AR5 Chapter 7 (Porter et al., 2014) stated with confidence that warmer temperatures have benefited agriculture in the high latitudes, and more evidence has been published to support this statement. Typical examples include pole-ward expansion of growing areas and reduction of cold stress in East Asia and North America (Table SM5.1).

Recent warming trends have generally shortened the life cycle of major crops (high confidence) (Zhang et al., 2014; Shen and Liu, 2015; Ahmed et al., 2018; Liu et al., 2018c; Tan et al., 2021). Some studies, however, observed prolonged crop growth duration despite the warming trends (Mueller et al., 2015; Tao et al., 2016; Butler et al., 2018; Zhu et al., 2018b) because of shifts in planting dates and/or adoption of longer-duration cultivars in mid-to-high latitudes. Conversely, in mid-to-low latitudes in Asia, a review study found that farmers favoured early maturing cultivars to reduce risks of damages due to drought, flood and/or heat (Shaffril et al., 2018), suggesting that region-specific adaptations are already occurring in different parts of the world (high confidence).

Global yields of major crops per unit land area have increased 2.5- to 3-fold since 1960. Plant breeding, fertilisation, irrigation and integrated pest management have been the major drivers, but many studies have found significant impacts from recent climate trends on crop yield (high confidence) (Figure 5.3; see Section 5.2.1 for the change attributable to anthropogenic climate change).

Climate impacts for the past 20–50 years differ by crops and regions. Positive effects have been identified for rice and wheat in Eastern Asia, and for wheat in Northern Europe. The effects are mostly negative in Sub-Saharan Africa, South America and Caribbean, Southern Asia, and Western and Southern Europe. Climate factors that affected long-term yield trends also differ between regions. For example, in Western Africa, 1°C warming above preindustrial climate has increased heat and rainfall extremes, and reduced yields by 10–20% for millet and 5–15% for sorghum (Sultan et al., 2019). In Australia, declined rainfall and increased temperatures reduced yield potential of wheat by 27%, accounting for the low yield growth between 1990 and 2015 (Hochman et al., 2017). In Southern Europe, climate warming has negatively impacted yields of almost all major crops, leading to recent yield stagnation (Moore and Lobell, 2015; Agnolucci and De Lipsis, 2020; Brás et al., 2021).

Ortiz-Bobea et al. (2021) analysed agricultural total factor productivity (TFP), defined as the ratio of all agricultural outputs to all agricultural inputs, and found that, while TFP has increased between 1961 and 2015, the climate change trends reduced global TFP growth by a cumulative 21% over a 55-year period relative to TFP growth under counterfactual non-climate change conditions. Greater effects (30–33%) were observed in Africa, Latin America and the Caribbean (Figure 5.3).

Climate variability is a major source of variation in crop production (Ray et al., 2015; Iizumi and Ramankutty, 2016; Frieler et al., 2017; Cottrell et al., 2019)(Table SM5.1). Weather signals in yield variability are generally stronger in productive regions than in the less productive regions (Frieler et al., 2017), where other yield constraints exist such as pests, diseases and poor soil fertility (Mills et al., 2018; 5.2.2). Nevertheless, yield variability in less productive regions has severe impacts on local food availability and livelihood (high confidence) (FAO, 2021).

Climate-related hazards that cause crop losses are increasing (medium evidence, high agreement ) (Cottrell et al., 2019; Mbow et al., 2019; Brás et al., 2021; FAO, 2021; Ranasinghe et al., 2021). Drought-related yield losses have occurred in about 75% of the global harvested area (Kim et al., 2019b) and increased in recent years (Lesk et al., 2016). Heatwaves have reduced yields of wheat (Zampieri et al., 2017) and rice (Liu et al., 2019b). The combined effects of heat and drought decreased global average yields of maize, soybeans and wheat by 11.6%, 12.4% and 9.2%, respectively (Matiu et al., 2017). In Europe, crop losses due to drought and heat have tripled over the last five decades (Brás et al., 2021), pointing to the importance of assessing multiple stresses. Globally, floods also increased in the past 50 years, causing direct damages to crops and indirectly reduced yields by delaying planting, which cost 4.5 billion USD in the 2010 flood in Pakistan and 572 million USD in the 2015 flood in Myanmar (FAO, 2021).

The impact of climate change on these diverse crop types is under-researched and uncertain (Manners and van Etten, 2018; Alae-Carew et al., 2020); there are reports of positive impacts in some cases, but overall the observed impacts are negative across all crop categories (Figure 5.3).

Above-ground annual crops consumed as vegetables, fruits or salad are essential for food security and nutrition (5.12). In temperate regions, climate change can result in higher yields (Potopová et al., 2017; Bisbis et al., 2018), while in subtropical/tropical regions, negative impacts from heat and drought take precedence (Scheelbeek et al., 2018). Different species have different sensitivities to heat and drought (Prasad et al., 2017; Scheelbeek et al., 2018) and to combinations of stresses (Zandalinas et al., 2018). Above-ground vegetables are especially vulnerable to heat and drought stress during pollination and fruit set, resulting in negitive impacts on yield (Daryanto et al., 2017; Sita et al., 2017; Brás et al., 2021) and harvest quality (Mattos et al., 2014; Bisbis et al., 2018). Growers have already seen negative impacts from the expansion of pest and disease agents due to warming (Section 5.4.1.3; Figure 5.3).

Below-ground vegetables include starchy roots and tubers that form a regular diet in many parts of the tropics and subtropics. Warming and climate variability has altered the rate of tuber development, with yield impacts varying by location, including yield increases in some cases (Shimoda et al., 2018; Ray et al., 2019). These crops are considered stress tolerant but are more sensitive to drought than cereals (Daryanto et al., 2017). Impacts on water supply are critical as root crops are water-demanding for long periods, and highly sensitive to drought and heat events during tuber initiation (Dua et al., 2013; Potopová et al., 2017; Brás et al., 2021).

Among perennial tree crops, only grapevine, olive, almond, apple, coffee and cocoa have received significant research attention. Concerns about climate impacts on harvest quality are widespread (Figure 5.3) (Barnuud et al., 2014; Bonada et al., 2015). In higher-latitude regions, the primary concern is the effect of temperature variability on harvest stability, pests and diseases and phenology (including fulfilment of winter chill requirements and risks due to early emergence in spring), (El Yaacoubi et al., 2014; Ramírez and Kallarackal, 2015; Santos et al., 2017; Gitea et al., 2019). In lower-latitude regions, information is limited, but studies are focused on increased tree mortality and yield loss due to drought, heat and impacts from variability in the timing of the wet and dry seasons (Glenn et al., 2013; Ramírez and Kallarackal, 2015); see Box 5.7). In fruit trees, warming and climate variability have already affected fruit quality, such as acidity and texture in apples, or skin colour in grape berries (Sugiura et al., 2013; Sugiura et al., 2018). The reliability and stability of harvests has been impacted by climate variability, changes in the distribution of pests and pathogens (Seidel, 2014; Bois et al., 2017), and the mismatch of important phenological events (such as bud emergence and flowering) (Guo and Shen, 2015; Legave et al., 2015; Ito et al., 2018; Vitasse et al., 2018). Perennial crops are particularly vulnerable to these impacts as they are exposed throughout the year, with little potential for growers to adjust planting date or location. Negative impacts via disruption to phenology and pest dynamics are best studied in grapevine (see Box 5.2).

Among the fibre crops, cotton is particularly well studied. As cotton is heat tolerant and yield increases with extra plant growth, positive effects of increasing temperature are expected, but observed impacts have been mixed due to negative impacts on phenology and plant water status (Traore et al., 2013; Chen et al., 2015a; Cho and McCarl, 2017). Negative impacts of climate change due to proliferation of the pest cotton bollworm are widely reported (Ouyang et al., 2014; Huang and Hao, 2020).

The impacts of climate change on water availability (rainfall and irrigation supply) are an emerging issue. Increased occurrence of drought combined with limited access to irrigation water is already a key constraint; for example, Californian almonds are predicted to increase their potential geographical range under climate warming (Parker, 2018), yet a trend of increasing drought has already resulted in trees being removed due to lack of access to irrigation water (Keppen and Dutcher, 2015; Kerr et al., 2018; Reisman, 2019).

AR5 and SRCCL (IPCC, 2019) indicated that more frequent outbreaks and area expansion of pests and diseases are serious concerns under climate change but are under-researched because of the difficulties in assessing multi-species interactions (Porter et al., 2014; Mbow et al., 2019). High-quality historical and current observational data to detect changes in pests and diseases attributable to recent trends in climate are still limited.

Bebber (2013) found significant poleward expansions of many important groups of crop pests and pathogens since 1960, with an average shift of 2.7 km yr −1. Different pest species populations respond differently to ongoing climate change, with some shifting, contracting or expanding their current distribution range and others persisting or disappearing in their current range (high confidence). These asymmetric distribution changes can create novel species combinations or decouple existing ones (Pecl et al., 2017; Hobbs et al., 2018), but their consequences on future crop production and food security are hard to predict. Multi-species climate change experiments are rare (Bonebrake et al., 2018), but one study shows that under future climates different pest assemblages of interacting species may alter levels of damage to crops compared with that by only one species (Crespo-Perez et al., 2015). Some studies highlight the importance of location-specific species interactions for more realistic projections of pest distribution, performance and damage to crops, which in turn would allow more effective prevention and pest control strategies (Wilson et al., 2015; Carrasco et al., 2018).

Weeds are recognised as a primary constraint on crop production (Oerke, 2006), rangelands (DiTomaso et al., 2017) and forests (Webster et al., 2006). Climate change could favour the growth and development of weeds over crops with negative consequences for desired plants in managed systems (medium evidence, high agreement ) (Peters et al., 2014; Ziska and McConnell, 2016). First, changes in temperature and precipitation alter the range, composition and competitiveness of native and invasive weeds (Bradley et al., 2010). Second, rising concentrations of CO2 enhance growth of C3 species (~85% of plant species, including many weeds) (Ogren and Chollet, 1982; Ziska, 2003), and increase plant water use efficiency with potentially strong effects on invasive plant species establishment (Smith et al., 2000; Belote et al., 2004; Blumenthal et al., 2013).

Some invasive species within unmanaged areas will expand further, proliferate and be more competitive under climate change as they may benefit from increased resource ability (e.g., additional CO2, enhanced precipitation) (Bradley et al., 2010; Kathiresan and Gualbert, 2016; Merow et al., 2017; Ramesh et al., 2017; Waryszak et al., 2018), which will make chemical weed control more problematic (medium evidence, high agreement ) (Waryszak et al., 2018; Ziska, 2020). The range of other invasive weeds may become static, or even decline (Bradley et al., 2016; Buckley and Csergo, 2017). A recent meta-analysis also supports that invasive plants respond more favourably to elevated CO2 concentrations and elevated temperatures than native plants (Korres et al., 2016; Liu et al., 2017). Movement of invasive species into low-fertility areas, however, could provide resource opportunities, especially if agriculture in those areas is limited (Randriambanona et al., 2019).

Rising CO2 concentrations and climate change could reduce herbicide efficacy (medium evidence, high agreement ). These reductions may be associated with physical environmental changes (precipitation, wind speed) that influence herbicide coverage (Ziska, 2016) as well as direct effects of CO2 on plant biochemistry and herbicide resistance (Refatti et al., 2019). Increasing CO2 levels and altered temperature and precipitation are therefore projected to affect all aspects of weed biology (Peters et al., 2014; Ziska and McConnell, 2016), including establishment (Bradley et al., 2016), competition (Fernando et al., 2019), distribution, (Castellanos-Frías et al., 2016) and management (Waryszak et al., 2018).

A warmer climate increases the need for pesticides (Shakhramanyan et al., 2013; Ziska, 2014; Delcour et al., 2015; Zhang et al., 2018). Increases in temperature and CO2 concentration may reduce pesticide efficiency by altering its metabolism, or accelerating detoxification (Matzrafi et al., 2016; Matzrafi, 2019). Intense rainfall also reduces persistence (Delcour et al., 2015). Invasive pests and pathogens impose an additional cost for the society (Bradshaw et al., 2016). Rapid and large-scale dispersal of pests is already a major threat to food security, as exemplified by the recent outbreak of desert locusts (see Box 5.8), indicating the importance of international cooperation. Taken together, the need for control of pests, disease and weeds will increase under climate change (medium evidence, high agreement ). The use of toxic agricultural chemicals also has human health and environmental risks (Whitmee et al., 2015; IPBES, 2019). Surveillance for monitoring pest distribution and damages, climate-relevant pest risk analysis, and climate-smart strategies for controlling pests with minimal impacts on human and environmental health are important tools in the face of climate change (IPPC Secretariat, 2021).

