Food Security

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).

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).

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 day¹ 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).

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).

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.

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.

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.

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).

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.

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.

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.

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).

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.

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.

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).

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.

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).

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).

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).

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).

Framing of adaptation in this section categorises and assesses adaptation measures as autonomous, incremental, and transformational (Glossary and Table 5.3). Adaptation responses can be reactive or anticipatory.

Autonomous. Autonomous adaptation in food systems does not constitute a conscious response to climatic stimuli but is triggered by changes in agroecosystems, markets, or welfare changes. It is also referred to as spontaneous adaptation (IPCC 2007⁴⁸¹). Examples of autonomous adaptation of rural populations have been documented in the Sahel (IRD 2017⁴⁸²). In India, farmers are changing sowing and harvesting timing, cultivating short duration varieties, inter-cropping, changing cropping patterns, investing in irrigation, and establishing agroforestry. These are considered as passive responses or autonomous adaptation, because they do not acknowledge that these steps are taken in response to perceived climatic changes (Tripathi and Mishra 2017⁴⁸³).

Incremental. Incremental adaptation maintains the essence and integrity of a system or process at a given scale (Park et al. 2012⁴⁸⁴). Incremental adaptation focuses on improvements to existing resources and management practices (IPCC 2014a⁴⁸⁵).

Transformational. Transformational adaptation changes the fundamental attributes of a socio-ecological system either in anticipation of, or in response to, climate change and its impacts (IPCC 2014a⁴⁸⁶). Transformational adaptation seeks alternative livelihoods and land-use strategies needed to develop new farming systems (Termeer et al. 2016⁴⁸⁷). For example, limitations in incremental adaptation among smallholder rice farmers in Northwest Costa Rica led to a shift from rice to sugarcane production due to decreasing market access and water scarcity (Warner et al. 2015⁴⁸⁸). Migration from the Oldman River Basin has been described as a transformational adaption to climate change in the Canadian agriculture sector (Hadarits et al. 2017⁴⁸⁹). If high-end scenarios of climate change eventuate, the food security of farmers and consumers will depend on how transformational change in food systems is managed. An integrated framework of adaptive transition – management of socio-technical transitions and adaptation to socio-ecological changes – may help build transformational adaptive capacity (Mockshell and Kamanda 2018⁴⁹⁰ and Pant et al. 2015). Rippke et al. (2016)⁴⁹¹ has suggested overlapping phases of adaptation needed to support transformational change in Africa.

Climate risks affect all pillars of food security, particularly stability because extreme events lead to strong variation to food access. The notion of risk is widely treated in IPCC reports (IPCC 2014c⁴⁹³) (see also Chapter 7 in this report). With food systems, many risks co-occur or reinforce each other, and this can limit effective adaptation planning as they require a comprehensive and dynamic policy approach covering a range of drivers and scales. For example, from the understanding by farmers of change in risk profiles to the establishment of efficient markets that facilitate response strategies will require more than systemic reviews of risk factors (Howden et al. 2007⁴⁹⁴).

Integration of Climate Change Adaptation (CCA) and Disaster Risk Reduction (DRR) helps to minimise the overlap and duplication of projects and programmes (Nalau et al. 2016⁴⁹⁵). Recently, countries started integrating the concept of DRR and CCA. For instance, the Philippines introduced new legislation calling for CCA and DRR integration, as current policy instruments had been largely unsuccessful in combining agencies and experts across the two areas (Leon and Pittock 2016⁴⁹⁶).

Studies reveal that the amplitude of interannual growing-season temperature variability is in general larger than that of long-term temperature change in many locations. Responding better to seasonal climate-induced food supply shocks therefore increases society’s capability to adapt to climate change. Given these backgrounds, seasonal crop forecasting and early response recommendations (based on seasonal climate forecasts), are emerging to strengthen existing operational systems for agricultural monitoring and forecasting (FAO 2016a⁴⁹⁷; Ceglar et al. 2018⁴⁹⁸ and Iizumi et al. 2018).

While adaptation and mitigation measures are intended to reduce the risk from climate change impacts in food systems, they can also be sources of risk themselves (e.g., investment risk, political risk) (IPCC 2014b⁴⁹⁹). Climate-related hazards are a necessary element of risks related to climate impacts but may have little or nothing to do with risks related to some climate policies/responses.

