Editing IPCC:AR6/WGII/Chapter-16 (section)

== 16.5 Key Risks across Sectors and Regions ==

<div id="h1-6-siblings" class="h1-siblings"></div>

This section builds on the analogous chapter in AR5 ( [[#Oppenheimer--2014|Oppenheimer et al., 2014]] ) to refine the definition of climate-related key risks (KRs) and criteria for identifying them ( [[#16.5.1|Section 16.5.1]] ), and describe a broad range of key risks by sector and region as identified by the authors of WGII AR6 ( [[#16.5.2|Section 16.5.2]] , SM16.4). Based on this, eight clusters of key risks (i.e., Representative Key Risks, RKRs) are identified and assessed in terms of the conditions under which they would become severe. In addition, the section assesses variation in KRs and RKRs by the level of global average warming, socioeconomic development pathways, and levels of adaptation, and illustrates the implications from resulting dynamics in all risk dimensions (hazard, exposure, vulnerability) along a case study of densely populated river deltas ( [[#16.5.3|Section 16.5.3]] ). Last, interactions among RKRs are discussed ( [[#16.5.4|Section 16.5.4]] ).

<div id="16.5.1" class="h2-container"></div>

<span id="defining-key-risks"></span>
=== 16.5.1 Defining Key Risks ===

<div id="h2-14-siblings" class="h2-siblings"></div>

A key risk is defined as a potentially severe risk and therefore especially relevant to the interpretation of dangerous anthropogenic interference (DAI) with the climate system, the prevention of which is the ultimate objective of the UNFCCC as stated in its Article 2 ( [[#Oppenheimer--2014|Oppenheimer et al., 2014]] ). Key risks are therefore a relevant lens for the interpretation of this policy framing. The severity of a risk is a context-specific judgement based on a number of criteria discussed below. KRs are ‘potentially’ severe because, while some could already reflect dangerous interference now, more typically they may become severe over time due to changes in the nature of hazards (or, more broadly, climatic impact drivers; [[#IPCC--2021|IPCC, 2021]] ) and/or of the exposure/vulnerability of societies or ecosystems to those hazards. They also may become severe due to the adverse consequences of adaptation or mitigation responses to the risk (on the former, see [[IPCC:Wg2:Chapter:Chapter-17#17.5.1|Section 17.5.1]] ; the latter is not assessed separately here, except as it contributes to risks from climate hazards). Dangerous interferences in this chapter are considered over the course of the 21st century.

KRs may be defined for a wide variety of systems at a range of scales. The broadest definition is for the global human system or planetary ecological system, but KRs may also apply to regions, specific sectors or communities, or to parts of a system rather than to the system as a whole. For example, the population at the lower end of the wealth distribution is often impacted by climate change much more severely than the rest of the population ( [[#Leichenko--2014|Leichenko and Silva, 2014]] ; [[#Hallegatte--2017|Hallegatte and Rozenberg, 2017]] ; [[#Hallegatte--2017|Hallegatte et al., 2017]] ; [[#Pelling--2019|Pelling and Garschagen, 2019]] ).

KRs are determined not just by the nature of hazards, exposure, vulnerability and response options, but also by values, which determine the importance of a risk. Importance is understood here as the degree of relevance to interpreting DAI at a given system’s level or scale, and was an explicit criterion for identifying key vulnerabilities and risks in AR5 ( [[#Oppenheimer--2014|Oppenheimer et al., 2014]] ). Because values can vary across individuals, communities or cultures, as well as over time, what constitutes a KR can vary widely from the perspective of each of these groups, or across individuals. For example, ecosystems providing indirect services and cultural assets such as historic buildings and archaeological sites may be considered very important to preserve by some people but not by others; and some types of infrastructure, such as a commuter rail, may be important to the well-being of some households but less so to others. Therefore, [https://www.ipcc.ch/report/ar6/wg2/chapter/chapter-16 Chapter 16] authors do not make their own judgements about the importance of particular risks. Instead, we highlight importance as an overarching factor but identify and evaluate KRs based on four other criteria for what may be considered potentially severe.

'''Magnitude of adverse consequences.''' Magnitude measures the degree to which particular dimensions of a system are affected, should the risk materialise. Magnitude can include the size or extent of the system, the ''pervasiveness of the consequences'' across the system (geographically or in terms of affected population), as well as the ''degree of consequences'' . Consequences can be measured by a wide range of characteristics. For example, risks to food security can be measured as uncertain consequences for food consumption, access or prices. The magnitude of these consequences would be the degree of change in these measures induced by climate change and accounting for the interaction with exposure and vulnerability. In addition to ''pervasiveness'' and ''degree of change'' , several other aspects can contribute to a judgement of magnitude, although they refer to concepts that are difficult to capture and highly context-specific:

''Irreversibility of consequences'' . Consequences that are irreversible, at least over long time scales, would be considered a higher risk than those that are temporary. For example, changes to the prevailing ecosystem in a given location may not be reversible on the decade to century scale.

''Potential for impact thresholds or tipping points'' . Higher risks are posed by the potential for exceeding a threshold beyond which the magnitude or rate of an impact substantially increases.

''Potential for cascading effects beyond system boundaries'' . Higher risks are posed by those with the potential to generate downstream cascading effects to other ecosystems, sectors or population groups within the affected system and/or to another system, whether neighbouring or distant (Cross-Chapter Box INTEREG in this Chapter).

'''Likelihood of adverse consequences.''' A higher probability of high-magnitude consequences poses a larger risk ''a priori'' , whatever the scale considered. This probability may not be quantifiable, and it may be conditional on assumptions about the hazard, exposure or vulnerability associated with the risk.

'''Temporal characteristics of the risk.''' Risks that occur sooner, or that increase more rapidly over time, present greater challenges to natural and societal adaptation. A persistent risk (due to the persistence of the hazard, exposure and vulnerability) may also pose a higher threat than a temporary risk due, for example, to a short-term increase in the vulnerability of a population (e.g., due to conflict or an economic downturn).

'''Ability to respond to the risk''' . Risks are more severe if the affected ecosystems or societies have limited ability to reduce hazards (e.g., for human systems, through mitigation, ecosystem management and possibly solar radiation management); to reduce exposure or vulnerability through various human or ecological adaptation options; or to cope with or respond to the consequences, should they occur.

The relative influence of these different criteria is case-specific and left to author judgement in the identification of KRs (groups of authors in regional and sectoral chapters) and the assessment of representative key risks (author teams, see SM16.4). But in general, the more criteria are met, the higher is the risk.

<div id="16.5.2" class="h2-container"></div>

<span id="identification-and-assessment-of-key-risks-and-representative-key-risks"></span>
=== 16.5.2 Identification and Assessment of Key Risks and Representative Key Risks ===

<div id="h2-15-siblings" class="h2-siblings"></div>

<div id="16.5.2.1" class="h3-container"></div>

<span id="identification-of-key-risks"></span>
==== 16.5.2.1 Identification of Key Risks ====

<div id="h3-31-siblings" class="h3-siblings"></div>

The authors of the sectoral and regional chapters and cross chapter papers of the WGII AR6 Report identified more than 120 key risks (SM16.7.4). Authors were asked to rely on the above definition and criteria to identify risks that could potentially become severe according to changes in the associated hazards, the study systems’ exposure and/or vulnerability, and important adaptation strategies that could reduce these risks (see SM16.3 for methodology). Wherever possible, identification is based on literature that includes projected future conditions for all three components of risk and adaptation. Where literature was insufficient, potential severity is based on current vulnerability and exposure to climate hazards and the expectation that hazards will increase in frequency and/or intensity in the future. This approach is more limited in that it does not consider future changes in exposure and vulnerability nor in adaptation, but has the benefit of being grounded in observed experience.

Table SM16.24 indicates that climate change presents a wide range of risks across scales, sectors and regions that could become severe under particular conditions of hazards, exposure and vulnerability, which may or may not occur. Some illustrations of the extent and diversity of KRs are provided here, and more detailed assessment can be found in the Chapters referenced in the table.

Global-scale KRs include threats to biodiversity in oceans, coastal regions and on land, particularly in biodiversity hotspots, as well as other ecological risks such as geographic shifts in vegetation, tree mortality, reduction in populations and reduction in growth (such as for shellfish). These ecological risks include cascading impacts on livelihoods and food security. Global-scale risks also include risks to people, property and infrastructure from river flooding and extreme heat (particularly in urban areas), risks to fisheries (with implications for living standards and food security) and some health risks from food-borne diseases as well as psychopathologies.

Many KRs are especially prominent in particular regions or systems, or for particular subgroups of the population. For example, coastal systems and small islands are a nexus of many KRs, including those to ecosystems and their services, especially coral reefs; people (health, livelihoods); and assets, including infrastructure. Risks to socio-ecological systems in polar regions are also identified as KRs, as are ecological risks to the Amazon Forest in South America and savannahs in Africa. For some regions, risks from wildfire are of particular concern, including in Australasia and North America. Vector-borne diseases are a particular concern in Africa and Asia. Loss of cultural heritage is identified as a KR in small islands, mountain regions, Africa, Australasia and North America.

For many risks, low-income populations are particularly vulnerable to KRs. Climate-related impacts on malnutrition and other forms of food insecurity will be larger for this group, along with small-holder farming households and Indigenous communities reliant on agriculture, and for women, children, the elderly and the socially isolated ( [[IPCC:Wg2:Chapter:Chapter-5#5.12|Section 5.12]] ). KRs in coastal communities are expected to affect low-income populations more strongly, including through risks to livelihoods of those reliant on coastal fisheries. KRs related to health are generally higher for low-income populations less likely to have adequate housing or access to infrastructure.

<div id="16.5.2.2" class="h3-container"></div>

<span id="identification-of-representative-key-risks"></span>
==== 16.5.2.2 Identification of Representative Key Risks ====

<div id="h3-32-siblings" class="h3-siblings"></div>

As in AR5 [[#Oppenheimer--2014|Oppenheimer et al. (2014)]] , major clusters of KRs are further analysed, and here referred to as ‘representative key risks’ (RKRs). RKRs were defined in a three-step process (SM16.3.1). First, half of [https://www.ipcc.ch/report/ar6/wg2/chapter/chapter-16 Chapter 16] authors independently mapped the KRs (SM16.7.4) to a set of candidate RKRs. Second, all [https://www.ipcc.ch/report/ar6/wg2/chapter/chapter-16 Chapter 16] authors discussed the set of independent results and proposed a list of RKRs, considering scope and overlap. Third, this proposal was discussed with a consultative group of about 20 WGII AR6 authors from other chapters closely involved in the KR identification process, and a final list of eight RKRs was identified (Table 16.6).

The RKRs are intended to capture the widest variety of KRs to human or ecological systems with a small number of categories that are easier to communicate and provide a manageable structure for further assessment. They expand the scope of some AR5 KR clusters (e.g., on coasts, health, food and water) and add new ones (e.g., on peace and human mobility). The RKRs encompass a diversity of types of systems, including an example of a geographically defined system (RKR ''-'' A on coastal regions), ecosystem well-being and integrity (RKR ''-'' B), a cross-cutting issue relevant to several outcomes of concern (RKR-C on critical infrastructure) and several topics focused directly on aspects of human well-being and security (RKR ''-'' D to RKR ''-'' H). This set of RKRs manages but does not eliminate overlap, instead providing alternative perspectives on underlying key risks that sometimes include complementary views on common risks. For example, the water security RKR highlights the many key risks mediated by water quantity or quality, which are sometimes manifested as risk to food security (RKR-F) or health (RKR-E).

'''Table 16.6 |''' Climate-related representative key risks (RKRs). The scope of each RKR is further described in the assessments in [[#16.5.2.3|Section 16.5.2.3]] . Relation to categories of overarching key risks identified in AR5 is provided for continuity.

{| class="wikitable"
|-
! Code

! Representative key risk

! Scope

! Relation to AR5 overarching key risks;

for definitions, refer to Oppenheimer et al.. (2014)

! Subsection assessment

|-
| RKR-A

| Risk to low-lying coastal socio-ecological systems

| Risks to ecosystem services, people, livelihoods and key infrastructure in low-lying coastal areas, and associated with a wide range of hazards, including sea level changes, ocean warming and acidification, weather extremes (storms, cyclones), sea ice loss, etc.

| Contains key risk (i), overlaps with key risks (iii) and (vii)

| 16.5.2.3.1

|-
| RKR-B

| Risk to terrestrial and ocean ecosystems

| Transformation of terrestrial and ocean/coastal ecosystems, including change in structure and/or functioning, and/or loss of biodiversity.

| Contained in key risks (vii) and (viii)

| 16.5.2.3.2

|-
| RKR-C

| Risks associated with critical physical infrastructure, networks and services

| Systemic risks due to extreme events leading to the breakdown of physical infrastructure and networks providing critical goods and services.

| Overlaps with key risk (iii)

| 16.5.2.3.3

|-
| RKR-D

| Risk to living standards

| Economic impacts across scales, including impacts on gross domestic product (GDP), poverty and livelihoods, as well as the exacerbating effects of impacts on socioeconomic inequality between and within countries.

| Broader version of key risk (ii)

| 16.5.2.3.4

|-
| RKR-E

| Risk to human health

| Human mortality and morbidity, including heat-related impacts and vector-borne and waterborne diseases.

| Broader version of key risk (iv)

| 16.5.2.3.5

|-
| RKR-F

| Risk to food security

| Food insecurity and the breakdown of food systems due to climate change effects on land or ocean resources.

| Overlaps with key risk (v)

| 16.5.2.3.6

|-
| RKR-G

| Risk to water security

| Risk from water-related hazards (floods and droughts) and water quality deterioration. Focus on water scarcity, water-related disasters and risk to indigenous and traditional cultures and ways of life.

| Overlaps with key risk (iv)

| 16.5.2.3.7

|-
| RKR-H

| Risks to peace and to human mobility

| Risks to peace within and among societies from armed conflict as well as risks to low-agency human mobility within and across state borders, including the potential for involuntarily immobile populations.

| New

| 16.5.2.3.8

|}

<div id="16.5.2.3" class="h3-container"></div>

<span id="assessment-of-representative-key-risks"></span>
==== 16.5.2.3 Assessment of Representative Key Risks ====

<div id="h3-33-siblings" class="h3-siblings"></div>

Each RKR was assessed by a team of four to nine members drawn from Chapter 16, other WGII AR6 chapters, and external contributing authors (SM16.4). The following subsections describe the scope of the category of risk (underlying KR considered) and the approach to defining ‘severe’ risks for each particular RKR. They also assess the conditions in terms of warming (more broadly, climatic impact drivers; ( [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ), exposure/vulnerability and adaptation under which the RKR would become severe. For each of these dimensions, RKR teams considered generic levels ranging from High to Medium and Low. For warming levels, in line with WGI framing, High refers to climate outcomes consistent with RCP8.5 or higher, Low refers to climate outcomes consistent with RCP2.6 or lower, and Medium refers to outcomes for scenarios between RCPs 2.6 and 8.5. For reference, the full range of warming levels (across all climate models) associated with RCP8.5 for the 2081–2100 period is 3.0–6.2°C; for RCP2.6 it is 0.9–2.3°C; and for intermediate RCPs it is 1.8–3.6°C (Cross-Chapter Box CLIMATE in Chapter 1). For Exposure-Vulnerability, levels are determined by the RKR teams relative to the range of future conditions considered in the literature, for example based on the Shared Socioeconomic Pathways (SSPs) in which future conditions based on SSPs 1 or 5 represent Low exposure or vulnerability and those based on SSPs 3 or 4 represent High exposure or vulnerability ( [[#O’Neill--2014|O’Neill et al., 2014]] ; [[#van%20Vuuren--2014|van Vuuren and Carter, 2014]] ). For Adaptation, two main levels have been considered: High refers to near maximum potential, and Low refers to the continuation of today’s trends. Despite being intertwined in reality, Exposure-Vulnerability and Adaptation conditions are distinguished to help understand their respective contributions to risk severity. Importantly, this assessment does not consider all risks, but only those that can be considered severe given the definition and criteria presented in [[#16.5.1|Section 16.5.1]] . The assessment does not exclude the possibility that severe risks are already observed in some contexts, and considers projected risks through the end of this century.

Each RKR assessment followed a common set of guidelines (SM16.3) that included broad criteria for defining severity ( [[#16.5.1|Section 16.5.1]] ), consideration of complex risks and interactions within and across RKRs, and consideration of risks across a range of scales, regions, and ecological and human development contexts. The specific definition of severity within each RKR was determined by the author teams of that assessment, applying different combinations of key risk criteria and metrics as judged appropriate in each case. Definitions are transparent and use common criteria, but are nonetheless based on the respective author team’s judgement. Conclusions about severity and associated confidence statements are therefore conditional on those definitions.

Assessments are based on different types of evidence depending on the nature of the literature. In some cases, quantitative projections of potential impacts are available. In others and as for KR identification, the potential for severe risk is inferred from high levels of current vulnerability and the expectation that the relevant climate hazards (climatic impact drivers, CIDs) will increase in frequency or intensity in the future.

<div id="16.5.2.3.1" class="h4-container"></div>

<span id="risk-to-the-integrity-of-low-lying-coastal-socio-ecological-systems-rkr-a"></span>
===== 16.5.2.3.1 Risk to the integrity of low-lying coastal socio-ecological systems (RKR-A) =====

<div id="h3-34-siblings" class="h4-siblings"></div>

RKR-A considers climate-change-related risks to low-lying coasts including their physical, ecological and human components. Low-lying systems are those occupying land below 10 m of elevation that is contiguous and hydrologically connected to the sea ( [[#McGranahan--2007|McGranahan et al., 2007]] ). The assessment builds on Key Risks identified in Chapters 3 and 15, Cross Chapter Paper 2 as well as in the SROCC ( [[#Magnan--2019|Magnan et al., 2019]] ; [[#Oppenheimer--2019|Oppenheimer et al., 2019]] ). It highlights risks to (i) natural coastal protection and habitats; (ii) lives, livelihoods, culture and well-being; and (iii) critical physical infrastructure; it therefore overlaps with several other RKRs (Figures 16.10, 16.11) but within a coastal focus. It encompasses all latitudes and considers multiple sources of climate hazards, including SLR, ocean warming and acidification, permafrost thaw, and sea ice loss and changes in weather extremes.

Severe risks to low-lying coasts involve irreversible long-term loss of land, critical ecosystem services, livelihoods, well-being or culture in relation to increasing combined drivers, including climate hazards and exposure and vulnerability conditions. The definition depends on the local context because of variation in the perception of tolerable risks and the limits to adaptation ( [[#Handmer--2019|Handmer and Nalau, 2019]] ). Accordingly, a qualitative range of consequences is presented here, in place of a quantitative global severe risk threshold.

The literature suggests that severe risks generally occur at the nexus of high levels and rates of anthropogenic-driven change in climate hazards ( [[#16.2.3.2|Section 16.2.3.2]] ), concentrations of people and tangible and intangible assets, non-climate hazards such as sediment mining and ecosystem degradation ( [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.1|Section 3.4.2.1]] ), and the reaching of adaptation limits ( [[#16.4|Section 16.4]] ) ( ''medium evidence'' , ''high agreement'' ). In some Arctic communities and in communities reliant on warm-water coral reefs, even 1.5–2°C warming will lead to severe risks from loss of ecosystem services ( [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.2|Section 3.4.2.2]] ; Cross-Chapter Paper 6) ( ''high confidence'' ). Loss of land is already underway globally due to accelerating coastal erosion and will be amplified by increased sea level extremes and permanent flooding ( ''high confidence'' ; Oppenheimer et al. 2019, Ranasinghe et al. 2021). Observed impacts of and projected increases in high-intensity extreme events (Ranasinghe et al. 2021) also provide evidence for severe risk to occur on livelihoods, infrastructure and well-being ( [[#16.5.2.3.3|Section 16.5.2.3.3]] ) by mid-century ( ''high confidence'' ). Consequently, the combination of high warming, continued coastal development and low adaptation levels will challenge the habitability of many low-lying coastal communities in both developing and developed countries over the course of this century ( ''limited evidence'' , ''high agreement'' ) ( [[#Duvat--2021|Duvat et al., 2021]] ; [[#Horton--2021|Horton et al., 2021]] ). In some contexts, climate risks are already considered severe ( ''medium evidence'' , ''medium agreement'' ), and in others, even lower warming will induce severe risks to habitability, which will not necessarily be offset by ambitious adaptation ( ''limited evidence'' , ''medium agreement'' ).