Tropospheric (i.e., the lowest 6–10 km of the atmosphere) ozone exacerbates negative impacts of climate change (high confidence) (Mattos et al., 2014; Chuwah et al., 2015; McGrath et al., 2015; Bisbis et al., 2018; Mills et al., 2018; Scheelbeek et al., 2018). Ozone is an air pollutant and short-lived GHG that affects air quality and global climate. It is a strong oxidant that reduces physiological functions, yield and quality of crops and animals. Surface ozone concentration has increased substantially since the late 19th century (Cooper et al., 2014; Forster et al., 2021; Gulev et al., 2021; Szopa et al., 2021) and in some locations and times reaches levels that harm plants, animals and human (high confidence) (Fleming et al., 2018).

Mills (2018) estimated global distributions of current yield losses of major crops due to ozone, pest and diseases, heat, and aridity (Figure 5.4). Ozone-induced yield losses in 2010–2012 averaged 12.4%, 7.1%, 4.4% and 6.1% for soybean, wheat, rice and maize, respectively. Spatial variation in yield losses is similar among different stresses; areas with a large loss due to ozone are also at high risk of yield losses due to pest and diseases and heat. Many vegetable crops are also susceptible to ozone, which will adversely impact quality and quantity (Mattos et al., 2014; Bisbis et al., 2018; Scheelbeek et al., 2018).

The estimated yield loss does not account for interactions with other climatic factors. Temperatures enhance not only ozone production but also ozone uptake by plants, exacerbating yield and quality damage. Burney (2014) estimated current yield losses due to the combined effects of ozone and heat in India at 36% for wheat and 20% for rice. Schauberger et al. (2019a) found global yield losses, ranging from 2% to 10% for soybean and 0% to 39% for wheat with a model that accounts for temperature, water and CO2 concentration on ozone uptake.

Drought is a major risk component in cropping systems globally, with substantial economic loss (Kim et al., 2019b), livelihood impacts (Shiferaw et al., 2014; Miyan, 2015) and ultimately health risks such as malnutrition (Phalkey et al., 2015; Cooper et al., 2019). Vulnerability to drought can be estimated with a range of indicators (Hagenlocher et al., 2019). Meza (2020) showed that drought risks could be exacerbated or moderated by regional differences in vulnerability (Figure 5.5). For instance, high-level risks observed in southern Africa, western Asia and central Asia result from high vulnerability (low coping capacity), whereas risk levels are relatively low despite the high exposure by relatively high adaptive capacity to drought in other regions.

Regional-scale assessment also highlights the importance of adaptive capacity. For instance, rice and maize production in Viet Nam Mekong Delta has high exposure to multiple climate hazards such as flooding, sea level rise, salinity intrusion and drought (Parker et al., 2019). Risks can be moderated by a relatively high adaptive capacity because of infrastructure, resources and high education levels (Parker et al., 2019). Another regional study demonstrated that erratic rains and high temperatures in southern and southeastern Africa increased the vulnerability of agricultural soils, thereby exacerbating impacts of prolonged and frequent droughts (Sonwa et al., 2017a; See also Box 5.4).

Farm-scale assessment exemplifies context-sensitive vulnerability to climate hazards. Studies of coffee growers in Central America demonstrated that key vulnerability indicators varied greatly between regions and between farms, ranging from a lack of labour, postharvest infrastructure, conservation practices and transport that limits access to market, technical and financial assistance (Baca et al., 2014; Bouroncle et al., 2017). These region- and scale-specific vulnerability indicators assist in identifying ways to enhance resilience to climate hazards (high confidence).

While those working with major crops have benefited from the release of new cultivars, those growing other crops are typically reliant on a heritage cultivars or landraces. While Indigenous knowledge and local smallholder knowledge and practices play an important role in supporting agrobiodiversity which provides genetic diversity resistant to climate-related stresses, a global and national focus in international research, subsidies and support for a few crop species has contributed to an overall decline in agrobiodiversity (FAO, 2019e; Song et al., 2019) Similarly, there is a lack of agronomic innovation and research to service ‘minor’ crops (Moriondo et al., 2015; Manners and van Etten, 2018). Even some high-value commodities grown outside high-income countries suffer from imbalances in the focus of available credit, research and innovation (Section 5.4.4.3; Glover, 2014; Fischer, 2016; Farrell et al., 2018). There is a possibility that a lack of adaptive capacity and policy support will drive these growers to move away from these diverse crops, further reducing the resilience of food systems by increasing risk of crop loss from pests, disease and drought and potential loss of Indigenous or local knowledge (Section 5.13.5, Table Box 5.1.1). In the Andean Altiplano of Bolivia, for example, Indigenous farmers have traditionally managed a diverse set of native crops which are drought and frost-tolerant, using cultural practices of seed selection and exchange, but have faced an increase in pests and diseases and a decline of traditional crops due to climate-change-related stresses, out-migration and intensification drivers (Meldrum et al., 2018).

Social inequities such as gender, ethnicity and income level, which vary by time and place and may overlap, can compound vulnerability to climate change for producers within cropping systems (high confidence) (Table 5.3, Arora-Jonsson, 2011; Djoudi et al., 2013; Carr and Thompson, 2014; Mbow et al., 2019; Rao et al., 2019a; Nyantakyi-Frimpong, 2020a). Rather than binary and static categories (i.e., men versus women), social vulnerabilities are dynamic and intersect; to understand vulnerability, the specific socio-cultural identities and political and environmental context need to be studied in relation to climate stress (Thompson-Hall et al., 2016; Rao et al., 2019a; Nyantakyi-Frimpong, 2020a).

Table 5.3 | Examples of social inequities in cropping systems that compound climate change vulnerability.

Elevated CO2 concentrations stimulate photosynthesis rates and biomass accumulation of C3 crops, and enhance crop water use efficiency of various crop species, including C4 crops (high confidence) (Kimball, 2016; Toreti et al., 2020). Perennial crops and root crops may have a greater capacity for enhanced biomass under elevated CO2 concentrations, although this does not always result in higher yields (Glenn et al., 2013; Kimball, 2016).

Recent FACE studies found that the effects of elevated CO2 are greater under water-limited conditions (medium confidence) (Manderscheid et al., 2014; Fitzgerald et al., 2016; Kimball, 2016), which was generally reproduced by crop models (Deryng et al., 2016). However, drought sometimes negates the CO2 effects (Jin et al., 2018).

There are significant interactions between CO2, temperature, cultivars, nitrogen and phosphorous nutrients (Kimball, 2016; Toreti et al., 2020): positive effects of rising CO2 on yield are significantly reduced by higher temperatures for soybean, wheat and rice (medium confidence) (Ruiz-Vera et al., 2013; Cai et al., 2016; Gray et al., 2016; Hasegawa et al., 2016; Obermeier et al., 2016; Purcell et al., 2018; Wang et al., 2018). In above-ground vegetables, elevated CO2 can in some cases reduce the impact of other climate stressors, while in others the negative impacts of other abiotic factors negate the potential benefit of elevated CO2 (Bourgault et al., 2017; Bourgault et al., 2018; Parvin et al., 2018; Parvin et al., 2019). Significant variation exists among cultivars in yield response to elevated CO2, which is positively correlated with yield potential in rice and soybean, suggesting the potential to develop cultivars for enhanced productivity under future elevated [CO2] (Ainsworth and Long, 2021).

Elevated CO2 reduces some important nutrients such as protein, iron, zinc and some grains, fruit or vegetables to varying degrees depending on crop species and cultivars (high confidence) (Mattos et al., 2014; Myers et al., 2014; Dong et al., 2018; Scheelbeek et al., 2018; Zhu et al., 2018a; Jin et al., 2019; Ujiie et al., 2019). This is of particular relevance for fruit and vegetable crops given their importance in human nutrition (high confidence) (see Section 5.12.4 for potential impacts on nutrition; Nelson et al., 2018; Springmann et al., 2018). Recent experimental studies (Section 5.3.2), however, show some complex and counteracting interactions between CO2 and temperature in wheat, soybean and rice; heat stress negates the adverse effect of elevated CO2 on some nutrient elements (Macabuhay et al., 2018; Kohler et al., 2019; Wang et al., 2019b). The CO2 by temperature interaction for grain quality needs to be better understood quantitatively to predict food nutritional security in the future.

AR5 Chapter 7 estimated global crop yield reduction due to climate change to be about 1% per decade (Porter et al., 2014), similar to the previous assessment reports (Porter et al., 2019). Additional research confirms that climate change will disproportionately affect crop yields among regions, with more negative than positive effects being expected in most areas, especially in currently warm regions, including Africa and Central and South America (high confidence).

A systematic literature search between 2014 and 2020 resulted in about 100 peer-reviewed papers that simulated crop yields of four major crops (maize, rice, soybean and wheat) using Coupled Model Intercomparison Project Phase 5 (CMIP5) data (Hasegawa et al., 2021b). Most studies focus on the relative change in crop yields due to climate change but do not consider technological advances. Nevertheless, they provide useful insights into time-, scenario- and warming-degree-dependent impacts of climate change.

The impact of climate change on crop yield without adaptation projected in the 21st century is generally negative even with the CO2 fertilisation effects, with the overall median per-decade effect being −2.3% for maize, −3.3% for soybean, −0.7% for rice and −1.3% for wheat, which is consistent with previous IPCC assessments (Porter et al., 2014). The effects vary greatly within each crop, timeframe and RCP, but show a few common features across crops (Figure 5.6a ). Differences in the projected impacts between RCPs are not pronounced by mid-century. From then onward, the negative effect becomes more pronounced under RCP8.5, notably in maize. Rice yields show less variation across models than other crops presumably because simulations are mostly under irrigated conditions. A part of the uncertainty in the projection is due to regional differences (Figure 5. 6b). Negative impacts on cereals are projected in Africa and Central and South America at the end of the century, which agrees with the previous studies (Aggarwal et al., 2019; Porter et al., 2019).

The differences due to regions, RCPs and timeframes are related to the current temperature level and degree of warming (Figure 5.7). The projected effects of climate change are positive where current annual mean temperatures (Tave) are below 10°C, but they become negative with Tave above around 15°C. At Tave> 20°C, even a small degree of warming could result in adverse effects. In maize, negative effects are apparent at almost all temperature zones. A new study using the latest climate scenarios (Coupled Model Intercomparison Project Phase 6, CMIP6) and global gridded crop model ensemble projected that climate change impacts on major crop yields appear sooner than previously anticipated, mainly because of warmer climate projections and improved crop model sensitivities (Jägermeyr et al., 2021).

As noted in Section 5.3.1, most simulations do not fully account for responses to pests, diseases, long-term change in soil, and some climate extremes (Rosenzweig et al., 2014), but studies are emerging to include some of these effects. For example, based on the temperature response of insect pest population and metabolic process, global yield losses of rice, maize and wheat are projected to increase by 10–25% per degree Celsius of warming (Deutsch et al., 2018). Rising temperatures reduce soil carbon and nitrogen, which in turn exacerbate the negative effects of +3°C warming on yield from 9% to 13% in wheat and from 14% to 19% in maize (Basso et al., 2018).

A few studies have examined possible occurrences of tele-connected yield losses (5.4.1.2) using future climate scenarios. Tigchelaar (2018) estimated that, for the top four maize-exporting countries, the probability that simultaneous production losses greater than 10% occur in any given year increases from 0% to 7% under 2°C warming and to 86% under 4°C warming. Gaupp (2019) estimated that risks of simultaneous failure in maize would increase from 6% to 40% at 1.5°C and to 54% at 2°C warming, relative to the historical baseline climate. Large-scale changes in SST are the major factors causing simultaneous variation in climate extremes, which are projected to intensify under global warming (Cai et al., 2014; Perry et al., 2017). Consequently, risks of simultaneous yield losses in major food-producing regions will also increase with global warming levels above 1.5°C (medium confidence). Further examination is needed for the effects of spatial patterns of these extremes on breadbaskets in relation to SST anomalies under more extreme climate scenarios.

Future surface ozone concentration is highly uncertain (Fiore et al., 2012; Turnock et al., 2018); it is projected to increase under RCP8.5 and decrease under other RCPs depending largely on different methane emission trajectories because methane is an important precursor of ozone. Methane, therefore, reduces crop yield both from climate warming and ozone increase (Avnery et al., 2013). Shindell (2016) estimated yield losses of four major crops (to be 25±11% by 2100 under RCP8.5, as a net balance of the positive effect of CO2 (15±2%) and negative effects of warming (35±10%) and ozone (4.0±1.3%), and that 62% of the yield loss was attributable to methane. This points to the importance of reducing methane and other precursors of ozone as an effective adaptation strategy (medium evidence, high agreement ).

Yield projections for crops other than cereals indicate mostly negative impacts on production due to a range of climate drivers (high confidence), with yield reductions similar to that of cereals expected in tropical, subtropical and semi-arid areas (Mbow et al., 2019). Springmann et al. (2016), compared the projected global food availability for different food groups under the SSP2 2050 scenario and found reductions in availability were similar in cereals, fruit and vegetables, and root and tubers (with legumes and oilseed crops showing a smaller reduction).

