a careful analysis of the production system is necessary to minimise the trade-offs among the multiple use of residues, especially where residues remain an integral part of livestock feeding (Sapkota et al. 2017b1130). Similarly, replacing mono-cropping systems with more diversified cropping systems and agroforestry, as well as afforestation and deforestation, can buffer temperatures as well as increase carbon storage (Mbow et al. 2014a1131; Bioversity International 20161132), and provide diversified and healthy diets in the face of climate change.

Adoption of conservation agriculture in Africa has been low despite more than three decades of implementation (Giller et al. 20091133), although there is promising uptake recently in east and southern Africa. This calls for a better understanding of the social and institutional aspects around CA adoption. Brown et al. (2017a)1134 found that institutional and community constraints hampered the use of financial, physical, human and informational resources to implement CA programmes.

Gender plays an important role at the intra-household level in regard to decision-making and distributing benefits. Conservation agriculture interventions have implications for labour requirements, labour allocation, and investment decisions, all of which impact the roles of men and women (Farnworth et al. 20161135) (Section 5.1.3). For example, in the Global South, CA generally reduces labour and production costs and generally leads to increased returns to family labour (Aryal et al. 20151136) although a gender shift of the labour burden to women have also been described (Giller et al. 20091137).

5.6.4.4

Sustainable intensification

The need to produce about 50% more food by 2050, required to feed the increasing world population (FAO 2018a1138), may come at the price of significant increases in GHG emissions and environmental impacts, including loss of biodiversity. For instance, land conversion for agriculture is responsible for an estimated 8–10% of all anthropogenic GHG emissions currently (Section 5.4). Recent calls for sustainable intensification (SI) are based on the premise that damage to the environment through extensification outweighs benefits of extra food produced on new lands (Godfray 20151139). However, increasing the net production area by restoring already degraded land may contribute to increased production on the one hand and increased carbon sequestration on the other (Jat et al. 20161140), thereby contributing to both increased agricultural production and improved natural capital outcomes (Pretty et al. 20181141).

Sustainable intensification is a goal but does not specify a priori how it could be attained, for example, which agricultural techniques to deploy (Garnett et al. 20131142). It can be combined with selected other improved management practices, for example, conservation agriculture (see above), or agroforestry, with additional economic, ecosystem services, and carbon benefits. Sustainable intensification, by improving nutrient, water, and other input-use efficiency, not only helps to close yield gaps and contribute to food security (Garnett et al. 20131143), but also reduces the loss of such production inputs and associated emissions (Sapkota et al. 2017c1144; Wollenberg et al. 20161145). Closing yield gaps is a way to become more efficient in use of land per unit production. Currently, most regions in Africa and South Asia have attained less than 40% of their potential crop production (Pradhan et al. 20151146). Integrated farming systems (e.g., mixed crop/livestock, crop/aquaculture) are strategies to produce more products per unit land, which in regard to food security, becomes highly relevant.

Sustainable intensification acknowledges that enhanced productivity needs to be accompanied by maintenance of other ecosystem services and enhanced resilience to shocks (Vanlauwe et al. 20141147). SI in intensively farmed areas may require a reduction in production in favour of increasing sustainability in the broad sense (Buckwell et al. 20141148) (Cross-Chapter Box 6 in Chapter 5). Hence, moving towards sustainability may imply lower yield growth rates than those maximally attainable in such situations. For areas that contain valuable natural ecosystems, such as the primary forest in the Congo basin, intensification of agriculture is one of the pillars of the strategy to conserve forest (Vanlauwe et al. 20141149). Intensification in agriculture is recognised as one of the pathways to meet food security and climate change adaptation and mitigation goals (Sapkota et al. 2017c1150).

However, SI does not always confer co-benefits in terms of food security and climate change adaption/mitigation. For example, in the case of Vietnam, intensified production of rice and pigs reduced GHG emissions in the short term through land sparing, but after two decades, the emissions associated with higher inputs were likely to outweigh the savings from land sparing (Thu Thuy et al. 2009). Intensification needs to be sustainable in all components of food system by curbing agricultural sprawl, rebuilding soils, restoring degraded lands, reducing agricultural pollution, increasing water use efficiency, and decreasing the use of external inputs (Cook et al. 20151151).

A study conducted by Palm et al. (2010)1152 in Sub-Saharan Africa, reported that, at low population densities and high land availability, food security and climate mitigation goals can be met with intensification scenarios, resulting in surplus crop area for reforestation. In contrast, for high population density and small farm sizes, attaining food security and reducing GHG emissions require the use of more mineral fertilisers to make land available for reforestation. However, some forms of intensification in drylands can increase rather than reduce vulnerability due to adverse effects such as environmental degradation and increased social inequity (Robinson et al. 20151153).

Sustainable intensification has been critiqued for considering food security only from the supply side, whereas global food security requires attention to all aspects of food system, including access, utilisation, and stability (Godfray 20151154). Further, adoption of high-input forms of agriculture under the guise of simultaneously improving yields and environmental performance will attract more investment leading to higher rate of adoption but with the environmental component of SI quickly abandoned (Godfray 20151155). Where adopted, SI needs to engage with the sustainable development agenda to (i) identify SI agricultural practices that strengthen rural communities, improve smallholder livelihoods and employment, and avoid negative social and cultural impacts, including loss of land tenure and forced migration; (ii) invest in the social, financial, natural, and physical capital needed to facilitate SI implementation; and (iii) develop mechanisms to pay poor farmers for undertaking sustainability measures (e.g., GHG emissions mitigation or biodiversity protection) that may carry economic costs (Garnett et al. 20131156).

In summary, integrated agricultural systems and practices can enhance food system resilience to climate change and reduce GHG emissions, while helping to achieve sustainability (high confidence).

Eamon Haughey (Ireland), Tim Benton (United Kingdom), Annette Cowie (Australia), Lennart Olsson (Sweden), Pete Smith (United Kingdom)

Introduction

The projected demand for more food, fuel and fibre for a growing human population necessitates intensification of current land use to avoid conversion of additional land to agriculture and potentially allow the sparing of land to provide other ecosystem services, including carbon sequestration, production of biomass for energy, and the protection of biodiversity (Benton et al. 20181157; Garnett et al. 20131158). Land-use intensity may be defined in terms of three components; (i) intensity of system inputs (land/soil, capital, labour, knowledge, nutrients and other chemicals), (ii) intensity of system outputs (yield per unit land area or per specific input) and (iii) the impacts of land use on ecosystem services such as changes in soil carbon or biodiversity (Erb et al. 20131159). Intensified land use can lead to ecological damage as well as degradation of soil, resulting in a loss of function which underpins many ecosystem services (Wilhelm and Smith 20181160; Smith et al. 20161161). Therefore, there is a risk that increased agricultural intensification could deliver short-term production goals at the expense of future productive potential, jeopardising long term food security (Tilman et al. 20111162).

Agroecosystems which maintain or improve the natural and human capital and services they provide may be defined as sustainable systems, while those which deplete these assets as unsustainable (Pretty and Bharucha 20141163). Producing more food, fuel and fibre without the conversion of additional non-agricultural land while simultaneously reducing environmental impacts requires what has been termed sustainable intensification (Godfray et al. 20101164; FAO 2011e1165) (Glossary and Figure 1 in this Cross-Chapter Box). Sustainable intensification (SI) may be achieved through a wide variety of means; from improved nutrient and water use efficiency via plant and animal breeding programmes, to the implementation of integrated soil fertility and pest management practices, as well as by smarter land-use allocation at a larger spatial scale: for example, matching land use to the context and specific capabilities of the land (Benton et al. 2018). However, implementation of SI is broader than simply increasing the technical efficiency of agriculture (‘doing more with less’). It sometimes may require a reduction of yields to raise sustainability, and successful implementation can be dependent on place and scale. Pretty et al. (2018)1166, following Hill (1985)1167, highlights three elements to SI: (i) increasing efficiency, (ii) substitution of less beneficial or efficient practices for better ones, and (iii) system redesign to adopt new practices and farming systems (Table 1 in this Cross-Chapter Box).

