FINAL DRAFT                       Annex I  IPCC WGII Sixth Assessment Report

 1 Captions

 2

 3 Figure AI.01: Risk in IPCC assessments.
 4 (a) An explicit risk framing emerged in the IPCC SREX and WGII AR5. (b) In the current AR6 assessment,
 5 the role of responses in modulating the determinants of risk is a new emphasis (the "wings" of the hazard,
 6 vulnerability, and exposure "propellers" represents the ways in which responses modulate each of these risk
 7 determinants {Figure 1.5}

 8

 9 Figure AI.02: Physical drivers of climate change: Temperature.
10 {AR6 WGI Interactive Atlas}

11

12 Figure AI.03: Physical drivers of climate change: Precipitation.
13 {AR6 WGI Interactive Atlas}

14

15 Figure AI.04: Physical drivers of climate change: Dissolved Oxygen in the Ocean.
16 {Assis et al., 2017}

17

18 Figure AI.05: Evidence of climate change impacts in many regions of the world.
19 Global density map shows climate impact evidence, derived by machine-learning from 77,785 studies. Bar
20 charts show the number of studies per continent and impact category. Bars are coloured by the climate
21 variable predicted to drive impacts. Colour intensity indicates the percentage of cells a study refers to where
22 a trend in the climate variable can be attributed (partially attributable: >0% of grid cells, mostly attributable:
23 >50% of grid cells) From Callaghan et al. (2021) {Figure 1.1}

24

25 Figure AI.6: Projected changes in global marine richness in 2100 compared to 2006.
26 Differences between current (year 2006) and projected (year 2100) cell species richness for Representative
27 Concentration Pathways (RCPs) RCP4.5 and RCP8.5 (Garca Molinos et al. 2016).

28

29 Figure AI.07: Observed shifts in distribution of plant functional types.
30 Observed shifts in the distribution of plant functional types over the 17002020. Shifts in plant functional
31 types are indicative of shift in biome function and structure {Box 2.1, Figure Box 2.1.1}

32

33 Figure AI.08: Projected responses of rangeland plants to CO2 fertilization.
34 Regional percent changes in land cover and soil carbon from ensemble simulation results and plant responses
35 to CO2 fertilisation. Regions as defined by the United Nations Statistics Division. (Boone et al., 2018)
36 {5.5.3; Figure 5.11}

37

38 Figure AI.09: People living in land area of high conservation importance:
39 {CCP1.2.1.3, Figures CCP1.1, CCP1.2}

40

41 Figure AI.10: Present & projected habitat losses of climatically suitable area in terrestrial biodiversity
42 hotspots.
43 Projected loss for present-day (around 1C warming) and at global warming levels of 1.5C, 2C and 3C.
44 Maps (right hand column) show the regional distribution of losses in five categories of loss (Very low loss
45 020%, Low loss 2040%, Medium loss 4060%, High loss 6080%, Very high loss 80100%). The
46 clusters of circles (middle column) show losses in the five categories of loss in each of the 143 hotspot areas
47 of high importance for terrestrial biodiversity conservation with circles scaled by area size. {CCP1, Figure
48 CCP1.6; Table CCP1.1}

49

50 Figure AI.11: Projected change in marine animal biomass.
51 Simulated global biomass changes of animals. Spatial patterns of simulated change by 20902099 are
52 calculated relative to 19952014 for SSP1-2.6 and SSP5-8.5. The ensemble projections of global changes in
53 total animal biomass updated based on Tittensor et al. (2021) include 69 published global fisheries and
54 marine ecosystem models from the Fisheries and Marine Ecosystem Model Intercomparison Project (Fish-
55 MIP, Tittensor et al., 2018; Tittensor et al., 2021), forced with standardised outputs from two CMIP6 Earth
56 System Models. {3.4.3; Fig. 3.21}

Do Not Cite, Quote or Distribute  AI-62    Total pages: 74
FINAL DRAFT                       Annex I  IPCC WGII Sixth Assessment Report

 1

 2 Figure AI.12: Projected change in marine zooplankton biomass.
 3 Simulated global biomass changes of zooplankton. In the multi-model mean (solid lines) and very likely
 4 range (envelope) over 20002100 relative to 19952014, for SSP1-2.6 and SSP5-8.5. Spatial patterns of
 5 simulated change by 20902099 are calculated relative to 19952014 for SSP1-2.6 and SSP5-8.5.
 6 Confidence intervals can be affected by the number of models available for the Coupled Model
 7 Intercomparison Project 6 (CMIP6) scenarios and for different variables.The ensemble projections of global
 8 changes in zooplankton biomasses updated based on Kwiatkowski et al. (2019) include, under SSP1-2.6 and
 9 SSP5-8.5, respectively, a total of nine and 10 CMIP6 Earth System Models (ESMs). {3.4.3.4., Figure 3.21}

10

11 Figure AI.13: Spatial patterns of simulated change in total phytoplankton biomass.
12 Simulated global biomass changes of surface phytoplankton. In the multi-model mean (solid lines) and very
13 likely range (envelope) over 20002100 relative to 19952014, for SSP1-2.6 and SSP5-8.5. Spatial patterns
14 of simulated change by 20902099 are calculated relative to 19952014 for SSP1-2.6 and SSP5-8.5.
15 Confidence intervals can be affected by the number of models available for the Coupled Model
16 Intercomparison Project 6 (CMIP6) scenarios and for different variables. The ensemble projections of global
17 changes in phytoplankton biomasses updated based on Kwiatkowski et al. (2019) include, under SSP1-2.6
18 and SSP5-8.5, respectively, a total of nine and 10 CMIP6 Earth System Models (ESMs). {3.4.3.4., Figure
19 3.21}