# Natural coastal protection and habitats—severe risks from the loss of shoreline protection from reductions in wave attenuation ( [[#Beck--2018|Beck et al., 2018]] , Sections 3.5.5.1, 3.5.4.5) and sediment delivery (Sections 3.4.2.5, 15.3.3) are already observed in some coastal systems ( [[#16.2.3.1|Section 16.2.3.1]] ) and occur broadly even with 1.5°C of global warming ( [[#Hoegh-Guldberg--2018a|Hoegh-Guldberg et al., 2018a]] ; [[#Bindoff--2019|Bindoff et al., 2019]] , [[IPCC:Wg2:Chapter:Chapter-3#3.4.2|Section 3.4.2]] ). These impacts are the consequence of warming and SLR on coastal ecosystems.

Warm-water coral reefs are at risk of widespread loss of structural complexity and reef accretion by 2050 under 1.5°C global warming ( [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.1|Section 3.4.2.1]] ) ( ''high confidence'' ). Kelp forests may experience shifts in community structure ( [[#Arafeh-Dalmau--2019|Arafeh-Dalmau et al., 2019]] ; [[#Rogers-Bennett--2019|Rogers-Bennett and Catton, 2019]] ; [[#Smale--2020|Smale, 2020]] ; [[#Smith--2021|Smith et al., 2021]] ) with &gt;2°C of global warming especially at lower latitudes ( [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.2|Section 3.4.2.2]] ) ( ''high confidence'' ). In addition, depending on the local tide and sediment conditions, SLR associated with &gt;1.5°C of global warming (SSP1–2.6; 3.4.2.5) is sufficient to initiate shifts to alternate states in some seagrass and coastal wetland systems ( [[#van%20Belzen--2017|van Belzen et al., 2017]] ; [[#El-Hacen--2018|El-Hacen et al., 2018]] , [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.5|Section 3.4.2.5]] , Cross-Chapter Box SLR in Chapter 3), and submergence of some mangrove forests ( [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.5|Section 3.4.2.5]] ). A striking example of risks becoming severe at higher levels of warming is the one of coral islands with low elevation ( [[IPCC:Wg2:Chapter:Chapter-15#15.3.4|Section 15.3.4]] , Box 15.1): the risk of loss of habitability transitions from Moderate-to-High under RCP2.6 for most island types (urban and rural) to High-to-Very High under RCP8.5 ( [[#Duvat--2021|Duvat et al., 2021]] ), even under a high adaptation scenario ( [[#Oppenheimer--2019|Oppenheimer et al., 2019]] ), partly due to declining sediment supply ( [[#Perry--2018|Perry et al., 2018]] ) and increased annual flooding ( [[#Giardino--2018|Giardino et al., 2018]] ; [[#Storlazzi--2018|Storlazzi et al., 2018]] ).

More broadly, about 28,000 km 2 of land have been lost globally since the 1980s due to anthropogenic factors (e.g., coastal structures, disruption of sediment fluxes) and coastal hazards ( [[#Mentaschi--2018|Mentaschi et al., 2018]] ), and an additional loss of 6000–17,000 km 2 is estimated by the end of the century due to coastal erosion alone associated with SLR in combination with other drivers ( [[#Hinkel--2013|Hinkel et al., 2013]] ).

# Impacts to lives, livelihoods, culture and well-being—in the absence of effective adaptation, changing extreme and slow-onset hazards combined with anthropogenic drivers (e.g., increased population pressure at the coast between +5% and +13.6% by 2100 compared with today, [[#Jones--2016|Jones and O’Neill, 2016]] ) will lead to loss of lives, livelihoods, health, well-being and/or culture ( [[#McGregor--2016|McGregor et al., 2016]] ; [[#Pinnegar--2019|Pinnegar et al., 2019]] ; [[#Pugatch--2019|Pugatch, 2019]] ; [[#Schneider--2020|Schneider and Asch, 2020]] ; [[#Thomas--2020|Thomas and Benjamin, 2020]] ; [[#McNamara--2021|McNamara et al., 2021]] ) ( ''high confidence'' ). Catastrophic examples that may foreshadow the future include Hurricane Sandy in 2012 ( [[#Strauss--2021|Strauss et al., 2021]] ) and Super Typhoon Haiyan in 2013 (&gt;6,000 deaths and inequities in access to safe housing; Trenberth et al. 2015) (Sections 6.2.2, 6.3.5.1). Although there is no unique definition of ‘intolerable’ loss, risks are generally expected to become severe over this century ( [[#Tschakert--2017|Tschakert et al., 2017]] ; [[#Dannenberg--2019|Dannenberg et al., 2019]] ; [[#Tschakert--2019|Tschakert et al., 2019]] ). Globally, with High warming, 90–380 million more people will be exposed to annual flood levels by the mid- and end-century, respectively, compared with 250 million people today ( [[#Kulp--2019|Kulp and Strauss, 2019]] ; [[#Kirezci--2020|Kirezci et al., 2020]] ), with potential implications on forced displacement or migration ( [[#Oppenheimer--2019|Oppenheimer et al., 2019]] ; [[#Wrathall--2019|Wrathall et al., 2019]] ; [[#Hauer--2020|Hauer et al., 2020]] ; [[#Lincke--2021|Lincke and Hinkel, 2021]] , [[#16.5.2.3.8|Section 16.5.2.3.8]] ). Some of the largest fish-producing and fish-dependent ecoregions have already experienced losses of up to 35% in marine fisheries productivity due to warming ( [[#Free--2019|Free et al., 2019]] ), and about 11% of the global population will face increasing nutritional risks if current trajectories continue ( [[#Golden--2016|Golden et al., 2016]] ). While difficult to measure, current climate-driven losses to (Indigenous) knowledge, traditions ( [[#Tschakert--2019|Tschakert et al., 2019]] ; [[#Pearson--2021|Pearson et al., 2021]] ) and well-being ( [[#Ebi--2017|Ebi et al., 2017]] ; [[#Cunsolo--2018|Cunsolo and Ellis, 2018]] ; [[#Jaakkola--2018|Jaakkola et al., 2018]] ) indicate such risk as already severe in some regions ( ''limited evidence'' , ''medium agreement'' ), jeopardising communities’ realisation of their rights to food, health and culture. In the Arctic, climate-driven changes to ice and weather regimes have substantially affected traditional coastal-based hunting and fishing activities ( [[#Fawcett--2018|Fawcett et al., 2018]] ; [[#Galappaththi--2019|Galappaththi et al., 2019]] ; [[#Huntington--2020|Huntington et al., 2020]] ; [[#Nuttall--2020|Nuttall, 2020]] , Cross-Chapter Paper 6), and where permafrost thaw, SLR and coastal erosion are contributing to threatening cultural sites ( [[#Hollesen--2018|Hollesen et al., 2018]] ; [[#Fenger-Nielsen--2020|Fenger-Nielsen et al., 2020]] ).
# Critical physical infrastructure—severe risks are also illustrated through damages that lead to possibly long-lasting disruption of key services like transportation as well as energy generation and distribution in coastal areas ( [[#16.5.2.3.3|Section 16.5.2.3.3]] ) under all RCPs (Section [https://www.ipcc.ch/chapter/16#CCP2.2 CCP2.2.3] ) and if no additional adaptation ( ''medium confidence'' ). Critical transport infrastructure is already suffering from structural failures in polar regions, for instance, due to permafrost thaw and increased erosion associated with ocean warming, storm surge flooding and loss of sea ice ( [[#Melvin--2017|Melvin et al., 2017]] ; [[#Fang--2018|Fang et al., 2018]] , Sections 14.5.2.8, 16.2.3.2, Cross-Chapter Paper 6). One hundred airports are projected to be below mean sea level in 2100 with 2°C of warming (i.e., 0.62 m SLR, [[#Yesudian--2021|Yesudian and Dawson, 2021]] ), including in small islands ( [[#Monioudi--2018|Monioudi et al., 2018]] ; [[#Storlazzi--2018|Storlazzi et al., 2018]] ) and megacities. Projections show San Francisco International Airport, for instance, to be inundated by 2100 under the upper likely range of SLR in RCP8.5 (also considering subsidence trends, [[#Shirzaei--2018|Shirzaei and Bürgmann, 2018]] ). On the energy side, it is estimated that with 1.8 m SLR, for example, 4 out of 13 US nuclear power plant facilities will become exposed to storm surges and 3 others will be surrounded or submerged by seawater ( [[#Jordaan--2019|Jordaan et al., 2019]] ; [[#Jenkins--2020|Jenkins et al., 2020]] ).

<div id="16.5.2.3.1" class="h4-container"></div>

<span id="risk-to-terrestrial-and-ocean-ecosystems-rkr-b"></span>
===== 16.5.2.3.2 Risk to terrestrial and ocean ecosystems (RKR-B) =====

<div id="h3-34-siblings" class="h4-siblings"></div>

This risk refers to transformations of terrestrial and ocean/coastal ecosystems that would include significant changes in structure and/or functioning, and/or loss of a substantial fraction of species richness (commonly used to indicate loss of biodiversity). These are sourced mainly from Chapters 2 and 3, Cross-Chapter Paper 1, and reference the 1.5C report, [[IPCC:Wg2:Chapter:Chapter-4|Chapter 4]] from WGII AR5, and [[IPCC:Wg2:Chapter:Chapter-4|Chapter 4]] from WGII AR4 Reports.

Severe adverse impacts on biodiversity include significant risk of species extinction (e.g., loss of a substantial fraction (one-tenth or more) of species from a local to global scale), mass population mortality (&gt;50% of individuals or colonies killed), ecological disruption (order-of-magnitude increases or abrupt reductions of population numbers or biomass), shifts in ecosystem structure and function (order-of magnitude increases or abrupt decreases in cover and/or biomass of novel growth forms or functional types) and/or a socioeconomically material increase in environmental risk (e.g., destruction by wildfire) or socioeconomically material decline in goods and services (e.g., carbon stock losses, loss of grazing, loss of pollination). Metrics relevant to SDGs are also germane.

A substantial proportion of biodiversity is at risk of being lost below 2°C of global warming (Chapter 2), due to range reductions and loss globally, with this risk amplified roughly three times in insular ecosystems and biodiversity hotspots, due to the increased vulnerability of endemic species ( [[#Manes--2021|Manes et al., 2021]] ). High-latitude, high-altitude, insular, freshwater, and coral reef ecosystems and biodiversity hotspots (Chapter 2, Cross-Chapter Paper 1) are at appreciable risk of substantial biodiversity loss due to climate change even under Low warming ( ''high confidence'' ). These systems comprise a large fraction of unique and endemic biodiversity, with species impacts often exacerbated by multiple drivers of global change (Chapter 2, Chapter 3). Roughly one-third of all known plant species are extremely rare, vulnerable to climate impacts, and clustered in areas of higher projected rates of anthropogenic climate change ( [[#Enquist--2019|Enquist et al., 2019]] ). Much evidence shows increased risk of the loss of 10% or more of terrestrial biodiversity with increasing anthropogenic climate change ( [[#Urban--2015|Urban, 2015]] ; [[#Smith--2018|Smith et al., 2018]] ) ( ''medium confidence'' ), ''likely'' with 2°C warming above pre-industrial level (Chapter 2), with consequent degradation of terrestrial, freshwater and ocean ecosystems ( [[#Oliver--2015|Oliver et al., 2015]] ) and adverse impacts on ecosystem services ( [[#Pecl--2017|Pecl et al., 2017]] ) and dependent human livelihoods ( [[#Dube--2016|Dube et al., 2016]] ). Adverse impacts on biodiversity may show lagged responses ( [[#Essl--2015|Essl et al., 2015]] ), and loss of a substantial fraction of species could occur abruptly, simultaneously across multiple taxa, below 4°C of global warming ( [[#Trisos--2020|Trisos et al., 2020]] ).

Mass population-level mortality (&gt;50% of individuals or colonies killed) and resulting abrupt ecological changes can be caused by simple or compound climate extreme events, such as exceedance of upper thermal limits by vulnerable terrestrial species ( [[#Fey--2015|Fey et al., 2015]] ), who also note reduced mass mortality trends due to extreme low thermal events; marine heatwaves that can cause mortality, enhance invasive alien species establishment, and damage coastal ecological communities and small-scale fisheries ( ''high confidence'' ) ( [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.7|Section 3.4.2.7]] ); and increased frequency and extent of wildfires that threaten populations dependent on habitat availability (like Koala Bears, [[#Lam--2020|Lam et al., 2020]] ). Abrupt ecological changes are widespread and increasing in frequency ( [[#Turner--2020|Turner et al., 2020]] ), and include tree mortality due to insect infestation exacerbated by drought, and ecosystem transformation due to wildfire ( [[#Vogt--2020|Vogt et al., 2020]] ). Freshwater ecosystems and their biodiversity are at high risk of biodiversity loss and turnover due to climate change (precipitation change and warming, including warming of water bodies), due to high sensitivity of processes and life histories to thermal conditions and water quality (Chapter 2) ( ''high confidence'' ). In marine systems, heatwaves cause damages in coastal systems, including extensive coral bleaching and mortality ( ''very high confidence'' ) ( [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.1|Section 3.4.2.1]] ), mass mortality of invertebrate species ( ''low'' to ''high confidence'' , depending on system) (Sections 3.4.2.2, [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.5|Section 3.4.2.5]] , [[IPCC:Wg2:Chapter:Chapter-3#3.4.4|Section 3.4.4.1]] ), and abrupt mortality of kelp-forest ( ''high confidence'' ) ( [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.3|Section 3.4.2.3]] ) and seagrass-meadow habitat ( ''high confidence'' ) ( [[IPCC:Wg2:Chapter:Chapter-3#3.4.4|Section 3.4.4.2]] ). The biodiversity of polar seas shows strong impacts of climate change on phenological timing of plankton activity, Arctic fish species range contractions and species community change (Table SM16.22) ( ''high confidence'' ). Extreme weather events and storm surges exacerbated by climate change have severe and sudden adverse impacts on coastal systems, including loss of seagrass meadows and mangrove forests ( ''high confidence'' ) (see [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.7|Section 3.4.2.7]] , [[IPCC:Wg2:Chapter:Chapter-3#3.4.2.8|Section 3.4.2.8]] , Cross-Chapter Box EXTREMES in Chapter 2).

Ecological disruption (order-of-magnitude increases or abrupt reductions of population numbers or biomass) can occur due to unprecedented inter-species interactions with unpredictable outcomes in ‘novel ecosystems’ (Chapter 2) as species shift geographic ranges idiosyncratically in response to climatic drivers (Table SM16.22). Idiosyncratic geographic shifts are now observed in an appreciable fraction of species studied (Chapter 2, Table 16.2). Commensal or parasitic diseases may infect immunologically naive hosts (e.g., chytrid fungus in amphibians). Atypical disturbance regimes may be enhanced, for example, with the spread of flammable plant species (e.g., [[#du%20Toit--2015|du Toit et al., 2015]] ), exacerbated by introduced species (e.g., [[#Martin--2015|Martin et al., 2015]] ), thus significantly increasing risk of losses and damages to infrastructure and livelihoods, as well as ecological degradation, and challenging existing management approaches.

Landscape- and larger-scale shifts in ecosystem structure and function (order-of-magnitude increases or abrupt decreases in cover and/or biomass of novel growth forms or functional types) are occurring in non-equilibrium ecosystems (systems which exist in multiple states, often disturbance-controlled) in response to changing disturbance regime, climate and rising CO 2 ( ''high confidence'' ) Woody plant encroachment has been occurring in multiple ecosystems, including subtropical and tropical fire driven grassland and savanna systems, upland grassland systems, arid grasslands and shrublands ( ''high confidence'' ), leading to large-scale biodiversity changes, albedo changes, and impacts on water delivery, grazing services and human livelihoods ( ''medium confidence'' ). Expansion of grasses (alien and native) into xeric shrublands is occurring, causing increasing fire prevalence in previous fire-free vegetation (Cross-Chapter Paper 3). In tropical forests, repeated droughts and recurrence of large-scale anthropogenic fires increase forest degradation, loss of biodiversity and ecosystem functioning ( ''high confidence'' ) ( [[#Anderson--2018b|Anderson et al., 2018b]] ; Longo et al., 2020). Accelerated growth rates and mortality of tropical trees is also adversely affecting tropical ecosystem functioning ( [[#McDowell--2018|McDowell et al., 2018]] ; [[#Aleixo--2019|Aleixo et al., 2019]] ). Projected changes in ecosystem functioning, such as via wildfire ( [[IPCC:Wg2:Chapter:Chapter-2#2.5|Section 2.5.5.2]] ), tree mortality ( [[IPCC:Wg2:Chapter:Chapter-2#2.5|Section 2.5.5.3]] ) and woody encroachment under climate change (Chapter 2) would alter hydrological processes, with adverse implications for water yields and water supplies ( [[#Sankey--2017|Sankey et al., 2017]] ; [[#Robinne--2018|Robinne et al., 2018]] ; [[#Rodrigues--2019|Rodrigues et al., 2019]] ; [[#Uzun--2020|Uzun et al., 2020]] ).

The loss of a substantial fraction of biodiversity globally, abrupt impacts such as significant local biodiversity loss and mass population mortality events, and ecological disruption due to novel species interactions have been observed or are projected at global warming levels below 2°C ( [[IPCC:Wg2:Chapter:Chapter-2|Chapter 2]] Table SM2.5, Cross Chapter Box: EXTREMES in Chapter 2, [[IPCC:Wg2:Chapter:Chapter-2#2.4.4.3.1|Section 2.4.4.3.1]] , [[IPCC:Wg2:Chapter:Chapter-2#2.4.2.3.3|Section 2.4.2.3.3]] ) ( ''medium confidence'' ). Simple and compound impacts of extreme climate events are already causing significant losses and damages in vulnerable ecosystems, including through the facilitation of important global change drivers of ecological disruption and homogenisation like invasive species ( ''high confidence'' ). Severe impacts on human livelihoods and infrastructure, and valuable ecosystem services, are all projected to accompany these changes. Adaptation potential for many of these risks is low due to the projected rate and magnitude of change, and to the requirement of significant amounts of land for terrestrial ecosystems ( [[#Hannah--2020|Hannah et al., 2020]] ). Biodiversity conservation efforts may be hampered due to climate change impacts on the effectiveness of protected areas, with high sensitivity of effectiveness to forcing scenario ( ''medium confidence'' ). In addition, climate-related risks to ecosystems pose challenges to ecosystem-based adaptation responses (‘nature-based solutions’) ( [[IPCC:Wg2:Chapter:Chapter-2#2.1|Section 2.1.3]] ) ( ''medium confidence'' ).

<div id="16.5.2.3.3" class="h4-container"></div>

<span id="risk-to-critical-physical-infrastructure-and-networks-rkr-c"></span>
===== 16.5.2.3.3 Risk to critical physical infrastructure and networks (RKR-C) =====

<div id="h4-8-siblings" class="h4-siblings"></div>

RKR-C includes risks associated with the breakdown of physical infrastructure and networks which provide goods and services considered critical to the functioning of societies. It encompasses infrastructure systems for energy, water, transportation, telecommunications, health care and emergency response, as well as compound, cascading and cross-boundary risks resulting from infrastructure interdependencies ( [[#Birkmann--2016|Birkmann et al., 2016]] ; [[#Fekete--2019|Fekete, 2019]] ). Critical infrastructures such as transport or energy supply also play a central role in coping with climate risks, especially in acute disaster situations in which the services of transport infrastructure, communication technologies or electricity are particularly needed, despite the fact that these very systems are themselves exposed to disaster impacts ( [[#Garschagen--2016|Garschagen et al., 2016]] ; [[#Pescaroli--2018|Pescaroli et al., 2018]] ). The major hazards driving such risks are acute extreme events such as cyclones, floods, droughts or fires ( ''high confidence'' ), but cumulative and chronic hazards such as SLR are also considered.