Fruit and vegetables have not been subject to extensive or coordinated yield projections (Figure 5.8). Yield projections have been performed for individual crops and locations (Ruane, 2014; Adhikari et al., 2015; Awoye et al., 2017; Ramachandran et al., 2017), but more often crop suitability models have been used (SM5.3). Zhao (2019) introduced a modelling approach that could be used to generate yield projections for a wider range of annual crops. The discussion here also draws on reviews of more restricted experimental studies. Negative impacts of climate change on crop production are expected across many cropping systems (Figure 5.8). Apart from the direct effects of elevated carbon dioxide, most changes are expected to have negative effects on crop production. Changes in temperature and rainfall are most often mentioned as drivers of climate impacts, but expected changes in phenology, pests and diseases are also raising concerns. Scheelbeek et al. (2018) synthesised projections for vegetables and legumes, based on their response to climate factors under experimental conditions; in most cases, the magnitude of the changes is comparable to the RCP8.5 2100 forecasts. Scheelbeek et al. (2018) projected yield changes of: +22.0% (+11.6% to +32.5%) for a 250 ppm increase in CO2 concentration; −34.7% (−44.6% to −24.9%) for a 50% reduction in water availability; −8.9% (−15.6% to −2.2%) for a 25% increase in ozone concentration; −31.5% for a 4°C increase in temperature (in papers with a baseline temperature of >20°C). Overall, impacts are expected to be largely negative in regions where the temperature is currently above 20°C, while some yield gains are expected in cooler regions (provided that water availability and other conditions are maintained). Scheelbeek et al. (2018) did not consider changes in pest and disease pressure, which are projected to increase with warming (see SM5.3).

Systematic assessments of climate response for root crops as a group are lacking (Raymundo et al., 2014; Knox et al., 2016; Manners and van Etten, 2018). Climate suitability is projected to increase for tropical root crops (SM5.3), and some studies have found that root crops will be less negatively impacted than cereals, but there is no consensus on this (Brassard and Singh, 2008; Adhikari et al., 2015; Schafleitner, 2016; Manners et al., 2021). For potato, Raymundo et al. (2018) projected global yield reductions of 2–6% by 2055 under different RCPs, but with important differences among regions; tuber dry weight may experience reductions of 50–100% in marginal growing areas such as central Asia, while increases of up to 25% are expected in many high-yielding environments. Projections show yield increases of 6% per 100 ppm elevation in CO2 but declines of 4.6% per degree Celsius and 2% per 10% decrease in rainfall (Fleisher et al., 2017). Jennings et al. (2020) projected an overall increase in global potato production, but only if widespread adoption of adaptation measures is achieved. Although increases in CO2 could produce positive yield responses, the effects of temperature may offset these potential benefits (Dua et al., 2013; Raymundo et al., 2014). Warming offers the potential of longer growing seasons but can also have negative impacts through disrupted phenology and interactions with pests (Figure 5.8, Bebber, 2015; Pulatov et al., 2015).

Global yield modelling is lacking for woody perennial crops. Experimental studies suggest negative impacts on yields due to reduced water supply and increased soil salinity, as well as from warming and ozone (although evidence was limited for these) (Alae-Carew et al., 2020). Increasing CO2 is expected to increase yields, but only where other factors, such as warming, do not become yield-limiting (Alae-Carew et al., 2020). Many local projections include large uncertainty because of a lack of observational data and reliable parametrisation (Moriondo et al., 2015; Mosedale et al., 2016; Kerr et al., 2018; Mayer et al., 2019b). Most perennial crop models have found large negative impacts on yield and suitability, although CO2 fertilisation and phenology are not always considered (Lobell and Field, 2011; Glenn et al., 2013). Perennial crops are often grown in dryland areas where rainfall or irrigation water can be critical (Mrabet et al., 2020). Valverde (2015) found that yield losses in the Mediterranean region were largely driven by reduced rainfall, with maximum estimated yield losses of 5.4% for grape, 14.9% for olive and 27.2% for almond under a relatively hot and dry scenario (by 2041–2070). Moriondo (2015) highlight the need for perennial crop models to incorporate phenology and extreme climate events. Equally challenging is the need to estimate the impact of biotic changes, particularly climate-driven movement of pests and diseases (Ponti et al., 2014; Bosso et al., 2016; Schulze-Sylvester and Reineke, 2019).

For cotton, experimental studies suggest positive impacts from rising CO2 and temperature (Zhang et al., 2017a; Jans et al., 2021), but projections show mixed impacts on yield, including large negative impacts in warmer regions due to heat, drought and the interaction of temperature with phenology (Yang et al., 2014; Williams et al., 2015; Adhikari et al., 2016; Rahman et al., 2018). Climate change is also expected to increase the demand for irrigation water, which will likely limit production (Jans et al., 2021). There are also concerns that fibre quality may deteriorate (e.g., air permeability of compressed cotton fibers) (Luo et al., 2016).

Higher temperatures and altered moisture levels are expected to present a food safety risk, particularly for above-ground harvested vegetables (Figures 5.8; 5.10). Warmer and wetter weather is anticipated to increase fungal and microbial growth on leaves and fruit, while altered flooding regimes increase the risk of crop contamination (Liu et al., 2013; Uyttendaele et al., 2015). This is also true for perennial crops; for example, warming and climate variability can increase fungal contamination of grapes, including that associated with mycotoxins (Battilani, 2016; Paterson, 2018).

Cultural ecosystem services (CES) are those non-material benefits, such as aesthetic experiences, recreation, spiritual enrichment, social relations, cultural identity, knowledge and other values (Millennium Ecosystem Assessment, 2005), which support physical and mental health and human well-being (Chan et al., 2012; Triguero-Mas et al., 2015). CES in agricultural and wild landscapes include recreational activities, access to wild or cultivated products, and cultural foods, spiritual rituals, heritage and memory dimensions, and aesthetic experiences (Daugstad et al., 2006; Calvet-Mir et al., 2012; Ruoso et al., 2015). Relative to other ecosystem services, CES in agricultural landscapes have been less researched (Merlín-Uribe et al., 2012; Milcu et al., 2013; Bernues et al., 2014; Plieninger et al., 2014; van Berkel and Verburg, 2014; Ruoso et al., 2015; Quintas-Soriano et al., 2016). Agricultural heritage is a key aspect of CES and plays an important role in maintaining agrobiodiversity (Hanaček and Rodríguez-Labajos, 2018).

Climate change is projected to have negative impacts on CES (medium confidence) (Table 5.4). There is limited evidence that climate change has been the main driver affecting CES of agroecosystems confounded by other drivers such as migration and changing farming patterns (Hanaček and Rodríguez-Labajos, 2018; Dhakal and Kattel, 2019). Recent studies observed declines in CES in alpine pastures and floodplains in Europe in part due to climate change impacts (Probstl-Haider et al., 2016; Schirpke et al., 2019). Another study estimated that the scenic beauty enjoyed by those who visit the vineyards in central Chile will decline by 18–28% by 2050 owing to a combination of reduced precipitation, increased temperatures and natural fire cycles (Martinez-Harms et al., 2017). More research is needed, however, particularly on cultural heritage and spiritually significant places and in low-income countries.

Table 5.4 | Projected impacts on CES from climate change.

Crop management practices are the most commonly studied adaptation measures (Shaffril et al., 2018; Hansen et al., 2019a; Muchuru and Nhamo, 2019), but quantitative assessments are mostly limited to existing agronomic options such as changes in planting schedules, cultivars and irrigation (Beveridge et al., 2018a; Aggarwal et al., 2019). This section draws on the global data set used in Section 5.4.3.2 (Hasegawa et al., 2021b) to estimate adaptation potential, defined as the difference in simulated yields with and without adaptations. A caveat to the analysis is that the data set includes management options if the literature treats them as adaptation. They include intensification measures such as fertilizer and water management, not allowing for physical and economic feasibility.

The overall adaptation potential of existing farm management practices to reduce yield losses averaged 8% in mid-century and 11% in end-century (Figure 5.9), which is insufficient to offset the negative impacts from climate change, particularly in currently warmer regions (Section 5.4.3.2). Emission scenarios, crop species, regions and adaptation options do not show discernible differences. Combinations of two or more options do not necessarily have greater adaptation potential than a single option, though a fair comparison is difficult in the data set from independent studies. One regional study in West Africa found that currently promising management would no longer be effective under future climate, suggesting the need to evaluate effectiveness under projected climate change.

A global-scale meta-analysis estimated a 3–7% yield loss per degree Celsius increase in temperature (Zhao et al., 2017). Two global-scale studies using multiple global gridded crop models found that growing-season adaptation through cultivar changes offsets global production losses up to 2°C of temperature increase (Minoli et al., 2019; Zabel et al., 2021). While these studies do not account for CO2 fertilisation effects, another global-scale study with the CO2 fertilisation effects (Iizumi et al., 2020) showed that residual damage (climate change impacts after adaptation) would start to increase almost exponentially from 2040 towards the end of the century under RCP8.5. The cost required for adaptation and due to residual damage is projected to rise from USD 63 billion at 1.5°C to USD 80 billion at 2°C and to USD 128 billion at 3°C (Iizumi et al., 2020). All these global studies project that risks and damages are greater in tropical and arid regions, where crops are exposed to heat and drought stresses more often than in temperate regions (Sun et al., 2019; Kummu et al., 2021; SM5.4). There are still large uncertainties in the crop model projections (Müller et al., 2021a), but these multiple lines of evidence suggest that warming beyond +2°C (projected to be reached by mid-century under high-emission scenarios) will substantially increase the cost of adaptation and the residual damage to major crops (high confidence). The residual damage will prevail much sooner in currently warmer regions, where the effect of even a modest temperature increase is greater (Section 5.4.3.2).

Most crop modelling studies on adaptation are still limited to a handful of options for each crop type (Beveridge et al., 2018a). A range of other options are possible not just to reduce yield losses but to diversify risks to livelihoods, which are partially assessed in Sections 5.4.4.4 and 5.14.1. Current modelling approaches are not suited for the assessment of multiple dimensions of adaptation options. New studies are emerging that evaluate multiple options for productivity, sustainability and GHG emission (Xin and Tao, 2019; Smith et al., 2020b), but local- and household-scale assessment, taking account of future climatic variability, needs to be enhanced (Beveridge et al., 2018a).

Across this diverse group of cropping systems, distinct adaptation options and adaptation limits have emerged (Figure 5.10; Acevedo et al., 2020; Berrang-Ford et al., 2021b). Some crop types have already seen widescale implementation of climate adaptation (e.g., grapevines), while others show little evidence of preparation for climate change (e.g., leafy salad crops). Many adaptation responses are shared with the major crops, but prominent options such as plant breeding are underutilised and there is a lack of evidence for assessing adaptation for many crops (Bisbis et al., 2018; Gunathilaka et al., 2018; Manners and van Etten, 2018). Figure 5.10 assesses several adaptation options based on the perceived importance of each in the literature. Fruit and vegetable crops tend to be more reliant on ecosystem services in the form of pollination, biocontrol and other resources (water, nutrients, microbes, etc.), and ecosystem-based adaptation options are prominent. The range of crops means that there is great potential for crop switching, but cultural and economic barriers will make such options difficult to implement, with barriers to entry for production and marketing (Waha et al., 2013; Magrini et al., 2016; Kongsager, 2017; Rhiney et al., 2018). Perennial crops are exposed to a wide range of climate factors throughout the year and have significant barriers to implementing some of the common adaptation options, such as relocation or replacing tree species/cultivar; agronomic interventions on-farm are well used in high-value tree crops and provide some climate resilience, but longer-term options will be needed (Glenn et al., 2013; Mosedale et al., 2016; Gunathilaka et al., 2018; Sugiura, 2019).

Many fruit and vegetable crops are water demanding, and adaptation responses relating to water management and access to irrigation water are crucial. Rainwater storage and deficit irrigation techniques are frequently mentioned as adaptation options and can minimise the burden on off-farm water supplies (Bisbis et al., 2018; Acevedo et al., 2020).

As stated in AR5, cultivar improvements are one effective countermeasure against climate change (Porter et al., 2014; Challinor et al., 2016; Atlin et al., 2017). Plant breeding biotechnology for climate change adaptation draws upon modern biotechnology and conventional breeding, with the latter often assisted by genomics and molecular markers. Plant breeding biotechnology will contribute to adaptation for large-scale producers (high confidence). However, in addition to inconsistencies in meeting farmer expectations, a variety of socioeconomic and political variables strongly influence, and limit, uptake of climate-resilient crops (Acevedo et al., 2020; Rhoné et al., 2020).

Genome sequencing significantly increases the rate and accuracy for identifying genes of agronomic traits that are relevant to climate change, including adaptation to stress from pests and disease, temperature and water extremes (high confidence) (Brozynska et al., 2016; Scheben et al., 2016; Voss-Fels and Snowdon, 2016). Access to this information where it is needed and in practical timeframes, as well as the expertise to use it, will limit the sharing of benefits by the most vulnerable groups and countries (high agreement , limited evidence) (Heinemann et al., 2018).