20

21 Figure AI.14: Spatial patterns of simulated change in total benthic animal biomass.
22 Simulated global biomass changes of seafloor benthos. In the multi-model mean (solid lines) and very likely
23 range (envelope) over 20002100 relative to 19952014, for SSP1-2.6 and SSP5-8.5. Spatial patterns of
24 simulated change by 20902099 are calculated relative to 19952014 for SSP1-2.6 and SSP5-8.5.
25 Confidence intervals can be affected by the number of models available for the Coupled Model
26 Intercomparison Project 6 (CMIP6) scenarios and for different variables. Globally integrated changes in total
27 seafloor biomass have been updated based on Yool et al. (2017) with one benthic model (Kelly-Gerreyn et
28 al., 2014) forced with the CMIP6 ESM.

29

30 Figure AI.15: Projected exposure of biodiversity.
31 Global warming levels (GMST) modelled across the ranges of more than 30,000 marine and terrestrial
32 species. Figure based on Trisos et al 2020. {CCP 1; Figure 3.20}.

33

34 Figure AI.16: Projected loss of terrestrial and freshwater biodiversity compared to pre-industrial
35 period.
36 Global warming levels (GSAT); change indicated by the proportion of species (modelled n=119,813 species
37 globally) for which the climate is projected to become unsuitable across their current distributions. {Figure
38 2.6}

39

40 Figure AI.17: Regional impacts to major crop yields and food production loss events.
41 Trends in food production shocks in different food supply sectors from 1961-2-13 (Cottrell et al., 2019).
42 Projected impacts are for RCP 4.5 mid 21st century, taking into account adaptation and CO2 fertilisation for
43 crop yield productivity {Figure 5.3; 5.5.3; 5.4.1; Figure FAQ 5.1; Figure 9.22; 15.3.4; 15.3.3}

44

45 Figure AI.18: Climatic and environmental stresses on global production of wheat.
46 The global effects of five climatic and environmental stresses on wheat yield. The combined effect of each
47 stress on yield is presented as a Yield Constraint Score (YCS) on a five-category scale from low stress to
48 high stress (Mills et al., 2018). Higher temperatures enhance not only ozone production but also ozone
49 uptake by plants thus exacerbating yield loss and quality damage. Data are available at Sharps et al., (2020).
50 All data are presented for the 1  1 (latitude and longitude) grid squares where the mean production of
51 wheat was >500 tonnes (0.0005 Tg). {5.4.1; Fig. 5.5}

52

53 Figure AI.19: Climatic and environmental stresses on global production of soybean.
54 The global effects of five climatic and environmental stresses on soybean yield. The combined effect of each
55 stress on yield is presented as a Yield Constraint Score (YCS) on a five-category scale from low stress to
56 high stress (Mills et al., 2018). Higher temperatures enhance not only ozone production but also ozone

Do Not Cite, Quote or Distribute  AI-63    Total pages: 74
FINAL DRAFT                       Annex I  IPCC WGII Sixth Assessment Report

 1 uptake by plants thus exacerbating yield loss and quality damage. Data are available at Sharps et al., (2020).
 2 All data are presented for the 1  1 (latitude and longitude) grid squares where the mean production of
 3 soybean was >500 tonnes (0.0005 Tg). {5.4.1; Fig. 5.5}

 4

 5 Figure AI.20: Climatic and environmental stresses on global production of rice.
 6 The global effects of five climatic and environmental stresses on rice yield. The combined effect of each
 7 stress on yield is presented as a Yield Constraint Score (YCS) on a five-category scale from low stress to
 8 high stress (Mills et al., 2018). Higher temperatures enhance not only ozone production but also ozone
 9 uptake by plants thus exacerbating yield loss and quality damage. Data are available at Sharps et al., (2020).
10 All data are presented for the 1  1 (latitude and longitude) grid squares. {5.4.1; Fig. 5.5}

11

12 Figure AI.21: Climatic and environmental stresses on global production of maize.
13 The global effects of five climatic and environmental stresses on maize yield. The combined effect of each
14 stress on yield is presented as a Yield Constraint Score (YCS) on a five-category scale from low stress to
15 high stress (Mills et al., 2018). Higher temperatures enhance not only ozone production but also ozone
16 uptake by plants thus exacerbating yield loss and quality damage. Data are available at Sharps et al., (2020).
17 All data are presented for the 1  1 (latitude and longitude) grid squares. {5.4.1; Fig. 5.5}

18

19 Figure AI.22: Projected changes in global maize production.
20 For maize production time series are shown as relative changes to the 1983-2013 reference period under
21 SSP126 (green) and SSP585 (yellow). Shaded ranges illustrate the interquartile range of all climate and crop
22 model combinations (5 GCMs x 8 GGCMs). The solid line shows the median climate and crop model
23 response (and a 30yr moving average). Horizontal dashed lines mark the 5th and 95th percentile of the
24 historical variability (1983-2013; ensemble median) and open circles highlight the "time of climate impact
25 emergence" (TCIE), the year in which the smoothed median response exceeds the historical envelope. For
26 context, the TCIE calculated from GC5 5 simulations is indicated in lighter shades above the TCIE based on
27 GC6 (>2099 if no TCIE occurs by 2099). The maps (c, d) show median yield changes (2069-2099) under
28 SSP585 across climate and crop models for current growing regions (>10 ha). Hatching indicates areas
29 where less than 70% of the climate-crop model combinations agree on the sign of impact. Regional
30 production time series (e) are similar to (a), but stratified for the four major KoeppenGeiger climate zones
31 (temperature limited, temperate/humid, subtropical, and tropical). The percentage of the total global
32 production contributed by each zone is indicated in the top right corner of the inlets. All data are shown for
33 the default [CO2] {Jgermeyr et al. 2021; 5.4.3.2}