RKR-C is considered severe when the functioning of critical infrastructure cannot be secured and maintained against climate change impacts, resulting in the frequent and widespread breakdown of service delivery and eventually a significant rise of detrimental impacts on people (lives, livelihoods and well-being), the economy (including averted growth) or the environment (disruption and loss of ecosystems) above historically observed levels. Severity in this RKR is assessed on two levels for (i) direct impacts of climate change on infrastructure assets and networks (e.g., amount of port infrastructure damaged or destroyed by SLR, flooding and storms) on which most of the literature focuses, as well as (ii) indirect and cascading downstream impacts to people, economy and environment ( [[#Markolf--2019|Markolf et al., 2019]] ; [[#Pyatkova--2019|Pyatkova et al., 2019]] ; [[#Chester--2020|Chester et al., 2020]] ), for which attribution is more difficult and uncertainties tend to be much higher. Overall, the literature with quantified assessments of climate change infrastructure risks remains to be less extensive than for many other risks, particularly with regard to assessments focusing on the Global South. While climate-related changes in hazards are widely considered in the literature, changes in future exposure and vulnerability conditions are often not treated explicitly. In addition, the severity of infrastructure risks also depends on future trends in the capacity to maintain, repair and rebuild infrastructure and adapt it to new hazard intensities ( ''medium evidence'' , ''high agreement'' ). These are mostly not quantified in a forward-looking manner in the literature; however, damage projections (see below) indicate a rapidly rising demand for investment, straining the financial capacity of countries ( ''medium evidence'' , ''high agreement'' ).

# Risks related to direct impacts on critical infrastructure would become severe with high warming, current infrastructure development regimes and minimal adaptation ( ''high confidence'' ), and in some contexts even with low warming, current vulnerability and no additional adaptation ( ''medium confidence'' ), with severity defined as infrastructure damage and required maintenance costs exceeding multiple times the current levels. Transport and energy infrastructure in coasts and polar systems and along rivers are projected to face a particularly steep rise in risk, resulting in severe risk even under medium warming ( ''high confidence'' ). Risk in relation to the increasing intensity and frequency of extreme events might become severe before the middle of the century ( ''medium confidence'' ). Damages from multiple climate hazards to transport, energy, industry and social infrastructure in Europe could increase 10-fold by the 2080s, from 3.4 € billion annually to date, and 15-fold for transport infrastructure, under Medium warming (A1B, ~3°C by 2100) and with current adaptation levels, even if no further extension of the infrastructure in exposed areas is considered ( [[#Forzieri--2018|Forzieri et al., 2018]] ). Under High warming (RCP8.5) in 2100, the percent of roads in the USA that require rehabilitation due to high temperatures and precipitation is expected to increase to 23–33%, relative to 14% in 2100 when no climate change is considered ( [[#Mallick--2018|Mallick et al., 2018]] ). Projections of climate-induced changes in exposure are an incomplete measure of risk but in the absence of other metrics can serve as a proxy for the potential for severe impacts. In the circumpolar Arctic, 14.8% of critical infrastructure assets would be affected by climate change under RCP8.5 by 2050, with lifecycle replacement costs projected to increase by 27.7% if infrastructure is to be preserved at current adaptation levels ( [[#Suter--2019|Suter et al., 2019]] ). Under RCP8.5, the number of ports under high risk will increase from 3.8% in the present day to 14.4% by 2100, as a result of increased coastal flooding and overtopping due to SLR, as well as the heat stress impacts of higher temperatures ( [[#Izaguirre--2021|Izaguirre et al., 2021]] ). In the UK under High warming (4°C), the number of clean and wastewater treatment sites located in the 1-in-75-year floodplain will increase by a third relative to today by the 2080s under current vulnerability and adaptation levels ( [[#Dawson--2018|Dawson et al., 2018]] ). A global assessment of changing climate and water resources for electricity generation finds considerable reductions in usable hydropower and thermoelectric capacity by 2050 for a range of warming scenarios from Low to High, with absolute declines on average for most (61–74%) of the world’s hydropower resources and monthly maximum reductions above 30% of usable capacity for over two-thirds of 1427 thermoelectric power plants worldwide ( [[#Van%20Vliet--2016|Van Vliet et al., 2016]] ). Many studies find large technical potential for coordinated adaptation–mitigation policies in the electricity sector to avoid a significant portion of projected climate change impacts (e.g., a two-thirds reduction, and in some cases fully offset) ( [[#Ciscar--2014|Ciscar and Dowling, 2014]] ; [[#Van%20Vliet--2016|Van Vliet et al., 2016]] ; [[#Gerlak--2018|Gerlak et al., 2018]] ; [[#Allen-Dumas--2019|Allen-Dumas et al., 2019]] ).
# Studies quantifying the indirect impacts of infrastructure failure on lives, livelihoods and economies are still rare but emerging, suggesting that risks would become severe in many contexts globally with high warming, current vulnerability and no additional adaptation ( ''medium confidence'' ). Severity in this context is defined as the potential to disrupt the lives, livelihoods and well-being of a significantly increased proportion of the population and to significantly forestall economic growth and development potential. Global risks to air travel from SLR, expressed in terms of expected annual route disruptions, could increase by a factor of between 17 and 69 by 2100 under the 1.5°C and the 95th percentile value of the RCP8.5 SLR scenario, respectively ( [[#Yesudian--2021|Yesudian and Dawson, 2021]] ). By 2050, up to 185,000 airline passengers per year may be grounded due to extreme heat (48°C) if no additional adaptation is taken, roughly 23 times more than today ( [[#McKinsey%20Global%20Institute--2020|McKinsey Global Institute, 2020]] ). In Africa, under RCP8.5 and without additional adaptation, a 250% increase in disruption time of the transport network is expected by 2050 due to extreme temperatures, a 76% increase due to precipitation, and 1400% increase due to flooding ( [[#Cervigni--2015|Cervigni et al., 2015]] ). On the Dawlish railway section (UK), the number of days with line restrictions is set to increase by up to 1170%, to as high as 84–120 yr –1 by 2100 due to 0.8 m SLR with High warming ( [[#Dawson--2016|Dawson et al., 2016]] ). Next to the limited number of projections or scenarios of indirect impacts, additional inferences from studies focusing on past and current impacts can be drawn. Already today, climate-related impacts on transport and energy infrastructure reach far beyond the direct impacts on physical infrastructure, triggering indirect impacts on, for example, health and income ( ''medium confidence'' ). A case study of future flood hazard in Europe found that the indirect impact of a power outage on the local economy is six to eight times greater than the direct flood damage and asset repair costs, due to the interruption of daily economic activity ( [[#Karagiannis--2019|Karagiannis et al., 2019]] ). In low- and middle-income countries, the annual costs from infrastructure disruptions reach up to 300 billion USD for firms and 90 billion USD for private households, with natural hazards such as floods being responsible for 10–70% of these disruptions, depending on the sectors and regions ( [[#Hallegatte--2019|Hallegatte et al., 2019]] ). Power outages triggered by floods or droughts have also been found to have substantial health implications, particularly among low-income populations ( [[#Klinger--2014|Klinger et al., 2014]] ), and shown to impede disaster recovery efforts and severely disrupt local economies ( [[#Karagiannis--2019|Karagiannis et al., 2019]] ; [[#Nicolas--2019|Nicolas et al., 2019]] ). In addition, risks associated with infrastructure have the potential to become particularly severe when hazard-driven infrastructure disruptions undermine the capacity of emergency response in disaster situations ( ''limited evidence'' , ''high agreement'' ). A study on the UK shows, for example, that even a small increase in minor road flooding leads to a disproportionately high disruption of the efficacy of emergency services ( [[#Yu--2020|Yu et al., 2020]] ). Similar risks have been found for rural areas, particularly in developing countries ( [[#Alegre--2020|Alegre et al., 2020]] ).

<div id="16.5.2.3.5" class="h4-container"></div>

<span id="risk-to-living-standards-rkr-d"></span>
===== 16.5.2.3.4 Risk to living standards (RKR-D) =====

<div id="h4-9-siblings" class="h4-siblings"></div>

This RKR includes risks to (i) aggregate economic output at the global and national levels, (ii) poverty and (iii) livelihoods, and their implications for economic inequality. It is informed by key risks identified by regional and sectoral chapters. Risks are potentially severe as measured by the magnitude of impacts in comparison with historical events or as inferred from the number of people currently vulnerable.

# Risks to aggregate economic output would become severe at the global scale with high warming and minimal adaptation ( ''medium confidence'' ), with severity defined as the potential for persistent annual economic losses due to climate change to match or exceed losses during the world’s worst historical economic recessions. With historically observed levels of adaptation, warming of ~4°C may cause a 10–23% decline in annual global GDP by 2100 relative to global GDP without warming, due to temperature impacts alone ( [[#Burke--2015|Burke et al., 2015]] ; [[#Kahn--2019|Kahn et al., 2019]] ; [[#Kalkuhl--2020|Kalkuhl and Wenz, 2020]] ). These magnitudes exceed economic losses during the Great Recession (2008–2009, ~5% decline in global GDP, up to 15–18% in some countries) and the COVID-19 pandemic (2020, ~3% decline globally, up to 10% in some countries) ( [[#IMF--2020|IMF, 2020]] ; [[#IMF--2021|IMF, 2021]] ). Unlike past recessions, climate change impacts would occur continuously every year. However, smaller effects (1–8%) are found when using alternative methodologies ( [[#Diaz--2017|Diaz and Moore, 2017]] ; [[#Nordhaus--2017|Nordhaus and Moffat, 2017]] ; [[#Kompas--2018|Kompas et al., 2018]] ; [[#Kalkuhl--2020|Kalkuhl and Wenz, 2020]] ), assuming less warming ( [[#Kahn--2019|Kahn et al., 2019]] ; Takakura et al., 2019), and assuming lower vulnerability and/or more adaptation ( [[#Diaz--2017|Diaz and Moore, 2017]] ); this literature is comprehensively summarised in Cross-Working Group Chapter Box ECONOMIC. Impacts at high levels of warming are particularly uncertain, as all methodologies require extrapolation and insufficiently incorporate possible tipping elements in the climate system ( [[#Kopp--2016|Kopp et al., 2016]] ).

Annual economic output losses in developing countries could exceed the worst country-level losses during historical economic recessions ( ''medium confidence'' ). Assuming global warming of ~4°C by 2100, historical adaptation levels and high vulnerability, losses across Sub-Saharan Africa may reach 12% of GDP by 2050 ( [[#Baarsch--2020|Baarsch et al., 2020]] ) and 80% by 2100 ( [[#Burke--2015|Burke et al., 2015]] ), and ~9% on average across developing countries by 2100 ( [[#Acevedo--2017|Acevedo et al., 2017]] ). The largest estimates are debated and depend on assumptions about development trends, adaptive capacity, and whether temperature impacts the level or growth rate of economic activity ( [[#Kalkuhl--2020|Kalkuhl and Wenz, 2020]] ). Severe risks are more likely in (typically hotter) developing countries because of nonlinearities in the relationship between economic damages and temperature ( [[#Burke--2015|Burke et al., 2015]] ; [[#Acevedo--2017|Acevedo et al., 2017]] ). These risks are highest in scenarios and countries with: a large portion of the workforce employed in highly exposed industries ( [[#Acevedo--2017|Acevedo et al., 2017]] ); a high concentration of population and economic activity on coastlines ( [[#Hsiang--2014|Hsiang and Jina, 2014]] ; [[#Acevedo--2017|Acevedo et al., 2017]] ); and an increase in the frequency or intensity of disasters triggered by natural hazards ( [[#Berlemann--2018|Berlemann and Wenzel, 2018]] ; [[#Botzen--2019|Botzen et al., 2019]] ). Whether baseline economic growth may help avoid severe future risks is highly uncertain ( [[#Dell--2012|Dell et al., 2012]] ; [[#Burke--2015|Burke et al., 2015]] ; [[#Acevedo--2017|Acevedo et al., 2017]] ; [[#Deryugina--2017|Deryugina and Hsiang, 2017]] ).

<div id="_idContainer033" class="Figure"></div>

[[File:74d211f7daf2c1526c89022ebf172002 IPCC_AR6_WGII_Figure_16_009.png]]

'''Figure 16.9 |''' '''Illustrative examples from individual studies of risks to living standards and the conditions under which they could become severe.''' Selected studies are not representative of the literature, but provide examples of potentially severe risks to aggregate economic output, poverty and livelihoods. High, medium and low levels of warming, exposure/vulnerability and adaptation are defined as in Figure 16.10.

# Under medium warming pathways, climate change risks to poverty would become severe if vulnerability is high and adaptation is low ( ''limited evidence'' , ''high agreement'' ). We define poverty in terms of absolute consumption levels and define severity as tens to hundreds of millions of additional people in poverty relative to the number without climate change (globally) or an absolute increase in the number of people living in poverty compared with today (nationally or locally). This global impact is comparable to the effect of the 2007 food price shock ( [[#De%20Hoyos--2009|De Hoyos and Medvedev, 2009]] ) and the 2020 COVID-19 pandemic ( [[#World%20Bank--2020|World Bank, 2020]] ) and can be compared to about 700 million in poverty in 2017, down from 1.9 billion in 1990 ( [[#World%20Bank--2020|World Bank, 2020]] ).

In a high-vulnerability development pathway, climate change in 2030 could push 35–132 million people into extreme poverty, in addition to the people already in poverty assuming climate is unchanged (disregarding impacts from natural variability; [[#Hallegatte--2017|Hallegatte and Rozenberg, 2017]] ; [[#Jafino--2020|Jafino et al., 2020]] ). In a low-warming pathway, risks from mitigation costs could also be severe if no progressive redistribution from carbon pricing revenues is applied (Soergel et al., 2021). At the national level, there is ''limited evidence'' of climate change causing an absolute increase in poverty (e.g., absolute increase of ~1–2% yr −1 through 2040, [[#Montaud--2017|Montaud et al., 2017]] ). Potentially severe risks to poverty are also supported by (1) the observed impacts of past disasters ( [[#Winsemius--2018|Winsemius et al., 2018]] ; [[#Hallegatte--2020|Hallegatte et al., 2020]] ; [[#Rentschler--2020|Rentschler and Melda, 2020]] ) and previous crises such as food price shocks ( [[#Ivanic--2008|Ivanic and Martin, 2008]] ) or current diseases ( [[#WHO--2018|WHO, 2018]] ) on poor people and on poverty; (2) the expectation that these events will become more intense or frequent in some regions (WGI Chapter 12, [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ); and (3) population growth and the low adaptive and coping capacities of the poor ( [[#Leichenko--2014|Leichenko and Silva, 2014]] ; [[#Huynh--2018|Huynh and Stringer, 2018]] ; [[#Thomas--2020|Thomas et al., 2020]] ). This literature provides indirect evidence that climate change will keep many people poor and may cause more than tens of millions to fall into poverty ( ''limited evidence'' , ''high agreement'' ).

# Climate change poses severe risks to livelihoods at low levels of warming, high exposure/vulnerability and low adaptation in climate-sensitive regions, ecosystems and economic sectors ( ''high confidence'' ), where severity refers to the disruption of livelihoods for tens to hundreds of millions of additional people ( [[#Arnell--2014|Arnell and Lloyd-Hughes, 2014]] ; [[#Liu--2018|Liu et al., 2018]] ). More widespread severe risks would occur at high levels of warming (with high exposure/vulnerability and low adaptation) where there is additional potential for one or more social or ecological tipping points to be triggered ( [[#Cai--2015|Cai et al., 2015]] ; [[#Cai--2016b|Cai et al., 2016b]] ; [[#Kopp--2016|Kopp et al., 2016]] ; [[#Steffen--2018|Steffen et al., 2018]] ; [[#Lenton--2019|Lenton et al., 2019]] ), and for severe impacts on livelihoods to cascade from relatively more climate-sensitive to relatively less climate-sensitive sectors and regions ( ''medium confidence'' ) ( [[#Lawrence--2020|Lawrence et al., 2020]] ). Severity assessment is based on the current magnitude of exposure and vulnerability across multiple social and ecological systems, projected future exposure and vulnerability, and the rate at which hazard frequency or intensity is expected to increase ( [[#Otto--2017|Otto et al., 2017]] ; [[#Roy--2018|Roy et al., 2018]] ; [[#Li--2019|Li et al., 2019]] , [[IPCC:Wg2:Chapter:Chapter-8#8.5|Section 8.5]] ). Without effective adaptation measures, regions with high dependence on climate-sensitive livelihoods—particularly agriculture and fisheries in the tropics and coastal regions—would be severely impacted even at low levels of warming ( ''high confidence'' ) ( [[#Hoegh-Guldberg--2018b|Hoegh-Guldberg et al., 2018b]] ; [[#Roy--2018|Roy et al., 2018]] ). For example, it is estimated that 330–396 million people could be exposed to lower agricultural yields and associated livelihood impacts at warming between 1.5°C and 2°C ( [[#Byers--2018|Byers et al., 2018]] ). Risks to the 200 million people with livelihoods derived from small-scale fisheries would also be severe, given sensitivity to ocean warming, acidification and coral reef loss occurring beyond 1.5°C ( [[#Cheung--2018b|Cheung et al., 2018b]] ; [[#Froehlich--2018|Froehlich et al., 2018]] ; [[#Free--2019|Free et al., 2019]] ; [[#Barnard--2021|Barnard et al., 2021]] ). Livelihoods in highly exposed locations, such as Small Island Developing States, low-lying coastal areas, arid or semiarid regions, the Arctic, and urban informal settlements or slums, are particularly vulnerable ( [[#Ford--2015c|Ford et al., 2015c]] ; [[#Hagenlocher--2018|Hagenlocher et al., 2018]] ; [[#Ahmadalipour--2019|Ahmadalipour et al., 2019]] ; [[#Tamura--2019|Tamura et al., 2019]] ). Within populations, the poor, women, children, the elderly and Indigenous populations are especially vulnerable due to a combination of factors, including gendered divisions of paid and/or unpaid labour, as well as barriers in access to information, skills, services or resources ( [[#Bose--2017|Bose, 2017]] ; [[#Thomas--2019b|Thomas et al., 2019b]] ; [[#Anderson--2020|Anderson and Singh, 2020]] ; [[#Adzawla--2021|Adzawla and Baumüller, 2021]] ) ( ''high confidence'' ). Future structural transformation could moderate risk severity by improving adaptive capacity, creating livelihoods in less climate-sensitive sectors, or by enabling sustainable migration to less climate-sensitive locations ( [[#Henderson--2017|Henderson et al., 2017]] ; [[#Roy--2018|Roy et al., 2018]] ). However, successful risk moderation would depend upon simultaneous avoidance of both climate-change-related and mitigation-related ( [[#Doelman--2019|Doelman et al., 2019]] ; [[#Fujimori--2019|Fujimori et al., 2019]] ; [[#Doelman--2020|Doelman et al., 2020]] ) or maladaptation-related risks ( [[#Magnan--2016|Magnan et al., 2016]] ; [[#Benveniste--2020|Benveniste et al., 2020]] ; [[#Schipper--2020|Schipper, 2020]] ).

