Final Government Distribution                                           Chapter11                                                        IPCC AR6 WGI

 1   Table Of Content
 2
 3   Executive Summary ................................................................................................................................... 6
 4
 5   11.1       Framing ........................................................................................................................................11
 6      11.1.1        Introduction to the chapter ...................................................................................................... 11
 7      11.1.2        What are extreme events and how are their changes studied? .................................................. 11
 8      11.1.3        Types of extremes assessed in this chapter .............................................................................. 12
 9      11.1.4        Effects of greenhouse gas and other external forcings on extremes .......................................... 13
10
11   BOX 11.1:           Thermodynamic and dynamic changes in extremes across scales ....................................15
12
13      11.1.5        Effects of large-scale circulation on changes in extremes ........................................................ 17
14      11.1.6        Effects of regional-scale processes and forcings and feedbacks on changes in extremes........... 18
15      11.1.7        Global-scale synthesis............................................................................................................. 19
16
17   BOX 11.2: Low-likelihood high-impact changes in extremes ..................................................................24
18
19   11.2       Data and Methods ........................................................................................................................26
20      11.2.1        Definition of extremes ............................................................................................................ 26
21      11.2.2        Data........................................................................................................................................ 27
22
23   BOX 11.3: Extremes in paleoclimate archives compared to instrumental records.................................28
24
25      11.2.3        Attribution of extremes ........................................................................................................... 31
26      11.2.4        Projecting changes in extremes as a function of global warming levels .................................... 32
27
28   Cross-Chapter Box 11.1: Translating between regional information at global warming levels vs
29   scenarios for end users  .......................................................................................................................34
30
31   11.3       Temperature extremes .................................................................................................................38
32      11.3.1        Mechanisms and drivers ......................................................................................................... 38
33      11.3.2        Observed trends ...................................................................................................................... 40
34      11.3.3        Model evaluation .................................................................................................................... 43
35      11.3.4        Detection and attribution, event attribution.............................................................................. 45
36      11.3.5        Projections .............................................................................................................................. 47
37
38   11.4       Heavy precipitation ......................................................................................................................51
39      11.4.1        Mechanisms and drivers ......................................................................................................... 51
40      11.4.2        Observed Trends ..................................................................................................................... 52
41      11.4.3        Model evaluation .................................................................................................................... 56
42      11.4.4        Detection and attribution, event attribution.............................................................................. 57
43      11.4.5        Projections .............................................................................................................................. 59

     Do Not Cite, Quote or Distribute                                            11-2                                                         Total pages: 345
     Final Government Distribution                                        Chapter11                                                       IPCC AR6 WGI

 1
 2   11.5    Floods............................................................................................................................................63
 3     11.5.1      Mechanisms and drivers ......................................................................................................... 64
 4     11.5.2      Observed trends ...................................................................................................................... 65
 5     11.5.3      Model evaluation .................................................................................................................... 65
 6     11.5.4      Attribution .............................................................................................................................. 66
 7     11.5.5      Future projections ................................................................................................................... 67
 8
 9   11.6    Droughts .......................................................................................................................................68
10     11.6.1      Mechanisms and drivers ......................................................................................................... 68
11     11.6.1.1       Precipitation deficits ........................................................................................................... 69
12     11.6.1.2       Atmospheric evaporative demand ....................................................................................... 69
13     11.6.1.3       Soil moisture deficits .......................................................................................................... 70
14     11.6.1.4       Hydrological deficits........................................................................................................... 71
15     11.6.1.5       Atmospheric-based drought indices ..................................................................................... 71
16     11.6.1.6       Relation of assessed variables and metrics for changes in different drought types ................ 71
17     11.6.2      Observed trends ...................................................................................................................... 72
18     11.6.2.1       Precipitation deficits ........................................................................................................... 72
19     11.6.2.2       Atmospheric evaporative demand ....................................................................................... 72
20     11.6.2.3       Soil moisture deficits .......................................................................................................... 73
21     11.6.2.4       Hydrological deficits........................................................................................................... 74
22     11.6.2.5       Atmospheric-based drought indices ..................................................................................... 74
23     11.6.2.6       Synthesis for different drought types ................................................................................... 75
24     11.6.3      Model evaluation .................................................................................................................... 75
25     11.6.3.1       Precipitation deficits ........................................................................................................... 75
26     11.6.3.2       Atmospheric evaporative demand ....................................................................................... 75
27     11.6.3.3       Soil moisture deficits .......................................................................................................... 76
28     11.6.3.4       Hydrological deficits........................................................................................................... 77
29     11.6.3.5       Atmospheric-based drought indices ..................................................................................... 77
30     11.6.3.6       Synthesis for different drought types ................................................................................... 77
31     11.6.4      Detection and attribution, event attribution.............................................................................. 78
32     11.6.4.1       Precipitation deficits ........................................................................................................... 78
33     11.6.4.2       Soil moisture deficits .......................................................................................................... 79
34     11.6.4.3       Hydrological deficits........................................................................................................... 79
35     11.6.4.4       Atmospheric-based drought indices ..................................................................................... 80
36     11.6.4.5       Synthesis for different drought types ................................................................................... 80
37     11.6.5      Projections .............................................................................................................................. 81
38     11.6.5.1       Precipitation deficits ........................................................................................................... 81
39     11.6.5.2       Atmospheric evaporative demand ....................................................................................... 82
40     11.6.5.3       Soil moisture deficits .......................................................................................................... 83

     Do Not Cite, Quote or Distribute                                         11-3                                                        Total pages: 345
     Final Government Distribution                                         Chapter11                                                       IPCC AR6 WGI

 1      11.6.5.4        Hydrological deficits........................................................................................................... 84
 2      11.6.5.5        Atmospheric-based drought indices ..................................................................................... 85
 3      11.6.5.6        Synthesis for different drought types ................................................................................... 86
 4
 5   11.7      Extreme storms ............................................................................................................................87
 6      11.7.1       Tropical cyclones.................................................................................................................... 88
 7      11.7.1.1        Mechanisms and drivers...................................................................................................... 88
 8      11.7.1.2        Observed trends .................................................................................................................. 88
 9      11.7.1.3        Model evaluation ................................................................................................................ 90
10      11.7.1.4        Detection and attribution, event attribution .......................................................................... 92
11      11.7.1.5        Projections .......................................................................................................................... 94
12      11.7.2       Extratropical storms ................................................................................................................ 97
13      11.7.2.1        Observed trends .................................................................................................................. 97
14      11.7.2.2        Model evaluation ................................................................................................................ 98
15      11.7.2.3        Detection and attribution, event attribution .......................................................................... 98
16      11.7.2.4        Projections .......................................................................................................................... 98
17      11.7.3       Severe convective storms ........................................................................................................ 99
18      11.7.3.1        Mechanisms and drivers.................................................................................................... 100
19      11.7.3.2        Observed trends ................................................................................................................ 101
20      11.7.3.3        Model evaluation .............................................................................................................. 102
21      11.7.3.4        Detection and attribution, event attribution ........................................................................ 103
22      11.7.3.5        Projections ........................................................................................................................ 103
23      11.7.4       Extreme winds ...................................................................................................................... 104
24
25   11.8      Compound events .......................................................................................................................106
26      11.8.1       Overview .............................................................................................................................. 106
27      11.8.2       Concurrent extremes in coastal and estuarine regions ............................................................ 107
28      11.8.3       Concurrent droughts and heat waves ..................................................................................... 108
29
30   BOX 11.4: Case study: Global-scale concurrent climate anomalies at the example of the 2015-2016
31   extreme El Niño and the 2018 boreal spring/summer extremes ............................................................109
32
33   11.9      Regional information on extremes .............................................................................................113
34      11.9.1       Overview .............................................................................................................................. 113
35      11.9.2       Temperature extremes........................................................................................................... 114
36      11.9.3       Heavy precipitation............................................................................................................... 114
37      11.9.4       Droughts............................................................................................................................... 115
38
39   Frequently Asked Questions ...................................................................................................................117
40      FAQ 11.1:           How do changes in climate extremes compare with changes in climate averages? .......... 117
41      FAQ 11.2:           Will unprecedented extremes occur as a result of human-induced climate change? ........ 119

     Do Not Cite, Quote or Distribute                                         11-4                                                         Total pages: 345
     Final Government Distribution                                            Chapter11                                                         IPCC AR6 WGI

 1      FAQ 11.3:            Did climate change cause that recent extreme event in my country?............................... 120
 2
 3   Large tables ..........................................................................................................................................122
 4
 5   Acknowledgements .................................................................................................................................233
 6
 7   References           ..........................................................................................................................................234
 8
 9   Appendix 11.A ........................................................................................................................................314
10
11   Figures              ..........................................................................................................................................316
12
13




     Do Not Cite, Quote or Distribute                                             11-5                                                          Total pages: 345
     Final Government Distribution                     Chapter11                                   IPCC AR6 WGI

 1   Executive Summary
 2
 3   This chapter assesses changes in weather and climate extremes on regional and global scales, including
 4   observed changes and their attribution, as well as projected changes. The extremes considered include
 5   temperature extremes, heavy precipitation and pluvial floods, river floods, droughts, storms (including
 6   tropical cyclones), as well as compound events (multivariate and concurrent extremes). Changes in marine
 7   extremes are addressed in Chapter 9 and Cross-Chapter Box 9.1. Assessments of past changes and their
 8   drivers are from 1950 onward, unless indicated otherwise. Projections for changes in extremes are presented
 9   for different levels of global warming, supplemented with information for the conversion to emission
10   scenario-based projections (Cross-Chapter Box 11.1; Chapter 4, Table 4.2). Since AR5, there have been
11   important new developments and knowledge advances on changes in weather and climate extremes, in
12   particular regarding human influence on individual extreme events, on changes in droughts, tropical
13   cyclones, and compound events, and on projections at different global warming levels (1.5°C–4°C). These,
14   together with new evidence at regional scales, provide a stronger basis and more regional information for the
15   AR6 assessment on weather and climate extremes.
16
17   It is an established fact that human-induced greenhouse gas emissions have led to an increased
18   frequency and/or intensity of some weather and climate extremes since pre-industrial time, in
19   particular for temperature extremes. Evidence of observed changes in extremes and their attribution to
20   human influence (including greenhouse gas and aerosol emissions and land-use changes) has strengthened
21   since AR5, in particular for extreme precipitation, droughts, tropical cyclones and compound extremes
22   (including dry/hot events and fire weather). Some recent hot extreme events would have been extremely
23   unlikely to occur without human influence on the climate system. {11.2, 11.3, 11.4, 11.6, 11.7, 11.8}
24
25   Regional changes in the intensity and frequency of climate extremes generally scale with global
26   warming. New evidence strengthens the conclusion from SR1.5 that even relatively small incremental
27   increases in global warming (+0.5°C) cause statistically significant changes in extremes on the global
28   scale and for large regions (high confidence). In particular, this is the case for temperature extremes
29   (very likely), the intensification of heavy precipitation (high confidence) including that associated with
30   tropical cyclones (medium confidence), and the worsening of droughts in some regions (high
31   confidence). The occurrence of extreme events unprecedented in the observed record will increase with
32   increasing global warming, even at 1.5°C of global warming. Projected percentage changes in frequency are
33   higher for the rarer extreme events (high confidence). {11.1, 11.2, 11.3, 11.4, 11.6, 11.9, CC-Box 11.1}
34
35   Methods and Data for Extremes
36
37   Since AR5, the confidence about past and future changes in weather and climate extremes has
38   increased due to better physical understanding of processes, an increasing proportion of the scientific
39   literature combining different lines of evidence, and improved accessibility to different types of climate
40   models (high confidence). There have been improvements in some observation-based datasets,
41   including reanalysis data (high confidence). Climate models can reproduce the sign of changes in
42   temperature extremes observed globally and in most regions, although the magnitude of the trends
43   may differ (high confidence). Models are able to capture the large-scale spatial distribution of precipitation
44   extremes over land (high confidence). The intensity and frequency of extreme precipitation simulated by
45   Coupled Model Intercomparison Project Phase 6 (CMIP6) models are similar to those simulated by CMIP5
46   models (high confidence). Higher horizontal model resolution improves the spatial representation of some
47   extreme events (e.g., heavy precipitation events), in particular in regions with highly varying topography
48   (high confidence). {11.2, 11.3, 11.4}
49
50   Temperature Extremes
51
52   The frequency and intensity of hot extremes have increased and those of cold extremes have decreased
53   on the global scale since 1950 (virtually certain). This also applies at regional scale, with more than


     Do Not Cite, Quote or Distribute                     11-6                                     Total pages: 345
     Final Government Distribution                        Chapter11                                      IPCC AR6 WGI

 1   80% of AR6 regions 1 showing similar changes assessed to be at least likely. In a few regions, limited
 2   evidence (data or literature) prevents the reliable estimation of trends. {11.3, 11.9}
 3
 4   Human-induced greenhouse gas forcing is the main driver of the observed changes in hot and cold
 5   extremes on the global scale (virtually certain) and on most continents (very likely). The effect of
 6   enhanced greenhouse gas concentrations on extreme temperatures is moderated or amplified at the regional
 7   scale by regional processes such as soil moisture or snow/ice-albedo feedbacks, by regional forcing from
 8   land use and land-cover changes, or aerosol concentrations, and decadal and multidecadal natural variability.
 9   Changes in anthropogenic aerosol concentrations have likely affected trends in hot extremes in some regions.
10   Irrigation and crop expansion have attenuated increases in summer hot extremes in some regions, such as the
11   U.S. Midwest (medium confidence). Urbanization has likely exacerbated changes in temperature extremes in
12   cities, in particular for night-time extremes. {11.1, 11.2, 11.3}
13
14   The frequency and intensity of hot extremes will continue to increase and those of cold extremes will
15   continue to decrease, at both global and continental scales and in nearly all inhabited regions1 with
16   increasing global warming levels. This will be the case even if global warming is stabilized at 1.5°C.
17   Relative to present-day conditions, changes in the intensity of extremes would be at least double at 2°C, and
18   quadruple at 3°C of global warming, compared to changes at 1.5°C of global warming. The number of hot
19   days and hot nights and the length, frequency, and/or intensity of warm spells or heat waves will increase
20   over most land areas (virtually certain). In most regions, future changes in the intensity of temperature
21   extremes will very likely be proportional to changes in global warming, and up to 2–3 times larger (high
22   confidence). The highest increase of temperature of hottest days is projected in some mid-latitude and semi-
23   arid regions, at about 1.5 time to twice the rate of global warming (high confidence). The highest increase of
24   temperature of coldest days is projected in Arctic regions, at about three times the rate of global warming
25   (high confidence). The frequency of hot temperature extreme events will very likely increase non-linearly
26   with increasing global warming, with larger percentage increases for rarer events. {11.2, 11.3, 11.9; Table
27   11.1; Figure 11.3}
28
29   Heavy Precipitation and Pluvial Floods
30
31   The frequency and intensity of heavy precipitation events have likely increased at the global scale over
32   a majority of land regions with good observational coverage. Heavy precipitation has likely increased
33   on the continental scale over three continents: North America, Europe, and Asia. Regional increases in
34   the frequency and/or intensity of heavy precipitation have been observed with at least medium confidence for
35   nearly half of AR6 regions, including WSAF, ESAF, WSB, SAS, ESB, REF, WCA, ECA, TIB, EAS, SEA,
36   NAU, NEU, EEU, GIC, WCE, SES, CNA, and ENA. {11.4, 11.9}
37
38   Human influence, in particular greenhouse gas emissions, is likely the main driver of the observed
39   global scale intensification of heavy precipitation in land regions. It is likely that human-induced climate
40   change has contributed to the observed intensification of heavy precipitation at the continental scale in North
41   America, Europe and Asia. Evidence of a human influence on heavy precipitation has emerged in some
42   regions. {11.4, 11.9, Table 11.1}
43
44   Heavy precipitation will generally become more frequent and more intense with additional global
45   warming. At global warming levels of 4°C relative to the pre-industrial, very rare (e.g., 1 in 10 or more

     1
      See Figure 1.18 in Chapter 1 for definition of AR6 regions. Acronyms for inhabited regions: ARP: Arabian Peninsula;
     CAF: C. Africa; CAR : Caribbean; CAU: C. Australia; CNA: C. North America; EAS: E. Asia; EAU: E. Australia;
     ECA: E. Central Asia; EEU: E. Europe; ENA: E. North America; ESAF: E. Southern Africa; ESB: E. Siberia; GIC:
     Greenland/Iceland; MDG: Madagascar; MED: Mediterranean; NAU: N. Australia; NCA: N. Central America; NEAF:
     N.E. Africa; NEN: N.E. North America; NES: N.E. South America; NEU: N. Europe; NSA: N. South America; NWN:
     N.W. North America; NWS: N.W. South America; NZ: New Zealand; RAR: Russian Arctic; RFE: Russian Far East;
     SAH: Sahara; SAM: South American Monsoon; SAS: South Asia; SAU: Southern Australia; SCA: S. Central America;
     SEAF: S.E. Africa; SES: S.E. South America; SSA: S. South America; SWS: S.W. South America; TIB: Tibetan
     Plateau; WAF: Western Africa; WCA: W. Central Asia; WCE: Western & Central Europe; WNA: W. North America;
     WSAF: W. Southern Africa; WSB: W. Siberia.
     Do Not Cite, Quote or Distribute                         11-7                                       Total pages: 345
     Final Government Distribution                     Chapter11                                   IPCC AR6 WGI

 1   years) heavy precipitation events would become more frequent and more intense than in the recent
 2   past, on the global scale (virtually certain) and in all continents and AR6 regions. The increase in
 3   frequency and intensity is extremely likely for most continents and very likely for most AR6 regions. At
 4   the global scale, the intensification of heavy precipitation will follow the rate of increase in the maximum
 5   amount of moisture that the atmosphere can hold as it warms (high confidence), of about 7% per 1°C of
 6   global warming. The increase in the frequency of heavy precipitation events will accelerate with more
 7   warming and will be higher for rarer events (high confidence), with a likely doubling and tripling in the
 8   frequency of 10-year and 50-year events, respectively, compared to the recent past at 4°C of global warming.
 9   Increases in the intensity of extreme precipitation at regional scales will vary, depending on the amount of
10   regional warming, changes in atmospheric circulation and storm dynamics (high confidence). {11.4, Box
11   11.1}
12
13   The projected increase in the intensity of extreme precipitation translates to an increase in the
14   frequency and magnitude of pluvial floods – surface water and flash floods – (high confidence), as
15   pluvial flooding results from precipitation intensity exceeding the capacity of natural and artificial
16   drainage systems. {11.4}
17
18   River Floods
19
20   Significant trends in peak streamflow have been observed in some regions over the past decades (high
21   confidence). This includes increases in RAR, NSA, and parts of SES, NEU, ENA and
22   decreases in NES, SAU, and parts of MED and EAS). The seasonality of river floods has changed in cold
23   regions where snow-melt is involved, with an earlier occurrence of peak streamflow (high confidence).
24   {11.5}
25
26   Global hydrological models project a larger fraction of land areas to be affected by an increase in
27   river floods than by a decrease in river floods (medium confidence). River floods are projected to become
28   more frequent and intense in some AR6 regions (RAR, SEA, SAS, NWS) (high confidence) and less
29   frequent and intense in others (WCE, EEU, MED) (high confidence). Regional changes in river floods are
30   more uncertain than changes in pluvial floods because complex hydrological processes and forcings,
31   including land cover change and human water management, are involved. {11.5}
32
33   Droughts
34
35   Different drought types exist, and they are associated with different impacts and respond differently to
36   increasing greenhouse gas concentrations. Precipitation deficits and changes in evapotranspiration (ET)
37   govern net water availability. A lack of sufficient soil moisture, sometimes amplified by increased
38   atmospheric evaporative demand (AED), results in agricultural and ecological drought. Lack of runoff and
39   surface water result in hydrological drought. {11.6}
40
41   Human-induced climate change has contributed to decreases in water availability during the dry
42   season over a predominant fraction of the land area due to evapotranspiration increases (medium
43   confidence). Increases in evapotranspiration have been driven by AED increases induced by increased
44   temperature, decreased relative humidity and increased net radiation (high confidence). Trends in
45   precipitation are not a main driver in affecting global-scale trends in drought (medium confidence), but have
46   induced drying trends in a few AR6 regions (NES: high confidence; WAF, CAF, ESAF, SAM, SWS, SSA,
47   SAS: medium confidence). Increasing trends in agricultural and ecological droughts have been observed on
48   all continents (WAF, CAF, WSAF, ESAF, WCA, ECA, EAS, SAU, MED, WCE, WNA, NES: medium
49   confidence), but decreases only in one AR6 region (NAU: medium confidence). Increasing trends in
50   hydrological droughts have been observed in a few AR6 regions (MED: high confidence; WAF, EAS, SAU:
51   medium confidence). Regional-scale attribution shows that human-induced climate change has contributed to
52   increased agricultural and ecological droughts (MED, WNA), and increased hydrological drought (MED) in
53   some regions (medium confidence). {11.6, 11.9}
54
55   The land area affected by increasing drought frequency and severity expands with increasing global
     Do Not Cite, Quote or Distribute                    11-8                                      Total pages: 345
     Final Government Distribution                     Chapter11                                   IPCC AR6 WGI

 1   warming (high confidence). Several regions will be affected by more severe agricultural and ecological
 2   droughts even if global warming is stabilized in a range of 1.5°C-2°C of global warming (high confidence),
 3   including WCE, MED, EAU, SAU, SCA, NSA, SAM, SWS, SSA, NCA, CAN, WSAF, ESAF and MDG
 4   (medium confidence). At 4°C of global warming, about 50% of all inhabited AR6 regions would be affected
 5   (WCE, MED, CAU, EAU, SAU, WCA, EAS, SCA, CAR, NSA, NES, SAM, SWS, SSA, NCA, CAN, ENA,
 6   WNA, WSAF, ESAF, MDG; medium confidence or higher), and only two regions (NEAF, SAS) would
 7   experience decreases in agricultural and ecological drought (medium confidence). There is high confidence
 8   that the projected increases in agricultural and ecological droughts are strongly affected by ET increases
 9   associated with enhanced AED. Several regions are projected to be more strongly affected by hydrological
10   droughts with increasing global warming (at 4°C of global warming: NEU, WCE, EEU, MED, SAU, WCA,
11   SCA, NSA, SAM, SWS, SSA, WNA, WSAF, ESAF, MDG; medium confidence or higher). There is low
12   confidence that effects of enhanced atmospheric CO2 concentrations on plant water-use efficiency alleviate
13   extreme agricultural and ecological droughts in conditions characterized by limited soil moisture
14   and enhanced AED. There is also low confidence that these effects will substantially reduce global plant
15   transpiration and the severity of hydrological droughts. There is high confidence that the land carbon sink
16   will become less efficient due to soil moisture limitations and associated drought conditions in some regions
17   in higher-emission scenarios, in particular under global warming levels above 4°C. {11.6, 11.9, CC-Box 5.1}
18
19   Extreme Storms, Including Tropical Cyclones (TCs)
20
21   The average and maximum rain rates associated with TCs, extratropical cyclones and atmospheric
22   rivers across the globe, and severe convective storms in some regions, increase in a warming world
23   (high confidence). Available event attribution studies of observed strong TCs provide medium confidence
24   for a human contribution to extreme TC rainfall. Peak TC rain rates increase with local warming at least at
25   the rate of mean water vapour increase over oceans (about 7% per 1°C of warming) and in some cases
26   exceeding this rate due to increased low-level moisture convergence caused by increases in TC wind
27   intensity (medium confidence). {11.7, 11.4, Box 11.1}
28
29   It is likely that the global proportion of major TC (Category 3–5) intensities over the past four decades
30   has increased. The average location where TCs reach their peak wind intensity has very likely migrated
31   poleward in the western North Pacific Ocean since the 1940s, and TC translation speed has likely slowed
32   over the conterminous USA since 1900. Evidence of similar trends in other regions is not robust. The global
33   frequency of TC rapid intensification events has likely increased over the past four decades. None of these
34   changes can be explained by natural variability alone (medium confidence).
35
36   The proportion of intense TCs, average peak TC wind speeds, and peak wind speeds of the most
37   intense TCs will increase on the global scale with increasing global warming (high confidence). The
38   total global frequency of TC formation will decrease or remain unchanged with increasing global warming
39   (medium confidence). {11.7.1}
40
41   There is low confidence in past changes of maximum wind speeds and other measures of dynamical
42   intensity of extratropical cyclones. Future wind speed changes are expected to be small, although
43   poleward shifts in the storm tracks could lead to substantial changes in extreme wind speeds in some
44   regions (medium confidence). There is low confidence in past trends in characteristics of severe convective
45   storms, such as hail and severe winds, beyond an increase in precipitation rates. The frequency of springtime
46   severe convective storms is projected to increase in the USA, leading to a lengthening of the severe
47   convective storm season (medium confidence); evidence in other regions is limited. {11.7.2, 11.7.3}.
48
49   Compound Events, Including Dry/Hot events, Fire Weather, Compound Flooding, and Concurrent
50   Extremes
51
52   The probability of compound events has likely increased in the past due to human-induced climate
53   change and will likely continue to increase with further global warming. Concurrent heat waves and
54   droughts have become more frequent and this trend will continue with higher global warming (high
55   confidence). Fire weather conditions (compound hot, dry and windy events) have become more probable in
     Do Not Cite, Quote or Distribute                    11-9                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   some regions (medium confidence) and there is high confidence that they will become more frequent in some
 2   regions at higher levels of global warming. The probability of compound flooding (storm surge, extreme
 3   rainfall and/or river flow) has increased in some locations, and will continue to increase due to both sea level
 4   rise and increases in heavy precipitation, including changes in precipitation intensity associated with TCs
 5   (high confidence). The land area affected by concurrent extremes has increased (high confidence).
 6   Concurrent extreme events at different locations, but possibly affecting similar sectors (e.g., critical crop-
 7   producing areas for global food supply) in different regions, will become more frequent with increasing
 8   global warming, in particular above 2°C of global warming (high confidence). {11.8, Box 11.3, Box 11.4}.
 9
10   Low-Likelihood High-Impact (LLHI) Events Associated With Climate Extremes
11
12   The future occurrence of LLHI events linked to climate extremes is generally associated with low
13   confidence, but cannot be excluded, especially at global warming levels above 4°C. Compound events,
14   including concurrent extremes, are a factor increasing the probability of LLHI events (high confidence).
15   With increasing global warming some compound events with low likelihood in past and current climate will
16   become more frequent, and there is a higher chance of occurrence of historically unprecedented events and
17   surprises (high confidence). However, even extreme events that do not have a particularly low probability in
18   the present climate (at more than 1°C of global warming) can be perceived as surprises because of the pace
19   of global warming (high confidence). {Box 11.2}
20




     Do Not Cite, Quote or Distribute                     11-10                                      Total pages: 345
     Final Government Distribution                          Chapter11                                IPCC AR6 WGI

 1   11.1 Framing
 2
 3   11.1.1 Introduction to the chapter
 4
 5   This chapter provides assessments of changes in weather and climate extremes (collectively referred to as
 6   extremes) framed in terms of the relevance to the Working Group II assessment. It assesses observed
 7   changes in extremes, their attribution to causes, and future projections, at three global warming levels: 1.5°C,
 8   2°C, 4°C. This chapter is also one of the four “regional chapters” of the WGI report (along with Chapters 10
 9   and 12 and the Atlas). Consequently, while it encompasses assessments of changes in extremes at global and
10   continental scales to provide a large-scale context, it also addresses changes in extremes at regional scales.
11
12   Extremes are climatic impact-drivers (Annex VII: Glossary, see Chapter 12 for a comprehensive
13   assessment). The IPCC risk framework (Chapter 1) articulates clearly that the exposure and vulnerability to
14   climatic impact-drivers, such as extremes, modulate the risk of adverse impacts of these drivers, and that
15   adaptation that reduces exposure and vulnerability will increase resilience resulting in a reduction in impacts.
16   Nonetheless, changes in extremes lead to changes in impacts not only as a direct consequence of changes in
17   their magnitude and frequency, but also through their influence on exposure and resilience.
18
19   The Special Report on Managing the Risks of Extreme Events and Disasters to Advance Climate Change
20   Adaptation (referred as the SREX report, IPCC, 2012) provided a comprehensive assessment on changes in
21   extremes and how exposure and vulnerability to extremes determine the impacts and likelihood of disasters.
22   Chapter 3 of that report (Seneviratne et al., 2012, hereafter also referred to as SREX Ch3) assessed physical
23   aspects of extremes, and laid a foundation for the follow-up IPCC assessments. Several chapters of the WGI
24   AR5 (IPCC AR5; IPCC, 2013) addressed climate extremes with respect to observed changes (Hartmann et
25   al., 2013), model evaluation (Flato et al., 2013), attribution (Bindoff et al., 2013), and projected long-term
26   changes (Collins et al., 2013). Assessments were also provided in the recent IPCC Special Reports on 1.5°C
27   global warming (SR15, IPCC, 2018; Hoegh-Guldberg et al., 2018), on climate change and land (IPCC,
28   2019), and on oceans and the cryosphere (IPCC, 2019). These assessments are the starting point of the
29   present assessment.
30
31   This chapter is structured as follows (Figure 11.1). This Section (11.1) provides the general framing and
32   introduction to the chapter, highlighting key aspects that underlie the confidence and uncertainty in the
33   assessment of changes in extremes, and introducing some main elements of the chapter. To provide readers a
34   quick overview of past and future changes in extremes, a synthesis of global scale assessment for different
35   types of extremes is included at the end of this Section (Tables 11.1 and 11.2). Section 11.2 introduces
36   methodological aspects of research on climate extremes. Sections 11.3 to 11.7 assess past changes and their
37   attribution to causes, and projected future changes in extremes, for different types of extremes, including
38   temperature extremes, heavy precipitation and pluvial floods, river floods, droughts, and storms, in separate
39   sections. Section 11.8 addresses compound events. Section 11.9 summarizes regional assessments of changes
40   in temperature extremes, in precipitation extremes and in droughts by continents in tables. The chapter also
41   includes several boxes and FAQs on more specific topics.
42
43
44   [START FIGURE 11.1 HERE]
45
46
47   Figure 11.1: Chapter 11 visual abstract of contents.
48
49
50   [END FIGURE 11.1 HERE]
51
52
53   11.1.2 What are extreme events and how are their changes studied?
54
55   Building on the SREX report and AR5, this Report defines an extreme weather event as “an event that is rare
56   at a particular place and time of year” and an extreme climate event as “a pattern of extreme weather that
     Do Not Cite, Quote or Distribute                        11-11                                   Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   persists for some time, such as a season” (Annex VII: Glossary). The definitions of rare are wide ranging,
 2   depending on applications. Some studies consider an event as an extreme if it is unprecedented; on the other
 3   hand, other studies consider events that occur several times a year as moderate extreme events. Rarity of an
 4   event with a fixed magnitude also changes under human-induced climate change, making events that are
 5   unprecedented so far rather probable under present conditions, but unique in the observational record – and
 6   thus often considered as “surprises” (see Box 11.2).
 7
 8   Various approaches are used to define extremes. These are generally based on the determination of relative
 9   (e.g. 90th percentile) or absolute (e.g. 35°C for a hot day) thresholds above which conditions are considered
10   extremes. Changes in extremes can be examined from two perspectives, either focusing on changes in
11   frequency of given extremes, or on changes in their intensity. These considerations in the definition of
12   extremes are further addressed in Section 11.2.1.
13
14
15   11.1.3 Types of extremes assessed in this chapter
16
17   The types of extremes assessed in this chapter include temperature extremes, heavy precipitation and pluvial
18   floods, river floods, droughts, and storms. The drought assessment addresses meteorological droughts,
19   agricultural and ecological droughts, and hydrological droughts (see Annex VII: Glossary). The storms
20   assessment addresses tropical cyclones, extratropical cyclones, and severe convective storms. In addition,
21   this chapter also assesses changes in compound events, that is, multivariate or concurrent extreme events,
22   because of their relevance to impacts as well as the emergence of new literature on the subject. Most of the
23   considered extremes were also assessed in the SREX and AR5. Compound events were not assessed in depth
24   in past IPCC reports (SREX Ch3; Section 11.8). Marine-related extremes such as marine heat waves and
25   extreme sea level, are assessed in Chapter 9 (Section 9.6.4 and Box 9.2) of this report.
26
27   Extremes and related phenomena are of various spatial and temporal scales. Tornadoes have a spatial scale
28   as small as less than 100 meters and a temporal scale as short as a few minutes. In contrast, a drought can last
29   for multiple years, affecting vast regions. The level of complexity of the involved processes differs from one
30   type of extreme to another, affecting our capability to detect, attribute and project changes in weather and
31   climate extremes. Temperature and precipitation extremes studied in the literature are often based on
32   extremes derived from daily values. Studies of events on longer time scales for both temperature or
33   precipitation, or on sub-daily extremes, are scarcer, which generally limits the assessment for such events.
34   Nevertheless, extremes on time scales different from daily are assessed for temperature extremes and heavy
35   precipitation, when possible (Sections 11.3, 11.4). Droughts, as well as tropical and extratropical cyclones,
36   are assessed as phenomena in general, not limited by their extreme forms, because these phenomena are
37   relevant to impacts (Sections 11.6, 11.7). Both precipitation and wind extremes associated with storms are
38   considered.
39
40   Multiple concomitant extremes can lead to stronger impacts than those resulting from the same extremes had
41   they happened in isolation. For this reason, the occurrence of multiple extremes that are multivariate and/or
42   concurrent and/or happening in succession, also called “compound events” (SREX Ch3), are assessed in this
43   chapter based on emerging literature on this topic (Section 11.8). Box 11.2 also provides an assessment on
44   low-likelihood high-impact scenarios associated with extremes.
45
46   The assessment of projected future changes in extremes is presented as function of different global warming
47   levels (Section 11.2.4 and CC-Box 11.1). On the one hand, this provides traceability and comparison to the
48   SR15 assessment (Hoegh-Guldberg et al., 2018, hereafter referred to as SR15 Ch3). On the other hand, this
49   is useful for decision makers as actionable information, as much of the mitigation policy discussion and
50   adaptation planning can be tied to the level of global warming. For example, regional changes in extremes,
51   and thus their impacts, can be linked to global mitigation efforts. Additionally, there is also an advantage of
52   separating uncertainty in future projections due to regional responses as function of global warming levels
53   from other factors such as differences in global climate sensitivity and emission scenarios (CC-Box 11.1).
54   However, information is also provided on the translation between information provided at global warming
55   levels and for single emissions scenarios (CC-Box 11.1) to facilitate easier comparison with the AR5
     Do Not Cite, Quote or Distribute                     11-12                                      Total pages: 345
     Final Government Distribution                         Chapter11                                        IPCC AR6 WGI

 1   assessment and with some analyses provided in other chapters as function of emissions scenarios.
 2
 3   A global-scale synthesis of this chapter’s assessments is provided in Section 11.1.7. In particular, Tables
 4   11.1 and 11.2 provide a synthesis for observed and attributed changes, and projected changes in extremes,
 5   respectively, at different global warming levels (1.5°C, 2°C, 4°C). Tables on regional-scale assessments for
 6   changes in temperature extremes, heavy precipitation and droughts, are provided in Section 11.9.
 7
 8
 9   11.1.4 Effects of greenhouse gas and other external forcings on extremes
10
11   SREX, AR5, and SR15 assessed that there is evidence from observations that some extremes have changed
12   since the mid 20th century, that some of the changes are a result of anthropogenic influences, and that some
13   observed changes are projected to continue into the future, while other changes are projected to emerge from
14   natural climate variability under enhanced global warming (SREX Chapter 3, AR5 Chapter 10).
15
16   At the global scale but also at the regional scale to some extent, many of the changes in extremes are a direct
17   consequence of enhanced radiative forcing, and the associated global warming and/or resultant increase in
18   the water-holding capacity of the atmosphere, as well as changes in vertical stability and meridional
19   temperature gradients that affect climate dynamics (see Box 11.1). Widespread observed and projected
20   increases in the intensity and frequency of hot extremes, together with decreases in the intensity and
21   frequency of cold extremes, are consistent with global and regional warming (Figure 11.2, Section 11.3).
22   Extreme temperatures on land tend to increase more than the global mean temperature (Figure 11.2), due in
23   large part to the land-sea contrast, and additionally to regional feedbacks in some regions (Section 11.1.6).
24   Increases in the intensity of temperature extremes scale robustly and in general linearly with global warming
25   across different geographical regions in projections up to 2100, with minimal dependence on emissions
26   scenarios (Figures 11.3 and 11.A.1; Seneviratne et al., 2016; Wartenburger et al., 2017; Kharin et al., 2018;
27   Section 11.2.4 and CC-Box 11.1). The frequency of hot temperature extremes (see Figure 11.6), the number
28   of heat wave days and the length of heat wave seasons in various regions also scale well, but non-linearly
29   (because of the threshold effect), with global mean temperatures (Wartenburger et al., 2017; Sun et al.,
30   2018a).
31
32   Changes in annual maximum one-day precipitation (Rx1day) are proportional to mean global surface
33   temperature changes, at about 7% increase per 1°C temperature increase, that is, following the Clausius-
34   Clapeyron relationship (Box 11.1), both in observations (Westra et al., 2013) and in future projections
35   (Kharin et al., 2013) at the global scale. Extreme short-duration precipitation in North America also scales
36   with global surface temperature (Li et al., 2018a; Prein et al., 2016b). At the local and regional scales,
37   changes in extremes are also strongly modulated and controlled by regional forcings and feedback
38   mechanisms (Section 11.1.6), whereby some regional forcings, for example, associated with changes in land
39   cover and land or aerosol emissions, can have non-local or some (non-homogeneous) global-scale effects. In
40   general, there is high confidence in changes in extremes due to global-scale thermodynamic processes (i.e.,
41   global warming, mean moistening of the air) as the processes are well understood, while the confidence in
42   those related to dynamic processes or regional and local forcing, including regional and local thermodynamic
43   processes, is much lower due to multiple factors (see following sub-section and Box 11.1).
44
45
46   [START FIGURE 11.2 HERE]
47
48   Figure 11.2: Time series of observed temperature anomalies for global average annual mean temperature (black), land
49                average annual mean temperature (green), land average annual hottest daily maximum temperature (TXx,
50                purple), and land average annual coldest daily minimum temperature (TNn, blue). Global and land mean
51                temperature anomalies are relative to their 1850-1900 means based on the multi-product mean annual
52                time series assessed in Section 2.3.1.1.3 (see text for references). TXx and TNn anomalies are relative to
53                their respective 1961-1990 means and are based on the HadEX3 dataset (Dunn et al., 2020) using values
54                for grid boxes with at least 90% temporal completeness over 1961-2018. Further details on data sources
55                and processing are available in the chapter data table (Table 11.SM.9).
56
     Do Not Cite, Quote or Distribute                         11-13                                        Total pages: 345
     Final Government Distribution                          Chapter11                                        IPCC AR6 WGI
 1
 2   [END FIGURE 11.2 HERE]
 3
 4
 5   [START FIGURE 11.3 HERE]
 6
 7
 8   Figure 11.3: Regional mean changes in annual hottest daily maximum temperature (TXx) for AR6 land regions and
 9                the global land, against changes in global mean surface air temperature (GSAT) as simulated by CMIP6
10                models under different forcing scenarios SSP1-1.9, SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5. (a)
11                shows individual models from the CMIP6 ensemble (grey), the multi-model median under three selected
12                SSPs (colours), and the multi-model median (black). (b) to (l) show the multi-model-median for the
13                pooled data for individual AR6 regions. Numbers in parantheses indicate the linear scaling between
14                regional TXx and GSAT. The black line indicates the 1:1 reference scaling between TXx and GSAT. See
15                Atlas.1.3.2 for the definition of regions. For details on the methods see Supplementary Material 11.SM.2.
16
17
18   [END FIGURE 11.3 HERE]
19
20
21   Since AR5, the attribution of extreme weather events, or the investigation of changes in the frequency and/or
22   magnitude of individual and local- and regional-scale extreme weather events due to various drivers (see
23   Cross-Working Group Box 1.1 (in Chapter 1) and Section 11.2.3) has provided evidence that greenhouse
24   gases and other external forcings have affected individual extreme weather events. The events that have been
25   studied are geographically uneven. A few events, for example, extreme rainfall events in the UK (Schaller et
26   al., 2016; Vautard et al., 2016; Otto et al., 2018b) or heat waves in Australia (King et al., 2014; Perkins-
27   Kirkpatrick et al., 2016; Lewis et al., 2017b), have spurred more studies than other events. Many highly
28   impactful extreme weather events have not been studied in the event attribution framework. Studies in the
29   developing world are also generally lacking. This is due to various reasons (Section 11.2) including lack of
30   observational data, lack of reliable climate models, and lack of scientific capacity (Otto et al., 2020). While
31   the events that have been studied are not representative of all extreme events that occurred and results from
32   these studies may also be subject to selection bias, the large number of event attribution studies provide
33   evidence that changes in the properties of these local and individual events are in line with expected
34   consequences of human influence on the climate and can be attributed to external drivers (Section 11.9).
35   Figure 11.4 summarizes assessments of observed changes in temperature extremes, in heavy precipitation
36   and in droughts, and their attribution in a map form.
37
38
39   [START FIGURE 11.4 HERE]
40
41   Figure 11.4: Overview of observed changes for cold, hot, and wet extremes and their potential human
42                contribution. Shown are the direction of change and the confidence in 1) the observed changes in how
43                cold and hot as well as wet extremes have already changed across the world and 2) in the contribution of
44                whether human-induced climate change contributed in causing to these changes (attribution). In each
45                region changes in extremes are indicated by colour (orange – increase in the type of extreme, blue –
46                decrease, both colours – there are changes of opposing direction within the region the signal depends on
47                the exact event definition, grey – there are no changes observed, and no fill – the data/evidence is too
48                sparse to make an assessment). The squares and dots next to the symbol indicate the level of confidence
49                for observing the trend and the human contribution, respectively. The more black dots/squares the higher
50                the level of confidence. The information on this figure is based on regional assessment of the literature on
51                observed trends, detection and attribution and event attribution in section 11.9.
52
53   [END FIGURE 11.4 HERE]
54
55
56
57
58
59
     Do Not Cite, Quote or Distribute                         11-14                                          Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   [START BOX 11.1 HERE]
 2
 3
 4   BOX 11.1: Thermodynamic and dynamic changes in extremes across scales
 5
 6   Changes in weather and climate extremes are determined by local exchanges in heat, moisture, and other
 7   related quantities (thermodynamic changes) and those associated with atmospheric and oceanic motions
 8   (dynamic changes). While thermodynamic and dynamic processes are interconnected, considering them
 9   separately helps to disentangle the roles of different processes contributing to changes in climate extremes
10   (e.g. Shepherd, 2014).
11
12   Temperature extremes
13   An increase in the concentration of greenhouse gases in the atmosphere leads to the warming of tropospheric
14   air and the Earth’s surface. This direct thermodynamic effect leads to warmer temperatures everywhere with
15   an increase in the frequency and intensity of warm extremes and a decrease in the frequency and intensity of
16   cold extremes. The initial increase in temperature in turn leads to other thermodynamic responses and
17   feedbacks affecting both the atmosphere and the surface. These include an increase in the water vapour
18   content of the atmosphere (water vapour feedback, see Section 7.4.2.2) and a change in the vertical profile of
19   temperature (e.g., lapse rate feedback, see Section 7.4.2.2). While the water vapour feedback always
20   amplifies the initial temperature increases (positive feedback), the lapse rate feedback amplifies near-surface
21   temperature increases (positive feedback) in mid- and high latitudes but reduces temperature increases
22   (negative feedback) in tropical regions (Pithan and Mauritsen, 2014).
23
24   Thermodynamic responses and feedbacks also occur through surface processes. For instance, observations
25   and model simulations show that temperature increases, including extreme temperatures, are amplified in
26   areas where seasonal snow cover is reduced due to decreases in surface albedo (see Section 11.3.1). In some
27   mid-latitude areas, temperature increases are amplified by the higher atmospheric evaporative demand (Fu
28   and Feng, 2014; Vicente-Serrano et al., 2020b) that results in a drying of soils in some regions (Section
29   11.6), leading to increased sensible heat fluxes (soil-moisture temperature feedback, see Sections 11.1.6 and
30   11.3.1). Other thermodynamic feedback processes include changes in the water-use efficiency of plants
31   under enhanced atmospheric CO2 concentrations that can reduce the overall transpiration, and thus also
32   enhance temperature in projections (Sections 8.2.3.3, 11.1.6, 11.3, and 11.6).
33
34   Changes in the spatial distribution of temperatures can also affect temperature extremes by modifying the
35   characteristics of weather patterns (e.g., Suarez-Gutierrez et al., 2020). For example, a robust thermodynamic
36   effect of polar amplification is a weakened north-south temperature gradient, which amplifies the warming
37   of cold extremes in the Northern Hemisphere mid- and high latitudes because of the reduction of cold air
38   advection (Holmes et al., 2015; Schneider et al., 2015; Gross et al., 2020). Much less robust is the dynamic
39   effect of polar amplification (Section 7.4.4.1) and the reduced low-altitude meridional temperature gradient
40   that has been linked to an increase in the persistence of weather patterns (e.g., heatwaves) and subsequent
41   increases in temperature extremes (Francis and Vavrus, 2012; Coumou et al., 2015, 2018; Mann et al., 2017)
42   (CC-Box 10.1).
43
44   Precipitation extremes
45   Changes in temperature also control changes in water vapour through increases in evaporation and in the
46   water-holding capacity of the atmosphere (Section 8.2.1). At the global scale, column-integrated water
47   vapour content increases roughly following the Clausius-Clapeyron (C-C) relation, with an increase of
48   approximately 7% for every degree celsius of global-mean surface warming (Section 8.2.1). Nonetheless, at
49   regional scales, water vapour increases differ from this C-C rate due to several reasons (Section 8.2.2),
50   including a change in weather regimes and limitations in moisture transport from the ocean, which warms
51   more slowly than land (Byrne and O’Gorman, 2018). Observational studies (Fischer and Knutti, 2016; Sun et
52   al., 2020) have shown the observed rate of increase of precipitation extremes is similar to the C-C scaling at
53   the global scale. Climate model projections show that the increase in water vapour leads to robust increases
54   in precipitation extremes everywhere, with a magnitude that varies between 4% and 8% per degree celsius of
55   surface warming (thermodynamic contribution, Box 11.1, Figure 1b). At regional scales, climate models
56   show that the dynamic contribution (Box 11.1, Figure 1c) can be substantial and strongly modify the
     Do Not Cite, Quote or Distribute                     11-15                                     Total pages: 345
     Final Government Distribution                          Chapter11                                         IPCC AR6 WGI