34

35 Figure AI.23: Projected changes in global wheat production.
36 Production time series are shown as relative changes to the 1983-2013 reference period under SSP126
37 (green) and SSP585 (yellow). Shaded ranges illustrate the interquartile range of all climate and crop model
38 combinations (5 GCMs x 8 GGCMs). The solid line shows the median climate and crop model response (and
39 a 30yr moving average). Horizontal dashed lines mark the 5th and 95th percentile of the historical variability
40 (1983-2013; ensemble median) and open circles highlight the "time of climate impact emergence" (TCIE),
41 the year in which the smoothed median response exceeds the historical envelope. For context, the TCIE
42 calculated from GC5 5 simulations is indicated in lighter shades above the TCIE based on GC6 (>2099 if no
43 TCIE occurs by 2099). The maps (c, d) show median yield changes (2069-2099) under SSP585 across
44 climate and crop models for current growing regions (>10 ha). Hatching indicates areas where less than 70%
45 of the climate-crop model combinations agree on the sign of impact. Regional production time series (e) are
46 similar to (a), but stratified for the four major KoeppenGeiger climate zones (temperature limited,
47 temperate/humid, subtropical, and tropical). The percentage of the total global production contributed by
48 each zone is indicated in the top right corner of the inlets. All data are shown for the default (CO2)
49 (Jgermeyr et al. 2021). {5.4.3.2}

50

51 Figure AI.24: Rainfed agriculture: drought risks, hazards, exposure & vulnerability indicators.
52 Hazard and exposure indicator score (a), vulnerability index (b) and drought risk index (c), for rainfed
53 agricultural systems between 1986 and 2015. Drought hazard indicator is defined as the ratio of actual crop
54 evapotranspiration to potential crop evapotranspiration, calculated for 24 crops. Vulnerability index is the
55 country-scale weighted average of a total of 64 indicators including social and ecological susceptibility

Do Not Cite, Quote or Distribute  AI-64    Total pages: 74
FINAL DRAFT                       Annex I  IPCC WGII Sixth Assessment Report

 1 indicators, and coping capacity. Risk index is calculated by multiplying hazard/exposure indicator score and
 2 vulnerability index (Meza et al., 2020). {Figure 5.5}

 3

 4 Figure AI.25: Extreme stress for livestock driven by temperature and humidity.
 5 Change in the number of days per year above "extreme stress" values from 2000 to the 2090s for livestock
 6 globally. Extreme stress conditions estimated using the Temperature Humidity Index (THI). Distributions of
 7 livestock in 2090s assumed to be the same as historical global distribution. {Fig 5.12}

 8

 9 Figure AI.26: Temperature and humidity-driven reduction in physical work capacity for humans
10 working outdoors
11 Projected increase in the number of days per year where physical work capacity is less than 50% based on
12 average daily air temperature and relative humidity. Physical work capacity is defined as the maximum
13 physical work output that can be reasonably expected from an individual performing moderate to heavy
14 work in a `cool' reference environment of 15oC. {Figure 5.17}

15

16 Figure AI.27: Full mortality risk and climate change.
17 Change in full risk mortality due to increases in temperatures. Estimates come from a model accounting for
18 both the costs and the benefits of adaptation, and the map shows the climate model weighted mean estimate
19 across Monte Carlo simulations conducted on 33 climate models (Carleton et al., 2018). {Figure 9.35,
20 9.10.1}

21

22 Figure AI.28: Projected geographical shift of the human temperature niche.
23 Geographical position of the human temperature niche projected on the current situation and the RCP8.5
24 projected 2070 climate. Those maps represent relative human distributions (summed to unity) for the
25 imaginary situation that humans would be distributed over temperatures following the stylized double
26 Gaussian model fitted to the modern data. Difference between the maps, visualizing potential source and
27 sink areas for the coming decades if humans were to be relocated in a way that would maintain this
28 historically stable distribution with respect to temperature. (Xu et al., 2020) {Table 8.7; 8.4.5.6}

29

30 Figure AI.29: Global population exposed to hyperthermia from extreme heat.
31 Global distribution of population exposed to hyperthermia from extreme heat and humidity. Maps indicate
32 the historical and projected number of days in a year in which conditions of air temperature and humidity
33 surpass a common threshold beyond which conditions turned deadly and pose a risk of death (Mora et al.,
34 2017). Largest fifteen urban areas by population size/number of citizens during 2020, 2050, and 2100
35 respectively as projected by Hoornweg and Pope (2017) {Figure 6.3; 6.2.3.1}"

36

37 Figure AI.30: Present-day global distribution of camps for refugees & internally displaced people.
38 The global distribution of the United Nations High Commissioner for Refugees (UNHCR) refugee and
39 internally displaced people (IDP) settlements (as of 2018) overlaid with annual mean near surface air
40 temperature (C) in 2040-2059 under RCP8.5. {Figure Box 8.1.1; Box 8.1}