Climate change also could increase income inequality between countries ( ''high confidence'' ) as well as within them ( ''medium evidence'' , ''high agreement'' ) resulting from and exacerbating impacts on aggregate economic activity, poverty and livelihoods. Increasing inequality implies larger impacts on the least well-off, threatens their ability to respond to climate hazards, compromises basic principles of fairness and established global development goals, and potentially threatens the functioning of society and long-term progress ( [[#Roe--2011|Roe and Siegel, 2011]] ; [[#Cingano--2014|Cingano, 2014]] ; [[#van%20der%20Weide--2018|van der Weide and Milanovic, 2018]] ). There is evidence that warming has slowed down the convergence in between-country income in recent decades ( [[#Diffenbaugh--2019|Diffenbaugh and Burke, 2019]] ). Future impacts may halt or even reverse this trend during this century owing to high sensitivity of developing economies ( [[#Burke--2015|Burke et al., 2015]] ; [[#Pretis--2018|Pretis et al., 2018]] ; [[#Baarsch--2020|Baarsch et al., 2020]] ), although projections depend as much or more on future socioeconomic development pathways and mitigation policies as on warming levels (Takakura et al., 2019; [[#Harding--2020|Harding et al., 2020]] ; [[#Taconet--2020|Taconet et al., 2020]] ). Within countries, studies that find adverse impacts on low-income groups imply an increase in inequality ( [[#Hallegatte--2017|Hallegatte and Rozenberg, 2017]] ; [[#Hsiang--2017|Hsiang et al., 2017]] ), although evidence for long-term climate impacts on within-country inequality at global scale remains limited.

<div id="16.5.2.3.5" class="h4-container"></div>

<span id="risk-to-human-health-rkr-e"></span>
===== 16.5.2.3.5 Risk to human health (RKR-E) =====

<div id="h4-9-siblings" class="h4-siblings"></div>

This RKR includes (i) mortality from heat, and morbidity and mortality from (ii) vector-borne diseases and (iii) waterborne diseases. It builds on KRs identified primarily in [[IPCC:Wg2:Chapter:Chapter-7|Chapter 7]] and health risks in regional chapters.

A severe risk to health is the potential for a widespread, substantial worsening of health conditions due to climate change. We measure severity in terms of the magnitude of mortality and morbidity. We consider a severe mortality impact to be a sustained increase in the crude mortality rate (CMR) of more than about 2–4 deaths per 10,000 people yr –1 . This range of increase is consistent with current mortality impacts with substantial global effects, including traffic fatalities (CMR of 1.6/10,000 yr −1 ; [[#IHME--2019|IHME, 2019]] ) and the COVID-19 pandemic (CMR of 4/10,000 yr −1 , as of April 2021, expressed as an annualised rate; [[#Ritchie--2021|Ritchie et al., 2021]] ). We use these global rates as thresholds in all cases, recognising that they reflect substantial variation across regions and sub-populations (for other points of comparison, see [[#IHME--2019|IHME, 2019]] ). Morbidity impacts are measured in numbers of disease cases or hospital admissions. We find that severe health impacts are projected to occur for particular sub-populations and regions where vulnerability is currently high and is assumed to persist into the future; we focus our assessment on these cases. In other cases, literature is either inadequate or does not support severe outcomes.

# Risks of heat-related mortality would become severe at global and regional scales with high levels of warming and vulnerability ( ''medium confidence'' ). Under these conditions (SSP3–8.5), accounting for adaptation, heat mortality would increase the global CMR by up to 7/10,000 yr −1 by 2100 ( [[#Carleton--2020|Carleton et al., 2020]] ). For example, the USA would experience a CMR increase of 2–4/10,000 yr −1 by the end of the century (medium vulnerability without adaptation, and recent vulnerability with adaptation, respectively) ( [[#Weinberger--2017|Weinberger et al., 2017]] ; [[#Shindell--2020|Shindell et al., 2020]] ). Also assuming no adaptation and recent vulnerability, most populations of the world would experience an increase of 2–10 percentage points in the percentage of deaths attributable to heat by the end of the century (RCP8.5) (Vicedo-Cabrera, 2018a; Gasparrini, 2017). Harmful conditions for health are expected to increase in frequency and intensity over all land areas along with the rising temperatures in the coming decades ( [[#Pal--2016|Pal and Eltahir, 2016]] ; [[#Russo--2017|Russo et al., 2017]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ; [[#Saeed--2021|Saeed et al., 2021]] ; [[#Schwingshackl--2021|Schwingshackl et al., 2021]] ). Projections of exposure are an incomplete measure of risk but suggest the potential for severe impacts. For example, the percent of global population exposed to deadly heat stress would increase from today’s 30% to 48–74% by the end of the century depending on level of warming and population distribution ( [[#Mora--2017|Mora et al., 2017]] ). Projected impacts are larger if exposure and/or vulnerability increases due to ageing of the population or increased inequality ( [[#Weinberger--2017|Weinberger et al., 2017]] ; [[#Chen--2020a|Chen et al., 2020a]] ; [[#IPCC--2021|IPCC, 2021]] ) and with limited adaptation capacity (e.g., poor infrastructure, limited air conditioning, few medical and public health resources) (SM16.7.4) ( [[#Carleton--2020|Carleton et al., 2020]] ). Higher risks are also expected in urban areas owing to hazard amplification (i.e., urban heat island effect) and in highly dense settlements with other environmental hazards such as air pollution ( [[#Zhao--2018|Zhao et al., 2018]] ; [[#Sera--2019|Sera et al., 2019]] ).
# Risks of vector-borne disease would become severe with high warming and current vulnerability, concentrated in children and in sensitive regions ( ''medium confidence'' ). Severity is defined by regionally substantial numbers of additional malaria deaths, disease cases and episodic hospitalisation demands (for dengue).

With high warming, the CMR for malaria among children under the age of 1 year could increase by 5.2–10.1/10,000 yr −1 in Africa under current vulnerability levels. This estimate assumes a net increase of 70–130 million more people exposed to potential disease transmission due to climate change in a high-warming scenario (RCP8.5, end of century) ( [[#Caminade--2014|Caminade et al., 2014]] ; Colón- [[#González--2021|González et al., 2021]] ; [[#Ryan--2020|Ryan et al., 2020]] ), representing a 14–27% increase in the current population at risk ( [[#Ryan--2020|Ryan et al., 2020]] ), and assumes children under 1 year of age are facing the same crude mortality in the future as for the African region today ( [[#IHME--2019|IHME, 2019]] ). The largest increase is observed in Eastern Africa, where the population exposed could nearly double by 2080 ( [[#Ryan--2020|Ryan et al., 2020]] ) without accounting for population growth, driven mainly by changes among previously unexposed populations at higher altitude areas (Colón- [[#González--2021|González et al., 2021]] ). Actual future disease burden of malaria will be highly sensitive to regional socioeconomic development and the effectiveness of malaria intervention programs.

For dengue, with high warming and current levels of vulnerability there could be as much as a doubling of cases and hospital admissions per year globally, relative to today, driven by both warming and population growth. These estimates are derived by assuming similar relative incidence rates as today ( [[#Shepard--2016|Shepard et al., 2016]] ) combined with projections of a more than doubling of the population exposed to potential disease transmission by the end of the century in a high-warming scenario (RCP8.5), although much of this increase is driven by population growth ( [[#Colón-González--2018|Colón-González et al., 2018]] ; [[#Monaghan--2018|Monaghan et al., 2018]] ; [[#Messina--2019|Messina et al., 2019]] ). There are around 3 billion people exposed to dengue today.

# Climate change would lead to severe risks of morbidity and mortality caused by waterborne diseases, particularly for diarrhoea in children in many lower- and middle-income countries (LMICs) and where vulnerability remains high ( ''medium confidence'' ). The global CMR for diarrhoea is 1.98 for all ages, but varies by region and age group, reaching as high as 53 for &lt;1-year-olds in Africa ( [[#IHME--2019|IHME, 2019]] ). In these vulnerable populations, even a small percentage increase can lead to substantial additional morbidity and mortality. For example, assuming no change in vulnerability or population, an increase in diarrhoea mortality of only 5% over 2019 baseline rates would create a severe risk (CMR of 2.0) for children under the age of 1 in the World Health Organization (WHO) Africa (AFRO) region. This percent increase due to climate change is plausible since diarrhoea incidence increases of 7% (95% confidence interval 3–10%) are associated with a 1°C increase in ambient temperature ( [[#WHO--2014|WHO, 2014]] ; [[#Carlton--2016|Carlton et al., 2016]] ), and diarrhoea is positively associated with heavy rainfall and flooding events ( [[#Levy--2016|Levy et al., 2016]] ), expected in some regions (WGI). Assuming vulnerability remains the same as today, mortality and morbidity rates would increase equivalently.

However, risks will be highly dependent on development trajectories, given that waterborne disease transmission is exacerbated by lack of clean drinking water and sanitation systems, inadequate food safety and hygiene conditions, lack of flood and drought protections, and interactions with other risks such as cholera outbreaks, food insecurity and infrastructure damage. Climate change threatens the progress that has been made towards reducing the burden of diarrhoea. For example, in Sub-Saharan Africa, while overall diarrhoea rates are expected to continue to decline (GBD 2016 Diarrhoeal Disease Collaborators, 2018), warming in 2030 (relative to the late 20th century) is projected to lead to diarrhoeal deaths in children under 15 equivalent to a CMR increase of 0.56/10,000 yr −1 (based on population projections for the region and age group; UN, 2020; [[#WHO--2014|WHO, 2014]] ). In China, by 2030, climate change could delay progress towards reducing waterborne disease burden by 8–85 months ( [[#Hodges--2014|Hodges et al., 2014]] ).

<div id="16.5.2.3.6" class="h4-container"></div>

<span id="risk-to-food-security-rkr-f"></span>
===== 16.5.2.3.6 Risk to food security (RKR-F) =====

<div id="h4-10-siblings" class="h4-siblings"></div>

Climate change affects food security primarily through impacts on food production, including crops, livestock and fisheries, as well as disruptions in food supply chains, linked to global warming, drought, flooding, precipitation variability and weather extremes ( [[#Myers--2017|Myers et al., 2017]] ; [[#FAO--2018|FAO et al., 2018]] ; [[#Mbow--2019|Mbow et al., 2019]] ). This RKR builds on Key Risks identified primarily in the Food, Fibre and Other Ecosystem Products Chapter, some sectoral (Health), and regional (Africa, Australasia, Central and South America, North America) chapters, as well as SR15, SRCCL and SROCC.

The severity of the risk to food security is defined here using a combination of criteria including the magnitude and likelihood of adverse consequences, affecting tens to hundreds of millions of people, timing of the risk and ability to respond to the risk. In this assessment, we use the number of undernourished people as a proxy outcome of these dimensions and their multiple interactions.

Climate change will pose severe risks in terms of increasing the number of undernourished people, affecting tens to hundreds of million people under High vulnerability and High warming, particularly among low-income populations in developing countries ( ''high confidence'' ). Extreme weather events will increase risks of undernutrition even on a regional scale, via spikes in food price and reduced income ( ''high confidence'' ) ( [[#FAO--2018|FAO et al., 2018]] , Hickey and Unwin, 2020; [[#Mbow--2019|Mbow et al., 2019]] ). The timing of these impacts and our ability to respond to them vary based on the level of GHG emissions and Shared Socioeconomic Pathways (SSP).. Under a low vulnerability development pathway (SSP1), climate change starts posing a moderate risk to food security above 1°C of global warming (i.e., impacts become detectable and attributable to climate-related factors), while beyond 2.5°C the risk becomes high (widespread impacts on larger numbers or proportion of population or area, but with the potential to adapt or recover) ( [[#Hurlbert--2019|Hurlbert et al., 2019]] ). Under high vulnerability–high warming scenario (i.e., SSP3-RCP6.0), up to 183 million additional people are projected to become undernourished in low-income countries owing to climate change by 2050 ( [[#Mbow--2019|Mbow et al., 2019]] ). Climate-related changes in food availability and diet quality are estimated to result in a crude mortality rate of about 54 deaths per million people with about 2°C warming by 2050 (SSP2, RCP8.5), most of them projected to occur in South and East Asia (67–231 deaths per million depending on the country) ( [[#Springmann--2016|Springmann et al., 2016]] ). In a medium vulnerability–high warming scenario (SSP2, RCP6.0), [[#Hasegawa--2018|Hasegawa et al. (2018)]] project that the number of undernourished people increases by 24 million in 2050, compared with outcomes without climate change and accounting for the CO 2 fertilisation effect. This number increases by around 78 million in a low-warming scenario (RCP2.6) accounting for the impacts of both climate change and mitigation policies. Caveats to these modelling studies are that most models (crop models in particular) are designed for long-term change in climate but not suited to project the impacts of short-term extreme events. The inclusion of adaptation measures into modelling estimates remains selective and partial.

Climate change risks of micronutrient deficiency will become severe in high-vulnerability development pathways and in the absence of societal adaptation, leading to hundreds of millions of additional people lacking key nutrients for atmospheric CO 2 levels above 500 ppm ( ''high confidence'' ) ( [[#Myers--2017|Myers et al., 2017]] ; [[#Nelson--2018|Nelson et al., 2018]] ; [[#Mbow--2019|Mbow et al., 2019]] ). For example, concentration of many micronutrients (e.g., phosphorus, potassium, calcium, sulphur, magnesium, iron, zinc, copper and manganese) can decrease by 5–10% under atmospheric CO 2 concentrations of 690 ppm (3.5°C warming). The decline in zinc content is projected to lead to an additional 150–220 million people affected by zinc deficiency with increases in existing deficiencies in more than 1 billion people ( [[#Myers--2017|Myers et al., 2017]] ). Similarly, decrease in protein and micronutrient content in rice due to a higher CO 2 concentration (568–590 ppm) can lead to 600 million people with rice as a staple at risk of micronutrient deficiency by 2050 ( [[#Zhu--2018|Zhu et al., 2018]] ). Additionally, the impact on protein content of increased CO 2 concentration (&gt;500 ppm) can lead an additional 150 million people with protein deficiency by 2050 (within the total of 1.4 billion people with protein deficiency) in comparison with the scenario without increased CO 2 concentration ( [[#Medek--2017|Medek et al., 2017]] ).

<div id="16.5.2.3.7" class="h4-container"></div>

<span id="risk-to-water-security-rkr-g"></span>
===== 16.5.2.3.7 Risk to water security (RKR-G) =====

<div id="h4-11-siblings" class="h4-siblings"></div>

Water security encompasses multiple dimensions: water for sanitation and hygiene, food production, economic activities, ecosystems, water-induced disasters, and use of water for cultural purposes (Chapter 4; Box 4.1; [[IPCC:Wg2:Chapter:Chapter-4#4.6.1|Section 4.6.1]] ). Water security risks are a combination of water-related hazards such as floods, droughts and water quality deterioration, and exposure of vulnerable groups exposed to too little, too much or contaminated water. Reasons for these can include both environmental conditions and issues of safety and access influenced by effectiveness of water governance ( [[#Sadoff--2020|Sadoff et al., 2020]] ). These are manifest through loss of lives, property, livelihoods and culture, and impacts on human health and nutrition, ecosystems and water-related conflicts which in turn can drive forced human displacement.

This RKR focuses on three types of risks with the potential to become severe: those associated with water scarcity, those driven by water-related disasters, and those impacting indigenous and traditional cultures and ways of life. Risk to water security constitutes a potentially severe risk because climate change could impact the hydrologic cycle in ways that would lead to substantial consequences for the health, livelihoods, property and cultures of large numbers of people. For those associated with water scarcity, ‘severe’ refers to magnitude (number of people in areas where water scarcity falls below recognised thresholds for adequate water supply per capita), along with the likelihood of unforeseen increases in water scarcity that outpace the ability to prepare for the increased risk by putting in place new large-scale infrastructure within the required time scale. For those associated with extreme events, ‘severe’ refers to magnitude (numbers of people affected, including deaths, physical health impacts including disease, mental health impacts, loss of livelihoods, loss of or damage to property) and timing (e.g., events coinciding with other stresses, e.g., a pandemic occurring at a time when local infrastructures are weakened by an extreme weather event). Important water-related extreme events include river flooding caused by heavy and/or prolonged rainfall, glacial lake outburst floods, and droughts. For those impacting cultures, ‘severe’ refers to the loss of key aspects of traditional ways of life. This includes consequences of the above two KRs.

Risks associated with water scarcity have the potential to become severe based on projections of large numbers of people becoming exposed to low levels of water availability per person, where ‘water availability’ includes fresh water in the landscape, including soil moisture and streamflows, available for all uses including agriculture as a dominant sector. Approximately 1.6 billion people currently experience ‘chronic’ water scarcity, defined as the availability of less than 1000 m 3 of renewable sources of fresh water per person per year ( [[#Gosling--2016|Gosling and Arnell, 2016]] ). In this context, we define a severe outcome as an additional 1 billion people experiencing ‘chronic’ water scarcity, relating to all uses of water, representing an increase of a magnitude comparable to current levels. The global number of people experiencing chronic water scarcity is projected to increase by approximately 800 million to 3 billion for 2°C global warming, and up to approximately 4 billion for 4°C global warming, considering the effects of climate change alone, with a 9 billion population ( [[#Gosling--2016|Gosling and Arnell, 2016]] ). Severe outcomes are projected to occur even with no changes in exposure: present-day exposure is defined here as ‘medium’ since either an increase or decrease in exposure could be possible. Vulnerability is not quantified in the literature assessed here, so in this assessment it is considered that severe outcomes could occur with present-day levels of vulnerability, again defined here as ‘medium’. Particularly severe outcomes (i.e., the high end of these ranges) are driven by regional patterns of climate change bringing severe reductions in precipitation and/or high levels of evapotranspiration in the most highly populated regions, leading to very substantial reductions in water availability compared with demand. There is strong consensus across models that water scarcity is projected to increase across substantial parts of the world even though projections disagree on which specific areas would see this impact. Moreover, a projected decrease in water scarcity in some regions does not prevent the increase in water scarcity in other regions becoming severe. Hence there is ''high confidence'' that risks to water scarcity have the potential to become severe due to climate change. Consequences of water scarcity include potential competition and conflicts between water users ( [[#Vanham--2018|Vanham et al., 2018]] ), damaging livelihoods, hindering socioeconomic development and reducing human well-being, for example through malnutrition resulting from inadequate water supplies leading to long-term health impacts such as child stunting ( [[#Cooper--2019|Cooper et al., 2019]] ). The avoidance of these consequences at high levels of water scarcity would require transformational adaptations including large-scale interventions such as dams and water transfer infrastructure ( [[#Greve--2018|Greve et al., 2018]] ). Since these require many years or even decades for planning and construction, and are also costly and irreversible and can potentially lead to lock-in and maladaptation, the potential for inadequate policy decisions made in the context of high uncertainties in regional climate changes brings the risk of a shortfall in adaptation. Around 2050, at approximately 2°C global warming, the risk of a substantial adaptation shortfall and hence severe outcomes for water scarcity have a relatively high likelihood across large parts of the southern USA and Mexico, northern Africa, parts of the Middle-East, northern China, and southern Australia, as well as many parts of Northwest India and Pakistan ( [[#Greve--2018|Greve et al., 2018]] ).

Risks associated with water-related extreme events and disasters have the potential to become severe based on projections of large numbers of people or high values of assets being affected. The risks to people from disasters can often only be quantified in terms of the hazard and exposure (the number of people affected), rather than the full consequences such as number of deaths, injuries or other health outcomes, as these often depend on complex or unpredictable factors such as the effectiveness of emergency and humanitarian responses or the access to healthcare. With approximately 50 million people per year currently affected by flooding ( [[#Alfieri--2017|Alfieri et al., 2017]] ), we define severe outcomes as more than 100 million people affected by flooding. At 2°C global warming, between approximately 50 million and 150 million people are projected to be affected by flooding, with figures rising to 110 million to 330 million at 4°C global warming. These projections assume present-day population and no additional adaptation, so no changes in exposure. Increased flood risk is projected by the WHO to lead to an additional 48,000 deaths of children under 15 years due to diarrhoea by 2030, with Sub-Saharan Africa impacted the most ( [[#WHO--2014|WHO, 2014]] ). Other consequences of floods that already occur include deaths by drowning, loss of access to fresh water, vector-borne diseases, mental health impacts, loss of livelihoods and loss of or damage to property. Many of these consequences depend on the vulnerability of individuals, households or communities to flooding impacts, for example through the presence or absence of measures to safeguard health and livelihoods, such as through infrastructure services, insurance or community support. The risks associated with these consequences could increase if there were no local adaptations to counter the effect of increased levels of hazard by reducing exposure and/or vulnerability. Climate-related changes to extreme events that would lead to these severe outcomes include increased frequency and/or magnitude of river floods of flash floods due to heavy or long-lasting precipitation, rapid snowmelt, or catastrophic failure of glacial lake moraine dams. These climate conditions are projected to increase with global warming.