Genetic improvements for climate change adaptation using modern biotechnology have not reliably translated into the field (Hu and Xiong, 2014; Nuccio et al., 2018; Napier et al., 2019), but good progress has been made by conventional breeding. Desirable traits that adapt plants to environmental stress are inherited as a complex of genes, each of which makes a small contribution to the trait (Negin and Moshelion, 2017). Adaptation by conventional breeding requires making rapid incremental changes in the best germplasm to keep pace with the environment (Millet et al., 2016; Atlin et al., 2017; Cobb et al., 2019). Further improvements would be difficult without in situ and ex situ conservation of plant genetic resources to maintain critical germplasm for breeding (Dempewolf et al., 2014; Castañeda-Álvarez et al., 2016).

Despite the advances in sequencing, phenotyping remains a significant bottleneck (Ghanem et al., 2015; Negin and Moshelion, 2017; Araus and Kefauver, 2018); the emergence of high-throughput phenotyping platforms may reduce this bottleneck in future. Emerging modern biotechnology such as gene/genome editing may in the future increase the ability to better translate genetic improvements into the field (medium agreement , limited evidence) (Puchta, 2017; Yamamoto et al., 2018; Friedrichs et al., 2019; Kawall, 2019; Zhang et al., 2019b).

Other breeding approaches assisted by genomics have been making steady gains in introducing traits that adapt crops to climate change (high confidence). DNA sequence information is used to identify markers of desirable traits that can be enriched in breeding programmes, as well as to quantify the genetic variability in species (Gepts, 2014; Brozynska et al., 2016; Voss-Fels and Snowdon, 2016). However, breeding for smallholder farmers and the stresses caused by climate change are unlikely to be addressed by the private sector and will require more public investment and adjusting to the local social-ecological system (Glover, 2014; Heinemann et al., 2014; Acevedo et al., 2020). Modern biotechnology has not demonstrated the scale neutrality needed to serve smallholder-dominated agroecosystems, due to a combination of the kinds of traits and restrictions that come from the predominant intellectual property rights instruments used in their commercialisation, as well as the focus on a small number of major crop species (medium confidence) (Fischer, 2016; Montenegro de Wit et al., 2020).

Globally, there is a notable lack of programmes aimed specifically at breeding for climate resilience in fruits and vegetables, although there have been calls to begin this process (Kole et al., 2015). Breeding for climate resilience in vegetables has great potential given the range of crop species available. Tolerance to abiotic stress is reasonably advanced in pulses (Araújo et al., 2015; Varshney et al., 2018), but examples of translation to commercial cultivars are still limited (Varshney et al., 2018; Varshney et al., 2019). The infrastructure for germplasm collection, maintenance, testing and breeding lags behind that of major crops (partly because of the large number of species involved) (Keatinge et al., 2016; Atlin et al., 2017).

Participatory plant breeding (PPB) facilitates interaction between Indigenous and local knowledge systems and scientific research and can be an effective adaptation strategy in generating varieties well adapted to the socio-ecological context and climate hazards (high confidence) (Table 5.5, Westengen and Brysting, 2014; Humphries et al., 2015; Anderson et al., 2016; Migliorini et al., 2016; Leitão et al., 2019; Ceccarelli and Grando, 2020; Singh et al., 2020).

Table 5.5 | PPB as cultivar improvement adaptation method.

Diversifying agricultural systems is an adaptation strategy that can strengthen resilience to climate change, with socioeconomic and environmental co-benefits, but trade-offs and benefits vary by socio-ecological context (high confidence) (Table 5.6, M’Kaibi et al., 2015; Bellon et al., 2016; Jones, 2017b; Schulte et al., 2017; Jarecki et al., 2018; Jones et al., 2018; Luna-Gonzalez and Sorensen, 2018; Sibhatu and Qaim, 2018; Renard and Tilman, 2019; Rosa-Schleich et al., 2019; Bozzola and Smale, 2020; Mulwa and Visser, 2020). Crop diversification alongside livestock, fish and other species can be applied at various scales in a range of systems, from rainfed or irrigated to urban and home gardens in multiple spatial and temporal arrangements such as mixed planting, intercrops, crop rotation, diversified management of field margins, agroforestry (Section 5.10.1.3) and integrated crop livestock systems (Section 5.10.1.1, Isbell et al., 2017; Kremen and Merenlender, 2018; Dainese et al., 2019; Rosa-Schleich et al., 2019; Hussain et al., 2020; Renwick et al., 2020; Tamburini et al., 2020; Snapp et al., 2021; see Section 5.1 4 and Cross-Chapter Box NATURAL in Chapter 2).

Table 5.6 | Agroecosystem diversification practices, climate change adaptation mechanisms, trade-offs, co-benefits and constraints to implementation.

Diversification improves regulating and supporting ecosystem services such as pest control, soil fertility and health, pollination, nutrient cycling, water regulation and buffering of temperature extremes (high confidence) (Barral et al., 2015; Prieto et al., 2015; Tiemann et al., 2015; Schulte et al., 2017; Beillouin et al., 2019a; Dainese et al., 2019; Kuyah et al., 2019; Tamburini et al., 2020), which can in turn mediate yield stability and reduced risk of crop loss according to socio-ecological contexts and time since adoption (high confidence) (Prieto et al., 2015; Roesch-McNally et al., 2018; Sida et al., 2018; Williams et al., 2018; Birthal and Hazrana, 2019; Degani et al., 2019; Amadu et al., 2020; Bowles et al., 2020; Li et al., 2020; Sanford et al., 2021).

Agroecosystem diversification often has variable impacts depending on crop combination, agro-ecological zone and soil types, and rigorous assessments of adaptive gains with traditional and locally diversified systems and potential trade-offs still need to be conducted across socio-ecological contexts. The quantitative upstanding will assist in enhancing multiple benefits of diversification tailored for each condition (Table 5.6). Progress is also needed via breeding and/or agronomy to adapt underutilised as well as major food crops to diversified agroecosystems and optimise management of nutrients, pest and disease pressure and other socio-ecological constraints (Araújo et al., 2015; Foyer et al., 2016; Adams et al., 2018; Pang et al., 2018).

Managing for diversity and flexibility at multiple scales is central to developing adaptive capacity. Policies to support diversification include shifting subsidies towards diversified systems, public procurement for diverse foods for schools and other public institutions, investment in shorter value chains, lower insurance premiums and payments for ecosystem services that include diversification (Sorensen et al., 2015; Guerra et al., 2017; Nehring et al., 2017; Valencia et al., 2019). Integrated landscape approaches involving multiple stakeholders (Reed et al., 2016) including urban governments can support diversification at a regional scale through public and private sector investment in extension services, regional supply chains, agritourism and other incentives for diversified landscapes (Milder et al., 2014; Münke et al., 2015; Sorensen et al., 2015; Pérez-Marin et al., 2017; Caron et al., 2018; 5.14.1.5).

Many grassland-based livestock systems are vulnerable to climate change and increases in climate variability (high confidence) (Dasgupta et al., 2014; Sloat et al., 2018; Stanimirova et al., 2019). Decadal vegetation changes from warming and drying trends have been detected in North American grasslands, with implications for species composition, rangeland quality and economic viability of grazing livestock (Rondeau et al., 2018; Reeves et al., 2020). Feed quality in South Asian grasslands has been negatively affected, reducing food security (Rasul et al., 2019). Increased grassland degradation has been observed in parts of Inner Mongolia (Nandintsetseg et al., 2021). Changing seasonality, increasing frequency of drought and rising temperatures are affecting pastoral systems globally (high confidence). These and other drivers are reducing herd mobility, decreasing productivity, increasing incidence of vector borne diseases and parasites, and reducing access to water and feed (high agreement , medium evidence) (López-i-Gelats et al., 2016; Vidal-González and Nahhass, 2018; de Leeuw et al., 2020).

There is limited evidence of observed distributional changes in livestock species due to climate changes. Asian buffalo and yak breeds in China over the past 50 years have shifted distribution partly because of increases in heat stress (Wu, 2015; Wu, 2016). Nepalese cattle numbers have declined, attributed to increases in the number of hot days (Koirala and Shrestha, 2017).

Climate variability has been identified as the primary cause of vegetation cover changes on the Tibetan Plateau since 2000 (Lehnert et al., 2016). Increasing inter-annual variability is a driver of farm extensification in Mediterranean dairy systems (Dono et al., 2016). In Australian rangelands (Godde et al., 2019) and dairy systems (Harrison et al., 2016; Harrison et al., 2017), increasing rainfall variability contributes more to stocking rate and profitability variability than changes in mean rainfall.

Climate change is affecting the transmission of vector-borne diseases (Hutter et al., 2018; Semenza and Suk, 2018) and parasites (Rinaldi et al., 2015) in high latitudes (high confidence). Different processes link climate change and infectious diseases in domesticated livestock. Some show a positive association between temperature and range expansion of arthropod vectors that spread the bluetongue virus. Others show a contraction, such as tsetse flies that transmit trypanosome parasites of several livestock species. Positive associations have been found between temperature and the spread of pathogens such as anthrax, and droughts and ENSO weather patterns and Rift Valley fever outbreaks in East Africa (Bett et al., 2017). Observed range expansion of economically important tick disease vectors in North America (Sonenshine, 2018) and Africa (Nyangiwe et al., 2018) are presenting new public health threats to humans and livestock.

Most domestic livestock have comfort zones in the range 10–30°C, depending on species and breed (Nardone et al., 2006). At higher temperatures, animals eat 3–5% less per additional degree of temperature, reducing their productivity and fertility. Heat stress suppresses the immune and endocrine system, enhancing susceptibility of the animal to disease (Das et al., 2016b). Recent stagnation in dairy production in West Africa and China may be associated with increased periods of high daily temperatures (low confidence) (Rahimi et al., 2020; Ranjitkar et al., 2020). Increases in the productive capacity of domestic animals can compromise thermal acclimation and plasticity, creating further loss. Escalating demand for livestock products in low-to-middle-income countries (LMICs) may necessitate considerable adaptation in the face of new thermal environments (medium confidence) (Collier and Gebremedhin, 2015; Theusme et al., 2021). Heat effects on productivity have been summarised for pigs (da Fonseca de Oliveira et al., 2019), sheep and goats (Sejian et al., 2018), and cattle (Herbut et al., 2019). The direct effects of higher temperatures on the smaller ruminants (sheep and goats) are relatively muted, compared with large ruminants; goats are better able to cope with multiple stressors than sheep (Sejian et al., 2018). Under SSP5-8.5 to mid-century, land suitability for livestock production will decrease because of increased heat stress prevalence in mid and lower latitudes (high confidence) (Thornton et al., 2021).

Livestock production may account for 30% of all water (blue, green and grey) used in agriculture (Mekonnen and Hoekstra, 2010) and can negatively affect water quality. Cropland feed production accounts for 38% of crop water consumption (Weindl et al., 2017). High-input livestock systems may consume more water than grazing or mixed systems, though water used per kg beef produced, for example, depends on country, context and system (Noya et al., 2019). In systems where feed production is rainfed, livestock and crop water productivity may be comparable (Haileslassie et al., 2009). Direct water consumption by livestock is <1–2% of global water consumption (Hejazi et al., 2014). Rising temperatures increase animal water needs, potentially affecting access of herders and livestock to drinking water sources (Flörke et al., 2018).

Climate change will have effects on future distribution, incidence and severity of climate-sensitive infectious diseases of livestock (high confidence) (Bett et al., 2017). In an assessment of climate sensitivity of European human and domestic animal infectious pathogens, 63% were sensitive to rainfall and temperature, and zoonotic pathogens were more climate-sensitive than human- or animal-only pathogens (McIntyre et al., 2017). Over the last 75 years, >220 emerging zoonotic diseases, some associated with domesticated livestock, have been identified, several of which may be affected by climate change, particularly vector-borne diseases (Vaillancourt and Ogden, 2016; see Cross-Chapter Box ILLNESS in Chapter 2). Walsh et al. (2018) identified both temperature and rainfall as influential factors in predicting increasing anthrax outbreaks in northern latitudes. Growing infectious disease burdens in domesticated animals may have wide-ranging impacts on the vulnerability of rural livestock producers in the future, particularly related to human health and projected increases in zoonoses (high confidence) (Bett et al., 2017; Heffernan, 2018; Rushton et al., 2018; Meade et al., 2019).

There is limited evidence about the role of livestock in addressing socioeconomic vulnerability. Although agriculture in parts of North America has become more sensitive to climate over the last 50 years, livestock have helped to moderate this effect, being less sensitive to increasing temperatures than some specialised crop systems (Ortiz-Bobea et al., 2018). Increasing frequency and severity of droughts will affect the future economic viability of grassland-based livestock production in the North American Great Plains (Briske et al., 2021). Purchasing more forage and selling more livestock have reduced household vulnerability in semi-arid parts of China over the last 35 years (Bai et al., 2019). A greater focus on sheep production away from cropping has increased the resilience of farming systems in Western Australia in low-rainfall years, although with mixed environmental effects (Ghahramani and Bowran, 2018). More insights are needed as to where and how livestock can affect the vulnerability of farmers and pastoralists.

