Since AR5, several new international agreements have come into effect that have implications for international responses to the climate risks assessed in this chapter. The 2015 Paris Agreement, which explicitly mentions health in three separate sections, set new goals for adaptation and established a working group to study the effects of climate change on population displacement. The 17 United Nations (UN) SDGs for 2030, adopted in 2015, are all important for building adaptive capacity in general, with goals 13 (‘Take urgent action to combat climate change and its impacts’) and 3 (‘Ensure healthy lives and promote well-being for all at all ages’) being directly relevant to this chapter. Other SDGs contain specific targets that are also relevant for this chapter, including Target 10.7 (‘Well-managed migration policies’), Target 8.3 (‘Decent work for all’) and Target 5.4 (‘Promotion of peaceful and inclusive societies’) (Piper, 2017). The 2015 Sendai Framework for Disaster Risk Reduction puts an emphasis on health and well-being (Aitsi-Selmi and Murray, 2016) In 2018, UN member states negotiated Global Compacts for Safe, Orderly and Regular Migration and on Refugees that, taken together with the Paris Agreement, provide pathways for coordinated international responses to climate-related migration and displacement (Warner, 2018).

All three post-AR5 IPCC Special Reports considered some of the research that is assessed here in greater detail. The 2018 report on 1.5°C (SR1.5) included a review of climate change and health literature published since AR5 and called for further efforts for protecting health and well-being of vulnerable people and regions (Ebi et al., 2018b) and highlighted links between climate change hazards, poverty, food security, migration and conflict. The 2019 Special Report on Climate Change and Land (SRCCL) (IPCC, 2019b) emphasised the impacts of climate change on food security; highlighted links between reduced resilience of dryland populations, land degradation, migration and conflict; and raised concerns about the impacts of climate extremes. The 2019 Special Report on the Ocean and Cryosphere in a Changing Climate (IPCC, 2019a) detailed how changes in the cryosphere and ocean systems have impacted people and ecosystem services, particularly food security, water resources, water quality, livelihoods, health and well-being, infrastructure, transportation, tourism and recreation as well as the culture of human societies, particularly for Indigenous Peoples. It also noted the risks of future displacements due to rising sea levels and associated coastal hazards.

This chapter uses the conceptual framing described in Chapter 1, in which risks emerging from climate change are described in terms of hazard, exposure and vulnerability, with adaptation and climate resilient development being responses that have the potential to reduce or modify risk. The observed and projected future risks to health well-being, involuntary population displacements and conflict identified in this chapter are associated with a range of hazards that are manifested at a variety of geographical and temporal scales. These include observed and projected changes in climate normals; changes in the frequency, duration, and/or severity of extreme events; and hazards such as rising sea levels and extreme temperatures where the impacts have only begun to be widely experienced. The 2021 report of IPCC WGI (IPCC, 2021) provides an assessment of observed and projected changes in these hazards and is the backdrop against which assessments of future risks and adaptation options identified in the present chapter should be considered.

The exposure to hazards of populations, infrastructure, ecosystem capital, socioeconomic systems and cultural assets critical to health and well-being varies considerably across and within regions (high confidence). Exposure is also projected to vary across and within regions over time, depending on future GHG emissions pathways, development trajectories and differential vulnerability, particularly for exposure to extreme events and conditions, such as floods and droughts (Figures 7.1a, 7.1b) (Winsemius et al. 2018). For this reason, region-specific assessments of climate-related risks for health, displacement and conflict are found in each of the regional chapters of this report in addition to the general assessment that appears in this chapter.

Vulnerability to climate change varies across time and location, across communities, and among individuals within communities; it reflects variations and changes in macro-scale non-climatic factors (such as changes in population, economic development, education, infrastructure, behaviour, technology and ecosystems) and individual- or household-specific characteristics, such as age, socioeconomic status, access to livelihood assets, pre-existing health conditions and ability (US Global Change Research Program, 2016; Chapter 1).

Many direct and indirect effects of climate change pose multiple threats to human health and well-being and can occur simultaneously, resulting in compounding or cascading impacts for vulnerable populations. For example, many of the long-term impacts of climate change on NCDs and injury described in Sections 7.2 and 7.3 are associated with future increases in air temperature and levels of air pollution; in many regions, and especially in large urban centres in Asia and Africa, these particular hazards are already causing substantial increases in morbidity and mortality due to respiratory illnesses (Tong et al., 2016). Climate change can therefore be expected to magnify such health risks over the long term.

At the same time, urban populations will also be experiencing indirect risks through climate change impacts on food and potable water systems, variations in the distribution and seasonality of infectious diseases and growing demand for shelter due to increased in-migration. The accumulation of these risks over time can be expected to generate accelerating declines in community resilience and health, with future vulnerability potentially expanding in a nonlinear fashion (Dilling et al., 2017; Liang and Gong, 2017; El-Zein and Tonmoy, 2017; see also Chapter 6). Further, although each individual risk in isolation may be transitory or temporary for the individuals or groups exposed, taken cumulatively, the impacts could create conditions of chronic lack of well-being, and early life experiences with specific illnesses and conditions could have lifelong consequences (Watts et al., 2015; Otto et al., 2017; WHO, 2018a). In this context, there is a distinct need for greater longitudinal research on vulnerability to multiple climatic and non-climatic health and well-being hazards over time (Fawcett et al., 2017). There is also need for more research to identify critical thresholds in social vulnerability to climate change (Otto et al., 2017); these include rapid, stepwise changes in vulnerability that emerge from changes in exposure (for example, air temperatures above which mortality rates or impacts on pre-natal health accelerate (Arroyo et al., 2016; Ngo and Horton, 2016; Abiona, 2017; Auger et al., 2017; Molina and Saldarriaga, 2017; Zhang et al., 2017b)) and thresholds in adaptation processes (such as when rural out-migration rates grow due to climate-related crop failures (McLeman, 2017)).

In virtually all of the research identifying particular climate-related risks to health, well-being, migration and conflict, specific types of individuals are identified as having higher levels of vulnerability and exposure to climate-related health hazards: people who are impoverished, undernourished, struggle with chronic or repeated illnesses, live in insecure housing in polluted or heavily degraded environments, work in unsafe conditions, are disabled, have limited education and/or have poor access to health and social infrastructure (WHO, 2018a). Their disproportionate exposure to ongoing climate hazards and their inability to recover from extreme events increase not only their own vulnerability but also that of the wider communities in which they live (US Global Change Research Program, 2016). Highly vulnerable populations are not evenly distributed across regions (Figure 7.2) nor within countries. Yet, even those fortunate enough to live in better neighbourhoods with greater financial means, higher-paying jobs and good access to resources and services, may experience adverse climate-related outcomes through community-level interactions and linkages (Haines and Ebi, 2019). Increased inequity itself threatens well-being and an effective response to climate change should not only avoid increased inequity but identify ways in which to reduce existing inequity.

Climate-sensitive VBDs include mosquito-borne diseases, rodent-borne diseases and tick-borne diseases. Many infectious agents, vectors, non-human reservoir hosts, and pathogen replication rates can be sensitive to ambient climatic conditions. Elevated proliferation and reproduction rates at higher temperatures, longer transmission season, changes in ecology and climate-related migration of vectors, reservoir hosts or human populations contribute to this climate sensitivity (Rocklöv & Dubrow, 2020; Semenza and Paz, 2021). Age-standardised DALY rates for many VBDs have decreased over the last decade due to factors unrelated to climate. Vulnerability to VBD is strongly determined by sociodemographic factors (e.g., children, the elderly and pregnant women are at greater risk) with exposure to vectors being strongly influenced by various factors including socioeconomic status, housing quality, healthcare access, susceptibility, occupational setting, recreational activity, conflicts and displacement ( Rocklöv & Dubrow, 2020; Semenza and Paz, 2021). Figure 7.5 illustrates how climatic and non-climatic drivers and responses determine VBD outcomes.

Evidence has increased since AR5 that the vectorial capacity has increased for dengue fever, malaria and other mosquito-borne diseases and that higher global average temperatures are making wider geographic areas more suitable for transmission (very high confidence). Transmission rates of malaria are directly influenced by climatic and weather variables such as temperature, with non-climatic socioeconomic factors and health system responses counteracting the climatic drivers (very high confidence). The burden of malaria is greatest in Africa, where more than 90% of all malaria-related deaths occur (M’Bra et al., 2018; Caminade et al., 2019). Between 2007 and 2017, DALYs for malaria have decreased by 39% globally. Malaria is mainly caused by five distinct species of plasmodium parasite (Plasmodium falciparum, Plasmodium vivax, Plasmodium malariae, Plasmodium ovale and Plasmodium knowlesi) and is transmitted by Anopheline mosquitoes. Evidence suggests that in highland areas of Colombia and Ethiopia, malaria has shifted in warmer years towards higher altitudes, indicating that, without intervention, malaria will increase at higher elevations as the climate warms (Siraj et al., 2014; Midekisa et al., 2015). Each year, local outbreaks of malaria occur due to importation in areas from which it was once eradicated, such as Europe, but the risk of re-establishment is considered low.

The transmission of dengue fever is linked to climatic and weather variables such as temperature, relative humidity and rainfall (high confidence). The dengue virus is carried and spread by Aedes mosquitoes, primarily Aedes aegypti. Dengue has the second highest burden of VBDs, with the majority of deaths occurring in Asia (Bhatt et al., 2013). Since 1950, global dengue burden has grown and is attributable to a combination of climate-associated expansion in the geographic range of the vector species and non-climatic factors such as globalised air traffic, urbanisation and ineffective vector abatement measures. Temperature, relative humidity and rainfall variables are significantly and positively associated with increased dengue case incidence and/or transmission rates globally, including in Vietnam (Phung et al., 2015; Xuan le et al., 2014), Thailand (Xu et al., 2019a), India (Mutheneni et al., 2017; Rao et al., 2018; Mala and Jat, 2019), Indonesia (Kesetyaningsih et al., 2018), the Philippines (Carvajal et al., 2018), the USA (Lopez et al., 2018; Pena-Garcia et al., 2017; Duarte et al., 2019; Rivas et al., 2018; Silva et al., 2016a), Jordan (Obaidat and Roess, 2018) and Timor-Leste (Wangdi et al., 2018). Variation in winds, sea surface temperatures and rain over the tropical eastern Pacific Ocean (El Niño-Southern Oscillation; ENSO) have been linked to increased dengue incidence in Colombia (Quintero-Herrera et al., 2015; McGregor and Ebi, 2018; Pramanik et al., 2020) and its interannual variation successfully forecasted in Ecuador using ENSO indices as predictors (Petrova et al., 2019). The observed time lag between climate exposures and increased dengue incidence is approximately 1–2 months (Chuang et al., 2017; Lai, 2018; Chang et al., 2018).

Changing climatic patterns are facilitating the spread of CHIKV, Zika, Japanese encephalitis and Rift Valley Fever in Asia, Latin America, North America and Europe (high confidence). Climate change may have facilitated the emergence of CHIKV as a significant public health challenge in some Latin American and Caribbean countries (Yactayo et al., 2016; Pineda et al., 2016) and contributed to chikungunya outbreaks in Europe (Rocklöv et al., 2019; Mascarenhas et al., 2018; Morens and Fauci, 2014). The Zika virus outbreak in South America in 2016 was preceded by 2007 outbreaks on Pacific islands and followed a period of record high temperatures and severe drought conditions in 2015 (Paz and Semenza, 2016; Tesla et al., 2018). Increased use of household water storage containers during the drought is correlated with a range expansion of Aedes aegypti during this period, increasing household exposure to the vector (Paz and Semenza, 2016). Changing climate also appears to be a risk factor for the spread of Japanese encephalitis to higher altitudes in Nepal (Ghimire and Dhakal, 2015) and in southwest China (Zhao et al., 2014). In eastern Africa, climate change may be a risk factor in the spread of Rift Valley Fever (Taylor et al., 2016a).

Changes in temperature, precipitation, and relative humidity have been implicated as drivers of West Nile fever in southeastern Europe (medium confidence). The average temperature and precipitation prior to the exceptional 2018 West Nile outbreak in Europe was above the 1981–2010 period average, which may have contributed to an early upsurge of the vector population (Marini et al., 2020; Haussig et al., 2018; Semenza and Paz, 2021). In 2019 and 2020, West Nile fever was first detected in birds and subsequently in humans in Germany and the Netherlands (Ziegler et al., 2020; Vlaskamp et al., 2020).

Climate change has contributed to the spread of the Lyme disease vectorIxodes scapularis , a corresponding increase in cases of Lyme disease in North America (high confidence) and the spread of the Lyme disease and tick-borne encephalitis vectorIxodes ricinus in Europe (medium confidence). In Canada, there has been a geographic range expansion of the black-legged tick I. scapularis, the main vector of Borrelia burgdorferi, the agent of Lyme disease. Vector surveillance of I. scapularis has identified strong correlation between temperatures and the emergence of tick populations, their range and recent geographic spread, with recent climate warming coinciding with a rapid increase in human Lyme disease cases (Clow et al., 2017; Cheng et al., 2017; Gasmi et al., 2017; Ebi et al., 2017). Ixodes ricinus, the primary vector in Europe for both Lyme borreliosis and tick-borne encephalitis is sensitive to humidity and temperature (Daniel et al., 2018; Estrada-Peña and Fernández-Ruiz, 2020) (high confidence). There has been an observed range expansion to higher latitudes in Sweden and to higher elevations in Austria and the Czech Republic.

Rodent-borne disease outbreaks have been linked to weather and climate conditions in a small number of studies published since AR5, but more research is needed in this area . In Kenya, a positive association exists between precipitation patterns and Theileria-infected rodents, but for Anaplasma, Theileria and Hepatozoon, the association between rainfall and pathogen varies according to rural land use types (Young et al., 2017). Weather variability plays a significant role in transmission rates of haemorrhagic fever with renal syndrome (HFRS) (Hansen et al., 2015; Xiang et al., 2018; Liang et al., 2018; Fei et al., 2015; Xiao et al., 2014; Vratnica et al., 2017; Roda Gracia et al., 2015; Monchatre-Leroy et al., 2017; Bai et al., 2019). In Chongqing, HFRS incidence has been positively associated with rodent density and rainfall (Bai et al., 2015).

Important waterborne diseases (WBDs) include diarrhoeal diseases (such as cholera, shigella, cryptosporidiosis and typhoid), schistosomiasis, leptospirosis, hepatitis A and E and poliomyelitis (Cisse, 2019; Houéménou et al., 2021; Hassan et al., 2021; Archer et al., 2020; Mbereko et al., 2020; Fan et al., 2021). The number of cases of WBDs is considerable, and even in high-income countries WBDs continue to be a concern (Cissé et al., 2018; Kirtman et al., 2014; Levy et al., 2018; Murphy et al., 2014; Brubacher et al., 2020; Lee et al., 2021). Nevertheless, diarrhoea mortality has declined substantially since 1990, although there are variations by country, and the global burden of WBDs has decreased in line with vaccination coverage of some WBDs (such as polio and cholera), poverty reduction and improved sanitation and hygiene (Jacob and Kazaura, 2021; Mutono et al., 2020; Lee et al., 2019; Semenza and Paz, 2021; Jacob and Kazaura, 2021; Mutono et al., 2020).

Drinking water containing pathogenic microorganisms is the main driver of the burden of WBDs (Murphy et al., 2014; Lee et al., 2021; Chen et al., 2021b; Musacchio et al., 2021). WBD outbreaks, particularly intestinal diseases, are attributable to a combination of the presence of particular pathogens (bacteria, protozoa, viruses or parasites) and the characteristics of drinking water systems in a given location (Bless et al., 2016; Ligon and Bartram, 2016; Mutono et al., 2021; Ferreira et al., 2021).

Since AR5 there is a growing body of evidence that increases in temperature (very high confidence), heavy rainfall (high confidence), flooding (medium confidence) and drought (low confidence) are associated with an increase of diarrhoeal diseases. In the majority of studies there is a significant positive association observed between WBDs and elevated temperatures, especially in areas where water, sanitation and hygiene (WASH) deficiencies are significant (Levy et al., 2018; Carlton et al., 2016; Levy et al., 2018; Sherpa et al., 2014; Guzman Herrador et al., 2015; Levy et al., 2016; Lo Iacono et al., 2017). In Ethiopia, South Africa and Senegal, increases in temperatures are associated with increases in diarrhoea, while in Ethiopia, Senegal and Mozambique, increases in monthly rainfall are associated with an increase in cases of childhood diarrhoea (Azage et al., 2015; Thiam et al., 2017; Horn et al., 2018). Similar associations between weather and diarrhoea have been observed in Cambodia, China, Bangladesh, Pacific Island countries and the Philippines (McIver et al., 2016a; McIver et al., 2016b; Liu et al., 2018; Wu et al., 2014; Matsushita et al., 2018). Heavy precipitation events have been consistently associated with outbreaks of WBDs in Europe, USA, UK and Canada (Guzman Herrador et al., 2015; Levy et al., 2016; Lo Iacono et al., 2017; Curriero et al., 2001; Guzman Herrador et al., 2016; Levy et al., 2018; Semenza and Paz, 2021).

Impacts of floods include outbreaks of WBDs, with such events disproportionately affecting the young, elderly and immunocompromised (Suk et al., 2020; Guzman Herrador et al., 2015; Levy et al., 2016; Lo Iacono et al., 2017; Zhang et al., 2019a). Water shortage and drought have been found associated with diarrhoeal disease peaks (Epstein et al., 2020b; Subiros et al., 2019; Boithias et al., 2016), while some reviews found insufficient evidence of the effects of drought on diarrhoea (Levy et al, 2016 ; Asmall et al., 2021; Epstein et al., 2020b; Subiros et al., 2019; Boithias et al., 2016; Ramesh et al., 2016).