Risks to cultural uses of water can become severe if there is permanent loss of aspects of communities’ cultures due to changes in water, including loss of areas of ice or snow with spiritual meanings, loss of culturally important places of access to such places, and loss of culturally important subsistence practices including by Indigenous People (Chapter 4). This includes mountain regions where changes in the cryosphere are having profound impacts (Cross-Chapter Paper 5). In these cases, severe outcomes would be defined locally rather than globally. Communities that lost a dominant environmental characteristic deeply associated with its cultural identity would be considered to be severely impacted. For example, due to the central role that travel on sea ice plays in the life of Inuit communities, providing freedom and mental well-being, loss of sea ice can be argued to represent environmental dispossession of these communities ( [[#Durkalec--2015|Durkalec et al., 2015]] ). Traditional ways of life are therefore threatened, and resulting changes would be transformative rather than adaptive. Similarly, changes in streamflow affecting the availability of species for traditional hunting can also negatively impact Indigenous communities (Norton-Smith et al.). Such changes are already being seen at current levels of warming, but studies remain somewhat limited in number, so this assessment is assigned ''medium confidence'' because of ''medium evidence'' and medium agreement. WGI conclude that it is ''virtually certain'' that further warming will lead to further reductions in Northern Hemisphere snow cover, and mass loss in individual glacier regions is projected to be between approximately 30% and 100% by 2100 under high-warming scenarios (Chapter 4). Streamflows are projected to change in most major river basins worldwide by several tens of percent at 4°C global warming (Chapter 4).

There is strong potential for increases in water scarcity, flooding, loss of snow and ice and changes in water bodies to lead to severe outcomes such as deaths from water-related diseases, drowning and starvation, long-term health impacts arising from malnutrition and diseases, loss of property, loss of existence or access to places of cultural significance, loss of livelihoods and loss of aspects of culture especially for Indigenous People with traditional lifestyles. The numbers of people affected are projected to range from hundreds of millions to several billion, depending on the level of global warming and socioeconomic futures. A key aspect of the risk is the high uncertainty in future regional precipitation changes in many regions of high vulnerability, including the potential for large and highly impactful changes, for which it may not be possible to provide adaptation measures before they become needed, leading to a high likelihood of adaptation deficits.

<div id="16.5.2.3.8" class="h4-container"></div>

<span id="risks-to-peace-and-to-human-mobility-rkr-h"></span>
===== 16.5.2.3.8 Risks to peace and to human mobility (RKR-H) =====

<div id="h4-12-siblings" class="h4-siblings"></div>

This RKR includes risks to peace within and among societies from armed conflict as well as risks to human mobility, epitomised by involuntary migration and displacement within and across state borders and involuntary immobility. Breakdown of peace and the inability of people to choose to move or stay challenge core elements of human security ( [[#Adger--2014|Adger et al., 2014]] ). Risks to peace also inform the agency and viability of mobility decisions. However, evidence does not indicate that human mobility constitutes a general risk to peace.

Breakdown of peace, materialised as overt or covert violence across social and spatial scales, constitutes a key risk because of its potential to cause widespread loss of life, livelihood and well-being. Such impacts are considered severe if they result in at least 1000 excess battle-related deaths in a country in a year. This threshold is consistent with the conventional definition of war ( [[#Pettersson--2020|Pettersson and Öberg, 2020]] ). However, because armed conflict routinely causes significant material destruction, triggers mass displacement, threatens health and food security, and undermines economic activity and living standards ( [[#Baumann--2016|Baumann and Kuemmerle, 2016]] ; [[#FAO--2017|FAO et al., 2017]] ; [[#de%20Waal--2018|de Waal, 2018]] ), risks to peace can be considered severe also when conflict has cascading effects on other aspects of well-being and amplifies vulnerability to other RKRs. Beyond the magnitude of such impacts, the rapidity with which armed conflict can escalate and the challenges of ending violence once it has broken out imply potentially very limited time and ability to respond for populations at risk.

Mobility is a universal strategy for pursuing well-being and managing household risks ( [[IPCC:Wg2:Chapter:Chapter-7#7.2.6|Section 7.2.6]] ; Cross-Chapter Box MIGRATE in Chapter 7, [[#UN--2018|UN, 2018]] ) and, where it occurs in a safe and orderly fashion, can reduce social inequality and facilitate sustainable development (Franco [[#Gavonel--2021|Gavonel et al., 2021]] ). Involuntary mobility constitutes a key risk because it implies reduced human agency with high potential for significant economic losses and non-material costs, an unequal gender burden, and amplified vulnerability to other RKRs ( [[#Schwerdtle--2018|Schwerdtle et al., 2018]] ; [[#Adger--2020|Adger et al., 2020]] ; [[#Maharjan--2020|Maharjan et al., 2020]] ; [[#Piggott-McKellar--2020|Piggott-McKellar et al., 2020]] ). Climate change also may erode or overwhelm human capacity to use mobility as a coping strategy, producing involuntarily immobile populations ( [[#Adams--2016|Adams, 2016]] ). A severe impact is when a large share of an affected population is forcibly displaced or prevented from moving, relative to normal mobility patterns, at local to global scale. However, because mobility may be a favourable mechanism for reducing risk or an adverse outcome of risk, depending on the circumstances under which it occurs, it is not possible to specify a simple quantitative threshold for when impacts become severe.

Complex causal pathways and lack of long-term projection studies presently prevent making confident quantitative judgements about how risks to peace and human mobility will materialise in response to specific warming levels, development pathways and adaptation scenarios. Literature concludes with ''medium confidence'' that risks to peace will increase with warming, with the largest impacts expected in weather-sensitive communities with low resilience to climate extremes and high prevalence of underlying risk factors ( [[#Theisen--2017|Theisen, 2017]] ; [[#Busby--2018|Busby, 2018]] ; [[#Koubi--2019|Koubi, 2019]] ; [[#von%20Uexkull--2021|von Uexkull and Buhaug, 2021]] ). However, climate-driven impacts on societies will depend critically on future political and socioeconomic development trajectories ( ''limited evidence'' , ''high agreement'' ), suggesting that risks due to climate change are relevant primarily for highly vulnerable populations and for pessimistic development scenarios. Overall risks to peace may decline despite warming if non-climatic determinants are reduced sufficiently in the future.

Regular human mobility will continue regardless of climate change, but mobility-related risks will increase with warming, notably in densely populated hazard-prone regions, in small islands and low-lying coastal zones, and among populations with limited coping capacity (RKR-A; Section [https://www.ipcc.ch/chapter/16#CCP2.2 CCP2.2.2] ; Chapter 7) ( ''high confidence'' ). Such risks can become severe even with limited levels of warming for populations with low adaptive capacity and whose settlements and livelihoods are critically sensitive to environmental conditions ( ''medium evidence'' , ''high agreement'' ). Likewise, risk of involuntary immobility could become severe for highly vulnerable populations with limited resources, even with moderate levels of warming ( ''limited evidence'' , ''high agreement'' ). Critically, population growth and shifting exposure will interact with warming to shape these risks ( [[#Davis--2018|Davis et al., 2018]] ; [[#Hauer--2020|Hauer et al., 2020]] ; [[#Robinson--2020a|Robinson, 2020a]] ). Although climate-driven human mobility generally does not increase risks to peace ( ''medium confidence'' ), armed conflict is a major driver of forced displacement ( ''high confidence'' ).

Expert elicitation estimates that 4°C warming above pre-industrial levels will have severe and widespread effects on armed conflict with 26% probability, assuming no change from present levels in non-climatic drivers ( [[#Mach--2019|Mach et al., 2019]] ). That judgement refers to impacts that exceed the threshold for severity considered here, suggesting that global warming of 4°C would produce severe risks to peace under present societal conditions ( ''low confidence'' ). Future risks to peace will remain strongly influenced by socioeconomic development ( [[#Hegre--2016|Hegre et al., 2016]] ). A study of Sub-Saharan Africa that accounts for both temperature and socioeconomic changes, 2015–2065, concludes that determinants other than rising temperatures, notably quality of governance, will remain most influential in shaping overall levels of violence even in the high-warming RCP8.5 scenario ( [[#Witmer--2017|Witmer et al., 2017]] ).

A larger empirical literature offers indirect evidence that climate change may produce severe risks to peace within this century by demonstrating how climate variability and extremes affect contemporary conflict dynamics, especially in contexts marked by low economic development, high economic dependence on climate-sensitive activities, high or increasing social marginalisation, and fragile governance ( ''medium confidence'' ) (Sections 7.2.7, 16.2, [[#Schleussner--2016a|Schleussner et al., 2016a]] ; [[#Von%20Uexkull--2016|Von Uexkull et al., 2016]] ; [[#Busby--2018|Busby, 2018]] ; [[#Harari--2018|Harari and Ferrara, 2018]] ; [[#Ide--2020|Ide et al., 2020]] ; [[#Scartozzi--2020|Scartozzi, 2020]] ).

Climatic risks interact with economic, political and social drivers to create risks to human mobility both directly (through the threat of physical harm and destruction of property and infrastructure) and indirectly (via adverse impacts on livelihood and well-being). Extreme weather events are leading causes of forced displacement (Cross-Chapter Box MIGRATE in Chapter 7, [[#IDMC--2020|IDMC, 2020]] ). Projected increases in the frequency and severity of extreme events ( [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ) in combination with future population growth in hazard-prone regions (e.g., [[#Merkens--2016|Merkens et al., 2016]] ) suggest that risks to mobility will increase in response to future global warming ( [[#Robalino--2015|Robalino et al., 2015]] ; [[#Davis--2018|Davis et al., 2018]] ; [[#Rigaud--2018|Rigaud et al., 2018]] ). For example, moving from RCP2.6 to RCP8.5 (entailing ~0.5°C additional global warming by 2050) is projected to increase internal migration by 2050 from 51 [31–72] million to 118 [92–143] million people across South Asia, Latin America and Africa ( [[#Rigaud--2018|Rigaud et al., 2018]] ), although those estimates principally comprise migrants, whose decisions are also informed by non-climatic drivers, rather than involuntarily displaced people. Global levels of flood displacement are estimated to increase by 50% with each 1°C warming ( [[#Kam--2021|Kam et al., 2021]] ). Should future warming reduce adaptation options for vulnerable populations ( [[#16.4|Section 16.4]] ), a consequence may be higher levels of involuntary migration and immobility ( [[#Grecequet--2017|Grecequet et al., 2017]] ; [[#Otto--2017|Otto et al., 2017]] ). There is little evidence that climate-driven mobility negatively affects peace ( [[#Brzoska--2016|Brzoska and]] [[#Fröhlich--2016|Fröhlich, 2016]] ; [[#Burrows--2016|Burrows and Kinney, 2016]] ; [[#Freeman--2017|Freeman, 2017]] ; [[#Petrova--2021|Petrova, 2021]] ).

There is ''high agreement'' that even moderate levels of future SLR will severely amplify involuntary migration and displacement in small islands and densely populated low-lying coastal areas in the absence of appropriate adaptive responses ( ''high confidence'' ) ( [[#Hauer--2017|Hauer, 2017]] ; [[#IPCC--2019b|IPCC, 2019b]] ; [[#Hauer--2020|Hauer et al., 2020]] ; [[#McMichael--2020|McMichael et al., 2020]] , Sections 15.3.4, 16.4). In some contexts, climate change also may accelerate migration towards high-exposure coastal areas ( [[#Bell--2021|Bell et al., 2021]] ). Under a high-emissions RCP8.5 scenario (global median 0.7 m SLR by 2100), the number of people exposed to annual coastal flooding may more than double by 2100 compared with present numbers ( [[#Kulp--2019|Kulp and Strauss, 2019]] ). In the USA alone, SLR of 0.9 m could potentially put 4.2 million people at risk of inundation by the end of this century ( [[#Hauer--2017|Hauer, 2017]] ). However, number of people exposed to SLR does not evenly translate to forcibly displaced populations ( [[#Hauer--2020|Hauer et al., 2020]] ). Ascertaining how many people will move forcibly or as an adaptive response to SLR is inherently challenging because of the complex and highly individual nature of migration decisions ( [[#Black--2013|Black et al., 2013]] ; [[#Boas--2019|Boas et al., 2019]] ; [[#Piguet--2019|Piguet, 2019]] ; [[#Bell--2021|Bell et al., 2021]] ). Implications of climate change for risks to human mobility across borders are even harder to quantify and highly uncertain, due to unknown developments in legal and political conditions that govern international migration ( [[#McLeman--2019|McLeman, 2019]] ; [[#Wrathall--2019|Wrathall et al., 2019]] ).

<div id="16.5.2.4" class="h3-container"></div>

<span id="synthesis-of-the-assessment-of-representative-key-risks"></span>
==== 16.5.2.4 Synthesis of the Assessment of Representative Key Risks ====

<div id="h3-35-siblings" class="h3-siblings"></div>

Figure 16.10 provides a synthesis of the RKRs and the conditions that lead to severe risks over the course of the 21st century, as assessed in Sections 16.5.2.3.1–16.5.2.3.8 (see Table SM16.14 for further description). It identifies sets of conditions—defined by levels of warming, exposure/vulnerability and adaptation—that would produce severe risk with a particular level of confidence. The risks are of two scopes: broadly applicable, meaning that the risks described by a particular KR or RKR would be severe pervasively and even globally; and specific, meaning that these risks would apply to particular areas, sectors or groups of people.

<div id="_idContainer035" class="Figure"></div>

[[File:1c46a100d72df7360311a29f3018687b IPCC_AR6_WGII_Figure_16_010.png]]

'''Figure 16.10 |''' '''Synthesis of the severity conditions for Representative Key Risks by the end of this century.''' The figure does not aim to describe severity conditions exhaustively for each RKR, but rather to illustrate the risks highlighted in this report (Sections 16.5.2.3.1 to 16.5.2.3.8). Coloured circles represent the levels of warming (climate), exposure/vulnerability, and adaptation that would lead to severe risks for particular key risks and RKRs. Each set of three circles represents a combination of conditions that would lead to severe risk with a particular level of confidence, indicated by the number of black dots to the right of the set, and for a particular scope, indicated by the number of stars to the left of the set. The two scopes are ‘broadly applicable’, meaning applicable pervasively and even globally, and ‘specific’, meaning applicable to particular areas, sectors or groups of people. Details of confidence levels and scopes can be found in [[#16.5.2.3|Section 16.5.2.3]] . In terms of severity condition levels ( [[#16.5.2.3|Section 16.5.2.3]] ), for warming levels (coloured circles labelled ‘C’ in the figure), High refers to climate outcomes consistent with RCP8.5 or higher, Low refers to climate outcomes consistent with RCP2.6 or lower, and Medium refers to intermediary climate scenarios. Exposure-Vulnerability levels are determined by the RKR teams relative to the range of future conditions considered in the literature. For Adaptation, High refers to near maximum potential and Low refers to the continuation of today’s trends. Despite being intertwined in reality, Exposure-Vulnerability and Adaptation conditions are distinguished to help understand their respective contributions to risk severity.

<div id="Five" class="h4-container"></div>

<span id="five-main-messages-arise-from-this-synthesis"></span>
===== Five main messages arise from this synthesis: =====

<div id="h4-13-siblings" class="h4-siblings"></div>

Severe risk is rarely driven by a single determinant (warming, exposure/vulnerability, adaptation), but rather by a combination of conditions that jointly produce the level of pervasiveness of consequences, irreversibility, thresholds, cascading effects, likelihood of consequences, temporal characteristics of risk and the systems’ ability to respond ( ''medium'' to ''high confidence'' ). In other words, climate risk is not a matter of changing CIDs only, but of the confrontation between changing CIDs and changing socio-ecological conditions.

In most of the RKRs, severe risk for broadly applicable situations requires high levels of warming or exposure/vulnerability, or low adaptation. In many cases, it is associated with several of these conditions occurring simultaneously (e.g., high warming and high vulnerability). Examples include low-lying coastal areas (RKR-A; ''medium confidence'' ), loss of livelihoods (RKR-D; ''medium confidence'' ) or armed conflicts (RKR-H; ''low confidence'' ).

High warming and exposure/vulnerability combined with low adaptation is, however, not necessarily required to lead to severe risk, and various other sets of conditions can lead to such an outcome. For example:

''Without high levels of warming'' . T his is especially the case for terrestrial and marine ecosystems (RKR-B) and water security (RKR-G) for which even medium to low levels of warming will generate severe risk, depending on the processes considered (e.g., mass population-level mortality and ecological disruption for ecosystems). This is also the case when more specific situations are considered, for example in the case of (in)voluntary mobility of vulnerable populations with limited resources (RKR-H), and for some critical infrastructure in already highly exposed and vulnerable contexts (RKR-C).

''With high levels of adaptation.'' H igh levels of adaptation will not necessarily avoid severe risk, as is illustrated by the cases of coral-dependent and arctic coastal communities (RKR-A), some terrestrial and marine ecosystems (RKR-B), and water scarcity and the cultural uses of water (RKR-G).

All RKR assessments indicate that risks are higher in high-vulnerability development pathways, and in some cases high vulnerability can occur in high-income societies. Examples include the possibility of increasing coastal settlement and the location of critical infrastructure in highly exposed locations (RKR-A, RKR-C), including to floods (RKR-G) and risks to terrestrial and marine ecosystems (RKR-B). The assessment therefore shows that, depending on socioeconomic trends especially in terms of equity, social justice and income sustainability, as well as on the ability to shift towards more climate-resilient economic and settlement systems (e.g., at the coast), higher-income societies also are at serious risk of being substantially affected in the decades to century to come.

In terms of the time frames, most of the RKRs conclude that severe risks to many dimensions (ecosystems, health, etc.) are expected to occur by the end of the 21st century and across the globe. Some RKRs, however, highlight that severe risk could occur far earlier, for example as soon as a warming level of 1.5°C or 2°C is reached, which means potentially well before mid-century ( [[#IPCC--2021|IPCC, 2021]] ). In some cases, risks are already considered severe, for example after major climatic events such as tropical storms (RKR-A).

<div id="16.5.3" class="h2-container"></div>

<span id="variation-of-key-risks-across-levels-of-global-warming-exposure-and-vulnerability-and-adaptation"></span>
=== 16.5.3 Variation of Key Risks across Levels of Global Warming, Exposure and Vulnerability, and Adaptation ===

<div id="h2-16-siblings" class="h2-siblings"></div>

This section builds on Sections 16.5.1 and 16.5.2 as well as on additional literature to illustrate how consequences associated with KRs and RKRs are projected to vary with three types of determinants: global average warming level, as a proxy for associated changes in climate hazards (CIDs, [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ); socioeconomic development pathway, as a means of capturing alternative future exposure and vulnerability conditions; and level of adaptation to reflect the extent to which successful adaptation is implemented. While these three dimensions are partly intertwined—for example, warming and adaptation scenarios are constrained by development pathways (Chapter 18)—this section assesses the influence of each dimension separately (Sections 16.5.3.2–16.5.3.4) to highlight how sensitivity varies across these dimensions for different KRs and RKRs. We then bring the dimensions together in an illustrative example (large deltas; [[#16.5.3|Section 16.5.3.5]] ).

<div id="16.5.3.1" class="h3-container"></div>

<span id="warming-level-including-risks-avoided-by-mitigation"></span>
==== 16.5.3.1 Warming Level, Including Risks Avoided by Mitigation ====

<div id="h3-36-siblings" class="h3-siblings"></div>

Studies illustrating sensitivity to warming level typically do so by contrasting projected impacts for the same socioeconomic conditions but different climate pathways or temperature levels, often based on Representative Concentration Pathways (RCPs) ( [[#van%20Vuuren--2014|van Vuuren and Carter, 2014]] ). We refer to future climate conditions either based on their global average warming level or as a ‘high warming’ scenario (based on RCP8.5), medium warming (RCP4.5 or RCP6.0) or low warming (RCP2.6 or 1.5°C scenarios). Because some of these scenarios assume no or minimal mitigation (RCP8.5, RCP6.0) while others do (RCP4.5, RCP2.6), differences in outcomes between them reflect risks avoided by mitigation (assuming consistent socioeconomic assumptions).