41

42 Figure AI.31: Estimated relative human dependence on marine ecosystems.
43 Relative human dependence on marine resources for coastal protection, nutrition, fisheries economic benefits
44 and overall. Each bar represents an index value that semi-quantitatively integrates the magnitude,
45 vulnerability to loss and substitutability of the benefit. Indices synthesize information on people's
46 consumption of marine protein and nutritional status, gross domestic product, fishing revenues,
47 unemployment, education, governance and coastal characteristics. Overall dependence is the mean of the
48 three index values after standardization from 01 (Details are found in Table 1 and supplementary material
49 of (Selig et al., 2019)). This index does not include the economic benefits from tourism or other ocean
50 industries, and data limitations prevented including artisanal or recreational fisheries or the protective impact
51 of saltmarshes (Selig et al., 2019). Values for reference regions established in the WGI AR6 Atlas (Gutirrez
52 et al., 2021) were computed as area-weighted means from original country-level data. {Figure 3.1}

53

Do Not Cite, Quote or Distribute  AI-65    Total pages: 74
FINAL DRAFT                       Annex I  IPCC WGII Sixth Assessment Report

 1 Figure AI.32: Regional vulnerabilities to impacts of current and projected climate change on marine
 2 fishery and terrestrial livestock resources.
 3 (a) Ocean areas are delineated into FAO (Food and Agricultural Organization of the United Nations) regions.
 4 Ocean sensitivity is calculated from aggregated sensitivities from Blasiak et al. (2017) S1 country data based
 5 on number of fishers, fisheries exports, proportions of economically active population working as fishers,
 6 total fisheries landings and nutritional dependence, which was subsequently reanalyzed for each FAO region
 7 depicted here. Arrows denote projected average commercial (light blue) and artisanal (orange arrows)
 8 fishing resource shifts in location under RCP2.6 and under RCP8.5 (dark blue and red arrows respectively)
 9 scenarios by 2100. Text boxes highlight examples of vulnerabilities (Bell et al., 2018a), conflicts (Miller et
10 al., 2013; Blasiak et al., 2017; sthagen et al., 2020), or opportunities for marine resource usage (Robinson
11 et al., 2015; Stuart-Smith et al., 2018; Meredith et al., 2019). (b) Projected changes in the number of extreme
12 heat stress days per year for cattle (Bos taurus, temperate sub-regions, grey background; Bos indicus,
13 tropical sub-regions, orange background) from 2000 to the 2090s, shown as arrows rooted in the most
14 affected area in each IPCC sub-region pointing to the nearest area of reduced or no extreme heat stress.
15 Arrows are shown only for sub-regions where > 1 million additional animals affected. Areas in green are
16 those with >5000 animals per 0.5 degree grid cell (Thornton et al., 2021). {Cross-Chapter Box MOVING
17 PLATE Figure 1}

18

19 Figure AI.33: Current fisheries adaptive capacity to climate change and regional dependence on
20 seafood micronutrients in human diets.
21 Global documented fisheries management adaptive capacity to climate change and regional dependencies on
22 micronutrients from fisheries. 1. Fisheries management adaptive capacity is a function of: averaged GDP
23 World Development Indicators for 2018 (World Bank, 2020); climate awareness assessments of 30 of the
24 FAO (Food and Agricultural Organization of the United Nations) recognized most recent Regional Fisheries
25 Management Organizations with direct fisheries linkages; governance effectiveness index based on six
26 aggregate indicators (voice and accountability, political stability and absence of violence / terrorism,
27 government effectiveness, regulatory quality, rule of law, control of corruption) from 2018 World
28 Governance Indicator (World Bank, 2019) data, and; heterogeneity of countries within each FAO zone
29 (highly heterogeneous regions are less likely to establish sustainable and efficient fisheries management for
30 the entire FAO zone). Adaptative capacity index ranges from 1 (high) to 0 (no adaptative capacity). Ocean
31 areas are delineated into FAO regions. 2. Nutritional dependence of regional human populations on
32 micronutrient supply from marine fisheries. Nutritional dependence scale ranges from 100 (full dependence)
33 to 0 (no dependence). (Beal et al. 2017). {Cross-Chapter Box MOVING PLATE Figure 3 in Chapter 5}

34

35 Figure AI.34: Climate change risk to fisheries in Africa.
36 Inland fisheries (panels a-e): (a) Countries' reliance on inland fisheries was estimated by catch (total, tonnes)
37 (FAO, 2018b; Fluet-Chouinard et al., 2018), per capita catch (kg/person/year) (FAO, 2018b), percent
38 reliance on fish for micronutrients, and percent consumption per household (Golden et al., 2016). Z-scores of
39 each metric were averaged for each country to create a composite index describing `current dependence on
40 freshwater fish' for each country with darker blue colours indicating higher dependence. (bc) Projected
41 concentrations (numbers) of vulnerable freshwater fishery species averaged within freshwater ecoregions
42 under >2C global warming (b) and >4C global warming (c) estimated from recent past (19611992) to the
43 end of the 21st century (2071 to 2100) (Nyboer et al., 2019). Numbers of vulnerable fish species translate to
44 an average of 5568% vulnerable at >2C and 7797% vulnerable at <4C global warming. Darker reds
45 indicate higher concentrations of vulnerable fish species. (de) Countries (in green) that have an overlap
46 between high dependence on freshwater fish and high concentrations of fishery species that are vulnerable to
47 climate change under two warming scenarios. Inland fisheries (panels fj) comparing countries' current
48 percent dependence on marine foods for nutrition compared with projected change in maximum catch
49 potential (MCP) from marine fisheries. (f) The percentage of animal sources foods consumed that originate
50 from a marine environment. Countries with higher dependence are indicated by darker shades of blue
51 (Golden et al., 2016). (gh) Projected percent change in maximum catch potential (MCP) of marine fisheries
52 under 1.6C global warming (g) and >4C global warming (h) from recent past (19862005) to end of 21st
53 century (2081-2100) in countries' Exclusive Economic Zones (EEZs) (Cheung William et al., 2016). Darker
54 red indicates greater percent reduction [negative values]. (ij) Countries (in green) that have overlap between
55 high nutritional dependence and high reduction in MCP under two warming scenarios. {Figure 9.25, Figure
56 9.26}