Heavy rainfall and higher than normal temperatures are associated with increased cholera risk in affected regions (very high confidence). Cholera is an acute diarrhoeal disease typically caused by the bacterium Vibrio cholerae that can result in severe morbidity and mortality. Maximum and minimum temperatures and precipitation have been negatively associated with cholera cases. Cholera outbreaks have occurred in several regions after natural disasters, including cholera incidence increasing three-fold in El Niño-sensitive regions of Africa (Mpandeli et al., 2018; Amegah et al., 2016; Escobar et al., 2015; Jutla et al., 2017; Asadgol et al., 2019; Moore et al., 2018; Moore et al., 2017; Camacho et al., 2018; IPCC, 2019a; Cross-Chapter Box ILLNESS in Chapter 2; Box 3.3).

Heavy rainfall, warmer weather and drought are linked to increased risks for other GI infections (high confidence). As temperature increases, bacterial causes of GI infection also appear to increase, and this association is variably influenced by humidity and rainfall (Ghazani et al., 2018; Levy, 2016). In New York it has been found that every 1°C increase in temperature was correlated with a 0.70–0.96% increase in daily hospitalisation for GI infections (Lin et al., 2016). In the Philippines, leptospirosis and typhoid fever showed an increase in incidence following heavy rainfall and flooding events (Matsushita et al., 2018).

FBDs refer to any illness resulting from ingesting food that is spoiled or contaminated by pathogenic bacteria, viruses, parasites, toxins, pesticides and/or medicines (WHO, 2018b). FBD risks are present throughout the food chain, from production to consumption, and most often arise due to contamination at source and from improper food handling, preparation and/or storage (Smith and Fazil, 2019; Semenza and Paz, 2021). As with WBDs, FBD outbreaks can follow multiple causal pathways as climatic risk factors interact with food production and distribution systems, urbanisation and population growth, resource and energy scarcity, decreasing agricultural productivity, price volatility, modification of diet trends, new technologies and the emergence of antimicrobial resistance (Lake, 2018; Yeni and Alpas, 2017). The burden of FBDs is also linked to malnutrition as reduced immunity increases susceptibility to various food-borne pathogens and toxins (FAO, 2020).

A strong association exists between increases in FBDs and high air and water temperatures and longer summer seasons (very high confidence). The risks occur through complex transmission pathways throughout the food chain and the wide range of food-borne pathogens (Cisse, 2019; Hellberg and Chu, 2016; Lake and Barker, 2018; Park et al., 2018b; Smith and Fazil, 2019). The food-borne pathogens of most concern are those having low infective doses, a significant persistence in the environment and high stress tolerance to temperature change (e.g., enteric viruses, Campylobacter spp., Shiga toxin-producing E. coli strains, Mycobacterium avium, tuberculosis complexes, parasitic protozoa and Salmonella) (Lake, 2018; Lake, 2017; Lake and Barker, 2018; Smith and Fazil, 2019; European Food Safety Authority 2020; Semenza and Paz, 2021). Priority risks include marine biotoxins, mycotoxins, salmonellosis, vibriosis, transfer of contaminants due to extreme precipitation, floods, increased use of chemicals in the food chain (plant protection products, fertilizers, veterinary drugs) and potential residues in food (European Food Safety Authority 2020; World Health Organization 2018b).

There is a strong association observed between the increase in average ambient temperature and increases in Salmonella infections (high confidence). Most types of Salmonella infections lead to salmonellosis, while some other types (SalmonellaTyphi and SalmonellaParatyphi) can lead to typhoid fever or paratyphoid fever. The transmission to humans of the non-typhoidal Salmonella infection, one of the most widespread FBDs, usually occurs through eating foods contaminated with animal faeces. Studies conducted in Australia (Milazzo et al., 2016), New Zealand (Lal et al., 2016), the UK (Lake, 2017), South Korea (Park et al., 2018a; Park et al., 2018c; Park et al., 2018a), Singapore (Aik et al., 2018) and Hong Kong, SAR of China (Wang et al., 2018a; Wang et al., 2018b), have shown that Salmonella outbreaks are strongly associated with temperature increases.

Significant associations exist between FBDs due toCampylobacter , precipitation and temperature (medium confidence). The timing of heat-associated Campylobacteriosis events varies across countries, with infection rates in the UK appearing to decline immediately after periods of high rainfall (Djennad et al., 2019; Lake et al., 2019; Rosenberg et al., 2018; Yun et al., 2016; Weisent et al., 2014). This suggests the association with climate may be indirect and due to weather conditions that encourage outdoor food preparation and recreational activities (Lake, 2017; Semenza and Paz, 2021).

Outbreaks of human and animal Cryptococcus have been reported as being associated with a combination of climatic factors and shifts in host and vector populations (Chang and Chen, 2015 ; Rickerts, 2019). The prevalence of childhood cryptosporidiosis, which is the second leading cause of moderate to severe diarrhoea among infants in the tropics and subtropics, shows associations with population density and rainfall, with contamination due to Cryptosporidium spp. being 2.61 times higher during and after heavy rain (Lal et al., 2019; Young et al., 2015; Khalil et al., 2018). Studies from Ghana, Guinea Bissau, Tanzania, Kenya and Zambia show a higher prevalence of Cryptosporidium during high rainfall seasons, with some peaks observed before, at the onset or at the end of the rainy season (Squire and Ryan, 2017).

Climatic risk factors for respiratory tract infections (RTIs) due to multiple pathogens (bacteria, viruses and fungi) include temperature and humidity extremes, dust storms, extreme precipitation events and increased climate variability. Amongst a range of RTIs, pneumonia and influenza represent a significant disease burden (Ferreira-Coimbra et al., 2020; Lafond et al., 2021; McAllister et al., 2019; Wang et al., 2020 c). The drivers of pneumonia incidence are complex and include a range of possible non-climate as well as climate factors. For example, chronic diseases (e.g., lung disease, chronic obstructive pulmonary disease (COPD) and asthma), other comorbidities, a weak immune system, age, gender, community, passive smoking, air pollution and childhood immunisation may confound the climate pneumonia relationship (Miyayo et al., 2021).

In temperate regions, the incidence of pneumonia is higher in winter months, but the exact causes of this seasonality remain debated (Mirsaeidi et al., 2016). With regards to temperature, various J-shaped, U-shaped or V-shaped temperature–pneumonia relationships have been reported in the literature (Huang et al., 2018 ; Kim et al., 2016; Liu et al., 2014; Qiu et al., 2016; Sohn et al., 2019) with such relationships dependent on location. Humidity also appears important but, like temperature, its effect is not consistent across studies – low temperatures and low humidity (Davis et al., 2016), high temperatures and high humidity (Lam et al., 2020) and low temperatures and high humidity (Miyayo et al., 2021) have all been found to be associated with an increased incidence of pneumonia.

Day-to-day variations in temperature also appear important. For Australia, increases in emergency room visits for childhood pneumonia are associated with sharp temperature drops (Xu et al., 2014). Large inter-daily changes in temperature are important for respiratory disease incidence in Guangzhou, China (Lin et al., 2013) and Shanghai (Lei et al., 2021) while rapidly changing and extreme temperatures during pregnancy have been linked to childhood pneumonia (Miao et al., 2017; Zeng et al., 2017; Zheng et al., 2021). In tropical and subtropical areas of Africa and Asia, pneumonia incidence has been reported to be higher during the rainy season, pointing to a positive association between pneumonia patterns and temperature and precipitation (Chowdhury et al., 2018a; Lim and Siow, 2018; Paynter et al., 2010).

The degree to which the timing, duration and magnitude of local influenza virus epidemics is dependent on climate factors is poorly understood (Lam et al., 2020). Further, a host of non-climate confounders are likely to influence the incidence of seasonal influenza (Caini et al., 2018). This poses a number of challenges for making reliable climate-based epidemiological forecasts for influenza (Gandon et al., 2016). Although no association between anomalous climate conditions and influenza have been reported in some locations (Lam et al., 2020), generally, low winter temperatures and humidity in temperate regions and periods of high humidity and precipitation in the tropical and subtropical regions have been linked to outbreaks of influenza (Deyle et al., 2016; Soebiyanto et al., 2015; Tamerius et al., 2013). However, the climate sensitivity of influenza may be more complex than this, with both high and low humidity; the amount and intensity of precipitation; solar activity and/or sunshine; and latitude also being important (Axelsen et al., 2014; Chong et al., 2020b; Geier et al., 2018; Park et al., 2019; Qu, 2016; Smith et al., 2017; Wang et al., 2017 c; Zhao et al., 2018a). Moreover, the shape of the climate variable influenza relationship may be conditioned on influenza type (Chong et al., 2020a). Further, distinct periods of weather variability characterised by rapid inter-daily changes in temperature may act as precursors to influenza epidemics as has been demonstrated for the marked 2017–2018 influenza season and others across the USA (Liu et al., 2020a; Zhao et al., 2018a). For the eastern Mediterranean, such rapid weather changes are associated with the ‘Cyprus Low’, with the timing and magnitude of seasonal influenza related to the interannual frequency of this particular weather regime (Hochman et al., 2021). Potentially, large-scale modes of climatic variability such as ENSO and the Indian Ocean Dipole, which strongly moderate the frequency of weather regimes in some parts of the world, could affect influenza pandemic dynamics. However, studies conducted to date report inconsistent results. Some point to an increased (decreased) severity of seasonal influenza during El Niño (La Niña) (Oluwole, 2015; Oluwole, 2017), while others find influenza to be more severe and frequent when coinciding with La Niña events (Chun et al., 2019; Flahault et al., 2016; Shaman and Lipsitch, 2013). This raises the possibly of non-stationary associations between large-scale modes of climatic variability and influenza dynamics (Onozuka and Hagihara, 2015) as found for other diseases (Kreppel et al., 2014), something that might be expected given El Niño’s time-varying impact on global precipitation and temperature fields and associated impacts on health outcomes (McGregor and Ebi, 2018).

Water shortage and drought are associated with skin diseases (Schachtel et al., 2021; Lundgren, 2018; Andersen and Davis, 2017; Kaffenberger et al., 2017; Andersen and Davis, 2017), trachoma (Ramesh et al., 2016) and violence (Epstein et al., 2020a); more research is warranted in these areas for future assessment.

CVDs are a group of disorders of the heart and blood vessels that include coronary heart disease, cerebrovascular disease, peripheral arterial disease, rheumatic heart disease, congenital heart disease, deep vein thrombosis and pulmonary embolism. CVDs are the leading cause of death globally and over three quarters of the world’s CVD deaths now occur in low- and middle-income countries (Roth et al., 2020).

Climate change affects the risk of CVD through high temperatures and extreme heat (assessed in Section 7.2.4.1) and through other mechanisms (medium confidence), though the degree to which non-temperature risks may increase remains unclear. For example, exposure to air pollutants including PM, ozone (via its precursors), black carbon, oxides of nitrogen, oxides of sulphur, hydrocarbons and metals can invoke pro-inflammatory and prothrombotic states, endothelial dysfunction and hypertensive responses (Giorgini et al., 2017; Stewart et al., 2017). Winter peaks in CVD events, associated with greater concentrations of air pollutants, have been reported in a range of countries and climates (Claeys et al., 2017; Stewart et al., 2017); however, the association between air pollution, weather and CVD events is complex and seems to differ between cold and warm months, particularly for gaseous pollutants such as ozone (Shi et al., 2020).

Climate change is projected to increase the number and severity of wildfires (Liu et al., 2015b; Youssouf et al., 2014) and the evidence for wildfire smoke-related CVD morbidity and mortality is suggestive of increased CVD morbidity and mortality risk (Chen et al., 2021a) including significant increases in certain cardiovascular outcomes (e.g., cardiac arrests) (Dennekamp et al., 2015). CVD risks to highly exposed populations, such as firefighters, are clearer (Navarro et al., 2019) and could increase with additional exposure driven by climate change.

Other climate-related mechanisms that may increase CVD risk include reductions in physical activity related to hot weather (Obradovich et al., 2017), sleep disturbance (Obradovich et al., 2017) and dehydration (Lim et al., 2015; Frumkin and Haines, 2019). There is little literature on how changes in winter weather may affect these risks. Saline intrusion of groundwater related to sea level rise (Taylor et al., 2012) may increase the salt intake of affected populations, a risk factor for hypertension that has been observed to increase blood pressure in exposed populations (Talukder et al., 2017; Al Nahian et al., 2018).

Lung diseases, including asthma, COPD and lung cancer, comprise the largest subsets of non-communicable pulmonary disease (Ferkol and Schraufnagel, 2014). Overall, the global burden of non-communicable lung disease including all chronic lung disease and lung cancer is substantial, being responsible for 10.6% of deaths and 5.9% of DALYs globally in 2019 (Vos et al., 2020).

Several non-communicable respiratory diseases are climate sensitive based on their exposure pathways (very high confidence). Multiple exposure pathways contribute to non-communicable respiratory disease (Deng et al., 2020), some of which are climate-related (Rice et al., 2014), including mobilisation and transport of dust (Schweitzer et al., 2018); changes in concentrations of air pollutants such as small particulates (PM2.5) and ozone formed by photochemical reactions sensitive to temperature (Hansel et al., 2016); increased wildland fires and related smoke exposure (Johnston et al., 2002; Reid et al., 2016); increased exposure to ambient heat driving reduced lung function and exacerbations of chronic lung disease (Collaco et al., 2018; Jehn et al., 2013; McCormack et al., 2016; Witt et al., 2015) and modification of aeroallergen production and duration of exposure (Ziska et al., 2019).

Burdens of allergic disease, particularly allergic rhinitis and allergic asthma may be changing in response to climate change (medium confidence) (D’Amato et al., 2020; Eguiluz-Gracia et al., 2020; Deng et al., 2020; Demain, 2018). This is supported by evidence showing an increase in the length of the North American pollen season attributable to climate change (Ziska et al., 2019), an association between timing of spring onset and higher asthma hospitalisations presumed to be due to higher pollen exposure (Sapkota et al., 2020) and other evidence linking aeroallergen exposure with a worsening burden of allergic disease (Demain, 2018; Poole et al., 2019).

Climate change is likely to increase the risk of several malignancies (high confidence), though the degree to which risks may increase remains unclear. Cancers, also known as malignant neoplasms, include a heterogeneous collection of diseases with various causal pathways, many with environmental influences. Malignant neoplasms impose a substantial burden of disease globally and are responsible for slightly more than 10 million deaths and 251 million DALYs globally in 2019 (Vos et al., 2020). Climatic hazards affect exposure pathways for several different chemical hazards associated with carcinogenesis (Portier et al., 2010). Most relevant literature has focused on elaborating potential pathways and producing qualitative or quantitative estimates of effect, though there is limited literature on current and projected impacts.

The vast majority of elaborated pathways point to increased risk; for example, there is concern that climate change may alter the fate and transport of carcinogenic polyaromatic hydrocarbons (Domínguez-Morueco et al., 2019) and increase mobilisation of carcinogens such as bromide (Regli et al., 2015), persistent organic pollutants (POPs) including polychlorinated-biphenyls that have accumulated in areas contaminated by industrial runoff (Miner et al., 2018) and radioactive material (Evangeliou et al., 2014). Exposure to these known carcinogens can occur through multiple environmental media and can be increased by climate change, for example through increased flooding related to extreme precipitation events and mobilisation of sediment where carcinogens have accumulated (León et al., 2017; Santiago and Rivas, 2012). In addition, there is concern that changes in ultraviolet light exposure related to shifts in precipitation may increase the incidence of malignant melanoma, particularly for outdoor workers (Modenese et al., 2018). Other harmful pathways include migration of and increased exposure to liver flukes, which cause hepatobiliary cancer (Prueksapanich et al., 2018) and the introduction of infectious diseases such as schistosomiasis that increase cancer risk due to climate-related migration (Ahmed et al., 2014). Increased exposure to carcinogenic toxins via multiple pathways is also a concern. Aflatoxin exposure, for example, is expected to increase in Europe (Moretti et al., 2019), India (Shekhar et al., 2018), Africa (Gnonlonfin et al., 2013; Bandyopadhyay et al., 2016) and North America (Wu et al., 2011). Other carcinogenic toxins originate from cyanobacteria blooms (Lee et al., 2017a), which are projected to increase in frequency and distribution with climate change (Wells et al., 2015; Paerl et al., 2016; Chapra et al., 2017).

Individuals suffering from diabetes are at higher risk of heat-related illness and death (medium confidence). Extreme weather events and rising temperatures have been found to increase morbidity and mortality in patients living with diabetes, especially in those with cardiovascular complications (Méndez-Lázaro et al., 2018; Zilbermint, 2020; Hajat et al., 2017). Evidence suggests that the local heat loss response of skin blood flow is affected by diabetes-related impairments, resulting in lower elevations in skin blood flow in response to a heat or pharmacological stimulus. Thermoregulatory sweating may also be diminished by type-2 diabetes, impairing the body’s ability to transfer heat from its core to the environment (Xu et al., 2019b). Higher rates of doctor consultations by patients with type-2 diabetes and diabetics with cardiovascular comorbidities have been observed during hot days, but without evidence of heightened risk of renal failure or neuropathy comorbidities (Xu et al., 2019b).