Adoption of agroecological practices could provide resilience for future shocks, spread farmer risk and mitigate the impact of droughts (Niles et al. 2018⁵⁰⁰) (Section 5.3.2.3). Traditionally, risk management is performed through multifunctional landscape approaches in which resource utilisation is planned across wide areas and local agreements on resource access. Multifunctionality permits vulnerable communities to access various resources at various times and under various risk conditions (Minang et al. 2015⁵⁰¹).

In many countries, governmental compensation for crop-failure and financial losses are used to protect against risk of severe yield reductions. Both public and private sector groups develop insurance markets and improve and disseminate index-based weather insurance programmes. Catastrophe bonds, microfinance, disaster contingency funds, and cash transfers are other available mechanisms for risk management.

In summary, risk management can be accomplished through agroecological landscape approaches and risk sharing and transfer mechanisms, such as development of insurance markets and improved index-based weather insurance programmes (high confidence).

Agroecological systems are integrated land-use systems that maintain species diversity in a range of productive niches. Diversified cropping systems and practicing traditional agroecosystems of crop production where a wide range of crop varieties are grown in various spatial and temporal arrangements, are less vulnerable to catastrophic loss (Zhu et al. 2011⁵⁰²). The use of local genetic diversity, soil organic matter enhancement, multiple-cropping or poly-culture systems, home gardening, and agroecological approaches can build resilience against extreme climate events (Altieri and Koohafkan 2008⁵⁰³).

However, Nie et al. (2016)⁵⁰⁴ argued that while integrated crop-livestock systems present some opportunities such as control of weeds, pests and diseases, and environmental benefits, there are some challenges, including yield reduction, difficulty in pasture-cropping, grazing, and groundcover maintenance in high rainfall zones, and development of persistent weeds and pests.

Adaptation measures based on agroecology entail enhancement of agrobiodiversity; improvement of ecological processes and delivery of ecosystem services. They also entail strengthening of local communities and recognition of the role and value of ILK. Such practices can enhance the sustainability and resilience of agricultural systems by buffering climate extremes, reducing degradation of soils, and reversing unsustainable use of resources; outbreak of pests and diseases and consequently increase yield without damaging biodiversity. Increasing and conserving biological diversity such as soil microorganisms can promote high crop yields and sustain the environment (Schmitz et al. 2015⁵⁰⁵; Bhattacharyya et al. 2016⁵⁰⁶; Garibaldi et al. 2017⁵⁰⁷).

Diversification of many components of the food system is a key element for increasing performance and efficiency that may translate into increased resilience and reduced risks (integrated land management systems, agrobiodiversity, ILK, local food systems, dietary diversity, the sustainable use of indigenous fruits, neglected and underutilised crops as a food source) (medium confidence) (Makate et al. 2016⁵⁰⁸; Lin 2011⁵⁰⁹; Awodoyin et al. 2015⁵¹⁰).

The more diverse the food systems are, the more resilient they are in enhancing food security in the face of biotic and abiotic stresses. Diverse production systems are important for providing regulatory ecosystem services such as nutrient cycling, carbon sequestration, soil erosion control, reduction of GHG emissions and control of hydrological processes (Chivenge et al. 2015⁵¹¹). Further options for adapting to change in both mean climate and extreme events are livelihood diversification (Michael 2017⁵¹²; Ford et al. 2015⁵¹³), and production diversity (Sibhatu et al. 2015⁵¹⁴).

Crop diversification, maintaining local genetic diversity, animal integration, soil organic matter management, water conservation, and harvesting the role of microbial assemblages. These types of farm management significantly affect communities in soil, plant structure, and crop growth in terms of number, type, and abundance of species (Morrison-Whittle et al. 2017⁵¹⁵). Complementary strategies towards sustainable agriculture (ecological intensification, strengthening existing diverse farming systems and investment in ecological infrastructure) also address important drivers of pollinator decline (IPBES 2016⁵¹⁶).

Evidence also shows that, together with other factors, on-farm agricultural diversity can translate into dietary diversity at the farm level and beyond (Pimbert and Lemke 2018⁵¹⁷; Kumar et al. 2015⁵¹⁸; Sibhatu et al. 2015⁵¹⁹). Dietary diversity is important but not enough as an adaptation option, but results in positive health outcomes by increasing the variety of healthy products in people’s diets and reducing exposure to unhealthy environments.