 1   projected rate of change of extreme precipitation (Box 11.1, Figure 1a) with large regions in the subtropics
 2   showing robust reductions and other areas (e.g., equatorial Pacific) showing robust amplifications (Box 11.1,
 3   Figure 1c). However, the dynamic contributions show large differences across models and are more
 4   uncertain than thermodynamic contributions (Shepherd, 2014; Trenberth et al., 2015; Pfahl et al., 2017; Box
 5   11.1, Figure 1c).
 6
 7   Dynamic contributions can occur in response to changes in the vertical and horizontal distribution of
 8   temperature (thermodynamics) and can affect the frequency and intensity of synoptic and subsynoptic
 9   phenomena including tropical cyclones, extratropical cyclones, fronts, mesoscale-convective systems and
10   thunderstorms. For example, the poleward shift and strengthening of the Southern Hemisphere mid-latitude
11   storm tracks (Section 4.5.1) can modify the frequency/intensity of extreme precipitation. However, the
12   precise way in which dynamic changes will affect precipitation extremes is unclear due to several competing
13   effects (Shaw et al., 2016; Allan et al., 2020).
14
15   Extreme precipitation can also be enhanced by dynamic responses and feedbacks occurring within storms
16   that result from the extra latent heat released from the thermodynamic increases in moisture (Lackmann,
17   2013; Willison et al., 2013; Marciano et al., 2015; Nie et al., 2018; Mizuta and Endo, 2020). The extra latent
18   heat released within storms has been shown to increase precipitation extremes by strengthening convective
19   updrafts and the intensity of the cyclonic circulation (e.g., Molnar et al., 2015; Nie et al., 2018), although
20   weakening effects have also been found in mid-latitude cyclones (e.g., Kirshbaum et al., 2017). Additionally,
21   the increase in latent heat can also suppress convection at larger scales due to atmospheric stabilization (Nie
22   et al., 2018; Tandon et al., 2018; Kendon et al., 2019). As these dynamic effects result from feedback
23   processes within storms where convective processes are crucial, their proper representation might require
24   improving the horizontal/vertical resolution, the formulation of parameterizations, or both, in current climate
25   models (i.e., Ban et al., 2015; Kendon et al., 2014; Meredith et al., 2015; Nie et al., 2018; Prein et al., 2015;
26   Westra et al., 2014).
27
28
29   [START BOX 11.1, FIGURE 1 HERE]
30
31   Box 11.1, Figure 1: Multi-model (CMIP5) mean fractional changes (in % per degree of warming) for (a) annual
32                       maximum precipitation (Rx1day), (b) changes in Rx1day due to the thermodynamic contribution
33                       and (c) changes in Rx1day due to the dynamic contribution estimated as the difference between
34                       the total changes and the thermodynamic contribution. Changes were derived from a linear
35                       regression for the period 1950–2100. Uncertainty is represented using the simple approach: no
36                       overlay indicates regions with high model agreement, where ≥80% of models (n=22) agree on sign
37                       of change; diagonal lines indicate regions with low model agreement, where <80% of models
38                       agree on sign of change. For more information on the simple approach, please refer to the Cross-
39                       Chapter Box Atlas 1. A detailed description of the estimation of dynamic and thermodynamic
40                       contributions is given in Pfahl et al. (2017). Adapted from (Pfahl et al., 2017), originally published
41                       in Nature Climate Change/ Springer Nature. Further details on data sources and processing are
42                       available in the chapter data table (Table 11.SM.9).
43
44   [END BOX 11.1, FIGURE 1 HERE]
45
46
47   Droughts
48   Droughts are also affected by both thermodynamic and dynamic processes (Sections 8.2.3.3 and 11.6).
49   Thermodynamic processes affect droughts by increasing atmospheric evaporative demand (Martin, 2018;
50   Gebremeskel Haile et al., 2020; Vicente-Serrano et al., 2020b) through changes in air temperature, radiation,
51   wind speed, and relative humidity. Dynamic processes affect droughts through changes in the occurrence,
52   duration and intensity of weather anomalies, which are related to precipitation and the amount of sunlight
53   (Section 11.6). While atmospheric evaporative demand increases with warming, regional changes in aridity
54   are affected by increasing land-ocean warming contrast, vegetation feedbacks and responses to rising CO2
55   concentrations and dynamic shifts in the location of the wet and dry parts of the atmospheric circulation in
56   response to climate change as well as internal variability (Byrne and O’Gorman, 2015; Kumar et al., 2015;
     Do Not Cite, Quote or Distribute                          11-16                                          Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   Allan et al., 2020).
 2
 3   In summary, both thermodynamic and dynamic processes are involved in the changes of extremes in
 4   response to warming. Anthropogenic forcing (e.g., increases in greenhouse gas concentrations) directly
 5   affects thermodynamic variables, including overall increases in high temperatures and atmospheric
 6   evaporative demand, and regional changes in atmospheric moisture, which intensify heatwaves, droughts and
 7   heavy precipitation events when they occur (high confidence). Dynamic processes are often indirect
 8   responses to thermodynamic changes, are strongly affected by internal climate variability and are also less
 9   well understood. As such, there is low confidence in how dynamic changes affect the location and magnitude
10   of extreme events in a warming climate.
11
12   [END BOX 11.1 HERE]
13
14
15   11.1.5 Effects of large-scale circulation on changes in extremes
16
17   Atmospheric large-scale circulation patterns and associated atmospheric dynamics are important
18   determinants of the regional climate (Chapter 10). As a result, they are also important to the magnitude,
19   frequency, and duration of extremes (Box 11.4). Aspects of changes in large-scale circulation patterns are
20   assessed in Chapters 2, 3, 4, and 8 and representative atmospheric and oceanic modes are described in Annex
21   IV. This subsection provides some general concepts, through a couple of examples, on why the uncertainty
22   in the response of large-scale circulation patterns to external forcing can cascade to uncertainty in the
23   response of extremes to external forcings. Details for specific types of extremes are covered in the relevant
24   subsections. For example, the occurrence of the El Niño-Southern Oscillation (ENSO) influences
25   precipitation regimes in many areas, favoring droughts in some regions and heavy rains in others (Box 11.4).
26   The extent and strength of the Hadley circulation influences regions where tropical and extra-tropical
27   cyclones occur, with important consequences for the characteristics of extreme precipitation, drought, and
28   winds (Section 11.7). Changes in circulation patterns associated with land-ocean heat contrast, which affect
29   the monsoon circulations (Section 8.4.2.4), lead to heavy precipitation along the coastal regions in East Asia
30   (Freychet et al., 2015). As a result, changes in the spatial and/or temporal variability of the atmospheric
31   circulation in response to warming affect characteristics of weather systems such as tropical cyclones
32   (Sharmila and Walsh, 2018), storm tracks (Shaw et al., 2016), and atmospheric rivers (Waliser and Guan,
33   2017) (e.g. Section 11.7). Changes in weather systems come with changes in the frequency and intensity of
34   extreme winds, extreme temperatures, and extreme precipitation, on the backdrop of thermodynamic
35   responses of extremes to warming (Box 11.1). Floods are also affected by large-scale circulation modes,
36   including ENSO, the North Atlantic Oscillation (NAO), the Atlantic Multi-decadal Variability (AMV), and
37   the Pacific Decadal Variability (PDV) (Kundzewicz et al., 2018; Annex IV). Aerosol forcing, through
38   changes in patterns of sea surface temperatures (SSTs), also affects circulation patterns and tropical cyclone
39   activities (Takahashi et al., 2017).
40
41   Changes in atmospheric large-scale circulation due to external forcing are uncertain in general, but there are
42   clear signals in some aspects (Chapter 2, 3, 4, and 8; Sections 2.3.1.4, 8.2.2.2). Among them, there has been
43   a very likely widening of the Hadley circulation since the 1980s and the extratropical jets and cyclone tracks
44   have likely been shifting poleward since the 1980s (Section 2.3.1.4). The poleward expansion affects drought
45   occurrence in some regions (Section 11.6), and results in poleward shifts of tropical cyclones and storm
46   tracks (Sections 11.7.1, 11.7.2). Although it is very likely that the amplitude of ENSO variability will not
47   robustly change over the 21st century (Section 4.3.3.2), the frequency of extreme El Niños (Box 11.4),
48   defined by precipitation threshold, is projected to increase with global warming (Section 6.5 of SROCC).
49   This would have implications for projected changes in extreme events affected by ENSO, including droughts
50   over wide areas (Section 11.6; Box 11.4) and tropical cyclones (Section 11.7.1). A case study is provided for
51   extreme ENSOs in 2015/2016 in Box 11.4 to highlight the influence of ENSO on extremes.
52
53   In summary, large-scale atmospheric circulation patterns are important drivers for local and regional
54   extremes. There is overall low confidence about future changes in the magnitude, frequency, and spatial
55   distribution of these patterns, which contributes to uncertainty in projected responses of extremes, especially
     Do Not Cite, Quote or Distribute                     11-17                                      Total pages: 345
     Final Government Distribution                     Chapter11                                    IPCC AR6 WGI

 1   in the near term.
 2
 3
 4   11.1.6 Effects of regional-scale processes and forcings and feedbacks on changes in extremes
 5
 6   At the local and regional scales, changes in extremes are strongly modulated by regional and local feedbacks
 7   (SRCCL, Jia et al., 2019; Seneviratne et al., 2013; Miralles et al., 2014; Lorenz et al., 2016; Vogel et al.,
 8   2017), changes in large-scale circulation patterns (11.1.5), and regional forcings such as changes in land use
 9   or aerosol concentrations (Chapters 3 and 7; Hirsch et al., 2017, 2018; Thiery et al., 2017; Wang et al.,
10   2017f; Findell et al., 2017). In some cases, such responses may also include non-local effects (e.g., Persad
11   and Caldeira, 2018; Miralles et al., 2019; de Vrese et al., 2016; Schumacher et al., 2019). Regional-scale
12   forcing and feedbacks often affect temperature distributions asymmetrically, with generally higher effects for
13   the hottest percentiles (Section 11.3).
14
15   Land use can affect regional extremes, in particular hot extremes, in several ways (high confidence). This
16   includes effects of land management (e.g. cropland intensification, irrigation, double cropping) and well as
17   of land cover changes (deforestation) (Section 11.3.2; see also 11.6). Some of these processes are not well
18   represented (e.g. effects of forest cover on diurnal temperature cycle) or not integrated (e.g. irrigation) in
19   climate models (Sections 11.3.2, 11.3.3). Overall, the effects of land use forcing may be particularly relevant
20   in the context of low-emissions scenarios, which include large land use modifications, for instance associated
21   with the expansion of biofuels, biofuels with carbon capture and storage (BECCS), or re-afforestation to
22   ensure negative emissions, as well as with the expansion of food production (e.g. SR15, Chapter 3; CC-Box
23   5.1; van Vuuren et al., 2011, Hirsch et al., 2018). There are also effects on the water cycle through
24   freshwater use (CC-Box 5.1; Section 11.6).
25
26   Aerosol forcing also has a strong regional footprint associated with regional emissions, which affects
27   temperature and precipitation extremes (high confidence; Sections 11.3, 11.4). From ca. the 1950s to 1980s,
28   enhanced aerosol loadings led to regional cooling due to decreased global solar radiation (“global dimming”)
29   which was followed by a phase of “global brightening” due to a reduction in aerosol loadings (Chapters 3
30   and 7; Wild et al., 2005). King et al. (2016a) show that aerosol-induced cooling delayed the timing of a
31   significant human contribution to record-breaking heat extremes in some regions. On the other hand, the
32   decreased aerosol loading since the 1990s has led to an accelerated warming of hot extremes in some
33   regions. Based on Earth System Model (ESM) simulations, Dong et al. (2017b) suggest that a substantial
34   fraction of the warming of the annual hottest days in Western Europe since the mid-1990s has been due to
35   decreases in aerosol concentrations in the region. Dong et al. (2016) also identify non-local effects of
36   decreases in aerosol concentrations in Western Europe, which they estimate played a dominant role in the
37   warming of the hottest daytime temperatures in Northeast Asia since the mid-1990s, via induced coupled
38   atmosphere-land surface and cloud feedbacks, rather than a direct impact of anthropogenic aerosol changes
39   on cloud condensation nuclei.
40
41   In addition to regional forcings, regional feedback mechanisms can also substantially affect extremes (high
42   confidence; Sections 11.3, 11.4, 11.6). In particular, soil moisture feedbacks play an important role for
43   extremes in several mid-latitude regions, leading in particular to a marked additional warming of hot
44   extremes compared to mean global warming (Seneviratne et al., 2016; Bathiany et al., 2018; Miralles et al.,
45   2019), which is superimposed on the known land-sea contrast in mean warming (Vogel et al., 2017). Soil
46   moisture-atmosphere feedbacks also affect drought development (Section 11.6). Additionally, effects of land
47   surface conditions on circulation patterns have also been reported (Koster et al., 2016; Sato and Nakamura,
48   2019). These regional feedbacks are also associated with substantial spread in models (Section 11.3), and
49   contribute to the identified higher spread of regional projections of temperature extremes as function of
50   global warming, compared with the spread resulting from the differences in projected global warming
51   (global transient climate responses) in climate models (Seneviratne and Hauser, 2020). In addition, there are
52   also feedbacks between soil moisture content and precipitation occurrence, generally characterized by
53   negative spatial feedbacks and positive local feedbacks (Taylor et al., 2012; Guillod et al., 2015). Climate
54   model projections suggest that these feedbacks are relevant for projected changes in heavy precipitation
55   (Seneviratne et al., 2013), however, there is evidence that climate models do not capture the correct sign of
     Do Not Cite, Quote or Distribute                    11-18                                      Total pages: 345
     Final Government Distribution                        Chapter11                                      IPCC AR6 WGI

 1   the soil moisture-precipitation feedbacks in several regions, in particular spatially and/or in some cases also
 2   temporally (Taylor et al., 2012; Moon et al., 2019). In the Northern Hemisphere high latitudes, the snow- and
 3   ice-albedo feedback, along with other factors, is projected to largely amplify temperature increases (e.g.,
 4   Pithan and Mauritsen, 2014), although the effect on temperature extremes is still unclear. It is also still
 5   unclear whether snow-albedo feedbacks in mountainous regions might have an effect on temperature and
 6   precipitation extremes (e.g. Gobiet et al., 2014), however these feedbacks play an important role in projected
 7   changes in high-latitude warming (Hall and Qu, 2006), and, in particular, in changes in cold extremes in
 8   these regions (Section 11.3).
 9
10   Finally, extreme events may also regionally amplify one another. This is, e.g., the case for heat waves and
11   droughts, with high temperatures and stronger radiative forcing leading to drying tendencies on land due to
12   increased evapotranspiration (Section 11.6), and drier soils then inducing decreased evapotranspiration and
13   higher sensible heat flux and hot temperatures (Seneviratne et al., 2013; Miralles et al., 2014; Vogel et al.,
14   2017; Zscheischler and Seneviratne, 2017; Zhou et al., 2019b; Kong et al., 2020; see Box 11.1, Section
15   11.8).
16
17   In summary, regional forcings and feedbacks, in particular associated with land use and aerosol forcings, and
18   soil moisture-temperature, soil moisture-precipitation, and snow/ice-albedo-temperature feedbacks, play an
19   important role in modulating regional changes in extremes. These can also lead to a higher warming of
20   extreme temperatures compared to mean temperature (high confidence), and possibly cooling in some
21   regions (medium confidence). However, there is only medium confidence in the representation of the
22   associated processes in state-of-the-art Earth System Models.
23
24
25   11.1.7 Global-scale synthesis
26
27   Tables 11.1 and 11.2 provide a synthesis for observed and attributed changes in extremes, and projected
28   changes in extremes, respectively, at different levels of global warming. This synthesis assessment focuses
29   on the more likely range of observed and projected changes. However, some low-likelihood high-impact
30   scenarios can also be of high relevance as addressed in Box 11.2.
31
32   Figure 11.5 provides a synthesis on the level of confidence in the attribution and projection of changes in
33   extremes, building on the assessments from Tables 11.1 and 11.2. In the case where the physical processes
34   underlying the changes in extremes in response to human forcing are well understood and the signal in the
35   observations is still relatively weak, confidence in the projections would be higher than in the attribution
36   because of an increase in the signal to noise ratio with higher global warming. On the other hand, when the
37   observed signal is already strong and when observational evidence is consistent with model simulated
38   responses, confidence in attribution may be higher than that in projections if certain physical processes could
39   be expected to behave differently in a much warmer world and under much higher greenhouse gas forcing,
40   and if such a behavior is poorly understood.
41
42   Further synthesis figures for regional assessments are provided in Figure 11.4 (event attribution), Figure 11.6
43   (projected change in hot temperature extremes) and Figure 11.7 (projected changes in precipitation
44   extremes), and a synthesis on regional assessments for observed, attributed and projected changes in
45   extremes is provided in Section 11.9 for all AR6 reference regions (See Chapter 1, section 1.4.5 and Figure
46   1.18 for definition of AR6 regions).
47
48   Confidence and likelihood of past changes and projected future changes at 2°C of global warming lon the
49   global scale. The information in this figure is based on Tables 11.1 and 11.2.
50   [START FIGURE 11.5 HERE]
51
52
53   Figure 11.5: Confidence and likelihood of past changes and projected future changes at 2°C of global warming on the
54                global scale. The information in this figure is based on Tables 11.1 and 11.2.
55
56
     Do Not Cite, Quote or Distribute                       11-19                                        Total pages: 345
     Final Government Distribution                               Chapter11                                          IPCC AR6 WGI

 1   [END FIGURE 11.5 HERE]
 2
 3
 4   [START FIGURE 11.6 HERE]
 5
 6
 7   Figure 11.6: Projected changes in the frequency of extreme temperature events under 1°C, 1.5°C, 2°C, 3°C, and 4°C
 8                global warming levels relative to the 1851-1900 baseline. Extreme temperatures are defined as the
 9                maximum daily temperatures that were exceeded on average once during a 10-year period (10-year event,
10                blue) and once during a 50-year period (50-year event, orange) during the 1851-1900 base period. Results
11                are shown for the global land and the AR6 regions. For each box plot, the horizontal line and the box
12                represent the median and central 66% uncertainty range, respectively, of the frequency changes across the
13                multi model ensemble, and the whiskers extend to the 90% uncertainty range. The dotted line indicates no
14                change in frequency. The results are based on the multi-model ensemble from simulations of global
15                climate models contributing to the sixth phase of the Coupled Model Intercomparison Project (CMIP6)
16                under different SSP forcing scenarios. Adapted from (Li et al., 2020a). Further details on data sources and
17                processing are available in the chapter data table (Table 11.SM.9).
18
19
20   [END FIGURE 11.6 HERE]
21
22
23   [START FIGURE 11.7 HERE]
24
25
26   Figure 11.7: Projected changes in the frequency of extreme precipitation events under 1°C, 1.5°C, 2°C, 3°C, and 4°C
27                global warming levels relative to the 1951-1990 baseline. Extreme precipitation is defined as the
28                maximum daily precipitation (Rx1day) that was exceeded on average once during a 10-year period (10-
29                year event, blue) and once during a 50-year period (50-year event, orange) during the 1851-1900 base
30                period. Results are shown for the global land and the AR6 regions. For each box plot, the horizontal line
31                and the box represent the median and central 66% uncertainty range, respectively, of the frequency
32                changes across the multi model ensemble, and the whiskers extend to the 90% uncertainty range. The
33                dotted line indicates no change in frequency. The results are based on the multi-model ensemble from
34                simulations of global climate models contributing to the sixth phase of the Coupled Model
35                Intercomparison Project (CMIP6) under different SSP forcing scenarios. Adapted from (Li et al., 2020a).
36                Further details on data sources and processing are available in the chapter data table (Table 11.SM.9).
37
38
39   [END FIGURE 11.7 HERE]
40
41
42   [START TABLE 11.1 HERE]
43
44
45   Table 11.1: Synthesis table on observed changes in extremes and contribution by human influences. Note that
46               observed changes in marine extremes are assessed in the Cross-Chapter Box 9.1 in Chapter 9.
47
      Phenomenon and direction of      Observed/detected trends since 1950 (for    Human contribution to the observed trends since
      trend                            +0.5°C global warming or higher)            1950 (for +0.5°C global warming or higher)

      Warmer and/or more frequent      Virtually certain on global scale {11.3}    Extremely likely main contributor on global scale
      hot days and nights over most                                                {11.3}
      land areas                       Continental-scale evidence:
      Warmer and/or fewer cold         Asia, Australasia, Europe, North America:   Continental-scale evidence:
                                       Very likely                                 North America, Europe, Australasia, Asia: Very
      days and nights over most
      land areas                       Central and South America: High             likely
                                       confidence                                  Central and South America: High confidence
      Warm spells/heat waves;          Africa: Medium confidence                   Africa: Medium confidence
      Increases in frequency or        {11.3, 11.9}
      intensity over most land areas                                               {11.3, 11.9}

      Cold spells/cold waves:
      Decreases in frequency or

     Do Not Cite, Quote or Distribute                                11-20                                          Total pages: 345
Final Government Distribution                                Chapter11                                             IPCC AR6 WGI

 intensity over most land areas

 Heavy precipitation events:      Likely on global scale, over majority of      Likely main contributor to the observed
 increase in the frequency,       land regions with good observational          intensification of heavy precipitation in land
 intensity, and/or amount of      coverage {11.3}                               regions on global scale.
 heavy precipitation                                                            {11.3}
                                  Continental-scale evidence:
                                                                                Continental-scale evidence:
                                  Asia, Europe, North America: Likely
                                                                                Asia, Europe, North America: Likely
                                  Africa, Australasia, Central and South
                                  America: Low confidence                       Africa, Australasia, Central and South America:
                                                                                Low confidence
                                  {11.3, 11.9}
                                                                                {11.3, 11.9}
 Agricultural and ecological      Medium confidence, in predominant             Medium confidence, in predominant fraction of
 drought events: Enhanced         fraction of land area                         land area
 drying in dry season             Observed decrease in water availability in    Human contribution to decrease in water
                                  the dry season due to increased               availability in the dry season in a predominant
                                  evapotranspiration (driven by increased       fraction of the land area (medium confidence)
                                  atmospheric evaporative demand) in a          {11.6}
                                  predominant fraction of the land area
                                  (medium confidence) {11.6}
                                  Increasing trends in agricultural and
                                  ecological droughts have been observed in
                                  AR6 regions on all continents (medium
                                  confidence) {11.6, 11.9}
 Increase in precipitation        Medium confidence                             High confidence
 associated with tropical         {11.7}                                        {11.7}
 cyclones
 Increase in likelihood that a    Likely                                        Medium confidence
 TC will be at major TC           {11.7}                                        {11.7}
 intensity (Cat. 3-5)
 Changes in frequency of          Likely                                        Medium confidence
 rapidly intensifying tropical    {11.7}                                        {11.7}
 cyclones

 Poleward migration of            Medium confidence                             Medium confidence
 tropical cyclones in the         {11.7}                                        {11.7}
 western Pacific

 Decrease in TC forward           It is likely that TC translation speed has    It is more likely than not that the slowdown of TC
 motion over the USA              slowed over the USA since 1900.               translation speed over the USA has contributions
                                  {11.7}                                        from anthropogenic forcing.
                                                                                {11.7}
 Severe convective storms         Low confidence in past trends in hail and     Low confidence.
 (tornadoes, hail, rainfall,      winds and tornado activity due to short       {11.7}
 wind, lightning)                 length of high-quality data records. {11.7}

 Increase in compound events      Likely increase in the probability of         Likely that human-induced climate change has
                                  compound events.                              increased the probability of compound events.

                                  High confidence that co-occurrent heat        High confidence that human influence has
                                  waves and droughts are becoming more          increased the frequency of co-occurrent heat
                                  frequent under enhanced greenhouse gas        waves and droughts.
                                  forcing at global scale.                      Medium confidence that human influence has
                                  Medium confidence that fire weather, i.e.     increased fire weather occurrence in some regions.
                                  compound hot, dry and windy events, have      Low confidence that human influences has
                                  become more frequent in some regions.         contributed to changes in compound events
                                  Medium confidence that compound               leading to flooding.
                                  flooding risk has increased along the USA     {11.8}
                                  coastline.
                                  {11.8}
Do Not Cite, Quote or Distribute                                11-21                                             Total pages: 345
     Final Government Distribution                               Chapter11                                               IPCC AR6 WGI



 1
 2   [END TABLE 11.1 HERE]
 3
 4
 5   [START TABLE 11.2 HERE]
 6
 7   Table 11.2: Synthesis table on projected changes in extremes. Note that projected changes in marine extremes are
 8               assessed in Chapter 9 and the Cross-chapter box 9.1 (marine heat waves). Assessments are provided
 9               compared to pre-industrial conditions.
10
      Phenomenon and        Projected changes at +1.5°C          Projected changes at +2°C global      Projected changes at +4°C global
      direction of trend    global warming                       warming                               warming

      Warmer and/or         Virtually certain compared to        Virtually certain compared to         Virtually certain compared to pre-
      more frequent hot     pre-industrial on global scale.      pre-industrial on global scale.       industrial on global scale.
      days and nights
      over most land        Extremely likely on all continents   Virtually certain on all              Virtually certain on all continents
      areas                                                      continents
                            Highest increase of temperature                                            Highest increase of temperature of
      Warmer and/or         of hottest days is projected in      Highest increase of temperature       hottest days is projected in some
      fewer cold days       some mid-latitude and semi-arid      of hottest days is projected in       mid-latitude and semi-arid
      and nights over       regions, at about 1.5 times to       some mid-latitude and semi-arid       regions, at about 1.5 times to
      most land areas       twice the rate of global warming     regions, at about 1.5 times to        twice the rate of global warming
                            (high confidence)                    twice the rate of global warming      (high confidence)
      Warm spells/heat                                           (high confidence)
      waves; Increases in   Highest increase of temperature                                            Highest increase of temperature of
      frequency or          of coldest days is projected in      Highest increase of temperature       coldest days is projected in Arctic
      intensity over most   Arctic regions, at about three       of coldest days is projected in       regions, at about three times the
      land areas            times the rate of global warming     Arctic regions, at about three        rate of global warming (high
                            (high confidence)                    times the rate of global warming      confidence)
      Cold spells/cold      {11.3}                               (high confidence)                     {11.3}
      waves: Decreases                                           {11.3}
      in frequency or       Continental-scale projections:                                             Continental-scale projections:
      intensity over most   Extremely likely: Africa, Asia,      Continental-scale projections:        Virtually certain: Africa, Asia,
      land areas            Australasia, Central and South       Virtually certain: Africa, Asia,      Australasia, Central and South
                            America, Europe, North America       Australasia, Central and South        America, Europe, North America
                            {11.3, 11.9}}                        America, Europe, North America        {11.3, 11.9}
                                                                 {11.3, 11.9}
      Heavy                 High confidence that increases       Likely that increases take place in   Very likely that increases take
      precipitation         take place in most land regions      most land regions                     place in most land regions
      events: increase in   {11.4}                               {11.4}                                {11.4}
      the frequency,
      intensity, and/or     Very likely: Asia, N. America        Extremely likely: Asia, N.            Virtually certain: Africa, Asia, N.
      amount of heavy       Likely: Africa, Europe               America                               America
      precipitation         High confidence: Central and         Very likely: Africa, Europe           Extremely likely: Central and
                            South America                        Likely: Australasia, Central and      South America, Europe
                            Medium confidence: Australasia       South America                         Very likely Australasia
                            {11.4, 11.9}                         {11.4, 11.9}                          {11.4, 11.9}




     Do Not Cite, Quote or Distribute                               11-22                                                Total pages: 345
     Final Government Distribution                                Chapter11                                               IPCC AR6 WGI

      Agricultural and      High confidence over                  Likely over predominant fraction      Very likely over predominant
      ecological            predominant fraction of land area     of land area                          fraction of land area
      droughts: Increases
      in intensity and/or   Land area affected by increasing      Land area affected by increasing      Land area affected by increasing
      duration of drought   drought frequency and severity        drought frequency and severity        drought frequency and severity
      events                expands with increasing global        expands with increasing global        expands with increasing global
                            warming (high confidence).            warming (likely). {11.6, 11.9}        warming (very likely). {11.6,
                            {11.6, 11.9}                                                                11.9}
                                                                  Precipitation decreases is going
                            Precipitation decreases is going      to increase the severity of           Precipitation decreases is going to
                            to increase the severity of           drought in some regions;              increase the severity of drought in
                            drought in some regions;              atmospheric evaporative demand        several regions; atmospheric
                            atmospheric evaporative demand        will continue to increase             evaporative demand will continue
                            will continue to increase             compared to pre-industrial            to increase compared to pre-
                            compared to pre-industrial            conditions and lead to further        industrial conditions and lead to
                            conditions and lead to further        increases in agricultural and         further increases in agricultural
                            increases in agricultural and         ecological droughts due to            and ecological droughts due to
                            ecological droughts due to            increased evapotranspiration in       increased evapotranspiration in
                            increased evapotranspiration in       some regions. (high confidence)       several regions. (high confidence)
                            some regions. (high confidence)       {11.6, 11.9}                          {11.6, 11.9}
                            {11.6, 11.9}

      Increase in           High confidence in a projected        High confidence in a projected        High confidence in a projected
      precipitation         increase of TC rain rates at the      increase of TC rain rates at the      increase of TC rain rates at the
      associated with       global scale; the median              global scale; the median              global scale; the median projected
      tropical cyclones     projected rate of increase due to     projected rate of increase due to     rate of increase due to human
      (TC)                  human emissions is about 11%.         human emissions is about 14%.         emissions is about 28%. {11.7}
                            {11.7}                                {11.7}
                                                                                                        Medium confidence that rain rates
                            Medium confidence that rain           Medium confidence that rain           will increase in every basin.
                            rates will increase in every basin.   rates will increase in every basin.   {11.7}
                            {11.7}                                {11.7}

      Increase in mean      Medium confidence                     High confidence                       High confidence
      tropical cyclone      {11.7}                                {11.7}                                {11.7}
      lifetime-maximum
      wind speed
      (intensity)

      Increase in           High confidence for an increase       High confidence for an increase       High confidence for an increase in
      likelihood that a     in the proportion of TCs that         in the proportion of TCs that         the proportion of TCs that reach
      TC will be at major   reach the strongest (Category 4-      reach the strongest (Category 4-      the strongest (Category 4-5)
      TC intensity (Cat.    5) levels. The median projected       5) levels. The median projected       levels. The median projected
      4-5)                  increase in this proportion is        increase in this proportion is        increase in this proportion is about
                            about 10%.                            about 13%.                            20%.
                            {11.7}                                {11.7}                                {11.7}

      Severe convective     There is medium confidence that the frequency of severe convective storms increases in the spring with
      storms                enhancement of convective available potential energy (CAPE), leading to extension of seasons of
                            occurrence of severe convective storms. There is high confidence of future intensification of precipitation
                            associated with severe convective storms. {11.7}

      Increase in           Likely that probability of compound events will continue to increase with global warming.
      compound events       High confidence that co-occurrent heat waves and droughts will continue to increase under higher levels of
      (frequency,           global warming, with higher frequency/intensity with every additional 0.5°C of global warming.
      intensity)
                            High confidence that fire weather, i.e. compound hot, dry and windy events, will become more frequent in
                            some regions at higher levels of global warming.
                            Medium confidence that compound flooding at the coastal zone will increase under higher levels of global
                            warming, with higher frequency/intensity with every additional 0.5°C of global warming.
                            {11.8}

 1
23

     Do Not Cite, Quote or Distribute                                11-23                                                Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   [END TABLE 11.2 HERE]
 2
 3
 4   [START BOX 11.2 HERE]
 5
 6
 7   BOX 11.2: Low-likelihood high-impact changes in extremes
 8
 9   SREX (Chapter 3) assigned low confidence to low-probability high-impact (LLHI) events. Such events are
10   often not anticipated and thus sometimes referred to as surprises. There are several types of LLHI events.
11   Abrupt changes in mean climate are addressed in Chapter 4. Unanticipated LLHI events can either result
12   from tipping points in the climate system (Section 1.4.4.3), such as the shutdown of the Atlantic
13   thermohaline circulation (SROCC Ch6; Collins et al., 2019) or the drydown of the Amazonian rainforest
14   (SR15 Ch3; Hoegh-Guldberg et al., 2018; Drijfhout et al. 2015), or from uncertainties in climate processes
15   including climate feedbacks that may enhance or damp extremes either related to global or regional climate
16   responses (Seneviratne et al., 2018b; Sutton, 2018). The low confidence does not by itself exclude the
17   possibility of such events to occur, it is instead an indication of a poor state of knowledge. Such outcomes,
18   while unlikely, could be associated with very high impacts, and are thus highly relevant from a risk
19   perspective (see Chapter 1, Section 1.4.3, Box 11.4; Sutton, 2018, 2019). Alternatively, high impacts can
20   occur when different extremes occur at the same time or in short succession at the same location or in several
21   regions with shared vulnerability (e.g. food-basket regions Gaupp et al., 2019). These “compound events”
22   are assessed in Section 11.8 and Box 11.4 provides a case-study example.
23
24   The difficulties in determining the likelihood of occurrence and time frame of potential tipping points and
25   LLHI events persist. However, new literature has emerged on unanticipated and low-probability high-impact
26   events more generally. There are events that are sufficiently rare that they have not been observed in
27   meteorological records, but whose occurrence is nonetheless plausible within the current state of the climate
28   system, see examples below and McCollum et al. (2020). The rare nature of such events and the limited
29   availability of relevant data makes it difficult to estimate their occurrence probability and thus gives little
30   evidence on whether to include such hypothetical events in planning decisions and risk assessments. The
31   estimation of such potential surprises is often limited to events that have historical analogues (including
32   before the instrumental records began, Wetter et al., 2014), albeit the magnitude of the event may differ.
33   Additionally, there is also a limitation of available resources to exhaust all plausible trajectories of the
34   climate system. As a result, there will still be events that cannot be anticipated. These events can be surprises
35   to many in that the events have not been experienced, although their occurrence could be inferred by
36   statistical means or physical modelling approaches (Chen et al., 2017; van Oldenborgh et al., 2017;
37   Harrington and Otto, 2018a). Another approach focusing on the estimation of low-probability events and of
38   events whose likelihood of occurrence is unknown consists in using physical climate models to create a
39   physically self-consistent storyline of plausible extreme events and assessing their impacts and driving
40   factors in past (Section 11.2.3) or future conditions (11.2.4) (Cheng et al., 2018; Schaller et al., 2020;
41   Shepherd, 2016; Shepherd et al., 2018; Sutton, 2018; Zappa and Shepherd, 2017; Wehrli et al., 2020;
42   Hazeleger et al., 2015).
43
44   In many parts of the world, observational data are limited to 50-60 years. This means that the chance to
45   observe an extreme event that occurs once in several hundred or more years is small. Thus, when a very
46   extreme event occurs, it becomes a surprise to many (Bao et al., 2017; McCollum et al., 2020), and very rare
47   events are often associated with high impacts (van Oldenborgh et al., 2017; Philip et al., 2018b; Tozer et al.,
48   2020). Attributing and projecting very rare events in a particular location by assessing their likelihood of
49   occurrence within the same larger region and climate thus provides another way to make quantitative
50   assessments regarding events that are extremely rare locally. Some examples of such events include for
51   instance:
52       • Hurricane Harvey, that made landfall in Houston, TX in August 2017 (Section 11.7.1.4.)
53       • The 2010-2011 extreme floods in Queensland, Australia (Christidis et al., 2013a)
54       • The 2018 concurrent heat waves across the northern Hemisphere (Box 11.4)
55       • Tropical cyclone Idai in Mozambique (Cross-Chapter Box Disaster in WGII AR6 Chapter 4)

     Do Not Cite, Quote or Distribute                      11-24                                      Total pages: 345
     Final Government Distribution                       Chapter11                                        IPCC AR6 WGI

 1       •   The California fires in 2018 and 2019
 2       •   The 2019-2020 Australia fires (Cross-Chapter Box Disaster in WGII AR6 Chapter 4)
 3
 4   One factor making such events hard to anticipate is the fact that we now live in a non-stationary climate, and
 5   that the framework of reference for adaptation is continuously moving. As an example, the concurrent heat
 6   waves that occurred across the Northern Hemisphere in the summer of 2018 were considered very unusual
 7   and were indeed unprecedented given the total area that was concurrently affected (Toreti et al., 2019; Vogel
 8   et al., 2019; Drouard et al., 2019; Kornhuber et al., 2019); however, the probability of this event under 1°C
 9   global warming was found to be about 16% (Vogel et al., 2019), which is not particularly low. Similarly, the
10   2013 summer temperature over eastern China was the hottest on record at the time, but it had an estimated
11   recurrence interval of about 4 years in the climate of 2013 (Sun et al., 2014). Furthermore, when other
12   aspects of the risk, vulnerability, and exposure are historically high or have recently increased (see WGII,
13   Chapter 16, Section 16.4), relatively moderate extremes can have very high impacts (Otto et al., 2015b;
14   Philip et al., 2018b). As warming continues, the climate moves further away from its historical state with
15   which we are familiar, resulting in an increased likelihood of unprecedented events and surprises. This is
16   particularly the case under high global warming levels e.g. such as the climate of the late 21st century under
17   high-emissions scenarios (above 4°C of global warming, CC-Box 11.1).
18
19   Another factor highlighted in Section 11.8 and Box 11.4 making events high-impact and difficult to
20   anticipate is that several locations under moderate warming levels could be affected simultaneously, or very
21   repeatedly by different types of extremes (Mora et al., 2018, Gaupp et al., 2019; Vogel et al., 2019). Box
22   11.4 shows that concurrent events at different locations, which can lead to major impacts across the world,
23   can also result from the combination of anomalous circulation or natural variability (ENSO) patterns with
24   amplification of resulting responses to human-induced global warming. Also multivariate extremes at single
25   locations pose specific challenges to anticipation (Section 11.8), with low-likelihoods in the current climate
26   but the probability of occurrence of such compound events strongly increasing with increasing global
27   warming levels (Vogel et al., 2020a). Therefore, in order to estimate whether and at what level of global
28   warming very high impacts arising from extremes would occur, the spatial extent of extremes and the
29   potential of compounding extremes need to be assessed. Sections 11.3, 11.4, 11.7 and 11.8 highlight
30   increasing evidence that temperature extremes, higher intensity precipitation accompanying tropical
31   cyclones, and compound events such as dry/hot conditions conducive to wildfire or storm surges resulting
32   from sea level rise and heavy precipitation events, pose widespread threats to societies already at relatively
33   low warming levels. Studies have already shown that the probability for some recent extreme events is so
34   small in the undisturbed world such that these events may not have been possible without human influence
35   (Section 11.2.4). Box 11.2, Table 1, provides examples of projected changes in LLHI extremes (single
36   extremes, compound events) of potential relevance for impact and adaptation assessments showing that
37   today’s very rare events can become commonplace in a warmer future.
38
39   In summary, the future occurrence of LLHI events linked to climate extremes is generally associated with
40   low confidence, but cannot be excluded, especially at global warming levels above 4°C. Compound events,
41   including concurrent extremes, are a factor increasing the probability of LLHI events (high confidence).
42   With increasing global warming some compound events with low likelihood in past and current climate
43   will become more frequent, and there is a higher chance of historically unprecedented events and
44   surprises (high confidence). However, even extreme events that do not have a particularly low probability
45   in the present climate (at more than 1°C of global warming) can be perceived as surprises because of the
46   pace of global warming (high confidence).
47
48   Box 11.2, Table 1: Examples of changes in LLHI extreme conditions (single extremes, compound events) at different
49                     global warming levels
50
                                                       +1°C (present-   +1.5°C           +2°C              +3°C and
                                                       day)                                                higher
      Risk ratio for annual hottest daytime            1                3.3 (i.e. 230%   8.2 (i.e. 720%    Not assessed
      temperature (TXx) with 1% of probability                          higher           higher
      under present-day warming (+1°C) (Kharin et                       probability)     probability)
     Do Not Cite, Quote or Distribute                      11-25                                          Total pages: 345
     Final Government Distribution                           Chapter11                                      IPCC AR6 WGI

      al., 2018): Global land
      Risk ratio for heavy precipitation events          1                1.2 (i.e. 20%    1.5 (i.e. 50%     Not assessed
      (Rx1day) with 1% of probability under present-                      higher           higher
      day warming (+1°C) (Kharin et al., 2018):                           probability)     probability)
      Global land
      Risk ratio for 1- 5 day duration extreme floods    Up to 3 in       Up to 5 in       2-6 in most       Up to 12 in
      with 1% of probability under present-day           individual       individual       locations         individual
      warming (+1°C) (Ali et al., 2019a)Indian           locations        locations                          locations
      subcontinent                                                                                           (4°C)
      Probability of “extremes extremes” hot days        ~20 days over    about ~50        about ~150        about ~500
      with 1/1000 probability at the end of 20th         20 years in      days in 20       days in 20        days in 20
      century (Vogel et al., 2020a): Global land         most locations   years in most    years in most     years in most
                                                                          locations        locations         locations
                                                                                                             (3°C)
      Probability of co-occurrence in the same week      0% probability   ~1 week          ~4-5 weeks        >9 weeks
      of hot days with 1/1000 probability and dry                         within 20        within 20         within 20
      days with 1/1000 probability at the end of 20th                     years            years             years (3°C)
      century (Vogel et al., 2020a): Amazon
      Projected soil moisture drought duration per       41 days (+46%    58 days          71 days           125 days
      year (Samaniego et al., 2018): Mediterranean       compared to      (+107%           (+154%            (+346%
      region                                             late 20th        compared to      compared to       compared to
                                                         century)         late 20th        late 20th         late 20th
                                                                          century)         century)          century)
                                                                                                             (3°C)
      Increase in days exposed to dangerous extreme      Not assessed,    1.6 times        2.3 times         ~ 80% of
      heat (measured in Health Heat Index (HHI)          baseline is      higher risk of   higher risk of    land area
      (Sun et al., 2019c) global land                    1981-2000        experiencing     experiencing      exposed to
                                                                          heat > 40.6      heat > 40.6       dangerous
                                                                                                             heat, tropical
                                                                                                             regions 1/3
                                                                                                             of the year
                                                                                                             (4°C)
      Increase in regional mean fire season length       Not assessed,    6.2 days         9.5 days          ~ 50 days
      (Sun et al., 2019c; Xu et al., 2020) global land   baseline is                                         (4°C)
                                                         1981-2000
 1
 2
 3   [END BOX 11.2 HERE]
 4
 5
 6   11.2 Data and Methods
 7
 8   This section provides an assessment of observational data and methods used in the analysis and attribution of
 9   climate change specific to weather and climate extremes, and also introduces some concepts used in
10   presenting future projections of extremes in the chapter. The main focus is on extreme events over land, as
11   extremes in the ocean are assessed in Chapter 9 of this Report. Later sections (11.3-11.8) also provide
12   additional assessments on relevant observational datasets and model validation specific for the type of
13   extremes to be assessed. General background on climate modelling is provided in Chapters 4 and 10.
14
15
16   11.2.1 Definition of extremes
17
18   In the literature, an event is generally considered extreme if the value of a variable exceeds (or lies below) a

     Do Not Cite, Quote or Distribute                         11-26                                         Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   threshold. The thresholds have been defined in different ways, leading to differences in the meaning of
 2   extremes that may share the same name. For example, two sets of frequency of hot/warm days have been
 3   used in the literature. One set counts the number of days when maximum daily temperature is above a
 4   relative threshold defined as the 90th or higher percentile of maximum daily temperature for the calendar day
 5   over a base period. An event based on such a definition can occur during any time of the year and the impact
 6   of such an event would differ depending on the season. The other set counts the number of days in which
 7   maximum daily temperature is above an absolute threshold such as 35°C, because exceedance of this
 8   temperature can sometimes cause health impacts (however, these impacts may depend on location and
 9   whether ecosystems and the population are adapted to such temperatures). While both types of hot extreme
10   indices have been used to analyze changes in the frequency of hot/warm events, they represent different
11   events that occur at different times of the year, possibly affected by different types of processes and
12   mechanisms, and possibly also associated with different impacts.
13
14   Changes in extremes have also been examined from two perspectives: changes in the frequency for a given
15   magnitude of extremes or changes in the magnitude for a particular return period (frequency). Changes in the
16   probability of extremes (e.g., temperature extremes) depend on the rarity of the extreme event that is
17   assessed, with a larger change in probability associated with a rarer event (e.g., Kharin et al., 2018). On the
18   other hand, changes in the magnitude represented by the return levels of the extreme events may not be as
19   sensitive to the rarity of the event. While the answers to the two different questions are related, their
20   relevance to different audiences may differ. Conclusions regarding the respective contribution of greenhouse
21   gas forcing to changes in magnitude versus frequency of extremes may also differ (Otto et al., 2012).
22   Correspondingly, the sensitivity of changes in extremes to increasing global warming is also dependent on
23   the definition of the considered extremes. In the case of temperature extremes, changes in magnitude have
24   been shown to often depend linearly on global surface temperature (Seneviratne et al., 2016; Wartenburger et
25   al., 2017), while changes in frequency tend to be non-linear and can, for example, be exponential for
26   increasing global warming levels (Fischer and Knutti, 2015; Kharin et al., 2018). When similar damage
27   occurs once a fixed threshold is exceeded, it is more important to ask a question regarding changes in the
28   frequency. But when the exceedance of this fixed threshold becomes a normal occurrence in the future, this
29   can lead to a saturation in the change of probability (Harrington and Otto, 2018a). On the other hand, if the
30   impact of an event increases with the intensity of the event, it would be more relevant to examine changes in
31   the magnitude. Finally, adaptation to climate change might change the relevant thresholds over time,
32   although such aspects are still rarely integrated in the assessment of projected changes in extremes. Framing,
33   including how extremes are defined and how the questions are asked in the literature, is considered when
34   forming the assessments of this chapter.
35
36
37   11.2.2 Data
38
39   Studies of past and future changes in weather and climate extremes and in the mean state of the climate use
40   the same original sources of weather and climate observations, including in-situ observations, remotely
41   sensed data, and derived data products such as reanalyses. Chapter 2 (Section 2.3) and Chapter 10 (Section
42   10.2) assess various aspects of these data sources and data products from the perspective of their general use
43   and in the analysis of changes in the mean state of the climate in particular. Building on these previous
44   chapters, this subsection highlights particular aspects that are related to extremes and that are most relevant
45   to the assessment of this chapter. The SREX (Chapter 3, Seneviratne et al., 2012) and AR5 (Chapter 2,
46   Hartmann et al., 2013) addressed critical issues regarding the quality and availability of observed data and
47   their relevance for the assessment of changes in extremes.
48
49   Extreme weather and climate events occur on time scales of hours (e.g., convective storms that produce
50   heavy precipitation) to days (e.g., tropical cyclones, heat waves), to seasons and years (e.g., droughts). A
51   robust determination of long-term changes in these events can have different requirements for the spatial and
52   temporal scales and sample size of the data. In general, it is more difficult to determine long-term changes
53   for events of fairly large temporal duration, such as “mega-droughts” that last several years or longer (e.g.,
54   Ault et al. 2014), because of the limitations of the observational sample size. Literature that study changes in
55   extreme precipitation and temperature often use indices representing specifics of extremes that are derived
     Do Not Cite, Quote or Distribute                     11-27                                      Total pages: 345
     Final Government Distribution                       Chapter11                                     IPCC AR6 WGI