57

Do Not Cite, Quote or Distribute  AI-66    Total pages: 74
FINAL DRAFT                       Annex I  IPCC WGII Sixth Assessment Report

 1 Figure AI.35: Regional synthesis of changes in water and consequent impacts on ecosystems and
 2 human systems.
 3 For physical changes, increase/decrease refers to changes in the amount or frequency of the measured
 4 variable, and the level of confidence refers to confidence that the change has occurred. For impacts on
 5 ecosystems and human systems, plus or minus marks depicts whether an observed impact of hydrological
 6 change is positive (beneficial) or negative (adverse), respectively, to the given system, and the level of
 7 confidence refers to confidence in attributing an impact on that system to a climate-induced hydrological
 8 change. Circles indicate that within that region, both increase and decrease of physical changes are found,
 9 but are not necessarily equal; or beneficial and adverse assessed impacts on ecosystems and human systems.
10 `na' indicates variables not assessed due to limited evidences. Agriculture refers to impacts on crop
11 production. Energy refers to impacts on hydro and thermoelectric power generation. {Figure 4.20}

12

13 Figure AI.36: Current global drought risk. Current global drought risk and its components.
14 (a) Drought hazard computed for the events between 1901 2010 by the probability of exceedance the
15 median of global severe precipitation deficits, using precipitation data from the Global Precipitation
16 Climatology Center (GPCC) for 19012010. (b) Drought vulnerability is derived from an arithmetic
17 composite model combining social, economic, and infrastructural factors proposed by UNISDR (2004). (c)
18 Drought exposure computed at the sub-national level with the non-compensatory DEA (Data Envelopment
19 Analysis) model (Cook et al., 2014). (d) Drought risk based on the above components of hazard,
20 vulnerability and exposure, scored on a scale of 0 (lowest risk) to 1(highest risk) with the lowest and highest
21 hazard, exposure, and vulnerability (Carro et al., 2016). {Figure 4.9}

22

23 Figure AI.37: Dependence of land surface areas and population on mountain water resources 1961
24 2050.
25 Results are shown as decadal averages for lowland population in each category of dependence on mountain
26 water from no surplus and negligible to essential. (a) Global mountain regions and their differentiated
27 importance for lowland water resources. (b) Lowland population and their differentiated dependence on
28 mountain water resources, both for the scenario combination SSP2-RCP6.0 and for the time period 2041
29 2050. (c) Number of lowland population and their differentiated dependence on mountain water resources
30 from the 1960's to the 2040's for three different scenario combinations (based on Viviroli et al., 2020).
31 {Figure CCP5.2}

32

33 Figure AI.38: Risk to livelihoods and the economy from changing mountain water resources.
34 The majority of studies assessed focus on impacts up to mid-century (20302060) and for RCP-2.6, RCP-4.5
35 and RCP-6.0, which was converted into the corresponding warming level range 1.5-2.0C GWL (see CCB
36 CLIMATE). Methodological details are provided in Section SMCCP5.4, Figure SMCCP5.1, Table
37 SMCCP5.16 and SMCCP5.18. Due to the limited evidence available to determine risks against high Global
38 Warming Levels (GLWs), and the relatively high uncertainties associated with future irrigation trends for the
39 second half of the century (see e.g. Viviroli et al., 2020), assessment of risks associated with GLWs greater
40 than 2.0C GWL was not conducted. {Figure CCP5.6}

41

42 Figure AI.39: The effect of regional sea level rise on extreme sea level events at coastal locations.
43 (a) Schematic illustration of extreme sea level events and their average recurrence in the recent past (1986
44 2005) and the future. As a consequence of mean sea level rise, local sea levels that historically occurred once
45 per century (historical centennial events, HCEs) are projected to recur more frequently in the future. (b) The
46 year in which HCEs are expected to recur once per year on average under RCP8.5 and RCP2.6, at the 439
47 individual coastal locations where the observational record is sufficient. The absence of a circle indicates an
48 inability to perform an assessment due to a lack of data but does not indicate absence of exposure and risk.
49 The darker the circle, the earlier this transition is expected. The likely range is 10 years for locations where
50 this transition is expected before 2100. White circles (33% of locations under RCP2.6 and 10% under
51 RCP8.5) indicate that HCEs are not expected to recur once per year before 2100. (c) An indication at which
52 locations this transition of HCEs to annual events is projected to occur more than 10 years later under
53 RCP2.6 compared to RCP8.5. As the scenarios lead to small differences by 2050 in many locations results
54 are not shown here for RCP4.5 but they are available in Chapter 4. {4.2.3, Figure 4.10, Figure 4.12}