Locally developed seeds and the concept of seed sovereignty can both help protect local agrobiodiversity and can often be more climate resilient than generic commercial varieties (Wattnem 2016⁵²⁰; Coomes et al. 2015⁵²¹; van Niekerk and Wynberg 2017; Vasconcelos et al. 2013⁵²²). Seed exchange networks and banks protect local agrobiodiversity and landraces, and can provide crucial lifelines when crop harvests fail (Coomes et al. 2015⁵²³; van Niekerk and Wynberg 2017⁵²⁴; Vasconcelos et al. 2013⁵²⁵).

Related to locally developed seeds, neglected and underutilised species (NUS) can play a key role in increasing dietary diversity (high confidence) (Baldermann et al. 2016⁵²⁶; van der Merwe et al. 2016⁵²⁷; Kahane et al. 2013⁵²⁸; Muhanji et al. 2011⁵²⁹) (Box 5.3). These species can also improve nutritional and economic security of excluded social groups, such as tribals (Nandal and Bhardwaj 2014⁵³⁰; Ghosh-Jerath et al. 2015⁵³¹), indigent (Kucich and Wicht 2016⁵³²) or rural populations (Ngadze et al. 2017⁵³³).

Dietary diversity has also been correlated (medium evidence, medium agreement) to agricultural diversity in small-holder and subsistence farms (Ayenew et al. 2018⁵³⁴; Jones et al. 2014⁵³⁵; Jones 2017⁵³⁶; Pimbert and Lemke 2018⁵³⁷), including both crops and animals, and has been proposed as a strategy to reduce micronutrient malnutrition in developing countries (Tontisirin et al. 2002⁵³⁸). In this regard, the capacity of subsistence farming to supply essential nutrients in reasonable balance to the people dependent on them has been considered as a means of overcoming their nutrient limitations in sound agronomic and sustainable ways (Graham et al. 2007⁵³⁹).

Ecosystem-based adaptation (EbA). EbA is a set of nature-based methods addressing climate change adaptation and food security by strengthening and conserving natural functions, goods and services that benefit people. EbA approaches to address food security provide co-benefits such as contributions to health and improved diet, sustainable land management, economic revenue and water security. EbA practices can reduce GHG emissions and increase carbon storage (USAID 2017⁵⁴⁰).

Food production and consumption are strongly influenced by cultures and beliefs. Culture, values and norms are primary factors in most climate change and food system policies. The benefits of integrating cultural beliefs and ILK into formal climate change mitigation and adaptation strategies can add value to the development of sustainable climate change, rich in local aspirations, planned with, and for, local people (Nyong et al. 2007⁵⁵¹).

There are many current agricultural management practices that can be optimised and scaled up to advance adaptation. Among the often-studied adaptation options are increased soil organic matter, improved cropland management, increased food productivity, prevention and reversal of soil erosion (see Chapter 6 for evaluation of these practices in regard to desertification and land degradation). Many analyses have demonstrated the effectiveness of soil management and changing sowing date, crop type or variety (Waongo et al. 2015⁵⁶⁵; Bodin et al. 2016⁵⁶⁶; Teixeira et al. 2017⁵⁶⁷; Waha et al. 2013⁵⁶⁸; Zimmermann et al. 2017⁵⁶⁹; Chalise and Naranpanawa 2016⁵⁷⁰; Moniruzzaman 2015⁵⁷¹; Sanz et al. 2017⁵⁷²). Biophysical adaptation options also include pest and disease management (Lamichhane et al. 2015⁵⁷³) and water management (Palmer et al. 2015⁵⁷⁴; Korbeľová and Kohnová 2017⁵⁷⁵).

Considering the benefits of higher temperature in temperate climates and the increase of pasture with incremental warming in some humid and temperate grasslands, as well as potential negative effects, can be useful in planning adaptation strategies to future climate change. Rivera-Ferre et al. (2016b)⁵⁹¹ characterize adaptation for different livestock systems as managerial, technical, behavioural and policy-related options. Managerial included production adjustments (e.g., intensification, integration with crops, shifting from grazing to browsing species, multispecies herds, mobility, soil and nutrient management, water management, pasture management, corralling, feed and food storage, farm diversification or cooling systems); and changes in labour allocation (diversifying livelihoods, shifting to irrigated farming, and labour flexibility). Technological options included breeding strategies and information technology research. Behavioural options are linked to cultural patterns and included encouraging social collaboration and reciprocity, for example, livestock loans, communal planning, food exchanges, and information sharing. Policy options are discussed in Section 5.7 and Chapter 7.