Under a land sparing strategy, intensification of land use in some areas, generating higher productivity per unit area of land, can allow other land to provide other ecosystem services, such as increased carbon sequestration and the conservation of natural ecosystems and biodiversity (Balmford et al. 20181168 and Strassburg et al. 20141169). Conversely under a land sharing strategy, less, or no, land is set aside, but lower levels of intensification are applied to agricultural land, providing a combination of provisioning and other functions such as biodiversity conservation from the same land (Green et al. 20051170). The two approaches are not mutually exclusive and the suitability of their application is generally system-, scale- and/or location-specific (Fischer et al. 2014). One crucial issue for the success of a land sparing strategy is that spared land is protected from further conversion. As the profits from the intensively managed land increase, there is an incentive for conversion of additional land for production (Byerlee et al. 20141171). Furthermore, it is implicit that there are limits to the SI of land at a local and also planetary boundary level (Rockström et al. 20091172). These may relate to the ‘health’ of soil, the presence of supporting services, such as pollination, local limits to water availability, or limits on air quality. This implies that it may not be possible to meet demand ‘sustainably’ if demand exceeds local and global limits. There are no single global solutions to these challenges and specific in situ responses for different farming systems and locations are required. Bajželj et al. (2014)1173 showed that implementation of SI, primarily through yield gap closure, had better environmental outcomes compared with ‘business as usual’ trajectories. However, SI alone will not be able to deliver the necessary environmental outcomes from the food system – dietary change and reduced food waste are also required (Springmann et al. 2018a1174; Bajželj et al. 20141175).

Cross-Chapter Box 6, Table 1 | Approaches to sustainable intensification of agriculture (Pretty et al. 2018; Hill 1985).

Improved efficiency – example of precision agriculture

Precision farming usually refers to optimising production in fields through site-specific choices of crop varieties, agrochemical application, precise water management (e.g., in given areas or threshold moistures) and management of crops at a small scale (or livestock as individuals) (Hedley 20151176). Precision agriculture has the potential to achieve higher yields in a more efficient and sustainable manner compared with traditional low-precision methods.

Precision agriculture

Precision agriculture is a technologically advanced approach that uses continual monitoring of crop and livestock performance to actively inform management practices. Precise monitoring of crop performance over the course of the growing season will enable farmers to economise on their inputs in terms of water, nutrients and pest management. Therefore, it can contribute to both the food security (by maintaining yields), sustainability (by reducing unnecessary inputs) and land sparing goals associated with SI. The site-specific management of weeds allows a more efficient application of herbicide to specific weed patches within crops (Jensen et al. 20121177). Such precision weed control has resulted in herbicide savings of 19–22% for winter oilseed rape, 46–57% for sugar beet and 60–77% for winter wheat production (Gutjahr and Gerhards 20101178). The use of on-farm sensors for real time management of crop and livestock performance can enhance farm efficiency (Aqeel-Ur-Rehman et al. 20141179). Mapping soil nutrition status can allow for more targeted, and therefore more effective, nutrient management practices (Hedley 20151180). Using wireless sensors to monitor environmental conditions, such as soil moisture, has the potential to allow more efficient crop irrigation (Srbinovska et al. 20151181). Controlled traffic farming, where farm machinery is confined to permanent tracks, using automatic steering and satellite guidance, increases yields by minimising soil compaction. However, barriers to the uptake of many of these high-tech precision agriculture technologies remain. In what is described as the ‘implementation problem’, despite the potential to collect vast quantities of data on crop or livestock performance, applying these data to inform management decisions remains a challenge (Lindblom et al. 20171182).

Low-tech precision agriculture

The principle of precision agriculture can be applied equally to low capital-input farming, in the form of low-tech precision agriculture (Conway 20131183). The principle is the same, but instead of adopting capital-heavy equipment (such as sensor technology connected to the ‘internet of things’, or large machinery and expensive inputs), farmers use knowledge and experience and re-purposed innovative approaches, such as a bottle cap as a fertiliser measure for each plant, applied by hand (Mondal and Basu 20091184). This type of precision agriculture is particularly relevant to small-scale farming in the Global South, where capital investment is major limiting factor. For example, the application of a simple seed priming technique resulted in a 20 to 30% increase in yields of pearl millet and sorghum in semi-arid West Africa (Aune et al. 20171185). Low-tech precision agriculture has the potential to increase the economic return per unit land area while also creating new employment opportunities.

Sustainable intensification through farming system redesign

Sustainable intensification requires equal weight to be placed on the sustainability and intensification components (Benton 20161186; Garnett et al. 20131187). Figure 1 in this Cross-Chapter Box outlines the trade-offs which SI necessitates between the intensity of land use against long-term sustainability. One approach to this challenge is through farming system redesign, including increased diversification.

Diversification of intensively managed systems

Incorporating higher levels of plant diversity in agroecosystems can improve the sustainability of farming systems (Isbell et al. 20171188). Where intensive land use has led to land degradation, more diverse land-use systems, such as intercropping, can provide a more sustainable land-use option with co-benefits for food security, adaptation and mitigation objectives. For example, in temperate regions, highly productive agricultural grasslands used to produce meat and dairy products are characterised by monoculture pastures with high agrochemical inputs. Multi-species grasslands may provide a route to SI, as even a modest increase in species richness in intensively managed grasslands can result in higher forage yields without increased inputs, such as chemical fertiliser (Finn et al. 20131189; Sanderson et al. 20131190; Tilman et al. 20111191). Recent evidence also indicates multispecies grasslands have greater resilience to drought, indicating co-benefits for adaptation (Hofer et al. 20161192; Haughey et al. 20181193).

Diversification of production systems

Agroforestry systems (see Glossary) can promote regional food security and provide many additional ecosystem services when compared with monoculture crop systems. Co-benefits for mitigation and adaptation include increased carbon sequestration in soils and biomass, improved water and nutrient use efficiency and the creation of favourable micro-climates (Waldron et al. 20171194). Silvopasture systems, which combine grazing of livestock and forestry, are particularly useful in reducing land degradation where the risk of soil erosion is high (Murgueitio et al. 20111195). Crop and livestock systems can also be combined to provide multiple services. Perennial wheat derivatives produced both high quality forage and substantial volumes of cereal grains (Newell and Hayes 2017), and show promise for integrating cereal and livestock production while sequestering soil carbon (Ryan et al. 20181196). A key feature of diverse production systems is the provision of multiple income streams for farming households, providing much needed economic resilience in the face of fluctuation of crop yields and prices.

Landscape approaches

The land sparing and land sharing approaches which may be used to implement SI are inherently ‘landscape approaches’ (e.g., Hodgson et al. 20101197). While the term landscape is by no means precise (Englund et al. 20171198), landscape approaches, focused, for example, at catchment scale, are generally agreed to be the best way to tackle competing demands for land (e.g., Sayer et al. 20131199), and are the appropriate scale at which to focus the implementation of sustainable intensification. The landscape approach allots land to various uses – cropping, intensive and extensive grazing, forestry, mining, conservation, recreation, urban, industry, infrastructure – through a planning process that seeks to balance conservation and production objectives. With respect to SI, a landscape approach is pertinent to achieving potential benefits for biodiversity conservation, ensuring that land ‘spared’ through SI remains protected, and that adverse impacts of agriculture on conservation land are minimised. Depending on the land governance mechanisms applied in the jurisdiction, different approaches will be appropriate/required. However, benefits are only assured if land-use restrictions are devised and enforced.