Vulnerability to the health impacts of climate change will be shaped by existing burdens of ill health and is expected to be highest in poor and socioeconomically marginalised populations (high agreement , limited evidence) (Labbé et al., 2016). In addition to projected changes in infectious disease burdens, labour capacity in a warming climate is anticipated to decrease further, beyond the >5% drop estimated since 2000 (Watts et al., 2018). Loss of labour capacity may greatly increase the vulnerability of subsistence livestock keepers (high agreement , limited evidence).

Vulnerability to climate change depends on demography and social roles (Mbow et al., 2019). Gender inequities can act as a risk multiplier, with women being more vulnerable than men to climate-change-induced food insecurity and related risks (high confidence) (Cross-Chapter Box GENDER in Chapter 18). Women and men often have differential and unequal control over different productive assets and the benefits they provide, such as income from livestock (Ngigi et al., 2017; Musinguzi et al., 2018). Indigenous livestock keepers can be more vulnerable to climate change, partly due to ongoing processes of land fragmentation (Hobbs et al., 2008), historical land dispossession, discrimination and colonialisation, creating greater levels of poverty and marginalisation (Stephen, 2018). Adaptation actions may also be affected by gender and other social inequities (Balehey et al., 2018; Dressler et al., 2019). Men and women heads of household may access institutional support for adaptation in different ways (Assan et al., 2018). Further research is warranted to evaluate alternative gendered and equity-based approaches that can address differences in adaptive capacity within communities.

Uncertainties persist regarding estimates of net primary productivity (NPP) in grazing lands (Fetzel et al., 2017; Chen et al., 2018b), so estimation of climate change impacts on grasslands is challenging. Mean global annual NPP is projected to decline 10 gC m −2 yr −1 in 2050 under RCP8.5, although herbaceous NPP is projected to increase slightly (Boone et al., 2018; see Figure 5.11). Similar estimates were made by Havlik et al. (2014): large increases in projected NPP in higher northern latitudes (21% increase in the USA and Canada) and large declines in western Africa (−46%) and Australia (−17%). The cumulative effects of impacts on forage productivity globally are projected to result in 7–10% declines in livestock numbers by 2050 for warming of ~2°C, representing a loss of livestock assets ranging from USD 10 to 13 billion (Boone et al., 2018). Changes to African grassland productivity will have substantial, negative impacts on the livelihoods of >180 million people.

Increases in above-ground NPP, and woody cover at the expense of grassland, are projected in some of the tropical and subtropical drylands (Doherty et al., 2010; Ravi et al., 2010; Saki et al., 2018), in Mediterranean wood pastures (Rolo and Moreno, 2019) and in the northern Great Plains of North America (Klemm et al., 2020). Godde et al. (2021) projected that woody encroachment would occur on 51% of global rangeland area by 2050 under RCP8.5. The future makeup of grasslands under climate change is uncertain, given the variation in responses of the component species, though this variation may provide a climate buffer (Jones, 2019) (low confidence). C4 grass species are regarded as less responsive to elevated carbon dioxide than C3 species, though this is not always the case (Reich et al., 2018).

There are other interactions between climate change and grazing effects on grasslands. Li (2018a) reported strong negative responses of NPP and species richness to 4°C warming, a 50% precipitation decrease, and high grazing intensity. Changes in grassland composition will inevitably change their suitability for different grazing animal species, with switches from herbaceous grazers such as cattle to goats and camels to take advantage of increases in shrubland (Kagunyu and Wanjohi, 2014). Rangeland feed quality may also be reduced via invasive species of lower quality than native species (Blumenthal et al., 2016).

Warming and water deficits impair the quality and digestibility of a C4 tropical forage grass, Panicum maximum, because of increases in leaf lignin (Habermann et al., 2019). A metanalysis by Dellar (2018) of climate change impacts on European pasture yield and quality found an increase in above-ground dry weight under increased CO2 concentrations for forbs, legumes, graminoids and shrubs with reductions in N concentrations in all plant functional groups. Temperature increases will increase yields in alpine and northern areas (+82.6%) but reduce N concentrations for shrubs (−13.6%) and forbs (−18.5%).

Increased temperatures and CO2 concentrations may increase herbaceous growth and favour legumes over grasses in mixed pastures (He et al., 2019). These effects may be modified by changes in rainfall patterns, plant competition, perennial growth habits and plant–animal interactions. The cumulative effect of these factors is uncertain. Large, persistent declines in forage quality are projected, irrespective of warming, under elevated CO2 conditions (600 ppm and +1.5°C day/3°C night temperature increases) in North American grasslands (Augustine et al., 2018). Rising CO2 concentrations may result in losses of iron, zinc and protein in plants by up to 8% by 2050 (Smith and Myers, 2018). Little information is available on possible impacts on carbon-based micronutrients, such as vitamins. About 57% of grasses globally are C3 plants and thus susceptible to CO2 effects on their nutritional quality (Osborne et al., 2014). These impacts will result in greater nutritional stress in grazing animals as well as reduced meat and milk production (quality and quantity) (high confidence, medium evidence).

Recent research confirms the seriousness of the heat stress issue (medium evidence, high agreement ). Considerable increases are projected during this century in the number of ‘extreme stress’ days per year for cattle, chicken, goat, pig and sheep populations with SSP5-8.5 but many fewer with SSP1-2.6 (Thornton et al., 2021: Figure 5.12; see Cross-Chapter Box MOVING PLATE in this chapter). Resulting impacts on livestock production and productivity may be large, particularly for cattle throughout the tropics and subtropics and for goats in parts of Latin America and much of Africa and Asia. Pigs are projected to be particularly affected in the mid-latitudes of Europe, East Asia and North America. Lallo et al. (2018) estimated that global warming of 1.5°C and 2°C may exceed limits for normal thermo-regulation of livestock animals and result in persistent heat stress for animals in the Caribbean. Breed differences in heat stress resistance in dairy animals are now being quantified (Gantner et al., 2017), as are effects on sow reproductive performance in temperate climates (Wegner et al., 2016). Estimates of losses in milk production due to heat stress in parts of the USA, UK and West Africa to the end of the century range from 1% to 17% (Hristov et al., 2018; Fodor et al., 2018; Wreford and Topp, 2020; Rahimi et al., 2020). Much larger losses in dairy and beef production due to heat stress are projected for many parts of the tropics and subtropics: these could amount to USD 9 billion per year for dairy and USD 31 billion per for beef to end-century under SSP5-8.5, approximately 5% and 14% of the global value of production of these commodities in constant 2005 dollars.

In many LMICs, poultry contribute significantly to rural livelihoods, including via modest improvements in nutritional outcomes of household children (de Bruyn et al., 2018). Rural poultry are generally assumed to be hardy and well adapted to stressful environments, but little information exists regarding their performance under warmer climates or interactions with other production challenges (Nyoni et al., 2019).

The impacts of climate change on livestock diseases remain highly uncertain (medium evidence, high agreement ). Bett et al. (2017) showed positive associations between rising temperature and expansion of the geographical ranges of arthropod vectors such asCulicoides imicola, which transmits the bluetongue virus. A 1-in-20-year bluetongue outbreak at present-day temperatures is projected to increase in frequency to 1-in-5 to 1-in-7 years by the 2050s, under RCP4.5 and RCP8.5, although animal movement restrictions can prevent devastating outbreaks (Jones et al., 2019).

The prevalence and occurrence of some livestock diseases are positively associated with extreme weather events (high confidence). There are high risks of future Rift Valley fever (RVF) outbreaks under both RCP4.5 and RCP8.5 this century in East Africa and beyond (Taylor et al., 2016; Mweya et al., 2017).

Few studies explicitly consider the biotic and abiotic factors that interact additively, multiplicatively or antagonistically to influence host–pathogen dynamics (Cable et al., 2017). Integrative concepts that aim to improve the health of people, animals and the environment such as One Health may offer a framework for enhancing understanding of these complex interactions (Zinsstag et al., 2018). Much remains unknown concerning disease transmission dynamics under a warming climate (Heffernan, 2018), highlighting the need for effective monitoring of livestock disease (Brito et al., 2017; Hristov et al., 2018).

Water resources for livestock may decrease in places because of increased runoff and reduced groundwater resources, as well as decreased groundwater availability in some environments (AR5). Increased temperatures will cause changes in river flow and the amount of water stored in basins, potentially leading to increased water stress in dry areas such as parts of the Volta River Basin (Mul et al., 2015). Toure (2017) estimated decreases in groundwater recharge rates of 49% and of stored groundwater by 24% to the 2030s in the Klela Basin in Mali under both RCP4.5 and RCP8.5, with potentially serious consequences for water availability for livestock and irrigation.

Water intake by livestock is related to species, breed, animal size, age, diet, animal activity, temperature and physiological status of animals (Henry et al., 2018). Direct water use by cattle may increase by 13% for a temperature increase of 2.7°C in a subtropical region (Harle et al., 2007). Changes in water availability may arise because of decreased supply or increased competition from other sectors. Availability changes may be accompanied by shifts in water quality, such as increased levels of microorganisms and algae, that can negatively affect livestock health (Naqvi et al., 2015). In arid lands, projected decreases in water availability will severely compromise reproductive performance and productivity in sheep (Naqvi et al., 2017). In higher-input livestock systems, water costs may increase substantially owing to increased competition for water (Rivera-Ferre et al., 2016).

Information on future climate variability changes on livestock system productivity does not exist yet. Increases in climate variability may increase food insecurity in the future, mediated through increased crop and livestock production variability (Thornton and Herrero, 2014) in LMICs. Rainfall variability increases in pastoral lands have been linked to declining cattle numbers (Megersa et al., 2014). Changes in future climate variability may have large negative impacts on livestock system outcomes (Sloat et al., 2018; Stanimirova et al., 2019); these effects can be larger than those associated with gradual climate change (limited evidence, medium agreement ) (Godde et al., 2019). In grasslands, Chang et al. (2017) (Europe) and Godde et al. (2020) (globally) projected increases in biomass inter-annual variability, the worst effects occurring in rangeland communities that are already vulnerable. Ways in which climate variability impacts have been addressed in the past, such as via herd mobility, may become increasingly unviable in the future (Hobbs et al., 2008).

Livestock play important social (Kitalyi et al., 2005) and cultural (Gandini and Villa, 2003) roles in many societies. Climate change will negatively affect the provisioning of social benefits in many of the world’s grasslands (medium confidence). Examples include moving to semi-private land ownership models, driven in part by climate change, that are changing social networks and limiting socio-ecological resilience in pastoral systems in East Africa (Kibet et al., 2016; Bruyere et al., 2018) and Asia (Cao et al., 2018a); altering traditional food, resource and medicine sharing mechanisms in West Africa (Boafo et al., 2016); and the limited ability of current livestock systems to satisfy societies’ demand for CES in Northwest Europe (Bengtsson et al., 2019). The societal impacts of climate change on livestock systems may interact with drivers of change and increase herders’ vulnerability via processes of sedentarisation and land fragmentation, both of which may result in decreased animal access to rangelands (Adhikari et al., 2015; Cross-Chapter Box MOVING PLATE this chapter). Stronger linkages are needed between ecosystem service and food security research and policy to address these challenges (Gentle and Thwaites, 2016; Bengtsson et al., 2019).

Indigenous knowledge has a role to play in helping livestock keepers adapt (medium confidence), though the transferability of this knowledge is often unclear. Pastoralists’ local knowledge of climate and ecological change can complement scientific research (Klein et al., 2014), and local knowledge can be mobilised to inform adaptation decision making (Klenk et al., 2017). While Indigenous weather forecasting systems among pastoralists in Ethiopia (Balehegn et al., 2019; Iticha and Husen, 2019) and Uganda (Nkuba et al., 2020) are effective, synergies can be gained by combining traditional and modern knowledge to help pastoralists adapt. Sophisticated knowledge of feed resources among agro-pastoralists in West Africa is being used to increase system resilience (Naah and Braun, 2019). Understanding local knowledge for adaptation can present research challenges, for which new multi-disciplinary research methods may be needed (Reyes-Garcia et al., 2016; Roncoli et al., 2016). In particular, the complexities of knowledge, practice, power, local governance and politics need to be addressed (Hopping et al., 2016; Scoville-Simonds et al., 2020).