55

56 Figure AI.40: Relative trends in projected regional shoreline change.

Do Not Cite, Quote or Distribute  AI-67    Total pages: 74
FINAL DRAFT                       Annex I  IPCC WGII Sixth Assessment Report

 1 Advance/retreat relative to 2010. Frequency distributions of median projected change by (a,c) 2050 and (b,d)
 2 2100 under (a,b) RCP4.5 and (c,d) RCP8.5. Projections account for both long-term shoreline dynamics and
 3 sea-level rise and assume no impediment to inland transgression of sandy beaches. Data for small island
 4 states are aggregated and plotted in the Caribbean. Data from Vousdoukas et al. (2020b). Values for
 5 reference regions established in the WGI AR6 Atlas (Gutirrez et al., 2021) were computed as area-weighted
 6 means from original country-level data. For model assumptions and associated debate, see Vousdoukas et al.
 7 (2020a) and Cooper et al. (2020a).{Figure 3.14}

 8

 9 Figure AI.41: Population living in small islands that may be exposed to coastal inundation.
10 Projected percentage of current population in selected small islands occupying vulnerable land (the number
11 of people on land that may be exposed to coastal inundation--either by permanently falling below Mean
12 Higher High Water, or temporarily falling below the local annual flood height) (adapted from Kulp et al.
13 2019, using the CoastalDEM_Perm_p50 model). Positions on the map are based on the capital city or largest
14 town. {Figure 15.3}

15

16 Figure AI.42: Projected number of people at risk of a 100-year coastal flood.
17 The size of the circle represents the number of people at risk per IPCC region and the colours show the
18 timing of risk based on projected sea-level rise (Haasnoot et al., 2021) under three different Shared
19 Socioeconomic Pathways (SSPs). Darker colours indicate earlier in setting risks. The left side of the circles
20 shows absolute population at risk and the right side the share of the population in percentage. {Figure
21 CCP2.4; Figure 13.6; Figure 15.3}.

22

23 Figure AI.43: Selected African cities exposed to sea level rise.
24 Selected African cities exposed to sea level rise include (a) Dar es Salaam, Bagamoyo, and Stone Town in
25 Tanzania (East Africa), (b) Lagos in Nigeria, and Cotonou and Porto-Novo in Benin (West Africa), and (c)
26 Cairo and Alexandria in Egypt (North Africa). Orange shows built-up area in 2014. Shades of blue show
27 permanent flooding due to sea level rise by 2050 and 2100 under low (RCP2.6), medium (RCP4.5) and high
28 (RCP8.5) greenhouse gas emissions scenarios. Darker colours for higher emissions scenarios show areas
29 projected to be flooded in addition to those for lower emissions scenarios. The figure assumes failure of
30 coastal defences in 2050 and 2100. Some areas are already below current sea level rise and coastal defences
31 need to be upgraded as sea level rises (e.g., in Egypt), others are just above mean sea levels and they do not
32 necessarily have high protection levels, so these defences need to be built (e.g., Dar Es Salam and Lagos).
33 Blue shading shows permanent inundation surfaces predicted by Coastal DEM and SRTM given the 95th
34 percentile K14/RCP2.6, RCP4.5, and RCP8.5, for present day, 2050, and 2100 sea level projection for
35 permanent inundation (inundation without a storm surge event), and RL10 (10-year return level storm) (Kulp
36 and Strauss, 2019). Low-lying areas isolated from the ocean are removed from the inundation surface using
37 connected components analysis. Current water bodies are derived from the SRTM Water Body Dataset.
38 Orange areas represent the extent of coastal human settlements in 2014 (Corbane et al., 2018). See Figure
39 CCP4.7 for projections including subsidence and worst-case scenario projections for 2100. {Figure 9.29}.

40

41 Figure AI.44: Risk of historical and projected river flooding.
42 (a) Vulnerability. Modelled mean global fluvial flood water depth (Tanoue et al., 2016; Tanoue et al., 2021)
43 based on a land surface model and a river and inundation model driven by reanalysis climate forcing of 5
44 CMIP5 GCMs (metres). The annual maximum daily river water was allocated along elevations, and
45 inundation depth was calculated for each year and averaged for the target period. (b) Hazard. Local flood
46 protection standard (return period) at sub-country scale (Scussolini et al., 2016) based on published reports
47 and documents, websites and personal communications with experts. Note that the vulnerability of this map
48 reflects local flood protection such as complex infrastructure and does not fully reflect the other source of
49 vulnerabilities, including exposure. (c) Exposure. Population distribution per 30 arc second grid cell (Klein
50 Goldewijk et al., 2010; Klein Goldewijk et al., 2011). (d) Risk as population exposed to flood (number of
51 people where inundation occurs) per 30 arc-second grid cell. Population under inundation depth > 0 m (a)
52 was counted when the return period of annual maximum daily river water exceeds the flood protection
53 standard (c). {Figure 4.8}

54

Do Not Cite, Quote or Distribute  AI-68    Total pages: 74
FINAL DRAFT                       Annex I  IPCC WGII Sixth Assessment Report