Some ecological risks (Chapter 2) are particularly sensitive to warming. For example, warm-water coral reefs are already experiencing High risk levels and are expected to face Very High risks under 1.5°C of global warming ( [[#Hoegh-Guldberg--2018a|Hoegh-Guldberg et al., 2018a]] ; [[#Bindoff--2019|Bindoff et al., 2019]] ). Some societal risks, such as human mortality due to extreme heat, also are sensitive to warming. A medium-warming scenario (relative to high warming) reduces projected global average mortality due to heat from seven deaths per 10,000 people yr –1 (7/10,000 yr −1 ) by 2100 to ~1/10,000 yr −1 , assuming high-vulnerability societal conditions ( [[#Carleton--2020|Carleton et al., 2020]] ). At the national level, without considering adaptation, reductions in a broader measure of mortality are projected across a range of countries including Colombia, the Philippines, and several in Europe ( [[#Guo--2018|Guo et al., 2018]] ), and exposure of the US population to high-mortality heatwaves is reduced by nearly half ( [[#Anderson--2018a|Anderson et al., 2018a]] ). Without considering changes in exposure or vulnerability, warming of 1.5–2°C (compared with 4–5°C) reduces global mortality impacts from an increase of 2.1–13.0% to 0.1–2.2% ( [[#Gasparrini--2017|Gasparrini et al., 2017]] ; [[#Vicedo-Cabrera--2018a|Vicedo-Cabrera et al., 2018a]] ) and impacts in China from up to 4/10,000 yr −1 ( [[#Weinberger--2017|Weinberger et al., 2017]] ) to 0.3–0.5/10,000 yr −1 ( [[#Wang--2019|Wang and Hijmans, 2019]] ).

A low-warming scenario (relative to high warming) reduces aggregate economic impacts from around 7% of global GDP to less than 1% (Takakura et al., 2019), and changes impacts on the number of people suffering from hunger from an increase (by 7–55 million) to a decrease (by up to 6 million) ( [[#Janssens--2020|Janssens et al., 2020]] ). Low versus high warming also reduces the coastal population at risk of flooding due to SLR from tripling by 2100 (relative to today) to doubling ( [[#Kulp--2019|Kulp and Strauss, 2019]] , [[#16.5.2.3|Section 16.5.2.3.2]] ). The SROCC estimates that SLR risks are reduced from Moderate-to-High to Moderate for large tropical agricultural deltas and resource-rich megacities, and from High and Very High to Moderate-to-High for Arctic human communities and urban atoll islands, respectively ( [[#Oppenheimer--2019|Oppenheimer et al., 2019]] ).

Higher levels of warming are projected to also generate higher income inequality between countries (e.g., [[#Pretis--2018|Pretis et al., 2018]] ; Takakura et al., 2019) as well as within them ( [[#Hallegatte--2016|Hallegatte et al., 2016]] ) even though other drivers will be more important ( [[#16.5.2.3.5|Section 16.5.2.3.5]] ). Similarly, climate and weather events are expected to play an increasing role in shaping risks to peace ( ''limited evidence, medium agreement)'' and migration ( ''medium evidence, high agreement'' ) in the future, but uncertainty is high due to complex causal pathways and non-climate factors likely dominate outcomes ( [[#16.5.2.3.8|Section 16.5.2.3.8]] ). There is ''high agreement'' that future SLR will amplify levels of forced migration from small islands and low-lying coastal areas in the absence of appropriate adaptive responses ( [[#Oppenheimer--2019|Oppenheimer et al., 2019]] ).

A synthesis of risk assessments in the recent IPCC Special Reports ( [[#Magnan--2021|Magnan et al., 2021]] ) concludes that an integrated measure of today’s global climate risk level will increase by the end of this century by two- to four-fold under low and high warming, respectively (based on aggregated scores developed in the study). An additional comparison of risk levels under +1.5°C and +2°C suggests that every additional 0.5°C of global warming will increase the risk level by about a third.

<div id="16.5.3.2" class="h3-container"></div>

<span id="exposure-and-vulnerability-trends"></span>
==== 16.5.3.2 Exposure and Vulnerability Trends ====

<div id="h3-37-siblings" class="h3-siblings"></div>

Development pathways describe plausible alternative futures of societal change and are critical to future risks because they affect outcomes of concern both through non-climate and climate-related channels ( ''very high confidence'' ).

Studies illustrating sensitivity to development pathways typically do so by contrasting projected impacts for the same climate pathway or temperature level but different levels of socioeconomic exposure and vulnerability, for example based on SSPs ( [[#O’Neill--2014|O’Neill et al., 2014]] ; [[#Van%20Vuuren--2014|Van Vuuren et al., 2014]] ). Or, they infer sensitivity to future development pathways based on differences in impacts across current populations with different levels of exposure or vulnerability. We refer to future conditions based on SSPs 1 or 5 as ‘low exposure’ or ‘low vulnerability’ conditions, and those based on SSPs 3 or 4 as ‘high exposure’ or ‘high vulnerability’ conditions ( [[#O’Neill--2014|O’Neill et al., 2014]] ; [[#van%20Vuuren--2014|van Vuuren and Carter, 2014]] ).

A wide range of climate change impacts depend strongly on development pathway ( ''high confidence'' ). A low (relative to high) exposure future, determined by limited population growth and urbanisation, results in about 30% fewer people exposed to extreme heat globally ( [[#Jones--2018b|Jones et al., 2018b]] ) and about 50% fewer in Africa ( [[#Rohat--2019a|Rohat et al., 2019a]] ), similar to the effect of a medium versus high level of global warming. Low-exposure conditions also reduce the fraction of the population in Europe at very high risk of heat stress from 39% to 11% ( [[#Rohat--2019b|Rohat et al., 2019b]] ). Demographic differences lead to a reduction in the global population exposed to mosquitos acting as viral disease vectors by more than half ( [[#Monaghan--2018|Monaghan et al., 2018]] ) and exposure to wildfire risk by nearly half ( [[#Knorr--2016|Knorr et al., 2016]] ).

Studies are increasingly going beyond exposure to incorporate future vulnerability, finding that it is often the dominant determinant of risk ( ''high confidence'' ). A low (relative to high) vulnerability future reduces the risk to global poverty by an order of magnitude, robustly across approaches that account for macroeconomic growth, structural change in the economy, inequality, and access to infrastructure services ( [[#Hallegatte--2017|Hallegatte and Rozenberg, 2017]] ), or for the exposure of vulnerable populations to multi-sector climate-related risks ( [[#Byers--2018|Byers et al., 2018]] ). A low (relative to high) vulnerability future also reduces the global mean number of temperature-attributable deaths in 2080–2095 due to enteric infections by an order of magnitude (from &gt;80,000 to &lt;7000; ( [[#Chua--2021|Chua et al., 2021]] )). Low future socioeconomic vulnerability to flooding reduces global fatalities and economic losses by 69–96% ( [[#Jongman--2015|Jongman et al., 2015]] ). Low vulnerability as measured by indicators including per capita GDP, education, governance, water demand and storage potential reduces water insecurity by a factor of three ( [[#Koutroulis--2019|Koutroulis et al., 2019]] ). A scenario with reduced barriers to trade reduces the number of people at risk of hunger due to climate change by 64% ( [[#Janssens--2020|Janssens et al., 2020]] ). Structural transformation of the economy (shift of the workforce from highly exposed sectors such as agriculture and fishing to less exposed sectors such as services) lowers GDP impact projections by 25–30% in today’s developing countries by 2100 ( [[#Acevedo--2017|Acevedo et al., 2017]] ).

The IPCC SRCCL supports the importance of societal conditions to climate-related risk ( [[#Hurlbert--2019|Hurlbert et al., 2019]] ), concluding that risks of water scarcity in drylands (i.e., desertification), land degradation and food insecurity are close to High [[#footnote-000|3]] beginning at 1.5°C under high-vulnerability conditions (SSP3), but remain close to Moderate up to slightly above 2°C for low-vulnerability conditions (SSP1). Specifically, risk of water scarcity in drylands (i.e., desertification) at 1.5°C warming is reduced in low vulnerability (relative to high vulnerability) conditions from High to Medium. Similarly, under a 2°C warming, risk is reduced from High to Moderate for food security and High to Moderate-to-High for land degradation.

While climate change will increase risk to society and ecosystems, future exposure and vulnerability conditions will also greatly impact outcomes of concern directly. Global economic damages to coastal assets from tropical cyclones are projected to increase by more than 300% due to coastal development alone, a much larger effect than projected climate change impacts through 2100 even in RCP8.5 ( [[#Gettelman--2018|Gettelman et al., 2018]] ). Similarly, global crop prices are more than three times more sensitive to alternative assumptions about changes in production technologies and demand than to alternative climate outcomes ( [[#Ren--2016|Ren et al., 2016]] ). Future water scarcity is driven mainly by both demographic change and socioeconomic changes affecting water demand and management. A measure of between-country inequality (Gini coefficient) would decline by more than 50% this century in low-vulnerability conditions, but would double in a high-vulnerability future ( [[#Crespo%20Cuaresma--2017|Crespo Cuaresma, 2017]] ), outweighing the effect of climate ( [[#Taconet--2020|Taconet et al., 2020]] ). Similarly, the global prevalence of armed conflict will roughly double this century in a high-vulnerability future, whereas it will drop by half in a low-vulnerability future ( [[#Hegre--2016|Hegre et al., 2016]] ). In Sub-Saharan Africa, assumptions about governance and political rights are estimated to be far more important to the future risk of violent conflict than climate change ( [[#Witmer--2017|Witmer et al., 2017]] ).

<div id="16.5.3.3" class="h3-container"></div>

<span id="climate-adaptation-scenarios"></span>
==== 16.5.3.3 Climate Adaptation Scenarios ====

<div id="h3-38-siblings" class="h3-siblings"></div>

One approach to understand adaptation benefits for risk reduction is to contrast projected impacts for the same climate and development conditions but different levels of adaptation. For example, global-scale coastal protection studies considering both RCPs and SSPs suggest that, under a given RCP, the total flooded area may be reduced by 40% by using 1-m height dykes, compared with a no-adaptation baseline ( [[#Tamura--2019|Tamura et al., 2019]] ). The global cost of SLR over the 21st century can be lowered by factor of two to four if local cost–benefit decisions consider migration an adaptation option, in addition to hard protection ( [[#Lincke--2021|Lincke and Hinkel, 2021]] ). Under a low-warming scenario, it is estimated that adaptation (i.e., changes in crop variety and planting dates) could reduce the total number of people at risk of hunger globally by about 4%, and by about 10% in a high-warming scenario ( [[#Hasegawa--2014|Hasegawa et al., 2014]] ). Impacts on heat-related mortality would be cut from 10 to 7 deaths per 10,000 people yr –1 in 2100 by adaptation actions beyond those assumed to be driven by income growth ( [[#Carleton--2020|Carleton et al., 2020]] ). In a regional example, proactive adaptation efforts on infrastructure (especially roads, runways, buildings and airports) in Alaska, USA, could reduce damage-related expenditure by 45% under medium or high warming ( [[#Melvin--2017|Melvin et al., 2017]] ).

Another approach infers the potential future effectiveness of adaptation based on current sensitivity of impacts to interventions. For example, the future disease burden of malaria is likely to be highly dependent on the future development of health services, deployment of malaria programs and adaptation. Investments in water and sanitation infrastructure are also recognised to have the potential to reduce severe risks of waterborne disease, although these improvements likely need to provide transformative change ( [[#Cumming--2019|Cumming et al., 2019]] ). The potential for severe risks may also be substantially reduced through the development of vaccines for specific enteric diseases ( [[#Riddle--2018|Riddle et al., 2018]] ), although most current vaccines target viral pathogens, incidence for which tends to be inversely correlated with ambient temperature ( [[#Carlton--2016|Carlton et al., 2016]] ). In addition, international migration as well as forced movement of people across borders will be influenced by developments in legal and political conditions ( [[#McLeman--2019|McLeman, 2019]] ; [[#Wrathall--2019|Wrathall et al., 2019]] ), but the fact that these developments are unknown strongly limits any forecasts on the magnitude of adaptation benefits ( [[#16.5.2.3.8|Section 16.5.2.3.8]] ).

Last, there is growing concern that even ambitious adaptation efforts will not eliminate residual risks from climate change ( [[#16.4.2|Section 16.4.2]] ). A synthesis of risk assessments in the recent IPCC Special Reports ( [[#Magnan--2021|Magnan et al., 2021]] ) concludes that high societal adaptation is expected to reduce the aggregated score—the proxy used in the study—of global risk from anthropogenic climate change by about 40% under all RCPs by the end of the century, compared with risk levels projected without adaptation. It, however, also shows that, even for the lowest warming scenario, a residual risk one-third greater than today’s risk level would still remain (with a doubling of today’s aggregated score under the high-emissions scenario).

<div id="16.5.3.4" class="h3-container"></div>

<span id="illustration-risk-and-adaptation-pathways-in-densely-populated-and-agricultural-deltas"></span>
==== 16.5.3.4 Illustration: Risk and Adaptation Pathways in Densely Populated and Agricultural Deltas ====

<div id="h3-39-siblings" class="h3-siblings"></div>

Large deltas, which are very dynamic risk hotspots of global importance and interest ( [[#Wigginton--2015|Wigginton, 2015]] ; [[#Hill--2020|Hill et al., 2020]] ; [[#Nicholls--2020|Nicholls et al., 2020]] ), serve well to illustrate how risk pathways develop over time, determined by climatic as well as non-climatic risk drivers and by adaptation. Deltas occupy less than 0.5% of the global land area but host over 5% of the global population ( [[#Dunn--2019|Dunn et al., 2019]] ) and contribute major fractions of food production in many world regions ( [[#Kuenzer--2020|Kuenzer et al., 2020]] ). Future risk in these areas is heavily driven by climate change but also greatly depends on past, current and future socioeconomic changes which influence future trends in exposure, vulnerability and adaptive capacity of natural and human systems ( ''high confidence'' ) ( [[#Oppenheimer--2019|Oppenheimer et al., 2019]] ). From a risk perspective, trends over the past decades have been unfavourable for many deltas, as most of them have experienced a simultaneous intensification of hazards, rise in exposure and stagnation or only limited reduction in vulnerability, particularly in low-income countries ( ''high confidence'' ) ( [[#Day--2016|Day et al., 2016]] ; [[#Tessler--2016|Tessler et al., 2016]] ; [[#Loucks--2019|Loucks, 2019]] ; [[#Oppenheimer--2019|Oppenheimer et al., 2019]] ; [[#Hill--2020|Hill et al., 2020]] ).

<div id="16.5.3.4.1" class="h4-container"></div>

<span id="hazard-trends-in-deltas"></span>
===== 16.5.3.4.1 Hazard trends in deltas =====

<div id="h4-14-siblings" class="h4-siblings"></div>

Deltas face multiple interacting hazards, many of which over the past decades have been intensified by local and regional anthropogenic developments (e.g., the construction of dams, groundwater extraction, or agricultural irrigation practices) and most of which are expected to be exacerbated by climate change ( ''high confidence'' ) ( [[#Giosan--2014|Giosan et al., 2014]] ; [[#Tessler--2015|Tessler et al., 2015]] ; [[#Tessler--2016|Tessler et al., 2016]] ; [[#Arto--2019|Arto et al., 2019]] ; [[#Oppenheimer--2019|Oppenheimer et al., 2019]] ). The most important hazards include SLR, inundation, salinity intrusion, cyclones, storms and erosion, many of which occur in combination. The potential for flooding and inundation depends on the relative sea level rise (RSLR) which results from global and regional SLR as well as local subsidence within the deltas. Subsidence caused by natural and human drivers (mainly compaction and groundwater extraction) is currently the most important cause for RSLR in many deltas and can exceed the rate of climate-induced SLR by an order of magnitude ( [[#Oppenheimer--2019|Oppenheimer et al., 2019]] ). But in higher warming scenarios the relative importance of climate-driven SLR is expected to increase over time ( [[#Oppenheimer--2019|Oppenheimer et al., 2019]] ). In a global study covering 47 major deltas and assessing future trends of sediment delivery across four RCPs, three SSPs (1,2,3) and a projection of future dam construction, [[#Dunn--2019|Dunn et al. (2019)]] find most deltas (33 out of the 47) will experience a mean decline of 38% in sediment flux by the end of the century when considering the average of the scenarios. [[#Nienhuis--2020|Nienhuis et al. (2020)]] find in a global assessment that some deltas have gained land through increased sediment load (e.g., through deforestation), but recent land gains are unlikely to be sustained if SLR continues to accelerate. According to the latest assessments, it is ''virtually certain'' that global mean sea level will continue to rise over the 21st century, with SLR by 2100 ''likely'' to reach 0.28–0.55 m in a an SSP1–1.9 and 0.63–1.01 m in an SSP5–8.5 scenario relative to 1995–2014 ( [[#IPCC--2021|IPCC, 2021]] ). The combined effects of local subsidence and GMSL rise result in a significant increase in the potential for inundation of low-lying deltas across all RCPs, with some variation according to regional sea level change rates, without significant further adaptation measures ( ''very high confidence'' ).

In terms of salt-water intrusion and salinisation, global comparative studies are still lacking but the general processes are well understood (e.g., [[#White--2017|White and Kaplan, 2017]] ), and research on individual deltas is on the rise. In the Mekong Delta of Vietnam, one of the main rice-producing deltas globally, salinity intrusion has been observed to extend around 15 km inland during the rainy season and around 50 km during the dry season ( [[#Gugliotta--2017|Gugliotta et al., 2017]] ), resulting in rice yield losses of up to 4 t ha −1 yr −1 ( [[#Khat--2018|Khat et al., 2018]] ). SLR, along with the expansion of dams and dry season irrigation upstream, is expected to further increase the salinity intrusion into the delta. This creates additional risk for food production as rice and other crops might be pushed beyond their adaptation limits in terms of salt tolerance, potentially affecting many of the 282,000 agriculture-based livelihoods in the Mekong Delta and increasing the pressure for cost-intensive adaptation ( [[#Smajgl--2015|Smajgl et al., 2015]] ). [[#Genua-Olmedo--2016|Genua-Olmedo et al. (2016)]] find for the Ebro that in high scenario (RCP8.5, and SLR of almost 1 m by 2100), SLR-induced salinity intrusion will lead to almost a doubling of salinity levels and a decrease of mean rice productivity by over 20% in a high-SLR scenario with almost 1 m of SLR by the end of the century.

<div id="16.5.3.4.2" class="h4-container"></div>

<span id="exposure-trends-in-deltas"></span>
===== 16.5.3.4.2 Exposure trends in deltas =====

<div id="h4-15-siblings" class="h4-siblings"></div>

Next to the trends in hazards, future exposure of and in deltas is shaped particularly by the increase of population and infrastructure and the intensification of land use. Over the recent years, the population has been rising in major deltas, roughly along with overall national population trends ( [[#Szabo--2016|Szabo et al., 2016]] ). In 2017, 339 million people lived in deltas with a high exposure to flooding, cyclones and other coastal hazards ( [[#Edmonds--2020|Edmonds et al., 2020]] ). Over 40% of the global population exposed to flooding from tropical cyclones lived in deltas, more than 90% of which in developing countries and emerging economies (ibid.). Looking into the future, population in low-elevation coastal zones is expected to increase by 2050 across all SSPs with diverging developments in the second half of the century, and at the end of the century will reach well over 1 billion people in SSP3 ( [[#Jones--2016|Jones and O’Neill, 2016]] ; [[#Merkens--2016|Merkens et al., 2016]] ). A major part of this population is expected to reside in deltas with large cities or mega-urban agglomerations such as the Pearl River Delta, China. One of the first studies using the SSP-RCP framework on the delta scale suggests a strong increase in intensive agricultural land by the middle of the century in three SSPs (2, 3, 5) in the Volta Delta, Ghana, while the Mahanadi, India, and the Ganges–Brahmaputra–Meghna do not show a significant further increase ( [[#Kebede--2018|Kebede et al., 2018]] ). Hence, the amount of population and infrastructure as well as agricultural land is expected to rise further under certain SSPs, further increasing the exposure to future climate hazards.