Summary

Intensification needs to be achieved sustainably, necessitating a balance between productivity today and future potential (high agreement, medium evidence). Improving the efficiency of agriculture systems can increase production per unit of land through more effective resource use. To achieve SI, some intensively managed agricultural systems may have to be diversified as they cannot be further intensified without land degradation. A combination of land sparing and sharing options can be utilised to achieve SI – their application is most likely to succeed if applied using a landscape approach.

5.6.5

Role of urban agriculture

Cities are an important actor in the food system through demand for food by urban dwellers and production of food in urban and peri-urban areas (Cross-Chapter Box 4 in Chapter 2). Both the demand side and supply side roles are important relative to climate change mitigation and adaptation strategies. Urban areas are home to more than half of the world’s population, and a minimal proportion of the production. Thus, they are important drivers for the development of the complex food systems in place today, especially with regard to supply chains and dietary preferences.

The increasing separation of urban and rural populations with regard to territory and culture is one of the factors favouring the nutrition transition towards urban diets (Weber and Matthews 20081200; Neira et al. 20161201). These are primarily based on a high diversity of food products, independent of season and local production, and on the extension of the distances that food travels between production and consumption. The transition of traditional diets to more homogeneous diets has also become tied to consumption of animal protein, which has increased GHG emissions globally (Section 5.4.6).

Cities are becoming key actors in developing strategies of mitigation to climate change, in their food procurement and in sustainable urban food policies alike (McPhearson et al. 20181202). These are being developed by big and medium-sized cities in the world, often integrated within climate change policies (Moragues et al. 20131203 and Calori and Magarini 20151204). A review of 100 cities shows that urban food consumption is one of the largest sources of urban material flows, urban carbon footprint, and land footprint (Goldstein et al. 20171205). Additionally, the urban poor have limited capacity to adapt to climate-related impacts, which place their food security at risk under climate change (Dubbeling and de Zeeuw 20111206).

Urban and peri-urban areas. In 2010, around 14% of the global population was nourished by food grown in urban and peri-urban areas (Kriewald et al. 20191207). A review study on Sub-Saharan Africa shows that urban and peri-urban agriculture contributes to climate change adaptation and mitigation (Lwasa et al. 20141208, 2015). Urban and peri-urban agriculture reduces the food carbon footprint by avoiding long distance food transport. These types of agriculture also limit GHG emissions by recycling organic waste and wastewater that would otherwise release methane from landfills and dumping sites (Lwasa et al. 2014). Urban and peri-urban agriculture also contribute in adapting to climate change, including extreme events, by reducing the urban heat island effect, increasing water infiltration and slowing down run-offs to prevent flooding, etc. (Lwasa et al. 2014, 2015; Kumar et al. 2017a1209). For example, a scenario analysis shows that urban gardens reduce the surface temperature up to 10°C in comparison to the temperature without vegetation (Tsilini et al. 20151210). Urban agriculture can also improve biodiversity and strengthen associated ecosystem services (Lin et al. 20151211).

Urban and peri-urban agriculture is exposed to climate risks and urban growth that may undermine its long-term potential to address urban food security (Padgham et al. 20151212). Therefore, there is a need to better understand the impact of urban sprawl on peri-urban agriculture; the contribution of urban and peri-urban agriculture to food self-sufficiency of cities; the risks posed by pollutants from urban areas to agriculture and vice-versa; the global and regional extent of urban agriculture; and the role that urban agriculture could play in climate resilience and abating malnutrition (Mok et al. 20141213; Hamilton et al. 20141214). Globally, urban sprawl is projected to consume 1.8–2.4% and 5% of the current cultivated land by 2030 and 2050 respectively, leading to crop calorie loss of 3–4% and 6–7%, respectively (Pradhan et al. 20141215 and Bren d’Amour et al. 2017). Kriewald et al. 2019 shows that the urban growth has the largest impact in many sub-continental regions (e.g., Western, Central, and Eastern Africa), while climate change will mostly reduce potential of urban and peri-urban agriculture in Southern Europe and North Africa.

In summary, urban and peri-urban agriculture can contribute to improving urban food security, reducing GHG emissions, and adapting to climate change impacts (robust evidence, medium agreement).

5.6.6

Links to the Sustainable Development Goals

5.7.1

Enabling policy environments

The scope for responses to make sustainable land use inclusive of climate change mitigation and adaptation, and the policies to implement them, are covered in detail in Chapters 6 and 7. Here we highlight some of the major policy areas that have shaped the food system, and might be able to shape responses in future. Although two families of policy – agriculture and trade – have been instrumental in shaping the food system in the past (and potentially have led to conditions that increase climate vulnerability) (Benton and Bailey 2019), a much wider family of policy instruments can be deployed to reconfigure the food system to deliver healthy diets in a sustainable way.

5.7.1.1

Agriculture and trade policy

Agriculture. The thrust of agricultural policies over the last 50 years has been to increase productivity, even if at the expense of environmental sustainability (Benton and Bailey 20191230). For example, in 2007–2009, 46% of OECD support for agriculture was based on measures of output (price support or payments based on yields), 37% of support was based on the current or historical area planted, herd size (or correlated measures of the notional costs of farming), and 13% was payments linked to input prices. In a similar vein, non-OECD countries have promoted productivity growth for their agricultural sectors.

Trade. Along with agricultural policy to grow productivity, the development of frameworks to liberalise trade (such as the General Agreement on Tariffs and Trade – GATT – Uruguay Round, now incorporated into the World Trade Organization) have been essential in stimulating the growth of a globalised food system. Almost every country has a reliance on trade to fulfil some or all of its local food needs, and trade networks have grown to be highly complex (Puma et al. 20151231; MacDonald et al. 20151232; Fader et al. 20131233 and Ercsey-Ravasz et al. 20121467). This is because many countries lack the capacity to produce sufficient food due to climatic conditions, soil quality, water constraints, and availability of farmland (FAO 2015b1234). In a world of liberalised trade, using comparative advantage to maximise production in high-yielding commodities, exporting excess production, and importing supplies of other goods supports economic growth.

City states as well as many small island states, do not have adequate farmland to feed their populations, while Sub-Saharan African countries are projected to experience high population growth as well as to be negatively impacted by climate change, and thus will likely find it difficult to produce all of their own food supplies (Agarwal et al. 20021235). One study estimates that some 66 countries are currently incapable of being self-sufficient in food (Pradhan et al. 20141236). Estimates of the proportion of people relying on trade for basic food security vary from about 16% to about 22% (Fader et al. 20131237; Pradhan et al. 20141238), with this figure rising to between 1.5 and 6 billion people by 2050, depending on dietary shifts, agricultural gains, and climate impacts (Pradhan et al. 20141239).

Global trade is therefore essential for achieving food and nutrition security under climate change because it provides a mechanism for enhancing the efficiency of supply chains, reducing the vulnerability of food availability to changes in local weather, and moving production from areas of surplus to areas of deficit (FAO 2018d1240). However, the benefits of trade will only be realised if trade is managed in ways that maximise broadened access to new markets while minimising the risks of increased exposure to international competition and market volatility (Challinor et al. 20181241; Brown et al. 2017b1242).

As described in Section 5.8.1, trade acts to buffer exposure to climate risks when the market works well. Under certain conditions – such as shocks, or the perception of a shock, coupled with a lack of food stocks or lack of transparency about stocks (Challinor et al. 20181243; Marchand et al. 20161244) – the market can fail and trade can expose countries to food price shocks.