 1 Figure AI.45: Projected changes in river flooding.
 2 Multi-model median return period (years) in the 2080s for the 20th-century 100-year river flood, based on a
 3 global river and inundation model, CaMa-Flood, driven by runoff output of 9 CMIP6 Models in the SSP1-
 4 2.6 (a), SSP2-4.5 (b) and SSP5-8.5 (c) scenario respectively. All changes are estimated in 2071-2100 relative
 5 to 1970-2000. A dot indicates regions with high model consistency (more than 7 models out of 9 show the
 6 same direction of change). (d) Global or regional potential exposure (% to the total population affected by
 7 flooding) under different warming levels with constant population scenario of CMIP5 (Alfieri et al., 2017)
 8 and with population scenario of SSP5 of CMIP6 (bar chart, (Hirabayashi et al., 2021b)). Inundation is
 9 calculated when the magnitude of flood exceeds current flood protection (Scussolini et al., 2016). Note that
10 number of GCMs used to calculate Global Warming Level (GWL) 4.0 is less than that for other SWLs, as
11 the global mean temperature of some GCMs did not exceed 4C. {Figure 4.17}

12

13 Figure AI.46 Burning ember diagrams of regional & global risk assessments.
14 {Reasons for concern: 16.6.3.1  16.6.3.5; 16.6.4; Table SM16.18 in Supplementary Material SM16.6
15 presents the consensus values of the transition range and median estimate in terms of global warming level
16 by risk level for each of the five RFC embers. Africa: 9.2; Table 9.2; For range of global warming levels for
17 each risk transition used to make this figure see Supplementary Material Table SM 9.1. Australia and New
18 Zealand/ Australia: The assessment is based on available literature and expert judgement, summarized in
19 Table 11.14 and described in Supplementary Material SM 11.2. Mediterranean: See CCP4.3.2-8 and
20 Supplementary Tables SMCCP4.2a-h for details. Europe: 13.10.2; More details on each burning ember are
21 provided in Sections 13.10.2.1-13.10.2.4 and SM13.10. North America: 14.6.2; 14.6.3; Table 14.3, see
22 SM14.4. for detailed information. Arctic: CCP6.3.1; Table CCP6.5; The supporting literature and methods
23 are provided in SMCCP6.6. Ecosystems: Terrestrial and freshwater: Tables 2.5 and 2.S.4 provide details of
24 the key risks and temperature levels for the risk transitions. Ocean: Special Report on the Ocean and
25 Cryosphere in a Changing Climate (SROCC). Health: 7.3.1; Based on (Ebi et al., 2021).}

26

27 Figure AI.47: Evidence of transformative adaptation by sector and region.
28 Evidence of transformational adaptation does not imply effectiveness, equity, or adequacy. Evidence of
29 transformative adaptation is assessed based on the scope, speed, depth, and ability to challenge limits of
30 responses reported in the scientific literature Studies relevant to multiple regions or sectors are included in
31 assessment for each relevant sector/region. {16.3.2; Figure 16.6}.

32

33 Figure AI.48: Drought is exacerbating water management challenges which vary across regions with
34 respect to anticipated water scarcity conditions by 2050.
35 Local levels of policy challenges for addressing water scarcity by 2050, considering both the central estimate
36 (median) and the changing uncertainty in projections of the Water Scarcity Index (WSI) from the present day
37 to 2050. Projections used five CMIP5 climate models, three global hydrological models from ISIMIP, and
38 three Shared Socioeconomic Pathways (SSPs).Reproduced from (Greve et al., 2018). {Figure Box 4.1.1;
39 Box 4.1}.

40

41 Figure AI.49: Observed water-related adaptation responses with positive outcomes.
42 (a) Location of case studies of water-related adaptation responses (996 data points from 319 studies). In
43 these 996 data points, at least one positive outcome was recorded in one of the five outcome indicators.
44 These outcome indicators are economic/financial, outcomes for vulnerable people, ecological/environmental,
45 water-related, and socio-cultural and institutional. (b) In most instances, the top six adaptation categories
46 include nearly 3/4th of the studies. (c) Due to a small number of studies in small island states, a spider
47 diagram was not generated for the Small Island States. {Figure 4.27}

48

49 Figure AI.50: Projected effectiveness of water-related adaptation options.
50 Effectiveness in returning the system to a study-specific baseline state relative to the projected climate
51 impact; and level of residual risk retained after adaptation, relative to baseline conditions. Regional
52 summaries are based on IPCC regions. Warming levels refer to the global mean temperature (GMT) increase
53 relative to a 1850-1900 baseline. For each data point, the study-specific GMT increase was calculated to
54 show effectiveness at 1.5C, 2C, 3C and 4C. Based on the ability of an implemented option to return the
55 system to its baseline state, the effectiveness is classified based on the share of risk the option can reduce:
56 Large (>80%); Moderate (80-50%); Small (<50-30%); Insufficient (<30%). Where the system state is

Do Not Cite, Quote or Distribute  AI-69    Total pages: 74
FINAL DRAFT                       Annex I  IPCC WGII Sixth Assessment Report

 1 improved relative to baseline, Co-benefits are identified. Residual impacts show the share of remaining
 2 impacts after adaptation has been implemented: Negligible (<5%); Small (5 to <20%); Moderate (20 to <50);
 3 Large (50% and more). Where risks increase after adaptation, data points are shown as maladaptation. All
 4 underlying data is provided in SM4.8. {Figure 4.28}

 5

 6 Figure AI.51: Evidence of observed adaptation across regions in food, fibre, and other ecosystem
 7 products.
 8 Stage of implementation; Type of adaptation; Inclusion of Indigenous knowledge and local knowledge (IK
 9 and LK) based on Global Adaptation Mapping Initiative (GAMI) database  (Berrang-Ford et al., 2021a).
10 The bars indicate the number of evidence for the options x region. {Figure 5.21}