Options may include livelihood diversification within and across sectors of fisheries, aquaculture and agriculture. Thus, adaptation options need to provide management approaches and policies that build the livelihood asset base, reducing vulnerability to multiple stressors with a multi-sector perspective (Badjeck et al. 2010⁵⁹²). In Bangladesh, fishing pressure on post-larval prawns has increased as displaced farmers have shifted to fishing following salt-water intrusion of agricultural land (Ahmed et al. 2013⁵⁹³). In West Africa, strategies to cope with sudden shifts in fisheries are wider-reaching and have included turning to seafood import (Gephart et al. 2017⁵⁹⁴) or terrestrial food production, including farming and bush-meat hunting on land (Brashares et al. 2004⁵⁹⁵).

Proposed actions for adaptation include effective governance, improved management and conservation, efforts to maximise societal and environmental benefits from trade, increased equitability of distribution and innovation in food production, and the continued development of low-input and low-impact aquaculture (FAO 2018c⁵⁹⁶).

Particular adaptation strategies proposed by FAO (2014a)⁵⁹⁷ include diverse and flexible livelihood strategies, such as introduction of fish ponds in areas susceptible to intermittent flood/drought periods; flood-friendly small-scale homestead bamboo pens with trap doors allowing seasonal floods to occur without loss of stocked fish; cage fish aquaculture development using plankton feed in reservoirs created by dam building; supporting the transition to different species, polyculture and integrated systems, allowing for diversified and more resilient systems; promotion of combined rice and fish farming systems that reduce overall water needs and provide integrated pest management; and supporting transitions to alternative livelihoods.

Risk reduction initiatives include innovative weather-based insurance schemes being tested for applicability in aquaculture and fisheries and climate risk assessments introduced for integrated coastal zone management. For aquaculture’s contribution to building resilient food systems, Troell et al. (2014)⁵⁹⁸ found that aquaculture could potentially enhance resilience through improved resource use efficiencies and increased diversification of farmed species, locales of production, and feeding strategies. Yet, its high reliance on terrestrial crops and wild fish for feeds, its dependence on freshwater and land for culture sites and its environmental impacts reduce this potential. For instance, the increase in aquaculture worldwide may enhance land competition for feed crops, increasing price levels and volatility and worsening food insecurity among the most vulnerable populations.

Fewer studies have been done on adaptation of food system transport and storage compared to the many studies on adaptation to climate in food production.

Adaptation measures are also being considered in trade, processing and packaging, other important components of the food system. These will enable availability, stability, and safety of food under changing climate conditions.

Trade. Brooks and Matthews (2015)⁶⁰⁴ found that food trade increases the availability of food by enabling products to flow from surplus to deficit areas, raises incomes and favours access to food, improves utilisation by increasing the diversity of national diets while pooling production risks across individual markets to maintain stability.

Processing. Growth of spoilage bacteria of red meat and poultry during storage due to increasing temperature has been demonstrated by European Food Safety Authority (EFSA Panel on Biological Hazards 2016⁶⁰⁵). In a recent experiment conducted on the optimisation of processing conditions of Chinese traditional smoke-cured bacon, Larou, Liu et al. (2018a) showed that the use of a new natural coating solution composed of lysozyme, sodium alginate, and chitosan during the storage period resulted in 99.69% rate of reducing deterioration after 30-day storage. Also, the use of High Hydrostatic Pressure (HHP) technology to inactivate pathogenic, spoilage microorganisms and enzymes (with little or no effects on the nutritional and sensory quality of foods) have been described by Wang et al. (2016)⁶⁰⁶ and Ali et al. (2018) ¹⁴³⁸as new advances in processing and packaging fruits, vegetables, meats, seafood, dairy, and egg products.

In summary, there are many practices that can be optimised and scaled up to advance supply-side adaptation. On-farm adaptation options include increased soil organic matter and erosion control in cropland, improved livestock and grazing land management, and transition to different species, polyculture and integrated systems in aquaculture. Crop and livestock genetic improvements include tolerance to heat, drought, and pests and diseases. Food transport, storage, trade, and processing will likely play increasingly important roles in adapting to climate change-induced food insecurity.