 1   from daily precipitation and temperature values. Station-based indices would have the same issues as those
 2   for the mean climate regarding the quality, availability, and homogeneity of the data. For the purpose of
 3   constructing regional information and/or for comparison with model outputs, such as model evaluation, and
 4   detection and attribution, these station-based indices are often interpolated onto regular grids. Two different
 5   approaches, involving two different orders of operation, have been used in producing such gridded datasets.
 6
 7   In some cases, such as for the HadEX3 dataset (Dunn et al., 2020), indices of extremes are computed using
 8   time series directly derived from stations first and are then gridded over the space. As the indices are
 9   computed at the station level, the gridded data products represent point estimates of the indices averaged
10   over the spatial scale of the grid box. In other instances, daily values of station observations are first gridded
11   (e.g., Contractor et al., 2020), and the interpolated values can then be used to compute various indices by the
12   users. Depending on the station density, values for extremes computed from data gridded this way represent
13   extremes of spatial scales anywhere from the size of the grid box to a point. In regions with high station
14   density (e.g., North America, Europe), the gridded values are closer to extremes of area means and are thus
15   more appropriate for comparisons with extremes estimated from climate model output, which is often
16   considered to represent areal means (Chen and Knutson, 2008; Gervais et al., 2014; Avila et al., 2015; Di
17   Luca et al., 2020a). In regions with very limited station density (e.g., Africa), the gridded values are closer to
18   point estimates of extremes. The difference in spatial scales among observational data products and model
19   simulations needs to be carefully accounted for when interpreting the comparison among different data
20   products. For example, the average annual maximum daily maximum temperature (TXx) over land
21   computed from the original ERA-interim reanalysis (at 0.75° resolution) is about 0.4°C warmer than that
22   computed when the ERA-interim dataset is upscaled to the resolution of 2.5° x 3.75° (Di Luca et al., 2020).
23
24   Extreme indices computed from various reanalysis data products have been used in some studies, but
25   reanalysis extreme statistics have not been rigorously compared to observations (Donat et al., 2016a).
26   In general, changes in temperature extremes from various reanalyses were most consistent with gridded
27   observations after about 1980, but larger differences were found during the pre-satellite era (Donat et al.,
28   2014b). Overall, lower agreement across reanalysis datasets was found for extreme precipitation changes,
29   although temporal and spatial correlations against observations were found to be still significant. In regions
30   with sparse observations (e.g., Africa and parts of South America), there is generally less agreement for
31   extreme precipitation between different reanalysis products, indicating a consequence of the lack of an
32   observational constraint in these regions (Donat et al., 2014b, 2016a). More recent reanalyses, such as ERA5
33   (Hersbach et al., 2020), seem to have improved over previous products, at least over some regions (e.g.,
34   Mahto and Mishra, 2019; Gleixner et al., 2020; Sheridan et al., 2020). Caution is needed when reanalysis
35   data products are used to provide additional information about past changes in these extremes in regions
36   where observations are generally lacking.
37
38   Satellite remote sensing data have been used to provide information about precipitation extremes because
39   several products provide data at sub-daily resolution for precipitation (e.g., TRMM; Maggioni et al. 2016)
40   and clouds (e.g., HIMAWARI; Bessho et al., 2016; Chen et al. 2019). However, satellites do not observe the
41   primary atmospheric state variables directly and polar orbiting satellites do not observe any given place at all
42   times. Hence, their utility as a substitute for high-frequency (i.e., daily) ground-based observations is limited.
43   For instance, Timmermans et al. (2019) found little relationship between the timing of extreme daily and
44   five-day precipitation in satellite and gridded station data products over the United States.
45
46
47   [START BOX 11.3 HERE]
48
49   BOX 11.3: Extremes in paleoclimate archives compared to instrumental records
50   Examining extremes in pre-instrumental information can help to put events occurring in the instrumental
51   record (referred to as ‘observed’) in a longer-term context. This box focuses on extremes in the Common Era
52   (CE, the last 2000 years), because there is generally higher confidence in pre-instrumental information
53   gathered from the more recent archives from the Common Era than from earlier evidence. It addresses
54   evidence of extreme events in paleo reconstructions, documentary evidence (such as grape harvest data,
55   religious documents, newspapers, and logbooks) and model-based analyses, and whether observed extremes
     Do Not Cite, Quote or Distribute                      11-28                                       Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   have or have not been exceeded in the Common Era. This box provides overviews of i) AR5 assessments
 2   and ii) types of evidence assessed here, evidence of iii) droughts, iv) temperature extremes, v) paleofloods,
 3   and vi) paleotempests, and vii) a summary.
 4
 5   AR5 (Chapter 5, Masson-Delmotte et al., 2013) concluded with high confidence that droughts of greater
 6   magnitude and of longer duration than those observed in the instrumental period occurred in many regions
 7   during the preceding millennium. There was high confidence in evidence that floods during the past five
 8   centuries in northern and Central Europe, the western Mediterranean region, and eastern Asia were of a
 9   greater magnitude than those observed instrumentally, and medium confidence in evidence that floods in the
10   near East, India and central North America were comparable to modern observed floods. While AR5
11   assessed 20th century summer temperatures compared to those reconstructed in the Common Era, it did not
12   assess shorter duration temperature extremes.
13
14   Many factors affect confidence in information on pre-instrumental extremes. First, the geographical coverage
15   of paleoclimate reconstructions of extremes is not spatially uniform (Smerdon and Pollack, 2016) and
16   depends on both the availability of archives and records, which are environmentally dependent, and also the
17   differing attention and focus from the scientific community. In Australia, for example, the paleoclimate
18   network is sparser than for other regions, such as Asia, Europe and North America, and synthesised products
19   rely on remote proxies and assumptions about the spatial coherence of precipitation between remote climates
20   (Cook et al., 2016c; Freund et al., 2017). Second, pre-instrumental evidence of extremes may be focused on
21   understanding archetypal extreme events, such as the climatic consequences of the 1815 eruption of Mount
22   Tambora, Indonesia (Brohan et al., 2016; Veale and Endfield, 2016). These studies provide narrow evidence
23   of extremes in response to specific forcings (Li, 2017) for specific epochs. Third, natural archives may
24   provide information about extremes in one season only and may not represent all extremes of the same types.
25
26   Evidence of shorter duration extreme event types, such as floods and tropical storms, is further restricted by
27   the comparatively low chronological controls and temporal resolution (e.g., monthly, seasonal, yearly,
28   multiple years) of most archives compared to the events (e.g., minutes to days). Natural archives may be
29   sensitive only to intense environmental disturbances, and so only sporadically record short-duration or small
30   spatial scale extremes. Interpreting sedimentary records as evidence of past short-duration extremes is also
31   complex and requires a clear understanding of natural processes. For example, paleoflood reconstructions of
32   flood recurrence and intensity produced from geological evidence (e.g., river and lake sediments),
33   speleothems (Denniston and Luetscher, 2017), botanical evidence (e.g., flood damage to trees, or tree ring
34   reconstructions), and floral and faunal evidence (e.g., diatom fossil assemblages) require understanding of
35   sediment sources and flood mechanisms. Pre-instrumental records of tropical storm intensity and frequency
36   (also called paleotempest records) derived from overwash deposits of coastal lake and marsh sediments are
37   difficult to interpret. Many factors impact whether disturbances are deposited in archives (Muller et al.,
38   2017) and deposits may provide sporadic and incomplete preservation histories (e.g., Tamura et al., 2018).
39
40   Overall, the most complete pre-instrumental evidence of extremes occurs for long-duration, large-spatial-
41   scale extremes, such as for multi-year meteorological droughts or seasonal- and regional-scale temperature
42   extremes. Additionally, more precise insights into recent extremes emerge where multiple studies have been
43   undertaken, compared to the confidence in extremes reported at single sites or in single studies, which may
44   not necessarily be representative of large-scale changes, or for reconstructions that synthesise multiple
45   proxies over large areas (e.g., drought atlases). Multiproxy synthesis products combine paleoclimate
46   temperature reconstructions and cover sub-continental- to hemispheric-scale regions to provide continuous
47   records of the Common Era (e.g. Ahmed et al., 2013; Neukom et al., 2014 for temperature).
48
49   There is high confidence in the occurrence of long-duration and severe drought events during the Common
50   Era for many locations, although their severity compared to recent drought events differs between locations
51   and the lengths of reconstruction provided. Recent observed drought extremes in some regions (such as the
52   Levant (Cook et al., 2016a), California in the United States (Cook et al., 2014; Griffin and Anchukaitis,
53   2014), and the Andes (Domínguez-Castro et al., 2018)) do not have precedents within the multi-century
54   periods reconstructed in these studies, in terms of duration and/or severity. In some regions (in Southwest
55   North America (Asmerom et al., 2013; Cook et al., 2015), the Great Plains region (Cook et al., 2004), the
     Do Not Cite, Quote or Distribute                     11-29                                      Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   Middle East (Kaniewski et al., 2012), and China (Gou et al., 2015), recent drought extremes may have been
 2   exceeded in the Common Era. In further locations, there is conflicting evidence for the severity of pre-
 3   instrumental droughts compared to observed extremes, depending on the length of the reconstruction and the
 4   seasonal perspective provided (see Cook et al., 2016b; Freund et al., 2017 for Australia). There can also be
 5   differing conclusions for the severity, or even the occurrence, of specific individual pre-instrumental
 6   droughts when different evidence is compared (e.g., Büntgen et al., 2015; Wetter et al., 2014).
 7
 8   There is medium confidence that the magnitude of large-scale, seasonal-scale extreme high temperatures in
 9   observed records exceed those reconstructed over the Common Era in some locations, such as Central
10   Europe. In one example, multiple studies have examined the unusualness of present-day European summer
11   temperature records in a long-term context, particularly in comparison to the exceptionally warm year of
12   1540 CE in Central Europe. Several studies indicate recent extreme summers (2003 and 2010) in Europe
13   have been unusually warm in the context of the last 500 years (Barriopedro et al., 2011; Wetter and Pfister,
14   2013; Wetter et al., 2014; Orth et al., 2016a), or longer (Luterbacher et al., 2016). Others studies show
15   summer temperatures in Central Europe in 1540 were warmer than the present-day (1966–2015) mean, but
16   note that it is difficult to assess whether or not the 1540 summer was for its part warmer than observed
17   record extreme temperatures (Orth et al., 2016a).
18
19   There is high confidence that the magnitude of floods over the Common Era has exceeded observed records
20   in some locations, including Central Europe and eastern Asia. Recent literature supports the AR5
21   assessments (Masson-Delmotte et al., 2013) of floods. High temporally resolved records provide evidence,
22   for example, of Common Era floods exceeding the probable maximum flood levels in the Upper Colorado
23   River, USA (Greenbaum et al., 2014) and peak discharges that are double gauge levels along the middle
24   Yellow River, China (Liu et al., 2014). Further studies demonstrate pre-instrumental or early instrumental
25   differences in flood frequency compared to the instrumental period, including reconstructions of high and
26   low flood frequency in the European Alps (e.g., Swierczynski et al., 2013; Amann et al., 2015) and
27   Himalayas (Ballesteros Cánovas et al., 2017). The combination of extreme historical flood episodes
28   determined from documentary evidence also increases confidence in the determination of flood frequency
29   and magnitude, compared to using geomorphological archives alone (Kjeldsen et al., 2014). In regions, such
30   as Europe and China, that have rich historical flood documents, there is strong evidence of high magnitude
31   flood events over pre-instrumental periods (Benito et al., 2015; Kjeldsen et al., 2014; Macdonald and
32   Sangster, 2017). A key feature of paleoflood records is variability in flood recurrence at centennial
33   timescales (Wilhelm et al., 2019), although constraining climate-flood relationships remains challenging.
34   Pre-instrumental floods often occurred in considerably different contexts in terms of land use, irrigation, and
35   infrastructure, and may not provide direct insight into modern river systems, which further prevents long-
36   term assessments of flood changes being made based on these sources.
37
38   There is medium confidence that periods of both more and less tropical cyclone activity (frequency or
39   intensity) than observed occurred over the Common Era in many regions. Paleotempest studies cover a
40   limited number of locations that are predominantly coastal, and hence provide information on specific
41   locations that cannot be extrapolated basin-wide (see Muller et al., 2017). In some locations, such as the Gulf
42   of Mexico and the New England coast, similarly intense storms to those observed recently have occurred
43   multiple times over centennial timescales (Donnelly et al., 2001; Bregy et al., 2018). Further research
44   focused on the frequency of tropical storm activity. Extreme storms occurred considerably more frequently
45   in particular periods of the Common Era, compared to the instrumental period in northeast Queensland,
46   Australia (Nott et al., 2009; Haig et al., 2014), and the Gulf Coast (e.g., Brandon et al., 2013; Lin et al.,
47   2014).
48
49   The probability of finding an unprecedented extreme event increases with an increased length of past record-
50   keeping, in the absence of longer-term trends. Thus, as a record is extended to the past based on paleo-
51   reconstruction, there is a higher chance of very rare extreme events having occurred at some time prior to
52   instrumental records. Such an occurrence is not, in itself, evidence of a change, or lack of a change, in the
53   magnitude or the likelihood of extremes in the past or in the instrumental period at regional and local scales.
54   Yet, the systematic collection of paleoclimate records over wide areas may provide evidence of changes in
55   extremes. In one study, extended evidence of the last millennium from observational data and paleoclimate
     Do Not Cite, Quote or Distribute                     11-30                                      Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   reconstructions using tree rings indicates human activities affected the worldwide occurrence of droughts as
 2   early as the beginning of the 20th century (Marvel et al., 2019).
 3
 4   In summary, there is low confidence in overall changes in extremes derived from paleo-archives. The most
 5   robust evidence is high confidence that high-duration and severe drought events occurred at many locations
 6   during the last 2000 years. There is also high confidence that high-magnitude flood events occurred at some
 7   locations during the last 2000 years, but overall changes in infrastructure and human water management
 8   make the comparison with present-day records difficult. But these isolated paleo-drought and paleo-flood
 9   events are not evidence of a change, or lack of a change, in the magnitude or the likelihood of relevant
10   extremes.
11
12   [END BOX 11.3 HERE]
13
14
15   11.2.3 Attribution of extremes
16
17   Attribution science concerns the identification of causes for changes in characteristics of the climate system
18   (e.g., trends, single extreme events). A general overview and summary of methods of attribution science is
19   provided in the Cross-Working Group Box 1.1 (in Chapter 1). Trend detection using optimal fingerprinting
20   methods is a well-established field, and has been assessed in the AR5 (Chapter 10, Bindoff et al., 2013), and
21   Chapter 3 in this Report (Section 3.2.1). There are specific challenges when applying optimal fingerprinting
22   to the detection and attribution of trends in extremes and on regional scales where the lower signal-to-noise
23   ratio is a challenge. In particular, the method generally requires the data to follow a Normal (Gaussian)
24   distribution, which is often not the case for extremes. Recent studies showed that extremes can, however, be
25   transformed to a Gaussian distribution, for example by averaging over space, so that optimal fingerprinting
26   techniques can still be used (Zhang et al., 2013; Wen et al., 2013; and Wan et al., 2019). Non-stationary
27   extreme value distributions, which allow for the detailed detection and attribution of regional trends in
28   temperature extremes, have also been used (Wang et al., 2017c).
29
30   Apart from the detection and attribution of trends in extremes, new approaches have been developed to
31   answer the question of whether and to what extent external drivers have altered the probability and intensity
32   of an individual extreme event (NASEM, 2016). In AR5, there was an emerging consensus that the role of
33   external drivers of climate change in specific extreme weather events could be estimated and quantified in
34   principle, but related assessments were still confined to particular case studies, often using a single model,
35   and typically focusing on high-impact events with a clear attributable signal.
36
37   However, since AR5, the attribution of extreme weather events has emerged as a growing field of climate
38   research with an increasing body of literature (see series of supplements to the annual State of the Climate
39   report (Peterson et al., 2012, 2013b, Herring et al., 2014, 2015, 2016, 2018), including the number of
40   approaches to examining extreme events (described in Easterling et al., 2016; Otto, 2017; Stott et al., 2016)).
41   A commonly-used approach, often called the risk-based approach in the literature and referred to here as the
42   “probability-based approach”, produces statements such as ‘anthropogenic climate change made this event
43   type twice as likely’ or ‘anthropogenic climate change made this event 15% more intense’. This is done by
44   estimating probability distributions of the index characterizing the event in today’s climate, as well as in a
45   counterfactual climate, and either comparing intensities for a given occurrence probability (e.g., 1-in-100
46   year event) or probabilities for a given magnitude (see FAQ 11.3). There are a number of different analytical
47   methods encompassed in the probability-based approach building on observations and statistical analyses
48   (e.g., van Oldenborgh et al., 2012), optimal fingerprint methods (Sun et al., 2014), regional climate and
49   weather forecast models (e.g., Schaller et al., 2016), global climate models (GCMs) (e.g., Lewis and Karoly,
50   2013), and large ensembles of atmosphere-only GCMs (e.g., Lott et al., 2013). A key component in any
51   event attribution analysis is the level of conditioning on the state of the climate system. In the least
52   conditional approach, the combined effect of the overall warming and changes in the large-scale atmospheric
53   circulation are considered and often utilize fully coupled climate models (Sun et al., 2014). Other more
54   conditional approaches involve prescribing certain aspects of the climate system. These range from
55   prescribing the pattern of the surface ocean change at the time of the event (e.g. Hoerling et al., 2013, 2014),
     Do Not Cite, Quote or Distribute                     11-31                                      Total pages: 345
     Final Government Distribution                       Chapter11                                      IPCC AR6 WGI

 1   often using AMIP-style global models, where the choice of sea surface temperature and ice patterns
 2   influences the attribution results (Sparrow et al., 2018), to prescribing the large-scale circulation of the
 3   atmosphere and using weather forecasting models or methods (e.g., Pall et al., 2017; Patricola and Wehner,
 4   2018; Wehner et al., 2018a). These highly conditional approaches have also been called “storylines”
 5   (Shepherd, 2016; Cross-Working Group Box 1.1 in Chapter 1) and can be useful when applied to extreme
 6   events that are too rare to otherwise analyse or where the specific atmospheric conditions were central to the
 7   impact. These methods are also used to enable the use of very-high-resolution simulations in cases were
 8   lower-resolution models do not simulate the regional atmospheric dynamics well (Shepherd, 2016; Shepherd
 9   et al., 2018). However, the imposed conditions limit an overall assessment of the anthropogenic influence on
10   an event, as the fixed aspects of the analysis may also have been affected by climate change. For instance,
11   the specified initial conditions in the highly conditional hindcast attribution approach often applied to
12   tropical cyclones (e.g., Patricola and Wehner, 2018; Takayabu et al., 2015) permit only a conditional
13   statement about the magnitude of the storm if similar large-scale meteorological patterns could have
14   occurred in a world without climate change, thus precluding any attribution statement about the change in
15   frequency if used in isolation. Combining conditional assessments of changes in the intensity with a multi-
16   model approach does allow for the latter as well (Shepherd, 2016).
17
18   The outcome of event attribution is dependent on the definition of the event (Leach et al., 2020), as well as
19   the framing (Christidis et al., 2018; Jézéquel et al., 2018; Otto et al., 2016) and uncertainties in observations
20   and modelling. Observational uncertainties arise both in estimating the magnitude of an event as well as its
21   rarity (Angélil et al., 2017). Results of attribution studies can also be very sensitive to the choice of climate
22   variables (Sippel and Otto, 2014; Wehner et al., 2016). Attribution statements are also dependent on the
23   spatial (Uhe et al., 2016; Cattiaux and Ribes, 2018; Kirchmeier‐Young et al., 2019) and temporal
24   (Harrington, 2017; Leach et al., 2020) extent of event definitions, as events of different scales involve
25   different processes (Zhang et al., 2020d) and large-scale averages generally yield higher attributable changes
26   in magnitude or probability due to the smoothing out of the noise. In general, confidence in attribution
27   statements for large-scale heat and lengthy extreme precipitation events have higher confidence than shorter
28   and more localized events, such as extreme storms, an aspect also relevant for determining the emergence of
29   signals in extremes or the confidence in projections (see also Cross-Chapter Box Atlas.1)
30
31   The reliability of the representation of the event in question in the climate models used in a study is essential
32   (Angélil et al., 2016; Herger et al., 2018). Extreme events characterized by atmospheric dynamics that stretch
33   the capabilities of current-generation models (see Section 10.3.3.4, Shepherd, 2014; Woollings et al., 2018)
34   limit the applicability of the probability-based approach of event attribution. The lack of model evaluation, in
35   particular in early event attribution studies, has led to criticism of the emerging field of attribution science as
36   a whole (Trenberth et al., 2015) and of individual studies (Angélil et al., 2017). In this regard, the storyline
37   approach (Shepherd, 2016) provides an alternative option that does not depend on the model’s ability to
38   represent the circulation reliably. In addition, several ways of quantifying statistical uncertainty (Paciorek et
39   al., 2018) and model evaluation (Lott and Stott, 2016; Philip et al., 2018b, 2020) have been employed to
40   evaluate the robustness of event attribution results. For the unconditional probability-based approach, multi-
41   model and multi-approach (e.g., combining observational analyses and model experiments) methods have
42   been used to improve the robustness of event attribution (Hauser et al., 2017; Otto et al., 2018a; Philip et al.,
43   2018b, 2019, 2020; van Oldenborgh et al., 2018; Kew et al., 2019).
44
45   In the regional tables provided in Section 11.9, the different lines of evidence from event attribution studies
46   and trend attributions are assessed alongside one another to provide an assessment of the human contribution
47   to observed changes in extremes in all AR6 regions .
48
49
50   11.2.4 Projecting changes in extremes as a function of global warming levels
51
52   The most important quantity used to characterize past and future climate change is global warming relative
53   to its pre-industrial level. On the one hand, changes in global warming are linked quasi-linearly to global
54   cumulative CO2 emissions (IPCC, 2013). On the other hand, changes in regional climate, including many
55   types of extremes, scale quasi-linearly with changes in global warming, often independently of the
     Do Not Cite, Quote or Distribute                      11-32                                       Total pages: 345
     Final Government Distribution                       Chapter11                                     IPCC AR6 WGI

 1   underlying emissions scenarios (SR15 Ch3; Seneviratne et al., 2016; Wartenburger et al., 2017; Matthews et
 2   al., 2017; Tebaldi and Knutti 2018, Sun et al., 2018a, Kharin et al., 2018, Beusch et al., 2020b; Li et al.,
 3   2020). Finally, the use of global warming levels in the context of global policy documents (in particular the
 4   2015 Paris Agreement, UNFCCC 2015), implies that information on changes in the climate system, and in
 5   particular extremes, as a function of global warming are of particular policy relevance. Cross-Chapter Box
 6   11.1 provides on overview on the translation between information at global warming levels (GWLs) and
 7   scenarios.
 8
 9   The assessment of projections of future changes in extremes as function of GWL has an advantage in
10   separating uncertainty associated with the global warming response (see Chapter 4) from the uncertainty
11   resulting from the regional climate response as a function of GWLs (Seneviratne and Hauser, 2020). If the
12   interest is in the projection of regional changes at certain GWLs, such as those defined by the Paris
13   Agreement, projections based on time periods and emission scenarios have unnecessarily larger uncertainty
14   due to differences in model global transient climate responses. To take advantage of this feature and to
15   provide easy comparison with SR15, assessments of projected changes in this chapter are largely provided in
16   relation to future GWLs, with a focus on changes at +1.5°C, +2°C, and +4°C of global warming above pre-
17   industrial levels (e.g. Tables 11.1, 11.2 and regional tables in Section 11.9). These encompass a scenario
18   compatible with the aim of the Paris Agreement (+1.5°C), a scenario slightly overshooting the aims of the
19   Paris Agreement (+2°C), and a “worst-case” scenario with no mitigation (+4°C). The CC-Box 11.1 provides
20   a background on the GWL sampling approach used in the AR6, both for the computation of GWL
21   projections from ESMs contributing to the 6th Phase of the Coupled Model Intercomparison Project (CMIP6)
22   as well as for the mapping of existing scenario-based literature for CMIP6 and the 5th Phase of CMIP
23   (CMIP5) to assessments as function of GWLs (see also Section 11.9. and Table 11.3 for an example).
24
25   While regional changes in many types of extremes do scale robustly with global surface temperature,
26   generally irrespective of emission scenarios (Section 11.1.4; Figures 11.3, 11.6, 11.7; CC-Box 11.1), effects
27   of local forcing can distort this relation. In particular, emission scenarios with the same radiative forcing can
28   have different regional extreme precipitation responses resulting from different aerosol forcing (Wang et al.,
29   2017d). Another example is related to forcing from land use and land cover changes (Section 11.1.6).
30   Climate models often either overestimate or underestimate observed changes in annual maximum daily
31   maximum temperature depending on the region and considered models (Donat et al., 2017; Vautard et al.,
32   9999). Part of the discrepancies may be due to the lack of representation of some land forcings, in particular
33   crop intensification and irrigation (Mueller et al., 2016b; Thiery et al., 2017; Findell et al., 2017; Thiery et
34   al., 2020). Since these local forcings are not represented and their future changes are difficult to project,
35   these can be important caveats when using GWL scaling to project future changes for these regions.
36   However, these caveats also apply to the use of scenario-based projections.
37
38   SR15 (Chapter 3) assessed different climate responses at +1.5°C of global warming, including transient
39   climate responses, short-term stabilization responses, and long-term equilibrium stabilization responses, and
40   their implications for future projections of different extremes. Indeed, the temporal dimension, that is, when
41   the given GWL occurs, also matters for projections, in particular beyond the 21st century and for some
42   climate variables with large inertia (e.g., sea level rise and associated extremes). Nonetheless, for
43   assessments focused on conditions within the next decades and for the main extremes considered in this
44   chapter, derived projections are relatively insensitive to details of climate scenarios and can be well
45   estimated based on transient simulations (CC-Box 11.1; see also SR15).
46
47   An important question is the identification of the GWL at which a given change in a climate extreme can
48   begin to emerge from climate noise. Figure 11.8 displays analyses of the GWLs at which emergence in hot
49   extremes (20-year return values of TXx, TXx_20yr) and heavy precipitation (20-year return values of
50   Rx1day, Rx1day_20yr) is identified in AR6 regions for the whole CMIP5 and CMIP6 ensembles). Overall,
51   signals for extremes emerge very early for TXx_20yr, already below 0.2°C in many regions (Fig. 11.8a,b),
52   and at around 0.5°C in most regions. This is consistent with conclusions from the SR15 Ch3 for less-rare
53   temperature extremes (TXx on the yearly time scale), which shows that a difference as small as 0.5°C of
54   global warming, e.g. between +1.5°C and +2°C of global warming, leads to detectable differences in
55   temperature extremes in TXx in most WGI AR6 regions in CMIP5 projections (e.g.,Wartenburger et al.,
     Do Not Cite, Quote or Distribute                      11-33                                       Total pages: 345
     Final Government Distribution                         Chapter11                                      IPCC AR6 WGI

 1   2017; Seneviratne et al., 2018b). The GWL emergence for Rx1day_20yr is also largely consistent with
 2   analyses for less-extreme heavy precipitation events (Rx5day on the yearly time scale) in the SR15 (see
 3   Chapter 3).
 4
 5   To some extent, analyses as functions of GWLs replace the time axis with a global surface temperature axis.
 6   Nonetheless, information on the timing of given changes in extremes is obviously also relevant. Regarding
 7   this information, that is, the time frame at which given global warming levels are reached, the readers are
 8   referred to Chapter 4 (Section 4.6; see also CC-Box 11.1). Figure 11.5 provides a synthesis of attributed and
 9   projected changes in extremes as function of GWLs (see also Figs. 11.3, 11.6, and 11.7 for regional
10   analyses).
11
12
13   [START FIGURE 11.8 HERE]
14
15   Figure 11.8: Global and regional-scale emergence of changes in temperature (a) and precipitation (b) extremes for the
16                globe (glob.), global oceans (oc.), global lands (land), and the AR6 regions. Colours indicate the multi-
17                model mean global warming level at which the difference in 20-year means of the annual maximum daily
18                maximum temperature (TXx) and the annual maximum daily precipitation (Rx1day) become significantly
19                different from their respective mean values during the 1851–1900 base period. Results are based on
20                simulations from the CMIP5 and CMIP6 multi-model ensembles. See Atlas.1.3.2 for the definition of
21                regions. Adapted from Seneviratne and Hauser, 2020) under the terms of the Creative Commons
22                Attribution license.
23
24   [END FIGURE 11.8 HERE]
25
26
27   [START CROSS-CHAPTER BOX 11.1 HERE]
28
29   Cross-Chapter Box 11.1:         Translating between regional information at global warming levels vs
30                                   scenarios for end users
31
32   Contributors: Erich Fischer (Switzerland), Mathias Hauser (Switzerland), Sonia I. Seneviratne
33   (Switzerland), Richard Betts (UK), José M. Gutiérrez (Spain), Richard G. Jones (UK), June-Yi Lee
34   (Republic of Korea), Malte Meinshausen (Australia/Germany), Friederike Otto (UK/Germany), Izidine Pinto
35   (Mozambique), Roshanka Ranasinghe (The Netherlands/Sri Lanka/Australia), Joeri Rogelj
36   (Germany/Belgium), Bjørn Samset (Norway), Claudia Tebaldi (USA), Laurent Terray (France)
37
38   Background
39   Traditionally, projections of climate variables are summarized and communicated as function of time and
40   scenario. Recently, quantifying global and regional climate at specific global warming levels (GWLs) has
41   become widespread, motivated by the inclusion of explicit GWLs in the long-term temperature goal of the
42   Paris Agreement (Section 1.6.2). GWLs, expressed as changes in global surface temperature relative to the
43   1850-1900 period (see CCBox 2.3), are used in the SR15 and in the assessment of Reasons for Concerns in
44   the WGII reports (see also CCBox 12.1). CCB 11.1, Figure 1 illustrates how the assessment of the climate
45   response at GWLs relates to the uncertainty in scenarios regarding the timing of the respective GWLs, as
46   well as to the uncertainty in the associated regional climate responses, including extremes and other climatic
47   impact-drivers (CIDs). For many (but not all) climate variables and CIDs the response pattern for a given
48   GWL is consistent across different scenarios (Chapters 1, 4, 9, 11 and Atlas). GWLs are defined as long-
49   term means (e.g. 20-year averages) compared to the pre-industrial period, are commonly used in the
50   literature and were also underlying main assessments of SR15 (Chapter 3).
51
52
53   [START CROSS-CHAPTER BOX 11.1, FIGURE 1 HERE]
54
55   Cross-Chapter Box 11.1, Figure 1: Schematic representation of relationship between emission scenarios, global

     Do Not Cite, Quote or Distribute                        11-34                                        Total pages: 345
     Final Government Distribution                         Chapter11                                        IPCC AR6 WGI

 1                                        warming levels (GWLs), regional climate responses, and impacts. The
 2                                        illustration shows the implied uncertainty problem associated with differentiating
 3                                        between 1.5, 2°C, and other GWLs. Focusing on GWL raises questions
 4                                        associated with emissions pathways to get to these temperatures (scenarios), as
 5                                        well as questions associated with regional climate responses and the associated
 6                                        impacts at the corresponding GWL (the impacts question). Adapted from (James
 7                                        et al., 2017) and (Rogelj, 2013) under the terms of the Creative Commons
 8                                        Attribution license.
 9
10   [END CROSS-CHAPTER BOX 11.1, FIGURE 1 HERE]
11
12
13   Numerous studies have compared the regional response to anthropogenic forcing at GWLs in annual and
14   seasonal mean values and extremes of different climate and impact variables across different multi-model
15   ensembles and/or different scenarios (e.g. Frieler et al., 2012; Schewe et al., 2014; Schleussner et al., 2016;
16   Seneviratne et al., 2016; Wartenburger et al., 2017; Dosio and Fischer, 2018; Tebaldi et al., 2020; (Herger et
17   al., 2015; Betts et al., 2018; Samset et al., 2019), see Sections 4.6.1, 8.5.3, 9.3.1, 9.5, 9.6.3, 10.4.3 and 11.2.4
18   for further details). The regional response patterns at given GWLs have been found to be consistent across
19   different scenarios for many climate variables (CC-Box 11.1 Fig.2) (Pendergrass et al., 2015; Seneviratne et
20   al., 2016; Wartenburger et al., 2017; Seneviratne and Hauser, 2020). The consistency tends to be higher for
21   temperature-related variables than for variables in the hydrological cycle or variables characterizing
22   atmospheric dynamics, and for intermediate to high emission scenarios than for low-emission scenarios (e.g.
23   for mean precipitation in the RCP2.6 scenario: Pendergrass et al., 2015; Wartenburger et al., 2017).
24   Nonetheless, CCB 11.1 Figure 2 illustrates that even for mean precipitation, which is known to be forcing-
25   dependent (Section 4.6.1 and Section 8.5.3), scenario differences in the response pattern at a given GWL are
26   smaller than model uncertainty and internal variability in many regions (Herger et al., 2015). The response
27   pattern is further found to be broadly consistent between models that reach a GWL relatively early and those
28   that reach it later under a given SSP (see CC Box 11.1 Fig.2 g, h)
29
30
31   [START CROSS-CHAPTER 11.1, FIGURE 2 HERE]
32
33   Cross-Chapter Box 11.1, Figure 2: (a-c) CMIP6 multi-model mean precipitation change at 2°C GWL (20-yr
34                                       mean) in three different SSP scenarios relative to 1850-1900. All models reaching
35                                       the corresponding GWL in the corresponding scenario are averaged. The number
36                                       of models averaged across is shown at the top right of the panel. The maps for the
37                                       other two SSP scenarios SSP1-1.9 (five models only) and SSP3-7.0 (not shown)
38                                       are consistent. (d-f) Same as (a-c) but for annual mean temperature. (g) Annual
39                                       mean temperature change at 2°C in CMIP6 models with high warming rate
40                                       reaching the GWL in the corresponding scenario before the earliest year of the
41                                       assessed very likely range (section 4.3.4) (h) Climate response at 2°C GWL across
42                                       all SSP1-1.9, SSP2-2.6, SSP2-4.5. SSP3-7.0 and SSP5-8.5 in all other models not
43                                       shown in (g). The good agreement of (g) and (h) demonstrate that the mean
44                                       temperature response at 2°C is not sensitive to the rate of warming and thereby the
45                                       GSAT warming of the respective models in 2081-2100. Uncertainty is represented
46                                       using the advanced approach: No overlay indicates regions with robust signal,
47                                       where ≥66% of models show change greater than variability threshold and ≥80%
48                                       of all models agree on sign of change; diagonal lines indicate regions with no
49                                       change or no robust signal, where <66% of models show a change greater than the
50                                       variability threshold; crossed lines indicate regions with conflicting signal, where
51                                       ≥66% of models show change greater than variability threshold and <80% of all
52                                       models agree on sign of change. For more information on the advanced approach,
53                                       please refer to the Cross-Chapter Box Atlas.1.
54
     Do Not Cite, Quote or Distribute                        11-35                                          Total pages: 345
     Final Government Distribution                       Chapter11                                      IPCC AR6 WGI

 1   [END CROSS-CHAPTER BOX 11.1, FIGURE 2 HERE]
 2
 3
 4   In contrast to linear pattern scaling (Mitchell, 2003; Collins et al., 2013a), the use of GWLs as a dimension
 5   of integration does not require linearity in the response of a climate variable. It is thus even useful for metrics
 6   which do not show a linear response, such as the frequency of heat extremes over land and oceans (Fischer
 7   and Knutti, 2015; Perkins-Kirkpatrick and Gibson, 2017; Frölicher et al., 2018; Kharin et al., 2018) if the
 8   relationship of the variable of interest to the GWL is scenario independent. The latter means that the
 9   response is independent of the pathway and relative contribution of various radiative forcings. For some
10   more complex indices like warm-spell duration or for regions with strong aerosol changes, discrepancies can
11   be larger (Wang et al., 2017d; King et al., 2018; Tebaldi et al., 2020) (see also subsection below on GWLs vs
12   scenarios for further caveats).
13
14   The limited scenario dependence of the GWL-based response for many variables implies that the regional
15   response to emissions scenarios can be split in almost independent contributions of 1) the transient global
16   warming response to scenarios (see Chapter 4), and 2) the regional response as function of a given GWL,
17   which has also been referred to as “regional climate sensitivity” (Seneviratne and Hauser, 2020). This
18   property has also been used to develop regionally-resolved emulators for global climate models, using global
19   surface temperature as input (Beusch et al., 2020; Tebaldi et al., 2020). Analyses of the CMIP6 and CMIP5
20   multi-model ensembles shows that the GWL-based responses are very similar for temperature and
21   precipitation extremes across the ensembles (Li et al., 2020a; Seneviratne and Hauser, 2020; Wehner, 2020).
22   This is despite their difference in global warming response (Chapter 4), confirming a substantial decoupling
23   between the two responses (global warming vs GWL-based regional response) for these variables. Thus, the
24   GWL approach isolates the uncertainty in the regional climate response from the global warming uncertainty
25   induced by scenario, global mean model response and internal variability (CCB Figure 1).
26
27   Mapping between GWL- and scenario-based responses in model analyses
28   To map scenario-based climate projections into changes at specific GWLs, first, all individual ESM
29   simulations that reach a certain GWL are identified. Second, the climate response patterns at the respective
30   GWL are calculated using an approach termed here “GWL-sampling approach” – sometimes also referred to
31   as epoch analysis, time shift, or time sampling approach –, taking into account all models and scenarios
32   (CCB Figure 3). Note that the range of years when a given GWL is reached in the CMIP6 ensemble is
33   different from the AR6 assessed range of projected global surface temperature (Table 4.5; Section 4.3.4).
34   The latter further takes into account different lines of evidence, including the assessed observed warming
35   between pre-industrial and present day, information from observational constraints on CMIP6, and emulators
36   using the assessed transient climate response (TCR) and equilibrium climate sensitivity (ECS) ranges
37   (Section 4.3.4). Hence the Chapter 4 assessed range (Table 4.5) is the reference to determine when a given
38   GWL is likely reached under given scenarios, while the mapping between scenarios/time frames and GWLs
39   is used to assess the respective regional responses happening at these time frames (which also allows to
40   account for the global surface temperature assessment rather than using scenarios analyses directly from
41   CMIP6 output).
42
43   In the model-based asssessment of Chapters 4, 8, 10, 11, 12 and the Atlas, the estimation of changes at
44   GWLs are generally defined as the 20-year time period in which the mean global surface air temperature
45   (GSAT; CCBox 2.3) first exceeds a certain anomaly relative to 1850-1900 (for simulations that start after
46   1850, relative to all years up to 1900 CCB Figure 3). The years when each individual model reaches a given
47   GWL for CMIP6 and CMIP5 can be found in Hauser et al. (2021). The changes at given GWLs are
48   identified for each ensemble member (for all scenarios) individually. Thereby, a given GWL is potentially
49   reached a few years earlier or later in different realizations of the same model due to internal variability, but
50   the temperature averaged across the 20-year period analysed in any simulation is consistent with the GWL.
51   Instead of blending the information from the different scenarios, the Interactive Atlas can be used to compare
52   the GWL spatial patterns and timings across the different scenarios (see Section Atlas 1.3.1).
     Do Not Cite, Quote or Distribute                      11-36                                        Total pages: 345
     Final Government Distribution                       Chapter11                                      IPCC AR6 WGI

 1
 2
 3   [START CROSS-CHAPTER BOX 11.1, FIGURE 3 HERE]
 4
 5   Cross-Chapter Box11.1, Figure 3: Illustration of the AR6 GWL sampling approach to derive the timing and the
 6                                     response at a given GWL for the case of CMIP6 data. For the mapping of
 7                                     scenarios/time slices into GWLs for CMIP6, please refer to Table 4.2. Respective
 8                                     numbers for the CMIP6 multi-model experiment are provided in the Chapter 11
 9                                     Supplementary Material (11.SM.1). Note that the time frames used to derived the
10                                     GWL time slices can also include different number of years (e.g. 30 years for
11                                     some analyses).
12
13
14   [END CROSS-CHAPTER BOX 11.1, FIGURE 3 HERE]
15
16
17   Mapping between GWL- and scenario-based responses for literature
18   A large fraction of the literature considers scenario-based analyses for given time slices. When GWL-based
19   information is required instead, an approximated mapping of the multi-model mean can be derived based on
20   the known GWL in the given experiments for a particular time period. As a rough approximation, CMIP6
21   multi-model mean projections for the near-term (2021-2040) correspond to changes at about 1.5°C, and
22   projections for the high-end scenario (SSP5-8.5) for the long-term (2081-2100) correspond to about 4-5°C of
23   global warming (see Table 4.2 for changes in the CMIP6 ensemble and the Chapter 11 Supplementary
24   Material (11.SM.1) and Hauser (2021) for details on other time periods and CMIP5). These approximated
25   changes are for instance used for some of the GWL-based assessments provided in the Chapter 11 regional
26   tables (Section 11.9; Table 11.3) when literature based on scenario projections is used to assessed estimated
27   changes at given GWLs.
28
29   GWLs vs scenarios
30   The use of scenarios remains a key element to inform mitigation decisions (Chapter 1, CCB1.4), to assess
31   which emission pathways are consistent with a certain GWL (CCB1.4 Figure1), to estimate when certain
32   GWLs are reached (Section 4.3.4), and to assess for which variables it is meaningful to use GWLs as a
33   dimension of integration. The use of scenarios is also essential for variables whose climate response strongly
34   depends on the contribution of radiative forcing (e.g. aerosols) and land use and land management changes,
35   and are time and warming rate dependent (e.g. sea level rise), or differ between transient and quasi-
36   equilibrium states. Furthermore, the use of concentration or emission-driven scenario simulations is required
37   if regional climate assessments need to account for the uncertainty in GSAT changes or climate-carbon
38   feedbacks.
39
40   Forcing dependence of the GWL response is found for global mean precipitation (Section 8.4.3), but less for
41   regional patterns of mean precipitation changes (CC-Box 11.1, Fig. 2). Limited dependence is found for
42   extremes, as highlighted above. In the cryosphere, elements that are quick to respond to warming like sea ice
43   area, permafrost, and snow show little scenario dependence (Chapter 9.3.1.1, 9.5.2.3, 9.5.3.3), whereas slow-
44   responding variables such as ice volumes of glaciers and ice sheets respond with a substantial delay and due
45   to their inertia, the response depends on when a certain GWL is reached. This also applies to some extent for
46   sea level rise where, for example, the contributions of melting glaciers and ice sheets depend on the pathway
47   followed to reach a given GWL (Chapter 9.6.3.4).
48
49   In addition to the lagged effect, the climate response at a given GWL may differ before and after a period of
50   overshoot, for example in the Atlantic Meridional Overturning Circulation (e.g. Palter et al. 2018). Finally,
51   as assessed in IPCC SR15, there is a difference in the response even for temperature-related variables if a
52   GWL is reached in a rapidly warming transient state or in an equilibrium state when the land-sea warming
53   contrast is less pronounced (e.g. King et al. 2020). However, in this report GWLs are used in the context of
54   projections for the 21st century when the climate response is mostly not in equilibrium and where projections
55   for many variables are less dependent on the pathway than for projections beyond 2100 (Section 9.6.3.4).
     Do Not Cite, Quote or Distribute                       11-37                                       Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1
 2   Key conclusions on assessments based on GWLs
 3
 4   GWL-based projections can inform society and policymakers on how climate would change under GWLs
 5   consistent with the aims of the Paris Agreement (stabilization at 1.5°C/well below 2°C), as well as on the
 6   consequences of missing these aims and reaching GWLs of 3°C or 4°C by the end of the century. The AR6
 7   assessment shows that every bit of global warming matters and that changes in global warming of 0.5°C lead
 8   to statistically significant changes in mean climate and climate extremes on global scale and for large regions
 9   (Sections 4.6.2, 11.2.4, 11.3, 11.4, 11.6, 11.9; Figs 11.8, 11.9, Atlas, Interactive Atlas), as also assessed in
10   the IPCC SR15.
11
12
13   [END CROSS-CHAPTER BOX 11.1 HERE]
14
15
16   11.3 Temperature extremes
17
18   This section assesses changes in temperature extremes at global, continental and regional scales. The main
19   focus is on the changes in the magnitude and frequency of moderate extreme temperatures (those that occur
20   several times a year) to very extreme temperatures (those that occur once-in-10-years or longer) of time
21   scales from a day to a season, though there is a strong emphasis on the daily scale where literature is most
22   concentrated.
23
24
25   11.3.1 Mechanisms and drivers
26
27   The SREX (IPCC, 2012) and AR5 (IPCC, 2014) concluded that greenhouse gas forcing is the dominant
28   factor for the increases in the intensity, frequency, and duration of warm extremes and the decrease in those
29   of cold extremes. This general global-scale warming is modulated by large-scale atmospheric circulation
30   patterns, as well as by feedbacks such as soil moisture-evapotranspiration-temperature and snow/ice-albedo-
31   temperature feedbacks, and local forcings such as land use change or changes in aerosol concentrations at the
32   regional and local scales (Box 11.1, Sections 11.1.5, 11.1.6). Therefore, changes in temperature extremes at
33   regional and local scales can have heterogeneous spatial distributions. Changes in the magnitudes (or
34   intensities) of extreme temperatures are often larger than changes in global surface temperature, because of
35   larger warming on land than on the ocean surface (2.3.1.1) and feedbacks, though they are of similar
36   magnitude to changes in the local mean temperature (Fig 11.2).
37
38   Extreme temperature events are associated with large-scale meteorological patterns (Grotjahn et al., 2016).
39   Quasi-stationary anticyclonic circulation anomalies or atmospheric blocking events are linked to temperature
40   extremes in many regions, such as in Australia (Parker et al., 2014; Perkins-Kirkpatrick et al., 2016), Europe
41   (Brunner et al., 2017, 2018; Schaller et al., 2018), Eurasia (Yao et al., 2017), Asia (Chen et al., 2016; Ratnam
42   et al., 2016; Rohini et al., 2016), and North America (Yu et al., 2018, 2019b; Zhang and Luo, 2019). Mid-
43   latitude planetary wave modulations affect short-duration temperature extremes such as heat waves (Perkins,
44   2015; Kornhuber et al., 2020). The large-scale modes of variability (Annex VI) affect the strength,
45   frequency, and persistence of these meteorological patterns and, hence, temperature extremes. For example,
46   cold and warm extremes in the mid-latitudes are associated with atmospheric circulation patterns such as the
47   Pacific-North American (PNA) pattern, as well as atmosphere-ocean coupled modes such as Pacific Decadal
48   Variability (PDV), the North Atlantic Oscillation (NAO), and Atlantic Multidecadal Variability (AMV)
49   (Kamae et al., 2014; Johnson et al., 2018; Ruprich-Robert et al., 2018; Yu et al., 2018, 2019a; Müller et al.,
50   2020; Section 11.1.5). Changes in the modes of variability in response to warming would therefore affect
51   temperature extremes (Clark and Brown, 2013; Horton et al., 2015). The level of confidence in those
52   changes, both in the observations and in future projections, varies, affecting the level of confidence in
53   changes in temperature extremes in different regions. As highlighted in Chapters 2-4 of this Report, it is
54   likely that there have been observational changes in the extratropical jets and mid-latitude jet meandering
55   (Section 2.3.1.4.3; Cross-Chapter Box 10.1). There is low confidence in possible effects of Arctic warming
     Do Not Cite, Quote or Distribute                     11-38                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   on mid-latitude temperature extremes (Cross-Chapter Box 10.1). A large portion of the multi-decadal
 2   changes in extreme temperature remains after the removal of the effect of these modes of variability and can
 3   be attributed to human influence (Kamae et al., 2017b; Wan et al., 2019). Thus, global warming dominates
 4   changes in temperature extremes at the regional scale and it is very unlikely that dynamic responses to
 5   greenhouse-gas induced warming would alter the direction of these changes.
 6
 7   Land-atmosphere feedbacks strongly modulate regional- and local-scale changes in temperature extremes
 8   (high confidence; Section 11.1.6; Seneviratne et al., 2013; Lemordant et al., 2016; Donat et al., 2017;
 9   Sillmann et al., 2017b; Hirsch et al., 2019). This effect is particularly notable in mid-latitude regions where
10   the drying of soil moisture amplifies high temperatures, in particular through increases in sensible heat flux
11   (Whan et al., 2015; Douville et al., 2016; Vogel et al., 2017). Land-atmosphere feedbacks amplifying
12   temperature extremes also include boundary-layer feedbacks and effects on atmospheric circulation
13   (Miralles et al., 2014a; Schumacher et al., 2019). Soil moisture-temperature feedbacks affect past and
14   present-day heat waves in observations and model simulations, both locally (Miralles et al. 2014; Hauser et
15   al. 2016; Meehl et al. 2016; Wehrli et al., 2019; Cowan et al., 2016) and beyond the regions of feedback
16   occurrence through changes in regional circulation patterns (Koster et al., 2016; Sato and Nakamura, 2019;
17   Stéfanon et al., 2014). The uncertainty due to the representation of land-atmosphere feedbacks in ESMs is a
18   cause of discrepancy between observations and simulations (Clark et al., 2006; Mueller and Seneviratne,
19   2014; Meehl et al., 2016). The decrease of plant transpiration or the increase of stomata resistance under
20   enhanced CO2 concentrations is a direct CO2 forcing of land temperatures (warming due to reduced
21   evaporative cooling), which contributes to higher warming on land (Lemordant et al., 2016; Vicente-Serrano
22   et al., 2020c). The snow/ice-albedo feedback plays an important role in amplifying temperature variability in
23   the high latitudes (Diro et al. 2018) and can be the largest contributor to the rapid warming of cold extremes
24   in the mid- and high latitudes of the Northern Hemisphere (Gross et al., 2020).
25
26   Regional external forcings, including land-use changes and emissions of anthropogenic aerosols, play an
27   important role in the changes of temperature extremes in some regions (high confidence, Section 11.1.6).
28   Deforestation may have contributed to about one third of the warming of hot extremes in some mid-latitude
29   regions since the pre-industrial time (Lejeune et al., 2018). Aspects of agricultural practice, including no-till
30   farming, irrigation, and overall cropland intensification, may cool hot temperature extremes (Davin et al.,
31   2014; Mueller et al., 2016b). For instance, cropland intensification has been suggested to be responsible for a
32   cooling of the highest temperature percentiles in the US Midwest (Mueller et al., 2016b). Irrigation has been
33   shown to be responsible for a cooling of hot temperature extremes of up to 1-2°C in many mid-latitude
34   regions in the present climate (Thiery et al., 2017; Thiery et al., 2020), a process not represented in most of
35   state-of-the-art ESMs (CMIP5, CMIP6). Double cropping may have led to increased hot extremes in the
36   inter-cropping season in part of China (Jeong et al., 2014). Rapid increases in summertime warming in
37   western Europe and northeast Asia since the 1980s are linked to a reduction in anthropogenic aerosol
38   precursor emissions over Europe (Dong et al., 2016, 2017; Nabat et al., 2014), in addition to the effect of
39   increased greenhouse gas forcing (see also Chapter 10, Section 10.1.3.1). This effect of aerosols on
40   temperature-related extremes is also noted for declines in short-lived anthropogenic aerosol emissions over
41   North America (Mascioli et al., 2016). On the local scale, the urban heat island (UHI) effect results in higher
42   temperatures in urban areas than in their surrounding regions and contributes to warming in regions of rapid
43   urbanization, in particular for night-time temperature extremes (Box 10.3; Phelan et al., 2015; Chapman et
44   al., 2017; Sun et al., 2019). But these local and regional forcings are generally not (well-) represented in the
45   CMIP5 and CMIP6 simulations (see also Section 11.3.3), contributing to uncertainty in model simulated
46   changes.
47
48   In summary, greenhouse gas forcing is the dominant driver leading to the warming of temperature extremes.
49   At regional scales, changes in temperature extremes are modulated by changes in large-scale patterns and
50   modes of variability, feedbacks including soil moisture-evapotranspiration-temperature or snow/ice-albedo-
51   temperature feedbacks, and local and regional forcings such as land use and land cover changes, or aerosol
52   concentrations, and decadal and multidecadal natural variability. This leads to heterogeneity in regional
53   changes and their associated uncertainties (high confidence). Urbanization has exacerbated the effects of
54   global warming in cities, in particular for night-time temperature extremes (high confidence).
55
     Do Not Cite, Quote or Distribute                      11-39                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1
 2   11.3.2 Observed trends
 3
 4   The SREX (IPCC, 2012) reported a very likely decrease in the number of cold days and nights and increase
 5   in the number of warm days and nights at the global scale. Confidence in trends was assessed as regionally
 6   variable (low to medium confidence) due to either a lack of observations or varying signals in sub-regions.
 7
 8   Since SREX (IPCC, 2012) and AR5 (IPCC, 2014), many regional-scale studies have examined trends in
 9   temperature extremes using different metrics that are based on daily temperatures, such as the
10   CCl/WCRP/JCOMM Expert Team on Climate Change Detection and Indices (ETCCDI) indices (Dunn et
11   al., 2020). The additional observational records, along with a stronger warming signal, show very clearly that
12   changes observed at the time of AR5 (IPCC, 2014) continued, providing strengthened evidence of an
13   increase in the intensity and frequency of hot extremes and decrease in the intensity and frequency of cold
14   extremes. While the magnitude of the observed trends in temperature-related extremes varies depending on
15   the region, spatial and temporal scales, and metric assessed, evidence of a warming effect is overwhelming,
16   robust, and consistent. In particular, an increase in the intensity and frequency of hot extremes is almost
17   always associated with an increase in the hottest temperatures and in the number of heatwave days. It is also
18   the case for changes in cold extremes. For this reason, and to simplify the presentation, the phrase “increase
19   in the intensity and frequency of hot extremes” is used to represent, collectively, an increase in the
20   magnitude of extreme day and/or night temperatures, in the number of warm days and/or nights, and in the
21   number of heat wave days. Changes in cold extremes are assessed similarly.
22
23   On the global scale, evidence of an increase in the number of warm days and nights and a decrease in the
24   number of cold days and nights, and an increase in the coldest and hottest extreme temperatures is very
25   robust and consistent among all variables. Figure 11.2 displays timeseries of globally-averaged annual
26   maximum daily maximum (TXx) and annual minimum daily minimum temperature (TNn) on land. Warming
27   of land mean TXx is similar to the mean land warming, which is about 45% higher than global warming
28   (Section 2.3.1). Warming of land mean TNn is even higher, with about 3°C of warming since 1960 (Figure
29   11.2). Figure 11.9 shows maps of linear trends over 1960-2018 in the annual maximum daily maximum
30   (TXx), the annual minimum daily minimum temperature (TNn), and frequency of warm days (TX90p). The
31   maps for TXx and TNn show trends consistent with overall warming in most regions, with a particularly
32   high warming of TXx in Europe and north-western South America, and a particularly high warming of TNn
33   in the Arctic. Consistent with the observed warming in global surface temperature (2.3.1.2) and the observed
34   trends in TXx and TNn, the frequency of TX90p has increased while that of cold nights (TN10p) has
35   decreased since the 1950s: Nearly all land regions showed statistically significant decreases in TN10p
36   (Alexander, 2016; Dunn et al., 2020), though trends in TX90p are variable with some decreases in southern
37   South America, mainly during austral summer (Rusticucci et al., 2017). A decrease in the number of cold
38   spell days is also observed over nearly all land surface areas (Easterling et al., 2016) and in the northern mid-
39   latitudes in particular (van Oldenborgh et al., 2019). These observed changes are also consistent when a new
40   global land surface daily air temperature dataset is analyzed (Zhang et al., 2019c). Consistent warming trends
41   in temperature extremes globally, and in most land areas, over the past century are also found in a range of
42   observation-based data sets (Fischer and Knutti, 2014; Donat et al., 2016a; Dunn et al., 2020), with the
43   extremes related to daily minimum temperatures changing faster than those related to daily maximum
44   temperatures (Dunn et al., 2020) (Fig. 11.2). Seasonal variations in trends in temperature-related extremes
45   have been demonstrated. A warming in warm-season temperature extremes is detected, even during the
46   “slower surface global warming” period from the late 1990s to early 2010s (Cross-Chapter Box 3.1) (Kamae
47   et al., 2014; Seneviratne et al., 2014; Imada et al., 2017). Many studies of past changes in temperature
48   extremes for particular regions or countries show trends consistent with this global picture, as summarized
49   below and in Tables 11.4, 11.7, 11.10, 11.13, 11.16 and 11.19 in Section 11.9.
50
51
52   [START FIGURE 11.9 HERE]
53
54   Figure 11.9: Linear trends over 1960-2018 in the annual maximum daily maximum temperature (TXx, a), the annual
55                minimum daily minimum temperature (TNn, b), and the annual number of days when daily maximum