People with chronic illnesses are at particular risk during and after extreme weather events due to treatment interruptions and lack of access to medication (medium confidence). The impacts of extreme weather events on the health of chronically ill people are due to a range of factors including disruption of transport, weakened health systems including drug supply chains, loss of power and evacuations of populations (Ryan et al., 2015a). Evacuations also pose specific health risks to older adults (especially those who are frail, medically incapacitated or residing in nursing or assisted living facilities) and may be complicated by the need for concurrent transfer of medical records, medications and medical equipment (Becquart et al., 2018; Quast and Feng, 2019; US Global Change Research Program, 2016). Emergency room visits after Hurricane Sandy rose among individuals with type-2 diabetes (Velez-Valle et al., 2016).

Extreme heat events and extreme temperature have well-documented, observed impacts on health, mortality (very high confidence) and morbidity (high confidence). AR5 described the thermoregulatory mechanisms and responses, including acclimatisation, linking heat, cold and health, and these have been further confirmed by recent studies and reviews (e.g., Giorgini et al., 2017; Ikaheimo, 2018; McGregor et al., 2015; Stewart et al., 2017; Schuster et al., 2017; Zhang et al., 2018b). The health impacts of heat manifest clearly in periods of extreme heat often codified as heatwaves. For example, heatwaves across Europe (2003), Russia (2010), India (2015) and Japan (2018) resulted in significant death tolls and hospitalisations (McGregor et al., 2017; Hayashida et al., 2019). Heat continues to pose a significant health risk due to increases in exposure, changes in the size and spatial distribution of the human population, mounting vulnerability and an increase in extreme heat events (high confidence) (Harrington et al., 2017; Liu et al., 2017; Mishra et al., 2017; Rohat et al., 2019a; Rohat et al., 2019b; Rohat et al., 2019c; Watts et al., 2019). Some regions are already experiencing heat stress conditions approaching the upper limits of labour productivity and human survivability (high confidence). These include the Persian Gulf and adjacent land areas, parts of the Indus River Valley, eastern coastal India, Pakistan, northwestern India, the shores of the Red Sea, the Gulf of California, the southern Gulf of Mexico and coastal Venezuela and Guyana (Krakauer et al., 2020; Li et al., 2020; Raymond et al., 2020; Saeed et al., 2021; Xu et al., 2020).

Under a variety of methods, estimates of the world’s population exposed to extreme heat indicate very large and growing numbers and an increase since pre-industrial times. For example, Li et al. (2020) estimate that globally, 1.28 billion people each year experience heatwave conditions similar to that of the lethal Chicago 1995 event, compared with 0.99 billion people that would be similarly exposed under a pre-industrial climate. Further, for the 150 most populated cities of the world, a 500% increase in the exposure to extreme heat events occurred over the 1980–2017 period (Li et al., 2021), while for the 1986–2005 period, the total exposure to dangerous heat in Africa’s 173 largest cities was 4.2 billion person-days yr –1 (Rohat et al., 2019a). Globally the present exposure to heatwave events is estimated to be 14.8 billion person-days yr –1, with the greatest cumulative exposures measured in person-days occurring across southern Asia (7.19 billion), sub-Saharan Africa (1.43 billion), and north Africa and the Middle East (1.33 billion) (Jones et al., 2018).

The country level percentage of mortality attributable to non-optimum temperature (heat and cold) has been found to range from 3.4 to 11% (Gasparrini et al., 2015; Zhang et al., 2019b). Heat as a health risk factor has largely been overlooked in low- and middle-income countries (Campbell et al., 2018; Green et al., 2019; Dimitrova et al., 2021). For 2019, the GBD report estimates the burden of DALYs attributable to low temperature was 2.2 times greater than the burden attributable to high temperature. However, this global figure obscures important regional variations. Countries with a high sociodemographic index—mainly mid-latitude high-income temperate to cool climate countries— were found to have a cold-related burden 15.4 times greater than the heat-related burden, while for warm lower-income regions, such as south Asia and sub-Saharan Africa, the heat-related burden was estimated to be 1.7 times and 3.6 times greater, respectively (Murray et al., 2020). For countries where data availability permits, there is evidence that extreme heat (and extreme cold) leads to higher rates of premature deaths (Armstrong et al., 2017; Cheng et al., 2018; Costa et al., 2017).

Rapid changes and variability in temperatures are observed to increase heat-related health and mortality risks, the outcomes varying across temperate and tropical regions (Guo et al., 2016; Cheng et al., 2019; Kim et al., 2019a; Tian et al., 2019; Zhang et al., 2018b; Zhao et al., 2019).

Several lines of evidence point to a possible decrease in population sensitivity to heat, albeit mainly for high-income countries (high confidence), arising from the implementation of heat warning systems, increased awareness and improved quality of life. (Sheridan and Allen, 2018). Evidence suggests a general decrease in the impact of heat on daily mortality (Diaz et al., 2018; Kinney, 2018; Miron et al., 2015), a decline in the relative risk attributable to heat (Åström et al., 2018; Barreca et al., 2016; Petkova et al., 2014) and an increase in the minimum mortality temperature (MMT) (Åström et al., 2018; Folkerts et al., 2020; Follos et al., 2021; Chung et al., 2018; Todd and Valleron, 2015; Yin et al., 2019). It is difficult to draw conclusions regarding trends in heat sensitivity for low- to middle-income countries and specific vulnerable groups as these are under-represented in the literature (Sheridan and Allen, 2018). Trends in heat sensitivity are generally believed to be scale and situation dependent, but there is considerable variability in changes in heat sensitivity as measured by trends in heat-related mortality or MMT (Follos et al., 2021; Kim et al., 2019a; Lee et al., 2021), with notable variability across different population groups (Lu et al., 2021).

Temperature interacts with heat-sensitive physiological mechanisms via multiple pathways to affect health. In the worst cases, these lead to organ failure and death (Mora et al., 2017a; Mora et al., 2017b). Excess deaths during extreme heat events occur predominantly in older individuals and are overwhelmingly cardiovascular in origin (very high confidence). A higher occurrence of CVD mortality in association with prolonged period of low temperatures has been well documented globally (Giorgini et al., 2017; Stewart et al., 2017); however, there is growing evidence that cardiovascular deaths are more related to heat events than cold spells (Chen et al., 2019; Liu et al., 2015a; Bunker et al., 2016). While there is strong association between ambient temperature and cardiovascular events globally, there are complex interactions and modulators of individual response (Wang et al., 2017 b). Further, some CVD morbidity sub-groups such as myocardial infarction (MI) and stroke hospitalisation display temperature sensitivity while others do not (Bao et al., 2019; Sun et al., 2018; Wang et al., 2016). Although older adults have inherent sensitivities to temperature-related health impacts (Bunker et al., 2016; Phung et al., 2016), children can also be affected by extreme heat (Xu et al., 2014). Cardiovascular capacity or health is also a critical determinant of individual health outcomes (Schuster et al., 2017). Medications to treat CVDs, such as diuretics and beta-blockers, may impair resilience to heat stress (Stewart et al., 2017). Other mediating factors in the causal pathway range from alcohol consumption (Cusack et al., 2011; Epstein and Yanovich, 2019) and obesity (Speakman, 2018) to pre-existing conditions, such as diabetes and hyperlipidaemia, and urban characteristics (Chen et al., 2019; Sera et al., 2019).

Under extreme heat conditions, increases in hospitalisations have been observed for fluid disorders, renal failure, urinary tract infections, septicaemia, general heat stroke as well as unintentional injuries (Borg et al., 2017; Phung et al., 2017; Goggins and Chan, 2017; Hayashida et al., 2019; Hopp et al., 2018; Ito et al., 2018; Kampe et al., 2016; McTavish et al., 2018; Ponjoan et al., 2017; van Loenhout et al., 2018). Hospitalisations and mortality due to respiratory disorders also occur during heat events with the interactive role of air quality being important for some locations but not others (Krug et al., 2019; Pascal et al., 2021; Patel et al., 2019). Increased levels of heat-related hospitalisation also manifest in elevated levels of emergency service calls (Cheng et al., 2016; Guo, 2017; Papadakis et al., 2018; Williams et al., 2020).

Heat- and cold-related health outcomes vary by location (Dialesandro et al., 2021; Hu et al., 2019; Phung et al., 2016), suggesting outcomes are highly moderated by socioeconomic, occupational and other non-climatic determinants of individual health and socioeconomic vulnerability (Åström et al., 2020; McGregor et al., 2017; McGregor et al., 2017; Schuster et al., 2017; Benmarhnia et al., 2015; Watts et al., 2019) (high confidence). For example, access to air conditioning is an important determinant of heat-related health outcomes for some locations (Guirguis et al., 2018; Ostro et al., 2010). Although there is a paucity of global level studies of the effectiveness of air conditioning for reducing heat-related mortality, a recent assessment indicates increases in air conditioning explains only part of the observed reduction in heat-related excess deaths, amounting to 16.7% in Canada, 20.0% in Japan, 14.3% in Spain and 16.7% in the US (Sera et al., 2020).

Significant effects of heat exposure are evident in sport and work settings with exertional heat illness leading to death and injury (Adams and Jardine, 2020). Although most studies of heat-related sports injuries refer to high-income countries, these point to an increasing number of heat injuries with widening participation in sport and an increasing frequency of extreme heat events. The highest rates of exertional heat illness are reported for endurance type events (running, cycling and adventure races), American football and athletics (Gamage et al., 2020; Grundstein et al., 2017; Kerr et al., 2020; McMahon et al., 2021; Yeargin et al., 2019). The health, safety and productivity consequences of working in extreme heat are widespread (Ma et al., 2019; Morabito et al., 2021; Kjellstrom et al., 2019; Orlov et al., 2020; Smith et al., 2021; Vanos et al., 2019; Varghese et al., 2020; Williams et al., 2020). Occupational heat strain in outdoor workers manifests as dehydration, mild reduction in kidney function, fatigue, dizziness, confusion, reduced brain function, loss of concentration and discomfort (Al-Bouwarthan et al., 2020; Boonruksa et al., 2020; Habibi et al., 2021; Levi et al., 2018; Venugopal et al., 2021; Xiang et al., 2014). In the case of armed forces, a global review of the available literature points to a slightly higher incidence of heat stroke in men compared to women but a higher proportion of heat intolerance and greater risk of exertional heat illness amongst women (Alele et al., 2020). There is also some evidence that for healthcare workers, the risk of occupational heat stress grew during the COVID-19 pandemic due to the need to wear personal protective equipment (Foster et al., 2020; Lee et al., 2020; Messeri et al., 2021). Based on a systematic review of the literature, one study estimates global costs from heat-related lost work time were USD 280 billion in 1995 and USD 311 billion in 2010, with low- and middle-income countries and countries with warmer climates experiencing greater losses as a proportion of gross domestic project (GDP) (Borg et al., 2021). Other global level assessments note an increase in the potential hours of work lost due to heat over the 2000–2018 period; in 2018, 133.6 billion potential work hours were lost, amounting to 45 billion hours more than in 2000 (Watts et al., 2019). For China, heat-related productivity losses have been estimated at 9.9 billion hours in 2019, equivalent to 0.5% of the total national work hours for that year, with Guangdong province, one of the warmest regions in China, accounting for almost a quarter of the losses (Cai et al., 2021).

Wide ranging knowledge regarding the specific detection of heat- and cold-related mortality/morbidity and its attribution to observed climate change is lacking . Although there has been an observed increase in winter-season temperatures for a number of regions, to date there is variable evidence for a consequential reduction in winter mortality and susceptibility to cold over time due to milder winters; some countries demonstrate decreasing trends, while other countries show stable or even increasing trends in cold-attributable mortality fractions over time (e.g., Arbuthnott et al. (2020); Åström et al. (2013); Diaz et al. (2019); Hajat (2017); Hanigan et al. (2021); Lee et al. (2018b)). While there is a burgeoning literature on the attribution of extreme heat events to climate change (e.g., Vautard et al. (2020)), the number of studies that assess the extent to which observed changes in heat-related mortality may be attributable to climate change is small (Ebi et al., 2020). During the 2003 European heatwave, anthropogenic climate change increased the risk of heat-related mortality by approximately 70% and 20% for London and Paris, respectively (Mitchell et al., 2016). For the severe heat event across Egypt in 2015, the impact on human discomfort was 69% (±17%) more likely due to anthropogenic climate change (Mitchell, 2016), and for Stockholm, Sweden, it has been estimated that mortality due to temperature extremes for 1980 to 2009 was double what would have occurred without climate change (Åström et al., 2013). To date there has only been one multi-country attempt to quantify the heat-related human health impacts that have already occurred due to climate change. Based on an analysis of 732 locations spanning 43 countries for the 1991–2018 period, the study found that on average 37.0% (inter-quartile range 20.5–76.3%) of warm-season heat-related deaths can be attributed to anthropogenic climate change, equivalent to an average mortality rate of 2.2/100,000 (median: 1.67/100,000; interquartile range: 1.08–2.34/100,000). Regions with a high attributed percentage (> 50%) include southern and western Asia (Iran and Kuwait), Southeast Asia (Philippines and Thailand) and several countries in Central and South America. Those with lower values (< 35%) include Western Europe (the Netherlands, Germany and Switzerland), eastern Europe (Moldova, the Czech Republic and Romania), southern Europe (Greece, Italy, Portugal and Spain), North America (USA) and eastern Asia (China, Japan and South Korea) (Vicedo-Cabrera et al., 2021). Due to data restrictions, some of the poorest and most susceptible regions to climate change and increases in heat exposure, such as west and east Africa (Asefi-Najafabady et al., 2018; Sylla et al., 2018) and south Asia, could not be included in the analysis (Mitchell, 2021).

Injuries comprise a substantial portion of the global burden of disease. In 2019, injuries comprised 9.82% of total global DALYs and 7.61% of deaths (Vos et al., 2020). The causal pathways for many injuries, particularly those from heat and extreme weather events, flooding and fires, exhibit clear climate sensitivity (Roberts and Arnold, 2007; Roberts and Hillman, 2005), as do some injuries occurring in occupational settings (Marinaccio et al., 2019; Sheng et al., 2018), but a comprehensive assessment of climate sensitivity in injury causal pathways has not been done. Certain groups, including Indigenous Peoples, children and the elderly (Ahmed et al., 2020) are at greater risk for a wide range of injuries. Extreme events impose substantial disease burden directly as a result of traumatic injuries, drowning and burns and large mental health burdens associated with displacement (Fullilove, 1996), depression and post-traumatic stress disorder (PTSD), but the overall injury burden associated with extreme weather is not known. It is known that the Asia-Pacific region has experienced the highest relative burden of injuries from extreme weather in recent decades (Hashim and Hashim, 2016).

Extreme weather imposes a substantial morbidity and mortality burden that is quite variable by location and hazard. The proportion of this burden related specifically to injuries is not established. From 1998 to 2017 there were 526,000 deaths from 11,500 extreme weather events, and the average annual attributable all-cause mortality incidence in the ten most affected countries was 3.5 per 100,000 population (Eckstein et al., 2017). Rates can be much higher; mortality incidence in Puerto Rico and Dominica from extreme weather were 90.2 and 43.7 per 100,000 population in 2017, respectively (Eckstein et al., 2017). Not all of these deaths are from injuries, and the proportion of mortality and morbidity associated with injuries varies by location and hazard. One review found that one-year post-event prevalence rates for injuries associated with extreme events (floods, droughts, heatwaves and storms) in developing countries ranged from 1.4% to 37.9% (Rataj et al., 2016). Other literature has documented an increase in the risk of motor vehicle accidents in association with extreme precipitation (Liu et al., 2017; Stevens et al., 2019), temperature (Leard and Roth, 2019) and sandstorms (Islam et al., 2019) and an increased risk of traumatic occupational injuries associated with temperature extremes, particularly extreme heat, likely from fatigue and decreased psychomotor performance (Varghese et al., 2019).

There is clear evidence of climate sensitivity for multiple injuries from floods, fires and storms, but there is a need for additional evidence regarding the current injury burden attributable to climate change. It is as likely as not that climate change has increased the current burden of disease from injuries related to extreme weather, particularly in low-income settings (low confidence). Approximately 120 million people are exposed to coastal flooding annually (Nicholls et al., 2007), causing an estimated 12,000 deaths (Shultz et al., 2005), and there is significant concern for worsening flooding associated with climate change (Shultz et al., 2018a; Shultz et al., 2018b; Woodward and Samet, 2018) but very limited quantification of attributable burden. A range of adverse health outcomes has been identified in a study of fires in sub-zero temperatures that are thought to be increasing in frequency due to climate change (Metallinou and Log, 2017).

Maternal and neonatal disorders accounted for 3.7% of total global deaths and 7.8% of global DALYs in 2019 (Vos et al., 2020). Children and pregnant women have potentially higher rates of vulnerability and/or exposure to climatic hazards, extreme weather events and undernutrition (Garcia and Sheehan, 2016; Sorensen et al., 2018; Chersich et al., 2018). Available evidence suggests that heat is associated with higher rates of pre-term birth (Wang et al., 2020), low birthweight, stillbirth, neonatal stress (Cil and Cameron, 2017; Kuehn and McCormick, 2017) and adverse child health (Kuehn and McCormick, 2017). Extreme weather events are associated with reduced access to prenatal care, unattended deliveries (Abdullah et al., 2019) and decreased paediatric healthcare access (Haque et al., 2019).