Climate change poses serious wide-ranging risks, requiring a broader approach in fighting the phenomenon. The United Nations Framework Convention on Climate Change (UNFCCC) and its annual Conferences of the Parties (COPs) has been instrumental in ensuring international cooperation in the field of tackling the impacts of climate change in a broader framework (Clémençon 2016⁶³¹). The National Adaptation Plan (NAP) programme under the UNFCCC was established to: identify vulnerable regions; assess the impacts of climate change on food security; and prioritise adaptation measures for implementation to increase resilience. The National Adaptation Programs of Action (NAPAs) was also established to support least-developed countries (LDCs) in addressing their particular challenges in adaptation, to enhance food security among other priorities.

The Paris Agreement (UNFCCC 2015⁶³²) is a major victory for small island states and vulnerable nations that face climate change-related impacts of floods and droughts resulting in food security challenges. Adaptation and mitigation targets set by the parties through their nationally determined commitments (NDCs) are reviewed internationally to ensure consistency and progress towards actions (Falkner 2016⁶³³).

The Food and Agriculture Organization of the United Nations (FAO) also plays a significant role in designing and coordinating national policies to increase adaptation and food security. The five key strategic objectives of FAO (help eliminate hunger, food insecurity and malnutrition; make agriculture, forestry and fisheries more productive and sustainable; reduce rural poverty; enable inclusive and efficient agricultural and food systems; and increase the resilience of livelihoods to climate threats) (FAO 2018e⁶³⁴), all relate to building resilience and increasing global adaptation to climate variability.

The successful development of food systems under climate change conditions requires a national-level management that involves the cooperation of a number of institutions and governance entities to enable more sustainable and beneficial production and consumption practices.

For example, Nepal has developed a novel multi-level institutional partnership, under the Local Adaptation Plan of Action (LAPA), which is an institutional innovation that aims to better integrate local adaptation planning processes and institutions into national adaptation processes. That includes collaboration with farmers and other non-governmental organisations (Chhetri et al. 2012⁶³⁶). By combining conventional technological innovation process with the tacit knowledge of farmers, this new alliance has been instrumental in the innovation of location-specific technologies thereby facilitating the adoption of technologies in a more efficient manner.

National Adaptation Planning of Indonesia was officially launched in 2014 and was an important basis for ministries and local governments to mainstream climate change adaptation into their respective sectoral and local development plans (Kawanishi et al. 2016⁶³⁷). Crop land-use policy – to switch from crops that are highly impacted by climate change to those that are less vulnerable – were suggested for improving climate change adaptation policy processes and outcomes in Nepal (Chalise and Naranpanawa 2016⁶³⁸).

Enhancement of representation, democratic and inclusive governance, as well as equity and fairness for improving climate change adaptation policy processes and outcomes in Nepal were also suggested as institutional measures by Ojha et al. (2015)⁶³⁹. Further, food, nutrition, and health policy adaptation options such as social safety nets and social protection have been implemented in India, Pakistan, Middle East and North Africa (Devereux 2015⁶⁴⁰; Mumtaz and Whiteford 2017⁶⁴¹; Narayanan and Gerber 2017⁶⁴²).

Financial incentives policies at the national scale used as adaptation options include taxes and subsidies; index-based weather insurance schemes; and catastrophe bonds (Zilberman et al. 2018⁶⁴³; Linnerooth-Bayer and Hochrainer-Stigler 2015⁶⁴⁴; Ruiter et al. 2017⁶⁴⁵ and Campillo et al. 2017). Microfinance, disaster contingency funds, and cash transfers are other mechanisms (Ozaki 2016⁶⁴⁵ and Kabir et al. 2016[reference number=1439).

Community-based adaptation (CBA) builds on social organisational capacities and resources to address food security and climate change. CBA represents bottom-up approaches and localised adaptation measures where social dynamics serve as the power to respond to the impacts of climate change (Ayers and Forsyth 2009⁶⁴⁶). It identifies, assists, and implements development activities that strengthen the capacity of local people to adapt to living in a riskier and less predictable climate, while ensuring their food security.

Klenk et al. (2017)⁶⁴⁷ 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 to climate change (Bocco and Napoletano, 2017⁶⁴⁸). Clemens et al. (2015)⁶⁴⁹ indicated that learning alliances provided social learning and knowledge-sharing in Vietnam through an open dialogue platform that provided incentives and horizontal exchange of ideas.