Furthermore, Clapp (2016) showed that trade, often supported by high levels of subsidy support to agriculture in some countries, can depress world prices and reduce incomes for other agricultural exporters. Lower food prices that result from subsidy support may benefit urban consumers in importing countries, but at the same time they may hurt farmers’ incomes in those same countries. The outmigration of smallholder farmers from the agriculture sector across the Global South is significantly attributed to these trade patterns of cheap food imports (Wittman 20111245; McMichael 20141246; Akram-Lodhi et al. 20131247). Food production and trade cartels, as well as financial speculation on food futures markets, affect low-income market-dependent populations.

Food sovereignty is a framing developed to conceptualise these issues (Reuter 20151248). They directly relate to the ability of local communities and nations to build their food systems, based, among other aspects, on diversified crops and ILK. If a country enters international markets by growing more commodity crops and reducing local crop varieties, it may get economic benefits, but may also expose itself to climate risks and food insecurity by increasing reliance on trade, which may be increasingly disrupted by climate risks. These include a local lack of resilience from reduced diversity of products, but also exposure to food price spikes, which can become amplified by market mechanisms such as speculation.

In summary, countries must determine the balance between locally produced versus imported food (and feed) such that it both minimises climate risks and ensures sustainable food security. There is medium evidence that trade has positive benefits but also creates exposure to risks (Section 5.3).

5.7.1.2

Scope for expanded policies

There are a range of ways that policy can intervene to stimulate change in the food system – through agriculture, research and development, food standards, manufacture and storage, changing the food environment and access to food, changing practices to encourage or discourage trade (Table 5.6). Novel incentives can stimulate the market, for example, through reduction in waste or changes in diets to gain benefits from a health or sustainability direction. Different contexts with different needs will require different set of policies at local, regional and national levels. See Supplementary Material Section SM5.7 for further discussion on expanded policies.

In summary, although agriculture is often thought to be shaped predominantly by agriculture and trade policies, there are over twenty families of policy areas that can shape agricultural production directly or indirectly (through environmental regulations or through markets, including by shaping consumer behaviour). Thus, delivering outcomes promoting climate change adaptation and mitigation can arise from policies across many departments, if suitably designed and aligned.

5.7.1.4

Multiple policy pathways

As discussed in more detail in Chapters 6 and 7, there is a wide potential suite of interventions and policies that can potentially enhance the adaptation of food systems to climate change, as well as enhance the mitigation potential of food systems on climate change. There is an increasing number of studies that argue that the key to sustainable land management is not in land management practices but in the factors that determine the demand for products from land (such as food). Public health policy, therefore, has the potential to affect dietary choice and thus the demand for different amounts of, and types of, food.

Obersteiner et al. (2016)1256 show that increasing the average price of food is an important policy lever that, by reducing demand, reduces food waste, pressure on land and water, impacts on biodiversity and through reducing emissions, mitigates climate change and potentially helps to achieve multiple SDGs. Whilst such policy responses – such as a carbon tax applied to goods including food – has the potential to be regressive, affecting the poor differentially (Frank et al. 20171257; Hasegawa et al. 20181258 and Kehlbacher et al. 20161259), and increasing food insecurity – further development of social safety nets can help to avoid the regressive nature (Hasegawa et al. 20181260). Hasegawa et al. (2018)1261 point out that such safety nets for vulnerable populations could be funded from the revenues arising from a carbon tax.

The evidence suggests, as with SR15 (IPCC 2018a1262) and its multiple pathways to climate change solutions, that there is no single solution that will address the problems of food and climate change, but instead there is a need to deploy many solutions, simultaneously adapted to the needs and options available in a given context. For example, Springmann et al. (2018a) indicate that maintaining the food system within planetary boundaries at mid-century, including equitable climate, requires increasing the production (and resilience) of agricultural outputs (i.e., closing yield gaps), reducing waste, and changes in diets towards ones often described as flexitarian (low-meat dietary patterns that are in line with available evidence on healthy eating). Such changes can have significant co-benefits for public health, as well as facing significant challenges to ensure equity (in terms of affordability for those in poverty).

Significant changes in the food system require them to be acceptable to the public (‘public license’), or they will be rejected. Focus groups with members of the public around the world, on the issue of changing diets, have shown that there is a general belief that the government plays a key role in leading efforts for change in consumption patterns (Wellesley et al. 20151263). If governments are not leading on an issue, or indicating the need for it through leading public dialogue, it signals to their citizens that the issue is unimportant or undeserving of concern.

In summary, there is significant potential (high confidence) that, through aligning multiple policy goals, multiple benefits can be realised that positively impact public health, mitigation and adaptation (e.g., adoption of healthier diets, reduction in waste, reduction in environmental impact). These benefits may not occur without the alignment across multiple policy areas (high confidence).

5.7.2

Enablers for changing markets and trade

‘Demand’ for food is not an exogenous variable to the food system but is shaped crucially by its ability to produce, market, and supply food of different types and prices. These market dynamics can be influenced by a variety of factors beyond consumer preferences (e.g., corporate power and marketing, transparency, the food environment more generally), and the ability to reshape the market can also depend on its internal resilience and/or external shocks (Challinor et al. 20181264; Oliver et al. 20181265).

5.7.2.1

Capital markets

Two areas are often discussed regarding the role of capital markets in shaping the food system. First, investment in disruptive technologies might stimulate climate-smart food systems (WEF/ McKinsey & Company 20181266 and Bailey and Wellesley 20171267), including alternative proteins, such as laboratory or ‘clean meat’ (which has significant ability to impact on land-use requirements) (Alexander et al. 20171268) (Section 5.5.1.6). An innovation environment through which disruptive technology can emerge typically requires the support of public policy, whether in directly financing small and emerging enterprises, or funding research and development via reducing tax burdens.

Second, widespread adoption of (and perhaps underpinned by regulation for) natural capital accounting as well as financial accounting are needed. Investors can then be aware of the risk exposure of institutions, which can undermine sustainability through externalising costs onto the environment. The prime example of this in the realm of climate change is the Carbon Disclosure Project, with around 2500 companies voluntarily disclosing their carbon footprint, representing nearly 60% of the world’s market capital (CDP 20181269).

5.7.2.2

Insurance and re-insurance

The insurance industry can incentivise actors’ behaviour towards greater climate mitigation or adaptation, including building resilience. For example, Lloyd’s of London analysed the implications of extreme weather for the insurance market, and conclude that the insurance industry needs to examine their exposure to risks through the food supply chain and develop innovative risk-sharing products that can make an important contribution to resilience of the global food system (Lloyd’s 20151270).

Many of these potential areas for enabling healthy and sustainable food systems are also knowledge gaps, in that, whilst the levers are widely known, their efficacy and the ability to scale-up, in any given context, are poorly understood.

5.7.3

Just Transitions to sustainability

Research is limited on how land-use transitions would proceed from ruminant production to other socio-ecological farming systems. Ruminants have been associated with humans since the early development of agriculture, and the role of ruminants in many agricultural systems and smallholder communities is substantial. Ruminant production systems have been adapted to a wide range of socioeconomic and environmental conditions in crop, forestry, and food processing settings (Čolović et al. 20191271), bioenergy production (de Souza et al. 2019), and food waste recycling (Westendorf 20001272). Pasture cultivation in succession to crops is recognised as important to management of pest and diseases cycles and to improve soil carbon stocks and soil quality (Carvalho and Dedieu 20141273). Grazing livestock is important as a reserve of food and economic stocks for some smallholders (Ouma et al. 20031274).

Possible land-use options for transitions away from livestock production in a range of systems include (a) retain land but reduce investments to run a more extensive production system; (b) change land use by adopting a different production activity; (c) abandon land (or part of the farm) to allow secondary vegetation regrowth (Carvalho et al. 20191275 and Laue and Arima 2016); and (d) invest in afforestation or reforestation (Baynes et al. 20171276). The extensification option could lead to increases rather than decreases in GHG emissions related to reduction in beef consumption. Large-scale abandonment, afforestation, or reforestation would probably have more positive environmental outcomes, but could result in economic and social issues that would require governmental subsidies to avoid decline and migration in some regions (Henderson et al. 20181277).