11

12 Figure AI.52: Who is responding, by geographic region and sector?
13 (a) Cell contents indicate the number of publications reporting engagement of each actor in adaptation-
14 related responses. Darker colours denote a high number of publications. (b) Percentages reflect the number
15 of articles mentioning each type of adaptation over the total number of articles for that region. Radar values
16 do not total 100% per region since publications frequently report multiple types of adaptation; for example,
17 construction of drainage systems (infrastructural), changing food storage practices by households
18 (behavioural), and planting of tree cover in flood prone areas (nature-based) in response to flood risk to
19 agricultural crops. Data updated and adapted from (Berrang-Ford et al., 2021b), based on 1682 scientific
20 publications reporting on adaptation-related responses in human systems. {Figure 16.4; Figure 16.5}

21

22 Figure AI.53: The Urban Adaptation Gap.
23 This is a qualitative assessment presenting individual, non-comparative data for world regions from 25 AR6
24 Contributing Lead Authors and Lead Authors, the majority from regional chapters. Respondents were asked
25 to make expert summary statements based on the data included within their chapters and across the AR6
26 report augmented by their expert knowledge. Multiple iterations allowed opportunity for individual and
27 group judgement. Urban populations and risks are very diverse within regions making the presented results
28 indicative only. Variability in data coverage leads to the overall analysis having medium agreement 
29 medium evidence. Major trends identified in 6.3.1 at least meet this level of confidence. Analysis is
30 presented for current observed climate change associated hazards and for three adaptation scenarios: (1)
31 current adaptation (based on current levels of risk management and climate adaptation), (2) planned
32 adaptation (assessing the level of adaptation that could be realised if all national, city and neighbourhood
33 plans and policies were fully enacted), (3) transformative adaptation (if all possible adaptation measures
34 were to be enacted). Assessments were made for the lowest and highest quintile by income. Residual risk
35 levels achieved for each income class under each adaptation scenario are indicated by five adaptation levels:
36 no risk, occasional discomfort, occasional impacts on wellbeing, frequent impacts on wellbeing, extreme
37 events and/or chronic risk. The urban adaptation gap is revealed when levels of achieved adaptation fall short
38 of delivering `no risk'. The graphic uses IPCC Regions, and has split Asia into two regions: North and East
39 Asia, and Central and South Asia. {Figure 6.4}

40

41 Figure AI.54: Evidence on constraints and limits to adaptation by region and sector.
42 Data from (Thomas et al. 2021), based on 1682 scientific publications reporting on adaptation-related
43 responses in human systems. See 16.A.1 for methods. Low evidence: <20% of assessed literature has
44 information on limits, literature mostly focuses on constraints to adaptation Medium evidence: between 20-
45 40% of assessed literature has information on limits, literature provides some evidence of constraints being
46 linked to limits High evidence: > 40% of assessed literature has information on limits, literature provides
47 broad evidence of constraints being linked to limits. {Figure 16.7}

48

49 Figure AI.55: Constraints associated with limits by region and sector.
50 Data from (Thomas et al. 2021), based on 1682 scientific publications reporting on adaptation-related
51 responses in human systems. See 16.A.1 for methods. Constraints are categorized as: (1) Economic: existing
52 livelihoods, economic structures, and economic mobility; (2) Social/cultural: social norms, identity, place
53 attachment, beliefs, worldviews, values, awareness, education, social justice, and social support; (3) Human
54 capacity: individual, organizational, and societal capabilities to set and achieve adaptation objectives over
55 time including training, education, and skill development; (4) Governance, Institutions & Policy: existing
56 laws, regulations, procedural requirements, governance scope, effectiveness, institutional arrangements,
57 adaptive capacity, and absorption capacity; (5) Financial: lack of financial resources; (6)

Do Not Cite, Quote or Distribute  AI-70    Total pages: 74
FINAL DRAFT                       Annex I  IPCC WGII Sixth Assessment Report

 1 Information/Awareness/Technology: lack of awareness or access to information or technology; (7) Physical:
 2 presence of physical barriers; and (8) Biologic/climatic: temperature, precipitation, salinity, acidity, and
 3 intensity and frequency of extreme events including storms, drought, and wind. Insufficient data: there is not
 4 enough literature to support an assessment (less than 5 studies available); Minor constraint: <20% of
 5 assessed literature identifies this constraint; Secondary constraint: 20-50% of assessed literature identifies
 6 this constraint; Primary constraint: >50% of assessed literature identifies this constraint. {Figure 16.8}

 7

 8 Figure AI.56: Distribution of adaptation finance across different regions and different types of finance.
 9 (a) Data for period 2015-2016, as tracked the Climate Policy Initiative. (b) Data for year 2018 from different
10 sources, through different instruments into different sectors and regions as collated by (CPI, 2020). Each
11 strand shows the relative proportion of finance flowing from one category to another (for example from
12 private or public sources to different instruments). Categories from left to right are: Use = whether the
13 finance is solely for adaptation or for adaptation and other objectives, including mitigation; Public/Private =
14 whether the finance comes from public or private sources; Instrument, the financing instrument; Sector = the
15 broad sectoral allocation; Region = the geographical distribution of funding (proportion of total in % and
16 per-capita allocation). {Figure Cross-Chapter Box FINANCE.2; Figure FAQ17.2.1}

17

18

19

Do Not Cite, Quote or Distribute  AI-71    Total pages: 74