     Do Not Cite, Quote or Distribute                     11-40                                       Total pages: 345
     Final Government Distribution                         Chapter11                                        IPCC AR6 WGI

 1                temperature exceeds its 90th percentile from a base period of 1961-1990 (TX90p, c), based on the
 2                HadEX3 data set (Dunn et al., 2020). Linear trends are calculated only for grid points with at least 66% of
 3                the annual values over the period and which extend to at least 2009. Areas without sufficient data are
 4                shown in grey. No overlay indicates regions where the trends are significant at p = 0.1 level. Crosses
 5                indicate regions where trends are not significant. For details on the methods see Supplementary Material
 6                11.SM.2. Further details on data sources and processing are available in the chapter data table (Table
 7                11.SM.9).
 8
 9   [END FIGURE 11.9 HERE]
10
11
12   In Africa (Table 11.4), while it is difficult to assess changes in temperature extremes in parts of the continent
13   because of a lack of data, evidence of an increase in the intensity and frequency of hot extremes and decrease
14   in the intensity and frequency of cold extremes is clear and robust in regions where data are available. These
15   include an increase in the frequency of warm days and nights and a decrease in the frequency of cold days
16   and nights with high confidence (Donat et al., 2013b, 2014b; Kruger and Sekele, 2013; Chaney et al., 2014;
17   Filahi et al., 2016; Moron et al., 2016; Ringard et al., 2016; Barry et al., 2018; Gebrechorkos et al., 2018) and
18   an increase in heat waves (Russo et al., 2016; Ceccherini et al., 2017). The increase in TNn is more notable
19   than in TXx (Figure 11.9). Cold spells occasionally strike subtropical areas, but are likely to have decreased
20   in frequency (Barry et al., 2018). The frequency of cold events has likely decreased in South Africa (Song et
21   al., 2014; Kruger and Nxumalo, 2017), North Africa (Driouech et al., 2021; Filahi et al., 2016), and the
22   Sahara (Donat et al., 2016a). Over the whole continent, there is medium confidence in an increase in the
23   intensity and frequency of hot extremes and decrease in the intensity and frequency of cold extremes; it is
24   likely that similar changes have also occurred in areas with poor data coverage, as warming is widespread
25   and as projected future changes are similar over all regions (11.3.5).
26
27   In Asia (Table 11.7), there is very robust evidence for a very likely increase in the intensity and frequency of
28   hot extremes and decrease in the intensity and frequency of cold extremes in recent decades. This is clear in
29   global studies (e.g. Alexander, 2016; Dunn et al., 2020), as well as in numerous regional studies (Table
30   11.7). The area fraction with extreme warmth in Asia increased during 1951–2016 (Imada et al., 2018). The
31   frequency of warm extremes increased and the frequency of cold extremes decreased in East Asia (Zhou et
32   al., 2016a; Chen and Zhai, 2017; Yin et al., 2017; Lee et al., 2018c; Qian et al., 2019) and west Asia (Acar
33   Deniz and Gönençgil, 2015; Erlat and Türkeş, 2016; Imada et al., 2017; Rahimi et al., 2018; Rahimi and
34   Hejabi, 2018) with high confidence. The duration of heat extremes has also lengthened in some regions, for
35   example, in southern China (Luo and Lau, 2016), but there is medium confidence of heat extremes increasing
36   in frequency in South Asia (AlSarmi and Washington, 2014; Sheikh et al., 2015; Mazdiyasni et al., 2017;
37   Zahid et al., 2017; Nasim et al., 2018; Khan et al., 2019; Roy, 2019). Warming trends in daily temperature
38   extremes indices have also been observed in central Asia (Hu et al., 2016; Feng et al., 2018), the Hindu Kush
39   Himalaya (Sun et al., 2017), and Southeast Asia (Supari et al., 2017; Cheong et al., 2018). The intensity and
40   frequency of cold spells in all Asian regions have been decreasing since the beginning of the 20th century
41   (high confidence) (Sheikh et al., 2015; Donat et al., 2016a; Dong et al., 2018; van Oldenborgh et al., 2019).
42
43   In Australasia (Table 11.10), there is very robust evidence for very likely increases in the number of warm
44   days and warm nights and decrease in the number of cold days and cold nights since 1950 (Lewis and King,
45   2015; Jakob and Walland, 2016; Alexander and Arblaster, 2017). The increase in extreme minimum
46   temperatures occurs in all seasons over most of Australia and typically exceeds the increase in extreme
47   maximum temperatures (Wang et al., 2013b; Jakob and Walland, 2016). However, some parts of southern
48   Australia have shown stable or increased numbers of frost days since the 1980s (Dittus et al., 2014) (see also
49   Section 11.3.4). Similar positive trends in extreme minimum and maximum temperatures have been
50   observed in New Zealand, in particular in the autumn-winter seasons, although they generally show higher
51   spatial variability (Caloiero, 2017). In the tropical Western Pacific region, spatially coherent warming trends
52   in maximum and minimum temperature extremes have been reported for the period of 1951–2011 (Whan et
53   al., 2014; McGree et al., 2019).
54
55   In Central and South America (Table 11.13), there is high confidence that observed hot extremes (TN90p,
56   TX90p) have increased and cold extremes (TN10p, TX10p) have decreased over recent decades, though
     Do Not Cite, Quote or Distribute                        11-41                                          Total pages: 345
     Final Government Distribution                     Chapter11                                    IPCC AR6 WGI

 1   trends vary among different extremes types, datasets, and regions (Dereczynski et al., 2020; Dittus et al.,
 2   2016; Dunn et al., 2020; Meseguer-Ruiz et al., 2018; Olmo et al., 2020; Rusticucci et al., 2017; Salvador and
 3   de Brito, 2018; Skansi et al., 2013). An increase in the intensity and frequency of heatwave events was also
 4   observed between 1961 and 2014, in an area covering most of South America (Ceccherini et al., 2016;
 5   Geirinhas et al., 2018). However, there is medium confidence that warm extremes (TXx and TX90p) have
 6   decreased in the last decades over the central region of SES during austral summer (Tencer, B.; Rusticucci,
 7   2012; Skansi et al., 2013; Rusticucci et al., 2017; Wu and Polvani, 2017). There is medium confidence that
 8   TNn extremes are increasing faster than TXx extremes, with the largest warming rates observed over
 9   Northeast Brazil (NEB) and North South America (NSA) for cold nights (Skansi et al., 2013).
10
11   In Europe (Table 11.16), there is very robust evidence for a very likely increase in maximum temperatures
12   and the frequency of heat waves. The increase in the magnitude and frequency of high maximum
13   temperatures has been observed consistently across regions including in central (Twardosz and Kossowska-
14   Cezak, 2013; Christidis et al., 2015; Lorenz et al., 2019) and southern Europe (Croitoru and Piticar, 2013; El
15   Kenawy et al., 2013; Christidis et al., 2015; Nastos and Kapsomenakis, 2015; Fioravanti et al., 2016; Ruml et
16   al., 2017). In northern Europe, a strong increase in extreme winter warming events has been observed
17   (Matthes et al., 2015; Vikhamar-Schuler et al., 2016). Temperature observations for wintertime cold spells
18   show a long-term decreasing frequency in Europe (Brunner et al., 2018; van Oldenborgh et al., 2019), and
19   typical cold spells such as that observed during the 2009/2010 winter had an occurrence probability that is
20   twice smaller currently than if climate change had not occurred (Christiansen et al., 2018).
21
22   In North America (Table 11.19), there is very robust evidence for a very likely increase in the intensity and
23   frequency of hot extremes and decrease in the intensity and frequency of cold extremes for the whole
24   continent, though there are substantial spatial and seasonal variations in the trends. Minimum temperatures
25   display warming consistently across the continent, while there are more contrasting trends in the annual
26   maximum daily temperatures in parts of the USA (Figure 11.9) (Lee et al., 2014; van Oldenborgh et al.,
27   2019; Dunn et al., 2020). In Canada, there is a clear increase in the intensity and frequency of hot extremes
28   and decrease in the intensity and frequency of cold extremes (Vincent et al., 2018). In Mexico, a clear
29   warming trend in TNn was found, particularly in the northern arid region (Montero-Martínez et al., 2018).
30   The number of warm days has increased and the number of cold days has decreased (García-Cueto et al.,
31   2019). Cold spells have undergone a reduction in magnitude and intensity in all regions of North America
32   (Bennett and Walsh, 2015; Donat et al., 2016a; Grotjahn et al., 2016; Vose et al., 2017; García-Cueto et al.,
33   2019; van Oldenborgh et al., 2019).
34
35   Extreme heat events have increased around the Arctic since 1979, particularly over Arctic North America
36   and Greenland (Matthes et al., 2015; Dobricic et al., 2020), which is consistent with summer melt (9.4.1).
37   Observations north of 60˚N show increases in wintertime warm days and nights over 1979-2015, while cold
38   days and nights declined (Sui et al., 2017). Extreme heat days are particularly strong in winter, with
39   observations showing the warmest mid-winter temperatures at the North Pole rising at twice the rate of mean
40   temperature (Moore, 2016), as well as increases in Arctic winter warm days (T>-10℃) (Vikhamar-Schuler et
41   al., 2016; Graham et al., 2017). Arctic annual minimum temperatures have increased at about three times the
42   rate of global surface temperature since the 1960s (Figs. 11.2, 11.9), consistent with the observed mean cold
43   season (October-May) warming of 3.1°C in the region (Atlas 11.2).
44
45   Trends in some measures of heat waves are also observed at the global scale. Globally-averaged heat wave
46   intensity, heat wave duration, and the number of heat wave days have significantly increased from 1950-
47   2011 (Perkins, 2015). There are some regional differences in trends in characteristics of heat waves with
48   significant increases reported in Europe (Russo et al., 2015; Forzieri et al., 2016; Sánchez-Benítez et al.,
49   2020) and Australia (CSIRO and BOM, 2016; Alexander and Arblaster, 2017). In Africa, there is medium
50   confidence that heat waves, regardless of the definition, have been becoming more frequent, longer-lasting,
51   and hotter over more than three decades (Fontaine et al., 2013; Mouhamed et al., 2013; Ceccherini et al.,
52   2016, 2017; Forzieri et al., 2016; Moron et al., 2016; Russo et al., 2016). The majority of heat wave
53   characteristics examined in China between 1961-2014 show increases in heat wave days, consistent with
54   warming (You et al., 2017; Xie et al., 2020). Increases in the frequency and duration of heat waves are also
55   observed in Mongolia (Erdenebat and Sato, 2016) and India (Ratnam et al., 2016; Rohini et al., 2016). In the
     Do Not Cite, Quote or Distribute                     11-42                                     Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   UK, the lengths of short heat waves have increased since the 1970s, while the lengths of long heat waves
 2   (over 10 days) have decreased over some stations in the southeast of England (Sanderson et al., 2017b). In
 3   Central and South America, there are increases in the frequency of heat waves (Barros et al., 2015;
 4   Bitencourt et al., 2016; Ceccherini et al., 2016; Piticar, 2018), although decreases in Excess Heat Factor
 5   (EHF), which is a metric for heat wave intensity, are observed in South America in data derived from
 6   HadGHCND (Cavanaugh and Shen, 2015).
 7
 8   In summary, it is virtually certain that there has been an increase in the number of warm days and nights and
 9   a decrease in the number of cold days and nights on the global scale since 1950. Both the coldest extremes
10   and hottest extremes display increasing temperatures. It is very likely that these changes have also occurred
11   at the regional scale in Europe, Australasia, Asia, and North America. It is virtually certain that there has
12   been increases in the intensity and duration of heat waves and in the number of heat wave days at the global
13   scale. These trends likely occur in Europe, Asia, and Australia. There is medium confidence in similar
14   changes in temperature extremes in Africa and high confidence in South America; the lower confidence is
15   due to reduced data availability and fewer studies. Annual minimum temperatures on land have increased
16   about three times more than global surface temperature since the 1960s, with particularly strong warming in
17   the Arctic (high confidence).
18
19
20   11.3.3 Model evaluation
21
22   AR5 assessed that CMIP3 and CMIP5 models generally captured the observed spatial distributions of the
23   mean state and that the inter-model range of simulated temperature extremes was similar to the spread
24   estimated from different observational datasets; the models generally captured trends in the second half of
25   the 20th century for indices of extreme temperature, although they tended to overestimate trends in hot
26   extremes and underestimate trends in cold extremes (Flato et al., 2013). Post-AR5 studies on the CMIP5
27   models’ performance in simulating mean and changes in temperature extremes continue to support the AR5
28   assessment (Fischer and Knutti, 2014; Sillmann et al., 2014; Ringard et al., 2016; Borodina et al., 2017b;
29   Donat et al., 2017; Di Luca et al., 2020a). Over Africa, the observed warming in temperature extremes is
30   captured by CMIP5 models, although it is underestimated in west and central Africa (Sherwood et al., 2014;
31   Diedhiou et al., 2018). Over East Asia, the CMIP5 ensemble performs well in reproducing the observed
32   trend in temperature extremes averaged over China (Dong et al., 2015). Over Australia, the multi-model
33   mean performs better than individual models in capturing observed trends in gridded station based ETCCDI
34   temperature indices (Alexander and Arblaster, 2017).
35
36   Initial analyses of CMIP6 simulations (Chen et al., 2020a; Di Luca et al., 2020b; Kim et al., 2020; Li et al.,
37   2020a; Thorarinsdottir et al., 2020; Wehner et al., 2020) indicate the CMIP6 models perform similarly to the
38   CMIP5 models regarding biases in hot and cold extremes. In general, CMIP5 and CMIP6 historical
39   simulations are similar in their performance in simulating the observed climatology of extreme temperatures
40   (high confidence). The general warm bias in hot extremes and cold bias in cold extremes reported for CMIP5
41   models (Kharin et al., 2013; Sillmann et al., 2013a) remain in CMIP6 models (Di Luca et al., 2020b).
42   However, there is some evidence that CMIP6 models better represent some of the underlying processes
43   leading to extreme temperatures, such as seasonal and diurnal variability and synoptic-scale variability (Di
44   Luca et al., 2020b). Whether these improvements are sufficient to enhance our understanding of past changes
45   or to reduce uncertainties in future projections remains unclear. The relative error estimates in the simulation
46   of various indices of temperature extremes in the available CMIP6 models show that no single model
47   performs the best on all indices and the multi-model ensemble seems to out-perform any individual model
48   due to its reduction in systematic bias (Kim et al., 2020). Figure 11.10 show errors in the 1979-2014 average
49   annual TXx and annual TNn simulated by available CMIP6 models in comparison with HadEX3 and ERA5
50   (Li et al., 2020; Kim et al., 2020, Wehner et al., 2020). While the magnitude of the model error depends on
51   the reference data set, the model evaluations drawn from different reference data sets are quite similar. In
52   general, models reproduce the spatial patterns and magnitudes of both cold and hot temperature extremes
53   quite well. There are also systematic biases. Hot extremes tend to be too cool in mountainous and high-
54   latitude regions, but too warm in the eastern United States and South America. For cold extremes, CMIP6
55   models are too cool, except in northeastern Eurasia and the southern mid-latitudes. Errors in seasonal mean
     Do Not Cite, Quote or Distribute                     11-43                                      Total pages: 345
     Final Government Distribution                        Chapter11                                       IPCC AR6 WGI

 1   temperatures are uncorrelated with errors in extreme temperatures and are often of opposite sign (Wehner et
 2   al., 2020).
 3
 4
 5   [START FIGURE 11.10 HERE]
 6
 7   Figure 11.10:Multi-model mean bias in temperature extremes (°C ) for the period 1979-2014, calculated as the
 8                difference between the CMIP6 multi-model mean and the average of observations from the values
 9                available in HadEX3 for (a) the annual hottest temperature (TXx) and (b) the annual coldest temperature
10                (TNn). Areas without sufficient data are shown in grey. Adapted from Wehner et al. (2020) under the
11                terms of the Creative Commons Attribution license. Further details on data sources and processing are
12                available in the chapter data table (Table 11.SM.9).
13
14   [END FIGURE 11.10 HERE]
15
16
17   Atmospheric model (AMIP) simulations are often used in event attribution studies to assess the influence of
18   global warming on observed temperature-related extremes. These simulations typically capture the observed
19   trends in temperature extremes, though some regional features, such as the lack of warming in daytime warm
20   temperature extremes over South America and parts of North America, are not reproduced in the model
21   simulations (Dittus et al., 2018), possibly due to internal variability, deficiencies in local surface processes,
22   or forcings that are not represented in the SSTs. Additionally, the AMIP models assessed tend to produce
23   overly persistent heat wave events. This bias in the duration of the events does not impact the reliability of
24   the models’ positive trends (Freychet et al., 2018).
25
26   Several regional climate models (RCMs) have also been evaluated in terms of their performance in
27   simulating the climatology of extremes in various regions of the Coordinated Regional Downscaling
28   Experiment (CORDEX) (Giorgi et al., 2009), especially in East Asia (Ji and Kang, 2015; Yu et al., 2015;
29   Park et al., 2016; Bucchignani et al., 2017; Gao et al., 2017a; Niu et al., 2018; Sun et al., 2018b; Wang et al.,
30   2018a), Europe (Cardoso et al., 2019; Gaertner et al., 2018; Jacob et al., 2020; Kim et al., 2020; Lorenz et
31   al., 2019; Smiatek et al., 2016; Vautard et al., 2013; Vautard et al., 2020b), and Africa (Kim et al., 2014b;
32   Diallo et al., 2015; Dosio, 2017; Samouly et al., 2018; Mostafa et al., 2019). Compared to GCMs, RCM
33   simulations show an added value in simulating temperature-related extremes, though this depends on
34   topographical complexity and the parameters employed (see Section 10.3.3). The improvement with
35   resolution is noted in East Asia (Park et al., 2016; Zhou et al., 2016b; Shi et al., 2017; Hui et al., 2018).
36   However, in the European CORDEX ensemble, different aerosol climatologies with various degrees of
37   complexity were used in projections (Bartók et al., 2017; Lorenz et al., 2019) and the land surface models
38   used in the RCMs do not account for physiological CO2 effects on photosynthesis leading to enhanced water-
39   use efficiency and decreased evapotranspiration (Schwingshackl et al., 2019), which could lead to biases in
40   the representation of temperature extremes in these projections (Boé et al., 2020). In addition, there are key
41   cold biases in temperature extremes over areas with complex topography (Niu et al., 2018). Over North
42   America, 12 RCMs were evaluated over the ARCTIC-CORDEX region (Diaconescu et al., 2018). Models
43   were able to simulate well climate indices related to mean air temperature and hot extremes over most of the
44   Canadian Arctic, with the exception of the Yukon region where models displayed the largest biases related to
45   topographic effects. Two RCMs were evaluated against observed extremes indices over North America over
46   the period 1989–2009, with a cool bias in minimum temperature extremes shown in both RCMs (Whan and
47   Zwiers, 2016). The most significant biases are found in TXx and TNn, with fewer differences in the
48   simulation of annual minimum daily maximum temperature (TXn) and annual maximum daily minimum
49   temperature (TNx) in central and western North America. Over Central and South America, maximum
50   temperatures from the Eta RCM are generally underestimated, although hot days, warm nights, and heat
51   waves are increasing in the period 1961-1990, in agreement with observations (Chou et al., 2014b; Tencer et
52   al., 2016; Bozkurt et al., 2019).
53
54   Some land forcings are not well represented in climate models. As highlighted in the IPCC SRCCL Ch2,
55   there is high agreement that temperate deforestation leads to summer warming and winter cooling (Bright et
56   al., 2017; Zhao and Jackson, 2014; Gálos et al., 2011, 2013; Wickham et al., 2013; Ahlswede and Thomas,
     Do Not Cite, Quote or Distribute                        11-44                                        Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   2017; Anderson-Teixeira et al., 2012; Anderson et al., 2011; Chen et al., 2012; Strandberg and Kjellström,
 2   2019), which has substantially contributed to the warming of hot extremes in the northern mid-latitudes over
 3   the course of the 20th century (Lejeune et al., 2018) and in recent years (Strandberg and Kjellström, 2019).
 4   However, observed forest effects on the seasonal and diurnal cycle of temperature are not well captured in
 5   several ESMs: while observations show a cooling effect of forest cover compared to non-forest vegetation
 6   during daytime (Li et al., 2015), in particular in arid, temperate, and tropical regions (Alkama and Cescatti,
 7   2016), several ESMs simulate a warming of daytime temperatures for regions with forest vs non-forest cover
 8   (Lejeune et al., 2017). Also irrigation effects, which can lead to regional cooling of temperature extremes,
 9   are generally not integrated in current-generations of ESMs (Section 11.3.1).
10
11   In summary, there is high confidence that climate models can reproduce the mean state and overall warming
12   of temperature extremes observed globally and in most regions, although the magnitude of the trends may
13   differ. The ability of models to capture observed trends in temperature-related extremes depends on the
14   metric evaluated, the way indices are calculated, and the time periods and spatial scales considered. Regional
15   climate models add value in simulating temperature-related extremes over GCMs in some regions. Some
16   land forcings on temperature extremes are not well captured (effects of deforestation) or generally not
17   representated (irrigation) in ESMs.
18
19
20   11.3.4 Detection and attribution, event attribution
21
22   SREX (IPCC, 2012) assessed that it is likely anthropogenic influences have led to the warming of extreme
23   daily minimum and maximum temperatures at the global scale. AR5 concluded that human influence has
24   very likely contributed to the observed changes in the intensity and frequency of daily temperature extremes
25   on the global scale in the second half of the 20th century (IPCC, 2014). With regard to individual, or
26   regionally- or locally-specific events, AR5 concluded that it is likely human influence has substantially
27   increased the probability of occurrence of heat waves in some locations.
28
29   Studies since AR5 continue to attribute the observed increase in the frequency or intensity of hot extremes
30   and the observed decrease in the frequency or intensity of cold extremes to human influence, dominated by
31   anthropogenic greenhouse gas emissions, on global and continental scales, and for many AR6 regions. These
32   include attribution of changes in the magnitude of annual TXx, TNx, TXn, and TNn, based on different
33   observational data sets including, HadEX2 and HadEX3, CMIP5 and CMIP6 simulations, and different
34   statistical methods (Kim et al., 2016; Wang et al., 2017c; Seong et al., 2020). As is the case for an increase in
35   mean temperature (3.3.1), an increase in extreme temperature is mostly due to greenhouse gas forcing, off-
36   set by aerosol forcing. The aerosols’ cooling effect is clearly detectable over Europe and Asia (Seong et al.,
37   2020). As much as 75% of the moderate daily hot extremes (above 99.9th percentile) over land are due to
38   anthropogenic warming (Fischer and Knutti, 2015). New results are found to be more robust due to the
39   extended period that improves the signal-to-noise ratio. The effect of anthropogenic forcing is clearly
40   detectable and attributable in the observed changes in these indicators of temperature extremes, even at
41   country and sub-country scales, such as in Canada (Wan et al., 2019). Changes in the number of warm
42   nights, warm days, cold nights, and cold days, and other indicators such as the Warm Spell Duration Index
43   (WSDI), are also attributed to anthropogenic influence (Hu et al., 2020; Christidis and Stott, 2016).
44
45   Regional studies, including for Asia (Dong et al., 2018; Lu et al., 2018), Australia (Alexander and Arblaster,
46   2017), and Europe (Christidis and Stott, 2016), found similar results. A clear anthropogenic signal is also
47   found in the trends in the Combined Extreme Index (CEI) for North America, Asia, Australia, and Europe
48   (Dittus et al., 2016). While various studies have described increasing trends in several heat wave metrics
49   (HWD, HWA, EHF, etc.) in different regions (e.g., Bandyopadhyay et al., 2016; Cowan et al., 2014;
50   Sanderson et al., 2017), few recent studies have explicitly attributed these changes to causes; most of them
51   stated that observed trends are consistent with anthropogenic warming. The detected anthropogenic signals
52   are clearly separable from the response to natural forcing, and the results are generally insensitive to the use
53   of different model samples, as well as different data availability, indicating robust attribution. Studies of
54   monthly, seasonal, and annual records in various regions (Kendon, 2014; Lewis and King, 2015; Bador et al.,
55   2016; Meehl et al., 2016; Zhou et al., 2019a) and globally (King, 2017) show an increase in the breaking of
     Do Not Cite, Quote or Distribute                      11-45                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   hot records and a decrease in the breaking of cold records (King, 2017). Changes in anthropogenically-
 2   attributable record-breaking rates are noted to be largest over the Northern Hemisphere land areas (Shiogama
 3   et al., 2016). Yin and Sun (2018) found clear evidence of an anthropogenic signal in the changes in the
 4   number of frost and icing days, when multiple model simulations were used. In some key wheat-producing
 5   regions of southern Australia, increases in frost days or frost season length have been reported (Dittus et al.,
 6   2014; Crimp et al., 2016); these changes are linked to decreases in rainfall, cloud-cover, and subtropical
 7   ridge strength, despite an overall increase in regional mean temperatures (Dittus et al., 2014; Pepler et al.,
 8   2018).
 9
10   A significant advance since AR5 has been a large number of studies focusing on extreme temperature events
11   at monthly and seasonal scales, using various extreme event attribution methods. Diffenbaugh et al. (2017)
12   found anthropogenic warming has increased the severity and probability of the hottest month over >80% of
13   the available observational area on the global scale. Christidis and Stott (2014) provide clear evidence that
14   warm events have become more probable because of anthropogenic forcings. Sun et al. (2014) found human
15   influence has caused a more than 60-fold increase in the probability of the extreme warm 2013 summer in
16   eastern China since the 1950s. Human influence is found to have increased the probability of the historically
17   hottest summers in many regions of the world, both in terms of mean temperature (Mueller et al., 2016a) and
18   wet-bulb globe temperature (WBGT) (Li et al., 2017a). In most regions of the Northern Hemisphere,
19   changes in the probability of extreme summer average WBGT were found to be about an order of magnitude
20   larger than changes in the probability of extreme hot summers estimated by surface air temperature (Li et al.,
21   2017a). In addition to these generalised, global-scale approaches, extreme event studies have found an
22   attributable increase in the probability of hot annual and seasonal temperatures in many locations, including
23   Australia (Knutson et al., 2014a; Lewis and Karoly, 2014), China (Sun et al., 2014; Sparrow et al., 2018;
24   Zhou et al., 2020), Korea (Kim et al., 2018c) and Europe (King et al., 2015b).
25
26   There have also been many extreme event attribution studies that examined short duration temperature
27   extremes, including daily temperatures, temperature indices, and heat wave metrics. Examples of these
28   events from different regions are summarised in various annual Explaining Extreme Events supplements of
29   the Bulletin of the American Meteorological Society (Peterson et al., 2012, 2013b, Herring et al., 2014,
30   2015, 2016, 2018, 2019, 2020), including a number of approaches to examine extreme events (described in
31   Easterling et al., 2016; Otto, 2017; Stott et al., 2016). Several studies of recent events from 2016 onwards
32   have determined an infinite risk ratio (fraction of attributable risk (FAR) of 1), indicating the occurrence
33   probability for such events is close to zero in model simulations without anthropogenic influences (see
34   Herring et al., 2018, 2019, 2020; Imada et al., 2019; Vogel et al., 2019). Though it is difficult to accurately
35   estimate the lower bound of the uncertainty range of the FAR in these cases (Paciorek et al., 2018), the fact
36   that those events are so far outside the envelop of the models with only natural forcing indicates that it is
37   extremely unlikely for those events to occur without human influence.
38
39   Studies that focused on the attributable signal in observed cold extreme events show human influence
40   reducing the probability of those events. Individual attribution studies on the extremely cold winter of 2011
41   in Europe (Peterson et al., 2012), in the eastern US during 2014 and 2015 (Trenary et al., 2015, 2016; Wolter
42   et al., 2015; Bellprat et al., 2016), in the cold spring of 2013 in the United Kingdom (Christidis et al., 2014),
43   and of 2016 in eastern China (Qian et al., 2018; Sun et al., 2018b) all showed a reduced probability due to
44   human influence on the climate. An exception is the study of Grose et al. (2018), who found an increase in
45   the probability of the severe western Australian frost of 2016 due to anthropogenically-driven changes in
46   circulation patterns that drive cold outbreaks and frost probability.
47
48   Different event attribution studies can produce a wide range of changes in the probability of event
49   occurrence because of different framing. The temperature event definition itself plays a crucial role in the
50   attributable signal (Fischer and Knutti, 2015; Kirchmeier‐Young et al., 2019). Large-scale, longer-duration
51   events tend to have notably larger attributable risk ratios (Angélil et al., 2014, 2018; Uhe et al., 2016;
52   Harrington, 2017; Kirchmeier‐Young et al., 2019), as natural variability is smaller. While uncertainty in the
53   best estimates of the risk ratios may be large, their lower bounds can be quite insensitive to uncertainties in
54   observations or model descriptions, thus increasing confidence in conservative attribution statements (Jeon et
55   al., 2016).
     Do Not Cite, Quote or Distribute                      11-46                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1
 2   The relative strength of anthropogenic influences on temperature extremes is regionally variable, in part due
 3   to differences in changes in atmospheric circulation, land surface feedbacks, and other external drivers like
 4   aerosols. For example, in the Mediterranean and over western Europe, risk ratios on the order of 100 have
 5   been found (Kew et al., 2019; Vautard et al., 2020a), whereas in the US, changes are much less pronounced.
 6   This is probably a reflection of the land-surface feedback enhanced extreme 1930s temperatures that reduce
 7   the rarity of recent extremes, in addition to the definition of the events and framing of attribution analyses
 8   (e.g., spatial and temporal scales considered). Local forcing may mask or enhance the warming effect of
 9   greenhouse gases. In India, short-lived aerosols or an increase in irrigation may be masking the warming
10   effect of greenhouse gases (Wehner et al., 2018c). Irrigation and crop intensification have been shown to
11   lead to a cooling in some regions, in particular in North America, Europe, and India (Mueller et al., 2016b;
12   Thiery et al., 2017, 2020; Chen and Dirmeyer, 2019),(high confidence). Deforestation has contributed about
13   one third of the total warming of hot extremes in some mid-latitude regions since pre-industrial times
14   (Lejeune et al., 2018). Despite all of these differences, and larger uncertainties at the regional scale, nearly
15   all studies demonstrated that human influence has contributed to an increase in the frequency or intensity of
16   hot extremes and to a decrease in the frequency or intensity of cold extremes.
17
18   In summary, long-term changes in various aspects of long- and short-duration extreme temperatures,
19   including intensity, frequency, and duration have been detected in observations and attributed to human
20   influence at global and continental scales. It is extremely likely that human influence is the main contributor
21   to the observed increase in the intensity and frequency of hot extremes and the observed decrease in the
22   intensity and frequency of cold extremes on the global scale. It is very likely that this applies on continental
23   scales as well. Some specific recent hot extreme events would have been extremely unlikely to occur without
24   human influence on the climate system. Changes in aerosol concentrations have affected trends in hot
25   extremes in some regions, with the presence of aerosols leading to attenuated warming, in particular from
26   1950-1980. Crop intensification, irrigation and no-till farming have attenuated increases in summer hot
27   extremes in some regions, such as central North America (medium confidence).
28
29
30   11.3.5 Projections
31
32   AR5 (Chapter12, Collins et al., 2013a) concluded it is virtually certain there will be more frequent hot
33   extremes and fewer cold extremes at the global scale and over most land areas in a future warmer climate
34   and it is very likely heat waves will occur with a higher frequency and longer duration . SR15 (Chapter 3,
35   Hoegh-Guldberg et al., 2018) assessment on projected changes in hot extremes at 1.5°C and 2°C global
36   warming is consistent with the AR5 assessment, concluding it is very likely a global warming of 2°C, when
37   compared with a 1.5°C warming, would lead to more frequent and more intense hot extremes on land, as
38   well as to longer warm spells, affecting many densely-inhabited regions. SR15 also assessed it is very likely
39   the strongest increases in the frequency of hot extremes are projected for the rarest events, while cold
40   extremes will become less intense and less frequent and cold spells will be shorter.
41
42   New studies since AR5 and SR15 confirm these assessments. New literature since AR5 includes projections
43   of temperature-related extremes in relation to changes in mean temperatures, projections based on CMIP6
44   simulations, projections based on stabilized global warming levels, and the use of new metrics. Constraints
45   for the projected changes in hot extremes were also provided (Borodina et al., 2017b; Sippel et al., 2017b;
46   Vogel et al., 2017). Overall, projected changes in the magnitude of extreme temperatures over land are larger
47   than changes in global mean temperature, over mid-latitude land regions in particular (Figures 11.3, 11.11),
48   (Fischer et al., 2014; Seneviratne et al., 2016; Sanderson et al., 2017a; Wehner et al., 2018b; Di Luca et al.,
49   2020a). Large warming in hot and cold extremes will occur even at the 1.5°C global warming level (Figure
50   11.11). At this level, widespread significant changes at the grid-box level occur for different temperature
51   indices (Aerenson et al., 2018). In agreement with CMIP5 projections, CMIP6 simulations show that a 0.5°C
52   increment in global warming will significantly increase the intensity and frequency of hot extremes and
53   decrease the intensity and frequency of cold extremes on the global scale (Figures 11.6, 11.8, 11.12). It takes
54   less than half of a degree for the changes in TXx to emerge above the level of natural variability (Figure
55   11.8) and the 66% ranges of the land medians of the 10-year or 50-year TXx events do not overlap between
     Do Not Cite, Quote or Distribute                      11-47                                      Total pages: 345
     Final Government Distribution                          Chapter11                                       IPCC AR6 WGI

 1   1.0°C and 1.5°C in the CMIP6 multi-model ensemble simulations (Figure 11.6, Li et al., 2020).
 2
 3
 4   [START FIGURE 11.11 HERE]
 5
 6
 7   Figure 11.11:Projected changes in (a-c) annual maximum temperature (TXx) and (d-f) annual minimum temperature
 8                (TNn) at 1.5°C, 2°C, and 4°C of global warming compared to the 1851-1900 baseline. Results are based
 9                on simulations from the CMIP6 multi-model ensemble under the SSP1-1.9, SSP1-2.6, SSP2-4.5, SSP3-
10                7.0, and SSP5-8.5 scenarios. The numbers in the top right indicate the number of simulations included.
11                Uncertainty is represented using the simple approach: no overlay indicates regions with high model
12                agreement, where ≥80% of models agree on sign of change; diagonal lines indicate regions with low
13                model agreement, where <80% of models agree on sign of change. For more information on the simple
14                approach, please refer to the Cross-Chapter Box Atlas 1. For details on the methods see Supplementary
15                Material 11.SM.2. Further details on data sources and processing are available in the chapter data table
16                (Table 11.SM.9).
17
18   [END FIGURE 11.11 HERE]
19
20
21   Projected warming is larger for TNn and exhibits strong equator-to-pole amplification similar to the warming
22   of boreal winter mean temperatures. The warming of TXx is more uniform over land and does not exhibit
23   this behaviour (Figure 11.11). The warming of temperature extremes on global and regional scales tends to
24   scale linearly with global warming (Section 11.1.4) (Fischer et al., 2014; Seneviratne et al., 2016,
25   Wartenburger et al., 2017; Li et al., 2020; see also SR15, Chapter 3). In the mid-latitudes, the rate of
26   warming of hot extremes can be as large as twice the rate of global warming (Figure 11.11). In the Arctic
27   winter, the rate of warming of the temperature of the coldest nights is about three times the rate of global
28   warming (Appendix Figure 11.A.1). Projected changes in temperature extremes can deviate from projected
29   changes in annual mean warming in the same regions (Figure 11.3, Figs. 11.A.1 and 11.A.2, Di Luca et al.,
30   2020a; Wehner, 2020) due to the additional processes that control the response of regional extremes,
31   including, in particular, soil moisture-evapotranspiration-temperature feedbacks for hot extremes in the mid-
32   latitudes and subtropical regions, and snow/ice-albedo-temperature feedbacks in high-latitude regions.
33
34
35   [START FIGURE 11.12 HERE]
36
37   Figure 11.12:Projected changes in the intensity of extreme temperature events under 1°C, 1.5°C, 2°C, 3°C, and 4°C
38                global warming levels relative to the 1851-1900 baseline. Extreme temperature events are defined as the
39                daily maximum temperatures (TXx) that were exceeded on average once during a 10-year period (10-year
40                event, blue) and that once during a 50-year period (50-year event, orange) during the 1851-1900 base
41                period. Results are shown for the global land. For each box plot, the horizontal line and the box represent
42                the median and central 66% uncertainty range, respectively, of the intensity changes across the space, and
43                the whiskers extend to the 90% uncertainty range. The results are based on the multi-model ensemble
44                median estimated from simulations of global climate models contributing to the sixth phase of the
45                Coupled Model Intercomparison Project (CMIP6) under different SSP forcing scenarios. Adapted from
46                (Li et al., 2020a). Further details on data sources and processing are available in the chapter data table
47                (Table 11.SM.9).
48
49   [END FIGURE 11.12 HERE]
50
51
52   The probability of exceeding a certain hot extreme threshold will increase, while those for cold extreme will
53   decrease with global warming (Mueller et al., 2016a; Lewis et al., 2017b; Suarez-Gutierrez et al., 2020b).
54   The changes tend to scale nonlinearly with the level of global warming, with larger changes for more rare
55   events (Section 11.2.4 and CCB 11.11; Figure. 11.6 and 11.12; e.g. Fischer and Knutti, 2015, Kharin et al.,
56   2018; Li et al., 2020). For example, the CMIP5 ensemble projects the frequency of the present-day climate
57   20-year hottest daily temperature to increase by 80% at the 1.5°C global warming level and by 180% at the
58   2.0°C global warming level, and the frequency of the present-day climate 100-year hottest daily temperature
     Do Not Cite, Quote or Distribute                         11-48                                         Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   to increase by 200% and more than 700% at the 1.5°C and 2.0°C warming levels, respectively (Kharin et al.,
 2   2018). CMIP6 simulations project similar changes (Li et al., 2020a).
 3
 4   Tebaldi and Wehner (2018) showed that at the middle of the 21st century, 66% of the land surface area would
 5   experience the present-day 20-year return values of TXx and the running 3-day average of the daily
 6   maximum temperature every other year on average under the RCP8.5 scenario, as opposed to only 34%
 7   under RCP4.5. By the end of the century, these area fractions increase to 92% and 62%, respectively. Such
 8   nonlinearities in the characteristics of future regional extremes are shown, for instance, for Europe (Lionello
 9   and Scarascia, 2020; Spinoni et al., 2018a; Dosio and Fischer, 2018), Asia (Guo et al., 2017; Harrington and
10   Otto, 2018b; King et al., 2018), and Australia (Lewis et al., 2017a) under various global warming thresholds.
11   The non-linear increase in fixed-threshold indices (e.g., percentile-based for a given reference period or
12   based on an absolute threshold) as a function of global warming is consistent with a linear warming of the
13   absolute temperature of the temperature extremes (e.g., Whan et al., 2015). Compared to the historical
14   climate, warming will result in strong increases in heat wave area, duration, and magnitude (Vogel et al.,
15   2020b). These changes are mostly due to the increase in mean seasonal temperature, rather than changes in
16   temperature variability, though the latter can have an effect in some regions (Di Luca et al., 2020a; Suarez-
17   Gutierrez et al., 2020a; Brown, 2020).
18
19   Projections of temperature-related extremes in RCMs in the CORDEX regions demonstrate robust increases
20   under future scenarios and can provide information on finer spatial scales than GCMs (e.g. Coppola et al.,
21   2021). Five RCMs in the CORDEX-East Asia region project decreases in the 20-year return values of
22   temperature extremes (summer maxima), with models that exhibit warm biases projecting stronger warming
23   (Park and Min, 2018). Similarly, in the African domain, future increases in TX90p and TN90p are projected
24   (Dosio, 2017; Mostafa et al., 2019). This regional-scale analysis provides fine scale information, such as
25   distinguishing the increase in TX90p over sub-equatorial Africa (Democratic Republic of Congo, Angola
26   and Zambia) with values over the Gulf of Guinea, Central African Republic, South Sudan, and Ethiopia.
27   Empirical-statistical downscaling has also been used to produce more robust estimates for future heat waves
28   compared to RCMs based on large multi-model ensembles (Furrer et al., 2010; Keellings and Waylen, 2014;
29   Wang et al., 2015; Benestad et al., 2018).
30
31   In all continental regions, including Africa (Table 11.4), Asia (Table 11.7), Australasia (Table 11.10),
32   Central and South America (Table 11.13), Europe (Table 11.16), North America (Table 11.19) and at the
33   continental scale, it is very likely the intensity and frequency of hot extremes will increase and the intensity
34   and frequency of cold extremes will decrease compared with the 1995-2014 baseline, even under 1.5°C
35   global warming, and those changes are virtually certain to occur under 4°C global warming. At the regional
36   scale and for almost all AR6 regions, it is likely the intensity and frequency of hot extremes will increase and
37   the intensity and frequency of cold extremes will decrease compared with the 1995-2014 baseline, even
38   under 1.5°C global warming and those changes will virtually certain to occur under 4°C global warming.
39   Exceptions include lower confidence in the projected decrease in the intensity and frequency of cold
40   extremes compared with the 1995-2014 baseline under 1.5°C of global warming (medium confidence) and
41   4°C of global warming (very likely) in North Central America, Central North America, and Western North
42   America.
43
44   In Africa (Table 11.4), evidence includes increases in the intensity and frequency of hot extremes, such as
45   warm days, warm nights, and heat waves, and decreases in the intensity and frequency of cold extremes,
46   such as cold days and cold nights, over the continent as projected by CMIP5, CMIP6, and CORDEX
47   simulations (Giorgi et al., 2014; Engelbrecht et al., 2015; Lelieveld et al., 2016; Russo et al., 2016; Dosio,
48   2017; Bathiany et al., 2018; Mba et al., 2018; Nangombe et al., 2018; Weber et al., 2018; Kruger et al., 2019;
49   Coppola et al., 2021; Li et al., 2020). Cold spells are projected to decrease under all RCPs and even at low
50   warming levels in West and Central Africa (Diedhiou et al., 2018) and the number of cold days is projected
51   to decrease in East Africa (Ongoma et al., 2018b).
52
53   In Asia (Table 11.7), evidence includes increases in the intensity and frequency of hot extremes, such as
54   warm days, warm nights, and heat waves, and decreases in the intensity and frequency of cold extremes,
55   such as cold days and cold nights, over the continent as projected by CMIP5, CMIP6, and CORDEX
     Do Not Cite, Quote or Distribute                     11-49                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   simulations (Gao et al., 2018; Han et al., 2018; Li et al., 2019b; Pal and Eltahir, 2016; Shin et al., 2018;
 2   Sillmann et al., 2013b; Singh and Goyal, 2016; Sui et al., 2018; Xu et al., 2017; Zhang et al., 2015c; Zhao et
 3   al., 2015; Zhou et al., 2014; Zhu et al., 2020). More intense heat waves of longer durations and occurring at a
 4   higher frequency are projected over India (Murari et al., 2015; Mishra et al., 2017) and Pakistan (Nasim et
 5   al., 2018). Future mid-latitude warm extremes, similar to those experienced during the 2010 event, are
 6   projected to become more extreme, with temperature extremes increasing potentially by 8.4°C (RCP8.5)
 7   over northwest Asia (van der Schrier et al., 2018). Over WSB, ESB and RFE, an increase in extreme heat
 8   durations is expected in all scenarios (Sillmann et al., 2013b; Kattsov et al., 2017; Reyer et al., 2017). In the
 9   MENA regions (ARP, WCA), extreme temperatures could increase by almost 7°C by 2100 under RCP8.5
10   (Lelieveld et al., 2016).
11
12   In Australasia (Table 11.10), evidence includes increases in the intensity and frequency of hot extremes, such
13   as warm days, warm nights, and heat waves, and decreases in the intensity and frequency of cold extremes,
14   such as cold days and cold nights, over the continent as projected by CMIP5, CMIP6, and CORDEX
15   simulations (Coppola et al., 2021; Alexander and Arblaster, 2017; CSIRO and BOM, 2015; Herold et al.,
16   2018; Lewis et al., 2017a; Evans et al., 2020). Over most of Australia, increases in the intensity and
17   frequency of hot extremes are projected to be predominantly driven by the long-term increase in mean
18   temperatures (Di Luca et al., 2020a). Future projections indicate a decrease in the number of frost days
19   regardless of the region and season considered (Alexander and Arblaster, 2017; Herold et al., 2018).
20
21   In Central and South America (Table 11.13), evidence includes increases in the intensity and frequency of
22   hot extremes, such as warm days, warm nights, and heat waves, and decreases in the intensity and frequency
23   of cold extremes, such as cold days and cold nights, over the continent as projected by CMIP5, CMIP6, and
24   CORDEX simulations (Chou et al., 2014a; Cabré et al., 2016; López-Franca et al., 2016; Stennett-Brown et
25   al., 2017; Li et al., 2020a; Coppola et al., 2021b; Vichot-Llano et al., 2021). Over SES during the austral
26   summer, the increase in the frequency of TN90p is larger than that projected for TX90p, consistent with
27   observed past changes (López-Franca et al., 2016). Under RCP8.5, the number of heat wave days are
28   projected to increase for the intra-Americas region for the end of the 21st century (Angeles-Malaspina et al.,
29   2018). A general decrease in the frequency of cold spells and frost days is projected as indicated by several
30   indices based on minimum temperature (López-Franca et al., 2016).
31
32   In Europe (Table 11.16), evidence includes increases in the intensity and frequency of hot extremes, such as
33   warm days, warm nights, and heat waves, and decreases in the intensity and frequency of cold extremes,
34   such as cold days and cold nights, over the continent as projected by CMIP5, CMIP6, and CORDEX
35   simulations (Coppola et al., 2021; Cardoso et al., 2019; Jacob et al., 2018; Lau and Nath, 2014; Lhotka et al.,
36   2018; Lionello and Scarascia, 2020; Molina et al., 2020; Ozturk et al., 2015; Rasmijn et al., 2018; Russo et
37   al., 2015; Schoetter et al., 2015; Suarez-Gutierrez et al., 2018; Vogel et al., 2017; Winter et al., 2017; Li et
38   al., 2020). Increases in heat waves are greater over the southern Mediterranean and Scandinavia (Forzieri et
39   al., 2016; Abaurrea et al., 2018; Dosio and Fischer, 2018; Rohat et al., 2019). The biggest increases in the
40   number of heat wave days are expected for southern European cities (Guerreiro et al., 2018a; Junk et al.,
41   2019), and Central European cities will see the biggest increases in maximum heat wave temperatures
42   (Guerreiro et al., 2018a).
43
44   In North America (Table 11.19), evidence includes increases in the intensity and frequency of hot extremes,
45   such as warm days, warm nights, and heat waves, and decreases in the intensity and frequency of cold
46   extremes, such as cold days and cold nights, over the continent as projected by CMIP5, CMIP6, and
47   CORDEX simulations (Li et al., 2020; Coppola et al., 2021; Alexandru, 2018; Grotjahn et al., 2016; Li et al.,
48   2018a; Vose et al., 2017a; Yang et al., 2018a; Zhang et al., 2019d). Projections of temperature extremes for
49   the end of the 21st century show that warm days and nights are very likely to increase and cold days and
50   nights are very likely to decrease in all regions. There is medium confidence in large increases in warm days
51   and warm nights in summer, particularly over the United States, and in large decreases in cold days in
52   Canada in fall and winter (Li et al., 2020; Coppola et al., 2021; Alexandru, 2018; Grotjahn et al., 2016; Li et
53   al., 2018a; Vose et al., 2017a; Yang et al., 2018a; Zhang et al., 2019d). Minimum winter temperatures are
54   projected to rise faster than mean winter temperatures (Underwood et al., 2017). Projections for the end of
55   the century under RCP8.5 showed the 4-day cold spell that happens on average once every 5 years is
     Do Not Cite, Quote or Distribute                      11-50                                      Total pages: 345
     Final Government Distribution                       Chapter11                                      IPCC AR6 WGI