A wide range of measures exist to adapt sustainably managed forests of the boreal and temperate zone to climate change (Kolström et al., 2011; Gauthier et al., 2014; Keenan, 2015). Evidence emerging since the last assessment report further bolstered the notion that adapting the tree species composition to more warm-tolerant and less disturbance-prone species can significantly mitigate climate change impacts (high confidence) (Duveneck and Scheller, 2015; Seidl et al., 2018). Assisting the establishment of species in suitable habitats is one option to achieve climate-adapted tree species compositions (Benito-Garzón and Fernández-Manjarrés, 2015; Iverson et al., 2019). Furthermore, increasing the diversity of tree species within stands can have positive effects on tree growth and reduce disturbance impacts (high confidence) (Neuner et al., 2015; Jactel et al., 2018; Ammer, 2019). Some studies also suggest a positive effect of increased structural diversity, such as on forest resilience (moderate confidence) (Lafond et al., 2013; Koontz et al., 2020). Managing for continuous forest cover can also help to maintain the forest microclimate and buffer tree regeneration and the forest floor community against climate change (high confidence) (De Frenne et al., 2013; Zellweger et al., 2020). Reducing stocking levels, such as through thinning, has been found to effectively mitigate drought stress (Gebhardt et al., 2014; Elkin et al., 2015; Bottero et al., 2017), yet effects vary with species and ecological context (robust evidence, medium agreement ) (Sohn et al., 2016; Castagneri et al., 2021). Also shortened rotation periods have been suggested in response to climate-induced increases in growth and disturbance (Jönsson et al., 2015; Schelhaas et al., 2015). However, recent evidence suggests that these measures diminish in efficiency under climate change and can have corollary effects on other important forest functions such as carbon storage and habitat quality (medium confidence) (Zimová et al., 2020). Also, measures targeting landscape structure and composition have proven effective for increasing the climate resilience of forest systems (medium confidence) (Aquilue et al., 2020; Honkaniemi et al., 2020). While an increasing number of adaptation measures exist for sustainably managed forests, many studies highlight that the lead times for adaptation in forestry are long and that some vulnerabilities might remain also after adaptation measures have been implemented. Furthermore, the costs and benefits of adaptation measures relative to other goals of sustainable forest management, such as the conservation of biological diversity, have to be considered (Felton et al., 2016; Zimová et al., 2020; see Cross-Chapter Paper 7.5 Adaptation Response Options).

Reducing Deforestation and Forest Degradation plus (REDD+) is a climate mitigation strategy which could also provide important climate change adaptation co-benefits; for example, sustainable forest management could provide long term livelihoods to local communities and enhance resilience to climate risks (Turnhout et al., 2017). However, major challenges related to REDD+ implementation and forest use remain such that it has not been implemented successfully at scale (Table 5.8).

Table 5.8 | Challenges and solutions for REDD+

Research is limited on the effects of climate change on the distribution, productivity or availability of medicinal plants (Applequist et al., 2020), but some are facing threats due to climate change (Phanxay et al., 2015; Chirwa et al., 2017; Chitale et al., 2018). Climate change is projected to impact some medicinal plant species through changes in temperature, precipitation, pests and pathogens; unsustainable harvest of high-value species will significantly exacerbate these impacts (medium evidence, high agreement ) (Applequist et al., 2020). Table 5.9 highlights that climate change impacts on medicinal plant species will vary greatly by species. Medicinal plants that grow in arid environments are also highly susceptible to climate-induced change (Applequist et al., 2020). Arctic medicinal species may also be particularly at risk due to climate change (Cavaliere, 2009).

Table 5.9 | Observed and predicted impacts of climate change on selected medicinal plant species.

Changes in range distribution will interact with detailed local knowledge and Indigenous knowledge needed to harvest and use medicinal plants. Northward range shifts, for example, may mean certain plants still exist, but not where they have traditionally been important as medicine, and possibly moving suitable ranges outside of areas where plants species have sufficient protection (Kaky and Gilbert, 2017). Climate-induced phenological changes are already observed as a threat to some species (Gaira et al., 2014; Maikhuri et al., 2018). Other major climate-induced impacts on medicinal plants will be via the phytochemical content and pharmacological properties of medical plants (Gairola et al., 2010; Das et al., 2016a). Experimental trials have shown that drought stresses increase phytochemical content, either by decreasing biomass or increasing metabolites production (high confidence) (Selmar and Kleinwachter, 2013; Al-Gabbiesh et al., 2015).

The importance of seafood in food security and nutrition is increasing, largely due to its contribution as high-quality food (high confidence) (Hicks et al., 2019), as seafood contains unique long-chain polyunsaturated fatty acids (LC-PUFAs) and highly bioavailable essential micronutrients—vitamins (A, B and D) and minerals (calcium, phosphorus, iodine, zinc, iron and selenium). These compounds, often not readily available elsewhere in diets, have beneficial effects for adult health and child cognitive development (HLPE, 2014). Changes in marine and freshwater fish production can have significant consequences for human nutrition (Colombo et al., 2020). These changes are of particular concern in regions with few nutrition alternatives, such as low-income countries in Africa, Asia, Australasia, and Central and South America (high confidence) (Ding et al., 2017; Kibria et al., 2017).

Freshwater ecosystems that support most inland fisheries are under continuing threat from changes in land use, water availability and pollution and other pressures that will be exacerbated by climate change (high confidence) (Section 4.3.5). Declines in dissolved oxygen in freshwater are 2.75–9.3 times greater than observed in the world’s oceans (Jane et al., 2021). These systems have a relatively low buffering capacity and are therefore more sensitive to climate-related shocks and variability (Harrod et al., 2018b). Freshwater faunae are projected to be highly vulnerable; in the tropics because organisms are closer to approaching their thermal physiological limits and in the northern hemisphere (30–50°N) because the rate of temperature change is faster (Comte and Olden, 2017). The worldwide spatial confluence of productive freshwater fisheries and low food security highlights the critical role of rivers and lakes in providing locally sourced, low-cost, nutritious food sources (McIntyre et al., 2016).

Deltas and other wetland fisheries are extremely vulnerable to climate change and home to a large and growing proportion of the world’s population. In India, Ghana and Bangladesh, where three of the most populated Deltaic systems are located, subsistence fisheries provide 12–60% of the animal protein in people’s diets (Lauria et al., 2018).

The concern over aquatic food products’ safety due to climate change is increasing (high confidence). A strong positive relationship exists between specific bacterial growth rates and temperature, including pathogenic species of the genera Vibrio, Listeria, Clostridium, Aeromonas, Salmonella, Escherichia and others, whose distributional area is expanding with changing climate conditions (Cross-Chapter Box ILLNESS in Chapter 2, Section 5.12.1).

There is high confidence that climate change is and will continue to be a threat to the livelihood of millions of fishers, with the most vulnerable being those with fewer opportunities and less income (Barange and Cochrane, 2018; Section 3.4.3). The social vulnerability can differ largely between locations, even between relatively close coastal or inland communities (Bennett et al., 2014; Maina et al., 2016; Ndhlovu et al., 2017; Martins et al., 2019) and among inhabitants within a location, depending on factors such as access to other economic activities, education, health, adults in the household, and political connections (high confidence) (Senapati and Gupta, 2017; Abu Samah et al., 2019; Lowe et al., 2019).

Indigenous coastal communities consume 1.5–2.8 million metric tonnes of fish per year (about 2% of global yearly commercial marine catch), and reach a per capita consumption estimated to be 15 times greater than that of non-Indigenous country populations (Cisneros-Montemayor et al., 2016). There is high confidence that some Indigenous fishing communities are particularly vulnerable to climate change through a reduced capacity to conduct traditional harvests because of limited access to, or availability of, fish resources (Weatherdon et al., 2016), with consequences that include dietary shifts with significant nutritional and health implications (Marushka et al., 2019), displacement and loss of cultural identity (Sullivan and Rosenberg, 2018) and loss of social, economic and cultural rights (Finkbeiner et al., 2018). Areas of high risk for Indigenous Peoples include the Arctic, coastal communities with a high dependency on marine and freshwater fisheries, and Small Island States and Territories (Finkbeiner et al., 2018; Hanich et al., 2018, Section CCP6.2.5.1).

Women play a crucial role along the entire fisheries value chain, providing labour force in industrialised and small-scale fisheries all around the world (FAO, 2020d). For small-scale fisheries alone, women represent about 11% of the labour force, and their activity is generally in subsistence fisheries, highlighting their role in household food security (Harper et al., 2020). In general, gendered division of labour tends to cause lower salaries for women and different perception and experience of risk to climate change impacts (high confidence) (Lokuge and Hilhorst, 2017).

Local, national, regional and international fisheries are mostly underprepared for geographic shifts in marine animals driven by climate change over the coming decades (high confidence) (Pinsky et al., 2018; Oremus et al., 2020; Pinsky et al., 2020). With fisheries distribution changes, sometimes into areas dedicated to different historical uses or new ventures, the current management regimes will face constraining legal frameworks (Farady and Bigford, 2019; Pinsky et al., 2020), which will demand interventions in the form of policies, programmes and actions, at multiple scales (Cross-Chapter Box MOVING PLATE this chapter). Coordinated fisheries management can substantially expand capacity to respond to a changing climate (Pinsky et al., 2020), but a great deal of political will, capacity building and collective action will be necessary (high confidence) (Teslić et al., 2017; Burden and Fujita, 2019; Section 5.8.4).

Today, approximately half the world’s population (~4 billion out of 7.8 billion people) are assessed as being currently subject to severe water scarcity for at least 1 month per year (medium confidence) (Box 4.1), and freshwater inland fisheries are particularly vulnerable as they are given lower priority for water resources than other sectors (high confidence). In some cases, this situation results in the total loss of freshwater fisheries. Examples include diversion of water for agriculture, shifts from food provision to recreational fisheries, conserving biodiversity, and the requirement for high-quality water for drinking water supply (Section 5.13, Harrod et al., 2018a).

There is high confidence that climate change increases the risk of conflicts due to the redistribution of stocks and their abundance fluctuations, with subsequent impacts on resource sharing (Spijkers and Boonstra, 2017; Pinsky et al., 2018; Spijkers et al., 2018; Mendenhall et al., 2020; Pinsky et al., 2020). High vulnerability and lack of adaptive capacity to climate change impacts (including fisheries-dependent livelihoods, attachment to place, and pre-existing tensions) increase the risk of conflicts, including among fishery area users and authorities (Ndhlovu et al., 2017; Shaffril et al., 2017; Spijkers and Boonstra, 2017; Mendenhall et al., 2020). Similarly, shifts in the distribution of transboundary fish stocks under climate change alter the current sharing of resources between countries and create conflicts as well as new opportunities (Cross-Chapter Box MOVING PLATE this chapter, Spijkers and Boonstra, 2017; Pinsky et al., 2018).

There are regional differences in women’s roles, responsibilities and involvement in adaptation strategies in the aquaculture sector. Women comprise 14% of the 2018 global aquaculture workforce of 20.5 million (FAO, 2020c), representing up to 42% of the salmon workforce in Chile (Chávez et al., 2019), predominantly in processing roles (Gopal et al., 2020). In the majority of lower-middle-income countries, seaweed culture is dominated by women in family-owned businesses as in Zanzibar and the Philippines (Brugere et al., 2020; Ramirez et al., 2020), where women are not always paid directly but contribute to family incomes (high confidence) (Msuya and Hurtado, 2017; Brugere et al., 2020; Ramirez et al., 2020). In India, women collect stocking juveniles and assist in pond construction; in Bangladesh, women do the same tasks as men; and in Ghana, women undertake post-harvest fishing activities (Lauria et al., 2018). Women employed in aquaculture cooperatives gained adaptive capacity, which reduced gender inequities (medium confidence) (Farquhar et al., 2018; Gonzal et al., 2019), but lack of financial access for women can create gender inequity at larger commercial scales (Gurung et al., 2016; Call and Sellers, 2019). Women in aquaculture experience competing roles between employment, childcare and home duties (high confidence) (Morgan et al., 2015; Lauria et al., 2018; Chávez et al., 2019; see Cross-Chapter Box GENDER in Chapter 18) and differ from men in terms of perceptions of environmental risk, climate change and adaptation behaviour, with limited contributions to decision making (medium confidence) (Barange and Cochrane, 2018). Therefore, effective climate aquaculture adaptation options need to address gender inequity, such as suitable technology designs that fit with social norms and access to credit to facilitate independent uptake (medium evidence, high agreement ) (Morgan et al., 2015; Oppenheimer et al., 2019). Generalised best practices for gender-sensitive approaches to adaptation are relevant for aquaculture (UNFCCC, 2013).

Predicted sea level and temperature rise will result in coastal inundation into brackish and inland aquaculture systems (high confidence) (Mehvar et al., 2019; Nhung et al., 2019; Oppenheimer et al., 2019; Fox-Kemper et al., 2021), with negative impacts on aquaculture production in Viet Nam, East Africa and Jamaica (medium confidence) (Lebel et al., 2018; Nguyen et al., 2018; Bornemann et al., 2019). Precipitation and temperature changes will cause drought and flooding, negatively affecting near-shore fishpond productivity (limited evidence) (Canevari-Luzardo et al., 2019), but provide competitive advantages to non-native shrimp in Australia (limited evidence) (Cerato et al., 2019). Warming and acidification will increase HAB toxicity in freshwater systems, but responses may be strain-specific (Griffith and Gobler, 2020; Hennon and Dyhrman, 2020). As for molluscs in marine systems, projected climate change in freshwater and brackish systems may limit the availability of wild-sourced juveniles from fisheries (Beveridge et al., 2018). Projected impact studies for the inland and small-scale aquatic sectors are very limited (Halpern et al., 2019; Galappaththi et al., 2020b); therefore, this is a noted knowledge gap.