Climate variability and change contribute to food insecurity that can lead to malnutrition, including undernutrition, overweight and obesity, and to disease susceptibility, particularly in low- and middle-income countries (high confidence). Since AR5, analyses of the links between climate change and food expanded beyond undernutrition to include the impacts of climate change on a wider set of diet- and weight-related risk factors and their impacts on NCDs, along with the role of dietary choices for GHG emissions (IPCC, 2019b) including dietary inadequacy (deficiencies, excesses or imbalances in energy, protein and micronutrients), infections and sociocultural factors (Global Nutrition Report 2020). Undernutrition exists when a combination of insufficient food intake, health, and care conditions results in one or more of the following: underweight for age, short for age (stunted), thin for height (wasted), or functionally deficient in vitamins and/or minerals (micronutrient malnutrition or ‘hidden hunger’). Food insecurity and poor access to nutrient-dense food contribute not only to undernutrition but also to obesity and susceptibility to NCDs in low- and middle-income countries (FAO et al., 2018; Swinburn et al., 2019).

Globally, more than 690 million people are undernourished, 144 million children are stunted (chronic undernutrition), 47 million children are wasted (acute undernutrition), and more than 2 billion people have micronutrient deficiencies (FAO, 2020). More than 135 million people across 55 countries experienced acute hunger requiring urgent food, nutrition and livelihood assistance in 2019 (FSIN/GNAFC, 2020). The COVID-19 pandemic is projected to increase the number of acutely food insecure people to 270 million people (FSIN, 2020) and worsen malnutrition levels (FAO et al., 2020; Rippin et al., 2020). The relationships between climate change and obesity vary based on geography, population sub-groups and/or stages of economic growth and population growth (An et al., 2017). Increasing temperatures could contribute to obesity through reduced physical activity, increased prices of produce or shifts in eating patterns of populations towards more processed foods (An et al., 2018). In the largest global study to date exploring the connections between child diet diversity and recent climate, data from 19 countries in six regions (Asia, Central America, South America, north Africa, southeast Africa and west Africa) indicated significant reductions in diet diversity associated with higher temperatures and significant increases in diet diversity associated with higher precipitation (Niles et al., 2021).

Climate change can affect the four aspects of food security: food production and availability, stability of food supplies, access to food and food utilisation (IPCC, 2019b). Access to sufficient food does not guarantee nutrition security. Extreme weather and climate events can result in inadequate or insufficient food consumption, increasing susceptibility to infectious diseases (Rodriguez-Llanes et al., 2016; Gari et al., 2017; Kumar et al., 2016; Lazzaroni and Wagner, 2016) but also to being overweight or obese and increasing susceptibility to non-communicable diseases in low- and middle-income countries (FAO, 2018; Swinburn et al., 2019).

Nearly half of all deaths in children under five years of age are attributable to undernutrition, putting children at greater risk of dying from common infections. Undernutrition in the first 1,000 days of a child’s life can lead to stunted growth, which can result in impaired cognitive ability and reduced future school and work performance and the associated costs of stunting in terms of lost economic growth can be of the order of 10% of GDP yr –1 in Africa (UNICEF/WHO/WBG, 2019).

At the same time, diseases associated with high-calorie, unhealthy diets are increasing globally, with 38.3 million overweight children under five years of age (Global Nutrition Report, 2018), 2.1 billion overweight or obese adults and the global prevalence of diabetes almost doubling in the past 30 years (Swinburn et al., 2019). Unbalanced diets, such as diets low in fruits and vegetables and high in red and processed meat, are the number one risk factor for mortality globally and in most regions (Gakidou et al., 2018; Afshin et al., 2019).

Socioeconomic factors that mediate the influence of climate change on nutrition include cultural and societal norms; governance, institutions, policies and fragility; human capital and potential; and social position and access to healthcare, education and food aid (Rozenberg, 2017; Alkerwi et al. 2015; Tirado, 2017; FAO et al., 2018; Global Nutrition Report 2020). Extreme events may affect access to adequate diets, leading to malnutrition and increasing the risk of disease (Beveridge et al., 2019; Rodriguez-Llanes et al., 2016; Gari et al., 2017; Kumar et al., 2016; Lazzaroni and Wagner, 2016; Thiede and Gray, 2020).

Climate change in northern regions, including Arctic ecosystems, is causing permafrost to thaw, creating the potential for mercury (Hg) to enter the food chain (medium agreement, low evidence)as methyl mercury (MeHg), which is highly neurotoxic and nephrotoxic and bioaccumulates and biomagnifies throughout the food chain via dietary uptake of fish, seafood and mammals. Mercury methylation processes in aquatic environments have been found to be exacerbated by ocean warming, coupled with more acidic and anoxic sediments (FAO, 2020). Consumption of mercury-contaminated fish has been found to be linked to neurological disorders due to methyl mercury poisoning (i.e., Minamata disease) that is associated with climate change-contaminant interactions that alter the bioaccumulation and biomagnification of toxic and fat-soluble persistent organic pollutants and polychlorinated biphenyls (PCBs) (Alava et al., 2017) in seafood and marine mammals (medium confidence). Indigenous Peoples have a higher exposure to such risks because of the accumulation of such toxins in traditional foods (J.J. et al., 2017). Contamination of food with PCBs and dioxins has a range of adverse health impacts (Lake et al., 2015).

Chapter 5 (Sections 5.4.3, 5.5.2.3, 5.8.1, 5.8.2, 5.8.3, 5.9.1, 5.11.1, 5.11.3, 5.12.3) discusses the possible impacts of climate change on food safety, including exposure to toxigenic fungi, PCBs and other POPs, mercury and harmful algal blooms.

Climate change may affect animal health management practices, potentially leading to an increased use of pesticides or veterinary drugs (such as preventive antimicrobials) that could result in increased levels of residues in foods (high agreement, medium/low evidence) (Beyene et al., 2015; FAO and WHO, 2018; European Food Safety Authority, 2020; MacFadden et al., 2018).

A wide range of climatic events and conditions have observed and detrimental impacts on mental health (very high confidence). The pathways through which climatic events affect mental health are varied, complex and inter-connected with other non-climatic influences that create vulnerability (Figure 7.6). The climatic exposure may be direct, such as experiencing an extreme weather event or prolonged high temperatures, or indirect, such as mental health consequences of undernutrition or displacement. Exposure may also be vicarious, with people experiencing decreased mental health associated with observing the impact of climate change on others or simply by learning about climate change. Non-climatic moderating influences range from an individual’s personality and pre-existing conditions, to social support, and to structural inequities (Gariepy et al., 2016; Hrabok et al., 2020; Nagy et al., 2018; Silva et al., 2016b). Depending on these background and contextual factors, similar climatic events may result in a range of potential mental health outcomes, including anxiety, depression, acute traumatic stress, PTSD, suicide, substance abuse and sleep problems, with conditions ranging from being mild in nature to those that require hospitalisation (Berry et al., 2010; Cianconi et al., 2020; Clayton et al., 2017; Ruszkiewicz et al., 2019; Bromet et al., 2017; Lowe, 2019). The line between mental health and more general well-being is permeable, but in this section we refer to diagnosable mental disorders—conditions that disrupt or impair normal functioning through impacts on mood, thinking or behaviour.

There is an observable association between high temperatures and mental health decrements (high confidence), with an additional possible influence of increased precipitation (medium agreement, medium evidence). Heat-associated mental health outcomes include suicide (Williams et al., 2015a; Carleton, 2017; Burke et al., 2018; Kim et al., 2019b; Thompson et al., 2018; Schneider et al., 2020; Cheng et al., 2021; Baylis et al., 2018; Obradovich et al., 2018); psychiatric hospital admissions and emergency room visits for mental disorders (Hansen et al., 2008; Wang et al., 2014; Chan et al., 2018; Mullins and White, 2019; Yoo et al., 2021); experiences of anxiety, depression and acute stress (Obradovich et al., 2018; Mullins and White, 2019); and self-reported mental health (Li et al., 2020). In Canada, Wang et al. (2014) found an association between mean heat exposure of 28°C and greater hospital admissions within 0 to 4 days for mood and behavioural disorders (including schizophrenia, mood and neurotic disorders). A US study found mental health problems increased by 0.5% when average temperatures exceed 30°C, compared to averages between 25°C and 30°C; a 1°C warming over five years was associated with a 2% increase in mental health problems (Obradovich et al., 2018). Another study found a 1°C rise in monthly average temperatures over several decades was associated with a 2.1% rise in suicide rates in Mexico and a 0.7% rise in suicide rates in the USA (Burke et al., 2018). A systematic review of published research using a variety of methodologies from 19 countries (Thompson et al., 2018) found an increased risk of suicide associated with a 1°C rise in ambient temperature.

Discrete climate hazards including storms (Kessler et al., 2008; Boscarino et al., 2013; Boscarino et al., 2017; Obradovich et al., 2018), floods (Baryshnikova and Pham, 2019), heatwaves, wildfires and drought (Hanigan et al., 2012; Carleton, 2017; Zhong et al., 2018; Charlson et al., 2021) have significant negative consequences for mental health (very high confidence). A large body of research identifies the impacts of extreme weather events on PTSD, anxiety and depression; much of the research has been done in the USA and the UK, but a growing number of studies find evidence for similar impacts on mental health in other countries, including Spain (Foudi et al., 2017), Brazil (Alpino et al., 2016), Chile (Navarro et al., 2016), Small Island Developing States (Kelman et al., 2021) and Vietnam (Pollack et al., 2016). Approximately 20–30% of those who live through a hurricane develop depression and/or PTSD within the first few months following the event (Obradovich et al., 2018; Schwartz et al., 2015; Whaley, 2009), with similar rates for people who have experienced flooding (Waite et al., 2017; Fernandez et al., 2015). Studies conducted in South America and Asia indicate an increase in PTSD and depressive disorders after extreme weather events (Rataj et al., 2016). Evidence is lacking for African countries (Otto et al., 2017). Children and adolescents are particularly vulnerable to post-traumatic stress after extreme weather events (Brown et al., 2017; Hellden et al., 2021; Kousky, 2016), and increased susceptibility to mental health problems may linger into adulthood (Maclean et al., 2016).

Wildfires have observed negative impacts on mental health (high confidence). This is due to the trauma of the immediate experience and/or subsequent displacement and evacuation (Dodd et al., 2018; Brown et al., 2019; Psarros et al., 2017; Silveira et al., 2021b). Sub-clinical outcomes, such as increases in anxiety, sleeplessness or substance abuse are reported in response to wildfires and extreme weather events, with impacts being pronounced among those who experience greater losses or are more directly exposed to the event; this may include first responders.

Mental health impacts can emerge as result of climate impacts on economic, social and food systems (high confidence). For example, malnutrition among children has been associated with a variety of mental health problems (Adhvaryu et al., 2019; Hock et al., 2018; Yan et al., 2018), as has food insecurity among adults (Lund et al., 2018). The economic impacts of droughts have been associated with increases in suicide, particularly among farmers (Carleton, 2017; Edwards et al., 2015; Vins et al., 2015); those whose occupations are likely to be affected by climate change report that it is a source of stress that is linked to substance abuse and suicidal ideation (Kabir, 2018). Studies of Indigenous Peoples often describe food insecurity or reduced access to traditional foods as a link between climate change and reduced mental health (Middleton et al., 2020b). The loss of family members, for example due to an extreme weather event, increases the risk of mental illness (Keyes et al., 2014). Individuals in low- and middle-income countries may be more severely impacted due to lesser access to mental health services and lower financial resources to help cope with impacts compared with high-income countries (Abramson et al., 2015).

Anxiety about the potential risks of climate change and awareness of climate change itself can affect mental health even in the absence of direct impacts (low confidence). There is a need for more evidence about the prevalence or severity of climate change-related anxiety, sometimes called ecoanxiety, but national surveys in the USA, Europe and Australia show that people express high levels of concern and perceived harm associated with climate change (Steentjes et al., 2017; Clayton and Karazsia, 2020; Cunsolo and Ellis, 2018; Helm et al., 2018; Leiserowitz et al., 2017; Reser et al., 2012; Steentjes et al., 2017). In a US sample, perceived ecological stress, defined as personal stress associated with environmental problems, predicted depressive symptoms (Helm et al., 2018); in a sample of Filipinos, climate anxiety was correlated with lower mental health (Reyes et al., 2021) and a non-random study in 25 countries showed positive correlations between negative emotions about climate change and self-rated mental health (Ogunbode et al., 2021). However, an earlier study found no correlation between climate change worry and mental health issues (Berry and Peel, 2015). Because the perceived threat of climate change is based on subjective perceptions of risk and coping ability as well as on experiences and knowledge (Bradley et al., 2014), even people who have not been directly affected may be stressed by a perception of looming danger (Clayton and Karazsia, 2020). Not surprisingly, those who have directly experienced some of the effects of climate change may be more likely to show such responses. Indigenous Peoples, whose culture and well-being tend to be strongly linked to local environments, may experience mental health effects associated with changes in environmental risks; studies suggest connections to an increase in depression, substance abuse or suicide in some Indigenous Peoples (Canu et al., 2017; Cunsolo Willox et al., 2013; Middleton et al., 2020b; Jaakkola et al., 2018).

Overall, research suggests that climate change has already had negative effects on subjective well-being (medium confidence). C limate change can affect well-being through a number of pathways, including loss of access to green and blue spaces due to damage from storms, coastal erosion, drought or wildfires; heat; decreased air quality; and disruptions to one’s normal pattern of behaviour, residence, occupation or social interactions (Hayward and Ayeb-Karlsson, 2021). For example, substantial evidence shows a negative correlation between air pollution and subjective well-being or happiness (Apergis, 2018; Cunado and de Gracia, 2013; Lu, 2020; Luechinger, 2010; Menz and Welsch, 2010; Orru et al., 2016; Yuan et al., 2018; Zhang et al., 2017a); in the reverse direction, there is evidence not only that time in nature but more specifically a feeling of connectedness to nature are both associated with well-being (Martin et al., 2020) and healthy ecosystems offer opportunities for health improvements (Pretty and Barton, 2020). Negative emotions such as grief—often termed ‘solastalgia’ (Albrecht et al., 2007)—are associated with the degradation of local or valued landscapes (Eisenman et al., 2015; Ellis and Albrecht, 2017; Polain et al., 2011; Tschakert et al., 2017; Tschakert et al., 2019), which may threaten cultural rituals, especially among Indigenous Peoples (Cunsolo and Ellis, 2018; Cunsolo et al., 2020). Studies conducted in the Solomon Islands and Tuvalu found qualitative and quantitative evidence of experiences of climate change and worry about the future, with negative impacts on respondents’ well-being (Asugeni et al., 2015; Gibson et al., 2020).

Heat is one of the best-studied aspects of climate change observed to reduce well-being (high confidence). Higher summer temperatures are associated with decreased happiness and ratings of well-being (Carleton and Hsiang, 2016; Miles-Novelo and Anderson, 2019; Connolly, 2013; Noelke et al., 2016; Baylis et al., 2018; Moore et al., 2019; Wang et al., 2020 b). A study of 1.9 million Americans (Noelke et al., 2016) found that exposure to one day averaging 21°C–27°C was associated with reduced well-being by 1.6% of a standard deviation and days above 32°C were associated with reduced well-being by 4.4% of a standard deviation relative to a reference interval of 10°C–16°C. A similar relationship between heat and mood has been observed in China, where expressed mood began to decrease when the average daily temperature was over 20°C (Wang et al., 2020 b). The causal mechanism is unclear but could be due to impacts on health, economic costs or social interactions (Belkin and Kouchaki, 2017; Osberghaus and Kühling, 2016) or reduced quality or quantity of sleep (Fujii et al., 2015; Obradovich et al., 2017; Obradovich and Migliorini, 2018). Heat has also been associated with inter-personal and inter-group aggression and increases in violent crime (Heilmann et al., 2021; Mapou et al., 2017; Tiihonen et al., 2017). For the most part, studies have measured daily response to average daily temperatures and are unable to predict whether the effect is cumulative in response to a sequence of unusually warm days. However, there is no evidence that adaptation occurs over time to eliminate the negative response to very warm temperatures (Moore et al., 2019). Some research has found a negative effect of extreme cold on well-being (Yoo et al., 2021); increasing winter temperatures associated with climate change could serve to compensate for the impact of increased summer temperatures. However, the effect of high temperatures is typically found to be stronger than the effect of low temperatures, and in some cases no detrimental impacts of cold weather are found (Almendra et al., 2019; Mullins and White, 2019).

Climate change also threatens well-being defined in terms of capabilities or the capacity to fulfil one’s potential and fully participate in society. Heat can limit labour capacity; one study estimated that 45 billion hours of labour productivity were lost in 2018 compared to 2000 due to high temperatures (Watts et al., 2019). Both heat and air pollution also impair human capabilities through a negative effect on cognitive performance (Taylor et al., 2016b) and even impair skills acquisition, reducing the ability to learn (Park et al., 2021) and affecting marginalised groups more strongly (Park et al., 2020), although findings are inconsistent and depend in part on the nature of the task (low confidence).

Systematic reviews have found an association between higher ambient levels of fine airborne particles with cognitive impairment in the elderly and with behavioural problems (related to impulsivity and attention problems) in children (Power et al., 2016; Yorifuji et al., 2017; Younan et al., 2018; Zhao et al., 2018b) (medium confidence). Malnutrition has also been associated with reduced educational achievement and long-term decrements in cognitive function (Acharya et al., 2019; Asmare et al., 2018; Na et al., 2020; Kim et al., 2017; Talhaoui et al., 2019).