 1   projected to warm by more than 10 ºC and CMIP5 models do not project current 1-in-20 year annual
 2   minimum temperature extremes to recur over much of the continent (Wuebbles et al., 2014).
 3
 4   In summary, it is virtually certain that further increases in the intensity and frequency of hot extremes and
 5   decreases in the intensity and frequency of cold extremes will occur throughout the 21st century and around
 6   the world. It is virtually certain the number of hot days and hot nights and the length, frequency, and/or
 7   intensity of warm spells or heat waves compared to 1995-2014 will increase over most land areas. In most
 8   regions, changes in the magnitude of temperature extremes are proportional to global warming levels (high
 9   confidence). The highest increase of temperature of hottest days is projected in some mid-latitude and semi-
10   arid regions, at about 1.5 time to twice the rate of global warming (high confidence). The highest increase of
11   temperature of coldest days is projected in Arctic regions, at about three times the rate of global warming
12   (high confidence). The probability of temperature extremes generally increases non-linearly with increasing
13   global warming levels (high confidence). Confidence in assessments depends on the spatial and temporal
14   scales of the extreme in question, with high confidence in projections of temperature-related extremes at
15   global and continental scales for daily to seasonal scales. There is high confidence that, on land, the
16   magnitude of temperature extremes increases more strongly than global mean temperature.
17
18
19   11.4 Heavy precipitation
20
21   This section assesses changes in heavy precipitation at global and regional scales. The main focus is on
22   extreme precipitation at a daily scale where literature is most concentrated, though extremes of shorter (sub-
23   daily) and longer (five-day or more) durations are also assessed to the extent the literature allows.
24
25
26   11.4.1 Mechanisms and drivers
27
28   SREX (Chapter 3, Seneviratne et al., 2012) assessed changes in heavy precipitation in the context of the
29   effects of thermodynamic and dynamic changes. Box 11.1 assesses thermodynamic and dynamic changes in
30   a warming world to aid the understanding of changes in observations and projections in some extremes and
31   the sources of uncertainties (See also Chapter 8, Section 8.2.3.2). In general, warming increases the
32   atmospheric water-holding capacity following the Clausius-Clapeyron (C-C) relation. This thermodynamic
33   effect results in an increase in extreme precipitation at a similar rate at the global scale. On a regional scale,
34   changes in extreme precipitation are further modulated by dynamic changes (Box 11.1).
35
36   Large-scale modes of variability, such as the North Atlantic Oscillation (NAO), El Niño-Southern
37   Oscillation (ENSO), Atlantic Multidecadal Variability (AMV), and Pacific Decadal Variability (PDV)
38   (Annex VI), modulate precipitation extremes through changes in environmental conditions or embedded
39   storms (Section 8.3.2). Latent heating can invigorate these storms (Nie et al., 2018; Zhang et al., 2019g);
40   changes in dynamics can increase precipitation intensity above that expected from the C-C scaling rate
41   (8.2.3.2, Box 11.1, and Section 11.7). Additionally, the efficiency of converting atmospheric moisture into
42   precipitation can change as a result of cloud microphysical adjustment to warming, resulting in changes in
43   the characteristics of extreme precipitation; but changes in precipitation efficiency in a warming world are
44   highly uncertain (Sui et al., 2020).
45
46   It is difficult to separate the effect of global warming from internal variability in the observed changes in the
47   modes of variability (Section 2.4). Future projections of modes of variability are highly uncertain (Section
48   4.3.3), resulting in uncertainty in regional projections of extreme precipitation. Future warming may amplify
49   monsoonal extreme precipitation. Changes in extreme storms, including tropical/extratropical cyclones and
50   severe convective storms, result in changes in extreme precipitation (Section 11.7). Also, changes in sea
51   surface temperatures (SSTs) alter land-sea contrast, leading to changes in precipitation extremes near coastal
52   regions. For example, the projected larger SST increase near the coasts of East Asia and India can result in
53   heavier rainfall near these coastal areas from tropical cyclones (Mei and Xie, 2016) or torrential rains
54   (Manda et al., 2014). The warming in the western Indian Ocean is associated with increases in moisture
55   surges on the low-level monsoon westerlies towards the Indian subcontinent, which may lead to an increase
     Do Not Cite, Quote or Distribute                       11-51                                       Total pages: 345
     Final Government Distribution                       Chapter11                                     IPCC AR6 WGI

 1   in the occurrence of precipitation extremes over central India (Krishnan et al., 2016; Roxy et al., 2017).
 2
 3   Decreases in atmospheric aerosols results in warming and thus an increase in extreme precipitation (Samset
 4   et al., 2018; Sillmann et al., 2019). Changes in atmospheric aerosols also result in dynamic changes such as
 5   changes in tropical cyclones (Takahashi et al., 2017; Strong et al., 2018). Uncertainty in the projections of
 6   future aerosol emissions results in additional uncertainty in the heavy precipitation projections of the 21st
 7   century (Lin et al., 2016).
 8
 9   There has been new evidence of the effect of local land use and land cover change on heavy precipitation.
10   There is a growing set of literature linking increases in heavy precipitation in urban centres to urbanization
11   (Argüeso et al., 2016; Zhang et al., 2019f). Urbanization intensifies extreme precipitation, especially in the
12   afternoon and early evening, over the urban area and its downwind region (medium confidence) (Box 10.3).
13   There are four possible mechanisms: a) increases in atmospheric moisture due to horizontal convergence of
14   air associated with the urban heat island effect (Shastri et al., 2015; Argüeso et al., 2016); b) increases in
15   condensation due to urban aerosol emissions (Han et al., 2011; Sarangi et al., 2017); c) aerosol pollution that
16   impacts cloud microphysics (Schmid and Niyogi, 2017) (Box 8.1); and d) urban structures that impede
17   atmospheric motion (Ganeshan and Murtugudde, 2015; Paul et al., 2018; Shepherd, 2013). Other local
18   forcing, including reservoirs (Woldemichael et al., 2012), irrigation (Devanand et al., 2019), or large-scale
19   land use and land cover change (Odoulami et al., 2019), can also affect local extreme precipitation.
20
21   In summary, precipitation extremes are controlled by both thermodynamic and dynamic processes.
22   Warming-induced thermodynamic change results in an increase in extreme precipitation, at a rate that
23   closely follows the Clausius-Clapeyron relationship at the global scale (high confidence). The effects of
24   warming-induced changes in dynamic drivers on extreme precipitation are more complicated, difficult to
25   quantify and are an uncertain aspect of projections. Precipitation extremes are also affected by forcings other
26   than changes in greenhouse gases, including changes in aerosols, land use and land cover change, and
27   urbanization (medium confidence).
28
29
30   11.4.2 Observed Trends
31
32   Both SREX (Chapter 3, Seneviratne et al., 2012) and AR5 (IPCC, 2014 Chapter 2) concluded it was likely
33   the number of heavy precipitation events over land had increased in more regions than it had decreased,
34   though there were wide regional and seasonal variations, and trends in many locations were not statistically
35   significant. This assessment has been strengthened with multiple studies finding robust evidence of the
36   intensification of extreme precipitation at global and continental scales, regardless of spatial and temporal
37   coverage of observations and the methods of data processing and analysis.
38
39   The average annual maximum precipitation amount in a day (Rx1day) has significantly increased since the
40   mid-20th century over land (Du et al., 2019; Dunn et al., 2020) and in the humid and dry regions of the globe
41   (Dunn et al., 2020). The percentage of observing stations with statistically significant increases in Rx1day is
42   larger than expected by chance, while the percentage of stations with statistically significant decreases is
43   smaller than expected by chance, over the global land as a whole and over North America, Europe, and Asia
44   (Figure 11.13, Sun et al., 2020) and over global monsoon regions (Zhang and Zhou, 2019) where data
45   coverage is relatively good. The addition of the past decade of observational data shows a more robust
46   increase in Rx1day over the global land region (Sun et al., 2020). Light, moderate, and heavy daily
47   precipitation has all intensified in a gridded daily precipitation data set (Contractor et al., 2020). Daily mean
48   precipitation intensities have increased since the mid-20th century in a majority of land regions (high
49   confidence, Section 8.3.1.3). The probability of precipitation exceeding 50 mm/day increased during 1961-
50   2018 (Benestad et al., 2019). The globally averaged annual fraction of precipitation from days in the top 5%
51   (R95pTOT) has also significantly increased (Dunn et al., 2020). The increase in the magnitude of Rx1day in
52   the 20th century is estimated to be at a rate consistent with C-C scaling with respect to global mean
53   temperature (Fischer and Knutti, 2016; Sun et al., 2020). Studies on past changes in extreme precipitation of
54   durations longer than a day are more limited, though there are some studies examining long-term trends in
55   annual maximum five-day precipitation (Rx5day). On global and continental scales, long-term changes in
     Do Not Cite, Quote or Distribute                      11-52                                       Total pages: 345
     Final Government Distribution                       Chapter11                                    IPCC AR6 WGI

 1   Rx5day are similar to those of Rx1day in many aspects (Zhang and Zhou 2019; Sun et al., 2020). As
 2   discussed below, at the regional scale, changes in Rx5day are also similar to those of Rx1day where there are
 3   analyses of changes in both Rx1day and Rx5day.
 4
 5   Overall, there is a lack of systematic analysis of long-term trends in sub-daily extreme precipitation at the
 6   global scale. Often, sub-daily precipitation data have only sporadic spatial coverage and are of limited
 7   length. Additionally, the available data records are far shorter than needed for a robust quantification of past
 8   changes in sub-daily extreme precipitation (Li et al., 2018b). Despite these limitations, there are studies in
 9   regions of almost all continents that generally indicate intensification of sub-daily extreme precipitation,
10   although confidence in an overall increase at the global scale remains very low. Studies include an increase in
11   extreme sub-daily rainfall in summer over South Africa (Sen Roy and Rouault, 2013), annually in Australia
12   (Guerreiro et al., 2018b), over 23 urban locations in India (Ali and Mishra, 2018), in Peninsular Malaysia
13   (Syafrina et al., 2015), and in eastern China in the summer season during 1971-2013 (Xiao et al., 2016). In
14   some regions in Italy (Arnone et al., 2013; Libertino et al., 2019) and in the US during 1950-2011 (Barbero
15   et al., 2017), there is also an increase. In general, an increase in sub-daily heavy precipitation results in an
16   increase in pluvial floods over smaller watersheds (Ghausi and Ghosh, 2020).
17
18   There is a considerable body of literature examining scaling of sub-daily precipitation extremes, conditional
19   on day-to-day air or dew-point temperatures (Westra et al., 2014; Fowler et al., 2021). This scaling, termed
20   apparent scaling (Fowler et al., 2020) is robust when different methodologies are used in different regions,
21   ranging between the C-C and two-times the C-C rate (e.g. Burdanowitz et al., 2019; Formayer and Fritz,
22   2017; Lenderink et al., 2017). This is confirmed when sub-daily precipitation data collected from multiple
23   continents (Lewis et al., 2019a) are analysed in a consistent manner using different methods (Ali et al.,
24   2021). It has been hoped that apparent scaling might be used to help understand past and future changes in
25   extreme sub-daily precipitation. However, apparent scaling samples multiple synoptic weather states, mixing
26   thermodynamic and dynamic factors that are not directly relevant for climate change responses (8.2.3.2)
27   (Prein et al., 2016b; Bao et al., 2017; Zhang et al., 2017c; Drobinski et al., 2018; Sun et al., 2019d). The
28   spatial pattern of apparent scaling is different from those of projected changes over Australia (Bao et al.,
29   2017) and North America (Sun et al., 2019) in regional climate model simulations. It thus remains difficult to
30   use the knowledge about apparent scaling to infer past and future changes in extreme sub-daily precipitation
31   according to observed and projected changes in local temperature.
32
33   In Africa (Table 11.5), evidence shows an increase in extreme daily precipitation for the late half of the 20th
34   century over the continent where data are available; there is a larger percentage of stations showing
35   significant increases in extreme daily precipitation than decreases (Sun et al., 2020). There are increases in
36   different metrics relevant to extreme precipitation in various regions of the continent (Chaney et al., 2014;
37   Harrison et al., 2019; Dunn et al., 2020; Sun et al., 2020). There is an increase in extreme precipitation
38   events in southern Africa (Weldon and Reason, 2014; Kruger et al., 2019) and a general increase in heavy
39   precipitation over East Africa, the Greater Horn of Africa (Omondi et al., 2014). Over sub-Saharan Africa,
40   increases in the frequency and intensity of extreme precipitation have been observed over the well-gauged
41   areas during 1950-2013; however, this covers only 15% of the total area of sub-Saharan Africa (Harrison et
42   al., 2019). Confidence about the increase in extreme precipitation for some regions where observations are
43   more abundant is medium, but for Africa as whole, it is low because of a general lack of continent-wide
44   systematic analysis, the sporadic nature of available precipitation data over the continent, and spatially non-
45   homogenous trends in places where data are available (Donat et al., 2014a; Mathbout et al., 2018; Alexander
46   et al., 2019; Funk et al., 2020)
47
48   In Asia (Table 11.8), there is robust evidence that extreme precipitation has increased since the 1950s (high
49   confidence), however this is dominated by high spatial variability. Increases in Rx1day and Rx5day during
50   1950-2018 are found over two thirds of stations and the percentage of stations with statistically significant
51   trends is significantly larger than can be expected by chance (Sun et al., 2020, also Fig 11.13). An increase
52   in extreme precipitation has also been observed in various regional studies based on different metrics of
53   extreme precipitation and different spatial and temporal coverage of the data. These include an increase in
54   daily precipitation extremes over central Asia (Hu et al., 2016), most of South Asia (Zahid and Rasul, 2012;
55   Pai et al., 2015; Sheikh et al., 2015; Adnan et al., 2016; Malik et al., 2016; Dimri et al., 2017; Priya et al.,
     Do Not Cite, Quote or Distribute                      11-53                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   2017; Roxy et al., 2017; Hunt et al., 2018; Kim et al., 2019; Wester et al., 2019), the Arabian Peninsula
 2   (Rahimi and Fatemi, 2019; Almazroui and Saeed, 2020; Atif et al., 2020), Southeast Asia (Siswanto et al.,
 3   2015; Supari et al., 2017; Cheong et al., 2018); the northwest Himalaya (Malik et al., 2016), parts of east
 4   Asia (Nayak et al., 2017; Baek et al., 2017; Ye and Li, 2017), the western Himalayas since the 1950s (Ridley
 5   et al., 2013; Dimri et al., 2015; Madhura et al., 2015), WSB, ESB and RFE (Donat et al., 2016a) and a
 6   decrease was found over the eastern Himalayas (Sheikh et al., 2015; Talchabhadel et al., 2018). Increases
 7   have been observed over Jakarta (Siswanto et al., 2015), but Rx1day over most parts of the Maritime
 8   Continent has decreased (Villafuerte and Matsumoto, 2015). Trends in extreme precipitation over China are
 9   mixed with increases and decreases (Fu et al., 2013a; Jiang et al., 2013; Ma et al., 2015; Yin et al., 2015;
10   Xiao et al., 2016) and are not significant over China as whole (Li et al., 2018c; Ge et al., 2017; Hu et al.,
11   2016; Jiang et al., 2013; Liu et al., 2019b; Chen et al., 2021; Deng et al., 2018; He and Zhai, 2018; Tao et al.,
12   2018). With few exceptions, most Southeast Asian countries have experienced an increase in rainfall
13   intensity, but with a reduced number of wet days (Donat et al., 2016a; Cheong et al., 2018; Naveendrakumar
14   et al., 2019), though large differences in trends exists if the trends are estimated from different datasets
15   including gauge-based, remotely-sensed, and reanalysis over a relatively short period (Kim et al. 2019).
16   There is a significant increase in heavy rainfall (>100 mm day-1) and a significant decrease in moderate
17   rainfall (5–100 mm day-1) in central India during the South Asian monsoon season (Deshpande et al., 2016;
18   Roxy et al., 2017).
19
20   In Australasia (Table 11.11), available evidence has not shown an increase or a decrease in heavy
21   precipitation over Australasia as a whole (medium confidence), but heavy precipitation tends to increase over
22   northern Australia (particularly the northwest) and decrease over the eastern and southern regions (e.g.,
23   Jakob and Walland, 2016; Dey et al., 2018; Guerreiro et al., 2018; Dunn et al., 2020; Sun et al., 2020).
24   Available studies that used long-term observations since the mid-20th century showed nearly as many
25   stations with an increase as those with a decrease in heavy precipitation (Jakob and Walland, 2016) or
26   slightly more stations with a decrease than with an increase in Rx1day and Rx5day (Sun et al., 2020), or
27   strong differences in Rx1day trends with increases over northern Australia and central Australia in general
28   but mostly decreases over southern Australia and eastern Australia (Dunn et al., 2020). Over New Zealand,
29   decreases are observed for moderate-heavy precipitation events, but there are no significant trends for very
30   heavy events (more than 64 mm in a day) for the period 1951-2012. The number of stations with an increase
31   in very wet days is similar to that with a decrease during 1960-2019 (MfE and Stats NZ, 2020). Overall,
32   there is low confidence in trends in the frequency of heavy rain days with mostly decreases over New
33   Zealand (Caloiero, 2015; Harrington and Renwick, 2014).
34
35   In Central and South America (Table 11.14), evidence shows an increase in extreme precipitation, but in
36   general there is low confidence; while continent-wide analyses produced wetting trends, trends are not
37   robust. Rx1day increased at more stations than it decreased in South America between 1950-2018 (Sun et al.,
38   2020). Over 1950-2010, both Rx5day and R99p increased over large regions of South America, including
39   NWS, NSA, and SES (Skansi et al., 2013). There are large regional differences. A decrease in daily extreme
40   precipitation is observed in northeastern Brazil (Bezerra et al., 2018; Dereczynski et al., 2020; Skansi et al.,
41   2013). Trends in extreme precipitation indices were not statistically significant over the period 1947-2012
42   within the São Francisco River basin in the Brazilian semi-arid region (Bezerra et al., 2018). An increase in
43   extreme rainfall is observed in AMZ with medium confidence (Skansi et al., 2013) and in SES with high
44   confidence (Ávila et al., 2016; Barros et al., 2015; Lovino et al., 2018; Skansi et al., 2013; Wu and Polvani,
45   2017; Dereczynski et al., 2020; Valverde and Marengo, 2014). Among all sub-regions, SES shows the
46   highest rate of increase for rainfall extremes, followed by AMZ (Skansi et al., 2013). Increases in the
47   intensity of heavy daily rainfall events have been observed in the southern Pacific and in the Titicaca basin
48   (Huerta and Lavado-Casimiro, 2020; Skansi et al., 2013). In SCA trends in annual precipitation are generally
49   not significant, although small (but significant) increases are found in Guatemala, El Salvador, and Panama
50   (Hidalgo et al., 2017). Small positive trends were found in multiple extreme precipitation indices over the
51   Caribbean region over a short time period (1986-2010) (Stephenson et al., 2014; McLean et al., 2015)
52
53   In Europe (Table 11.17), there is robust evidence that the magnitude and intensity of extreme precipitation
54   has very likely increased since the 1950s. There is a significant increase in Rx1day and Rx5day during 1950-
55   2018 in Europe as whole (Sun et al., 2020, also Figure 11.13). The number of stations with increases far
     Do Not Cite, Quote or Distribute                      11-54                                      Total pages: 345
     Final Government Distribution                           Chapter11                                         IPCC AR6 WGI

 1   exceeds those with decreases in the frequency of daily rainfall exceeding its 90th or 95th percentile in century-
 2   long series (Cioffi et al., 2015). The 5-, 10-, and 20-year events of one-day and five-day precipitation during
 3   1951-1960 became more common since the 1950s (van den Besselaar et al., 2013). There can be large
 4   discrepancies among studies and regions and seasons (Croitoru et al., 2013; Willems, 2013; Casanueva et al.,
 5   2014; Roth et al., 2014; Fischer et al., 2015); evidence for increasing extreme precipitation is more
 6   frequently observed for summer and winter, but not in other seasons (Madsen et al., 2014; Helama et al.,
 7   2018). An increase is observed in central Europe (Volosciuk et al., 2016; Zeder and Fischer, 2020), and in
 8   Romania (Croitoru et al., 2016). Trends in the Mediterranean region are in general not spatially (Reale and
 9   Lionello, 2013), with decreases in the western Mediterranean and some increases in the eastern
10   Mediterranean (Rajczak et al., 2013; Casanueva et al., 2014; de Lima et al., 2015; Gajić-Čapka et al., 2015;
11   Sunyer et al., 2015; Pedron et al., 2017; Serrano-Notivoli et al., 2018; Ribes et al., 2019). In the Netherlands,
12   the total precipitation contributed from extremes higher than the 99th percentile doubles per degree C
13   increase in warming (Myhre et al., 2019), though extreme rainfall trends in northern Europe may differ in
14   different seasons (Irannezhad et al., 2017).
15
16   In North America (Table 11.20), there is robust evidence that the magnitude and intensity of extreme
17   precipitation has very likely increased since the 1950s. Both Rx1day and Rx5day have significantly increased
18   in North America during 1950-2018 (Sun et al., 2020, also Figure 11.13). There is, however, regional
19   diversity. In Canada, there is a lack of detectable trends in observed annual maximum daily (or shorter
20   duration) precipitation (Shephard et al., 2014; Mekis et al., 2015; Vincent et al., 2018). In the United States,
21   there is an overall increase in one-day heavy precipitation, both in terms of intensity and frequency (Sun et
22   al., 2020; Donat et al., 2013; Huang et al., 2017; Villarini et al., 2012; Easterling et al., 2017; Wu, 2015;
23   Howarth et al., 2019), except for the southern part of the US (Hoerling et al., 2016) where internal variability
24   may have played a substantial role in the lack of observed increases. In Mexico, increases are observed in
25   R10mm and R95p (Donat et al., 2016a), very wet days over the cities (García-Cueto et al., 2019) and in
26   PRCPTOT and Rx1day (Donat et al., 2016b).
27
28   In Small Islands, there is a lack of evidence showing changes in heavy precipitation overall. There were
29   increases in extreme precipitation in Tobago from 1985–2015 (Stephenson et al., 2014; Dookie et al., 2019)
30   and decreases in southwestern French Polynesia and the southern subtropics (low confidence; Atlas.10; Table
31   11.5). Extreme precipitation leading to flooding in the small islands has been attributed in part to TCs, as
32   well as being influenced by ENSO (Khouakhi et al., 2016; Hoegh-Guldberg et al., 2018) (Box 11.5).
33
34
35   [START FIGURE 11.13 HERE]
36
37   Figure 11.13:Signs and significance of the observed trends in annual maximum daily precipitation (Rx1day) during
38                1950–2018 at 8345 stations with sufficient data. (a) Percentage of stations with statistically significant
39                trends in Rx1day; green dots show positive trends and brown dots negative trends. Box-and-whisker plots
40                indicate the expected percentage of stations with significant trends due to chance estimated from 1000
41                bootstrap realizations under a no-trend null hypothesis. The boxes mark the median, 25th percentile, and
42                75th percentile. The upper and lower whiskers show the 97.5th and the 2.5th percentiles, respectively. Maps
43                of stations with positive (b) and negative (c) trends. The light color indicates stations with non-significant
44                trends and the dark color stations with significant trends. Significance is determined by a two-tailed test
45                conducted at the 5% level. Adapted from Sun et al., (2020). © American Meteorological Society. Used
46                with permission. Further details on data sources and processing are available in the chapter data table
47                (Table 11.SM.9).
48
49   [END FIGURE 11.13 HERE]
50
51
52   In summary, the frequency and intensity of heavy precipitation have likely increased at the global scale over
53   a majority of land regions with good observational coverage. Since 1950, the annual maximum amount of
54   precipitation falling in a day or over five consecutive days has likely increased over land regions with
55   sufficient observational coverage for assessment, with increases in more regions than there are decreases.
56   Heavy precipitation has likely increased on the continental scale over three continents, including North
57   America, Europe, and Asia where observational data are more abundant. There is very low confidence about
     Do Not Cite, Quote or Distribute                          11-55                                          Total pages: 345
     Final Government Distribution                       Chapter11                                     IPCC AR6 WGI

 1   changes in sub-daily extreme precipitation due to a limited number of studies and the data used in these
 2   studies are often limited.
 3
 4
 5   11.4.3 Model evaluation
 6
 7   The evaluation of the skill of climate models to simulate heavy precipitation extremes is challenging due to a
 8   number of factors, including the lack of reliable observations and the spatial scale mismatch between
 9   simulated and observed data (Avila et al., 2015, Alexander et al., 2019). Simulated precipitation represents
10   areal means, but station-based observations are conducted at point locations and are often sparse. The areal-
11   reduction factor, the ratio between pointwise station estimates of extreme precipitation and extremes of the
12   areal mean, can be as large as 130% at CMIP6 resolutions (~100km) (Gervais et al., 2014). Hence, the order
13   in which gridded station based extreme values are constructed (i.e., if the extreme values are extracted at the
14   station first and then gridded or if the daily station values are gridded and then the extreme values are
15   extracted) represents different spatial scales of extreme precipitation and needs to be taken into account in
16   model evaluation (Wehner et al. 2020). This aspect has been considered in some studies. Reanalysis products
17   are used in place of station observations for their spatial completeness as well as spatial-scale comparability
18   (Sillmann et al., 2013a; Kim et al., 2020; Li et al., 2020). However, reanalyses share similar
19   parameterizations to the models themselves, reducing the objectivity of the comparison.
20
21   Different generations of the Coupled Model Intercomparison Project (CMIP) models have improved over
22   time, though quite modestly (Flato et al., 2013; Watterson et al., 2014). Improvements in the representation
23   of the magnitude of the ETCCDI indices in CMIP5 over CMIP3 (Sillmann et al., 2013a; Chen and Sun,
24   2015a) have been attributed to higher resolution as higher-resolution models represent smaller areas at
25   individual grid boxes. Additionally, the spatial distribution of extreme rainfall simulated by high-resolution
26   models (CMIP5 median resolution ∼ 180 ×96) is generally more comparable to observations (Sillmann et al.,
27   2013b; Kusunoki, 2017, 2018b; Scher et al., 2017) as these models tend to produce more realistic storms
28   compared to coarser models (11.7.2). Higher horizontal resolution alone improves simulation of extreme
29   precipitation in some models (Wehner et al., 2014; Kusunoki, 2017, 2018), but this is insufficient in other
30   models (Bador et al., 2020) as model parameterization also plays a significant role (Wu et al., 2020a). A
31   simple comparison of climatology may not fully reflect the improvements of the new models that have more
32   comprehensive formulations of processes (Di Luca et al., 2015). Dittus et al. (2016) found that many of the
33   eight CMIP5 models they evaluated reproduced the observed increase in the difference between areas
34   experiencing an extreme high (90%) and an extreme low (10%) proportion of the annual total precipitation
35   from heavy precipitation (R95p/PRCPTOT) for Northern Hemisphere regions. Additionally, CMIP5 models
36   reproduced the relation between changes in extreme and non-extreme precipitation: an increase in extreme
37   precipitation is at the cost of a decrease in non-extreme precipitation (Thackeray et al., 2018), a characteristic
38   found in the observational record (Gu and Adler, 2018).
39
40   CMIP6 models perform reasonably well in capturing large-scale features of precipitation extremes, including
41   intense precipitation extremes in the intertropical convergence zone (ITCZ), and weak precipitation extremes
42   in dry areas in the tropical regions (Li et al., 2020) but a double-ITCZ bias over the equatorial central and
43   eastern Pacific that appeared in CMIP5 models remains (3.3.2.1). There are also regional biases in the
44   magnitude of precipitation extremes (Kim et al., 2020). The models also have difficulties in reproducing
45   detailed regional patterns of extreme precipitation such as over the northeast US (Agel and Barlow, 2020),
46   though they performed better for summer extremes over the US (Akinsanola et al., 2020). The comparison
47   between climatologies in the observations and in model simulations shows that the CMIP6 and CMIP5
48   models that have similar horizontal resolutions also have similar model evaluation scores and their error
49   patterns are highly correlated (Wehner et al., 2020). In general, extreme precipitation in CMIP6 models tends
50   to be somewhat larger than in CMIP5 models (Li et al., 2020a), reflecting smaller spatial scales of extreme
51   precipitation represented by slightly higher resolution models (Gervais et al., 2014). This is confirmed by
52   Kim et al. (2020), who showed that Rx1day and Rx5day simulated by CMIP6 models tend to be closer to
53   point estimates of HadEX3 data (Dunn et al., 2020) than those simulated by CMIP5. Figure 11.14 shows the
54   multi-model ensemble bias in mean Rx1day over the period 1979-2014 from 21 available CMIP6 models
55   when compared with observations and reanalyses. Measured by global land root mean square error, the
     Do Not Cite, Quote or Distribute                      11-56                                       Total pages: 345
     Final Government Distribution                        Chapter11                                       IPCC AR6 WGI

 1   model performance is generally consistent across different observed/reanalysis data products for the extreme
 2   precipitation metric (Figure 11.14). The magnitude of extreme area-mean precipitation simulated by the
 3   CMIP6 models is consistently smaller than the point estimates of HadEX3, but the model values are more
 4   comparable to those of areal-mean values (Figure 11.14) of the ERA5 reanalysis or REGEN (Contractor et
 5   al., 2020b). Taylor-plot-based performance metrics reveal strong similarities in the patterns of extreme
 6   precipitation errors over land regions between CMIP5 and CMIP6 (Srivastava et al., 2020; Wehner et al.,
 7   2020) and between annual mean precipitation errors and Rx1day errors for both generations of models
 8   (Wehner et al., 2020).
 9
10   In general, there is high confidence that historical simulations by CMIP5 and CMIP6 models of similar
11   horizontal resolutions are interchangeable in their performance in simulating the observed climatology of
12   extreme precipitation.
13
14
15   [START FIGURE 11.14 HERE]
16
17   Figure 11.14:Multi-model mean bias in annual maximum daily precipitation (Rx1day, %) for the period 1979-2014,
18                calculated as the difference between the CMIP6 multi-model mean and the average of available
19                observational or reanalysis products including (a) ERA5, (b) HadEX3, and (c) and REGEN. Bias is
20                expressed as the percent error relative to the long-term mean of the respective observational data
21                products. Brown indicates that models are too dry, while green indicates that they are too wet. Areas
22                without sufficient observational data are shown in grey. Adapted from Wehner et al. (2020) under the
23                terms of the Creative Commons Attribution license. Further details on data sources and processing are
24                available in the chapter data table (Table 11.SM.9).
25
26   [END FIGURE 11.14 HERE]
27
28
29   Studies using regional climate models (RCMs), for example, CORDEX (Giorgi et al., 2009) over Africa
30   (Dosio et al., 2015; Klutse et al., 2016; Pinto et al., 2016; Gibba et al., 2019), Australia, East Asia (Park et
31   al., 2016), Europe (Prein et al., 2016a; Fantini et al., 2018), and parts of North America (Diaconescu et al.,
32   2018) suggest that extreme rainfall events are better captured in RCMs compared to their host GCMs due to
33   their ability to address regional characteristics, for example, topography and coastlines. However, CORDEX
34   simulations do not show good skill over south Asia for heavy precipitation and do not add value with respect
35   to their GCM source of boundary conditions (Mishra et al., 2014a; Singh et al., 2017b). The evaluation of
36   models in simulating regional processes is discussed in detail in Chapter 10 (Section 10.3.3.4). The high-
37   resolution simulation of mid-latitude winter extreme precipitation over land is of similar magnitude to point
38   observations. Simulation of summer extreme precipitation has a high bias when compared with observations
39   at the same spatial scale. Simulated extreme precipitation in the tropics also appears to be too large,
40   indicating possible deficiencies in the parameterization of cumulus convection at this resolution. Indeed,
41   precipitation distributions at both daily and sub-daily time scales are much improved with a convection-
42   permitting model (Belušić et al., 2020) over west Africa (Berthou et al., 2019b), East Africa (Finney et al.,
43   2019), North America and Canada (Cannon and Innocenti, 2019; Innocenti et al., 2019) and over Belgium in
44   Europe (Vanden Broucke et al., 2019).
45
46   In summary, there is high confidence in the ability of models to capture the large-scale spatial distribution of
47   precipitation extremes over land. The magnitude and frequency of extreme precipitation simulated by
48   CMIP6 models are similar to those simulated by CMIP5 models (high confidence).
49
50
51   11.4.4 Detection and attribution, event attribution
52
53   Both SREX (Chapter 3, Seneviratne et al., 2012) and AR5 (Chapter 10, IPCC, 2014) concluded with medium
54   confidence that anthropogenic forcing has contributed to a global-scale intensification of heavy precipitation
55   over the second half of the 20th century. These assessments were based on the evidence of anthropogenic
56   influence on aspects of the global hydrological cycle, in particular, the human contribution to the warming-
     Do Not Cite, Quote or Distribute                        11-57                                        Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   induced observed increase in atmospheric moisture that leads to an increase in heavy precipitation, and
 2   limited evidence of anthropogenic influence on extreme precipitation of durations of one and five days.
 3
 4   Since AR5 there has been new and robust evidence and improved understanding of human influence on
 5   extreme precipitation. In particular, detection and attribution analyses have provided consistent and robust
 6   evidence of human influence on extreme precipitation of one- and five-day durations at global to continental
 7   scales. The observed increases in Rx1day and Rx5day over the Northern Hemisphere land area during 1951-
 8   2005 can be attributed to the effect of combined anthropogenic forcing, including greenhouse gases and
 9   anthropogenic aerosols, as simulated by CMIP5 models and the rate of intensification with regard to
10   warming is consistent with C-C scaling (Zhang et al., 2013). This is confirmed to be robust when an
11   additional nine years of observational data and the CMIP6 model simulations were used (Paik et al., 2020;
12   CCB3.2, Figure 1). Additionally, the influence of greenhouse gases is attributed as the dominant contributor
13   to the observed intensification. The global average of Rx1day in the observations is consistent with
14   simulations by both CMIP5 and CMIP6 models under anthropogenic forcing, but not under natural forcing
15   (CCB3.2, Figure 1). The observed increase in the fraction of annual total precipitation falling into the top 5th
16   or top 1st percentiles of daily precipitation can also be attributed to human influence at the global scale (Dong
17   et al., 2020). CMIP5 models were able to capture the fraction of land experiencing a strong intensification of
18   heavy precipitation during 1960-2010 under anthropogenic forcing, but not in unforced simulations (Fischer
19   et al., 2014)). But the models underestimated the observed trends (Borodina et al., 2017a). Human influence
20   also significantly contributed to the historical changes in record-breaking one-day precipitation (Shiogama et
21   al., 2016). There is also limited evidence of the influences of natural forcing. Substantial reductions in
22   Rx5day and SDII (daily precipitation intensity) over the global summer monsoon regions occurred during
23   1957-2000 after explosive volcanic eruptions (Paik and Min, 2018). The reduction in post-volcanic eruption
24   extreme precipitation in the simulations is closely linked to the decrease in mean precipitation, for which
25   both thermodynamic effects (moisture reduction due to surface cooling) and dynamic effects (monsoon
26   circulation weakening) play important roles.
27
28   There has been new evidence of human influence on extreme precipitation at continental scales, including
29   the detection of the combined effect of greenhouse gases and aerosol forcing on Rx1day and Rx5day over
30   North America, Eurasia, and mid-latitude land regions (Zhang et al., 2013) and of greenhouse gas forcing in
31   Rx1day and Rx5day in the mid-to-high latitudes, western and eastern Eurasia, and the global dry regions
32   (Paik et al., 2020). These findings are corroborated by the detection of human influence in the fraction of
33   extreme precipitation in the total precipitation over Asia, Europe, and North America (Dong et al., 2020).
34   Human influence was found to have contributed to the increase in frequency and intensity of regional
35   precipitation extremes in North America during 1961-2010, based on both optimal fingerprinting and event
36   attribution approaches (Kirchmeier-Young and Zhang, 2020). Tabari et al. (2020) found the observed
37   latitudinal increase in extreme precipitation over Europe to be consistent with model-simulated responses to
38   anthropogenic forcing.
39
40   Evidence of human influence on extreme precipitation at regional scales is more limited and less robust. In
41   northwest Australia, the increase in extreme rainfall since 1950 can be related to increased monsoonal flow
42   due to increased aerosol emissions, but cannot be attributed to an increase in greenhouse gases (Dey et al.,
43   2019a). Anthropogenic influence on extreme precipitation in China was detected in one study (Li et al.,
44   2017), but it was not detected in another study (Li et al., 2018e) using different detection and data-processing
45   procedures, indicating the lack of robustness in the detection results. A still weak signal-to-noise ratio seems
46   to be the main cause for the lack of robustness, as detection would become robust 20 years in the future (Li
47   et al., 2018e). Krishnan et al. (2016) attributed the observed increase in heavy rain events (intensity > 100
48   mm/day) in the post-1950s over central India to the combined effects of greenhouse gases, aerosols, land use
49   and land cover changes, and rapid warming of the equatorial Indian Ocean SSTs. Roxy et al. (2017) and
50   Devanand et al. (2019) showed the increase in widespread extremes over the South Asian Monsoon during
51   1950-2015 is due to the combined impacts of the warming of the Western Indian Ocean (Arabian Sea) and
52   the intensification of irrigation water management over India
53
54   Anthropogenic influence may have affected the large-scale meteorological processes necessary for extreme
55   precipitation and the localized thermodynamic and dynamic processes, both contributing to changes in
     Do Not Cite, Quote or Distribute                     11-58                                       Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   extreme precipitation events. Several new methods have been proposed to disentangle these effects by either
 2   conditioning on the circulation state or attributing analogues. In particular, the extremely wet winter of
 3   2013/2014 in the UK can be attributed, approximately to the same degree, to both temperature-induced
 4   increases in saturation vapour pressure and changes in the large-scale circulation (Vautard et al., 2016; Yiou
 5   et al., 2017). There are multiple cases indicating that very extreme precipitation may increase at a rate more
 6   than the C-C rate (6-7%/ °C) (Pall et al., 2017; Risser and Wehner, 2017; van der Wiel et al., 2017; van
 7   Oldenborgh et al., 2017;Wang et al., 2018).
 8
 9   Event attribution studies found an influence of anthropogenic activities on the probability or magnitude of
10   observed extreme precipitation events, including European winters (Schaller et al., 2016; Otto et al., 2018b),
11   extreme 2014 precipitation over the northern Mediterranean (Vautard et al., 2015), parts of the US for
12   individual events (Knutson et al., 2014b; Szeto et al., 2015; Eden et al., 2016; van Oldenborgh et al., 2017),
13   extreme rainfall in 2014 over Northland, New Zealand (Rosier et al., 2016) or China (Burke et al., 2016; Sun
14   and Miao, 2018; Yuan et al., 2018b; Zhou et al., 2018). For other heavy rainfall events, however, studies
15   identified a lack of evidence about anthropogenic influences (Imada et al., 2013; Schaller et al., 2014; Otto et
16   al., 2015c; Siswanto et al., 2015). There are also studies whose results are inconclusive because of limited
17   reliable simulations (Christidis et al., 2013b; Angélil et al., 2016). Overall, both the spatial and temporal
18   scales on which extreme precipitation events are defined are important for attribution; events defined on
19   larger scales have larger signal-to-noise ratios and thus the signal is more readily detectable. At the current
20   level of global warming, there is a strong enough signal to be detectable for large-scale extreme precipitation
21   events, but the chance to detect such signals for smaller-scale events becomes smaller (Kirchmeier‐Young et
22   al., 2019).
23
24   In summary, most of the observed intensification of heavy precipitation over land regions is likely due to
25   anthropogenic influence, for which greenhouse gases emissions are the main contributor. New and robust
26   evidence since AR5 includes attribution of the observed increase in annual maximum one-day and five-day
27   precipitation and in the fraction of annual precipitation due to heavy events to human influence. It also
28   includes a larger fraction of land showing enhanced extreme precipitation and a larger probability of record-
29   breaking one-day precipitation than expected by chance, both of which can only be explained when
30   anthropogenic greenhouse gas forcing is considered. Human influence has contributed to the intensification
31   of heavy precipitation in three continents where observational data are more abundant, including North
32   America, Europe and Asia (high confidence). On the spatial scale of AR6 regions, evidence of human
33   influence on extreme precipitation is limited, but new evidence is emerging; in particular, studies attributing
34   individual heavy precipitation events found that human influence was a significant driver of the events,
35   particularly in the winter season.
36
37
38   11.4.5 Projections
39
40   AR5 concluded it is very likely that extreme precipitation events will be more frequent and more intense over
41   most of the mid-latitude land masses and wet tropics in a warmer world (Collins et al., 2013). Post-AR5
42   studies provide more and robust evidence to support the previous assessments. These include an observed
43   increase in extreme precipitation (11.4.3) and human causes of past changes (11.4.4), as well as projections
44   based on either GCM and/or RCM simulations. CMIP5 models project the rate of increase in Rx1day with
45   warming is independent of the forcing scenario (Pendergrass et al., 2015, Chapter 8, Section 8.5.3.1) or
46   forcing mechanism (Sillmann et al., 2017). This is confirmed in CMIP6 simulations (Li et al., 2020, and
47   Sillmann et al., 2019). In particular, for extreme precipitation that occurs once a year or less frequently, the
48   magnitudes of the rates of change per 1°C change in global mean temperature are similar regardless of
49   whether the temperature change is caused by increases in CO2, CH4, solar forcing, or SO4 (Sillmann et al.,
50   2019). In some models, CESM1 in particular, the extreme precipitation response to warming may follow a
51   quadratic relation (Pendergrass et al., 2019). Figure 11.15 shows changes in the 10-year and 50-year return
52   values of Rx1day at different warming levels as simulated by the CMIP6 models. The median value of the
53   scaling over land, across all SSP scenarios and all models, is close to 7%/°C for the 50-year return value of
54   Rx1day. It is just slightly smaller for the 10-year and 50-year return values of Rx5day (Li et al., 2020a). The
55   90% ranges of the multi-model ensemble changes across all land grid boxes in the 50-yr return values for
     Do Not Cite, Quote or Distribute                     11-59                                      Total pages: 345
     Final Government Distribution                         Chapter11                                       IPCC AR6 WGI

 1   Rx1day and Rx5day do not overlap between 1.5°C and 2°C warming levels (Li et al., 2020), indicating that a
 2   small increment such as 0.5°C in global warming can result in a significant increase in extreme precipitation.
 3   Projected long-period Rx1day return value changes are larger than changes in mean Rx1day and increase
 4   with increasing rarity (Pendergrass, 2018; Mizuta and Endo, 2020; Wehner, 2020). The rate of change of
 5   moderate extreme precipitation may depend more on the forcing agent, similar to the mean precipitation
 6   response to warming (Lin et al., 2016, 2018a). Thus, there is high confidence that extreme precipitation that
 7   occurs once a year or less frequently increases proportionally to the amount of surface warming and the rate
 8   of change in precipitation is not dependent on the underlying forcing agents of warming.
 9
10
11   [START FIGURE 11.15 HERE]
12
13
14   Figure 11.15:Projected changes in the intensity of extreme precipitation events under 1°C, 1.5°C, 2°C, 3°C, and 4°C
15                global warming levels relative to the 1851-1900 baseline. Extreme precipitation events are defined as the
16                daily precipitation (Rx1day) that was exceeded on average once during a 10-year period (10-year event,
17                blue) and once during a 50-year period (50-year event, orange) during the 1851-1900 base period. Results
18                are shown for the global land. For each box plot, the horizontal line and the box represent the median and
19                central 66% uncertainty range, respectively, of the intensity changes across the space, and the whiskers
20                extend to the 90% uncertainty range. The results are based on the multi-model ensemble median
21                estimated from simulations of global climate models contributing to the sixth phase of the Coupled Model
22                Intercomparison Project (CMIP6) under different SSP forcing scenarios. Based on (Li et al., 2020a).
23                Further details on data sources and processing are available in the chapter data table (Table 11.SM.9).
24
25
26   [END FIGURE 11.15 HERE]
27
28
29   The spatial patterns of the projected changes across different warming levels are quite similar, as shown in
30   Figure 11.16 and confirmed by near-linear scaling between extreme precipitation and global warming levels
31   at regional scales (Seneviratne and Hauser, 2020). Internal variability modulates changes in heavy rainfall
32   (Wood and Ludwig, 2020), resulting in different changes in different regions (Seneviratne and Hauser,
33   2020). Extreme precipitation nearly always increases across land areas with larger increases at higher global
34   warming levels, except in very few regions, such as southern Europe around the Mediterranean Basin in
35   some seasons. The very likely ranges of the multi-model ensemble changes across all land grid boxes in the
36   50-yr return values for Rx1day and Rx5day between 1.5°C and 1°C warming levels are above zero for all
37   continents expect Europe, with likely range above zero over Europe (Li et al., 2020). Decreases in extreme
38   precipitation are confined mostly to subtropical ocean areas and are highly correlated to decreases in mean
39   precipitation due to storm track shifts. These subtropical decreases can extend to nearby land areas in
40   individual realizations.
41
42   Projected increases in the probability of extreme precipitation of fixed magnitudes are non-linear and show
43   larger increases for more rare events (Figures 11.7 and 11.15, Fischer and Knutti, 2015, Li et al., 2020,
44   Kharin et al., 2018). CMIP5-model-projected increases in the probability of high (99th and 99.9th) percentile
45   precipitation between 1.5°C and 2°C warming scenarios are consistent with what can be expected based on
46   observed changes (Fischer and Knutti, 2015), providing confidence in the projections. CMIP5 model
47   simulations show that the frequency for present-day climate 20-year extreme precipitation is projected to
48   increase by 10% at the 1.5°C global warming level and by 22% at the 2.0°C global warming level, while the
49   increase in the frequency for present-day climate 100-year extreme precipitation is projected to increase by
50   20% and more than 45% at the 1.5°C and 2.0°C warming levels, respectively (Kharin et al., 2018). CMIP6
51   simulations with SSP scenarios show the frequency of 10-year and 50-year events will be approximately
52   doubled and tripled, respectively, at a very high warming level of 4°C (Figure 11.7, Li et al., 2020).
53
54   The number of studies on the projections of extreme hourly precipitation are limited. The ability of GCMs to
55   simulate hourly precipitation extremes is limited (Morrison et al., 2019) and very few modelling centres
56   archive sub-daily and hourly precipitation prior to CMIP6 experiments. RCM simulations project an increase
57   in extreme sub-daily precipitation in North America (Li et al., 2019a) and over Sweden (Olsson and Foster,
     Do Not Cite, Quote or Distribute                        11-60                                         Total pages: 345
     Final Government Distribution                       Chapter11                                     IPCC AR6 WGI