Alternative economic use of land, such as bioenergy production, could balance the negative socioeconomic impact of reducing beef output, reduce the tax values needed to reduce consumption, and avoid extensification of ruminant production systems (Wirsenius et al. 20111278). However, the analysis of the transition of land use for ruminants to other agricultural production systems is still a literature gap (Cross-Chapter Box 7 in Chapter 6).

Finally, it is important to recognise that, while energy alternatives produce the same function for the consumer, it is questionable that providing the same nutritional value through an optimised mix of dietary ingredients provides the same utility for humans. Food has a central role in human pleasure, socialisation, cultural identity, and health (Röös et al. 20171279), including some of the most vulnerable groups, so Just Transitions and their costs need to be taken into account. Pilot projects are important to provide greater insights for large-scale policy design, implementation, and enforcement.

In summary, more research is needed on how land-use transitions would proceed from ruminant production to other farming systems and affect the farmers and other food system actors involved. There is limited evidence on what the decisions of farmers under lower beef demand would be.

5.7.4

Mobilising knowledge

Addressing climate change-related challenges and ensuring food security requires all types of knowledge (formal/non-formal, scientific/ indigenous, women, youth, technological). Miles et al. (2017) stated that a research and policy feedback that allows transitions to sustainable food systems must take a whole system approach. Currently, in transmitting knowledge for food security and land sustainability under climate change there are three major approaches: (i) public technology transfer with demonstration (extension agents); (ii) public and private advisory services (for intensification techniques) and; (iii) non-formal education with many different variants such as farmer field schools, rural resource centres; facilitation extension where front-line agents primarily work as ‘knowledge brokers’ in facilitating the teaching-learning process among all types of farmers (including women and rural young people), or farmer-to-farmer, where farmers act themselves as knowledge transfer and sharing actors through peer processes.

5.7.4.1

Indigenous and local knowledge

Recent discourse has a strong orientation towards scaling-up innovation and adoption by local farmers. However, autonomous adaptation, indigenous knowledge and local knowledge are both important for agricultural adaptation (Biggs et al.20131280)(Section5.3).These involve the promotion of farmer participation in governance structures, research, and the design of systems for the generation and dissemination of knowledge and technology, so that farmers’ needs and knowledge can be taken into consideration. Klenk et al. (2017)1281 found that mobilisation of local knowledge can inform adaptation decision-making and may facilitate greater flexibility in government-funded research. As an example, rural innovation in terrace agriculture developed on the basis of a local coping mechanism and adopted by peasant farmers in Latin America may serve as an adaptation option or starting place for learning about climate change responses (Bocco and Napoletano 20171282). Clemens et al. (2015)1283 found that an open dialogue platform enabled horizontal exchange of ideas and alliances for social learning and knowledge-sharing in Vietnam. Improving local technologies in a participatory manner, through on-farm experimentation, farmer-to- farmer exchange, consideration of women and youths, is also relevant in mobilising knowledge and technologies.

Citizen science has been tested as a useful tool with potential for biodiversity conservation (Schmitz et al. 20151284) and mobilising knowledge from society. In food systems, knowledge-holders (e.g., farmers and pastoralists) are trained to gather scientific data in order to promote conservation and resource management (Fulton et al. 20191285) or to conserve and use traditional knowledge in developed countries relevant to climate change adaptation and mitigation through the use of ICT (Calvet-Mir et al. 20181286).

5.7.4.3

Capacity building and education

Mobilising knowledge may also require significant efforts on capacity building and education to scale up food system responses to climate change. This may involve increasing the capacity of farmers to manage current climate risks and to mitigate and adapt in their local contexts, and of citizens and consumers to understand the links between food demand and climate change emissions and impacts, as well as policy makers to take a systemic view of the issues. Capacity building may also require institutional change. For example, alignment of policies towards sustainable and healthy food systems may require building institutional capacity across policy silos.

As a tool for societal transformation, education is a powerful strategy to accelerate changes in the way we produce and consume food. Education refers to early learning and lifelong acquisition of skills for higher awareness and actions for solving food system challenges (FAO 20051287). Education also entails vocational training, research and institutional strengthening (Hollinger 20151288). Educational focus changes according to the supply side (e.g., crop selection, input resource management, yield improvement, and diversification) and the demand since (nutrition and dietary health implications). Education on food loss and waste spans both the supply and demand sides.

In developing countries, extension learning such as farmer field schools – also known asrural resources centers – are established to promote experiential learning on improved production and food transformation (FAO 2016c1289). In developed countries, education campaigns are being undertaken to reduce food waste, improve diets and redefine acceptable food (e.g., “less than perfect” fruits and vegetables), and ultimately can contribute to changes in the structure of food industries (Heller 20191290; UNCCD 20171291).

The design of new education modules from primary to secondary to tertiary education could help create new jobs in the realm of sustainability (e.g., certification programmes). For example, one area could be educating managers of recycling programmes for food-efficient cities where food and organic waste are recycled to become fertilisers (Jara-Samaniego et al. 20171292). Research and education need to be coordinated so that knowledge gaps can be filled and greater trust established in shifting behaviour of individuals to be more sustainable. Education campaigns can also influence policy and legislation, and help to advance successful outcomes for climate change mitigation and adaptation regarding supply-side innovations, technologies, trade, and investment, and demand-side evolution of food choices for health and sustainability, and greater gender equality throughout the entire food system (Heller 20191293).

5.7.5

Knowledge gaps and key research areas

Knowledge gaps around options and solutions and their (co-)benefits and trade-offs are increasingly important now that implementation of mitigation and adaptation measures is scaling up.

Research is needed on how a changing climate and interventions to respond to it will affect all aspects of food security, including access, utilisation and stability, not just availability. Knowledge gaps across all the food security pillars are one of the barriers hindering mitigation and adaptation to climate change in the food system and its capacity to deliver food security. The key areas for climate change, food systems, and food security research are enlisted below.

5.7.5.1

Impacts and adaptation

Climate Services (food availability). Agriculture and food security is a priority area for the Global Framework for Climate Services (GFCS) a programme of the World Meteorological Organization (WMO). The GFCS enables vulnerable sectors and populations to better manage climate variability and adapt to climate change (Hansen et al. 20181294). Global precipitation datasets and remote sensing technologies can be used to detect local to regional anomalies in precipitation as a tool for devising early-warning systems for drought-related impacts, such as famine (Huntington et al. 20171295).

Crop and livestock genetics (food availability, utilisation). Advances in plant breeding are crucial for enhancing food security under changing climate for a wide variety of crops including fruits and vegetables as well as staples. Genetics improvement is needed in order to breed crops and livestock that can both reduce GHG emissions, increase drought and heat tolerance (e.g., rice), and enhance nutrition and food security (Nankishore and Farrell 20161296; Kole et al. 20151297). Many of these characteristics already exist in traditional varieties, including orphan crops and indigenous and local breeds, so research is needed to recuperate such varieties and evaluate their potential for adaptation and mitigation.

Phenomics-assisted breeding appears to be a promising tool for deciphering the stress responsiveness of crop and animal species (Papageorgiou 20171298; Kole et al. 20151299; Lopes et al. 20151300; Boettcher et al. 20151301). Initially discovered in bacteria and archaea, CRISPR–Cas9 is an adaptive immune system found in prokaryotes and since 2013 has been used as a genome editing tool in plants. The main use of CRISPR systems is to achieve improved yield performance, biofortification, biotic and abiotic stress tolerance, with rice (Oryza sativa) being the most studied crop (Gao 20181302 and Ricroch et al. 20171303).