Community-based adaptation generates strategies through participatory processes, involving local stakeholders and development and disaster risk reduction practitioners. Fostering collaboration and community stewardship is central to the success of CBA (Scott et al. 2017⁶⁵⁰). Preparedness behaviours that are encouraged include social connectedness, education, training, and messaging; CBA also can encompass beliefs that might improve household preparedness to climate disaster risk (Thomas et al. 2015⁶⁵¹). Reliance on social networks, social groups connectivities, or moral economies reflect the importance of collaboration within communities (Reuter 2018⁶⁵²; Schramski et al. 2017⁶⁵³).

Yet, community-based adaptation also needs to consider methods that engage with the drivers of vulnerability as part of community-based approaches, particularly questions of power, culture, identity and practice (Ensor et al. 2018⁶⁵⁴). The goal is to avoid maladaptation or exacerbation of existing inequalities within the communities (Buggy and McNamara 2016⁶⁵⁵). For example, in the Pacific Islands, elements considered in a CBA plan included people’s development aspirations; immediate economic, social and environmental benefits; dynamics of village governance, social rules and protocols; and traditional forms of knowledge that could inform sustainable solutions (Remling and Veitayaki 2016⁶⁵⁶).

Many countries and regions in the world have adopted early warning systems (EWS) to cope with climate variability and change as it helps to reduce interruptions and improve response times before and after extreme weather events (Ibrahim and Kruczkiewicz 2016⁶⁶⁰). The Early Warning and Early Action (EW/EA) framework has been implemented in West Africa (Red Cross 2011) and Mozambique (DKNC 2012⁶⁶¹). Bangladesh has constructed cyclone shelters where cyclone warnings are disseminated and responses organised (Mallick et al. 2013⁶⁶²). In Benin, a Standard Operating Procedure is used to issue early warnings through the UNDP Climate Information and Early Warning Systems Project (UNDP 2016⁶⁶³).

However, there are some barriers to building effective early warning systems in Africa, such as lack of reliable data and distribution systems, lack of credibility, and limited relationships with media and government agencies (UNDP 2016⁶⁶⁴). Mainstreaming early warning systems in adaptation planning could present a significant opportunity for climate disaster risk reduction (Zia and Wagner 2015⁶⁶⁵). Enenkel et al. (2015) suggested that the use of smartphone applications that concentrate on food and nutrition security could help with more frequent and effective monitoring of food prices, availability of fertilisers and drought-resistant seeds, and could help to turn data streams into useful information for decision support and resilience building.

GIS and remote sensing technology are used for monitoring and risk quantification for broad-spectrum stresses such as drought, heat, cold, salinity, flooding, and pests (Skakun et al. 2017⁶⁶⁶; Senay et al. 2015⁶⁶⁷; Hossain et al. 2015⁶⁶⁸ and; Brown 2016⁶⁶⁹), while site-specific applications, such as drones, for nutrient management, precision fertilisers, and residue management can help devise context-specific adaptations (Campbell et al. 2016⁶⁷⁰ and; Baker et al. 2016⁶⁷¹). Systematic monitoring and remote sensing options, as argued by Aghakouchak et al. (2015)⁶⁷², showed that satellite observations provide opportunities to improve early drought warning. Waldner et al. (2015)⁶⁷³ found that cropland mapping allows strategic food and nutrition security monitoring and climate modelling.

Access to a wide range of adaptation technologies for precipitation change is important, such as rainwater harvesting, wastewater treatment, stormwater management and bioswales, water demand reduction, water-use efficiency, water recycling and reuse, aquifer recharge, inter-basin water transfer, desalination, and surface-water storage (ADB 2014⁶⁷⁴).

Financial instruments such as micro-insurance, index-based insurance, provision of post-disaster finances for recovery and pre-disaster payment are fundamental means to reduce lower and medium level risks (Linnerooth-Bayer and Hochrainer-Stigler 2014⁶⁷⁵). Fenton & Paavola, 2015⁶⁷⁶; Dowla, 2018⁶⁷⁷). Hammill et al. (2010)⁶⁷⁸ found that microfinance services (MFS) are especially helpful for the poor. MFS can provide poor people with the means to diversify, accumulate and manage the assets needed to become less susceptible to shocks and stresses. As a result, MFS plays an important role in vulnerability reduction and climate change adaptation among some of the poor. The provision of small-scale financial products to low-income and otherwise disadvantaged groups by financial institutions can serve as adaptation to climate change. Access to finance in the context of climate change adaptation that focuses on poor households and women in particular is bringing encouraging results (Agrawala and Carraro 2010⁶⁷⁹).