Aquaculture is often viewed as an adaptation option for fisheries declines, thereby alleviating food security from losses of other climate change impacts (Sowman and Raemaekers, 2018; Johnson et al., 2020) such as Pacific Islands freshwater aquaculture, Bangladesh crop-aquaculture systems or Viet Nam rice–fish cultivations (Soto et al., 2018). Many adaptations are specific to regions, countries or sectors, implemented on a regional to national scale (FAO, 2018c; Galappaththi et al., 2020b). Adaptation likelihood (potential), effectiveness and risk of maladaptation was assessed per major FAO production region for inland, brackish and marine aquaculture (Figure 5.16) production systems. Potential adaptation measures to reduce production loss can be built upon existing adaptation planning and guidelines, to reduce the risk of maladaptation including feedback loops (e.g., FAO, 2015; Bueno and Soto, 2017; Dabbadie et al., 2018; FAO, 2018c; Poulain et al., 2018; Brugère et al., 2019; Pham et al., 2021; Soto et al., 2021). Large climate change adaptation strategies for the aquaculture sector exist, such as in the USA (Link et al., 2015), Australia (Hobday et al., 2017) and South Africa (Department of Environmental Affairs, 2016). Lower-income countries often lack financial, technical or institutional capacity for adaptation planning (Galappaththi et al., 2020b), but examples include Bangladesh and Myanmar (FAO, 2018c), with programmes offering adaptation funding (Dabbadie et al., 2018). Early participation of stakeholders in adaptive planning has promoted action and ownership of results (high confidence), such as in India and the USA (Link et al., 2015; FAO, 2018c; Soto et al., 2018) Early outreach, education and knowledge gap assessments raise awareness, where utilisation of local knowledge and Indigenous knowledge and scientific involvement support informed adaptive planning and uptake for all stakeholders (high confidence) (Cooley et al., 2016; FAO, 2018c; Rybråten et al., 2018; Soto et al., 2018; McDonald et al., 2019; Galappaththi et al., 2020b), as perceptions of climate risk and capacity will vary (Tiller and Richards, 2018). Supporting the active involvement of women helps address gender inequity and perceived risk, particularly for smallholder farmers (high confidence) (Morgan et al., 2015; Barange and Cochrane, 2018; FAO, 2018c; Avila-Forcada et al., 2020). However, regional and national political influences, financial and technical capacity, governance planning and policy development will ultimately support or hinder adaptation for aquaculture (high confidence) (Cooley et al., 2016; FAO, 2018c; Galappaththi et al., 2020b; Greenhill et al., 2020).

Adaptation options at the operational level include species selections, such as cultivation of brackish species (shrimp, crabs) during dry seasons, and rice-finfish in wetter seasons in Thailand (Chiayarak et al., 2019), use of salt-tolerant plants in Viet Nam (Nhung et al., 2019; Paik et al., 2020), converting inundated rice paddies into aquaculture, rotating shrimp, and rice culture (high confidence) (Chiayarak et al., 2019). Species diversification through co-culture, integrated aquaculture–agriculture (e.g., rice–fish) or integrated multi-trophic culture (e.g., shrimp–tilapia–seaweed or finfish–bivalve–seaweed) may maintain farm long-term performance and viability by: creating new aquaculture opportunities; promoting societal and environmental stability; reducing GHG emissions through reduced feed usage and waste; and carbon sequestration (medium confidence) (see Section 5.10, Ahmed et al., 2017; Bunting et al., 2017; Gasco et al., 2018, Soto et al., 2018; Ahmed et al., 2019; Dubois et al., 2019; FAO, 2019c; Li et al., 2019; Freed et al., 2020; Galappaththi et al., 2020b; Prasko et al., 2020; Tran et al., 2020). In practice, most aquaculture operations concentrate on single-species systems (Metian et al., 2020), and barriers such as land availability, freshwater resources and lack of credit access may limit the uptake and success of integrated adaptation approaches to climate change (Ahmed et al., 2019; Tran et al., 2020; Kais and Islam, 2021).

Selective breeding can promote climate resilience (medium confidence) (Klinger et al., 2017; Fitzer et al., 2019), and operations have already intentionally, or unintentionally, selected for production traits for changing conditions (de Melo et al., 2016; Tan and Zheng, 2020). Exposure of broodstock to future climate conditions may or may not confer advantages to offspring (moderate evidence, low agreement ) (Parker et al., 2015; Griffith and Gobler, 2017; Thomsen et al., 2017; Durland et al., 2019). Traditional pedigree developments require extensive phenotypic data, but genomic selections can rapidly select for robust climate-associated traits (Sae-Lim et al., 2017; Gutierrez et al., 2018; Zenger et al., 2018; Houston et al., 2020; Tan and Zheng, 2020). Genomic resources are available for salmon, rainbow trout, coho, carp, tilapia, seabass, bream, turbot, flounder, catfish, yellow drum, scallops, oysters and shrimp, but have been developed for disease and growth selections rather than climate resistance (Dégremont et al., 2015a; Dégremont et al., 2015b; Abdelrahman et al., 2017; Gjedrem and Rye, 2018; Gutierrez et al., 2018; Guo et al., 2018; Liu et al., 2018a; FAO, 2019d; Houston et al., 2020), although bivalve selections for ocean acidification and warming resiliency are underway (Tan and Zheng, 2020). Targeted genome editing could modify phenotypes of major aquaculture species (Li et al., 2014 a; Elaswad et al., 2018; Yu et al., 2019; Houston et al., 2020), but uptake is dependent upon national regulatory and public approvals. Local adaptations within species with higher climate resiliencies may assist in selections (Thomsen et al., 2017; Falkenberg et al., 2019; Scanes et al., 2020; Toomey et al., 2020), but highlight the need to consider specific farming environments for selective processes (Houston et al., 2020). Projections of climate on aquaculture production traits are not well understood (Lhorente et al., 2019); therefore, genetic diversity needs to be maintained to ensure population fitness (high confidence) (Bitter et al., 2019; Lhorente et al., 2019; Visch et al., 2019; Houston et al., 2020; Mantri et al., 2020).

Land-based aquaculture systems including hatcheries may reduce exposure to climatic extremes (due to better control of the culture environment), limit water usage, reduce juvenile reliance and buffer climate effects using optimal diets (high confidence) (Barton et al., 2015; Reid et al., 2019; Cominassi et al., 2020). However, land-based aquaculture requires large capital and operational costs and use of land, increasing conflicts between land and water use, have increased energy demands (increasing GHG if fossil fuels are the primary energy source), require necessary expertise and will not reduce outgrowing exposures (high confidence) (see Section 5.13, Beveridge et al., 2018b; Soto et al., 2018; Tillotson et al., 2019; Costello et al., 2020; Prakoso et al., 2020).

Geographical selection of marine farm sites may prevent climate productivity declines (medium confidence) (Froehlich et al., 2018a; Sainz et al., 2019; Oyinlola et al., 2020), particularly for temperature-related mortality hotspots (Garrabou et al., 2019), HAB occurrences (Dabbadie et al., 2018) or extreme events (Liu et al., 2020; Wu et al., 2020). However, while downscaled climate forecasts facilitate localised adaptation planning (Falconer et al., 2020a), such projections are rare (Whitney et al., 2020). GIS can be used for climate adaptive planning along with routine site assessments (Falconer et al., 2020b; Galappaththi et al., 2020b; Jayanthi et al., 2020). Building coastal protection, stronger cages and mooring systems, and deeper ponds and using sheltered bays can reduce escapees and mortalities related to flooding, increased storms and extreme events (medium confidence) (Dabbadie et al., 2018; Bricknell et al., 2021; Kais and Islam, 2021). Inshore aquaculture in low-lying areas prone to sea level salinity intrusion (e.g., Mekong delta and Viet Nam) have already implemented adaptation measures, such as conversion of land to mixed plant–animal systems (Nguyen et al., 2019a), conversion of freshwater ponds to brackish or saline aquaculture (Galappaththi et al., 2020b), building of dams and dykes (Renaud et al., 2015) and intensification of shrimp or fish pond culture to reduce water and land usage (Nguyen et al., 2019b; Johnson et al., 2020). Other adaptation options for limited water supply are government equitable water allocations and water storage (high confidence) (Bunting et al., 2017; Galappaththi et al., 2020b).

Feed formulations and improved feed conversion can reduce climate-associated stress for freshwater species, significantly reducing waste and increase sustainability (medium confidence) (FAO, 2018c; Gasco et al., 2018; Chen and Villoria, 2019). Projected decreases in fish meal and global targets of limiting warming to under 2°C may increase the ratio of plant-based diets but reduce fish nutritional content (see Sections 5.10 and 5.13, Hasan and Soto, 2017; Johnson et al., 2020). Companies provide insurance in major production areas, but aquaculture is considered high risk with large levels of small claims (Secretan et al., 2007). Insurance covers natural disasters and disease, helping to reduce and cope with climate-induced risk, enabling faster livelihood recoveries and preventing poverty (high agreement , limited evidence) (Xinhua et al., 2017; Kalikoski et al., 2018; Soto et al., 2018). For example, small-scale shrimp farmers were willing to pay higher premiums to manage risk, after participation in government pilot insurance schemes, ensuring greater pay-outs if a mortality event occurred (Nyguyen and Pongthanapanic, 2016; Pongthanapanic et al., 2019). Technological innovations are more widely implemented in larger operations, with Internet access promoting adoption at the farm site (Joffre et al., 2017; Salazar et al., 2018). Improved farm management is a key opportunity (high confidence) to reduce climate risks on aquaculture, where Best Management Practices can increase resiliency (Soto et al., 2018) and lower additional risk from non-climatic stressors (Gattuso et al., 2018; Smith and Bernard, 2020), and decision-tree frameworks can provide adaptation choices when events occur (Nguyen et al., 2016).

Globally, monitoring is increasing to fill scientific uncertainties (Goldsmith et al., 2019) but is not often at spatial scales which facilitate farm or regional adaptation management (Whitney et al., 2020) or data complexities prevent direct uptake by operators, resource managers and policymakers (medium confidence) (Soto et al., 2018; Gallo et al., 2019). Specialised industry portals (Pacific shellfish) and government-established monitoring programmes (Chilean salmon) and other observational networks (e.g., Global Ocean Acidification Observing Network (GOA-ON)) can provide real-time monitoring and early-warning event alerts and facilitate aquaculture decision making (medium confidence) ( Cross et al., 2019; Farcy et al., 2019; Soto et al., 2019; Tilbrook et al., 2019; Bresnahan et al., 2020; Peck et al., 2020). Seasonal forecasting, downscaled models and early-warning systems provide valuable regional or farm site risk information (Hobday et al., 2018; Galappaththi et al., 2020b; Whitney et al., 2020), but monitoring will need to be useful for farmers, involve farmers, and be accurate, timely, cost-effective, reviewed and maintained in order to ensure uptake (high confidence) (Soto et al., 2018). Early-warning systems for HABs enable rapid decision making and risk mitigation (medium confidence), such as ocean colour monitoring in South Africa (Smith and Bernard, 2020), where early harvesting and additional husbandry were used to minimise production and economic losses (Pitcher et al., 2019). New tools, strategies and observations are needed to predict HAB occurrences and range shifts with changing climate (high confidence) (Schaefer et al., 2019; Tester et al., 2020), as there is uncertainty on drivers of incidence and toxicity (Wells et al., 2020).

Overall, there is high confidence that farm strategies that integrate mixed crop–livestock systems can improve farm productivity and have positive sustainability outcomes (Havet et al., 2014; Thornton and Herrero, 2014; Herrero et al., 2015; Thornton and Herrero, 2015; HLPE, 2019). The scale of the improvement varies between regions and systems and is moderated by overall demand in specific food products and the policy context. Integrated crop–livestock systems present opportunities for the control of weeds, pests and diseases. They can also provide a range of environmental benefits, such as increased soil carbon and soil water retention, increased biodiversity and reduced need for inorganic fertilizers (Havet et al., 2014; Thornton and Herrero, 2014; Herrero et al., 2015; Thornton and Herrero, 2015; HLPE, 2019).

Research indicates that mixed crop–livestock systems are often more resilient to climate change (medium confidence). In the southern Afar region of Ethiopia, crop–livestock households were more resilient than livestock-only households to climate-induced shock (Mekuyie et al., 2018). However, the benefits of managing crop–livestock interactions in response to climate change depend on local context. For example, in higher-rainfall zones in Australia, Nie et al. (2016) found some yield reductions and difficulty in maintaining groundcover. The systematic review of Gil et al. (2017) concluded that the integration of crop and livestock enterprises as an adaptation measure can enhance resilience (FAQ 5.1).