Reliable global estimates of voluntary climate-related migration within and between countries are not available due to a general absence of data of this specific nature, with existing national and global datasets often lacking information on migration causation or motivation. Better data are available for involuntary displacements within countries for reasons associated with weather-related hazards. Data collected annually since 2008 on internal displacements attributed to extreme weather events by the IDMC indicate that extreme storms and floods are the two most significant weather-related drivers of population displacements globally. Because of improvements in collection sources and methods since it first began reporting data in 2008, upward trends since that year in the total reported annual number of people displaced should be treated cautiously. However, it is reasonable to conclude that the average annual rate currently exceeds 20 million people globally, with considerable interannual variation due to the frequency and severity of extreme events in heavily populated areas. Regional distribution of displacement events has been consistent throughout the IDMC data collection period (high confidence), with displacement events occurring most often in East, Southeast and south Asia; sub-Saharan Africa; the USA; and the Caribbean region (Figure 7.7). Relative to their absolute population size, small island states experience a disproportionate risk of climate-related population displacements (Desai et al., 2021) (high confidence).

Tropical cyclones and extreme storms are a particularly significant displacement risk in East and Southeast Asia, the Caribbean region, the Bay of Bengal region and southeast Africa (IDMC 2020) (high confidence). The scale of immediate displacement from any given storm and potential for post-event migration depend heavily on the extent of damage to housing and livelihood assets and the responsive capacity of governments and humanitarian relief agencies (Saha, 2016; Islam et al., 2018; Mahajan, 2020; Spencer and Urquhart, 2018). In Bangladesh, the rural poor are most often displaced, with initial increases in short-term, labour-seeking migration followed by more permanent migration by some groups (Saha, 2016; Islam and Hasan, 2016; Islam and Shamsuddoha, 2017). Past hurricanes in the Caribbean basin have generated internal and inter-state migration within the region, typically along pre-existing social networks, and to the USA (Loebach, 2016; Chort and de la Rupelle, 2016). In 2017, Hurricanes Irma and Maria caused widespread damage to infrastructure and health services, and a slow recovery response by authorities was followed by the migration of tens of thousands of Puerto Ricans to Florida and New York (Zorrilla, 2017; Echenique and Melgar, 2018). In the US, coastal counties experience increased out-migration after hurricanes that flows along existing social networks (Hauer, 2017), with post-disaster reconstruction employment opportunities potentially attracting new labour migrants to affected areas (Ouattara and Strobl, 2014; Curtis et al., 2015; DeWaard et al., 2016; Fussell et al., 2018).

Riverine flood displacement can lead to increases or decreases in temporary or short-distance migration flows, depending on the local context (medium confidence) (Robalino et al., 2015; Ocello et al., 2015; Afifi et al., 2016; Koubi et al., 2016b). Floods are a particularly important driver of displacement in river valleys and deltas in Asia and Africa, although large flood-related displacements have been recorded by the IDMC in all regions. In Africa, populations exposed to low flood risks, as compared with other regions, are observed to have a greater vulnerability to displacement due to limited economic resources and adaptive capacity (Kakinuma et al., 2020). In areas where flooding is especially frequent, in situ adaptations may be more common, and out-migration may temporarily decline after a flood (Afifi et al., 2016; Chen et al., 2017; Call et al., 2017). Rates of indefinite or permanent migration tend not to change following riverine floods unless damage to homes and livelihood assets is especially severe and widespread, with household perceptions of short- and longer-term risks playing an important role (Koubi et al., 2016a).

Displacements due to droughts, extreme heat and associated impacts on food and water security are most frequent in east Africa and, to a lesser extent, south Asia and west and southern Africa (IDMC, 2020). Since droughts unfold progressively and typically do not cause permanent damage to housing or livelihood assets, there is greater opportunity for government and non-governmental organisation (NGO) interventions and greater use of in situ adaptation options (Koubi et al., 2016b; Koubi et al., 2016a; Cattaneo et al., 2019). Drought-related population movements are most common in dryland rural areas of low-income countries and occur after a threshold is crossed and in situ adaptation options are exhausted (Gautier et al., 2016; Wiederkehr et al., 2018; McLeman, 2017). Observed population movements may occur for an extended period after the event; one study of Mexican data found this lag to be up to 36 months (Nawrotzki et al., 2017). The most common response to drought is an increase in short-distance, rural–urban migration (medium confidence), with examples being documented in Bangladesh, Ethiopia, Pakistan, sub-Saharan Africa, Latin America and Brazil (Neumann and Hermans, 2015; Gautier et al., 2016; Gautier et al., 2016; Mastrorillo et al., 2016; Baez et al., 2017; Call et al., 2017; Nawrotzki et al., 2017; Jessoe et al., 2018 ; Carrico and Donato, 2019; Hermans and Garbe, 2019).

Few assessable studies were identified that examine links between wildfires and migration. Wildfire events are often associated with urgent evacuations and temporary relocations, which place significant stress on receiving communities (Spearing and Faust., 2020), but research in the USA suggests fires have only a modest influence on future migration patterns in exposed areas (Winkler and Rouleau., 2021). More research, particularly in other regions, is needed.

Immobility in the context of climatic risks can reflect vulnerability and lack of agency (an inability to migrate), but can also be a deliberate choice (high confidence). Research since AR5 shows that immobility is best described as a continuum from people who are financially or physically unable to move away from hazards (involuntary immobility) to people who choose not to move (voluntary immobility) because of strong attachments to place, culture and people (Nawrotzki and DeWaard, 2018; Adams, 2016; Farbotko and McMichael, 2019; Zickgraf, 2019; Neef et al., 2018; Suckall et al., 2017; Ayeb-Karlsson et al., 2018; Zickgraf, 2018; Mallick and Schanze, 2020). Involuntary immobility is associated with individuals and households with low adaptive capacity and high exposure to hazard, and can exacerbate inequality and future vulnerability to climate change (Sheller, 2018), including through impacts on health (Schwerdtle et al., 2018). Voluntary immobility represents an assertion of the importance of culture, livelihoods and people to well-being, and is of particular relevance for Indigenous Peoples (Suliman et al., 2019).

Planned relocations by governments of settlements and populations exposed to climatic hazards are not presently commonplace, although the need is expected to grow in coming decades (Hino et al 2017). Examples include relocations of coastal settlements exposed to storm and erosion hazards as well as smaller numbers of cases of flood-prone settlements in river valleys; these examples suggest that organised relocations are expensive, contentious, create multiple challenges for governments and generate short- and longer-term disruptions for the people involved (high agreement, medium evidence) (Ajibade et al., 2020; Henrique and Tschakert, 2020; Desai et al., 2021).

Examples of relocations of small indigenous communities in coastal Alaska and villages in the Solomon Islands and Fiji suggest that relocated people experience significant financial and emotional distress as cultural and spiritual bonds to place and livelihoods are disrupted (Albert et al., 2018; Neef et al., 2018; McMichael and Katonivualiku, 2020; McMichael and Katonivualiku, 2020; McMichael et al., 2021; Piggott-McKellar et al., 2019; Bertana, 2020). Voluntary relocation programmes offered by US state governments in communities damaged by Hurricane Sandy in 2012 have been subject to multiple studies, and these show longer-term economic outcomes, social connections and mental well-being vary for a range of reasons unrelated to the impacts of the hazard event itself (Bukvic and Owen, 2017; Binder et al., 2019; Koslov and Merdjanoff, 2021).

The number of assessable peer-reviewed studies that make connections between climate-related migration and health and well-being is small. The health outcomes of migrants generally, and of climate-migrants in particular, vary according to geographical context, country and the particular circumstances of migration or immobility (Hunter and Simon, 2017; Hunter et al., 2021; Schwerdtle et al., 2020). Such linkages are ‘multi-directional’, with studies suggesting that healthy individuals may be more likely to migrate internationally in search of economic opportunities than people in poorer health, except during adverse climatic conditions when migration rates may change across all groups, and that migrants may have different long-term health outcomes than people born in destination areas, potentially displaying a range of positive and negative health outcomes compared to non-migrants (Kennedy et al., 2015; Dodd et al., 2017; Hunter and Simon, 2017; Riosmena et al., 2017). Refugees and other involuntary migrants often experience higher exposure to disease and malnutrition, adverse indirect health effects of changes in diet or activity and increased rates of mental health concerns. These latter may be attributable to a sense of loss or fear (Schwerdtle et al., 2018; Torres and Casey, 2017) as well as due to the interruption of healthcare; occupational injuries; sleep deprivation; non-hygienic lodgings and insufficient sanitary facilities; heightened exposure to vector- and WBDs; vulnerability to psychosocial, sexual and reproductive issues; behavioural disorders; substance abuse; and violence (Farhat et al., 2018; Wickramage et al., 2018) (high confidence). Linkages between climate migration and the spread of infectious disease are bidirectional; migrants may be exposed to diseases at the destination to which they have lower immunity than the host community; in other cases, migrants could introduce diseases to the receiving community (McMichael, 2015). Thus, receiving areas may have to pay greater attention to building migrant sensitive health systems and services (Hunter and Simon, 2017). The risk of migration leading to disease transmission is exacerbated by weak governance and lack of policy to support public health measures and access to medicines (Pottie et al., 2015).

In AR5, conflict was addressed in WGII Chapter 12 on human security. The chapter concluded that some of the factors that increase the risk of violent conflict within states are sensitive to climate change (medium evidence, medium agreement ), that people living in places affected by violent conflict are particularly vulnerable to climate change (medium confidence)and that climate change will lead to new challenges to states and will increasingly shape both conditions of security and national security policies (medium evidence, medium agreement ). AR5 characterised a major debate within the field as: authors supporting an association between climate anomalies and conflict that can be extrapolated into the future (e.g., Hsiang et al. (2013); Hsiang and Marshall (2014); Burke et al. (2015a)) and authors arguing that these associations are not universal and break down when contextual, scale and political factors are introduced (e.g., Buhaug et al. (2014); Buhaug (2016)).

Consistent with AR5 findings, there continues to be little observed evidence that climatic variability or change cause violent inter-state conflict. In intra-state settings, climate change has been associated with the onset of conflict, civil unrest or riots in urban settings (high agreement, medium evidence) (Ide (2020), and changes in the duration and severity of existing violent conflicts (Koubi, 2019). Climate change is conceptualised as one of many factors that interact to raise tensions (Boas and Rothe, 2016) through diverse causal mechanisms (Mach et al., 2019; Ide et al., 2020) and as part of the peace-vulnerability-development nexus (Barnett, 2019; Abrahams, 2020; Buhaug and von Uexkull, 2021). New areas of the literature assessed in this report include the security implications of responses to climate change, the gendered dynamics of conflict and exposure to violence under climate change and civil unrest in urban settings. The impact of violent conflict on vulnerability is not addressed in this chapter but does arise in other chapters (Sections 8.3.2.3, 17.2.2.2). Other chapters address non-violent conflict over changing availability and distribution of resources, for example, competing land uses and fish stocks migrating to different territories (Sections 5.8.2.3; 5.8.3, 5.9.3, 5.13, 9.8.1.1, 9.8.5.1). A commonly used definition of armed conflict is conflicts involving greater than 25 battle-related deaths in a year; this number represents the Uppsala Conflict Data Program threshold for inclusion in their database, a core resource in this field.

Climatic conditions have affected armed conflict within countries, but their influence has been small compared to socioeconomic, political and cultural factors (Mach et al., 2019) (high agreement, medium evidence). Inter-group inequality, and consequent relative deprivation can lead to conflict and the negative impacts of climate change lower the opportunity cost of involvement in conflict (Buhaug et al., 2020; Vestby, 2019). Potential pathways linking climate and conflict include direct impacts on physiology from heat or resource scarcity; indirect impacts of climatic variability on economic output, agricultural incomes, higher food prices and increasing migration flows; and the unintended effects of climate mitigation and adaptation policies (Koubi, 2019; Busby, 2018; Sawas et al., 2018). Relative deprivation, political exclusion and ethnic fractionalisation and ethnic grievances are other key variables (Schleussner et al., 2016; Theisen, 2017). Research shows that factors such as land tenure and competing land uses interacting with market-driven pressures and existing ethnic divisions produce conflict over land resources rather than a scarcity of natural resources caused by climate impacts such as drought (high agreement, medium evidence) (Theisen, 2017; Balestri and Maggioni, 2017; Kuusaana and Bukari, 2015; Box 8.3).

Positive temperature anomalies and average increases in temperature over time have been associated with collective violent conflict in certain settings (medium agreement, low evidence). Helman and Zaitchik (2020) find statistical associations between temperature and violent conflict in Africa and the Middle East that are stronger in warmer places and identify seasonal temperature effects on violence. However, they are unable to detect the impact of regional temperature increases on violence. For Africa, Van Weezel (2019) found associations between average increases in temperature and conflict risk. Caruso et al. (2016) found an association between rises in minimum temperature and violence through the impact of temperature on rice yields (Box 9.4). However, the associations between temperature and violence are weak compared to those with political and social factors (e.g., Owain and Maslin (2018)) and research focuses on areas where conflict is already present and, as such, is sensitive to selection bias (Adams et al., 2018). There is a body of literature that finds statistical associations between temperature anomalies and inter-personal violence, crime and aggression in the Global North, predominantly in the USA (e.g., Ranson (2014); Mares and Moffett (2019); Tiihonen et al. (2017); Parks et al. (2020); Section 14.4.8). However, authors have cautioned against extrapolating seasonal associations into long-term trends and against focusing on individual crimes rather than wider social injustices associated with climate change and its impacts (Lynch et al., 2020).

Variation in availability of water has been associated with international political tension and intra-national collective violence (low agreement, medium evidence). Drought conditions have been associated with violence due to impacts on income from agriculture and water and food security, with studies focusing predominantly on sub-Saharan Africa and the Middle East (Ide and Frohlich, 2015; De Juan, 2015; Von Uexkull et al., 2016; Waha et al., 2017; Abbott et al., 2017; D’Odorico et al., 2018). A small set of published studies has argued inconclusively over the role of drought in causing the Syrian civil war (Gleick, 2014; Kelley et al., 2015; Selby et al., 2017; 16.2.3.9). In general, research stresses the underlying economic, social and political drivers of conflict. For example, research on conflict in the Lake Chad region has demonstrated that the lake drying was only one of many factors including lack of development and infrastructure (Okpara et al., 2016; Nagarajan et al., 2018; Tayimlong, 2020). Fewer studies examine the relationship between flooding (excess water) and violence and often rely on migration as the causal factor (see below). However, some studies have shown an association between flooding and civil unrest (Ide et al., 2020; Section 4.3.6; Section 12.5.3; Box 9.4).

Extreme weather events can be associated with increased conflict risk (low agreement, medium evidence). There is the potential for extreme weather events and disasters to cause political instability and increase the risk of violent conflict, although not conclusively (Brzoska, 2018). Post-disaster settings can be used to intensify state repression (Wood and Wright, 2016) and to alter insurgent groups’ behaviour (Walch, 2018). Different stakeholders use disasters to establish new narratives and alter public opinion (Venugopal and Yasir, 2017). Some research has demonstrated how post-disaster activities have had positive impacts on the social contract between people and the state, reducing the risk of conflict by strengthening relations between government and citizens and strengthening the citizenship of marginalised communities (Siddiqi, 2018; Pelling and Dill, 2010; Siddiqi, 2019). However, post-disaster and disaster risk-related activities themselves have limited capacity to support diplomatic efforts to build peace (Kelman et al., 2018).

Increases in food prices due to reduced agricultural production and global food price shocks are associated with conflict risk and represent a key pathway linking climate variability and conflict (medium confidence). Increases in food prices are associated with civil unrest in urban areas among populations unable to afford or produce their own food and in rural populations due to changes in availability of agricultural employment with shifting commodity prices (Martin-Shields and Stojetz, 2019). Under such conditions, locally specific grievances, hunger and social inequalities can initiate or exacerbate conflicts. Food price volatility in general is not associated with violence, but sudden food price hikes have been linked to civil unrest in some circumstances (Bellemare, 2015; McGuirk and Burke, 2020; Winne and Peersman, 2019). In urban settings in Kenya, Koren et al. (2021) found an association between food and water insecurity that is mutually reinforcing and associated with social unrest (although insecurity in either food or water on its own was not). Analysing the global food riots in 2007/2008 and 2011, Heslin (2021) stresses the role of local politics and pre-existing grievances in determining whether people mobilise around food insecurity (Chapter 5).

Climate-related internal migration has been associated with the experience of violence by migrants, the prolongation of conflicts in migrant receiving areas and civil unrest in urban areas (medium agreement, low evidence). Research points to the potential for conflict to serve as an intervening factor between climate and migration. However, the nature of the relationship is diverse and context specific. For example, displaced people and migrants may be associated with heightened social tensions in receiving areas through mechanisms such as ecological degradation, reduced access to services and a disturbed demographic balance in the host area (Rüegger and Bohnet, 2020). Ghimire et al. (2015) observed that an influx of flood-displaced people prolonged conflict by causing a lack of access to services for some of the host population and feelings of grievance. Further, migration from drought-stricken areas to local urban centres has been used to suggest a climate trigger for the Syrian conflict (e.g., Ash and Obradovich (2020)). However, this link has been strongly contested by research that contextualises the drought in wider political economic approaches and existing migration patterns (De Châtel, 2014; Fröhlich, 2016; Selby, 2019; 16.2.3.9).