Climate impact models (food availability). Understanding the full range of climate impacts on staple crops (especially those important in developing countries, such as fruits and vegetables) is missing in the current climate impact models. Further, the CO2 effects on nutrition quality of different crops are just beginning to be parameterised in the models (Müller et al. 20141304). Bridging these gaps is essential for projecting future dietary diversity, healthy diets, and food security (Bisbis et al. 20181305). Crop model improvements are needed for simulation of evapotranspiration to guide crop water management in future climate conditions (Cammarano et al. 20161306). Similarly, mores studies are needed to understand the impacts of climate change on global rangelands, livestock and aquaculture, which have received comparatively less attention than the impacts on crop production.

Resilience to extreme events (food availability, access, utilisation, and stability). On the adaptation side, knowledge gaps include impacts of climate shocks (Rodríguez Osuna et al. 20141307) as opposed to impacts of slow-onset climate change, how climate-related harvest failures in one continent may influence food security outcomes in others, impacts of climate change on fruits and vegetables and their nutrient contents.

5.7.5.2

Emissions and mitigation

GHG emissions inventory techniques (food utilisation). Knowledge gaps include food consumption-based emissions at national scales, embedded emissions (overseas footprints) of food systems, comparison of GHG emissions per type of food systems (e.g., smallholder and large-scale commercial food systems), and GHG emissions from land-based aquaculture. An additional knowledge gap is the need for more socio-economic assessments of the potential of various integrated practices to deliver the mitigation potential estimated from a biophysical perspective. This needs to be effectively monitored, verified, and implemented, once barriers and incentives to adoption of the techniques, practices, and technologies are considered. Thus, future research needs fill the gaps on evaluation of climate actions in the food system.

Food supply chains (food availability). The expansion of the cold chain into developing economies means increased energy consumption and GHG emissions at the consumer stages of the food system, but its net impact on GHG emissions for food systems as a whole, is complex and uncertain (Heard and Miller 20161308). Further understanding of negative side effects in intensive food processing systems is still needed.

Blockchains, as a distributed digital ledger technology which ensures transparency, traceability, and security, is showing promise for easing some global food supply chain management challenges, including the need for documentation of sustainability and the circular economy for stakeholders including governments, communities, and consumers to meet sustainability goals. Blockchain-led transformation of food supply chains is still in its early stages; research is needed on overcoming barriers to adoption (Tripoli and Schmidhuber 20181309; Casado-Vara et al. 20181310; Mao et al. 20181311; Saberi et al. 20191312).

5.7.5.3

Synergies and trade-offs

Supply-side and demand-side mitigation and adaptation (food availability, utilisation). Knowledge gaps exist in characterising the potential and risks associated with novel mitigation technologies on the supply side (e.g., inhibitors, targeted breeding, cellular agriculture, etc.). Additionally, most integrated assessment models (IAMs) currently have limited regional data on BECCS projects because of little BECCS implementation (Lenzi et al. 20181313). Hence, several BECCS scenarios rely on assumptions regarding regional climate, soils and infrastructure suitability (Köberle et al. 20191314) as well as international trade (Lamers et al. 20111315).

Areas for study include how to incentivise, regulate, and raise awareness of the co-benefits of healthy consumption patterns and climate change mitigation and adaptation; to improve access to healthy diets for vulnerable groups through food assistance programmes; and to implement policies and campaigns to reduce food loss and food waste. Knowledge gaps also exist on the role of different policies, and underlying uncertainties, to promote changes in food habits towards climate resilience and healthy diets.

Food systems, land-use change, and telecoupling (food availability, access, utilisation). The analytical framework of telecoupling has recently been proposed to address this complexity, particularly the connections, flows, and feedbacks characterising food systems (Friis et al. 20161316; Easter et al. 20181317). For example, how will climate-induced shifts in livestock and crop diseases affect food production and consumption in the future. Investigating the social and ecological consequences of these changes will contribute to decision-making under uncertainty in the future. Research areas include food systems and their boundaries, hierarchies, and scales through metabolism studies, political ecology and cultural anthropology.

Food-Energy-Water Nexus (food availability, utilisation, stability). Emerging interdisciplinary science efforts are providing new understanding of the interdependence of food, energy, and water systems. These interdependencies are beginning to take into account climate change, food security, and AFOLU assessments (Scanlon et al. 20171318; Liu et al. 20171319). These science advances, in turn, provide critical information for coordinated management to improve the affordability, reliability, and environmental sustainability of food, energy, and water systems. Despite significant advances within the past decade, there are still many challenges for the scientific community. These include the need for interdisciplinary science related to the food-energy-water nexus; ground-based monitoring and modelling at local-to-regional scales (Van Gaelen et al. 2017); incorporating human and institutional behaviour in models; partnerships among universities, industry, and government to develop policy-relevant data; and systems modelling to evaluate trade-offs associated with food-energy-water decisions (Scanlon et al. 20171320).

However, the nexus approach, as a conceptual framework, requires the recognition that, although land and the goods and services it provides is finite, potential demand for the goods and services may be greater than the ability to supply them sustainably (Benton et al. 20181321). By addressing demand-side issues, as well as supply-side efficiencies, it provides a potential route for minimising trade-offs for different goods and services (Benton et al. 20181322) (Section 5.6).

5.8.1

Food price spikes

Under average conditions, global food system markets may function well, and equilibrium approaches can estimate demand and supply with some confidence; however, if there is a significant shock, the market can fail to smoothly link demand and supply through price, and a range of factors can act to amplify the effects of the shock, and transmit it across the world (Box 5.5). Given the potential for shocks driven by changing patterns of extreme weather to increase with climate change, there is the potential for market volatility to disrupt food supply through creating food price spikes. This potential is exacerbated by the interconnectedness of the food system (Puma et al. 20151323) with other sectors (i.e., the food system depends on water, energy, and transport) (Homer-Dixon et al. 20151324), so the impact of shocks can propagate across sectors and geographies (Homer-Dixon et al. 20151325). There is also less spare land globally than there has been in the past, such that if prices spike, there are fewer options to bring new production on stream (Marianela et al. 20161326).

Increasing extreme weather events can disrupt production and transport logistics. For example, in 2012 the USA Corn Belt suffered a widespread drought; USA corn yield declined 16% compared to 2011 and 25% compared to 2009. In 2016, a record yield loss in France that is attributed to a conjunction of abnormal warmness in late autumn and abnormal wet in the following spring (Ben-Ari et al. 20181327) is another well-documented example. To the extent that such supply shocks are associated with climate change, they may become more frequent and contribute to greater instability in agricultural markets in the future.

Furthermore, analogue conditions of past extremes might create significantly greater impacts in a warmer world. A study simulating analogous conditions to the Dust Bowl drought in today’s agriculture suggests that Dust Bowl-type droughts today would have unprecedented consequences, with yield losses about 50% larger than the severe drought of 2012 (Glotter and Elliott 20161328). Damages at these extremes are highly sensitive to temperature, worsening by about 25% with each degree centigrade of warming. By mid-century, over 80% of summers are projected to have average temperatures that are likely to exceed the hottest summer in the Dust Bowl years (1936) (Glotter and Elliott 20161329).

How a shortfall in production – or an interruption in trade due to an event affecting a logistics choke-point (Wellesley et al. 20171330) – of any given magnitude may create impacts depends on many interacting factors (Homer-Dixon et al. 20151331; Tadasse et al. 20161332; Challinor et al. 20181333). The principal route is by affecting agricultural commodity markets, which respond to a perturbation through multiple routes as in Figure 5.17. This includes pressures from other sectors (such as, if biofuels policy is incentivising crops for the production of ethanol, as happened in 2007–2008). The market response can be amplified by poor policies, setting up trade and non-trade barriers to exports, from countries seeking to ensure their local food security (Bailey et al. 20151334). Furthermore, the perception of problems can fuel panic buying on the markets that in turn drives up prices.