Reconfiguration of mixed farming systems is occurring. In semi-arid eastern Senegal, Brottem and Brooks (2018) found increasing reliance on livestock production mostly because of changing climate conditions. Many poorer households are having to rely on migration to compensate for shortfalls in crop production arising from a changing climate. Some farmers have successfully shifted to crop–livestock systems in Australia, where they have allocated land and forage resources in response to climate and price trends (Bell et al., 2014).

Mixed livestock–crop systems may increase burdens on women, require managing competing uses of crop residues, and have higher requirements of capital and management skills. These factors can be challenging in many lower-income countries (Rufino et al., 2013; Thornton and Herrero, 2015; Jost et al., 2016; Thornton, 2018). The policy actions needed for the successful operation of mixed crop–livestock systems may be similar across widely different situations: good access to credit inputs and capacity building needed to facilitate uptake (Hassen et al., 2017; Marcos-Martinez et al., 2017), and good levels of market infrastructure (Ouédraogo et al., 2017; Iiyama et al., 2018).

Households may have a mix of aquatic and land-based food production, contributing to food security and nutrition and income generation (Freed et al., 2020; see also discussion of aquaponics and hydroponics in Section 5.10.4.3. and combined rice–aquatic species production in Section 5.9.4). Failures in agricultural outputs due to climate-associated factors may result in diversification to fisheries as a way of alleviating food production shortfalls; for example, fisheries landings may dramatically increase after agricultural failures following hurricanes, which can subsequently create overfishing collapses (Cottrell et al., 2019). Where climatic impact drivers affect multiple sectors, adaptation may become more difficult because of the interacting challenges (Cottrell et al., 2019). One study of 12 countries with high food insecurity levels found that fish-reliant households utilised as much land as those not reliant on fish (Fisher et al., 2017). To meet food security requirements, most of these households needed to both farm and fish, illustrating the interdependence of aquatic–terrestrial food systems.

Agroforestry is frequently mentioned as a strategy to adapt to and mitigate climate change and address food security (de Coninck et al., 2018; Smith et al., 2019c). There is strong evidence of net positive biophysical and socioeconomic effects of agroforestry systems under both smallholder and large-scale mechanised production systems (Quandt et al., 2017; Hoegh-Guldberg et al., 2018; Sida et al., 2018; Wood and Baudron, 2018; Table 5.10; Cross-Chapter Box NATURAL in Chapter 2; Quandt et al., 2019). Many of these effects also reduce climate risk. At the same time, agroforestry systems are subject to impacts from climate change, potentially reducing the benefits they provide. Still, there is limited evidence of observed climate impacts on agroforestry systems, and modelling climate impacts is more complex for agroforestry than for single cropping systems (Luedeling et al., 2014).

Important information gaps exist concerning the costs and benefits of many adaptation options in mixed systems, where the interactions between farming enterprises may be complex. Among communal crop–livestock farmers in Eastern Cape Province of South Africa, Bahta (2016) reported high levels of vulnerability to drought and highlighted the need for more coordination between monitoring agencies in terms of reliable early-warning information that can be communicated appropriately, between farmers’ organisations and the private sector to facilitate adaptation options that can overcome feed shortages such as fodder purchases in times of drought, and between government departments at the national and provincial level that address the concerns and needs of affected communities. Nyamushamba (2017) reviewed the use of indigenous beef cattle breeds in smallholder mixed production systems in southern Africa. Some of these breeds exhibit adaptive traits such as drought and heat tolerance and resistance to tick-borne diseases. However, their adaptation potential in crossbreeding programmes is essentially unknown, as most African cattle populations are still largely uncharacterised.

As in other production systems, Indigenous groups, gender, race and other social categories can result in heightened vulnerability to climate change in mixed production systems owing to historical and current marginalisation and discrimination (high confidence) (Parraguez-Vergara et al., 2016; Baptiste and Devonish, 2019; Moulton and Machado, 2019; Popke and Rhiney, 2019; Fagundes et al., 2020). A study of the Mapuche Indigenous group in Chile found that marginalisation and discrimination worsened their vulnerability and observed impacts of climate change because they had less access to services and lower incomes and were not as high a priority as other groups (Parraguez-Vergara et al., 2016). Among fisherfolk on Lake Wamala, Uganda, Musinguzi (2018) found evidence of considerable diversification to crop and livestock production as a means of increasing households’ food security and income, but women had greater workloads and less control over new income sources than men. Ngigi (2017) evaluated adaptation actions within households in rural Kenya and found that women tended to adopt adaptation strategies related to crops, and men to livestock and agroforestry activities. Chingala (2017) found substantial gender- and age-related differences in control of access to animal feed, animal health and water resources in beef producers in mixed crop–livestock systems in Malawi. In a review of agriculture–aquaculture systems in coastal Bangladesh, Hossain et al. (2018) showed that existing policies and adaptation mechanisms are not adequately addressing gender power imbalances, and women continue to be marginalised, leading to increasing feminisation of food insecurity. Such studies highlight the need to consider gender and other social inequities when examining adaptation in mixed production systems, particularly in situations in which men and women have different levels of control over productive assets (Cross-Chapter Box GENDER in Chapter 18).

There is medium confidence in the effectiveness of changing the nature of the integration between crops and livestock as an adaptation: moving from crops to livestock, moving from livestock to crops, and moving from one species of livestock to others, for example (Roy et al., 2018). Such transitions that increase integration between farm enterprises may contribute to risk reduction and increased food security. In areas with adequate rainfall and relatively limited rainfall variability under climate change, where agricultural diversity is the greatest, transitions towards more diverse and integrated systems may bring substantial adaptation benefits (Waha et al., 2018).

Barriers to increasing integration and diversification include policies which support cereals and crop specialisation, lack of markets, limited post-harvest processing, limited technical or biophysical research on implementation and poor market infrastructure (Keatinge et al., 2015; Bodin et al., 2016; Garibaldi et al., 2016; Bassett and Koné, 2017; Kongsager, 2017; Rhiney et al., 2018; Roesch-McNally et al., 2018; Clay and King, 2019; Ickowitz et al., 2019). Proactive policy and market development are needed to reduce these barriers (Clay and King, 2019; Ickowitz et al., 2019; See 5.14.3.8 for Insurance).

Agroforestry, the purposeful integration of trees or shrubs with crop or livestock systems, increases resilience against climate risks through a range of biophysical and economic effects (high confidence). Traditional agroforestry has been practiced for millennia and provides prime examples of sustainable agroecological production systems meeting the production, income and socio-cultural needs of farming communities within their ecological niches, but market forces have often led to their demise (McNeely and Schroth, 2006; Plieninger and Schaar, 2008; García-Martínez et al., 2016; Krčmářová and Jeleček, 2016; Coq-Huelva et al., 2017; Paudel et al., 2017; Doddabasawa et al., 2018; Maezumi et al., 2018; Lincoln, 2020). The wide range of options to associate different trees with crops, livestock and aquaculture allows agroforestry to be practiced in most regions, including those with precipitation regimes ranging from semi-arid to humid. While most agroforestry systems occur in smallholder settings, there are examples of successful industrial-scale mechanised agroforestry systems (Feliciano et al., 2018; Lovell et al., 2018). Agroforestry delivers medium to large benefits to all five land challenges described in the SRCCL—climate change mitigation, adaptation, desertification, land degradation and food security—and is considered to have broad adaptation and moderate mitigation potential compared with other land challenges (Smith et al., 2019c). Agroforestry is also able to deliver multiple biophysical and socioeconomic benefits (Table 5.12).

Table 5.12 | Some of the biophysical and socioeconomic benefits of agroforestry.

The adoption and maintenance of agroforestry practices require appropriate incentives or the removal of barriers (high confidence). Agroforestry adoption has been limited to date in both higher-income and lower-income countries. Several constraints need to be carefully addressed for successful scaling-up of agroforestry systems, including costs of establishment, limited short-term benefits, lack of reliable financial support to incentivise longer-term returns on investments, land tenure, knowledge of and experience with trees and the management of multiple component systems, and inadequate market access, (Coulibaly et al., 2017; Iiyama et al., 2017; Jacobi et al., 2017; Kongsager, 2017; Hernández-Morcillo et al., 2018; Iiyama et al., 2018; Lincoln, 2019). Kongsager (2017) and Roupsard et al. (2020) also highlight the need for vertical integration of measures from local to national scales to successfully address local barriers to adoption. Although there are few studies evaluating the long-term performance of agroforestry systems (Coe et al., 2014; Meijer et al., 2015; Brockington et al., 2016; Kongsager, 2017; Toth et al., 2017), the available results suggest that successful adoption of agroforestry practices depends strongly on the local enabling environment, including appropriate markets, technologies and delivery systems (medium evidence, high agreement ).

Hydroponic systems produce plants in a soilless environment requiring mineral fertilizers to meet plant nutritional needs, whereas aquaponics combines an aquaculture production system with hydroponics, where fish waste provides nitrogen, phosphorous and potassium for plant growth and nitrifying and mineralising bacteria act as filters (Goddek et al., 2015; Pérez-Urrestarazu et al., 2019; Ghamkhar et al., 2020). The relative environmental impact of hydroponic systems is lower compared with conventional systems owing to the significant reductions in land use and fertilizer usage (high confidence) (Goddek et al., 2015; Datta et al., 2018; Pantanella, 2018; Suhl et al., 2018; El-Essawy et al., 2019; Jaeger et al., 2019; Monsees et al., 2019; Mupambwa et al., 2019; Pérez-Urrestarazu et al., 2019; Ghamkhar et al., 2020). While studies indicate that aquaponics and hydroponics have higher yields and a lower environmental footprint than conventional agriculture (medium confidence), aquaculture and heated greenhouse production (Pantanella, 2018; Romeo et al., 2018), aquaponic production may need to be coupled or decoupled or have double-recirculation systems to meet the different requirements of farmed fish and crop species (Pantanella, 2018; Suhl et al., 2018; Mupambwa et al., 2019). Aquaponics and hydroponics are a promising adaptation option for urban agriculture, with benefits including a protected growing environment from climate extremes, reduced GHG emissions related to food transportation, reduced food waste, rainwater harvesting and use of food waste (medium agreement , limited evidence) (Goddek et al., 2015; Al-Kodmany, 2018; Clinton et al., 2018; Weidner and Yang, 2020). Such systems show promise for reducing food production environmental footprints and increasing food security, particularly in arid or water-stressed environments (Doyle et al., 2018; Mupambwa et al., 2019). Barriers to aquaponics and hydroponics adoption include market acceptance of cultured fish species and desirability of plant crops, lack of expertise, legal constraints or high investment costs and financial feasibility (Bosma et al., 2017; Al-Kodmany, 2018; Datta et al., 2018; Pantanella, 2018; El-Essawy et al., 2019; Martin and Molin, 2019; Pérez-Urrestarazu et al., 2019; Specht et al., 2019). There ishigh confidence (high agreement , medium evidence) that a major barrier to hydroponic and aquaponics adoption is the requirement for skilled operators (Goddek et al., 2015; Bosma et al., 2017; Datta et al., 2018; McHunu et al., 2018; Pantanella, 2018), which could be mitigated by decoupling systems and disciplines (Pantanella, 2018). As yet, these systems are not widely implemented and information on their climate change impacts is limited.

Transitions in and between the different elements of integrated agricultural systems can be an effective adaptation option (medium confidence). Havlik et al. (2014) projected that, by 2030, market-driven autonomous transitions towards more efficient production systems would increase ruminant meat and milk productivity by up to 20% and decrease emissions by 736 MtCO2 e y −1, most of this arising through avoided emissions from the conversion of 162 Mha of natural land. Weindl et al. (2015) assessed the implications of several climate projections on land use change to 2045 and found that shifts in livestock production towards mixed crop–livestock systems would represent a resource- and cost-efficient adaptation option, reducing global agricultural adaptation costs and abating deforestation by about 76 million ha globally. Both studies suggest that public policy support for transitioning livestock production systems to increase their efficiency could be an important lever for reducing adaptation costs and contributing to emissions reductions. This policy support could include modified regulatory and certification frameworks that incentivise livestock producers to adapt and mitigate (Weindl et al., 2015).

Recent reviews have summarised literature on production system transitions, driven at least partly by a changing climate or changing climate variability, that sometimes involves substantial shifts in enterprises and land configurations. These reviews found several cases of transitions affecting pastoral and mixed systems, with a range of responses including intensification, diversification and sedentarisation as well as the abandonment of agriculture (see Section 5.1 4.3.1, Vermeulen et al., 2018; Thornton et al., 2019). The consequences of these system transitions have been mixed; in some cases, the household-level outcomes have been beneficial, while in others not. Policy environments, defined in terms of multi-level governance structures and institutions, are critical enablers of change. The vulnerability of many crop–livestock keepers to climate change is particularly affected by property and grazing rights (high confidence). Identifying the winners and losers from changes in land ownership and the use of communal lands in the coming decades is a key challenge for the research agenda, particularly as climate change impacts in the marginal lands intensify (Reid et al., 2014).