In summary, effective adaptation strategies can reduce the negative impacts of climate change. Food security under changing climate conditions depends on adaptation throughout the entire food system – production, supply chain, and consumption/demand, as well as reduction of food loss and waste. Adaptation can be autonomous, incremental, or transformative, and can reduce vulnerability and enhance resilience. Local food systems are embedded in culture, beliefs and values, and ILK can contribute to enhancing food system resilience to climate change (high confidence). Institutional and capacity-building measures are needed to scale up adaptation measures across local, national, regional, and global scales.

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).

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.

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.

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).

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).

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.

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.

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.

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).

Both reducing ruminant meat consumption and increasing its efficiency are often identified as the main options to reduce GHG emissions (GHGE) and to lessen pressure on land (Westhoek et al. 2014) (see Section 5.6 for synergies and trade-offs with health and Section 5.7 for discussion of Just Transitions). However, analysing ruminant meat production is highly complex because of the extreme heterogeneity of production systems and due to the numerous products and services associated with ruminants (Gerber et al. 2015⁸⁹³). See Supplementary Material Section SM5.5 for further discussion of uncertainties in estimates of livestock mitigation technical potential. Further, current market mechanisms are regarded as insufficient to decrease consumption or increase efficiency, and governmental intervention is often suggested to encourage mitigation in both the supply-side and demand-side of the food system (Section 5.7) (Wirsenius et al. 2011⁸⁹⁴; Henderson et al. 2018⁸⁹⁵).

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.

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.

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.

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%.

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).

‘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).

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).

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).

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).

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.

The co-benefits arising from mitigating climate change through changing dietary patterns, and thus demand, have potentially important economic impacts (high confidence). The gross value added from agriculture to the global economy (GVA) was 1.9 trillion USD2013 (FAO 2015c¹²⁴⁹), from a global agriculture economy (GDP) of 2.7 trillion USD2016. In 2013, the FAO estimated an annual cost of 3.5 trillion USD for malnutrition (FAO 2013a¹²⁵⁰).

However, this is likely to be an underestimate of the economic health costs of current food systems for several reasons: (i) lack of data – for example there is little robust data in the UK on the prevalence of malnutrition in the general population (beyond estimates of obesity and surveys of malnourishment of patients in hospital and care homes, from which estimates over 3 million people in the UK are undernourished (BAPEN 2012); (ii) lack of robust methodology to determine, for example, the exact relationship between over-consumption of poor diets, obesity and non-communicable diseases like diabetes, cardiovascular disease, a range of cancers or Alzheimer’s disease (Pedditizi et al. 2016¹²⁵¹), and (iii) unequal healthcare spending around the world.

In the USA, the economic cost of diabetes, a disease strongly associated with obesity and affecting about 23 million Americans, is estimated at 327 billion USD2017 (American Diabetes Association 2018¹²⁵²), with direct healthcare costs of 9600 USD per person. By 2025, it is estimated that, globally, there will be over 700 million people with diabetes (NCD-RisC 2016b¹²⁵³), over 30 times the number in the USA. Even if a global average cost of diabetes per capita were a quarter of that in the USA, the total economic cost of diabetes would be approximately the same as global agricultural GDP. Finally, (iv) the role of agriculture in causing ill-health beyond dietary health, such as through degrading air quality (e.g., Paulot and Jacob 2014¹²⁵⁴).

Whilst data of the healthcare costs associated with the food system and diets are scattered and the proportion of costs directly attributable to diets and food consumption is uncertain, there is potential for more preventative healthcare systems to save significant costs that could incentivise agricultural business models to change what is grown, and how. The potential of moving towards more preventative healthcare is widely discussed in health economics literature, particularly in order to reduce the life-style-related (including dietary-related) disease component in aging populations (e.g., Bloom et al. 2015¹²⁵⁵).

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).

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).

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.

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.

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).

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.

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).

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).

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).

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}