 1   2013), but these models still do not explicitly resolve convective processes that are important for properly
 2   simulating extreme sub-daily precipitation. Simulations by RCMs that explicitly resolve convective
 3   processes (convection-permitting models) are limited in length and only available in a few regions because
 4   of high computing costs. Yet, a majority of the available convection-permitting simulations project increases
 5   in the intensities of extreme sub-daily precipitation events with the amount similar to or higher than the C-C
 6   scaling rate (Ban et al., 2015; Helsen et al., 2020; Kendon et al., 2014, 2019; Prein et al., 2016b; Fowler et
 7   al., 2020). An increase is projected in extreme sub-daily precipitation over Africa (Kendon et al., 2019); over
 8   East Africa (Finney et al., 2020) and West Africa (Berthou et al., 2019a; Fitzpatrick et al., 2020), even for
 9   areas where parameterized RCMs project a decrease; in Europe (Chan et al., 2020 and Hodnebrog et al.,
10   2019); as well as in the continental US (Prein et al., 2016). Overall, available evidence, while limited, points
11   to an increase in extreme sub-daily precipitation in the future. Studies on future changes in extreme
12   precipitation for a month or longer are limited. One study projects an increase in extreme monthly
13   precipitation in Japan under 4°C global warming for around 80% of stations in the summer (Hatsuzuka and
14   Sato, 2019).
15
16   In Africa (Table 11.5), extreme precipitation will likely increase under warming levels of 2°C or below
17   (compared to pre-industrial values) and very likely increase at higher warming levels. Simulations by
18   CMIP5, CMIP6 and CORDEX regional models project an increase in daily extreme precipitation between
19   1.5°C and 2.0°C warming levels. The pattern of change in heavy precipitation under different scenarios or
20   warming levels is similar with larger increases for higher warming levels (e.g., Nikulin et al., 2018; Li et al.,
21   2020). With increases in warming, extreme precipitation is projected to increase in the majority of land
22   regions in Africa (Mtongori et al., 2016; Pfahl et al., 2017; Diedhiou et al., 2018; Dunning et al., 2018;
23   Akinyemi and Abiodun, 2019; Giorgi et al., 2019). Over southern Africa, heavy precipitation will likely
24   increase by the end of the 21st century under RCP 8.5 (Dosio, 2016; Pinto et al., 2016; Abiodun et al., 2017;
25   Dosio et al., 2019). However, heavy rainfall amounts are projected to decrease over western South Africa
26   (Pinto et al., 2018) as a result of a projected decrease in the frequency of the prevailing westerly winds south
27   of the continent that translates into fewer cold fronts and closed mid-latitudes cyclones (Engelbrecht et al.,
28   2009; Pinto et al., 2018). Heavy precipitation will likely increase by the end of the century under RCP8.5 in
29   West Africa (Diallo et al., 2016; Dosio, 2016; Sylla et al., 2016; Abiodun et al., 2017; Akinsanola and Zhou,
30   2018; Dosio et al., 2019) and is projected to increase (medium confidence) in central Africa (Fotso-Nguemo
31   et al., 2018, 2019; Sonkoué et al., 2019) and eastern Africa (Thiery et al., 2016; Ongoma et al., 2018a). In
32   northeast and central east Africa, extreme precipitation intensity is projected to increase across CMIP5,
33   CMIP6 and CORDEX-CORE (high confidence) in most areas annually (Coppola et al., 2021a), but the
34   trends differ from season to season in all future scenarios (Dosio et al., 2019). In northern Africa, there is low
35   confidence in the projected changes in heavy precipitation, either due to a lack of agreement among studies
36   on the sign of changes (Sillmann et al., 2013a; Giorgi et al., 2014) or due to insufficient evidence.
37
38   In Asia (Table 11.8), extreme precipitation will likely increase at global warming levels of 2°C and below,
39   but very likely increase at higher warming levels for the region as whole. The CMIP6 multi-model median
40   projects an increase in the 10- and 50-yr return values of Rx1day and Rx5day over more than 95% of
41   regions, even at the 2°C warming level, with larger increases at higher warming levels, independent of
42   emission scenarios (Li et al., 2020, also Figure 11.7). CMIP5 models produced similar projections. Both
43   heavy rainfall and rainfall intensity are projected to increase (Endo et al., 2017; Guo et al., 2016, 2018; Han
44   et al., 2018; Kim et al., 2018; Xu et al., 2016; Zhou et al., 2014). A half-degree difference in warming
45   between the 1.5°C and 2.0°C warming levels can result in a detectable increase in extreme precipitation over
46   the region (Li et al., 2020), in the Asian-Australian monsoon region (Chevuturi et al., 2018), and over South
47   Asia and China (Lee et al., 2018b; Li et al., 2018f). While there are regional differences, extreme
48   precipitation is projected to increase in almost all sub-regions, though there can be spatial heterogeneity
49   within sub-regions, such as in India (Shashikanth et al., 2018) and Southeast Asia (Ohba and Sugimoto,
50   2019). In East and Southeast Asia, there is high confidence that extreme precipitation is projected to intensify
51   (Guo et al., 2018; Li et al., 2018a; Seo et al., 2014; Sui et al., 2018; Wang et al., 2017b, 2017c; Xu et al.,
52   2016; Zhou et al., 2014, Nayak et al., 2017; Mandapaka and Lo, 2018; Raghavan et al., 2018; Tangang et al.,
53   2018; Supari et al., 2020). Extreme daily precipitation is also projected to increase in South Asia
54   (Shashikanth et al., 2018; Han et al., 2018; Xu et al., 2017). The extreme precipitation indices, including
55   Rx5day, R95p, and days of heavy precipitation (i.e., R10mm), are all projected to increase under the RCP4.5
     Do Not Cite, Quote or Distribute                      11-61                                       Total pages: 345
     Final Government Distribution                       Chapter11                                      IPCC AR6 WGI

 1   and RCP8.5 scenarios in central and northern Asia (Xu et al., 2017; Han et al., 2018). A general wetting
 2   across the whole Tibetan Plateau and the Himalaya is projected, with increases in heavy precipitation in the
 3   21st century (Zhou et al., 2014; Zhang et al., 2015c; Gao et al., 2018; Palazzi et al., 2013; Rajbhandari et al.,
 4   2015; Wu et al., 2017; Paltan et al., 2018). Agreement in projected changes by different models is low in
 5   regions of complex topography such as Hindu-Kush-Himalaya (Wester et al., 2019), but CMIP5, CMIP6 and
 6   CORDEX-CORE simulations consistently project an increase in heavy precipitation in higher latitude areas
 7   (WSB, ESB, RFE) (Coppola et al., 2021a) (high confidence).
 8
 9   In Australasia (Table 11.11), most CMIP5 models project an increase in Rx1day under RCP4.5 and RCP8.5
10   scenarios for the late 21st century (CSIRO and BOM, 2015; Alexander and Arblaster, 2017; Grose et al.,
11   2020) and the CMIP6 multi-model median projects an increase in the 10- and 50-yr return values of Rx1day
12   and Rx5day at a rate between 5-6% per degree celsius of near-surface global mean warming (Li et al., 2020,
13   also Figure 11.7). Yet, there is large uncertainty in the increase because projected changes in dynamic
14   processes lead to a decrease in Rx1day that can offsets the thermodynamic increase over a large portion of
15   the region (Pfahl et al., 2017, see also Box 11.1 Figure 1). Projected changes in moderate extreme
16   precipitation (the 99th percentile of daily precipitation) by RCMs under RCP8.5 for 2070-2099 are mixed,
17   with more regions showing decreases than increases (Evans et al., 2020). It is likely that daily rainfall
18   extremes such as Rx1day will increase at the continental scale for global warming levels at or above 3°C,
19   daily rainfall extremes are projected to increase at the 2.0°C global warming level (medium confidence), and
20   there is low confidence in changes at the 1.5°C. Projected changes show important regional differences with
21   very likely increases over NAU (Alexander and Arblaster, 2017; Herold et al., 2018; Grose et al., 2020) and
22   NZ (MfE, 2018) where projected dynamic contributions are small (Pfahl et al., 2017), see also Box 11.1
23   Figure 1) and medium confidence on increases over central, eastern, and southern Australia where dynamic
24   contributions are substantial and can affect local phenomena (CSIRO and BOM, 2015; Pepler et al., 2016;
25   Bell et al., 2019; Dowdy et al., 2019).
26
27   In Central and South America (Table 11.14), extreme precipitation will likely increase at global warming
28   levels of 2°C and below, but very likely increase at higher warming levels for the region as whole. A larger
29   increase in global surface temperature leads to a larger increase in extreme precipitation, independent of
30   emission scenarios (Li et al., 2020a). But there are regional differences in the projection and projected
31   changes for more moderate extreme precipitation are also more uncertain. Extreme precipitation, represented
32   by the R50mm and R90p extreme indices, is projected to increase on the eastern coast of SCA, but to decrease
33   along the Pacific coasts of El Salvador and Guatemala (Imbach et al., 2018). Chou et al. (2014) and Giorgi et
34   al. (2014) projected an increase in extreme precipitation over southeastern South America and the Amazon.
35   Projected changes in moderate extreme precipitation represented by the 99th percentile of daily precipitation
36   by different models under different emission scenarios, even at high warming levels, are mixed, with
37   increases projected for all regions by the CORDEX-CORE and CMIP5 simulations, but increases for some
38   regions and decreases for other regions by CMIP6 simulations (Coppola et al., 2021a). Extreme precipitation
39   is projected to increase in the La Plata basin (Cavalcanti et al., 2015; Carril et al., 2016). Taylor et al. (2018)
40   projected a decrease in days with intense rainfall in the Caribbean under 2°C global warming by the 2050s
41   under RCP4.5 relative to 1971-2000.
42
43   In Europe (Table 11.17), extreme precipitation will likely increase at global warming levels of 2°C and
44   below, but very likely increase for higher warming levels for the region as whole. The CMIP6 multi-model
45   median projects an increase in the 10- and 50-yr return values of Rx1day and Rx5day over a majority of the
46   region at the 2°C global warming level, with more than 95% of the region showing an increase at higher
47   warming levels (Li et al., 2020, also Figure 11.7). The most intense precipitation events observed today in
48   Europe are projected to almost double in occurrence for each degree celsius of further global warming
49   (Myhre et al., 2019). Extreme precipitation is projected to increase in both boreal winter and summer over
50   Europe (Madsen et al., 2014; OB et al., 2015; Nissen and Ulbrich, 2017). There are regional differences,
51   with decreases or no change for the southern part of Europe, such as the southern Mediterranean (Lionello
52   and Scarascia, 2020; Tramblay and Somot, 2018; Coppola et al., 2020), uncertain changes over central
53   Europe (Argüeso et al., 2012; Croitoru et al., 2013; Rajczak et al., 2013; Casanueva et al., 2014; Patarčić et
54   al., 2014; Paxian et al., 2014; Roth et al., 2014; Fischer and Knutti, 2015; Monjo et al., 2016) and a strong
55   increase in the remaining parts, including the Alps region (Gobiet et al., 2014; Donnelly et al., 2017),
     Do Not Cite, Quote or Distribute                       11-62                                       Total pages: 345
     Final Government Distribution                        Chapter11                                      IPCC AR6 WGI

 1   particularly in winter (Fischer et al., 2015), and northern Europe. In a 3°C warmer world, there will be a
 2   robust increase in extreme rainfall over 80% of land areas in northern Europe (Madsen et al., 2014; Donnelly
 3   et al., 2017; Cardell et al., 2020).
 4
 5   In North America (Table 11.20), the intensity and frequency of extreme precipitation will likely increase at
 6   the global warming levels of 2°C and below and very likely increase at higher warming levels. An increase is
 7   projected by CMIP6 model simulations (Li et al., 2020) and by previous model generations (Easterling et al.,
 8   2017; Wu, 2015; Zhang et al. 2018f; Innocenti et al., 2019b), as well as by RCMs (Coppola et al., 2020).
 9   Projections of extreme precipitation over the southern portion of the continent and over Mexico in particular
10   are more uncertain, with decreases possible (Alexandru, 2018; Sillmann et al., 2013b; Coppola et al., 2020).
11
12
13   [START FIGURE 11.16 HERE]
14
15   Figure 11.16:Projected changes in annual maximum daily precipitation at (a) 1.5°C, (b) 2°C, and (c) 4°C of global
16                warming compared to the 1851-1900 baseline. Results are based on simulations from the CMIP6 multi-
17                model ensemble under the SSP1-1.9, SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5 scenarios. The
18                numbers on the top right indicate the number of simulations included. Uncertainty is represented using
19                the simple approach: no overlay indicates regions with high model agreement, where ≥80% of models
20                agree on sign of change; diagonal lines indicate regions with low model agreement, where <80% of
21                models agree on sign of change. For more information on the simple approach, please refer to the Cross-
22                Chapter Box Atlas 1. For details on the methods see Supplementary Material 11.SM.2. Further details on
23                data sources and processing are available in the chapter data table (Table 11.SM.9).
24
25
26   [END FIGURE 11.16 HERE]
27
28
29   In summary, heavy precipitation will generally become more frequent and more intense with additional
30   global warming. At global warming levels of 4°C relative to the pre-industrial, very rare (e.g., 1 in 10 or
31   more years) heavy precipitation events would become more frequent and more intense than in the recent
32   past, on the global scale (virtually certain), and in all continents and AR6 regions: The increase in frequency
33   and intensity is extremely likely for most continents and very likely for most AR6 regions. The likelihood is
34   lower at lower global warming levels and for less-rare heavy precipitation events. At the global scale, the
35   intensification of heavy precipitation will follow the rate of increase in the maximum amount of moisture
36   that the atmosphere can hold as it warms (high confidence), of about 7% per °C of global warming.The
37   increase in the frequency of heavy precipitation events will accelerate with more warming and will be higher
38   for rarer events (high confidence), with 10-year and 50-year events to be approximately double and triple,
39   respectively, at the 4°C warming level. Increases in the intensity of extreme precipitation events at regional
40   scales will depend on the amount of regional warming as well as changes in atmospheric circulation and
41   storm dynamics leading to regional differences in the rate of heavy precipitation changes (high confidence).
42
43
44   11.5 Floods
45
46   Floods are the inundation of normally dry land and are classified into types (e.g., pluvial floods, flash floods,
47   river floods, groundwater floods, surge floods, coastal floods) depending on the space and time scales and
48   the major factors and processes involved (Chapter 8, Section 8.2.3.2, Nied et al., 2014; Aerts et al., 2018).
49   Flooded area is difficult to measure or quantify and, for this reason, many of the existing studies on changes
50   in floods focus on streamflow. Thus, this section assesses changes in flow as a proxy for river floods, in
51   addition to some types of flash floods. Pluvial and urban floods, types of flash floods resulting from the
52   precipitation intensity exceeding the capacity of natural and artificial drainage systems, are directly linked to
53   extreme precipitation. Because of this link, changes in extreme precipitation are the main proxy for inferring
54   changes in pluvial and urban floods (see also Section 12.4, REF Chapter 12), assuming there is no additional
55   change in the surface condition. Changes in these types of floods are not assessed in this section, but can be
56   inferred from the assessment of changes in heavy precipitation in Section 11.4. Coastal floods due to extreme
57   sea levels and flood changes at regional scales are assessed in Chapter 12 (12.4).
     Do Not Cite, Quote or Distribute                       11-63                                        Total pages: 345
     Final Government Distribution                       Chapter11                                     IPCC AR6 WGI

 1
 2
 3   11.5.1 Mechanisms and drivers
 4
 5   Since AR5, the number of studies on understanding how floods may have changed and will change in the
 6   future has substantially increased. Floods are a complex interplay of hydrology, climate, and human
 7   management, and the relative importance of these factors is different for different flood types and regions.
 8
 9   In addition to the amount and intensity of precipitation, the main factors for river floods include antecedent
10   soil moisture (Paschalis et al., 2014; Berghuijs et al., 2016; Grillakis et al., 2016; Woldemeskel and Sharma,
11   2016) and snow water-equivalent in cold regions (Sikorska et al., 2015; Berghuijs et al., 2016). Other factors
12   are also important, including stream morphology (Borga et al., 2014; Slater et al., 2015), river and catchment
13   engineering (Pisaniello et al., 2012; Nakayama and Shankman, 2013; Kim and Sanders, 2016), land-use and
14   land-cover characteristics (Aich et al., 2016; Rogger et al., 2017) and changes (Knighton et al., 2019), and
15   feedbacks between climate, soil, snow, vegetation, etc. (Hall et al., 2014; Ortega et al., 2014; Berghuijs et al.,
16   2016; Buttle et al., 2016; Teufel et al., 2019). Water regulation and management have, in general, increased
17   resilience to flooding (Formetta and Feyen, 2019), masking effects of an increase in extreme precipitation on
18   flood probability in some regions, even though they do not eliminate very extreme floods (Vicente-Serrano
19   et al., 2017). This means that an increase in precipitation extremes may not always result in an increase in
20   river floods (Sharma et al., 2018; Do et al., 2020). Yet, as very extreme precipitation can become a dominant
21   factor for river floods, there can then be some correspondence in the changes in very extreme precipitation
22   and river floods (Ivancic and Shaw, 2015; Wasko and Sharma, 2017; Wasko and Nathan, 2019). This has
23   been observed in the western Mediterranean (Llasat et al., 2016), in China (Zhang et al., 2015a) and in the
24   US (Peterson et al., 2013a; Berghuijs et al., 2016; Slater and Villarini, 2016).
25
26   In regions with a seasonal snow cover, snowmelt is the main cause of extreme river flooding over large areas
27   (Pall et al., 2019). Extensive snowmelt combined with heavy and/or long-duration precipitation can cause
28   significant floods (Li et al., 2019b; Krug et al., 2020). Changes in floods in these regions can be uncertain
29   because of the compounding and competing effects of the responses of snow and rain to warming that affect
30   snowpack size: warming results in an increase in precipitation, but also a reduction in the time period of
31   snowfall accumulation (Teufel et al., 2019). An increase in atmospheric CO2 enhances water-use efficiency
32   by plants (Roderick et al., 2015; Milly and Dunne, 2016; Swann et al., 2016; Swann, 2018); this could
33   reduce evapotranspiration and contribute to the maintenance of soil moisture and streamflow levels under
34   enhanced atmospheric CO2 concentrations (Yang et al., 2019). This mechanism would suggest an increase in
35   the magnitude of some floods in the future (Kooperman et al., 2018). But this effect is uncertain as an
36   increase in leaf area index and vegetation coverage could also result in overall larger water consumption
37   (Mátyás and Sun, 2014; Mankin et al., 2019; Teuling et al., 2019), and there are also other CO2-related
38   mechanisms that come into play (Chapter 5, CC Box 5.1).
39
40   Various factors, such as extreme precipitation (Cho et al., 2016; Archer and Fowler, 2018), glacier lake
41   outbursts (Schneider et al., 2014; Schwanghart et al., 2016), or dam breaks (Biscarini et al., 2016) can cause
42   flash floods. Very intense rainfall, along with a high fraction of impervious surfaces can result in flash floods
43   in urban areas (Hettiarachchi et al., 2018). Because of this direct connection, changes in very intense
44   precipitation can translate to changes in urban flood potential (Rosenzweig et al., 2018), though there can be
45   a spectrum of urban flood responses to this flood potential (Smith et al., 2013), as many factors such as the
46   overland flow rate and the design of urban (Falconer et al., 2009) and storm water drainage systems
47   (Maksimović et al., 2009) can play an important role. Nevertheless, changes in extreme precipitation are the
48   main proxy for inferring changes in some types of flash floods , which are addressed in Chapter 12 (Section
49   12.4)), given the relation between extreme precipitation and pluvial floods, the very limited literature on
50   urban and pluvial floods (e.g., Skougaard Kaspersen et al., 2017), and limitations of existing methodologies
51   for assessing changes in floods (Archer et al., 2016).
52
53   In summary, there is not always a one-to-one correspondence between an extreme precipitation event and a
54   flood event, or between changes in extreme precipitation and changes in floods, because floods are affected
55   by many factors in addition to heavy precipitation (high confidence). Changes in extreme precipitation may
     Do Not Cite, Quote or Distribute                      11-64                                       Total pages: 345
     Final Government Distribution                        Chapter11                                      IPCC AR6 WGI

 1   be used as a proxy to infer changes in some types of flash floods that are more directly related to extreme
 2   precipitation (high confidence).
 3
 4
 5   11.5.2 Observed trends
 6
 7   The SREX (Seneviratne et al., 2012) assessed low confidence for observed changes in the magnitude or
 8   frequency of floods at the global scale. This assessment was confirmed by the AR5 report (Hartmann et al.,
 9   2013). The SR15 (Hoegh-Guldberg et al., 2018) found increases in flood frequency and extreme streamflow
10   in some regions, but decreases in other regions. While the number of studies on flood trends has increased
11   since the AR5 report, and there were also new analyses after the release of SR15 (Berghuijs et al., 2017;
12   Blöschl et al., 2019; Gudmundsson et al., 2019), hydrological literature on observed flood changes is
13   heterogeneous, focusing at regional and sub-regional basin scales, making it difficult to synthesise at the
14   global and sometimes regional scales. The vast majority of studies focus on river floods using streamflow as
15   a proxy, with limited attention to urban floods. Streamflow measurements are not evenly distributed over
16   space, with gaps in spatial coverage, and their coverage in many regions of Africa, South America, and parts
17   of Asia is poor (e.g. Do et al., 2017), leading to difficulties in detecting long-term changes in floods (Slater
18   and Villarini, 2017). See also Chapter 8, Section 8.3.1.5.
19
20   Peak flow trends are characterized by high regional variability and lack overall statistical significance of a
21   decrease or an increase over the globe as a whole. Of more than 3500 streamflow stations in the US, central
22   and northern Europe, Africa, Brazil, and Australia, 7.1% stations showed a significant increase and 11.9%
23   stations showed a significant decrease in annual maximum peak flow during 1961-2005 (Do et al., 2017).
24   This is in direct contrast to the global and continental scale intensification of short-duration extreme
25   precipitation (11.4.2). There may be some consistency over large regions (see Gudmundsson et al., 2019), in
26   high streamflows (> 90th percentile), including a decrease in some regions (e.g., in the Mediterranean) and an
27   increase in others (e.g., northern Asia), but gauge coverage is often limited. On a continental scale, a
28   decrease seems to dominate in Africa (Tramblay et al., 2020) and Australia (Ishak et al., 2013; Wasko and
29   Nathan, 2019), an increase in the Amazon (Barichivich et al., 2018), and trends are spatially variable in other
30   continents (Do et al., 2017; Hodgkins et al., 2017; Bai et al., 2016; Zhang et al., 2015b). In Europe, flow
31   trends have large spatial differences (Hall et al., 2014; Mediero et al., 2015; Kundzewicz et al., 2018;
32   Mangini et al., 2018), but there appears to be a pattern of increase in northwestern Europe and a decrease in
33   southern and eastern Europe in annual peak flow during 1960-2000 (Blöschl et al., 2019). In North America,
34   peak flow has increased in the northeast US and decreased in the southwest US (Peterson et al., 2013a;
35   Armstrong et al., 2014; Mallakpour and Villarini, 2015; Archfield et al., 2016; Burn and Whitfield, 2016;
36   Wehner et al., 2017; Neri et al., 2019). There are important changes in the seasonality of peak flows in
37   regions where snowmelt dominates, such as northern North America (Burn and Whitfield, 2016; Dudley et
38   al., 2017) and northern Europe (Blöschl et al., 2017), corresponding to strong winter and spring warming.
39
40   In summary, the seasonality of floods has changed in cold regions where snowmelt dominates the flow
41   regime in response to warming (high confidence). Confidence about peak flow trends over past decades on
42   the global scale is low, but there are regions experiencing increases, including parts of Asia, southern South
43   America, the northeast USA, northwestern Europe, and the Amazon, and regions experiencing decreases,
44   including parts of the Mediterranean, Australia, Africa, and the southwestern USA.
45
46
47   11.5.3 Model evaluation
48
49   Hydrological models used to simulate floods are structurally diverse (Dankers et al., 2014; Mateo et al.,
50   2017; Şen, 2018), often requiring extensive calibration since sub-grid processes and land-surface properties
51   need to be parameterized, irrespective of the spatial resolutions (Döll et al., 2016; Krysanova et al., 2017).
52   The data that are used to drive and calibrate the models are usually of coarse resolution, necessitating the use
53   of a wide variety of downscaling techniques (Muerth et al., 2013). This adds uncertainty not only to the
54   models, but also to the reliability of the calibrations. The quality of the flood simulations also depends on the
55   spatial scale, as flood processes are different for catchments of different sizes. It is more difficult to replicate
     Do Not Cite, Quote or Distribute                       11-65                                        Total pages: 345
     Final Government Distribution                        Chapter11                                      IPCC AR6 WGI

 1   flood processes for large basins, as water management and water use are often more complex for these
 2   basins.
 3
 4   Studies that use different regional hydrological models show large spread in flood simulations (Dankers et
 5   al., 2014; Roudier et al., 2016; Trigg et al., 2016; Krysanova et al., 2017). Regional models reproduce
 6   moderate and high flows (0.02 – 0.1 flow annual exceedance probabilities) reasonably well, but there are
 7   large biases for the most extreme flows (0 - 0.02 annual flow exceedance probability), independent of the
 8   climatic and physiographic characteristics of the basins (Huang et al., 2017)). Global-scale hydrological
 9   models have even more challenges, as they struggle to reproduce the magnitude of the flood hazard (Trigg et
10   al., 2016). Additionally, the ensemble mean of multiple models does not perform better than individual
11   models (Zaherpour et al., 2018).
12
13   The use of hydrological models for assessing changes in floods, especially for future projections, adds
14   another dimension of uncertainty on top of uncertainty in the driving climate projections, including emission
15   scenarios, and uncertainty in the driving climate models (both RCMs and GCMs) (Arnell and Gosling, 2016;
16   Hundecha et al., 2016; Krysanova et al., 2017). The differences in hydrological models (Roudier et al., 2016;
17   Thober et al., 2018), as well as post-processing of climate model output for the hydrological models (Muerth
18   et al., 2013; Maier et al., 2018), both add to uncertainty for flood projections.
19
20   In summary, there is medium confidence that simulations for the most extreme flows by regional
21   hydrological models can have large biases. Global-scale hydrological models still struggle with reproducing
22   the magnitude of floods. Projections of future floods are hampered by these difficulties and cascading
23   uncertainties, including uncertainties in emission scenarios and the climate models that generate inputs.
24
25
26   11.5.4 Attribution
27
28   There are very few studies focused on the attribution of long-term changes in floods, but there are studies on
29   changes in flood events. Most of the studies focus on flash floods and urban floods, which are closely related
30   to intense precipitation events (Hannaford, 2015). In other cases, event attribution focused on runoff using
31   hydrological models, and examples include river basins in the UK (Schaller et al., 2016; Kay et al., 2018)
32   (See Section 11.4.4), the Okavango river in Africa (Wolski et al., 2014), and the Brahmaputra in Bangladesh
33   (Philip et al., 2019). Findings about anthropogenic influences vary between different regions and basins. For
34   some flood events, the probability of high floods in the current climate is lower than in a climate without an
35   anthropogenic influence (Wolski et al., 2014), while in other cases anthropogenic influence leads to more
36   intense floods (Cho et al., 2016; Pall et al., 2017; van der Wiel et al., 2017; Philip et al., 2018a; Teufel et al.,
37   2019). Factors such as land cover change and river management can also increase the probability of high
38   floods (Ji et al., 2020). These, along with model uncertainties and the lack of studies overall, suggest a low
39   confidence in general statements to attribute changes in flood events to anthropogenic climate change. Some
40   individual regions have been well studied, which allows for high confidence in the attribution of increased
41   flooding in these cases (Section 11.9 table). For example, flooding in the UK following increased winter
42   precipitation (Schaller et al., 2016; Kay et al., 2018) can be attributed to anthropogenic climate change
43   (Schaller et al., 2016; Vautard et al., 2016; Yiou et al., 2017; Otto et al., 2018b).
44
45   Attributing changes in heavy precipitation to anthropogenic activities (Section 11.4.4) cannot be readily
46   translated to attributing changes in floods to human activities, because precipitation is only one of the
47   multiple factors, albeit an important one, that affect floods. For example, Teufel et al. (2017) showed that
48   while human influence increased the odds of the flood-producing rainfall for the 2013 Alberta flood in
49   Canada, it was not detected to have influenced the probability of the flood itself. Schaller et al. (2016)
50   showed human influence on the increase in the probability of heavy precipitation translated linearly into an
51   increase in the resulting river flow of the Thames in winter 2014, but its contribution to the inundation was
52   inconclusive.
53
54   Gudmundsson et al. (2021) compared the spatial pattern of the observed regional trends in high river flows
55   (> 90th percentile) over 1971-2010 with those simulated by global hydrological models driven by outputs of
     Do Not Cite, Quote or Distribute                       11-66                                        Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   climate models under all historical forcing and with pre-industrial climate model simulations. They found
 2   complex spatial patterns of extreme river flow trends. They also found the observed spatial patterns of trends
 3   can be reproduced only if anthropogenic climate change is considered and that simulated effects of water and
 4   land management cannot reproduce the observed spatial pattern of trends. As there is only one study and
 5   multiple caveats, including relatively poor observational data coverage, there is low confidence about human
 6   influence on the changes in high river flows on the global scale.
 7
 8   In summary there is low confidence in the human influence on the changes in high river flows on the global
 9   scale. Confidence is in general low in attributing changes in the probability or magnitude of flood events to
10   human influence because of a limited number of studies and differences in the results of these studies, and
11   large modelling uncertainties.
12
13
14   11.5.5 Future projections
15
16   The SREX report (Chapter 3, Seneviratne et al., 2012) stressed the low availability of studies on flood
17   projections under different emission scenarios and concluded there was low confidence in projections of
18   flood events given the complexity of the mechanisms driving floods at the regional scale. The AR5 WGII
19   report (Chapter 3, Jimenez Cisneros et al., 2014) assessed with medium confidence the pattern of future flood
20   changes, including flood hazards increasing over about half of the globe (parts of southern and Southeast
21   Asia, tropical Africa, northeast Eurasia, and South America) and flood hazards decreasing in other parts of
22   the world, despite uncertainties in GCMs and their coupling to hydrological models. SR15 (Chapter 3,
23   Hoegh-Guldberg et al., 2018) assessed with medium confidence that global warming of 2°C would lead to an
24   expansion of the fraction of global area affected by flood hazards, compared to conditions at 1.5°C of global
25   warming, as a consequence of changes in heavy precipitation.
26
27   The majority of new studies that produce future flood projections based on hydrological models do not
28   typically consider aspects that are also important to actual flood severity or damages, such as flood
29   prevention measures (Neumann et al., 2015; Şen, 2018), flood control policies (Barraqué, 2017), and future
30   changes in land cover (see also Chapter 8, Section 8.4.1.5). At the global scale, Alfieri et al. (2017a) used
31   downscaled projections from seven GCMs as input to drive a hydrodynamic model. They found successive
32   increases in the frequency of high floods in all continents except Europe, associated with increasing levels of
33   global warming (1.5°C, 2°C, 4°C). These results are supported by Paltan et al. (2018), who applied a
34   simplified runoff aggregation model forced by outputs from four GCMs. Huang et al. (2018b) used three
35   hydrological models forced with bias-adjusted outputs from four GCMs to produce projections for four river
36   basins including the Rhine, Upper Mississippi, Upper Yellow, and Upper Niger under 1.5ºC, 2ºC, and 3°C
37   global warming. This study found diverse projections for different basins, including a shift towards earlier
38   flooding for the Rhine and the Upper Mississippi, a substantial increase in flood frequency in the Rhine only
39   under the 1.5ºC and 2°C scenarios, and a decrease in flood frequency in the Upper Mississippi under all
40   scenarios.
41
42   At the continental and regional scales, the projected changes in floods are uneven in different parts of the
43   world, but there is a larger fraction of regions with an increase than with a decrease over the 21st century
44   (Hirabayashi et al., 2013; Dankers et al., 2014; Arnell and Gosling, 2016; Döll et al., 2018). These results
45   suggest medium confidence in flood trends at the global scale, but low confidence in projected regional
46   changes. Increases in flood frequency or magnitude are identified for southeastern and northern Asia and
47   India (high agreement across studies), eastern and tropical Africa, and the high latitudes of North America
48   (medium agreement), while decreasing frequency or magnitude is found for central and eastern Europe and
49   the Mediterranean (high confidence), and parts of South America, southern and central North America, and
50   southwest Africa (Hirabayashi et al., 2013; Dankers et al., 2014; Arnell and Gosling, 2016; Döll et al., 2018).
51   Over South America, most studies based on global and regional hydrological models show an increase in the
52   magnitude and frequency of high flows in the western Amazon (Sorribas et al., 2016; Langerwisch et al.,
53   2013; Guimberteau et al., 2013; Zulkafli et al., 2016) and the Andes (Hirabayashi et al., 2013; Bozkurt et al.,
54   2018). Chapter 12, Section 12.4, provides a detailed assessment of regional flood projections.
55
     Do Not Cite, Quote or Distribute                     11-67                                      Total pages: 345
     Final Government Distribution                     Chapter11                                    IPCC AR6 WGI

 1   In summary, global hydrological models project a larger fraction of land areas to be affected by an increase
 2   in river floods than by a decrease in river floods (medium confidence). There is medium confidence that river
 3   floods will increase in the western Amazon, the Andes, and southeastern and northern Asia. Regional
 4   changes in river floods are more uncertain than changes in pluvial floods because complex hydrological
 5   processes and forcings, including land cover change and human water management, are involved.
 6
 7
 8   11.6 Droughts
 9
10   Droughts refer to periods of time with substantially below-average moisture conditions, usually covering
11   large areas, during which limitations in water availability result in negative impacts for various components
12   of natural systems and economic sectors (Wilhite and Pulwarty, 2017; Ault, 2020). Depending on the
13   variables used to characterize it and the systems or sectors being impacted, drought may be classified in
14   different types (Figure 8.6; Table 11.A.1) such as meteorological (precipitation deficits), agricultural (e.g.,
15   crop yield reductions or failure, often related to soil moisture deficits), ecological (related to plant water
16   stress that causes e.g., tree mortality), or hydrological droughts (e.g., water shortage in streams or storages
17   such as reservoirs, lakes, lagoons, and groundwater) (See Annex VII: Glossary). The distinction of drought
18   types is not absolute as drought can affect different sub-domains of the Earth system concomitantly, but
19   sometimes also asynchronously, including propagation from one drought type to another (Brunner and
20   Tallaksen, 2019). Because of this, drought cannot be characterized using a single universal definition (Lloyd-
21   Hughes, 2014) or directly measured based on a single variable (SREX Chapter 3; Wilhite and Pulwarty,
22   2017). Drought can happen on a wide range of timescales - from "flash droughts" on a scale of weeks, and
23   characterized by a sudden onset and rapid intensification of drought conditions (Hunt et al., 2014; Otkin et
24   al., 2018; Pendergrass et al., 2020) to multi-year or decadal rainfall deficits (sometimes termed
25   “megadroughts”; Annex VII: Glossary) (Ault et al., 2014; Cook et al., 2016b; Garreaud et al., 2017).
26   Droughts are often analysed using indices that are measures of drought severity, duration and frequency
27   (Table 11.A.1; Chapter 8, Sections 8.3.1.6, 8.4.1.6, Chapter 12, Sections 12.3.2.6 and 12.3.2.7). There are
28   many drought indices published in the scientific literature, as also highlighted in the IPCC SREX report
29   (SREX Chapter 3). These can range from anomalies in single variables (e.g., precipitation, soil moisture,
30   runoff, evapotranspiration) to indices combining different atmospheric variables.
31
32   This assessment is focused on changes in physical conditions and metrics of direct relevance to droughts
33   (Table 11.A.1): a) precipitation deficits, b) excess of atmospheric evaporative demand (AED), c) soil
34   moisture deficits, d) hydrological deficits, and e) atmospheric-based indices combining precipitation and
35   AED. In the regional tables (Section 11.9), the assessment is structured by drought types, addressing i)
36   meteorological, ii) agricultural and ecological, and iii) hydrological droughts. Note that the latter two
37   assessments are directly informing the Chapter 12 assessment on projected regional changes in these climatic
38   impact-drivers (Chapter 12, Section 12.4). The text refers to AR6 regions acronyms (Section 11.9, see
39   Chapter 1, Section 1.4.5) when referring to changes in AR6 regions.
40
41
42   11.6.1 Mechanisms and drivers
43
44   Similar to many other extreme events, droughts occur as a combination of thermodynamic and dynamic
45   processes (Box 11.1). Thermodynamic processes contributing to drought, which are modified by greenhouse
46   gas forcing both at global and regional scales, are mostly related to heat and moisture exchanges and also
47   partly modulated by plant coverage and physiology. They affect, for instance, atmospheric humidity,
48   temperature, and radiation, which in turn affect precipitation and/or evapotranspiration in some regions and
49   time frames. On the other hand, dynamic processes are particularly important to explain drought variability
50   on different time scales, from a few weeks (flash droughts) to multiannual (megadroughts). There is low
51   confidence in the effects of greenhouse gas forcing on changes in atmospheric dynamic (Chapter 2, Section
52   2.4; Chapter 4, Section 4.3.3), and, hence, on associated changes in drought occurrence. Thermodynamic
53   processes are thus the main driver of drought changes in a warming climate (high confidence).
54
55
     Do Not Cite, Quote or Distribute                     11-68                                     Total pages: 345
     Final Government Distribution                       Chapter11                                      IPCC AR6 WGI