Thus, the impact of an extreme weather event on markets has both a trigger component (the event) and a risk perception component (Challinor et al. 2016, 2018). Through commodity markets, prices change across the world because almost every country depends, to a greater or lesser extent, on trade to fulfil local needs. Commodity prices can also affect local market prices by altering input prices, changing the cost of food aid, and through spill-over effects. For example, in 2007– 2008 the grain affected by extreme weather was wheat, but there was a significant price spike in rice markets (Dawe 20101335).

As discussed by Bailey et al. (2015)1336, there are a range of adaptation measures that can be put in place to reduce the impact of climate-related production shortfalls. These include (i) ensuring transparency of public and private stocks, as well as improved seasonal forecasting to signal forthcoming yield shortfalls (FAO 2016a1337; Ceglar et al. 20181338; Iizumi et al. 20181339), (ii) building real or virtual stockholdings, (iii) increasing local productivity and diversity (as a hedge against a reliance on trade) and (iv) ensuring smoother market responses, through, for example, avoiding the imposition of export bans.

In summary, given the likelihood that extreme weather will increase, in both frequency and magnitude (Hansen et al. 20121340; Coumou et al. 20141341; Mann et al. 20171342; Bailey et al. 20151343), and the current state of global and cross-sectoral interconnectedness, the food system is at increasing risk of disruption (medium evidence, medium agreement), with large uncertainty about how this could manifest. There is, therefore, a need to build resilience into international trade as well as local supplies.

The 2010–2011 food price spike was initially triggered by the exceptional heat in summer 2010, with an extent from Europe to the Ukraine and Western Russia (Barriopedro et al. 20111344; Watanabe et al. 20131345; Hoag 20141346). The heatwave in Russia was extreme in both temperature (over 40°C) and duration (from July to mid-August in 2010). This reduced wheat yields by approximately one third (Wegren 20111347; Marchand et al. 20161348). Simultaneously, in the Indus Valley in Pakistan, unprecedented rainfall led to flooding, affecting the lives and livelihoods of 20 million people. There is evidence that these effects were both linked and made more likely through climate change (Mann et al. 20171349).

In response to its shortfall in yields, Russia imposed an export ban in order to maintain local food supplies. Other countries responded in a largely uncoordinated ways, each of them driven by internal politics as well as national self-interests (Jones and Hiller 20171350). Overall, these measures led to rapid price rises on the global markets (Welton 20111351), partly through panic buying, but also through financial speculation (Spratt 20131352).

Analysis of responses to higher food prices in the developing world showed that lower-income groups responded by taking on more employment, reducing food intake, limiting expenditures, spending savings (if available), and participating in demonstrations. People often identified their problems as stemming from collusion between powerful incumbent interests (e.g., of politicians and big business) and disregard for the poor (Hossain and Green 20111353). This politicised social response helped spark food-related civil protest, including riots, across a range of countries in 2010–2011 (Natalini et al. 20171354). In Pakistan, food price rises were exacerbated by the economic impacts of the floods, which further contributed to food-related riots in 2010.

Price spikes also impact on food security in the developed world. In the UK, global commodity price inflation influenced local food prices, increasing food-price inflation by about five times at the end of 2010. Comparing household purchases over the five-year period from 2007 to 2011 showed that the amount of food bought declined, on average, by 4.2%, whilst paying 12% more for it. The lowest income decile spent 17% more by 2011 than they did in 2007 (Holding et al. 20131355; Tadasse et al. 20161356). Consumers also saved money by trading down for cheaper alternatives. For the poorest, in the extreme situation, food became unaffordable: the Trussell Trust, a charity supplying emergency food handouts for people in crisis, noted a 50% increase in handouts in 2010.

5.8.2

Migration and conflict

Since the IPCC AR5 (Porter et al. 20141357; Cramer et al. 20141358), new work has advanced multi-factor methodological issues related to migration and conflict (e.g., Kelley et al. 20151469, 20171470; Werrell et al. 20151359; Challinor et al. 20181360; Pasini et al. 20181361). These in particular have addressed systemic risks to food security that result from cascading impacts triggered by droughts and floods and how these are related to a broad range of societal influences.

Climate variability and extremes have short-, medium – and long-term impacts on livelihoods and livelihood assets – especially of the poor – contributing to greater risk of food insecurity and malnutrition (FAO et al. 20181362). Drought threatens local food security and nutrition and aggravates humanitarian conditions, which can trigger large-scale human displacement and create a breeding ground for conflict (Maystadt and Ecker 20141363). There is medium agreement that existing patterns of conflict could be reinforced under climate change, affecting food security and livelihood opportunities, for example, in already fragile regions with ethnic divides such as North and Central Africa as well as Central Asia (Buhaug 20161364; Schleussner et al. 20161365) (Box 5.6).

Challinor et al. (2018) have developed a typology for transboundary and transboundary risk transmission that distinguishes the roles of climate and social and economic systems. To understand these complex interactions, they recommend a combination of methods that include expert judgement; interactive scenario building; global systems science and big data; and innovative use of climate and integrated assessment models; and social science techniques (e.g., surveys, interviews, and focus groups).

5.8.2.1

Migration

There has been a surge in international migration in recent years, with around five million people migrating permanently in 2016 (OECD 20171366). Though the initial driver of migration may differ across populations, countries and contexts, migrants tend to seek the same fundamental objective: to provide security and adequate living conditions for their families and themselves. Food insecurity is a critical ‘push’ factor driving international migration, along with conflict, income inequality, and population growth. The act of migration itself causes food insecurity, given the lack of income opportunities and adverse conditions compounded by conflict situations.

Warner et al. (2012)1367 found the interrelationships between changing rainfall patterns, food and livelihood security in eight countries in Asia, Africa and Latin America. Several studies in Africa have found that persistent droughts and land degradation contributed to both seasonal and permanent migration (Gray 20111368; Gray and Mueller 20121369; Hummel 20151379; Henry et al. 20041371; Folami and Folami 20131372), worsening the vulnerability of different households (Dasgupta et al. 20141373).

Dependency on rainfed agriculture ranges from 13% in Mexico to more than 30% in Guatemala, Honduras, and Nicaragua, suggesting a high degree of sensitivity to climate variability and change, and undermined food security (Warner et al. 20091374). Studies have demonstrated that Mexican migration (Feng et al. 20101375; Nawrotzki et al. 20131376) and Central American migration (WFP 20171377) fluctuate in response to climate variability. The food system is heavily dependent on maize and bean production and long-term climate change and variability significantly affect the productivity of these crops and the livelihoods of smallholder farmers (WFP 20171378). In rural Ecuador, adverse environmental conditions prompt out-migration, although households respond to these challenges in diverse ways resulting in complex migratory responses (Gray and Bilsborrow 20131471).

Migration patterns have been linked to heat stress in Pakistan (Mueller et al. 20141380) and climate variability in the Sundarbans due to decline in food security (Guha and Roy 20161381). In Bangladesh, the impacts of climate change have been on the rise throughout the last three decades with increasing migration, mostly of men leaving women and children to cope with increasing effects of natural disasters (Rabbani et al. 20151382).

Small islands are very sensitive to climate change impacts (high confidence) (Nurse et al. 20141383) and impacted by multiple climatic stressors (IPCC 2018a1384 and SROCC). Food security in the Pacific, especially in Micronesia, has worsened in the past half century and climate change is likely to further hamper local food production, especially in low-lying atolls (Connell 2016). Migration in small islands (internally and internationally) occurs for multiple reasons and purposes, mostly for better livelihood opportunities (high confidence).