<div id="16.5.3.4.3" class="h4-container"></div>

<span id="vulnerability-trends-in-deltas"></span>
===== 16.5.3.4.3 Vulnerability trends in deltas =====

<div id="h4-16-siblings" class="h4-siblings"></div>

Deltas are characterised by multi-faceted vulnerabilities of their environment and human populations. Over 200 indicators are being used in the literature to characterise and analyse vulnerability in deltas, spanning social, ecological and economic aspects ( [[#Sebesvari--2016|Sebesvari et al., 2016]] ). However, only a few studies model or dynamically assess trends in vulnerability, particularly for the future, at global scale, or take a comparative approach. But overall, a global trend assessment suggests that social vulnerability to climate hazards has been improving over the past years in all world regions hosting major deltas apart from Oceania, yet with emerging economies and developing countries in Africa showing less improvement than the Americas, Asia and Europe ( [[#Feldmeyer--2017|Feldmeyer et al., 2017]] ). An analysis of 48 major deltas finds that vulnerability therefore is a less dominant source of future increase in risk than exposure ( [[#Haasnoot--2012|Haasnoot et al., 2012]] ). However, case study research from individual deltas suggests that delta populations, particularly those with agriculture-based livelihoods, have seen more limited vulnerability reduction due in particular to the impacts of environmental hazards, stress and disasters ( ''high confidence'' ). In the Mekong Delta, for instance, the strong economic growth since the beginning of Vietnam’s reform process has not led to a reduction of vulnerability across the board for all socioeconomic groups ( [[#Garschagen--2015|Garschagen, 2015]] ). Rather, issues such as widespread landlessness or continued poverty have maintained and, in some respect, increased social vulnerability.

<div id="16.5.4" class="h2-container"></div>

<span id="rkr-interactions"></span>
=== 16.5.4 RKR Interactions ===

<div id="h2-17-siblings" class="h2-siblings"></div>

Multiple feedbacks between individual risks exist that have the potential to create cascades ( [[#WEF--2018|WEF, 2018]] ; [[#IPCC--2019c|IPCC, 2019c]] p. 680; [[#Simpson--2021|Simpson et al., 2021]] ) and then to amplify systemic risks and impacts far beyond the level of individual RKRs ( ''medium confidence'' ). Scientific research, however, remains limited on whether such interactions would result in increasing or decreasing the initial impact(s), and hence risk severity across systems. Given the scope of this chapter on increasing risk severity, here we focus on assessing RKR interactions that lead to increasing risk. Drawing directly on RKR assessments (Sections 16.5.2.3.2–16.5.2.3.8), this section cites those assessments rather than primary literature. The arrows in Figure 16.11 are derived from a qualitative analysis by three authors of [https://www.ipcc.ch/report/ar6/wg2/chapter/chapter-16 Chapter 16] of the material provided by chapters on KRs and RKR assessments ( [[#16.5.2.3|Section 16.5.2.3]] ), and do not result from any systematic and quantitative approach as done in some recent studies (e.g., [[#WEF--2018|WEF, 2018]] ; [[#Yokohata--2019|Yokohata et al., 2019]] ).

<div id="_idContainer038" class="Figure"></div>

[[File:e62e9bd366add2d9fd79c9e2fe619b19 IPCC_AR6_WGII_Figure_16_011.png]]

'''Figure 16.11 |''' '''Illustration of some connections across key risks.''' Panel A describes all the cross-RKR risk cascades that are described in RKR assessments (Sections 16.5.2.3.2–16.5.2.3.8). Panel B builds on [[#16.5.2|Section 16.5.2]] and Table SM16.24 to provide an illustration of such interactions at the key risk level, for example from ecological risk to key dimensions for human societies. The arrows are representative of interactions as qualitatively identified in this chapter; they do not result from any quantitative modelling exercise.

''Interactions at the RKR level'' (Figure 16.11, panel A)—climate change will combine with pre-existing socioeconomic and ecological conditions (grey blocks on the left-hand-side of panel A in Figure 16.10) to generate direct and second-order effects (black plain arrows) both on the structure and/or functioning of ecosystems (RKR-B) and on some natural processes such as the hydrologic cycle (RKR-G), for example. This then translates into implications not only for biodiversity but also for natural resources that support livelihoods, which will in turn affect food security (especially food availability; RKR-F), water security (especially access to adequate quantities of acceptable quality water; RKR-G) and the living standards of already vulnerable groups and aggregate economic outputs at the global level (RKR-D). CIDs ( [[#IPCC--2021|IPCC, 2021]] ) will also directly affect infrastructure that are critical to ensure some basic conditions for economies to function (RKR-C), for example through transportation within and outside the country, energy production and international trade. Such disturbances to socioecological systems and economies pose climate-related risks to human health (RKR-E) as well as to peace and human mobility (RKR-H). Indeed, while health is concerned with direct influence of climate change, for example through hotter air temperatures impacting morbidity and mortality or the spatial distribution of disease vectors such as mosquitos, it is also at risk of being stressed by direct and secondary climate impacts on living standards, food security and water security (RKR-D, RKR-F, RKR-G, respectively). Increased poverty, increased hunger and limited access to drinkable water are well-known drivers of poor health conditions. The role of impact cascades is even more prominent in the case of peace and human mobility (RKR-H), even though the scientific literature does not conclude on any clear and direct climate influence on armed conflict and human migration. Rather, climate-induced degradation of natural resources that are vital for subsistence agriculture and fisheries, transformational and long-term consequences on livelihoods (e.g., new risks, increasing precarious living conditions, gendered inequity, etc.), and erosion of social capital due to exacerbated tension within and between communities are considered among the main drivers of armed conflicts and forced displacement, therefore highlighting links with water security (RKR-G) and living standards (RKR-D), for example.

RKR assessments also suggest that some feedback effects are at work (arrows moving from the right to the left in panel A) that contribute to the potentially long-lasting effects of climate risks. RKR-H assessment, for example, states that there is ''robust evidence'' that major armed conflicts routinely trigger mass displacement, threaten health and food security, and undermine economic activity and livelihoods, often with lasting negative consequences for living standards and socioeconomic development, therefore linking back to risks to living standards (RKR-D), human health (RKR-E) and food security (RKR-F).

''Interactions at the KR level'' (Figure 16.11, panel B)—panel B illustrates risk connections at the Key Risk level ( [[#16.5.2.1|Section 16.5.2.1]] ) and as described in RKR assessments ( [[#16.5.2.3|Section 16.5.2.3]] ). To only take one example here, risk to livelihoods and economies is influenced by the loss of ecosystem services (RKR-B) and the loss or breakdown of critical infrastructures (RKR-C), and it influences risks to human lives and health (RKR-E), food and water security (RKR-F, RKR-G), poverty (RKR-D) and peace and human mobility (RKR-H). As a third-order sequence, RKR assessments show that increased risk to peace and human mobility affects lives and health as well as food security, which in turn threaten livelihoods and economies.

The above suggests that some vicious cycle effects play a central role in explaining impact processes. Cascading effects can indeed lead to cumulative risks that partly feed various drivers of the emergence of severe risks ( [[#16.5.1|Section 16.5.1]] ), such as the acceleration of ecosystem degradation, or the reaching of thresholds and irreversible states in human systems at a decade-to-century time horizon (e.g., when permanent inundation questions the habitability of some low-lying coasts; RKR-A). The extent and duration of risk cascades are, however, expected to substantially vary depending on warming levels and development pathways, both separately ( [[#16.5.3|Section 16.5.3]] ) and when combined (Sections 16.6.1, 16.6.2) (Figure 16.10).

In addition, RKR assessments converge to suggest that regions that are already experiencing climate change impacts will experience severe impact cascades first (e.g., RKR-F), because they are in areas (i) that face development constraints and associated challenges such as poverty, inequity and social discrimination for example, and (ii) where climate change projections are the most intense for the next decades. That is especially a concern for Africa (RKR-F, RKR-G), Asia and Latin America (Chapters 9, 10, 12). RKR-E, for example, concludes that the likelihood of severe risks to human health is especially high for highly susceptible populations, particularly the poor and otherwise marginalised. RKR assessments, however, emphasise that middle- and high-income regions are also to be considered at serious risk because climate change is accelerating at the global level ( [[#IPCC--2021|IPCC, 2021]] ), and because critical dimensions are exposed to severe risks such as major transportation (e.g., international airports) and energy (e.g., nuclear power plants) infrastructure for instance (RKR-C), and because of the interconnectedness of economies.

Finally, all RKR assessments suggest that enhanced adaptation has the potential to contain such feedback effects and cascading processes more broadly, and reduce the duration of the impacts on the system as a whole. There are, however, knowledge gaps on such a potential, as well as on the nature of impact cascades (positive, negative, neutral, mixed).

<div id="cross-working-group-box-srm" class="h2-container box-container"></div>

'''Cross-Working Group Box SRM | Solar Radiation Modification'''
<div id="h2-25-siblings" class="h2-siblings"></div>

Authors: Christopher H. Trisos (South Africa), Oliver Geden (Germany), Sonia I. Seneviratne (Switzerland), Masahiro Sugiyama (Japan), Maarten van Aalst (the Netherlands), Govindasamy Bala (India), Katharine J. Mach (USA), Veronika Ginzburg (Russia), Heleen de Coninck (the Netherlands), Anthony Patt (Switzerland)

'''Proposed solar radiation modification schemes'''
This cross-working group box assesses solar radiation modification (SRM) proposals, their potential contribution to reducing or increasing climate risk, as well as other risks they may pose (categorised as risks from responses to climate change in the IPCC AR6 risk definition in 1.2.1.1), and related perception, ethics and governance questions.

SRM refers to proposals to increase the reflection of shortwave radiation (sunlight) back to space to counteract anthropogenic warming and some of its harmful impacts ( [[#de%20Coninck--2018|de Coninck et al., 2018]] ) (Cross-Chapter Box 10; WGI Chapters 4, 5). A number of SRM options have been proposed, including: stratospheric aerosol interventions (SAI), marine cloud brightening (MCB), ground-based albedo modifications (GBAM) and ocean albedo change (OAC). Although not strictly a form of SRM, cirrus cloud thinning (CCT) has been proposed to cool the planet by increasing the escape of longwave thermal radiation to space and is included here for consistency with previous assessments ( [[#de%20Coninck--2018|de Coninck et al., 2018]] ). SAI is the most-researched proposal. Modelling studies show SRM could reduce surface temperatures and potentially ameliorate some climate change risks (with more confidence for SAI than other options), but SRM could also introduce a range of new risks.

There is ''high agreement'' in the literature that for addressing climate change risks SRM cannot be the main policy response to climate change and is, at best, a supplement to achieving sustained net zero or net negative CO 2 emission levels globally ( [[#de%20Coninck--2018|de Coninck et al., 2018]] ; [[#MacMartin--2018|MacMartin et al., 2018]] ; [[#Buck--2020|Buck et al., 2020]] ; National Academies of Sciences and Medicine, 2021b). SRM contrasts with climate change mitigation activities, such as emission reductions and carbon dioxide removal (CDR), as it introduces a ‘mask’ to the climate change problem by altering the Earth’s radiation budget, rather than attempting to address the root cause of the problem, which is the increase in greenhouse gases (GHGs) in the atmosphere. In addition, the effects of proposed SRM options would only last as long as a deployment is maintained—for example, requiring ca. yearly injection of aerosols in the case of SAI as the lifetime of aerosols in the stratosphere is 1–3 years ( [[#Niemeier--2011|Niemeier et al., 2011]] ) or continuous spraying of sea salt in the case of MCB as the lifetime of sea salt aerosols in the atmosphere is only about 10 d—which contrasts with the long lifetime of CO 2 and its climate effects, with global warming resulting from CO 2 emissions ''likely'' remaining at a similar level for a hundred years or more ( [[#MacDougall--2020|MacDougall et al., 2020]] ) and long-term climate effects of emitted CO 2 remaining for several hundreds to thousands of years ( [[#Solomon--2009|Solomon et al., 2009]] ).

'''Which scenarios?'''
The choice of SRM deployment scenarios and reference scenarios is crucial in assessment of SRM risks and its effectiveness in attenuating climate change risks ( [[#Keith--2015|Keith and MacMartin, 2015]] ; [[#Honegger--2021|Honegger et al., 2021]] ). Most climate model simulations have used scenarios with highly stylised large SRM forcing to fully counteract large amounts of warming in order to enhance the signal-to-noise ratio of climate responses to SRM ( [[#Kravitz--2015|Kravitz et al., 2015]] ; [[#Sugiyama--2018a|Sugiyama et al., 2018a]] ; [[#Tilmes--2018|Tilmes et al., 2018]] ; [[#Krishna-Pillai--2019|Krishna-Pillai et al., 2019]] ).

The effects of SRM fundamentally depend on a variety of choices about deployment ( [[#Sugiyama--2018b|Sugiyama et al., 2018b]] ), including: its position in the portfolio of human responses to climate change (e.g., the magnitude of SRM used against the background radiative forcing), governance of research and potential deployment strategies, and technical details (latitude, materials, and season, among others, see WGI [[IPCC:Wg2:Chapter:Chapter-4#4.6.3|Section 4.6.3.3]] ). The plausibility of many SRM scenarios is highly contested, and not all scenarios are equally plausible because of socio-political considerations ( [[#Talberg--2018b|Talberg et al., 2018b]] ), as with, for example, CDR ( [[#Fuss--2014|Fuss et al., 2014]] ; [[#Fuss--2018|Fuss et al., 2018]] ). Development of scenarios and their selection in assessments should reflect a diverse set of societal values with public and stakeholder inputs ( [[#Sugiyama--2018a|Sugiyama et al., 2018a]] ; [[#Low--2020|Low and Honegger, 2020]] ), as depending on the focus of a limited climate model simulation, SRM could look grossly risky or highly beneficial (Pereira and al., 2021).

In the context of reaching the long-term global temperature goal of the Paris Agreement, there are different hypothetical scenarios of SRM deployment: early, substantial mitigation with no SRM, more limited or delayed mitigation with moderate SRM, unchecked emissions with total reliance on SRM, and regionally heterogeneous SRM. Each scenario presents different levels and distributions of SRM benefits, side effects and risks. The more intense the SRM deployment, the larger is the likelihood for the risks of side effects and environmental risks (e.g., [[#Heutel--2018|Heutel et al., 2018]] ). Regional disparities in climate hazards may result from both regionally deployed SRM options such as GBAM, and more globally uniform SRM such as SAI ( [[#Jones--2018a|Jones et al., 2018a]] ; [[#Seneviratne--2018b|Seneviratne et al., 2018b]] ). There is an emerging literature on smaller forcings of SAI to reduce global average warming, for instance, to hold global warming to 1.5°C or 2°C alongside ambitious conventional mitigation ( [[#Jones--2018a|Jones et al., 2018a]] ; [[#MacMartin--2018|MacMartin et al., 2018]] ), or bring down temperature after an overshoot ( [[#Tilmes--2020|Tilmes et al., 2020]] ). If emissions reductions and CDR are deemed insufficient, SRM may be seen by some as the only option left to ensure the achievement of the Paris Agreement’s temperature goal by 2100.

'''Table Cross-Working Group Box SRM.1''' '''|''' SRM options and their potential climate and non-climate impacts. Description, potential climate impacts, potential impacts on human and natural systems, and termination effects of a number of SRM options: stratospheric aerosol interventions (SAI), marine cloud brightening (MCB), ocean albedo change (OAC), ground-based albedo modifications (GBAM) and cirrus cloud thinning (CCT).

{| class="wikitable"
|-
! ''SRM option''

! ''SAI''

! ''MCB''

! ''OAC''

! ''GBAM''

! ''CCT''

|-
| Description

| Injection of reflective aerosol particles directly into the stratosphere or a gas which then

converts to aerosols that reflect sunlight

| Spraying sea salt or other particles in marine clouds, making them more reflective

| Increase surface albedo of the ocean (e.g., by creating microbubbles or placing reflective foam on the surface)

| Whitening roofs, changes in land use management (e.g., no-till farming, bioengineering to make crop leaves more reflective), desert albedo enhancement, covering glaciers with reflective sheeting

| Seeding to promote nucleation of cirrus clouds, reducing optical thickness and cloud lifetime to allow more outgoing longwave radiation to escape to space

|-
| Potential climate impacts ''other than reduced warming''

| Change precipitation and runoff

pattern; reduced temperature and precipitation extremes; precipitation reduction in some monsoon regions; decrease in direct and increase in diffuse sunlight at surface; changes to stratospheric dynamics and chemistry; potential

delay in ozone hole recovery; changes in surface ozone and UV radiation

| Change in land-sea contrast in temperature and precipitation, regional precipitation and runoff changes

| Change in land–sea contrast in temperature and precipitation, regional, precipitation and runoff changes

| Changes in regional precipitation pattern, regional extremes and regional circulation

| Changes in temperature and precipitation pattern, altered regional water cycle, increase in sunlight reaching the surface

|-
| Potential impacts on human and natural systems

| Changes in crop yields, changes in land and ocean ecosystem productivity, acid rain (if using sulphate), reduced risk of heat stress to corals

| Changes in regional ocean productivity, changes in crop yields, reduced heat stress for corals, changes in ecosystem productivity on land, sea salt deposition over land

| Unresearched

| Altered photosynthesis, carbon uptake and side effects on biodiversity

| Altered photosynthesis and carbon uptake

|-
| Termination effects

| Sudden and sustained termination would result in rapid warming, and abrupt changes to water cycle. Magnitude of termination depends on the degree of warming offset

| Sudden and sustained

termination would result in rapid warming, and abrupt changes to water cycle. Magnitude of termination depends on the degree of warming offset

| Sudden and sustained

termination would result in rapid warming. Magnitude of termination depends on the degree of warming offset

| GBAM can be maintained over several years without major termination effects because of its regional scale of application.

Magnitude of termination depends on the degree of warming offset

| Sudden and sustained

termination would result in rapid warming. Magnitude of termination depends on the degree of warming offset

|-
| References (also see main text of this box)

| [[#Tilmes--2018|Tilmes et al. (2018)]] ; [[#Simpson--2019|Simpson et al. (2019)]] ; Visioni et al. (2017)

| [[#Latham--2012|Latham et al. (2012)]] ; [[#Ahlm--2017|Ahlm et al. (2017)]] ; [[#Stjern--2018|Stjern et al. (2018)]]

| [[#Evans--2010|Evans et al. (2010)]] ; [[#Crook--2015a|Crook et al. (2015a)]]

| [[#Zhang--2016|Zhang et al. (2016)]] ; [[#Field--2018|Field et al. (2018)]] ; [[#Seneviratne--2018a|Seneviratne et al. (2018a)]] ; [[#Davin--2014|Davin et al. (2014)]] ; [[#Crook--2015a|Crook et al. (2015a)]]

| [[#Storelvmo--2014|Storelvmo and Herger (2014)]] ; [[#Crook--2015a|Crook et al. (2015a)]] ; Jackson et al. (2016); [[#Gasparini--2020|Gasparini et al. (2020)]] ; [[#Duan--2020|Duan et al. (2020)]]

|}

'''SRM risks to human and natural systems and potential for risk reduction'''
Since AR5, hundreds of climate modelling studies have simulated effects of SRM on climate hazards ( [[#Kravitz--2015|Kravitz et al., 2015]] ; [[#Tilmes--2018|Tilmes et al., 2018]] ). Modelling studies have shown SRM has the potential to offset some effects of increasing GHGs on global and regional climate, including the increase in frequency and intensity of extremes of temperature and precipitation, melting of Arctic sea ice and mountain glaciers, weakening of Atlantic meridional overturning circulation, changes in frequency and intensity of tropical cyclones, and decrease in soil moisture (WGI, Chapter 4). However, while SRM may be effective in alleviating anthropogenic climate warming either locally or globally, it would neither maintain the climate in its present-day state nor return the climate to a pre-industrial state (climate averaged over 1850–1900, see WGI Chapter 1, Box 1.2) in all regions and in all seasons even when used to fully offset the global mean warming ( ''high confidence'' ) (WGI Chapter 4). This is because the climate forcing and response to SRM options are different from the forcing and response to GHG increase. Because of these differences in climate forcing and response patterns, the regional and seasonal climates of a world with a global mean warming of 1.5°C or 2°C achieved via SRM would be different from a world with similar global mean warming but achieved through mitigation (MacMartin et al.., 2019). At the regional scale and seasonal time scale, there could be considerable residual climate change and/or overcompensating change (e.g., more cooling, wetting or drying than just what is needed to offset warming, drying or wetting due to anthropogenic greenhouse gas emissions), and there is ''low confidence'' in understanding of the climate response to SRM at the regional scale (WGI, Chapter 4).