 1   11.6.1.1 Precipitation deficits
 2
 3   Lack of precipitation is generally the main factor controlling drought onset. There is high confidence that
 4   atmospheric dynamics, which varies on interannual, decadal and longer time scales, is the dominant
 5   contributor to variations in precipitation deficits in the majority of the world regions (Dai, 2013; Seager and
 6   Hoerling, 2014; Miralles et al., 2014b; Burgman and Jang, 2015; Dong and Dai, 2015; Schubert et al., 2016;
 7   Raymond et al., 2018; Baek et al., 2019; Drumond et al., 2019; Herrera-Estrada et al., 2019; Gimeno et al.,
 8   2020; Mishra, 2020). Precipitation deficits are driven by dynamic mechanisms taking place on different
 9   spatial scales, including synoptic processes –atmospheric rivers and extratropical cyclones, blocking and
10   ridges (Section 11.7; Sousa et al., 2017), dominant large-scale circulation patterns (Kingston et al., 2015),
11   and global ocean-atmosphere coupled patterns such as IPO, AMO and ENSO (Dai and Zhao, 2017). These
12   various mechanisms occur on different scales, are not independent, and substantially interact with one
13   another. Also regional moisture recycling and land-atmosphere feedbacks play an important role for some
14   precipitation anomalies (see below).
15
16   There is high confidence that land-atmosphere feedbacks play a substantial or dominant role in affecting
17   precipitation deficits in some regions (SREX, Chapter3; Gimeno et al., 2012; Guillod et al., 2015; Haslinger
18   et al., 2019; Herrera-Estrada et al., 2019; Koster et al., 2011; Santanello Jr. et al., 2018; Taylor et al., 2012;
19   Tuttle and Salvucci, 2016). The sign of the feedbacks can be either positive or negative, as well as local or
20   non-local (Taylor et al., 2012; Guillod et al., 2015; Tuttle and Salvucci, 2016). ESMs tend to underestimate
21   non-local negative soil moisture-precipitation feedbacks (Taylor et al., 2012) and also show high variations
22   in their representation in some regions (Berg et al., 2017a). Soil moisture-precipitation feedbacks contribute
23   to changes in precipitation in climate model projections in some regions, but ESMs display substantial
24   uncertainties in their representation, and there is thus only low confidence in these contributions (Berg et al.,
25   2017a; Vogel et al., 2017, 2018).
26
27
28   11.6.1.2 Atmospheric evaporative demand
29
30   Atmospheric evaporative demand (AED) quantifies the maximum amount of actual evapotranspiration (ET)
31   that can happen from land surfaces if they are not limited by water availability (Table 11.A.1). AED is
32   affected by both radiative and aerodynamic components. For this reason, the atmospheric dryness, often
33   quantified with the relative humidity or the vapor pressure deficit (VPD), is not equivalent to the AED, as
34   other variables are also highly relevant, including solar radiation and wind speed (Hobbins et al., 2012;
35   McVicar et al., 2012b; Sheffield et al., 2012). AED can be estimated using different methods (McMahon et
36   al., 2013). Methods solely based on air temperature (e.g. Hargreaves, Thornthwaite) usually overestimate it
37   in terms of magnitude and temporal trends (Sheffield et al., 2012), in particular in the context of substantial
38   background warming. Physically-based combination methods such as the Penman-Monteith equation are
39   more adequate and recommended since 1998 by the Food and Agriculture Oganization (Pereira et al., 2015).
40   For this reason, the assessment of this chapter, when considering atmospheric-based drought indices, only
41   includes AED estimates using the latter (see also Section 11.9). AED is generally higher than ET, since it
42   represents an upper bound for it. Hence, an AED increase does not necessarily lead to increased ET (Milly
43   and Dunne, 2016), in particular under drought conditions given soil moisture limitation (Bonan et al., 2014;
44   Berg et al., 2016; Konings et al., 2017; Stocker et al., 2018). In general, AED is highest in regions where ET
45   is lowest (e.g., desert areas), further illustrating the decoupling between the two variables under limited soil
46   moisture.
47
48   The influence of AED on drought depends on the drought type, background climate, the environmental
49   conditions and the moisture availability (Hobbins et al., 2016, 2017; Vicente-Serrano et al., 2020b). This
50   influence also includes effects not related to increased ET. Under low soil moisture conditions, increased
51   AED increases plant stress, enhancing the severity of agricultural and ecological droughts (Williams et al.,
52   2013; Allen et al., 2015; McDowell et al., 2016; Grossiord et al., 2020). Moreover, high VPD impacts
53   overall plant physiology; it affects the leaf and xylem safety margins, and decreases the sap velocity and
54   plant hydraulic conductance (Fontes et al., 2018). VPD also affects the plant metabolism of carbon and if
55   prolongued, it may cause plant mortality via carbon starvation (Breshears et al., 2013; Hartmann, 2015).
     Do Not Cite, Quote or Distribute                      11-69                                        Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   Drought projections based exclusively on AED metrics overestimate changes in soil moisture and runoff
 2   deficits. Nevertheless, AED also directly impacts hydrological drought, as ET from surface waters is not
 3   limited (Wurbs and Ayala, 2014; Friedrich et al., 2018; Hogeboom et al., 2018; Xiao et al., 2018a), and this
 4   effect increases under climate change projections (Wang et al., 2018c; Althoff et al., 2020). In addition, high
 5   AED increases crop water consumptions in irrigated lands (García-Garizábal et al., 2014), contributing to
 6   intensifying hydrological droughts downstream (Fazel et al., 2017; Vicente-Serrano et al., 2017).
 7
 8   On subseasonal to decadal scales, temporal variations in AED are strongly controlled by circulation
 9   variability (Williams et al., 2014; Chai et al., 2018; Martens et al., 2018), but thermodynamic processes also
10   play a fundamental role and under human-induced climate change dominate the changes in AED.
11   Atmospheric warming due to increased atmospheric CO2 concentrations increases AED by means of
12   enhanced VPD in the absence of other influences (Scheff and Frierson, 2015). Indeed, because of the greater
13   warming over land than over oceans (Chapter 2, Section 2.3.1.1; Section 11.3), the saturation pressure of
14   water vapor increases more over land than over oceans; oceanic air masses advected over land thus contain
15   insufficient water vapour to keep pace with the greater increase in saturation vapour pressure over land
16   (Sherwood and Fu, 2014; Byrne and O’Gorman, 2018; Findell et al., 2019). Land-atmosphere feedbacks are
17   also important in affecting atmospheric moisture content and temperature, with resulting effects on relative
18   humidity and VPD (Berg et al., 2016; Haslinger et al., 2019; Zhou et al., 2019; Box 11.1).
19
20
21   11.6.1.3 Soil moisture deficits
22
23   Soil moisture shows an important correlation with precipitation variability (Khong et al., 2015; Seager et al.,
24   2019), but ET also plays a substantial role in further depleting moisture from soils, in particular in humid
25   regions during periods of precipitation deficits (Padrón et al., 2020; Teuling et al., 2013). In addition, soil
26   moisture plays a role in drought self-intensification under dry conditions in which ET is decreased and leads
27   to higher AED (Miralles et al., 2019), an effect that can also contribute to trigger “flash droughts” (Otkin et
28   al., 2016, 2018; DeAngelis et al., 2020; Pendergrass et al., 2020). If soil moisture becomes limited, ET is
29   reduced, which on one hand may decrease the rate of soil drying, but on the other hand can lead to further
30   atmospheric dryness through various feedback loops (Seneviratne et al., 2010; Miralles et al., 2014a, 2019;
31   Teuling, 2018; Vogel et al., 2018; Zhou et al., 2019b; Liu et al., 2020). The process is complex since
32   vegetation cover plays a role in modulating albedo and in providing access to deeper stores of water (both in
33   the soil and groundwater), and changes in land cover and in plant phenology may alter ET (Sterling et al.,
34   2013; Woodward et al., 2014; Frank et al., 2015; Döll et al., 2016; Ukkola et al., 2016; Trancoso et al., 2017;
35   Hao et al., 2019; Lian et al., 2020). Snow depth has strong and direct impacts on soil moisture in many
36   systems (Gergel et al., 2017; Williams et al., 2020).
37
38   Soil moisture directly affects plant water stress and ET. Soil moisture is the primary factor that controls
39   xylem hydraulic conductance, i.e. plant water uptake in plants (Sperry et al., 2016; Hayat et al., 2019; Chen
40   et al., 2020d). For this reason, soil moisture deficits are the main driver of xylem embolism, the primary
41   mechanism of plant mortality (Anderegg et al., 2012, 2016; Rowland et al., 2015). Also carbon assimilation
42   by plants strongly depend on soil moisture (Hartzell et al., 2017), with implications for carbon starvation and
43   plant dying if soil moisture deficits are prolongued (Sevanto et al., 2014). These mechanisms explain that
44   soil moisture deficits are usually more relevant than AED excess to explain gross primary production
45   anomalies and vegetation stress, mostly in sub-humid and semi-arid regions (Stocker et al. 2018; Liu et al.,
46   2020b). CO2 concentrations are shown to potentially decrease plant ET and increase plant water-use
47   efficiency, affecting soil moisture levels, although this effect interacts with other CO2 physiological and
48   radiative effects (Section 11.6.5.2; Chapter 5, CC Box 5.1), and has less relevance under low soil moisture
49   (Morgan et al., 2011; Xu et al., 2016b; Nackley et al., 2018; Dikšaitytė et al., 2019). ESMs represent both
50   surface (ca. 10cm) and total column soil moisture, whereby total soil moisture is of more direct relevance for
51   root water uptake, in particular by trees. There is evidence that surface soil moisture projections are
52   substantially drier than total soil moisture projections, and may thus overestimate drying of relevance for
53   most vegetation (Berg et al., 2017b).
54
55
     Do Not Cite, Quote or Distribute                     11-70                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   11.6.1.4 Hydrological deficits
 2
 3   Drivers of streamflow and surface water deficits are complex and strongly depend on the hydrological
 4   system analysed (e.g., streamflows in the headwaters, medium course of the rivers, groundwater, highly
 5   regulated hydrological basins). Soil hydrological processes, which control the propagation of meteorological
 6   droughts throughout different parts of the hydrological cycle (Van Loon and Van Lanen, 2012), are spatially
 7   and temporally complex (Herrera‐Estrada et al., 2017; Huang et al., 2017c) and difficult to quantify (Van
 8   Lanen et al., 2016; Apurv et al., 2017; Caillouet et al., 2017; Konapala and Mishra, 2017; Hasan et al.,
 9   2019). The physiographic characteristics of the basins also affect how droughts propagate throughout the
10   hydrological cycle (Van Loon and Van Lanen, 2012; Van Lanen et al., 2013; Van Loon, 2015; Konapala and
11   Mishra, 2020; Valiya Veettil and Mishra, 2020). In addition, the assessment of groundwater deficits is very
12   difficult given the complexity of processes that involve natural and human-driven feedbacks and interactions
13   with the climate system (Taylor et al., 2013). Streamflow and surface water deficits are affected by land
14   cover, groundwater and soil characteristics (Van Lanen et al., 2013; Van Loon and Laaha, 2015; Barker et
15   al., 2016; Tijdeman et al., 2018), as well as human activities (water management and demand, damming) and
16   land use changes (He et al., 2017; Jehanzaib et al., 2020; Van Loon et al., 2016; Veldkamp et al., 2017; Wu
17   et al., 2018; Xu et al., 2019b; Section 11.6.4.3). Finally, snow and glaciers are relevant for water resources in
18   some regions. For instance, warming affects snowpack levels (Dierauer et al., 2019; Huning and
19   AghaKouchak, 2020), as well as the timing of snow melt, thus potentially affecting the seasonality and
20   magnitude of low flows (Barnhart et al., 2016).
21
22
23   11.6.1.5 Atmospheric-based drought indices
24
25   Given difficulties of drought quantification and data constraints, atmospheric-based drought indices
26   combining both precipitation and AED have been developed, as they can be derived from meteorological
27   data that is available in most regions with few exceptions. These demand/supply indices are not intended to
28   be metrics of soil moisture, streamflow or vegetation water stress. Because of their reliance on precipitation
29   and AED, they are mostly related to the actual water balance in humid regions, in which ET is not limited by
30   soil moisture and tends towards AED. In water-limited regions and in dry periods everywhere, they
31   constitute an upper bound for overall water-balance deficits (e.g. of surface waters) but are also related
32   to conditions conducive to vegetation stress, particularly under soil moisture limitation (Section 11.6.1.2).
33
34   Although there are many atmospheric-based drought indices, two are assessed in this chapter: the Palmer
35   Drought Severity Index (PDSI) and the Standardized Precipitation Evapotranspiration Index (SPEI). The
36   PDSI has been widely used to monitor and quantify drought severity (Dai et al., 2018), but is affected by
37   some constraints (SREX Chapter 3; Mukherjee et al., 2018). Although the calculation of the PDSI is based
38   on a soil water budget, the PDSI is essentially a climate drought index that mostly responds to the
39   precipitation and the AED (van der Schrier et al., 2013; Vicente-Serrano et al., 2015; Dai et al., 2018). The
40   SPEI also combines precipitation and AED, being equally sensitive to these two variables (Vicente-Serrano
41   et al., 2015). The SPEI is more sensitive to AED than the PDSI (Cook et al., 2014a; Vicente-Serrano et al.,
42   2015), although under humid and normal precipitation conditions, the effects of AED on the SPEI are small
43   (Tomas-Burguera et al., 2020). Given the limitations associated with temperature-based AED estimates
44   (Section 11.6.1.2), only studies using the Penman-Monteith-based SPEI and PDSI (hereafter SPEI-PM and
45   PDSI-PM) are considered in this assessment and in the regional tables in Section 11.9.
46
47
48   11.6.1.6 Relation of assessed variables and metrics for changes in different drought types
49
50   This chapter assesses changes in meteorological drought, agricultural and ecological droughts, and
51   hydrological droughts. Precipitation-based indices are used for the estimation of changes in meteorological
52   droughts, such as the Standardized Precipitation Index (SPI) and the Consecutive Dry Days (CDD). Changes
53   in total soil moisture and soil moisture-based drought events are used for the estimation of changes in
54   agricultural and ecological droughts, complemented by changes in surface soil moisture, water-balance
55   estimates (precipitation minus ET), and SPEI-PM and PDSI-PM. For hydrological droughts, changes in low
     Do Not Cite, Quote or Distribute                      11-71                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   flows are assessed, sometimes complemented by changes in mean streamflow.
 2
 3   In summary, different drought types exist and they are associated with different impacts and respond
 4   differently to increasing greenhouse gas concentrations. Precipitation deficits and changes in
 5   evapotranspiration govern net water availability. A lack of sufficient soil moisture, sometimes amplified by
 6   increased atmospheric evaporative demand, result in agricultural and ecological drought. Lack of runoff and
 7   surface water result in hydrological drought. Drought events are both the result of dynamic and/or
 8   thermodynamic processes, with thermodynamic processes being the main driver of drought changes under
 9   human-induced climate change (high confidence).
10
11
12   11.6.2 Observed trends
13
14   Evidence on observed drought trends at the time of the SREX (Chapter 3) and AR5 (Chapter 2) was limited.
15   SREX concluded that “There is medium confidence that since the 1950s some regions of the world have
16   experienced a trend to more intense and longer droughts, in particular in southern Europe and West Africa,
17   but in some regions droughts have become less frequent, less intense, or shorter, for example, in central
18   North America and northwestern Australia”. The assessment at the time did not distinguish between different
19   drought types. This chapter includes numerous updates on observed drought trends, associated with
20   extensive new literature and longer datasets since the AR5.
21
22
23   11.6.2.1 Precipitation deficits
24
25   Strong precipitation deficits have been recorded in recent decades in the Amazon (2005, 2010), southwestern
26   China (2009-2010), southwestern North America (2011-2014), Australia (1997-2009), California (2014), the
27   middle East (2012-2016), Chile (2010-2015), the Great Horn of Africa (2011), among others (van Dijk et al.,
28   2013; Mann and Gleick, 2015; Rowell et al., 2015; Marengo and Espinoza, 2016; Dai and Zhao, 2017;
29   Garreaud et al., 2017, 2020; Marengo et al., 2017; Brito et al., 2018; Cook et al., 2018). Global studies
30   generally show no significant trends in SPI time series (Orlowsky and Seneviratne, 2013; Spinoni et al.,
31   2014), and in derived drought frequency and severity data (Spinoni et al., 2019), with very few regional
32   exceptions (Figure 11.17 and Section 11.9). Long-term decreases in precipitation are found in some AR6
33   regions in Africa (CAF, ESAF), and several regions in South America (NES, SAM, SWS, SSA) (Section
34   11.9). Evidence of precipitation-based drying trends is also found in Western Africa (WAF), consistent with
35   studies based on CDD trends (Chaney et al., 2014; Donat et al., 2014b; Barry et al., 2018; Dunn et al.,
36   2020)(Figure 11.17), however there is a partial recovery of the rainfall trends since the 1980s in this region
37   (Chapter 10, 10.4.2.1). Some AR6 regions show a decrease in meteorological drought, including NAU,
38   CAU, NEU and CNA (Section 11.9). Other regions do not show substantial trends in long-term
39   meteorological drought, or display mixed signals depending on the considered time frame and subregions,
40   such as in Southern Australia (SAU; Gallant et al., 2013; Delworth and Zeng, 2014; Alexander and
41   Arblaster, 2017; Spinoni et al., 2019; Dunn et al., 2020; Rauniyar and Power, 2020) and the Mediterranean
42   (MED; Camuffo et al., 2013; Gudmundsson and Seneviratne, 2016; Spinoni et al., 2017; Stagge et al., 2017;
43   Caloiero et al., 2018; Peña-Angulo et al., 2020) (see also Section 11.9 and Atlas 8.2).
44
45
46   11.6.2.2 Atmospheric evaporative demand
47
48   In several regions, AED increases have intensified recent drought events (Williams et al., 2014, 2020; Seager
49   et al., 2015b; Basara et al., 2019; García-Herrera et al., 2019), enhanced vegetation stress (Allen et al., 2015;
50   Sanginés de Cárcer et al., 2018; Yuan et al., 2019), or contributed to the depletion of soil moisture or runoff
51   through enhanced ET (Teuling et al., 2013; Padrón et al., 2020) (high confidence). Trends in pan evaporation
52   measurements and Penman-Monteith AED estimates provide an indication of possible trends in the influence
53   of AED on drought. Given the observed global temperature increases (Chapter 2; Section 2.3.1.1; Section
54   11.3) and dominant decrease in relative humidity over land areas (Simmons et al., 2010; Willett et al., 2014),
55   VPD has increased globally (Barkhordarian et al., 2019; Yuan et al., 2019). Pan evaporation has increased as
     Do Not Cite, Quote or Distribute                     11-72                                       Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   a consequence of VPD changes in several AR6 regions such as East Asia (EAS; Li et al., 2013; Sun et al.,
 2   2018; Yang et al., 2018a), West Central Europe (WCE; Mozny et al., 2020), MED; Azorin-Molina et al.,
 3   2015) and Central and Southern Australia (CAU, SAU; Stephens et al., 2018). Nevertheless, there is an
 4   important regional variability in observed trends, and in other AR6 regions pan evaporation has decreased
 5   (e.g. in North Central America, NCA (Breña-Naranjo et al., 2016) and in the Tibetan Plateau, TIB (Zhang et
 6   al., 2018a)). Physical models also show an important regional diversity, with an increase in New Zealand
 7   (NZ; Salinger, 2013) and the Mediterranean (MED; Gocic and Trajkovic, 2014; Azorin-Molina et al., 2015;
 8   Piticar et al., 2016), a decrease in SAS (Jhajharia et al., 2015), and strong spatial variability in North
 9   America (Seager et al., 2015b). This variability is driven by the role of other meteorological variables
10   affecting AED. Changes in solar radiation as a consequence of solar dimming and brightening may affect
11   trends (Kambezidis et al., 2012; Sanchez-Lorenzo et al., 2015; Wang and Yang, 2014; Chapter 7, Section
12   7.2.2.2). Wind speed is also relevant (McVicar et al., 2012a), and studies suggest a reduction of the wind
13   speed in some regions (Zhang et al., 2019h) that could compensate the role of the VPD increase.
14   Nevertheless, the VPD trend seems to dominate the overall AED trends, compared to the effects of trends in
15   wind speed and solar radiation (Wang et al., 2012; Park Williams et al., 2017; Vicente-Serrano et al., 2020b).
16
17
18   11.6.2.3 Soil moisture deficits
19
20   There are limited long-term measurements of soil moisture from ground observations (Dorigo et al., 2011;
21   Qiu et al., 2016; Quiring et al., 2016), which impedes their use in the analysis of trends. Among the few
22   existing observational studies covering at least two decades, several studies have investigated trends in
23   ground soil moisture in East Asia (Section 11.9; (Chen and Sun, 2015b; Liu et al., 2015; Qiu et al., 2016)).
24   Alternatively, microwave-based satellite measurements of surface soil moisture have also been used to
25   analyse trends (Dorigo et al., 2012; Jia et al., 2018). Although there is regional evidence that microwave-
26   based soil moisture estimates can capture well drying trends in comparison with ground soil moisture
27   observations (Jia et al., 2018), there is only medium confidence in the derived trends, since satellite soil
28   moisture data are affected by inhomogeneities (Dorigo et al., 2015; Rodell et al., 2018; Preimesberger et al.,
29   2020). Furthermore, microwave-based satellites only sense surface soil moisture, which differs from root-
30   zone soil moisture (Berg et al., 2017b), although relationships can be derived between the two (Brocca et al.,
31   2011). Several studies have also analysed long-term soil moisture timeseries from observations-driven land-
32   surface or hydrological models, including land-based reanalysis products (Albergel et al., 2013; Jia et al.,
33   2018; Gu et al., 2019b; Markonis et al., 2021). Such models have also been used to assess changes in land
34   water availability, estimated as precipitation minus ET, which is equal to the sum of soil moisture and runoff
35   (Greve et al., 2014; Padrón et al., 2020).
36
37   Overall, evidence from global studies suggests that several land regions have been affected by increased soil
38   drying or water-balance in past decades, despite some spread among products (Albergel et al., 2013; Greve
39   et al., 2014; Gu et al., 2019b; Padrón et al., 2020). Drying has not only occurred in dry regions, but also in
40   humid regions (Greve et al., 2014). Some studies have specifically addressed changes in soil moisture at
41   regional scale (Section 11.9). For AR6 regions, several studies suggest an increase in the frequency and areal
42   extent of soil moisture deficits, with examples in East Asia (EAS; Cheng et al., 2015; Qin et al., 2015; Jia et
43   al., 2018), Western and Central Europe (WCE; Trnka et al., 2015b), and the Mediterranean (MED; Hanel et
44   al., 2018; Moravec et al., 2019; Markonis et al., 2021). Nonetheless, some analyses also show no long-term
45   trends in soil drying in some AR6 regions, e.g. in Eastern (ENA; Park Williams et al., 2017) and Central
46   North America (CNA; Seager et al., 2019), as well as in North-Eastern Africa (NEAF; Kew et al., 2021).
47   The soil moisture drying trends identified in both global and regional studies are generally related to
48   increases in ET (associated with higher AED) rather than decreases in precipitation, as identified on global
49   land for trends in water-balance in the dry season (Padrón et al., 2020), as well as for some regions (Teuling
50   et al., 2013; Cheng et al., 2015; Trnka et al., 2015a; Van Der Linden et al., 2019; Li et al., 2020c).
51
52   Evidence from observed or observations-derived trends in soil moisture and precipitation minus ET, are
53   combined with evidence from SPEI and PDSI-PM studies to derive regional assessments of changes in
54   agricultural and ecological droughts (Section 11.9). This assessment is summarized in Section 11.6.2.6.
55
     Do Not Cite, Quote or Distribute                     11-73                                     Total pages: 345
     Final Government Distribution                           Chapter11                                          IPCC AR6 WGI

 1
 2   11.6.2.4 Hydrological deficits
 3
 4   There is evidence based on streamflow records of increased hydrological droughts in East Asia (Zhang et al.,
 5   2018b) and southern Africa (Gudmundsson et al., 2019). In areas of Western and Central Europe and of
 6   Northern Europe, there is no evidence of changes in the severity of hydrological droughts since 1950 based
 7   on flow reconstructions (Caillouet et al., 2017; Barker et al., 2019) and observations (Vicente-Serrano et al.,
 8   2019). In the Mediterranean region, there is high confidence in hydrological drought intensification (Giuntoli
 9   et al., 2013; Gudmundsson et al., 2019; Lorenzo-Lacruz et al., 2013; Masseroni et al., 2020; Section 11.9). In
10   Southeastern South America there is a decrease in the severity of hydrological droughts (Rivera and Penalba,
11   2018). In North America, depending on the methods, datasets and study periods, there are differences
12   between studies that suggest an increase (Shukla et al., 2015; Udall and Overpeck, 2017) vs a decrease in
13   hydrological drought frequency (Mo and Lettenmaier, 2018), but in general there is strong spatial variability
14   (Poshtiri and Pal, 2016). Streamflow observation reference networks of near-natural catchments have also
15   been used to isolate the effect of climate trends on hydrological drought trends in a few regions, but these
16   show limited trends in Northern Europe and Western and Central Europe (Stahl et al., 2010; Bard et al.,
17   2015; Harrigan et al., 2018), North America (Dudley et al., 2020) and most of Australia with the exception
18   of Eastern and Southern Australia (Zhang et al., 2016c). Given the low availability of observations, there are
19   few studies analysing trends of drought severity in the groundwater. Nevertheless, some studies suggest a
20   noticeable response of groundwater droughts to climate variability (Lorenzo-Lacruz et al., 2017) and
21   increased drought frequency and severity associated with warming, probably as a consequence of enhanced
22   ET induced by higher AED (Maxwell and Condon, 2016). This is supported by studies in Northern Europe
23   (Bloomfield et al., 2019) and North America (Condon et al., 2020).
24
25
26   11.6.2.5 Atmospheric-based drought indices
27
28   Globally, trends in SPEI-PM and PDSI-PM suggest slightly higher increases of drought frequency and
29   severity in regions affected by drying over the last decades in comparison to the SPI (Dai and Zhao, 2017;
30   Spinoni et al., 2019; Song et al., 2020), mainly in regions of West and Southern Africa, the Mediterranean
31   and East Asia (Figure 11.17), which is consistent with observed soil moisture trends (Section 11.6.2.3).
32   These indices suggest that AED has contributed to increase the severity of agricultural and ecological
33   droughts compared to meteorological droughts (García-Herrera et al., 2019; Williams et al., 2020), reduce
34   soil moisture during the dry season (Padrón et al., 2020), increase plant water stress (Allen et al., 2015;
35   Grossiord et al., 2020; Solander et al., 2020) and trigger more severe forest fires (Abatzoglou and Williams,
36   2016; Turco et al., 2019; Nolan et al., 2020). A number of regional studies based on these drought indices
37   have also shown stronger drying trends in comparison to trends in precipitation-based indices in the
38   following AR6 regions (see also 11.9): NSA (Fu et al., 2013b; Marengo and Espinoza, 2016), SCA (Hidalgo
39   et al., 2017), WCA (Tabari and Aghajanloo, 2013; Sharafati et al., 2020), SAS (Niranjan Kumar et al.,
40   2013), NEAF (Zeleke et al., 2017), WSAF (Edossa et al., 2016), NWN and NEN (Bonsal et al., 2013), EAS
41   (Yu et al., 2014; Chen and Sun, 2015b; Li et al., 2020b; Liang et al., 2020; Wu et al., 2020b) and MED
42   (Kelley et al., 2015; Stagge et al., 2017; González-Hidalgo et al., 2018; Mathbout et al., 2018a).
43
44
45   [START FIGURE 11.17 HERE]
46
47   Figure 11.17:Observed linear trend for (a) consecutive dry days (CDD) during 1960-2018, (b) standardized
48                precipitation index (SPI) and (c) standardized precipitation-evapotranspiration index (SPEI) during 1951-
49                2016. CDD data are from the HadEx3 dataset (Dunn et al., 2020), trend calculation of CDD as in Figure
50                11.9 Drought severity is estimated using 12-month SPI (SPI-12) and 12-month SPEI (SPEI-12). SPI and
51                SPEI datasets are from Spinoni et al. (2019). The threshold to identify drought episodes was set at -1
52                SPI/SPEI units. Areas without sufficient data are shown in grey. No overlay indicates regions where the
53                trends are significant at p = 0.1 level. Crosses indicate regions where trends are not significant. For details
54                on the methods see Supplementary Material 11.SM.2. Further details on data sources and processing are
55                available in the chapter data table (Table 11.SM.9).
56
     Do Not Cite, Quote or Distribute                           11-74                                          Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   [END FIGURE 11.17 HERE]
 2
 3
 4   11.6.2.6 Synthesis for different drought types
 5
 6   Few AR6 regions show observed increases in meteorological drought (Section 11.9), mostly in Africa and
 7   South America (NES: high confidence; WAF, CAF, ESAF, SAM, SWS, SSA, SAS: medium confidence); a
 8   few others show a decrease (WSB, ESB, NAU, CAU, NEU, CNA: medium confidence). There are stronger
 9   signals indicating observed increases in agricultural and ecological drought (Section 11.9), which highlights
10   the role of increased ET, driven by increased AED, for these trends (Sections 11.6.2.3, 11.6.2.5). Past
11   increases in agricultural and ecological droughts are found on all continents and several regions (WAF, CAF,
12   WSAF, ESAF, WCA, ECA, EAS, SAU, MED, WCE, NES: medium confidence), while decreases are found
13   only in one AR6 region (NAU: medium confidence). The more limited availability of datasets makes it more
14   difficult to assess historical trends in hydrological drought at regional scale (Section 11.9). Increasing (MED:
15   high confidence; WAF, EAS, SAU: medium confidence) and decreasing (NEU, SES: medium confidence)
16   trends in hydrological droughts have only been observed in a few regions.
17
18   In summary, there is high confidence that AED has increased on average on continents, contributing to
19   increased ET and resulting water stress during periods with precipitation deficits, in particular during dry
20   seasons. There is medium confidence in increases in precipitation deficits in a few regions of Africa and
21   South America. Based on multiple evidence, there is medium confidence that agricultural and ecological
22   droughts have increased in several regions on all continents (WAF, CAF, WSAF, ESAF, WCA, ECA, EAS,
23   SAU, MED, WCE, NES: medium confidence), while there is only medium confidence in decreases in one
24   AR6 region (NAU). More frequent hydrological droughts are found in fewer regions (MED: high
25   confidence; WAF, EAS, SAU: medium confidence).
26
27
28   11.6.3 Model evaluation
29
30   11.6.3.1 Precipitation deficits
31
32   ESMs generally show limited performance and large spread in identifying precipitation deficits and
33   associated long-term trends in comparison with observations (Nasrollahi et al., 2015). Meteorological
34   drought trends in the CMIP5 ensemble showed substantial disagreements compared with observations
35   (Orlowsky and Seneviratne, 2013; Knutson and Zeng, 2018) including a tendency to overestimate drying, in
36   particular in mid- to high latitudes (Knutson and Zeng, 2018). CMIP6 models display a better performance in
37   reproducing long-term precipitation trends or seasonal dynamics in some studies in southern South America
38   (Rivera and Arnould, 2020), East Asia (Xin et al., 2020), southern Asia (Gusain et al., 2020), and
39   southwestern Europe (Peña-Angulo et al., 2020b), but there is still too limited evidence to allow for an
40   assessment of possible differences in performance between CMIP5 and CMIP6. Furthermore, ESMs are
41   generally found to underestimate the severity of precipitation deficits and the dry day frequencies in
42   comparison to observations (Fantini et al., 2018; Ukkola et al., 2018). This is probably related to
43   shortcomings in the simulation of persistent weather events in the mid-latitudes (Chapter 10, Section
44   10.3.3.3). In addition, ESMs also show a tendency to underestimate precipitation-based drought persistence
45   at monthly to decadal time scales (Ault et al., 2014; Moon et al., 2018). The overall inter-model spread in
46   the projected frequency of precipitation deficits is also substantial (Touma et al., 2015; Zhao et al., 2016;
47   Engström and Keellings, 2018). Moreover, there are spatial differences in the spread, which is higher in the
48   regions where enhanced drought conditions are projected and under high-emission scenarios (Orlowsky and
49   Seneviratne, 2013). Nonetheless, some event attribution studies have concluded that droughts at regional
50   scales can be adequately simulated by some climate models (Schaller et al., 2016; Otto et al., 2018c).
51
52
53   11.6.3.2 Atmospheric evaporative demand
54
55   There is only limited evidence on the evaluation of AED in state-of-the-art ESMs, which is performed on
     Do Not Cite, Quote or Distribute                     11-75                                      Total pages: 345
     Final Government Distribution                       Chapter11                                      IPCC AR6 WGI

 1   externally computed AED based on model output (Scheff and Frierson, 2015; Liu and Sun, 2016, 2017). An
 2   evaluation of average AED in 17 CMIP5 ESMs for 1981-1999 based on potential evaporation show that the
 3   models’ spatial patterns resemble the observations, but that the magnitude of potential evaporation displays
 4   strong divergence among models globally and regionally (Scheff and Frierson, 2015). The evaluation of
 5   AED in 12 CMIP5 ESMs with pan evaporation observations in East Asia for 1961-2000 (Liu and Sun, 2016,
 6   2017) show that the ESMs capture seasonal cycles well, but that regional AED averages are underestimated
 7   due to biases in the meteorological variables controlling the aerodynamic and radiative components of AED.
 8   CMIP5 ESMs also show a strong underestimation of atmospheric drying trends compared to reanalysis data
 9   (Douville and Plazzotta, 2017).
10
11
12   11.6.3.3 Soil moisture deficits
13
14   The performance of climate models for representing soil moisture deficits shows more uncertainty than for
15   precipitation deficits since in addition to the uncertainties related to cloud and precipitation processes, there
16   is uncertainty related to the representation of complex soil hydrological and boundary-layer processes (Van
17   Den Hurk et al., 2011; Lu et al., 2019; Quintana-Seguí et al., 2020). A limitation is also the lack of
18   observations, and in particular soil moisture, in most regions (Section 11.6.2.3), and the paucity of land
19   surface property data to parameterize land surface models, in particular soil types, soil properties and depth
20   (Xia et al., 2015). The spatial resolution of models is an additional limitation since the representation of
21   some land-atmosphere feedbacks and topographic effects requires detailed resolution (Nicolai‐Shaw et al.,
22   2015; Van Der Linden et al., 2019). Beside climate models, also land surface and hydrological models are
23   used to derive historical and projected trends in soil moisture and related land water variables (Albergel et
24   al., 2013; Cheng et al., 2015; Gu et al., 2019b; Padrón et al., 2020; Markonis et al., 2021; Pokhrel et al.,
25   2021).
26
27   Overall, there are contrasting results on the performance of land surface models and climate models in
28   representing soil moisture. Some studies suggest that soil moisture anomalies are well captured by land
29   surface models driven with observation-based forcing (Dirmeyer et al., 2006; Albergel et al., 2013; Xia et al.,
30   2014; Balsamo et al., 2015; Reichle et al., 2017; Spennemann et al., 2020), but other studies report limited
31   agreement in the representation of interannual soil moisture variability (Stillman et al., 2016; Yuan and
32   Quiring, 2017; Ford and Quiring, 2019) and noticeable seasonal differences in model skill (Xia et al., 2014,
33   2015) in some regions. Models with good skill can nonetheless display biases in absolute soil moisture (Xia
34   et al., 2014; Gu et al., 2019a), but these are not necessarily of relevance for the simulation of surface water
35   fluxes and drought anomalies (Koster et al., 2009). There is also substantial intermodel spread (Albergel et
36   al. 2013), particularly for the root-zone soil moisture (Berg et al., 2017b).
37
38   Regarding the performance of regional and global climate models, an evaluation of an ensemble of RCM
39   simulations for Europe (Stegehuis et al., 2013) shows that these models display too strong drying in early
40   summer, resulting in an excessive decrease of latent heat fluxes, with potential implications for more severe
41   droughts in dry environments (Teuling, 2018; Van Der Linden et al., 2019). Compared with a range of
42   observational ET estimates, CMIP5 models show an overestimation of ET on annual scale, but an ET
43   underestimation in boreal summer in many North-Hemisphere mid-latitude regions, also suggesting a
44   tendency towards excessive soil drying (Mueller and Seneviratne, 2014), consistent with identified biases in
45   soil moisture-temperature coupling (Donat et al., 2018; Vogel et al., 2018; Selten et al., 2020). Land surface
46   models used in ESM display a bias in their representation of the sensitivity of interannual land carbon uptake
47   to soil moisture conditions, which appears related to a limited range of soil moisture variations compared to
48   observations (Humphrey et al., 2018).
49
50   For future projections, the spread of soil moisture outputs among different ESMs is more important than
51   internal variability and scenario uncertainty, and the bias is strongly related to the sign of the projected
52   change (Ukkola et al., 2018; Lu et al., 2019; Selten et al., 2020). CMIP5 ESMs that project more drying and
53   warming in mid-latitude regions show a substantial bias in soil moisture-temperature coupling (Donat et al.,
54   2018; Vogel et al., 2018). Although CMIP6 and CMIP5 simulations for soil moisture changes are overall
55   similar, some differences are found in projections in a few regions (Cook et al., 2020)(see also Section 11.9).
     Do Not Cite, Quote or Distribute                      11-76                                       Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   There is still limited evidence to assess whether there are substantial differences in model performance in the
 2   two ensembles, but improvements in modeling aspects relevant for soil moisture have been reported for
 3   precipitation (11.6.3.2), and a better performance has been found in CMIP6 for the representation of long-
 4   term trends in soil moisture in the continental USA (Yuan et al., 2021). Despite the mentioned model
 5   limitations, the representation of soil moisture processes in ESMs uses physical and biological understanding
 6   of the underlying processes, which can represent well the temporal anomalies associated with temporal
 7   variability and trends in climate. In summary, there is medium confidence in the representation of soil
 8   moisture deficits in ESMs and related land surface and hydrological models.
 9
10
11   11.6.3.4 Hydrological deficits
12
13   Streamflow and groundwater are not directly simulated by ESMs, which only simulate runoff, but they are
14   generallyn represented in hydrological models (Prudhomme et al., 2014; Giuntoli et al., 2015), which are
15   typically driven in a stand-alone manner by observed or simulated climate forcing. The simulation of
16   hydrological deficits is much more problematic than the simulation of mean streamflow or peak flows
17   (Fundel et al., 2013; Stoelzle et al., 2013; Velázquez et al., 2013; Staudinger et al., 2015), since models tend
18   to be too responsive to the climate forcing and do not satisfactorily capture low flows (Tallaksen and Stahl,
19   2014). Simulations of hydrological drought metrics show uncertainties related to the contribution of both
20   GCMs and hydrological models (Bosshard et al., 2013; Giuntoli et al., 2015; Samaniego et al., 2017; Vetter
21   et al., 2017), but hydrological models forced by the same climate input data also show a large spread (Van
22   Huijgevoort et al., 2013; Ukkola et al., 2018). At the catchment scale, the hydrological model uncertainty is
23   higher than both GCM and downscaling uncertainty (Vidal et al., 2016), and the hydrological models show
24   issues in representing drought propagation throughout the hydrological cycle (Barella-Ortiz and Quintana
25   Seguí, 2019). A study on the evaluation of streamflow droughts in seven global (hydrological and land
26   surface) models compared with observations in near-natural catchments of Europe showed a substantial
27   spread among models, an overestimation of the number of drought events, and an underestimation of drought
28   duration and drought-affected area (Tallaksen and Stahl, 2014).
29
30
31   11.6.3.5 Atmospheric-based drought indices
32
33   A number of studies have analysed the ability of models to capture drought severity and trends based on
34   climatic drought indices. Given the limitations of ESMs in reproducing the dynamic of precipitation deficits
35   and AED (11.6.3.1, 11.6.3.2), atmospheric-based drought indices derived from ESM data for these two
36   variables are also affected by uncertainties and biases. A comparison of historical trends in PDSI-PM for
37   1950-2014 derived from CMIP3 and CMIP5 with respective estimates derived from observations (Dai and
38   Zhao, 2017) show a similar behaviour at global scale (long-term decrease), but low spatial agreement in the
39   trends except in a few regions (Mediterranean, South Asia, northwestern US). In future projections there is
40   an important spread in PDSI-PM and SPEI-PM among different models (Cook et al., 2014a).
41
42
43   11.6.3.6 Synthesis for different drought types
44
45   The performance of ESMs used to assessed changes in variables related to meteorological droughts,
46   agricultural and ecological droughts, and hydrological droughts, show the presence of biases and
47   uncertainties compared to observations, but there is medium confidence in their overall performance for
48   assessing drought projections given process understanding. Given the substantial inter-model spread
49   documented for all related variables, the consideration of multi-model projections increases the confidence
50   of model-based assessments, with only low confidence in assessments based on single models.
51
52   In summary, the evaluation of ESMs, land surface and hydrological models for the simulation of droughts is
53   complex, due to the regional scale of drought trends, their overall low signal-to-noise ratio, and the lack of
54   observations in several regions, in particular for soil moisture and streamflow. There is medium confidence
55   in the ability of ESMs to simulate trends and anomalies in precipitation deficits and AED, and also medium
     Do Not Cite, Quote or Distribute                     11-77                                      Total pages: 345
     Final Government Distribution                     Chapter11                                    IPCC AR6 WGI

 1   confidence in the ability of ESMs and hydrological models to simulate trends and anomalies in soil moisture
 2   and streamflow deficits, on global and regional scales.
 3
 4
 5   11.6.4 Detection and attribution, event attribution
 6
 7   11.6.4.1 Precipitation deficits
 8
 9   There are only two AR6 regions in which there is at least medium confidence that human-induced climate
10   change has contributed to changes in meteorological droughts (Section 11.9). In South-western South
11   America (SSW), there is medium confidence that human-induced climate change has contributed to an
12   increase in meteorological droughts (Boisier et al., 2016; Garreaud et al., 2020), while in Northern Europe
13   (NEU), there is medium confidence that it has contributed to a decrease in meteorological droughts
14   (Gudmundsson and Seneviratne, 2016) (Section 11.9). In other AR6 regions, there is inconclusive evidence
15   in the attribution of long-term trends, but a human contribution to single meteorological events or
16   subregional trends has been identified in some instances (Section 11.9; see also below). In the Mediterranean
17   (MED) region, some studies have identified a precipitation decline or increase in meteorological drought
18   probability for time frames since the early or mid 20th century and a possible human contribution to these
19   trends (Hoerling et al., 2012; Gudmundsson and Seneviratne, 2016; Knutson and Zeng, 2018), also on
20   subregional scale in Syria from 1930 to 2010 (Kelley et al., 2015). On the contrary, other studies have not
21   identified precipitation and meteorological drought trends in the region for the long-term (Camuffo et al.,
22   2013; Paulo et al., 2016; Vicente‐Serrano et al., 2021) and also from the mid 20th century (Norrant and
23   Douguédroit, 2006; Stagge et al., 2017). There is evidence of substantial internal variability in long-term
24   precipitation trends in the region (Section 11.6.2.1), which limits the attribution of human influence on
25   variability and trends of meteorological droughts from observational records (Kelley et al., 2012; Peña-
26   Angulo et al., 2020b). In addition, there are important subregional trends showing mixed signals (MedECC,
27   2020)(Section 11.9). The evidence thus leads to an assessment of low confidence in the attribution of
28   observed short-term changes in meteorological droughts in the region (Section 11.9). In North America, the
29   human influence on precipitation deficits is complex (Wehner et al., 2017), with low confidence in the
30   attribution of long-term changes in meteorological drought in AR6 regions (Lehner et al., 2018; Section
31   11.9). In Africa there is low confidence that human influence has contributed to the observed long-term
32   meteorological drought increase in Western Africa (Section 11.9; Chapter 10, Section 10.6.2). There is low
33   confidence in the attribution of the observed increasing trends in meteorological drought in Eastern Southern
34   Africa, but evidence that human-induced climate change has affected recent meteorological drought events
35   in the region (11.9).
36
37   Attribution studies for recent meteorological drought events are available for various regions. In Central and
38   Western Europe, a multi-method and multi-model attribution study on the 2015 Central European drought
39   did not find conclusive evidence for whether human-induced climate change was a driver of the rainfall
40   deficit, as the results depended on model and method used (Hauser et al., 2017). In the Mediterranean region,
41   a human contribution was found in the case of the 2014 meteorological drought in the southern Levant based
42   on a single-model study (Bergaoui et al., 2015). In Africa, there is some evidence of a contribution of human
43   emissions to single meteorological drought events, such as the 2015-2017 southern African drought (Funk et
44   al., 2018a; Yuan et al., 2018a; Pascale et al., 2020), and the three-year 2015-2017 drought in the western
45   Cape Town region of South Africa (Otto et al., 2018c). An attributable signal was not found in droughts that
46   occurred in different years with different spatial extents in the last decade in Northern and Southern East
47   Africa (Marthews et al., 2015; Uhe et al., 2017; Otto et al., 2018a; Philip et al., 2018b; Kew et al., 2021).
48   However, an attributable increase in 2011 long rain failure was identified (Lott et al., 2013). Further studies
49   have attributed some African meteorological drought events to large-scale modes of variability, such as the
50   strong 2015 El Niño (Philip et al., 2018; Box 11.4) and increased SSTs overall (Funk et al., 2015b, 2018b).
51   Natural variability was dominant in the California droughts of 2011/12-2013/14 (Seager et al., 2015a). In
52   Asia, no climate change signal was found in the record dry spell over Singapore-Malaysia in 2014 (Mcbride
53   et al., 2015) or the drought in central southwest Asia in 2013/2014 (Barlow and Hoell, 2015). Nevertheless,
54   the South East Asia drought of 2015 has been attributed to anthropogenic warming effects (Shiogama et al.,
55   2020). Recent droughts occurring in South America, specifically in the southern Amazon region in 2010
     Do Not Cite, Quote or Distribute                     11-78                                     Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   (Shiogama et al., 2013) and in Northeast South America in 2014 (Otto et al., 2015) and 2016 (Martins et al.,
 2   2018) were not attributed to anthropogenic climate change. Nevertheless, the central Chile drought between
 3   2010 and 2018 has been suggested to be partly associated to global warming (Boisier et al., 2016; Garreaud
 4   et al., 2020). The 2013 New Zealand meteorological drought was attributed to human influence by
 5   Harrington et al. (2014, 2016) based on fully coupled CMIP5 models, but, no corresponding change in the
 6   dry end of simulated precipitation from a stand-alone atmospheric model was found by Angélil et al. (2017).
 7
 8   Event attribution studies also highlight a complex interplay of anthropogenic and non-anthropogenic
 9   climatological factors for some events. For example, anthropogenic warming contributed to the 2014 drought
10   in North Eastern-Africa by increasing east African and west Pacific temperatures, and increasing the gradient
11   between standardized western and central Pacific SSTs causing reduced rainfall (Funk et al., 2015b). As
12   different methodologies, models and data sources have been used for the attribution of precipitation deficits,
13   Angélil et al. (2017) reexamined several events using a single analytical approach and climate model and
14   observational datasets. Their results showed a disagreement in the original anthropogenic attribution in a
15   number of precipitation deficit events, which increased uncertainty in the attribution of meteorological
16   droughts events.
17
18
19   11.6.4.2 Soil moisture deficits
20
21   There is a growing number of studies on the detection and attribution of long-term changes in soil moisture
22   deficits. Mueller and Zhang (2016) concluded that anthropogenic forcing contributed significantly to an
23   increase in the land surface area affected by soil moisture deficits, which can be reproduced by CMIP5
24   models only if anthropogenic forcings are involved. A similar assessment was provided globally by Gu et al.
25   (2019b) also using CMIP5 models. Padrón et al. (2019) analyzed long-term reconstructed and CMIP5
26   simulated dry season water availability, defined as precipitation minus ET (i.e., equivalent to soil moisture
27   and runoff availability), and found that patterns of changes in dry-season deficits in the recent three last
28   decades can only be explained by anthropogenic forcing and are mostly related to changes in ET. Similarly
29   Williams et al. (2020) concluded human-induced climate change contributed to the strong soil moisture
30   deficits recorded in the last two decades in western North America through VPD increases associated with
31   higher air temperatures and lower air humidity. There are few studies analysing the attribution of particular
32   episodes of soil moisture deficits to anthropogenic influence. Nevertheless, the available modeling studies
33   coincide in supporting an anthropogenic attribution associated with more extreme temperatures, exacerbating
34   AED and increasing ET, and thus depleting soil moisture, as observed in southern Europe in 2017 (García-
35   Herrera et al., 2019) and in Australia in 2018 (Lewis et al., 2019b) and 2019 (van Oldenborgh et al., 2021),
36   the latter event having strong implications in the propagation of widespread mega-fires (Nolan et al., 2020).
37
38
39   11.6.4.3 Hydrological deficits
40
41   It is often difficult to separate the role of climate trends from changes in land use, water management and
42   demand for changes in hydrological deficits, especially on regional scale. However, a global study based on
43   a recent multi-model experiment with global hydrological models and covering several AR6 regions
44   suggests a dominant role of anthropogenic radiative forcing for trends in low, mean and high flows, while
45   simulated effects of water and land management do not suffice to reproduce the observed spatial pattern of
46   trends (Gudmundsson et al., 2021). Regional studies also suggest that climate trends have been dominant
47   compared to land use and human water management for explaining trends in hydrological droughts in some
48   regions, for instance in Ethiopia (Fenta et al., 2017), in China (Xie et al., 2015), and in North America for the
49   Missouri and Colorado basins, as well as in California (Shukla et al., 2015; Udall and Overpeck, 2017;
50   Ficklin et al., 2018; Xiao et al., 2018a; Glas et al., 2019; Martin et al., 2020; Milly and Dunne, 2020).
51
52   In other regions the influence of human water uses can be more important to explain hydrological drought
53   trends (Liu et al., 2016b; Mohammed and Scholz, 2016). There is medium confidence that human-induced
54   climate change has contributed to an increase of hydrological droughts in the Mediterranean (Giuntoli et al.,
55   2013; Vicente-Serrano et al., 2014; Gudmundsson et al., 2017), but also medium confidence that changes in
     Do Not Cite, Quote or Distribute                     11-79                                       Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   land use and terrestrial water management contributed to these trends as well (Teuling et al., 2019; Vicente-
 2   Serrano et al., 2019; Section 11.9). A global study with a single hydrological model estimated that human
 3   water consumption has intensified the magnitude of hydrological droughts by 20%-40% over the last 50
 4   years, and that the human water use contribution to hydrological droughts was more important than climatic
 5   factors in the Mediterranean, and the central US, as well as in parts of Brazil (Wada et al., 2013). However,
 6   Gudmundsson et al. (2021) concluded that the contribution of human water use is smaller than that of
 7   anthropogenic climate change to explain spatial differences in the trends of low flows based on a multi-
 8   model analysis. There is still limited evidence and thus low confidence in assessing these trends at the scale
 9   of single regions, with few exceptions (Section 11.9).
10
11
12   11.6.4.4 Atmospheric-based drought indices
13
14   Different studies using atmospheric-based drought indices suggest an attributable anthropogenic signal,
15   characterized by the increased frequency and severity of droughts (Cook et al., 2018), associated to increased
16   AED (Section 11.6.4.2). The majority of studies are based on the PDSI-PM. Williams et al. (2015) and
17   Griffin and Anchukaitis (2014) concluded that increased AED has had an increased contribution to drought
18   severity over the last decades, and played a dominant role in the intensification of the 2012-2014 drought in
19   California. The same temporal pattern and physical mechanism was stressed by Li et al. (2017) in Central
20   Asia. Marvel et al. (2019) compared tree ring-based reconstructions of the PDSI-PM over the past
21   millennium with PDSI-PM estimates based on output from CMIP5 models, suggesting a contribution of
22   greenhouse gas forcing to the changes since the beginning of the 20th century, although characterized with
23   temporal differences that could be driven by temporal variations in the aerosol forcing, in agreement with the
24   dominant external forcings of aridification at global scale between 1950 and 2014 (Bonfils et al., 2020). In
25   the Mediterranean region there is medium confidence of drying attributable to antropogenic forcing as a
26   consequence of the strong AED increase (Gocic and Trajkovic, 2014; Liuzzo et al., 2014; Azorin-Molina et
27   al., 2015; Maček et al., 2018), which has enhanced the severity of drought events (Vicente-Serrano et al.,
28   2014; Stagge et al., 2017; González-Hidalgo et al., 2018). In particular, this effect was identified to be the
29   main driver of the intensification of the 2017 drought that affected southwestern Europe, and was attributed
30   to the human forcing (García-Herrera et al., 2019). Nangombe et al. (2020) and Zhang et al. (2020)
31   concluded from differences between precipitation and AED that anthropogenic forcing contributed to 2018
32   droughts that affected southern Africa and southeastern China, respectively, principally as consequence of
33   the high AED that characterised these two events.
34
35
36   11.6.4.5 Synthesis for different drought types
37
38   The regional evidence on attribution for single AR6 regions generally shows low confidence for a human
39   contribution to observed trends in meteorological droughts at regional scale, with few exceptions (Section
40   11.9). There is medium confidence that human influence has contributed to changes in agricultural and
41   ecological droughts and has led to an increase in the overall affected land area. At regional scales, there is
42   medium confidence in a contribution of human-induced climate change to increases in agricultural and
43   ecological droughts in the Mediterranean (MED) and Western North America (WNA) (Section 11.9). There
44   is medium confidence that human-induced climate change has contributed to an increase in hydrological
45   droughts in the Mediterranean region, but also medium confidence in contributions from other human
46   influences, including water management and land use (Section 11.9). Several meteorological and agricultural
47   and ecological drought events have been attributed to human-induced climate change, even in regions where
48   no long-term changes are detected (medium confidence). However, a lack of attribution to human-induced
49   climate change has also been shown for some events (medium confidence).
50
51   In summary, human influence has contributed to changes in water availability during the dry season over
52   land areas, including decreases over several regions due to increases in evapotranspiration (medium
53   confidence). The increases in evapotranspiration have been driven by increases in atmospheric evaporative
54   demand induced by increased temperature, decreased relative humidity and increased net radiation over
55   affected land areas (high confidence). There is low confidence that human influence has affected trends in
     Do Not Cite, Quote or Distribute                     11-80                                     Total pages: 345
     Final Government Distribution                       Chapter11                                     IPCC AR6 WGI

 1   meteorological droughts in most regions, but medium confidence that they have contributed to the severity of
 2   some single events. There is medium confidence that human-induced climate change has contributed to
 3   increasing trends in the probability or intensity of recent agricultural and ecological droughts, leading to an
 4   increase of the affected land area. Human-induced climate change has contributed to global-scale change in
 5   low flow, but human water management and land use changes are also important drivers (medium
 6   confidence).
 7
 8
 9   11.6.5 Projections
10
11   SREX (Chapter 3) asssessed with medium confidence projections of increased drought severity in some
12   regions, including southern Europe and the Mediterranean, central Europe, Central America and Mexico,
13   northeast Brazil, and southern Africa, and. low confidence elsewhere given large inter-model spread. AR5
14   (Chapters 11 and 12) also assessed large uncertainties in drought projections at the regional and global
15   scales. The assessment of drought mechanisms under future climate change scenarios depends on the model
16   used (Section 11.6.3). Moreover, uncertainties in drought projections are affected by the consideration of
17   plant physiological responses to increasing atmospheric CO2 (Greve et al., 2019; Mankin et al., 2019; Milly
18   and Dunne, 2016; Yang et al., 2020; Chapter 5, Cross-Chapter Box 5.1), the role of soil moisture-atmosphere
19   feedbacks for changes in water-balance and aridity (Berg et al., 2016; Zhou et al., 2021), and statistical
20   issues related to considered drought time scales (Vicente-Serrano et al., 2020a). Nonetheless, the extensive
21   literature available since AR5 allows a substantially more robust assessment of projected changes in
22   droughts, also subdivided in different drought types (meteorological drought, agricultural and ecological
23   drought, and hydrological drought). This includes assessments of projected changes in droughts, including
24   changes at 1.5°C, 2°C and 4°C of global warming, for all AR6 regions (Section 11.9). Projected changes
25   show increases in drought frequency and intensity in several regions as function of global warming (high
26   confidence). There are also substantial increases in drought hazard probability from 1.5°C to 2°C global
27   warming as well as for further additional increments of global warming (Figs. 11.18 and 11.19) (high
28   confidence). These findings are based both on CMIP5 and CMIP6 analyses (Section 11.9; Greve et al., 2018;
29   Wartenburger et al., 2017; Xu et al., 2019a), and strengthen the conclusions of the SR15 Ch3.
30
31
32   11.6.5.1 Precipitation deficits
33
34   Studies based on CMIP5, CMIP6 and CORDEX projections show a consistent signal in the sign and spatial
35   pattern of projections of precipitation deficits. Global studies based on these multi-model ensemble
36   projections (Orlowsky and Seneviratne, 2013; Martin, 2018; Spinoni et al., 2020; Ukkola et al., 2020;
37   Coppola et al., 2021b) show particularly strong signal-to-noise ratios for increasing meteorological droughts
38   in the following AR6 regions: MED, ESAF, WSAF, SAU, CAU, NCA, SCA, NSA and NES (Section 11.9).
39   There is also substantial evidence of changes in meteorological droughts at 1.5°C vs 2°C of global warming
40   from global studies (Wartenburger et al., 2017; Xu et al., 2019a). The patterns of projected changes in mean
41   precipitation are consistent with the changes in the drought duration, but they are not consistent with the
42   changes in drought intensity (Ukkola et al., 2020). In general, CMIP6 projections suggest a stronger increase
43   of the probability of precipitation deficits than CMIP5 projections (Cook et al., 2020; Ukkola et al., 2020).
44   Projections for the number of CDDs in CMIP6 (Figure 11.19) for different levels of global warming relative
45   to 1850-1900 show similar spatial patterns as projected precipitation deficits. The robustness of the patterns
46   in projected precipitation deficits identified in the global studies is also consistent with results from regional
47   studies (Giorgi et al., 2014; Marengo and Espinoza, 2016; Pinto et al., 2016; Huang et al., 2018a; Maúre et
48   al., 2018; Nangombe et al., 2018; Tabari and Willems, 2018; Abiodun et al., 2019; Dosio et al., 2019).
49
50   In Africa, a strong increase in the length of dry spells (CDD) is projected for 4°C of global warming over
51   most of the continent with the exception of central and eastern Africa (Giorgi et al., 2014; Han et al., 2019;
52   Sillmann et al., 2013; Section 11.9). In West Africa, a strong reduction of precipitation is projected
53   (Sillmann et al., 2013a; Diallo et al., 2016; Akinsanola and Zhou, 2018; Han et al., 2019; Todzo et al., 2020)
54   at 4°C of global warming, and CDD would increase with stronger global warming levels (Klutse et al.,
55   2018). The regions most strongly affected are Southern Africa (ESAF, WSAF; (Nangombe et al., 2018;
     Do Not Cite, Quote or Distribute                      11-81                                       Total pages: 345
     Final Government Distribution                     Chapter11                                   IPCC AR6 WGI