Beyond rising sea levels, the effects of increasing frequency and intensity of extreme events such as severe tropical cyclones are likely to affect human migration in the Pacific (Connell 20151385; Krishnapillai and Gavenda 20141386; Charan et al. 20171387; Krishnapillai 20171388). On Yap Island, extreme weather events are affecting every aspect of atoll communities’ existence, mainly due to the islands’ small size, their low elevation, and extensive coastal areas (Krishnapillai 20181389). Displaced atoll communities on Yap Island grow a variety of nutritious vegetables and use alternative crop production methods such as small-plot intensive farming, raised bed gardening, as part of a community-based adaptation programme (Krishnapillai and Gavenda 20141390; Krishnapillai 20181391).

Recurrences of natural disasters and crises threaten food security through impacts on traditional agriculture, causing the forced migration and displacement of coastal communities to highlands in search of better living conditions. Although considerable differences occur in the physical manifestations of severe storms, such climate stressors threaten the life-support systems of many atoll communities (Campbell et al. 20141392). The failure of these systems resulting from climate disasters propel vulnerable atoll communities into poverty traps, and low adaptive capacity could eventually force these communities to migrate.

Climate change-induced displacement and migration in the Pacific has received wide attention in the scientific discourse (Fröhlich and Klepp 20191472). The processes of climate change and their effects in the region have serious implications for Pacific Island nations as they influence the environments that are their ‘life-support systems’ (Campbell 20141473). Climate variability poses significant threats to both agricultural production and food security. Rising temperatures and reductions in groundwater availability, as well as increasing frequency and severity of disaster events translate into substantial impacts on food security, causing human displacement, a trend that will be aggravated by future climate impacts (ADB 20171474). Declining soil productivity, groundwater depletion, and non-availability of freshwater threatens agricultural production in many remote atolls.

Many countries in the Pacific devote a large share of available land area to agricultural production. For example, more than 60% of land area is cultivated in the Marshall Islands and Tuvalu and more than 40% in Kiribati and Tonga. With few options to expand agricultural area, the projected impacts of climate change on food production are of particular concern (ADB 20131475, 20171476). The degradation of available land area for traditional agriculture, adverse disruptions of agricultural productivity and diminishing livelihood opportunities through climate change impacts leads to increasing poverty and food insecurity, incentivising migration to urban agglomerations (ADB 20171477; FAO et al. 20181478).

Campbell (20141479) describe the trends that lead to migration. First, climate change, including rising sea levels, affects communities’ land security, which is the physical presence on which to live and sustain livelihoods. Second, they impinge on livelihood security (especially food security) of island communities where the productivity of both subsistence and commercial food production systems is reduced. Third, the effects of climate change are especially severe on small-island environments since they result in declining ecological habitat. The effects on island systems are mostly manifested in atolls through erosion and inundation, and on human populations through migration. Population growth and scenarios of climate change are likely to further induce food stress as impacts unfold in the coming decades (Campbell 20151480).

While the populations of several islands and island groups in the Pacific (e.g., Tuvalu, Carteret Islands, and Kiribati) have been perceived as the first probable victims of rising seas so that their inhabitants would become, and in some quarters already are seen to be, the first ‘environmental’ or ‘climate change refugees’, migration patterns vary. Especially in small islands, the range and nature of the interactions among economic, social, and/or political drivers are complex. For example, in the Maldives, Stojanov et al. (20171481) show that while collective perceptions support climate change impacts as being one of the key factors prompting migration, individual perceptions give more credence to other cultural, religious, economic or social factors.

In the Pacific, Tuvalu has long been a prime candidate to disappear due to rising sea levels, forcing human migration. However, results of a recent study (Kench et al. 20181482) challenge perceptions of island loss in Tuvalu, reporting that there is a net increase in land area of 73.5 ha. The findings suggest that islands are dynamic features likely to persist as habitation sites over the next century, presenting opportunities for adaptation that embrace the heterogeneity of island types and processes. Farbotko (20101483) and Farbotko and Lazrus (20121484) present Tuvalu as a site of ‘wishful sinking’, in the climate change discourse. These authors argue that representations of Tuvalu as a laboratory for global climate change migration are visualisations by non-locals.

In Nanumea (Tuvalu), forced displacements and voluntary migrations are complex decisions made by individuals, families and communities in response to discourses on risk, deteriorating infrastructure and other economic and social pressures (Marino and Lazrus 20151485). In many atoll nations in the Western Pacific, migration has increasingly become a sustainable livelihood strategy, irrespective of climate change (Connell 20151486).

In Lamen Bay, Vanuatu, migration is both a cause and consequence of local vulnerabilities. While migration provides an opportunity for households to meet their immediate economic needs, it limits the ability of the community to foster longer-term economic development. At the same time, migration adversely affects the ability of the community to maintain food security due to lost labour and changing attitudes towards traditional ways of life among community members (Craven 20151487).

5.8.2.2

Conflict

While climate change will not alone cause conflict, it is often acknowledged as having the potential to exacerbate or catalyse conflict in conjunction with other factors. Increased resource competition can aggravate the potential for migration to lead to conflict. When populations continue to increase, competition for resources will also increase, and resources will become even scarcer due to climate change (Hendrix and Glaser 20071393). In agriculture-dependent communities in low-income contexts, droughts have been found to increase the likelihood of violence and prolonged conflict at the local level, which eventually pose a threat to societal stability and peace (FAO et al. 20171394). In contrast, conflicts can also have diverging effects on agriculture due to land abandonment, resulting in forest growth, or agriculture expansion causing deforestation, for example, in Colombia (Landholm et al. 20191395).

Several studies have explored the causal links among climate change, drought, impacts on agricultural production, livelihoods, and civil unrest in Syria from 2007–2010, but without agreement as to the role played by climate in subsequent migration (Kelley et al. 2015, 2017; Challinor et al. 20181396; Selby et al. 20171397; Hendrix 20181398). Contributing factors that have been examined include rainfall deficits, population growth, agricultural policies, and the influx of refugees that had placed burdens on the region’s water resources (Kelley et al. 20151399). Drought may have played a role as a trigger, as this drought was the longest and the most intense in the last 900 years (Cook et al. 20161400; Mathbout et al. 20181401). Some studies linked the drought to widespread crop failure, but the climate hypothesis has been contested (Selby et al. 20171402; Hendrix 20181403). Recent evidence shows that the severe drought triggered agricultural collapse and displacement of rural farm families, with approximately 300,000 families going to Damascus, Aleppo and other cities (Kelley et al. 20171404).

Persistent drought in Morocco during the early 1980s resulted in food riots and contributed to an economic collapse (El-Said and Harrigan 20141405). A drought in Somalia that fuelled conflict through livestock price changes, establishing livestock markets as the primary channel of impact (Maystadt and Ecker 20141406). Cattle raiding as a normal means of restocking during drought in the Great Horn of Africa led to conflict (ICPAC and WFP 20171407) whereas a region-wide drought in northern Mali in 2012 wiped out thousands of livestock and devastated the livelihoods of pastoralists, in turn swelling the ranks of armed rebel factions and forcing others to steal and loot for survival (Breisinger et al. 20151408).

On the other hand, inter-annual adjustments in international trade can play an important role in shifting supplies from food surplus regions to regions facing food deficits which emerge as a consequence of extreme weather events, civil strife, and/or other disruptions (Baldos and Hertel 20151409). A more freely functioning global trading system is tested for its ability to deliver improved long run food security in 2050.

In summary, given increasing extreme events and global and cross-sectoral interconnectedness, the food system is at increasing risk of disruption, for example, via migration and conflict (high confidence). {5.2.3, 5.2.4}