There is some evidence of an association between climate-related rural-to-urban migration and the risk of civil unrest (medium agreement, low evidence). Petrova (2021) found that while migration in general was associated with increased protests in urban receiving areas, the relationship did not hold for hazard-related migration. In other settings, the association of civil unrest with in-migration was found to depend on the political alignment of the host state with the capital (Bhavnani and Lacina, 2015), previous experience of extreme climate hazards (Koubi et al., 2021) and previous experience of violence among migrants (Linke et al., 2018). Climate-related migrants have reported higher levels of perception and experience of violence in their destination (Linke et al., 2018; Koubi et al., 2018). There has been no association established between international migration and conflict. The literature highlights how unjust racial logics may generate spurious links between climate migration and security (Fröhlich, 2016; Telford, 2018).

Structural inequalities play out at an individual level to create gendered experiences of violence (high agreement, medium evidence). Violent conflict is experienced differently by men and women because of gender norms that already exist in society and shape vulnerabilities. For example, conflict deepens gendered vulnerabilities to climate change related to unequal access to land and livelihood opportunities (Chandra et al., 2017). Motivations for inter-group violence may be influenced by constructions of masculinity, for example the responsibility to secure their family’s survival or pay dowries (Myrttinen et al., 2017), and gendered roles may incentivise young men to protest or to join non-state armed groups during periods of adverse climate (Myrttinen et al., 2015; Myrttinen et al., 2017; Anwar et al., 2019; Hendrix and Haggard, 2015; Koren and Bagozzi, 2017). Research has found a positive correlation between crop failures and suicides by male farmers who could not adapt their livelihoods to rising temperatures (Bryant and Garnham 2015; Kennedy and King, 2014; Carleton, 2017).

Extreme weather and climate impacts are associated with increased violence against women, girls and vulnerable groups (high agreement, medium evidence). During and after extreme weather events, women, girls and LGBTQI people are at increased risk of domestic violence, harassment, sexual violence and trafficking (Le Masson et al., 2019; Nguyen, 2019;Myrttinen et al., 2015; Chindarkar, 2012). For example, early marriage is used as a coping strategy for managing the effects of extreme weather events (Ahmed et al., 2019) and women are exposed to increase risk of harassment and sexual assault as scarcity and gender-based roles cause them to walk longer distances to fetch water and fuel (Le Masson et al., 2019). Within the household, violent backlash or heightened tensions may arise from changing gender norms as men migrate to find work in post-disaster settings (Stork et al., 2015) and men’s use of negative coping mechanisms, such as alcoholism, when unable to meet norms of providing for the household (Anwar et al., 2019; Stork et al., 2015). Rates of intimate partner violence have been found to increase with higher temperatures (Sanz-Barbero et al., 2018).

Climate adaptation and mitigation projects implemented without taking local interests and dynamics into account have the potential to cause conflict (high agreement, medium evidence). Reforestation or forest management programmes driven by reducing emissions through deforestation, land zoning and managed retreat due to sea level rise have been identified as having the potential to cause friction and conflict within and between groups and communities (de la Vega-Leinert et al., 2018; Froese and Schilling, 2019). Conflict may arise when there is resistance to a proposed project, where interventions favour one group over another, or when projects undermine livelihoods or displace populations (e.g., Nightingale (2017); Sovacool et al. (2015); Sovacool (2018); Corbera (2017); Hunsberger (2018); Sections 4.6.8, 5.13.4, 14.4.7.3). In addition to conflict generated by the poor implementation of land-based climate mitigation and adaptation projects, Gilmore and Buhaug (2021) highlight the links between climate policy and conflict through the potential effects of unequal distribution of economic burdens and fossil fuel markets on economic growth. There is a small literature that draws attention to the potential security of nuclear proliferation, if nuclear energy is increasingly employed as a low-carbon energy source (e.g., Parthemore et al. (2018); Bunn, (2019)).

Economic and social changes due to changes in sea ice extent in the Arctic are anticipated to be managed as part of existing governance structures (high agreement, medium evidence). The opening-up of the Arctic and associated geopolitical manoeuvring for access to shipping routes and sub-sea hydrocarbons is often highlighted as a potential source of climate conflict (e.g., Koivurova (2009); Åtland (2013); Tamnes and Offerdal (2014)). Research assessed in AR5 focused on the potential for resource wars and Arctic land grabs. However, research since AR5 is less sensationalist in its approach to Arctic security, focusing instead on the practicalities of polycentric Arctic governance under climate change, the economic impacts of climate change, protecting the human security of Arctic populations whose autonomy is at risk (Heininen and Exner-Pirot, 2020), understanding how different regions (e.g., the EU) are positioning themselves more prominently in the Arctic space (Raspotnik and Østhagen, 2019) and Arctic Indigenous Peoples’ understanding of security (Hossain, 2016; Chapter 3; Chapter 14; CCP6).

Climate change is expected to significantly increase the health risks resulting from a range of climate-sensitive diseases and conditions, with the scale of impacts depending on emissions and adaptation pathways in coming decades (very high confidence). Sections 7.3.1.2 to 7.3.1.11 assess the available studies on future projections for risks associated with specific climate-sensitive diseases and conditions previously described in Section 7.2.1. In the case of diabetes, cancer, injuries, mosquito-borne diseases other than dengue and malaria, rodent-borne diseases and most mental illnesses, insufficient literature was found to allow for assessment. Adaptation pathways and options for managing such risks are detailed in Section 7.4.

Even in the absence of further warming beyond current levels, the proportion of the overall global deaths caused by climate-sensitive diseases and conditions would increase marginally by mid-century (high confidence). Two global projections of climate change health impacts have been conducted since AR5. The first focused on cause-specific mortality for eight exposures for 2030 and 2050 for a mid-range emissions scenario (A1b) and three scenarios of economic growth (WHO, 2014). The study estimated that the climate change projected to occur by 2050 (compared to 1961–1990) could result in an excess of approximately 250,000 deaths yr –1, dominated by increases in deaths due to heat (94,000, mainly in Asia and high-income countries), childhood undernutrition (85,000, mainly in Africa but also in Asia), malaria (33,000, mainly in Africa) and diarrhoeal disease (33,000, mainly in Africa and Asia). Overall, more than half of this excess mortality is projected for Africa. Near-term projections (for 2030) are predominantly for childhood undernutrition (95,200 out of 241,000 total excess deaths) (Figure 7.8). The second study (Carleton et al. 2020) focused on all-cause mortality associated with warming under both a high emissions scenario (RCP8.5) and a middle emissions scenario (RCP4.5). The analyses created a metric of death equivalents that accounted for hot and cold temperature-related mortality and the costs of individual level adaptation; no acclimatization or community-level adaptation, such as early warning systems, were incorporated. Average annual temperature-mortality-income per capita relationships estimated from pooled data from 40 predominantly middle- and high-income countries (38% of the world population) were applied worldwide. Under the high emissions scenario, climate change was projected to result in approximately 85 deaths equivalents per 100,000 population.

Temperature increases are projected to exceed critical risk thresholds for six key climate-sensitive health outcomes, highlighting the criticality of building adaptive capacity in health systems and in other sectors that influence health and well-being (high confidence). Recently reported research illustrates the temperature thresholds under three adaptation scenarios describing the effectiveness of health systems to manage additional risks from climate change for heat-related morbidity and mortality; ozone-related mortality; malaria incidence rates; incidence rates of Dengue and other diseases spread by Aedes sp. mosquitos; Lyme disease; and West Nile fever (Ebi et al., 2021a). As shown in Figure 7.9, these adaptation scenarios significantly alter the warming thresholds at which risks accelerate, with the proactive adaptation scenario, a scenario that emphasises international cooperation towards achieving sustainable development, having the greatest potential to avoid significant increases in risks under all but the highest levels of warming. The incomplete adaptation scenario describes a world with moderate challenges to adaptation and mitigation. The limited adaptation scenario describes a world with high challenges to adaptation and mitigation. In the figure, transitions are based on the peer-reviewed literature projecting risks for each of the health outcomes. Projections for time intervals were changed to temperature increase above pre-industrial levels based on the climate models and scenarios used in the projections.The assessed projections were based on a range of scenarios, including SRES, CMIP5, and ISIMIP, and, in some cases, demographic trends. The black dots are levels of confidence, from very high (four dots) to low (one dot). The diagrams for the proactive and incomplete adaptation scenarios are truncated at the nearest whole °C within the range of temperature change in 2100 under three SSP scenarios used in panel (a) of SPM.3.

This section considers the broad impacts of projected changes in heat- and cold-related exposure and related outcomes including mortality and work productivity. Several of the most common heat- and cold-related specific health outcomes (e.g., CVD) are assessed individually in later sections of this chapter.

Population heat exposure will increase under climate change (very high confidence). Since AR5 there has been considerable progress with quantifying future human exposure to extreme heat (Schwingshackl et al., 2021), especially as determined by different combinations of SSPs and RCPs (Chambers, 2020; Cheng et al., 2020; Jones et al., 2018; Liu et al., 2017; Ma and Yuan, 2021; Russo et al., 2019). For example, Table 7.1 shows projections of population exposure to heatwaves, as expressed by the number of person-days, for the 2061–2080 period aggregated by geographical region and SSP/RCP. At the global level, projected future exposure increases from approximately 15 million person-days for the current period to 535 billion person-days for high population growth under the high GHG emission SSP3-RCP8.5 scenario, while for the low population growth/high urbanisation and business as usual SSP5-RCP4.5 scenario, the exposure is substantially lower at 170 billion person-days. Spatial variations in future heatwave frequency and population growth play out in the form of significant geographical contrasts in exposure, with the largest increases projected for low latitude regions such as India and significant portions of sub-Saharan Africa, where increases in heatwave frequency and population are expected. Over East Asia and especially eastern China, exposures are projected to rise, with the effect of increases in heatwave frequency exceeding the countering effect of projected reductions in population, especially in non-urban areas. Further, for North America and Europe, where rural depopulation is projected, the predominant driver of increases in exposure is urban growth (Jones et al., 2018).

Table 7.1 | Projected exposure to heatwaves in millions of person-days by region under different SSP/RCP combinations.

Comparisons of heatwave exposure for 1.5°C and 2.0°C warming for different SSPs indicate strong geographical contrasts in potential heatwave risk (high confidence). One global level assessment for a 1.5°C warming projects that low human development index countries will experience exposure levels equal to or greater than the exposure levels for very high human development index countries under a 2°C warming (Russo, 2019). The same assessment also finds that holding global warming below 1.5°C in tandem with achieving sustainable socioeconomic development (e.g., SSP1 as opposed to SSP4) yields reduced levels of heatwave exposure, especially for low human development index countries, particularly across sub-Saharan Africa. Similar findings were found in other global level assessments. Global exposure to extreme heat increases almost 30 times under a SSP3-8.5 combination, with the average exposure for Africa 118 times greater than historical levels, in stark contrast to the four-fold increase projected for Europe. Compared to a SSP3-8.5 scenario, exposure was reduced by 65% and 85% under the SSP2-4.5 and SSP1-2.6 scenarios, respectively (Liu et al., 2017).

Regional level assessments of changes in population heat exposure for Africa, Europe, the USA, China and India corroborate general findings at the global level, that the impact of warming is amplified under divergent regional development pathways (e.g., SSP4 – inequality) compared to those fostering sustainable development (e.g., SSP1 – sustainability) (high confidence) (Rohat et al., 2019a; Weber et al., 2020; Broadbent et al., 2020; Dahl et al., 2019; Harrington and Otto, 2018; Rohat et al., 2019b; Vahmani et al., 2019; Huang and et al., 2018; Zhang et al., 2020a; Liu et al., 2017). For some regions, such as Europe, changes in exposure are projected to be largely a consequence of climate change, while for others, such as Africa and to a lesser extent Asia, Oceania, North America and South America, the interactive effects of demographic and climate change are projected to be important (Jones et al., 2018; Liu et al., 2017; Russo et al., 2016; Ma and Yuan, 2021) (medium confidence).

Compared to research that estimates the temperature only impacts of climate change on heat-related mortality (see below), the number of studies that explicitly model mortality responses considering various combinations of SSPs and RCPs is small and mostly restricted to the country or regional level. These studies point to increases in heat-related mortality especially amongst the elderly across a range of SSPs, with the greatest increases under SSP5 and RCP8.5 (Rail et al., 2019; Yang et al., 2021).

Estimates of heat-related mortality based solely on changes in temperature point to elevated levels of global and regional level mortality compared to the present, with the magnitude of this increasing from RCP4.5 through to RCP8.5 (high confidence) (Ahmadalipour and Moradkhani, 2018; Cheng et al., 2019; Kendrovski et al., 2017; Lee et al., 2020; Limaye et al., 2018; Morefield et al., 2018). Further support comes from the projection that heat-related health impacts for a 2°C increase in global temperatures will be greater than those for 1.5°C warming (very high confidence) (Dosio et al., 2018; Mitchell et al., 2018; King and Karoly, 2017; Vicedo-Cabrera et al., 2018a).

Estimates of future mortality that incorporate adaptation in addition to temperature change point to increases in heat-related mortality under global warming, albeit at lower levels than the case of no adaptation (high confidence) (Anderson et al., 2018; Gosling et al., 2017; Guo et al., 2018; Honda and Onozuka, 2020; Vicedo-Cabrera et al., 2018b; Wang et al., 2018b). Whether adaptation is considered or not, the consensus is Central and South America, southern Europe, southern and Southeast Asia and Africa will be the most affected by climate change in terms of heat-related mortality (high confidence). Similarly, projections of the impacts of future heat on occupational health, worker productivity and workability point to these regions as problematic under climate change (high confidence) (Andrews et al., 2018; de Lima et al., 2021; Dillender, 2021; Kjellstrom et al., 2018; Orlov et al., 2020; Rao et al., 2020; Tigchelaar et al., 2020), especially for occupations with high exposure to heat, such as agriculture and construction. This accords with the findings from independent projections of population heat exposure as outlined above (high confidence).

The effect of climate change on productivity is projected to reduce GDP at a range of geographical scales (high confidence) (Borg et al., 2021; Oppermann et al., 2021; Orlov et al., 2020). For example, measuring economic costs using occupational health and safety recommendations, it was estimated that RCP8.5 would result in a 2.4% reduction in global GDP compared to a 0.5% reduction under RCP2.6 (Orlov et al., 2020). For the USA, it was estimated that the total hours of labour supplied declined ∼0.11% (±0.004%) per degree Celsius increase in global mean surface temperature for low-risk workers and 0.53% (±0.01%) per degree Celsius increase for high-risk workers exposed to outdoor temperatures (Hsiang et al., 2017). Further, a systematic review of the literature indicates that extreme heat exacts a substantial economic burden on health systems, which bears implications for future heat-attributable healthcare costs (Wondmagegn et al., 2019).

Since AR5, there has been an increase in the understanding of the extent to which a warming world is likely to affect cold- or winter-related health impacts. Future increases in heat-related deaths are expected to outweigh those related to cold (high confidence) (Aboubakri et al., 2020; Achebak et al., 2020; Burkart et al., 2021; Huber et al., 2020b; Martinez et al., 2018; Rodrigues et al., 2020; Vardoulakis et al., 2014; Weinberger et al., 2017; Weinberger et al., 2018a; Weitensfelder and Moshammer, 2020). However, strong regional contrasts in heat- and cold-related mortality trends are likely under a RCP8.5 scenario, with countries in the Global North experiencing minimal to moderate decreases in cold-related mortality while warm climate countries in the Global South are projected to experience increases in heat-attributable deaths by the end of the century (Gasparrini et al., 2017; Burkart et al., 2021). Projections of the magnitude of change in the temperature-related burden of disease do, however, demonstrate great variability, due to the application of a wide range of climate change, adaptation and demographic scenarios (Cheng et al., 2019).

A particular focus since AR5 has been the impact of climate change on cities (see AR6 Chapter 6). Heat risks are expected to be greater in urban areas due to changes in regional heat exacerbated by ‘heat island’ effects (high confidence) (Doan and Kusaka, 2018; Heaviside et al., 2016; Li et al., 2021; Rohat et al., 2019a; Rohat et al., 2019c; Varquez et al., 2020; Wouters et al., 2017; Zhao et al., 2021), with intra-urban scale variations in heat exposure attributable to land cover contrasts and urban form and function (Avashia et al., 2021; Jang et al., 2020; Macintyre et al., 2018; Schinasi et al., 2018). However, further research is required to establish the health implications of increasing chronic slow-onset extreme heat (Oppermann et al., 2021) in addition to the acute health outcomes of UHI–heatwave synergies under climate change. The latter is particularly important as studies that address UHI–heatwave interactions have mainly focused on changes in UHI intensity (e.g., Ramamurthy and Bou-Zeid (2017); Scott et al. (2018)). Whether significant urban mortality anomalies arise from the interplay of heatwaves and UHIs largely remains an open question although at least one study demonstrated higher urban compared to rural mortality rates during heatwaves (Ruuhela et al., 2021). The benefits of the winter UHI effect for cold-related mortality remain largely unexplored, but one study for Birmingham, UK, indicates the winter UHI will continue to have a protective effect in future climate (Macintyre et al., 2021).

The distribution and abundance of disease vectors, and the transmission of the infections that they carry, are influenced both by changes in climate and by trends such as human population growth and migration, urbanisation, land use change, biodiversity loss and public health measures. Each of these may increase or decrease risk, interact with climate effects and may contribute to the emergence of infectious disease, although there are few studies assessing future risk of emergence (Gibb et al., 2020). Unless stated otherwise, the assessments below are specifically for the effects of climate change on individual diseases, assuming other determinants remain constant.