SAI implemented to partially offset warming (e.g., offsetting half of global warming) may have potential to ameliorate hazards in multiple regions and reduce negative residual change, such as drying compared with present-day climate, that is associated with fully offsetting global mean warming ( [[#Irvine--2020|Irvine and Keith, 2020]] ), but may also increase flood and drought risk in Europe compared with unmitigated warming ( [[#Jones--2021|Jones et al., 2021]] ). Recent modelling studies suggest it is conceptually possible to meet multiple climate objectives through optimally designed SRM strategies (WGI, Chapter 4). Nevertheless, large uncertainties still exist for climate processes associated with SRM options (e.g., aerosol–cloud–radiation interaction) (WGI, Chapter 4) ( [[#Kravitz--2020|Kravitz and MacMartin, 2020]] ).

Compared with climate hazards, many fewer studies have examined SRM risks—the potential adverse consequences to people and ecosystems from the combination of climate hazards, exposure and vulnerability—or the potential for SRM to reduce risk ( [[#Curry--2014|Curry et al., 2014]] ; [[#Irvine--2017|Irvine et al., 2017]] ). Risk analyses have often used inputs from climate models forced with stylised representations of SRM, such as dimming the sun. Fewer have used inputs from climate models that explicitly simulated injection of gases or aerosols into the atmosphere, which include more complex cloud radiative feedbacks. Most studies have used scenarios where SAI is deployed to hold average global temperature constant despite high emissions.

There is ''low confidence'' and large uncertainty in projected impacts of SRM on crop yields due in part to a limited number of studies. Because SRM would result in only a slight reduction in CO 2 concentrations relative to the emission scenario without SRM (Chapter 5, WGI), the CO 2 fertilisation effect on plant productivity is nearly the same in emissions scenarios with and without SRM. Nevertheless, changes in climate due to SRM are likely to have some impacts on crop yields. A single study indicates MCB may reduce crop failure rates compared with climate change from a doubling of CO 2 pre-industrial concentrations ( [[#Parkes--2015|Parkes et al., 2015]] ). Models suggest SAI cooling would reduce crop productivity at higher latitudes compared with a scenario without SRM by reducing the growing season length, but benefit crop productivity in lower latitudes by reducing heat stress ( [[#Pongratz--2012|Pongratz et al., 2012]] ; [[#Xia--2014|Xia et al., 2014]] ; [[#Zhan--2019|Zhan et al., 2019]] ). Crop productivity is also projected to be reduced where SAI reduces rainfall relative to the scenario without SRM, including a case where reduced Asian summer monsoon rainfall causes a reduction in groundnut yields ( [[#Xia--2014|Xia et al., 2014]] ; [[#Yang--2016|Yang et al., 2016]] ). SAI will increase the fraction of diffuse sunlight, which is projected to increase photosynthesis in forested canopy, but will reduce the direct and total available sunlight, which tends to reduce photosynthesis. As total sunlight is reduced, there is a net reduction in crop photosynthesis with the result that any benefits to crops from avoided heat stress may be offset by reduced photosynthesis, as indicated by a single statistical modelling study ( [[#Proctor--2018|Proctor et al., 2018]] ). SAI would reduce average surface ozone concentration ( [[#Xia--2017|Xia et al., 2017]] ) mainly as a result of aerosol-induced reduction in stratospheric ozone in polar regions, resulting in reduced downward transport of ozone to the troposphere ( [[#Pitari--2014|Pitari et al., 2014]] ; [[#Tilmes--2018|Tilmes et al., 2018]] ). The reduction in stratospheric ozone also allows more UV radiation to reach the surface. The reduction in surface ozone, together with an increase in surface UV radiation, would have important implications for crop yields but there is ''low confidence'' in our understanding of the net impact.

Few studies have assessed potential SRM impacts on human health and well-being. SAI using sulphate aerosols is projected to deplete the ozone layer, increasing mortality from skin cancer, and SAI could increase particulate matter due to offsetting warming, reduced precipitation and deposition of SAI aerosols, which would increase mortality, but SAI also reduces surface-level ozone exposure, which would reduce mortality from air pollution, with net changes in mortality uncertain and depending on aerosol type and deployment scenario ( [[#Effiong--2016|Effiong and Neitzel, 2016]] ; [[#Eastham--2018|Eastham et al., 2018]] ; [[#Dai--2020|Dai et al., 2020]] ). However, these effects may be small compared with changes in risk from infectious disease (e.g., mosquito-borne illnesses) or food security due to SRM influences on climate ( [[#Carlson--2020|Carlson et al., 2020]] ). Using volcanic eruptions as a natural analogue, a sudden implementation of SAI that forced the El Niño-Southern Oscillation (ENSO) system may increase risk of severe cholera outbreaks in Bengal ( [[#Trisos--2018|Trisos et al., 2018]] ; [[#Pinke--2019|Pinke et al., 2019]] ). Considering only mean annual temperature and precipitation, SAI that stabilises global temperature at its present-day level is projected to reduce income inequality between countries compared with the highest warming pathway (RCP8.5) ( [[#Harding--2020|Harding et al., 2020]] ). Some integrated assessment model scenarios have included SAI ( [[#Arino--2016|Arino et al., 2016]] ; [[#Emmerling--2018|Emmerling and Tavoni, 2018]] ; [[#Heutel--2018|Heutel et al., 2018]] ; [[#Helwegen--2019|Helwegen et al., 2019]] ; [[#Rickels--2020|Rickels et al., 2020]] ) showing the indirect costs and benefits to welfare dominate, since the direct economic cost of SAI itself is expected to be relatively low ( [[#Moriyama--2017|Moriyama et al., 2017]] ; [[#Smith--2018|Smith and Wagner, 2018]] ). There is a general lack of research on the wide scope of potential risk or risk reduction to human health, well-being and sustainable development from SRM and on their distribution across countries and vulnerable groups ( [[#Carlson--2020|Carlson et al., 2020]] ; [[#Honegger--2021|Honegger et al., 2021]] ).

SRM may also introduce novel risks for international collaboration and peace. Conflicting temperature preferences between countries may lead to counter-geoengineering measures such as deliberate release of warming agents or destruction of deployment equipment ( [[#Parker--2018|Parker et al., 2018]] ). Game-theoretic models and laboratory experiments indicate that a powerful actor or group with a higher preference for SRM may use SAI to cool the planet beyond what is socially optimal, imposing welfare losses on others, although this cooling does not necessarily imply that excluded countries would be worse off relative to a world of unmitigated warming ( [[#Ricke--2013|Ricke et al., 2013]] ; [[#Weitzman--2015|Weitzman, 2015]] ; [[#Abatayo--2020|Abatayo et al., 2020]] ). In this context, counter-geoengineering may promote international cooperation or lead to large welfare losses ( [[#Heyen--2019|Heyen et al., 2019]] ; [[#Abatayo--2020|Abatayo et al., 2020]] ).

Cooling caused by SRM would increase the global land and ocean CO 2 sinks ( ''medium confidence'' ), but this would not stop CO 2 from increasing in the atmosphere or affect the resulting ocean acidification under continued anthropogenic emissions ( ''high confidence'' ) (WGI Chapter 5).

Few studies have assessed potential SRM impacts on ecosystems. SAI and MCB may reduce risk of coral reef bleaching compared with global warming with no SAI ( [[#Latham--2013|Latham et al., 2013]] ; [[#Kwiatkowski--2015|Kwiatkowski et al., 2015]] ), but risks to marine life from ocean acidification would remain, because SRM proposals do not reduce elevated levels of anthropogenic atmospheric CO 2 concentrations. MCB could cause changes in marine net primary productivity by reducing light availability in deployment regions, with important fishing regions off the west coast of South America showing both large increases and decreases in productivity ( [[#Partanen--2016|Partanen et al., 2016]] ; [[#Keller--2018|Keller, 2018]] ).

There is large uncertainty in terrestrial ecosystem responses to SRM. By decoupling increases in atmospheric greenhouse gas concentrations and temperature, SAI could generate substantial impacts on large-scale biogeochemical cycles, with feedbacks to regional and global climate variability and change ( [[#Zarnetske--2021|Zarnetske et al., 2021]] ). Compared with a high-CO 2 world without SRM, global-scale SRM simulations indicate reducing heat stress in low latitudes would increase plant productivity, but cooling would also slow down the process of nitrogen mineralisation, which could decrease plant productivity ( [[#Glienke--2015|Glienke et al., 2015]] ; [[#Duan--2020|Duan et al., 2020]] ). In high-latitude and polar regions, SRM may limit vegetation growth compared with a high-CO 2 world without SRM, but net primary productivity may still be higher than pre-industrial climate ( [[#Glienke--2015|Glienke et al., 2015]] ). Tropical forests cycle more carbon and water than other terrestrial biomes, but large areas of the tropics may tip between savanna and tropical forest depending on rainfall and fire ( [[#Beer--2010|Beer et al., 2010]] ; [[#Staver--2011|Staver et al., 2011]] ). Thus, SAI-induced reductions in precipitation in Amazonia and central Africa are expected to change the biogeography of tropical ecosystems in ways different from both present-day climate and global warming without SAI ( [[#Simpson--2019|Simpson et al., 2019]] ; [[#Zarnetske--2021|Zarnetske et al., 2021]] ). This would have potentially large consequences for ecosystem services ( [[IPCC:Wg2:Chapter:Chapter-2|Chapter 2]] and Chapter 9). When designing and evaluating SAI scenarios, biome-specific responses need to be considered if SAI approaches are to benefit rather than harm ecosystems. Regional precipitation change and sea salt deposition over land from MCB may increase or decrease primary productivity in tropical rainforests ( [[#Muri--2015|Muri et al., 2015]] ). SRM that fully offsets warming could reduce the dispersal velocity required for species to track shifting temperature niches, whereas partially offsetting warming with SAI would not reduce this risk unless rates of warming were also reduced ( [[#Trisos--2018|Trisos et al., 2018]] ; [[#Dagon--2019|Dagon and Schrag, 2019]] ). SAI may reduce high-fire-risk weather in Australia, Europe and parts of the Americas, compared with global warming without SAI ( [[#Burton--2018|Burton et al., 2018]] ). Yet SAI using sulphur injection could shift the spatial distribution of acid-induced aluminium soil toxicity into relatively undisturbed ecosystems in Europe and North America ( [[#Visioni--2020|Visioni et al., 2020]] ). For the same amount of global mean cooling, SAI, MCB and CCT would have different effects on gross and net primary productivity because of different spatial patterns of temperature, available sunlight, and hydrological cycle changes ( [[#Duan--2020|Duan et al., 2020]] ). Large-scale modification of land surfaces for GBAM may have strong trade-offs with biodiversity and other ecosystem services, including food security ( [[#Seneviratne--2018a|Seneviratne et al., 2018a]] ). Although existing studies indicate SRM will have widespread impacts on ecosystems, risks and potential for risk reduction for marine and terrestrial ecosystems and biodiversity remain largely unknown.

A sudden and sustained termination of SRM in a high CO 2 emissions scenario would cause rapid climate change ( ''high confidence'' ; WGI Chapter 4). More scenario analysis is needed on the potential likelihood of sudden termination ( [[#Kosugi--2013|Kosugi, 2013]] ; [[#Irvine--2020|Irvine and Keith, 2020]] ). A gradual phase-out of SRM combined with emission reduction and CDR could avoid these termination effects ( ''medium confidence'' ) ( [[#MacMartin--2014|MacMartin et al., 2014]] ; [[#Keith--2015|Keith and MacMartin, 2015]] ; [[#Tilmes--2016|Tilmes et al., 2016]] ). Several studies find that large and extremely rapid warming and abrupt changes to the water cycle would occur within a decade if a sudden termination of SAI occurred ( [[#McCusker--2014|McCusker et al., 2014]] ; [[#Crook--2015b|Crook et al., 2015b]] ). The size of this ‘termination shock’ is proportional to the amount of radiative forcing being masked by SAI. A sudden termination of SAI could place many thousands of species at risk of extinction, because the resulting rapid warming would be too fast for species to track the changing climate ( [[#Trisos--2018|Trisos et al., 2018]] ).

'''Public perceptions of SRM'''
Studies on the public perception of SRM have used multiple methods: questionnaire surveys, workshops, and focus group interviews ( [[#Burns--2016|Burns et al., 2016]] ; [[#Cummings--2017|Cummings et al., 2017]] ). Most studies have been limited to Western societies, with some exceptions. Studies have repeatedly found that respondents are largely unaware of SRM ( [[#Merk--2015|Merk et al., 2015]] ). In the context of this general lack of familiarity, the public prefers CDR to SRM ( [[#Pidgeon--2012|Pidgeon et al., 2012]] ), is very cautious about SRM deployment because of potential environmental side effects and governance concerns, and mostly rejects deployment for the foreseeable future. Studies also suggest conditional and reluctant support for research, including proposed field experiments, with conditions of proper governance ( [[#Sugiyama--2020|Sugiyama et al., 2020]] ). Recent studies show that the perception varies with the intensity of deliberation ( [[#Merk--2019|Merk et al., 2019]] ), and that the public distinguishes different funding sources ( [[#Nelson--2021|Nelson et al., 2021]] ). Limited studies for developing countries show a tendency for respondents to be more open to SRM ( [[#Visschers--2017|Visschers et al., 2017]] ; [[#Sugiyama--2020|Sugiyama et al., 2020]] ), perhaps because they experience climate change more directly ( [[#Carr--2018|Carr and Yung, 2018]] ). In some Anglophone countries, a small portion of the public believes in chemtrail conspiracy theories, which are easily found in social media ( [[#Tingley--2017|Tingley and Wagner, 2017]] ; [[#Allgaier--2019|Allgaier, 2019]] ). Since researchers rarely distinguish different SRM options in engagement studies, there remains uncertainty in public perception.

'''Ethics'''
There is broad literature on ethical considerations around SRM, mainly stemming from philosophy or political theory, and mainly focused on SAI ( [[#Flegal--2019|Flegal et al., 2019]] ). There is concern that publicly debating, researching and potentially deploying SAI could involve a ‘moral hazard’, with potential to obstruct ongoing and future mitigation efforts ( [[#Morrow--2014|Morrow, 2014]] ; [[#Baatz--2016|Baatz, 2016]] ; [[#McLaren--2016|McLaren, 2016]] ), while empirical evidence is limited and mostly at the individual, not societal, level ( [[#Burns--2016|Burns et al., 2016]] ; [[#Merk--2016|Merk et al., 2016]] ; [[#Merk--2019|Merk et al., 2019]] ). There is ''low agreement'' whether research and outdoors experimentation will create a ‘slippery slope’ towards eventual deployment, leading to a lock-in to long-term SRM, or can be effectively regulated at a later stage to avoid undesirable outcomes ( [[#Hulme--2014|Hulme, 2014]] ; [[#Parker--2014|Parker, 2014]] ; [[#Callies--2019|Callies, 2019]] ; [[#McKinnon--2019|McKinnon, 2019]] ). Regarding potential deployment of SRM, procedural, distributive and recognitional conceptions of justice are being explored ( [[#Svoboda--2014|Svoboda and Irvine, 2014]] ; [[#Svoboda--2017|Svoboda, 2017]] ; [[#Preston--2018|Preston and Carr, 2018]] ; [[#Hourdequin--2019|Hourdequin, 2019]] ). With the SRM research community’s increasing focus on distributional impacts of SAI, researchers have started more explicitly considering inequality in participation and inclusion of vulnerable countries and marginalised social groups ( [[#Flegal--2018|Flegal and Gupta, 2018]] ; [[#Whyte--2018|Whyte, 2018]] ; [[#Táíwò--2021|Táíwò and Talati, 2021]] ), including considering stopping research ( [[#Stephens--2020|Stephens and Surprise, 2020]] ; National Academies of Sciences and Medicine, 2021a). There is recognition that SRM research has been conducted predominantly by a relatively small number of experts in the Global North, and that more can be done to enable participation from diverse peoples and geographies in setting research agendas and research governance priorities, and undertaking research, with initial efforts to this effect (e.g., [[#Rahman--2018|Rahman et al., 2018]] ), noting unequal power relations in participation could influence SRM research governance and potential implications for policy ( [[#Whyte--2018|Whyte, 2018]] ; [[#Táíwò--2021|Táíwò and Talati, 2021]] ; [[#Winickoff--2015|Winickoff et al., 2015]] ; [[#Frumhoff--2018|Frumhoff and Stephens, 2018]] ; [[#Biermann--2019|Biermann and Möller, 2019]] ; [[#McLaren--2021|McLaren and Corry, 2021]] ; National Academies of Sciences and Medicine, 2021b)

'''Governance of research and of deployment'''
Currently, there is no dedicated, formal international SRM governance for research, development, demonstration or deployment (see WGIII Chapter 14). Some multilateral agreements—such as the UN Convention on Biological Diversity or the Vienna Convention on the Protection of the Ozone Layer—indirectly and partially cover SRM, but none is comprehensive, and the lack of robust and formal SRM governance poses risks ( [[#Ricke--2013|Ricke et al., 2013]] ; [[#Talberg--2018a|Talberg et al., 2018a]] ; [[#Reynolds--2019a|Reynolds, 2019a]] ). While governance objectives range broadly, from prohibition to enabling research and potentially deployment ( [[#Sugiyama--2018b|Sugiyama et al., 2018b]] ; [[#Gupta--2020|Gupta et al., 2020]] ), there is agreement that SRM governance should cover all interacting stages of research through to any potential, eventual deployment with rules, institutions and norms ( [[#Reynolds--2019b|Reynolds, 2019b]] ). Accordingly, governance arrangements are co-evolving with respective SRM technologies across the interacting stages of research, development, demonstration and—potentially—deployment ( [[#Rayner--2013|Rayner et al., 2013]] ; [[#Parker--2014|Parker, 2014]] ; [[#Parson--2014|Parson, 2014]] ). Stakeholders are developing governance already in outdoors research, for example for MCB and OAC experiments on the Great Barrier Reef ( [[#McDonald--2019|McDonald et al., 2019]] ). Co-evolution of governance and SRM research provides a chance for responsibly developing SRM technologies with broader public participation and political legitimacy, guarding against potential risks and harms relevant across a full range of scenarios, and ensuring that SRM is considered only as a part of a broader portfolio of responses to climate change ( [[#Stilgoe--2015|Stilgoe, 2015]] ; [[#Nicholson--2018|Nicholson et al., 2018]] ). For SAI, large-scale outdoor experiments even with low radiative forcing could be transboundary, and those with deployment-scale radiative forcing may not be distinguished from deployment, such that [[#MacMartin--2019|MacMartin and Kravitz (2019)]] argue for continued reliance on modelling until a decision on whether and how to deploy is made, with modelling helping governance development. For further discussion of SRM governance, see Chapter 14, WGIII.

<div id="16.6" class="h1-container"></div>

<span id="reasons-for-concern-across-scales-1"></span>