There is a high likelihood that climate change will contribute to increased distributional range and vectorial capacity of malaria vectors in parts of sub-Saharan Africa, Asia and South America (high confidence). In Nigeria, the range and abundance of Anopheles mosquitoes are projected to increase under both lower (RCP2.6) and especially under higher emissions scenarios (RCP8.5) due to increasing and fluctuating temperature, longer tropical rainfall seasons and rapid land use changes (Akpan et al., 2018). Similarly, vegetation acclimation due to elevated atmospheric CO2 under climate change will likely increase the abundance of Anopheles vectors in Kenya (Le et al., 2019). Distribution of Anopheles may decrease in parts of India and Southeast Asia, but there is an expected increase in vectorial capacity in China (Khormi and Kumar, 2016). In South America, climate change is projected to expand the distributions of malaria vectors to 35–46% of the continent by 2070, particularly species of the Albitarsis complex (Laporta et al., 2015).

Malaria infections have significant potential to increase in parts of sub-Saharan Africa and Asia, with risk varying according to the warming scenario (medium confidence). In Africa, where most malaria is due to the more deadly Plasmodium falciparum parasite, climate change is likely to increase the overall transmission risk due to the likely expansion of vector distribution and increase in biting rates (Bouma et al., 2016; M’Bra et al., 2018; Nkumama et al., 2017; Ryan et al., 2015b; Tompkins and Caporaso, 2016a). The projected effect of climate change varies markedly by region, with projections for west Africa tending to indicate a shortening of transmission seasons and neutral or small net reductions in overall risk, whereas studies consistently project increases in southern and eastern Africa, with potentially an additional 76 million people at risk of endemic exposure (10–12 months yr –1) by the 2080s (Nkumama et al., 2017; Ryan et al., 2015b; Semakula et al., 2017; Zaitchik, 2019; Leedale et al., 2016; Murdock et al., 2016; Yamana et al., 2016; Ryan et al., 2020). In sub-Saharan Africa, malaria case incidence associated with dams in malaria-endemic regions will likely be exacerbated by climate change, with significantly higher rates projected under RCP8.5 in comparison to lower-emission scenarios (Kibret et al., 2016). Incidence of malaria in Madagascar is projected to increase under RCP4.5 through RCP8.5 (Rakotoarison et al., 2018). Distribution of P. vivax and P. falciparum malaria in China is likely to increase under RCPs higher than 2.6, especially RCP8.5 (Hundessa et al., 2018). In India, projected scenarios for the 2030s under RCP4.5 indicate changes in the spatial distribution of malaria, with new foci and potential outbreaks in the Himalayan region, southern and eastern states, and an overall increase in months suitable for transmission overall, with some other areas experiencing a reduction in transmission months (Sarkar et al., 2019).

Rising temperatures are likely to cause poleward shifts and overall expansion in the distribution of mosquitoesAedes aegypti andAedes albopictus , the principal vectors of dengue, yellow fever, chikungunya and Zika (high confidence). Globally, the population exposed to disease transmission by one of these vectors is expected to increase significantly due to the combination of climate change and non-climatic processes including urbanisation and socioeconomic inter-connectivity, with exposure rates rising under higher warming scenarios (Kamal et al., 2018; Kraemer et al., 2019). For example, approximately 50% of the global population is projected to be exposed to these vectors by 2050 under RCP6.0 (Kraemer et al., 2019). The effect of climate change alone is projected to increase the population exposed to Aedes aegypti by 8–12% by 2061–2080 (Monaghan et al., 2018), and its abundance is projected to increase by 20% under RCP2.6 and 30% under RCP8.5 by the end of the century (Liu-Helmersson et al., 2019; Figure 7.10). Exposure to transmission by Aedes albopictus specifically would be highest at intermediate climate change scenarios and would decrease in the warmest scenarios (Ryan et al., 2019). Under scenarios other than RCP2.6, most of Europe would experience significant increases in exposure to viruses transmitted by both vectors (Liu-Helmersson et al., 2019).

Climate change is expected to increase dengue risk and facilitate its global spread, with the risk being greatest under high emissions scenarios (high confidence). Future exposure to risk will be influenced by the combined effects of climate change and non-climatic factors such as population density and economic development (Akter et al., 2017). Overall, risk levels are expected to rise on all continents (Akter et al., 2017; Messina et al., 2015; Rogers, 2015; Liu-Helmersson et al., 2016; Messina et al., 2019). Compared to 2015, an additional 1 billion people are projected to be at risk of dengue exposure by 2080 under an SSP1-4.5 scenario, 2.25 billion under SSP2-6.0, and 5 billion under SSP3-8.5 (Messina et al., 2019). In North America, risk is projected to expand in north-central Mexico, with annual dengue incidence in Mexico increasing by up to 40% by 2080, and expand from US southern states to mid-western regions (Proestos et al., 2015; Colon-Gonzalez et al., 2013). In China, under RCP8.5, dengue exposure would increase from 168 million people in 142 counties to 490 million people in 456 counties by the late 2100s (Fan and Liu, 2019). In Nepal, dengue fever is expected to expand throughout the 2050s and 2070s under all RCPs (Acharya et al., 2018). In Tanzania, there is a projected shift in distribution towards central and northeastern areas and risk intensification in nearly all parts of the country by 2050 (Mweya et al., 2016). Dengue vectorial capacity is projected to increase in Korea under higher RCP scenarios (Lee et al., 2018a).

There are insufficient studies for assessment of projected effects of climate change on other arboviral diseases, such as chikungunya and Zika. Zika virus transmits under different temperature optimums than does dengue, suggesting environmental suitability for Zika transmission could expand with future warming (low confidence) (Tesla et al., 2018).

Climate change can be expected to continue to contribute to the geographical spread of the Lyme disease vectorIxodes scapularis (high confidence)and the spread of tick-borne encephalitis and Lyme disease vectorIxodes ricinus in Europe (medium confidence). In Canada, vector surveillance of the black-legged tick I. scapularis identified strong temperature effects on the limits of their occurrence, on recent geographic spread, temporal coincidence in emergence of tick populations and acceleration of the speed of spread (Clow et al., 2017; Cheng et al., 2017). In Europe, increasing temperatures over the 1950–2018 period significantly accelerated the life cycle of Ixodes ricinus and contributed to its spread (Estrada-Peña and Fernández-Ruiz, 2020). Under RCP4.5 and RCP8.5 scenarios, projections indicate a northward and eastward shift of the distribution of I. persulcatus and I. ricinus, vectors of Lyme disease and tick-borne encephalitis in northern Europe and Russia, with an overall large increase in distribution in the second half of the current century (Popov and Yasyukevich, 2014; Yasjukevich et al., 2018) and increases in intensity of tick-borne encephalitis transmission in central Europe (Nah et al., 2020).

Climate change is projected to increase the incidence of Lyme disease and tick-borne encephalitis in the Northern Hemisphere (high confidence) (Figure 7.9). The basic reproduction number (R0) of I. scapularis in at least some regions of Canada is projected to increase under all RCP scenarios (McPherson et al., 2017). In the USA, a 2°C warming could increase the number of Lyme disease cases by over 20% over the coming decades and lead to an earlier onset and longer length of the annual Lyme disease season (Dumic and Severnini, 2018; Monaghan et al., 2015).

Climate change is projected to change the distribution of schistosomiasis in Africa and Asia (high confidence), with a possible increase in global land area suitable for transmission (medium confidence). A global increase in land area with temperatures suitable for transmission by the three main species of Schistosoma (S. japonicum, S. mansoni and S. haematobium) is projected under the RCP4.5 scenario for the 2021–2050 and 2071–2100 periods (Yang and Bergquist, 2018), but regional outcomes are expected to vary. In Africa, shifting temperature regimes associated with climate change are expected to lead to reduced snail populations in areas with already high temperatures and higher populations in areas with currently low winter temperatures (Kalinda et al., 2017; McCreesh and Booth, 2014). Infection risk with Schistosoma mansoni may increase by up to 20% over most of eastern Africa over the next 20–50 years but decrease by more than 50% in parts of north and east Kenya, southern South Sudan and eastern People’s Democratic Republic of Congo (PDRC) (McCreesh et al., 2015), with a possible overall net contraction (Stensgaard et al., 2013). In China, currently endemic areas in Sichuan Province may become unsuitable for snail habitats, but currently non-endemic areas in Sichuan and Hunan/Hubei provinces may see a new emergence (Yang and Bergquist, 2018). In addition to the projected effects of temperature described above, distribution and transmission of schistosomiasis will also be affected positively or negatively by changes in the availability of freshwater bodies, which were not included in these models.

Climate change will contribute to additional deaths and mortality due to diarrhoeal diseases in the absence of adaptation (medium confidence) (see Figure 7.8). Risk factors for future excess deaths due to diarrhoeal diseases are highly mediated by future levels of socioeconomic development and adaptation. An additional 1°C increase in mean average temperature is expected to result in a 7% (95% CI, 3–10%) increase in all-cause diarrhoea (Carlton et al., 2016), an 8% (95% CI, 5–11%) increase in the incidence of diarrheic E. coli (Philipsborn et al., 2016) and a 3–11% increase in deaths attributable to diarrhoea (WHO, 2014). WHO Quantitative Risk Assessments for the effects of climate change on selected causes of death for the 2030s and 2050s project that overall deaths from diarrhoea should fall due to socioeconomic development but that the effect of climate change under higher emission scenarios could cause an additional 48,000 deaths in children aged under 15 years in 2030 and 33,000 deaths for 2050, particularly in Africa and parts of Asia. In Ecuador, projected increases in rainfall variability and heavy rainfall events may increase diarrhoea burden in urban regions (Deshpande et al., 2020). A limit in the assessable literature is a lack of studies in the highest risk areas (Liang and Gong, 2017; UNEP, 2018).

Climate change is expected to increase future health risks associated with a range of other WBDs and parasites, with effects varying by region (medium confidence). WBDs attributable to protozoan parasites including Cryptosporidium spp. and Giardia duodenalis (intestinalis) are expected to increase in Africa due to increasing temperatures and drought (Ahmed et al., 2018; Efstratiou et al., 2017). Recent data suggest a poleward expansion of Vibrios to areas with no previous incidence, particularly in mid- to high-latitude regions in areas where rapid warming is taking place (Baker-Austin et al., 2017). The number of Vibrio-induced diarrhoea cases yr –1 increased in past decades in the Baltic Sea region, and the projected risk of vibriosis will increase in northern areas, where waters are expected to become warmer and more saline due to reduced precipitation and have higher chlorophyll concentrations (Escobar et al., 2015; Semenza et al., 2017).

The risk of Campylobacteriosis and other enteric pathogens could rise in regions where heavy precipitation events or flooding are projected to increase (medium confidence). In Europe, the risk of Campylobacteriosis and diseases caused by other enteric pathogens could rise in regions where precipitation or extreme flooding are projected to increase (European Environment Agency, 2017), although incidence rates may be further mediated by seasonal social activities (Rushton et al., 2019; Williams et al., 2015b). Accelerated releases of dissolved organic matter to inland and coastal waters through increases in precipitation are expected to reduce the potential for solar ultraviolet inactivation of pathogens and increase risks for associated WBDs (Williamson et al., 2017). The combined relative risk for waterborne campylobacteriosis, salmonellosis and diseases due to Verotoxin-producing Escherichia coli was estimated to be 1.1 (i.e., a 10% increase) for every 1°C in mean annual temperature, while by the 2080s, under RCP8.5, annual rates of cryptosporidiosis and giardiasis could rise by approximately 16% due to more severe precipitation events (Brubacher et al., 2020; Chhetri et al., 2019).

The prevalence of Salmonella infections is expected to rise as higher temperatures enable more rapid replication (medium confidence). Research from Canada finds a very strong association of salmonellosis and other FBDs with higher temperatures, suggesting that climate change could increase food safety risks ranging from increased public health burden to emergent risks not currently seen in the food chain (Smith and Fazil, 2019). In Europe, the average annual number of temperature-related cases of salmonellosis under high emissions scenarios could increase by up to 50% more than would be expected on the basis of on population change alone by 2100 (Lake, 2017; European Environment Agency, 2017). Warming trends in the southern USA may lead to increased rates of Salmonella infections (Akil et al., 2014).

Global air pollution-related mortality attributable directly to climate change—the human health climate penalty associated with climate-induced changes in air quality—is likely to increase and partially counteract any decreases in air pollution-related mortality achieved through ambitious emission reduction scenarios or stabilisation of global temperature change at 2°C (medium confidence). Demographic trends in aging and more vulnerable population are likely to be important determinants of future air quality—a human health climate penalty (high confidence).

Poor air quality contributes to a range of NCDs, including cardiovascular, respiratory and neurological, commonly resulting in hospitalisation or death. This section considers the possible risks for health of future climate-related changes in ozone and PM. The climate penalty, the degree to which global warming could affect future air quality, is better understood for ozone than for PM (von Schneidemesser et al., 2020). This is because increases in air temperature enhance ozone formation via associated photochemical processes (Archibald et al., 2020; Fu and Tian, 2019). The association between climate and PM is complex and moderated by a diverse range of PM components as well as formation and removal mechanisms (von Schneidemesser et al., 2020), added to which is uncertainty about future climate-related PM sources such as wildfires (Ford et al., 2018) and changes in aridity (Achakulwisut et al., 2019). As noted in AR6 WGI Chapter 6 (Naik et al 2021), future air quality will largely depend on precursor emissions, with climate change projected to have mixed effects. Because of the uncertainty in how natural processes will respond, there is low confidence in the projections of surface ozone and PM under climate change (Naik et al., 2021). This has implications on the levels of confidence in the projections of the health climate penalty associated with climate-induced changes in air quality (Orru et al., 2017; Orru et al., 2019; Silva et al., 2017).

There is a rich literature on global and regional level projections of air quality-related health effects arising from changes in emissions. Comparatively few studies assess how changes in air pollution directly attributable to climate change are likely to affect future mortality levels. Projections indicate that emission reduction scenarios consistent with stabilisation of global temperature change at 2°C or below would yield substantial co-benefits for air quality-related health outcomes (Chowdhury et al., 2018b; von Schneidemesser et al., 2020; Silva et al., 2016c; Markandya et al., 2018; Orru et al., 2019; Shindell et al., 2018) (high confidence). For example, by 2030, compared to 2000, it was estimated that globally and annually 289,000 PM2.5-related premature deaths could have been avoided under RCP4.5 compared to 17,200 PM2.5-related excess premature deaths under RCP8.5 (Silva et al., 2016c). Further, and notwithstanding estimated reductions in global PM2.5 levels and an associated increase in the number of avoidable deaths, the benefits of following a low emissions pathway are expected to be apparent by 2100, with avoidable deaths estimated at 2.39 million deaths yr –1 under RCP4.5. This contrasts with the 1.31 million deaths estimated under RCP8.5. A few projections of the health-related climate penalty indicate a possible increase in ozone and PM2.5-associated mortality under RCP8.5 (Doherty et al., 2017; Orru et al., 2019; Silva et al., 2017).

At the global level for PM2.5, annual premature deaths due to climate change were projected to be 55,600 (−34,300 to 164,000) and 215,000 (−76,100 to 595,000) in 2030 and 2100, respectively, countering by 16% the projected decline in PM2.5-related mortality between 2000 and 2100 without climate change (Silva et al., 2017). Similarly for ozone, the number of annual premature ozone-related deaths due to climate change was projected to be 3,340 in 2030 and 43,600 in 2050, with climate change accounting for 1.2% (14%) of the annual premature deaths in 2030 (2100) (Silva et al., 2017). These global level projections average over considerable geographical variations (Silva et al., 2017). Projections of the climate change effect on ozone mortality in 2100 were greatest for East Asia (41 deaths yr –1 per million people), India (8 deaths yr –1 per million people) and North America (13 deaths yr –1 per million people). For PM2.5, mortality was projected to increase across all regions except Africa (−25,200 deaths yr –1 per million people) by 2100, with estimated increases greatest for India (40 deaths yr –1 per million people), the Middle East (45 deaths yr –1 per million people), East Asia (43 deaths yr –1 per million people) and the Former Soviet Union (57 deaths yr –1 per million people). Overall, higher ozone-related health burdens were projected to occur in highly populated regions, and greater PM2.5 health burdens were projected in high PM emission regions (Doherty et al., 2017).

For central and southern Europe, climate change alone could result in an 11% increase in ozone-associated mortality by 2050. However, projected declines in ozone precursor emissions could reduce the EU-wide climate change effect on ozone-related mortality by up to 30%; the reduction was projected to be approximately 24% if aging and an increasingly susceptible population were accounted for in projections to 2050 (Orru et al., 2019). For the USA in 2069, the impact of climate change alone on annual PM2.5- and ozone-related deaths was estimated to be 13,000 and 3,000 deaths, respectively, with heat-driven adaptation of air conditioning accounting for 645 and 315 of the PM2.5- and ozone-related annual excess deaths, respectively (Abel et al., 2018). An aging population is a determinant of future air quality-related mortality levels. An aging population along with an increase in the number of vulnerable people may work to offset the decrease in deaths associated with a low emission pathway (RCP4.5) and possibly dominate the net increase in deaths under a business as usual pathway (RCP8.5) (Chen et al., 2020; Doherty et al., 2017; Hong et al., 2019; Schucht et al., 2015).

Complementing the longer-term changes in air quality arising from climate change are those associated with air pollution sensitive short-term meteorological events, such as heatwaves. Studies of individual heat events (Garrido-Perez et al., 2019; Johansson et al., 2020; Kalisa et al., 2018; Pu et al., 2017; Pyrgou et al., 2018; Schnell and Prather, 2017; Varotsos et al., 2019) and systematic reviews (Anenberg et al., 2020) provide evidence for synergistic effects of heat and air pollution. However, the health consequences of a possible additive effect of air pollutants during heatwave events were heterogeneous, varying by location and moderated by socioeconomic factors at the intra-urban scale (Analitis et al., 2014; Fenech et al., 2019; Krug et al., 2020; Pascal et al., 2021; Schwarz et al., 2021; Scortichini et al., 2018). This, combined with the challenges associated with projecting future concentrations of health-relevant pollutants during heatwave events (Jahn and Hertig, 2021; Meehl et al., 2018), makes it difficult to say with any certainty that synergistic effects of heat and poor air quality will result in a heatwave–air pollution health penalty under climate change.

The burden of disease associated with aeroallergens is anticipated to grow due to climate change (high confidence). The incidence of pollen allergy and associated allergic disease increases with pollen exposure, and the timing of the pollen season and pollen concentrations are expected to change under climate change (Beggs, 2021; Ziska et al., 2019; Ziska, 2020). The overall length of the pollen season and total seasonal pollen counts/concentrations for allergenic species such as birch (Betula) and ragweed (Ambrosia) are expected to increase as a result of CO2 fertilisation and warming, leading to greater sensitisation (Hamaoui-Laguel et al., 2015; Lake et al., 2017; Zhang et al., 2013). Changes in pollen levels for several species of trees and grasses are projected to increase annual emergency department visits in the USA by between 8% for RCP4.5 and 14% for RCP8.5 by the year 2090 (Neumann et al., 2019) with the exposure to some pollen types estimated to double beyond present levels in Europe by 2041–2060 (Lake et al., 2017). The prospect of increases in summer thunderstorm events under climate change (Brooks, 2013) may hold implications for changes in the occurrence of epidemic thunderstorm asthma (Bannister et al., 2021; Emmerson et al., 2021; Price et al., 2021). Similarly, projected alterations in hydroclimate under climate change may bear implications for increased exposure to mould allergens in some climates (D’Amato et al., 2020; Paudel et al., 2021).

Climate change is expected to increase heat-related CVD mortality by the end of the 21st century, particularly under higher emission scenarios (high confidence). Most modelling studies conducted since AR5 project higher rates of heat-related CVD mortality throughout the remainder of this century (Huang and et al., 2018; Li et al., 2015; Li et al., 2018; Limaye et al., 2018; Zhang et al., 2018a; Silveira et al., 2021a; Yang et al., 2021). CVD mortality in Beijing, China, could increase by an average of 18.4%, 47.8% and 69.0% in the 2020s, 2050s and 2080s, respectively, under RCP4.5 and by 16.6%, 73.8% and 134%, respectively, under RCP8.5 relative to a 1980s baseline (Li et al., 2015). Projections of temperature-related mortality from CVD for Beijing in the 2080s vary depending on RCP and population assumptions (Zhang et al., 2018a). Projections for Ningo, China, suggest heat-related years of life lost (YLL) could increase significantly in the month of August by between 3 and 11.5 times over current baselines by the 2070s, even with adaptation (Huang and et al., 2018). Yang and colleagues project that heat-related excess CVD mortality in China could increase to approximately 6% (from a 2010 baseline of under 2%) by the end of the century under RCP8.5 and to over 3% under RCP4.5 (Yang et al., 2021). The future burden of temperature-related myocardial infarctions in Germany is projected to rise under high emissions scenarios (Chen et al., 2019), while in the eastern USA, Limaye et al. (2018) projected an additional 11,562 annual deaths (95% CI: 2,641–20,095) by mid-century due to cardiovascular stress in the population 65 years of age and above. CVD mortality in Brazil is projected to increase up to 8.6% by the end of the century under RCP8.5, compared with an increase of 0.7% for RCP4.5 (Silveira et al., 2021a).

It is important to note that the assessed studies typically take an observed epidemiological relationship and apply future temperature projections (often derived from regional climate projections) to these relationships. Because the relationships between temperature and CVD deaths are influenced by both climatic and non-climatic factors (such as population fitness and aging), future projections are highly sensitive to assumptions about interactions between climate, population characteristics and adaptation pathways. Changes in air quality because of climate change are an additional important factor. For example, an assessment of future annual and seasonal excess mortality from short-term exposure to higher levels of ambient ozone in Chinese cities under RCP8.5 projected approximately 1,500 excess annual CVD deaths in 2050 (Chen et al., 2018). To the extent possible, the relationships reported above reflect changes derived from changes in heat exposure driven by climate change and not changes in population demographics or air pollution exposure.

Climate change could impact CVD through other pathways, including exposure to fine dust. For example, adult mortality attributable to fine dust exposure in the American southwest could increase by 750 deaths yr –1 (a 130% increase over baseline) by the end of the century under RCP8.5 (Achakulwisut et al., 2018).

Additional research is needed on future impacts of climate change on maternal, foetal and neonatal health . Maternal heat exposure is a risk factor for several adverse maternal, foetal and neonatal outcomes (Kuehn and McCormick, 2017), including foetal growth (Sun et al., 2019) and congenital anomalies (Haghighi et al., 2021). There is very limited research on this subject, an exception being Zhang et al. (2020), which projected a 34% increase in congenital health disease risk in the USA in 2025 and 2035 based on increased maternal extreme heat exposure.

Harmful algal blooms are projected to increase globally, thus increasing the risk of seafood contamination with marine toxins (high confidence) (European Food Safety Authority et al., 2020; Gobler et al., 2017; Barange et al., 2018; IPCC, 2019b; Wells et al., 2020). Climate change impacts on oceans could generate increased risks of ciguatera poisoning in some regions (medium confidence). Studies suggest that rising sea surface temperatures could increase rates of ciguatera poisoning in Spain (Botana, 2016) and other parts of Europe (European Food Safety Authority et al., 2020).

Mycotoxins and aflatoxins may become more prevalent due to climate change (medium agreement, low evidence). Models of aflatoxin occurrence in maize under climate change scenarios of +2°C and +5°C in Europe over the next 100 years project that aflatoxin B1 may become a major food safety issue in maize, especially in Eastern Europe, the Balkan Peninsula and the Mediterranean regions (Battilani, 2016). The occurrence of toxin-producing fungal phytopathogens has the potential to increase and expand from tropical and subtropical regions into regions where such contamination does not currently occur (Battilani, 2016).

Climate change may alter regional and local exposures to anthropogenic chemical contaminants (medium agreement, low evidence). Changes in future occurrences of wildfires could lead to a 14% increase in global emissions of mercury by 2050, depending on the scenarios used (Kumar et al., 2018a). Mercury exposure via consumption of fish may be affected by warming waters. Warming trends in the Gulf of Maine could increase the methyl mercury levels in resident tuna by 30% between 2015 and 2030 (Schartup et al., 2019). An observed annual 3.5% increase in mercury levels was attributed to fish having higher metabolism in warmer waters, leading them to consume more prey. The combined impacts of climate change and the presence of arsenic in paddy fields are projected to potentially double the toxic heavy metal content of rice in some regions, potentially leading to a 39% reduction in overall production by 2100 under some models (Muehe et al., 2019).

Climate change is expected to have adverse impacts on well-being, some of which will become serious enough to threaten mental health (very high confidence). However, changes (Hayes and Poland, 2018) in extreme events due to climate change, including floods (Baryshnikova, 2019), droughts (Carleton, 2017) and hurricanes (Kessler et al., 2008; Boscarino et al., 2013, Boscarino et al., 2017; Obradovich et al., 2018), which are projected to increase due to climate change, directly worsen mental health and well-being and increase anxiety (high confidence). Projections suggest that sub-Saharan African children and adolescents, particularly girls, are extremely vulnerable to negative direct and indirect impacts on their mental health and well-being (Atkinson and Bruce, 2015; Owen et al., 2016). The direct risks are greatest for people with existing mental disorders, physical injuries, and compromised respiratory, cardiovascular and reproductive systems, with indirect impacts potentially arising from displacement, migration, famine and malnutrition, degradation or destruction of health and social care systems, conflict, and climate-related economic and social losses (high to very high confidence) (Burke et al., 2018; Curtis et al., 2017; Hayes et al., 2018; Serdeczny et al., 2017; Watts et al., 2019). Demographic factors increasing vulnerability include age, gender and low socioeconomic status, though the effect of these will vary depending on the specific manifestation of climate change; overall, climate change is predicted to increase inequality in mental health across the globe (Cianconi et al., 2020). Based on evidence assessed in Section 7.2, future direct impacts of increased heat risks and associated illnesses can be expected to have negative implications for mental health and well-being, with outcomes being highly mediated by adaptation, but there are no assessable studies that quantify such risks. There may be some benefits to mental health and well-being associated with fewer very cold days in the winter; however, research is inconsistent. Any positive effect associated with reduced low-temperature days is projected to be outweighed by the negative effects of increased high temperatures (Cianconi et al., 2020).

Human behaviours and systems will be disrupted by climate change in a myriad of ways, and the potential consequences for mental health and well-being are correspondingly large in number and complex in mechanism (high confidence). For example, climate change may alter human physical activity and mobility patterns, in turn producing alterations in the mental health statuses promoted by regular physical activity (Obradovich and Fowler, 2017; Obradovich and Rahwan, 2019). Climate change may affect labour capacity, because heat can compromise the ability to engage in manual labour as well as cognitive functioning, with impacts on the economic status of individual households as well as societies (Kjellstrom et al., 2016; Liu, 2020). Migrations and displacement caused by climate change may worsen the well-being of those affected (Vins et al., 2015; Missirian and Schlenker, 2017). Climate change is expected to increase aggression through both direct and indirect mechanisms, with one study predicting a 6% increase in homicides globally for a 1°C temperature increase, although noting significant variability across countries (Mares and Moffett, 2016). Broad societal outcomes such as economic unrest, political conflict or governmental dysfunction assessed in Section 7.3.5 may undermine the mental health of populations in the future (medium confidence). Food insecurity presents its own severe risks for mental health and cognitive function (Jones, 2017).

As outlined in 7.2, the most common drivers of observed climate-related migration and displacement are extreme storms (particularly tropical cyclones), floods and droughts (high confidence). The future frequency and/or severity of such events due to anthropogenic climate change are expected to vary by region according to future GHG emission pathways (Naik et al 2021; Regional Chapters, this report), with there being an increased potential for compound effects of successive or multiple hazards (e.g., tropical storms accompanied by extreme heat events (Matthews et al., 2019)). Table 7.2 summarises anticipated changes in future migration and displacement risks due to sudden-onset climate events by region (and by sub-regions for Africa and Asia, where climatic risks vary within the region).

Table 7.2 | Projected changes in sudden-onset climate events associated with migration and displacement by region.

In low-lying coastal areas of most regions, future increases in mean sea levels will amplify the impacts of coastal hazards on settlements, including erosion, inland penetration of storm surges and groundwater contamination by salt water, and eventually lead to inundation of very low-lying coastal settlements (high confidence) (Diaz, 2016; Hauer et al., 2016; Neumann et al., 2015; Rahman et al., 2019; IPCC, 2019a). Projections of the number of people at risk of future displacement by sea level rise range from tens of millions to hundreds of millions by the end of this century, depending on (a) the sea level rise scenario or RCP selected, (b) projections of future population growth in exposed areas and (c) the criteria used for identifying exposure. These latter measures can include estimates of populations situated within selected elevations above sea level (with 1 m, 2 m and 10 m being common parameters), populations situated in 1-in-100 year floodplains or populations in areas likely to be entirely inundated under specific RCPs (Neumann et al., 2015; Hauer et al., 2016; Merkens et al., 2018; McMichael et al., 2020; Hooijer and Vernimmen, 2021). As an illustrative example, an estimated 267 million people (error range = 197–347 million at 68% confidence level) worldwide lived within 2 m of sea level in 2020, 59% of whom reside in tropical regions of Asia (Hooijer and Vernimmen, 2021). At a 1 m increase in sea level and holding coastal population numbers constant, the number of people worldwide living within 2 m of sea level expands to 410 million (error range = 341–473 million). However, it is unlikely that coastal population growth rates will remain constant at global or regional scales in future decades. At present, coastal cities in many regions have relatively high rates of population growth due to the combined effects of in-migration from other regions and natural increase, with coastal areas of Africa having the highest projected future population growth rates (Neumann et al., 2015; Hooijer and Vernimmen, 2021; Box 7.5). Further complicating future estimates is that many large coastal cities are situated in deltas with high rates of subsidence, meaning that locally experienced changes in relative sea level may be much greater than sea level rise attributable to climate change, thereby further increasing the number of people exposed (Edmonds et al., 2020; Nicholls et al., 2021).

Sea level rise is not presently a significant driver of migration in comparison with hazards assessed in Section 7.2.6, but it has been attributed as a factor necessitating the near-term resettlement of small coastal settlements in Alaska, Louisiana, Fiji, Tuvalu and the Carteret Islands of Papua New Guinea (Marino and Lazrus, 2015; Connell, 2016; Hamilton et al., 2016; Nichols, 2019). In coastal Louisiana, communities tend to resist leaving exposed settlements until approximately 50% of available land has been lost (Hauer et al., 2019). Movements away from highly exposed areas may have longer-term demographic implications for inland settlements (Hauer, 2017), but this requires further study. Based on the available empirical evidence, sea level rise does not appear to currently be a primary motivation for international migration originating in small island states in the Indian and Pacific Oceans; rather, economic considerations and family reunification appear to be the dominant drivers (McCubbin et al., 2015; Stojanov and Du, 2016; Heslin, 2019; Kelman et al., 2019). However, climatic drivers of migration are anticipated to take on a much greater causal role in migration decisions in coming decades (Thomas et al., 2020) and may discourage return migration to small island states (van der Geest et al., 2020). Even under best-case sustainable development scenarios, rising sea levels and associated hazards create risks of involuntary displacement in low-lying coastal areas and should be expected to generate a need for organised relocation of populations where protective infrastructure cannot be constructed (Horton and de Sherbinin, 2021; Hamilton et al., 2016). In high emissions scenarios, low-lying island states may face the long-term risk of becoming uninhabitable, creating the potential for a new phenomenon of climate-induced statelessness (Piguet, 2019; Desai et al., 2021).

Increased frequency of extreme heat events and long-term increases in average temperatures pose future risks to the habitability of settlements in tropical and subtropical regions, and may in the long term affect migration patterns in exposed areas, especially under high emissions scenarios (medium agreement, low evidence). Greater research into the specific dynamics between extreme heat and population movements is required in order to make an accurate assessment of this risk. Recent studies suggest that future increases in average temperatures could expose populations across wide areas of the tropics and subtropics to ambient temperatures for extended periods each year that are beyond the threshold for human habitability (Pal and Eltahir, 2016; Im et al., 2017; Xu et al., 2020). This effect would be amplified in urban settings where heat-island effects occur and create a heightened need for air conditioning and other adaptation measures. In addition to risks associated with average temperature changes, Dosio et al. (2018) project that at 1.5°C warming, between 9% and 18% of the global population will be regularly exposed to extreme heat events at least once in five years, with the exposure rate nearly tripling with 2°C warming. How these changes in exposure to high temperatures will affect future migration patterns, particularly among vulnerable groups, will depend heavily on future adaptation responses (Horton and de Sherbinin, 2021). Multiple country-level studies assessed in Section 7.2 observe existing associations between extreme heat, its impacts on agricultural livelihoods and changes in rural-to-urban migration flows in parts of south Asia and sub-Saharan Africa. A study conducted in Indonesia, Malaysia and the Philippines suggests that an increased risk of heat stress would likely influence migration intentions of significant numbers of people (Zander et al., 2019).

Only a very small number of studies have attempted to make systematic projections of future regional or global migration and displacement numbers under climate change. Key methodological challenges for making such projections include the availability of reliable data on migration within and between countries, definitional ambiguity in distinguishing climate-related migration from migration undertaken for other reasons, and accounting for the future influence of non-climatic factors. The most reliable example of such studies to date is a World Bank report by Rigaud et al. (2018) that generated projections of future internal population displacements in south Asia, sub-Saharan Africa and Latin America by 2050 using multiple climate and development scenarios, resulting in a very large range of possible outcomes (from 31 to 143 million people being displaced, depending on assumptions). An important outcome is the study’s emphasis on how the potential for future migration and displacement will be strongly mediated by socioeconomic development pathways in low- and middle-income countries. Hoffmann et al. (2020) used meta-regression-based analyses to project that future environmental influences on migration are likely to be greatest in low- and middle-income countries in Latin America and the Caribbean, sub-Saharan Africa, the Middle East and most of continental Asia.

Research reviewed in AR4 and AR5 observed that at higher rates of socioeconomic development, the in situ adaptive capacity of households and institutions rises, and climatic influences on migration correspondingly decline. Recent evidence adds further support for such conclusions (high confidence) (Kumar et al., 2018b; Mallick, 2019; Gray et al., 2020; Box 7.5). Population growth rates are currently highest in low-income countries (UN DESA Population Division, 2019), many of which have high rates of exposure to climatic hazards associated with population displacement, further emphasising the importance of socioeconomic development and adaptive capacity-building. Although country-specific scenarios for socioeconomic development and population are embedded in SSPs, research into future migration flows under climate change has not made great use of these. One of the few studies to do so found that safe and orderly international migration tends to increase wealth at regional and global scales in all SSP narratives, which in turn reduces income inequality between countries (Benveniste et al., 2021). International barriers to safe and orderly migration may potentially impede progress towards attainment of the objectives described in the SDGs and increase exposure to climatic hazards in low- and middle-income countries (McLeman, 2019; Benveniste et al., 2020).