 1   Abiodun et al., 2019) and Northern Africa (part of MED region), with increases in meteorological droughts
 2   already at 1.5°C of global warming, and further increases with increasing global warming (Section 11.9).
 3   CDD is projected to increase more in the southern Mediterranean (northern Africa) than in the northern part
 4   of the Mediterranean region (Lionello and Scarascia, 2020).
 5
 6   In Asia, most AR6 regions show low confidence in projected changes in meteorological droughts at 1.5°C
 7   and 2°C of global warming, with a few regions displaying a decrease in meteorological droughts at 4°C of
 8   global warming (RAR, ESB, RFE, ECA; medium confidence), although there is a projected increase in
 9   meteorological droughts in Southeast Asia (SEA) at 4°C (medium confidence) (Section 11.9). In Southeast
10   Asia, an increasing frequency of precipitation deficits is projected as a consequence of an increasing
11   frequency of extreme El Niño (Cai et al., 2014a, 2015, 2018).
12
13   In central America, projections suggest an increase in mid-summer meteorological drought (Imbach et al.,
14   2018) and increased CDD (Nakaegawa et al., 2013; Chou et al., 2014a; Giorgi et al., 2014). In the Amazon,
15   there is also a projected increase in dryness (Marengo and Espinoza, 2016), which is the combination of a
16   projected increase in the frequency and geographic extent of meteorological drought in the eastern Amazon,
17   and an opposite trend in the West (Duffy et al., 2015). In southwestern South America, there is a projected
18   increase of the CDD (Chou et al., 2014a; Giorgi et al., 2014) and in Chile, drying is projected to prevail
19   (Boisier et al., 2018). In the South America monsoon region, an increase in CDD is projected (Chou et al.,
20   2014a; Giorgi et al., 2014), but a decrease is projected in southeastern and southern South America (Giorgi et
21   al., 2014). In Central America, mid summer meteorological drought is projected to intensify during 2071-
22   2095 for the RCP8.5 scenario (Corrales-Suastegui et al., 2019).
23
24   An increase in the frequency, duration and intensity of meteorological droughts is projected in southwest,
25   south and east Australia (Kirono et al., 2020; Shi et al., 2020). In Canada and most of the USA, and based on
26   the SPI, Swain and Hayhoe (2015) identified drier summer conditions in projections over most of the region,
27   and there is a consistent signal toward an increase in duration and intensity of droughts in southern North
28   America (Pascale et al., 2016; Escalante-Sandoval and Nuñez-Garcia, 2017). In California, more
29   precipitation variability is projected, characterised by increased frequency of consecutive drought and humid
30   periods (Swain et al., 2018).
31
32   Substantial increases in meteorological drought are projected in Europe, in particular in the Mediterranean
33   region already at 1.5°C of global warming (Section 11.9). In southern Europe, model projections display a
34   consistent drying among models (Russo et al., 2013; Hertig and Tramblay, 2017; Guerreiro et al., 2018a;
35   Raymond et al., 2019). In Western and Central Europe there is some spread in CMIP5 projections, with
36   some models projecting very strong drying and others close to no trend (Vogel et al., 2018), although CDD
37   is projected to increase in CMIP5 projections under the RCP 8.5 scenario (Hari et al., 2020). The overall
38   evidence suggests an increase in meteorological drought at 4°C in the WCE region (medium confidence;
39   Section 11.9).
40
41   Overall, based on both global and regional studies, several hot spot regions are identified displaying more
42   frequent and severe meteorological droughts with increasing with global warming, including several AR6
43   regions at 1.5°C (WSAF, ESAF, SAU, MED, NES) and 2°C of global warming (WSAF, ESAF, EAU, SAU,
44   MED, NCA, SCA, NSA, NES) (Section 11.9). At 4°C of global warming, there is also confidence in
45   increases in meteorological droughts in further regions (WAF, WCE, ENA, CAR, NWS, SAM, SWS, SSA;
46   Section 11.9), showing a geographical expansion of meteorological drought with increasing global warming.
47   Only few regions are projected to have less intense or frequent meteorological droughts (Section 11.9).
48
49
50   11.6.5.2 Atmospheric evaporative demand
51
52   Effects of AED on droughts in future projections is under debate. CMIP5 models project an AED increase
53   over the majority of the world with increasing global warming, mostly as a consequence of strong VPD
54   increases (Scheff and Frierson, 2015; Vicente-Serrano et al., 2020b). However, ET is projected to increase
55   less than AED in many regions, due to plant physiological responses related to i) CO2 effects on plant
     Do Not Cite, Quote or Distribute                    11-82                                     Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   photosynthesis and ii) soil moisture control on ET.
 2
 3   Several studies suggest that increasing atmospheric CO2 could lead to reduced leaf stomatal conductance,
 4   which would increase water-use efficiency and reduce plant water needs, thus limiting ET (Chapter 5, Cross-
 5   Chapter Box 5.1; Greve et al., 2017; Lemordant et al., 2018; Milly and Dunne, 2016; Roderick et al., 2015;
 6   Scheff et al., 2017; Swann, 2018; Swann et al., 2016). The implemention of a CO2-dependent land resistance
 7   parameter has been suggested for the estimation of AED (Yang et al., 2019). Nevertheless, there are other
 8   relevant mechanisms, as soil moisture deficits and VPD also play an importantrole in the control of the leaf
 9   stomatal conductance (Xu et al., 2016b; Menezes-Silva et al., 2019; Grossiord et al., 2020) and a number of
10   ecophysiological and anatomical processes affect the response of plant physiology under higheratmospheric
11   CO2 concentrations (Mankin et al., 2019; Menezes-Silva et al., 2019; Chapter 5, Cross-Chapter Box 5.1).
12   The benefits of the atmospheric CO2 for plant stress and agricultural and ecological droughts would be
13   minimal precisely during dry periods given stomatal closure in response to limited soil moisture (Allen et al.,
14   2015; Xu et al., 2016b). In addition, CO2 effects on plant stomatal conductance could not entirely
15   compensate the increased demand associated to warming (Liu and Sun, 2017); in large tropical and
16   subtropical regions (e.g. southern Africa, the Amazon, the Mediterranean and southern North America),
17   AED is projected to increase even considering the possible CO2 effects on the land resistance (Vicente-
18   Serrano et al., 2020b). Moreover, these CO2 effects would not affect the direct evaporation from soils and
19   water bodies, which is very relevant in the reservoirs of warm areas (Friedrich et al., 2018). Because of these
20   uncertainties, there is low confidence whether increased CO2-induced water-use efficiency in vegetation will
21   substantially reduce global plant transpiration and will diminish the frequency and severity of soil moisture
22   and streamflow deficits associated with the radiative effect of higher CO2 concentrations (Chapter 5, CC Box
23   5.1).
24
25   Another mechanism reducing the ET response to increased AED in projections is the control of soil moisture
26   limitations on ET, which leads to reduced stomatal conductance under water stress (Berg and Sheffield,
27   2018; Stocker et al., 2018; Zhou et al., 2021). This response may be further amplified through VPD-induced
28   decreases in stomatal conductance (Anderegg et al., 2020). However, the decreased stomatal conductance in
29   response to both soil moisture limitation and enhanced CO2 would further enhance AED (Sherwood and Fu,
30   2014; Berg et al., 2016; Teuling, 2018; Miralles et al., 2019), whereby the overall effects on AED in ESMs
31   are found to be of similar magnitude for soil moisture limitation and CO2 physiological effects on stomatal
32   conductance (Berg et al., 2016). Increased AED is thus both a driver and a feedback with respect to changes
33   in ET, complicating the interpretation of its role on drought changes with increasing CO2 concentrations and
34   global warming.
35
36
37   11.6.5.3 Soil moisture deficits
38
39   Areas with projected soil moisture decreases do not fully coincide with areas with projected precipitation
40   decreases, although there is substantial consistency in the respective patterns (Dirmeyer et al., 2013; Berg
41   and Sheffield, 2018). There are, however, more regions affected by increased soil moisture deficits (Figure
42   11.19) than precipitation deficits (CC-Box 11.1, Figures 2a,b,c), as a consequence of enhanced AED and the
43   associated increased ET, as highlighted by some studies (Dai et al., 2018; Orlowsky and Seneviratne, 2013;
44   Chapter 8, Section 8.2.2.1). Moisture in the top soil layer is projected to decrease more than precipitation at
45   all warming levels (Lu et al., 2019), extending the regions affected by severe soil moisture deficits over most
46   of south and central Europe (Lehner et al., 2017; Ruosteenoja et al., 2018; Samaniego et al., 2018; Van Der
47   Linden et al., 2019), southern North America (Cook et al., 2019), South America (Orlowsky and
48   Seneviratne, 2013), southern Africa (Lu et al., 2019), East Africa (Rowell et al., 2015), southern Australia
49   (Kirono et al., 2020), India (Mishra et al., 2014b) and East Asia (Cheng et al., 2015) (Figure 11.19).
50   Projected changes in total soil moisture display less widespread drying than those for surface soil moisture
51   (Berg et al., 2017b), but still more than for precipitation (CC-Box 11.1, Figures 2a,b,c). The severity of
52   droughts based on surface soil moisture in future projections is stronger than projections based on
53   precipitation and runoff (Dai et al., 2018; Vicente-Serrano et al., 2020a). Nevertheless, in many parts of the
54   world in which soil moisture is projected to decrease, the signal to noise ratio among models is low and only
55   in the Mediterranean, Europe, the southwestern United States, and southern Africa the projections show a
     Do Not Cite, Quote or Distribute                      11-83                                    Total pages: 345
     Final Government Distribution                         Chapter11                                      IPCC AR6 WGI

 1   high signal to noise ratio in soil moisture projections (Lu et al., 2019; (Figure 11.19). Increases in soil
 2   moisture deficits are found to be statistically signicant at regional scale in the Mediterranean region,
 3   Southern Africa and Western South America for changes as small as 0.5°C in global warming, based on
 4   differences between +1.5°C and +2°C of global warming (Wartenburger et al., 2017). Several other regions
 5   are affected when considering changes in droughts for higher changes in global warming (Figure 11.19;
 6   Section 11.9). Seasonal projections of drought frequency for boreal winter (DJF) and summer (JJA), from
 7   CMIP6 multimodel ensemble for 1.5ºC, 2ºC and 4ºC global warming levels, show contrasting trends (Fig
 8   11.19). In the boreal winter in the Northern Hemisphere, the areas affected by drying show high agreement
 9   with those characterized by increase in meteorological drought projections (Chapter 8, Figure 8.14; Chapter
10   12, Figure 12.4). On the contrary, in the boreal summer the drought frequency increases worldwide in
11   comparison to meteorological drought projections, with large areas of the Northern Hemisphere displaying a
12   high signal to noise ratio (low spead between models). This stresses the dominant influence of ET (as a result
13   of increased AED) in intensifying agricultural and ecological droughts in the warm season in many locations,
14   including mid- to high latitudes.
15
16   Increased soil moisture limitation and associated changes in droughts are projected to lead to increased
17   vegetation stress affecting the global land carbon sink in ESM projections (Green et al., 2019), with
18   implications for projected global warming (Cross-Chapter Box 5). There is high confidence that the global
19   land sink will become less efficient due to soil moisture limitations and associated agricultural and
20   ecological drought conditions in some regions in higher emission scenarios specially under global warming
21   levels above 4°C ; however, there is low confidence on how these water cycle feedbacks will play out in
22   lower emission scenarios (at 2°C global warming or lower) (Cross-Chapter Box5.1).
23
24
25   [START FIGURE 11.18 HERE]
26
27   Figure 11.18:Projected changes in the intensity (a) and frequency (b) of drought under 1°C, 1.5°C, 2°C, 3°C, and 4°C
28                global warming levels relative to the 1850-1900 baseline. Summaries are computed for the AR6 regions
29                in which there is at least medium confidence in increase in agriculture/ ecological drought at the 2°C
30                warming level (“drying regions”), including W. North-America, C. North-America, N. Central-America,
31                S. Central-America, N. South-America, N. E. South-America, South-American-Monsoon, S.W.South-
32                America, S.South-America, West & Central-Europe, Mediterranean, W.Southern-Africa, E.Southern-
33                Africa, Madagascar, E.Australia, S.Australia (c). A drought event is defined as a 10-year drought event
34                whose annual mean soil moisture was below its 10th percentile from the 1850-1900 base period. For each
35                box plot, the horizontal line and the box represent the median and central 66% uncertainty range,
36                respectively, of the frequency or the intensity changes across the multi-model ensemble, and the whiskers
37                extend to the 90% uncertainty range. The line of zero in (a) indicates no change in intensity, while the
38                line of one in (b) indicates no change in frequency. The results are based on the multi-model ensemble
39                estimated from simulations of global climate models contributing to the sixth phase of the Coupled Model
40                Intercomparison Project (CMIP6) under different SSP forcing scenarios. Intensity changes in (a) are
41                expressed as standard deviations of the interannualvariability in the period 1850-1900 of the
42                corresponding modelFor details on the methods see Supplementary Material 11.SM.2. Further details on
43                data sources and processing are available in the chapter data table (Table 11.SM.9).
44
45   [END FIGURE 11.18 HERE]
46
47
48   11.6.5.4 Hydrological deficits
49
50   Some studies support wetting tendencies as a response to a warmer climate when considering globally-
51   averaged changes in runoff over land (Roderick et al., 2015; Greve et al., 2017; Yang et al., 2018e), and
52   streamflow projections respond to enhanced CO2 concentrations in CMIP5 models (Yang et al., 2019).
53   Nevertheless, when focusing regionally on low-runoff periods, model projections also show an increase of
54   hydrological droughts in large world regions (Wanders and Van Lanen, 2015; Dai et al., 2018; Vicente-
55   Serrano et al., 2020a). In general, the frequency of hydrological deficits is projected to increase over most of
56   the continents, although with regionally and seasonally differentiated effects (Section 11.9), with medium
57   confidence of increase in the following AR6 regions:WCE, MED, SAU, WCA, WNA, SCA, NSA, SAM,
     Do Not Cite, Quote or Distribute                        11-84                                        Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   SWS, SSA, WSAF, ESAF and MDG (Section 11.9; Cook et al., 2019; Forzieri et al., 2014; Giuntoli et al.,
 2   2015; Marx et al., 2018; Prudhomme et al., 2014; Roudier et al., 2016; Wanders and Van Lanen, 2015; Zhao
 3   et al., 2020). There are, however, large uncertainties related to the hydrological/impact model used
 4   (Prudhomme et al., 2014; Schewe et al., 2014; Gosling et al., 2017), limited signal-to-noise ratio (due to
 5   model spread) in several regions (Giuntoli et al., 2015), and also uncertainties in the projection of future
 6   human activities including water demand and land cover changes, which may represent more than 50% of
 7   the projected changes in hydrological droughts in some regions (Wanders and Wada, 2015).
 8
 9   Regions dependent on mountainous snowpack as a temporary reservoir may be affected by severe
10   hydrological droughts in a warmer world. In the southern European Alps, both winter and summer low flows
11   are projected to be more severe, with a 25% decrease in the 2050s (Vidal et al., 2016). In the western United
12   States, a 22% reduction in winter snow water equivalent is projected at around 2°C of global warming with a
13   further decrease of a 70% reduction at 4°C global warming (Rhoades et al., 2018). This decline would cause
14   less predictable hydrological droughts in snowmelt-dominated areas of North America (Livneh and Badger,
15   2020). The exact magnitude of the influence of higher temperatures on snow-related droughts is, however,
16   difficult to estimate (Mote et al., 2016), since the streamflow changes could affect the timing of peak
17   streamflows but not necessarily their magnitude. In addition, projected changes in hydrological droughts
18   downstream of declining glaciers can be very complex to assess (Chapter 9, see also SROCC).
19
20
21   11.6.5.5 Atmospheric-based drought indices
22
23   Studies show a stronger drying in projections based on atmospheric-based drought indices compared to ESM
24   projections of changes in soil moisture (Berg and Sheffield, 2018) and runoff (Yang et al., 2019). It has been
25   suggested that this difference is due to physiological CO2 effects (Greve et al., 2019; Lemordant et al., 2018;
26   Milly and Dunne, 2016; Roderick et al., 2015; Scheff, 2018; Swann, 2018; Swann et al., 2016; Yang et al.,
27   2020; Section 11.6.5.2). Nonetheless, there is evidence that differences in projections between atmospheric-
28   based drought indices and water-balance metrics from ESMs are not alone due to CO2-plant effects (Berg et
29   al., 2016; Scheff et al., 2021), and can be also related to the fact that AED is an upper bound for ET in dry
30   regions and conditions (Section 11.6.1.2) and that soil moisture stress limits increases in ET in projections
31   (Berg et al., 2016; Zhou et al., 2021; Section 11.6.5.2). Atmospheric-based indices show in general more
32   drying than total column soil moisture (Berg and Sheffield, 2018; Cook et al., 2020; Scheff et al., 2021), but
33   are more consistent with projected increases in surface soil moisture deficits (Dirmeyer et al., 2013; Dai et
34   al., 2018; Lu et al., 2019; Cook et al., 2020; Vicente-Serrano et al., 2020a).
35
36   Atmospheric-based drought indices are not metrics of soil moisture or runoff (11.6.1.5) so their projections
37   may not necessarily reflect the same trend of online simulated soil moisture and runoff. Independently of
38   effects on the land water balance, atmospheric-based drought indices will reflect the potential vegetation
39   stress resulting from deficits between available water and enhanced AED, even in conditions with no or only
40   low ET. Under dry conditions, the enhanced AED associated to the human forcing would increase plant
41   water stress (Brodribb et al., 2020), with effects on widespread forest dieback and mortality (Anderegg et al.,
42   2013; Williams et al., 2013; Allen et al., 2015; McDowell and Allen, 2015; McDowell et al., 2016, 2020),
43   and stronger risk of megafires (Flannigan et al., 2016; Podschwit et al., 2018; Clarke and Evans, 2019;
44   Varela et al., 2019). For these reasons, there is high confidence that the future projections of enhanced
45   drought severity showed by the PDSI-PM and the SPEI-PM are representative of more frequent and severe
46   plant stress episodes and more severe agricultural and ecological drought impacts in some regions.
47
48   Global tendencies towards more severe and frequent agricultural and ecological drought conditions are
49   identified in future projections when focusing on atmospheric-based drought indices such as the PDSI-PM or
50   the SPEI-PM. They expand the spatial extent of drought conditions compared to meteorological drought to
51   most of North America, Europe, Africa, Central and East Asia and southern Australia (Cook et al., 2014a;
52   Chen and Sun, 2017b, 2017a; Zhao and Dai, 2017; Gao et al., 2017b; Lehner et al., 2017; Dai et al., 2018;
53   Naumann et al., 2018; Potopová et al., 2018; Vicente-Serrano et al., 2020a; Gu et al., 2020; Dai, 2021).
54   Projections in PDSI-PM and SPEI-PM are used in complement to changes in total soil moisture for the
55   assessed projected changes in agricultural and ecological drought (Section 11.9).
     Do Not Cite, Quote or Distribute                     11-85                                     Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1
 2
 3   11.6.5.6 Synthesis for different drought types
 4
 5   The tables in Section 11.9 provide assessed projected changes in metorological drought, agricultural and
 6   ecological drought, and hydrological droughts. The assessment shows that several regions will be affected by
 7   more severe agricultural and ecological droughts even if global warming is stabilized at well below 2°C, and
 8   1.5°C, within the bounds of the Paris Agreement (high confidence). The most affected regions include WCE,
 9   MED, EAU, SAU, SCA, NSA, SAM, SWS, SSA, NCA, CAN, WSAF, ESAF and MDG (medium
10   confidence). At 4°C of global warming, even more regions would be affected by agricultural and ecological
11   droughts (WCE, MED, CAU, EAU, SAU, WCA, EAS, SCA, CAR, NSA, NES, SAM, SWS, SSA, NCA,
12   CAN, ENA, WNA, WSAF, ESAF and MDG). NEAF, SAS are also projected to experience less agricultural
13   and ecological drought with global warming (medium confidence). Projected changes in meteorological
14   droughts are overall less extended but also affect several AR6 regions, at 1.5°C and 2°C (MED, EAU, SAU,
15   SCA, NSA, NCA, WSAF, ESAF, MDG) and 4ºC of global warming (WCE, MED, EAU, SAU, SEA, SCA,
16   CAR, NWS, NSA, NES, SAM, SWS, SSA, NCA, ENA, WAF, WSAF, ESAF, MDG). Several regions are
17   also projected to be affected by more hydrological droughts at 1.5°C and 2°C (WCE, MED, WNA, WSAF,
18   ESAF) and 4ºC of global warming (NEU, WCE, EEU, MED, SAU, WCA, SCA, NSA, SAM, SWS, SSA,
19   WNA, WSAF, ESAF, MDG). To illustrate the changes in both intensity and frequency of drought in the
20   regions where strongest changes are projected, Figure 11.18 displays changes in the intensity and frequency
21   of soil moisture drought under different global warming levels (1.5°C, 2°C, 4°C) relative to the 1851-1900
22   baseline based on CMIP6 simulations under different SSP forcing scenarios. The 90% uncertainty ranges for
23   the projected changes in both intensity and frequency are above zero, indicating significant increase in both
24   intensity and frequency of drought in these regions as whole.
25
26   In summary, the land area affected by increasing drought frequency and severity expands with increasing
27   global warming (high confidence). New evidence strengthens the SR15 conclusion that even relatively small
28   incremental increases in global warming (+0.5°C) cause a worsening of droughts in some regions (high
29   confidence). Several regions will be affected by more frequent and severe agricultural and ecological
30   droughts even if global warming is stabilized at 1.5-2°C (high confidence). The most affected regions
31   include WCE, MED, EAU, SAU, SCA, NSA, SAM, SWS, SSA, NCA, CAN, WSAF, ESAF and MDG
32   (medium confidence). At 4°C of global warming, even more regions would be affected by agricultural and
33   ecological droughts (WCE, MED, CAU, EAU, SAU, WCA, EAS, SCA, CAR, NSA, NES, SAM, SWS,
34   SSA, NCA, CAN, ENA, WNA, WSAF, ESAF and MDG). Some regions are also projected to experience
35   less agricultural and ecological drought with global warming (medium confidence; (NEAF, SAS)). There is
36   high confidence that the projected increases in agricultural and ecological droughts are strongly affected by
37   AED increases in a warming climate, although ET increases are projected to be smaller than those in AED
38   due to soil moisture limitations and CO2 effects on leaf stomatal conductance. Enhanced atmospheric CO2
39   concentrations lead to enhanced water-use efficiency in plants (medium confidence), but there is low
40   confidence that it can ameliorate agricultural and ecological droughts, or hydrological droughts, at higher
41   global warming levels characterized by limited soil moisture and enhanced AED.
42
43   Projected changes in meteorological droughts are overall less extended than for agricultural and ecological
44   droughts, but also affect several AR6 regions, even at 1.5°C and 2°C of global warming. Several regions are
45   also projected to be more strongly affected by hydrological droughts with increasing global warming (NEU,
46   WCE, EEU, MED, SAU, WCA, SCA, NSA, SAM, SWS, SSA, WNA, WSAF, ESAF, MDG). Increased
47   soil moisture limitation and associated changes in droughts are projected to lead to increased vegetation
48   stress in many regions, with implications for the global land carbon sink (CC-Box 5). There is high
49   confidence that the global land sink will become less efficient due to soil moisture limitations and associated
50   drought conditions in some regions in higher emission scenarios specially under global warming levels
51   above 4°C ; however, there is low confidence on how these water cycle feedbacks will play out in lower
52   emission scenarios (at 2°C global warming or lower) (Cross-Chapter Box5.1).
53
54
55   [START FIGURE 11.19 HERE]
     Do Not Cite, Quote or Distribute                     11-86                                      Total pages: 345
     Final Government Distribution                         Chapter11                                       IPCC AR6 WGI

 1
 2   Figure 11.19:Projected changes in (a-c) the number of consecutive dry days (CDD), (d-f) annual mean soil moisture
 3                over the total column, and (g-l) the frequency and intensity of one-in-ten year soil moisture drought for
 4                the June-to-August and December-to-February seasons at 1.5°C, 2°C, and 4°C of global warming
 5                compared to the 1851-1900 baseline. Results are based on simulations from the CMIP6 multi-model
 6                ensemble under the SSP1-1.9, SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5 scenarios. The numbers in
 7                the top right indicate the number of simulations included. Uncertainty is represented using the simple
 8                approach: no overlay indicates regions with high model agreement, where ≥80% of models agree on sign
 9                of change; diagonal lines indicate regions with low model agreement, where <80% of models agree on
10                sign of change. For more information on the simple approach, please refer to the Cross-Chapter Box
11                Atlas 1. For details on the methods see Supplementary Material 11.SM.2. Further details on data sources
12                and processing are available in the chapter data table (Table 11.SM.9).
13
14
15   [END FIGURE 11.19 HERE]
16
17
18   11.7 Extreme storms
19
20   Extreme storms, such as tropical cyclones (TCs), extratropical cyclones (ETCs), and severe convective
21   storms often have substantial societal impacts. Quantifying the effect of climate change on extreme storms is
22   challenging, partly because extreme storms are rare, short-lived, and local, and individual events are largely
23   influenced by stochastic variability. The high degree of random variability makes detection and attribution of
24   extreme storm trends more uncertain than detection and attribution of trends in other aspects of the
25   environment in which the storms evolve (e.g., larger-scale temperature trends). Projecting changes in
26   extreme storms is also challenging because of constraints in the models' ability to accurately represent the
27   small-scale physical processes that can drive these changes. Despite the challenges, progress has been made
28   since AR5.
29
30   SREX (Chapter 3) concluded that there is low confidence in observed long-term (40 years or more) trends in
31   TC intensity, frequency, and duration, and any observed trends in phenomena such as tornadoes and hail; it
32   is likely that extratropical storm tracks have shifted poleward in both the Northern and Southern
33   Hemispheres and that heavy rainfalls and mean maximum wind speeds associated with TCs will increase
34   with continued greenhouse gas (GHG) warming; it is likely that the global frequency of TCs will either
35   decrease or remain essentially unchanged, while it is more likely than not that the frequency of the most
36   intense storms will increase substantially in some ocean basins; there is low confidence in projections of
37   small-scale phenomena such as tornadoes and hail storms; and there is medium confidence that there will be
38   a reduced frequency and a poleward shift of mid-latitude cyclones due to future anthropogenic climate
39   change.
40
41   Since SREX, several IPCC assessments also assessed storms. AR5 (Chapter 2, Hartmann et al., 2013)
42   assessment with low confidence observed long-term trends in TC metrics, but revised the statement from
43   SREX to state that it is virtually certain that there are increasing trends in North Atlantic TC activity since
44   the 1970s, with medium confidence that anthropogenic aerosol forcing has contributed to these trends. AR5
45   concluded that it is likely that TC precipitation and mean intensity will increase and more likely than not that
46   the frequency of the strongest storms increases with continued GHG warming. Confidence in projected
47   trends in overall TC frequency remained low. Confidence in observed and projected trends in hail storm and
48   tornado events also remained low. SROCC (Chapter 6, Collins et al., 2019) assessed past and projected TCs
49   and ETCs supporting the conclusions of AR5 with some additional detail. Literature subsequent to AR5 adds
50   support to the likelihood of increasing trends in TC intensity, precipitation, and frequency of the most intense
51   storms, while some newer studies have added uncertainty to projected trends in overall frequency. A
52   growing body literature since AR5 on the poleward migration of TCs led to a new assessment in SROCC of
53   low confidence that the migration in the western North Pacific represents a detectable climate change
54   contribution from anthropogenic forcing. SR15 (Chapter 3, Hoegh-Guldberg et al., 2018) essentially
55   confirmed the AR5 assessment of TCs and ETCs, adding that heavy precipitation associated with TCs is
56   projected to be higher at 2°C compared to 1.5°C global warming (medium confidence).
     Do Not Cite, Quote or Distribute                        11-87                                         Total pages: 345
     Final Government Distribution                        Chapter11                                      IPCC AR6 WGI

 1
 2   SREX, AR5, SROCC, and SR15, do not provide assessments of the atmospheric rivers and SROCC and
 3   SR15 do not assess severe convective storms and extreme winds. This section assesses the state of
 4   knowledge on the four phenomena of TCs, ETCs, severe convective storms, and extreme winds.
 5   Atmospheric rivers are addressed in Chapter 8. In this respect, this assessment closely mirrors the SROCC
 6   assessment of TCs and ETCs, while updating SREX and AR5 assessments of severe convective storms and
 7   extreme winds.
 8
 9
10   11.7.1 Tropical cyclones
11
12   11.7.1.1 Mechanisms and drivers
13
14   The genesis, development, and tracks of TCs depend on conditions of the larger-scale circulations of the
15   atmosphere and ocean (Christensen et al., 2013). Large-scale atmospheric circulations (Annex VI), such as
16   the Hadley and Walker circulations and the monsoon circulations, and internal variability acting on various
17   time-scales, from intra-seasonal (e.g., the Madden-Julian and Boreal Summer Intraseasonal oscillations
18   (MJO, BSISO), and equatorial waves) and inter-annual (e.g., the El Niño-Southern Oscillation (ENSO) and
19   Pacific and Atlantic Meridional Modes (PMM, AMM)), to inter-decadal (e.g., Atlantic Multidecadal
20   Variability and Pacific Decadal Variability (PDV)) can all significantly affect TCs. This broad range of
21   natural variability makes detection of anthropogenic effects difficult, and uncertainties in the projected
22   changes of these modes of variability increase uncertainty in the projected changes in TC activity. Aerosol
23   forcing also affects SST patterns and cloud microphysics, and it is likely that observed changes in TC activity
24   are partly caused by changes in aerosol forcing (Evan et al., 2011; Ting et al., 2015; Sobel et al., 2016, 2019;
25   Takahashi et al., 2017; Zhao et al., 2018; Reed et al., 2019). Among possible changes from these drivers,
26   there is medium confidence that the Hadley cell has widened and will continue to widen in the future
27   (Chapter 2.3, 3.3, 4.5). This likely causes latitudinal shifts of TC tracks (Sharmila and Walsh, 2018).
28   Regional TC activity changes are also strongly affected by projected changes in SST warming patterns
29   (Yoshida et al., 2017), which are highly uncertain (Chapter 4, 9).
30
31
32   11.7.1.2 Observed trends
33
34   Identifying past trends in TC metrics remains a challenge due to the heterogeneous character of the historical
35   instrumental data, which are known as “best-track” data (Schreck et al., 2014). There is low confidence in
36   most reported long-term (multidecadal to centennial) trends in TC frequency- or intensity-based metrics due
37   to changes in the technology used to collect the best-track data. This should not be interpreted as implying
38   that no physical (real) trends exist, but rather as indicating that either the quality or the temporal length of the
39   data is not adequate to provide robust trend detection statements, particularly in the presence of multidecadal
40   variability.
41
42   There are previous and ongoing efforts to homogenize the best-track data (Elsner et al., 2008; Kossin et al.,
43   2013, 2020; Choy et al., 2015; Landsea, 2015; Emanuel et al., 2018) and there is substantial literature that
44   finds positive trends in intensity-related metrics in the best-track during the “satellite period”, which is
45   generally limited to the past ~40 years (Kang and Elsner, 2012; Kishtawal et al., 2012; Kossin et al., 2013,
46   2020; Mei and Xie, 2016; Zhao et al., 2018; Tauvale and Tsuboki, 2019). When best-track trends are tested
47   using homogenized data, the intensity trends generally remain positive, but are smaller in amplitude (Kossin
48   et al., 2013; Holland and Bruyère, 2014). Kossin et al. (2020) extended the homogenized TC intensity record
49   to the period 1979–2017 and identified significant global increases in major TC exceedance probability of
50   about 6% per decade. In addition to trends in TC intensity, there is evidence that TC intensification rates and
51   the frequency of rapid intensification events have increased within the satellite era (Kishtawal et al. 2012;
52   Balaguru et al., 2018; Bhatia et al., 2018). The increase in intensification rates is found in the best-track as
53   well as the homogenized intensity data.
54
55   A subset of the best-track data corresponding to hurricanes that have directly impacted the United States
     Do Not Cite, Quote or Distribute                       11-88                                       Total pages: 345
     Final Government Distribution                        Chapter11                                     IPCC AR6 WGI

 1   since 1900 is considered to be reliable, and shows no trend in the frequency of U.S. landfall events (Knutson
 2   et al., 2019). However, in this period since 1900, an increasing trend in normalized U.S. hurricane damage,
 3   which accounts for temporal changes in exposed wealth (Grinsted et al., 2019), and a decreasing trend in TC
 4   translation speed over the U.S. (Kossin, 2019) have been identified. A similarly reliable subset of the data
 5   representing TC landfall frequency over Australia shows a decreasing trend in eastern Australia since the
 6   1800s (Callaghan and Power, 2011), as well as in other parts of Australia since 1982 (Chand et al., 2019;
 7   Knutson et al., 2019), and a paleoclimate proxy reconstruction shows that recent levels of TC interactions
 8   along parts of the Australian coastline are the lowest in the past 550–1,500 years (Haig et al., 2014). Existing
 9   TC datasets show substantial interdecadal variations in basin-wide TC frequency and intensity in the western
10   North Pacific, but a statistically significant northwestward shift in the western North Pacific TC tracks since
11   the 1980s (Lee et al., 2020b). In the case of the North Indian Ocean, analyses of trends are highly dependent
12   on the details of each analysis (e.g., pre- and/or post-monsoon season period, or Bay of Bengal and/or
13   Arabian Sea region). The most consistent trends are an increase in the occurrence of the most intense TCs
14   and a decrease in the overall TC frequency, in particular in the Bay of Bengal (Sahoo and Bhaskaran, 2016;
15   Balaji et al., 2018; Singh et al., 2019; Baburaj et al., 2020). In the South Indian Ocean (SIO), an increase in
16   the occurrence of the most intense TCs has been noted, however there are well-known data quality issues
17   there (Kuleshov et al., 2010; Fitchett, 2018). When the SIO data are homogenized, a significant increase is
18   found in the fractional proportion of global category 3-5 TC estimates to all category 1-5 estimates (Kossin
19   et al., 2020).
20
21   As with all confined regional analyses of TC frequency, it is generally unclear whether any identified
22   changes are due to a basin-wide change in TC frequency, or to systematic track shifts (or both). From an
23   impacts perspective, however, these changes over land are highly relevant and emphasize that large-scale
24   modifications in TC behaviour can have a broad spectrum of impacts on a regional scale.
25
26   Subsequent to AR5, two metrics that are argued to be comparatively less sensitive to data issues than
27   frequency- and intensity-based metrics have been analysed. Trends in these metrics have been identified over
28   the past ~70 years or more (Knutson et al., 2019). The first metric, the mean latitude where TCs reach their
29   peak intensity, exhibits a global and regional poleward migration during the satellite period (Kossin et al.,
30   2014). The poleward migration can influence TC hazard exposure and risk (Kossin et al., 2016a) and is
31   consistent with the independently-observed expansion of the tropics (Lucas et al., 2014). The migration has
32   been linked to changes in the Hadley circulation (Altman et al., 2018; Sharmila and Walsh, 2018; Studholme
33   and Gulev, 2018). The migration is also apparent in the mean locations where TCs exhibit eyes (Knapp et al.,
34   2018), which is when TCs are most intense. Part of the Northern Hemisphere poleward migration is due to
35   inter-basin changes in TC frequency (Kossin et al., 2014, 2016b, Moon et al., 2015, 2016) and the trends, as
36   expected, can be sensitive to the time period chosen (Tennille and Ellis, 2017; Kossin, 2018; Song and
37   Klotzbach, 2018) and to subsetting of the data by intensity (Zhan and Wang, 2017). The poleward migration
38   is particularly pronounced and well-documented in the western North Pacific basin (Kossin et al., 2016a;
39   Oey and Chou, 2016; Liang et al., 2017; Nakamura et al., 2017; Altman et al., 2018; Daloz and Camargo,
40   2018; Sun et al., 2019b; Lee et al., 2020b; Yamaguchi and Maeda, 2020a; Kubota et al., 2021).
41
42
43   [START FIGURE 11.20 HERE]
44
45   Figure 11.20:Summary schematic of past and projected changes in tropical cyclone (TC), extratropical cyclone (ETC),
46                atmospheric river (AR), and severe convective storm (SCS) behaviour. Global changes (blue shading)
47                from top to bottom: 1) Increased mean and maximum rain-rates in TCs, ETCs, and ARs [past (low
48                confidence due to lack of reliable data) & projected (high confidence)]. 2) Increased proportion of
49                stronger TCs [past (medium confidence) & projected (high confidence)]. 3) Decrease or no change in
50                global frequency of TC genesis [past (low confidence due to lack of reliable data) & projected (medium
51                confidence)]. 4) Increased and decreased ETC wind-speed, depending on the region, as storm-tracks
52                change [past (low confidence due to lack of reliable data) & projected (medium confidence)]. Regional
53                changes, from left to right: 1) Poleward TC migration in the western North Pacific and subsequent
54                changes in TC exposure [past (medium confidence) & projected (medium confidence)]. 2) Slowdown of
55                TC forward translation speed over the contiguous US and subsequent increase in TC rainfall [past
56                (medium confidence) & projected (low confidence due to lack of directed studies)]. 3) Increase in mean
     Do Not Cite, Quote or Distribute                       11-89                                       Total pages: 345
     Final Government Distribution                        Chapter11                                       IPCC AR6 WGI

 1                and maximum SCS rain-rate and increase in springtime SCS frequency and season length over the
 2                contiguous US [past (low confidence due to lack of reliable data) & projected (medium confidence)].
 3
 4   [END FIGURE 11.20 HERE]
 5
 6
 7   A second metric that is argued to be comparatively less sensitive to data issues than frequency- and intensity-
 8   based metrics is TC translation speed (Kossin, 2018), which exhibits a global slowdown in the best-track
 9   data over the period 1949-2016. TC translation speed is a measure of the speed at which TCs move across
10   the Earth’s surface and is very closely related to local rainfall amounts (i.e., a slower translation speed causes
11   greater local rainfall). TC translation speed also affects structural wind damage and coastal storm surge by
12   changing the hazard event duration. The slowdown is observed in the best-track data from all basins except
13   the Northern Indian Ocean and is also found in a number of regions where TCs interact directly with land.
14   The slowing trends identified in the best-track data by Kossin (2018) have been argued to be largely due to
15   data heterogeneity. Moon et al. (2019) and Lanzante (2019) provide evidence that meridional TC track shifts
16   project onto the slowing trends and argue that these shifts are due to the introduction of satellite data. Kossin
17   (2019) provides evidence that the slowing trend is real by focusing on Atlantic TC track data over the
18   contiguous United States in the 118-year period 1900–2017, which are generally considered reliable. In this
19   period, mean TC translation speed has decreased by 17%. The slowing TC translation speed is expected to
20   increase local rainfall amounts, which would increase coastal and inland flooding. In combination with
21   slowing translation speed, abrupt TC track direction changes – that can be associated with track “meanders”
22   or “stalls” – have become increasingly common along the North American coast since the mid-20th century,
23   leading to more rainfall in the region (Hall and Kossin, 2019).
24
25   In summary, there is mounting evidence that a variety of TC characteristics have changed over various time
26   periods. It is likely that the proportion of major TC intensities and the frequency of rapid intensification
27   events have both increased globally over the past 40 years. It is very likely that the average location where
28   TCs reach their peak wind-intensity has migrated poleward in the western North Pacific Ocean since the
29   1940s. It is likely that TC translation speed has slowed over the U.S. since 1900.
30
31
32   11.7.1.3 Model evaluation
33
34   Accurate projections of future TC activity have two principal requirements: accurate representation of
35   changes in the relevant environmental factors (e.g., SSTs) that can affect TC activity, and accurate
36   representation of actual TC activity in given environmental conditions. In particular, models’ capacity to
37   reproduce historical trends or interannual variabilities of TC activity is relevant to the confidence in future
38   projections. One test of the models is to evaluate their ability to reproduce the dependency of the TC
39   statistics in the different basins in the real world, in addition to their capability of reproducing atmospheric
40   and ocean environmental conditions. For the evaluation of projections of TC-relevant environmental
41   variables, AR5 confidence statements were based on global surface temperature and moisture, but not on the
42   detailed regional structure of SST and atmospheric circulation changes such as steering flows and vertical
43   shear, which affect characteristics of TCs (genesis, intensity, tracks, etc.). Various aspects of TC metrics are
44   used to evaluate how capable models are of simulating present-day TC climatologies and variability (e.g. TC
45   frequency, wind-intensity, precipitation, size, tracks, and their seasonal and interannual changes) (Camargo
46   and Wing, 2016; Knutson et al., 2019, 2020; Walsh et al., 2015). Other examples of TC
47   climatology/variability metrics are spatial distributions of TC occurrence and genesis (Walsh et al., 2015),
48   seasonal cycles and interannual variability of basin-wide activity (Zhao et al., 2009; Shaevitz et al., 2014;
49   Kodama et al., 2015; Murakami et al., 2015; Yamada et al., 2017) or landfalling activity (Lok and Chan,
50   2018), as well as newly developed process-diagnostics designed specifically for TCs in climate models (Kim
51   et al., 2018a; Wing et al., 2019; Moon et al., 2020).
52
53   Confidence in the projection of intense TCs, such as those of Category 4-5, generally becomes higher as the
54   resolution of the models becomes higher. CMIP5/6 class climate models (~100-200 km grid spacing) cannot
55   simulate TCs of Category 4-5 intensity. They do simulate storms of relatively high vorticity that are at best
     Do Not Cite, Quote or Distribute                       11-90                                         Total pages: 345
     Final Government Distribution                     Chapter11                                    IPCC AR6 WGI

 1   described as “TC-like”, but metrics like storm counts are highly dependent on tracking algorithms (Wehner
 2   et al., 2015; Zarzycki and Ullrich, 2017; Roberts et al., 2020a). High-resolution global climate models (~10-
 3   60 km grid spacing) as used in HighResMIP (Haarsma et al., 2016; Roberts et al., 2020a) begin to capture
 4   some structures of TCs more realistically, as well as produce intense TCs of Category 4-5 despite the effects
 5   of parameterized deep cumulus convection processes (Murakami et al., 2015; Wehner et al., 2015; Yamada
 6   et al., 2017; Roberts et al., 2018; Moon et al., 2020). Convection-permitting models (~1-10 km grid-
 7   spacing), such as used in some dynamical downscaling studies, provide further realism with capturing eye
 8   wall structures (Tsuboki et al., 2015). Model characteristics besides resolution, especially details of
 9   convective parameterization, can influence a model’s ability to simulate intense TCs (Reed and Jablonowski,
10   2011; Zhao et al., 2012; He and Posselt, 2015; Kim et al., 2018a; Zhang and Wang, 2018; Camargo et al.,
11   2020). However, models’ dynamical cores and other physics also affect simulated TC properties (Reed et al.,
12   2015; Vidale et al., 2021). Both wide-area regional and global convection-permitting models without the
13   need for parameterized convection are becoming more useful for TC regional model projection studies
14   (Tsuboki et al., 2015; Kanada et al., 2017a; Gutmann et al., 2018) and global model projection studies (Satoh
15   et al., 2015, 2017; Yamada et al., 2017), as they capture more realistic eye-wall structures of TCs (Kinter et
16   al., 2013) and are becoming more useful for investigating changes in TC structures (Kanada et al., 2013;
17   Yamada et al., 2017). Large ensemble simulations of global climate models with 60 km grid spacing provide
18   TC statistics that allow more reliable detection of changes in the projections, which are not well captured in
19   any single experiment (Yoshida et al., 2017; Yamaguchi et al., 2020). Variable resolution global models
20   offer an alternative to regional models for individual TC or basin-wide simulations (Yanase et al., 2012;
21   Zarzycki et al., 2014; Harris et al., 2016; Reed et al., 2020; Stansfield et al., 2020). Computationally less
22   intense than equivalent uniform resolution global models, they also do not require lateral boundary
23   conditions, thus reducing this source of error (Hashimoto et al., 2016). Confidence in the projection of TC
24   statistics and properties is increased by the higher-resolution models with more realistic simulations.
25
26   Operational forecasting models also reproduce TCs and their use for climate projection studies shows
27   promise. However, there is limited application for future projections as they are specifically developed for
28   operational purposes and TC climatology is not necessarily well evaluated. Intercomparison of operational
29   models indicates that enhancement of horizontal resolution can provide more credible projections of TCs
30   (Nakano et al., 2017). Likewise, high-resolution climate models show promise as TC forecast tools (Zarzycki
31   and Jablonowski, 2015; Reed et al., 2020), further narrowing the continuum of weather and climate models
32   and increasing confidence in projections of future TC behaviour. However, higher horizontal resolution does
33   not necessarily lead to an improved TC climatology (Camargo et al., 2020).
34
35   Atmosphere-ocean interaction is an important process in TC evolution. Atmosphere-ocean coupled models
36   are generally better than atmosphere-only models at capturing realistic processes related to TCs (Murakami
37   et al., 2015; Ogata et al., 2015, 2016; Zarzycki, 2016; Kanada et al., 2017b; Scoccimarro et al., 2017),
38   although the basin-scale SST biases commonly found in atmosphere-ocean models can introduce substantial
39   errors in the simulated TC number (Hsu et al., 2019). Higher-resolution ocean models improve the
40   simulation of TCs by reducing the SST climatology bias (Li and Sriver, 2018; Roberts et al., 2020a). Coarse
41   resolution atmospheric models may degrade coupled model performance as well. For example, in a case
42   study of Hurricane Harvey, Trenberth et al. (2018) suggested that the lack of realistic hurricane activity
43   within coupled climate models hampers the models’ ability to simulate SST and ocean heat content and their
44   changes.
45
46   Even with higher-resolution atmosphere-ocean coupled models, TC projection studies still rely on
47   assumptions in experimental design that introduce uncertainties. Computational constraints often limit the
48   number of simulations, resulting in relatively small ensemble sizes and incomplete analyses of possible
49   future SST magnitude and pattern changes (Zhao and Held, 2011; Knutson et al., 2013a). Uncertainties in
50   aerosol forcing also are reflected in TC projection uncertainty (Wang et al., 2014).
51
52   Regional climate models (RCM) with grid spacing around 15-50 km can be used to study the projection of
53   TCs. RCMs are run with lateral and surface boundary conditions, which are specified by the atmospheric
54   state and SSTs simulated by GCMs. Various combinations of the lateral and surface boundary conditions can
55   be chosen for RCM studies, and uncertainties in the projection can be further examined in general. They are
     Do Not Cite, Quote or Distribute                    11-91                                     Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   used for studying changes in TC characteristics in a specific area, such as Vietnam (Redmond et al., 2015)
 2   and the Philippines (Gallo et al., 2019).
 3
 4   Less computationally-expensive downscaling approaches that allow larger ensembles and long-term studies
 5   are also used in the projection of TCs (Emanuel et al., 2006; Lee et al., 2018a). A statistical–dynamical TC
 6   downscaling method requires assumptions of the rate of seeding of random initial disturbances, which are
 7   generally assumed to not change with climate change (Emanuel et al., 2008; Emanuel, 2013). The results
 8   with the downscaling approach might depend on the assumptions which are required for the simplification of
 9   the methods.
10
11   In summary, various types of models are useful to study climate changes of TCs, and there is no unique
12   solution for choosing a model type. However, higher-resolution models generally capture TC properties
13   more realistically (high confidence). In particular, models with horizontal resolutions of 10-60 km are
14   capable of reproducing strong TCs with Category 4-5 and those of 1-10km are capable of the eyewall
15   structure of TCs. Uncertainties in TC simulations come from details of the model configuration of both
16   dynamical and physical processes. Models with realistic atmosphere-ocean interactions are generally better
17   than atmosphere-only models at reproducing realistic TC evolutions (high confidence).
18
19
20   11.7.1.4 Detection and attribution, event attribution
21
22   There is general agreement in the literature that anthropogenic greenhouse gases and aerosols have
23   measurably affected observed oceanic and atmospheric variability in TC-prone regions (see Chapter 3). This
24   underpinned the SROCC assessment of medium confidence that humans have contributed to the observed
25   increase in Atlantic hurricane activity since the 1970s (Chapter 5, Bindoff et al., 2013). Literature subsequent
26   to the AR5 lends further support to this statement (Knutson et al., 2019). However, there is still no consensus
27   on the relative magnitude of human and natural influences on past changes in Atlantic hurricane activity, and
28   particularly on which factor has dominated the observed increase (Ting et al., 2015) and it remains uncertain
29   whether past changes in Atlantic TC activity are outside the range of natural variability. A recent result using
30   high-resolution dynamical model experiments suggested that the observed spatial contrast in TC trends
31   cannot be explained only by multi-decadal natural variability, and that external forcing plays an important
32   role (Murakami et al., 2020). Observational evidence for significant global increases in the proportion of
33   major TC intensities (Kossin et al., 2020) is consistent with both theory and numerical modeling simulations,
34   which generally indicate an increase in mean TC peak intensity and the proportion of very intense TCs in a
35   warming world (Knutson et al., 2015, 2020, Walsh et al., 2015, 2016). In addition, high-resolution coupled
36   model simulations provide support that natural variability alone is unlikely to explain the magnitude of the
37   observed increase in TC intensification rates and upward TC intensity trend in the Atlantic basin since the
38   early 1980s (Bhatia et al., 2019; Murakami et al. 2020).
39
40   The cause of the observed slowdown in TC translation speed is not yet clear. Yamaguchi et al. (2020) used
41   large ensemble simulations to argue that part of the slowdown is due to actual latitudinal shifts of TC tracks,
42   rather than data artefacts, in addition to atmospheric circulation changes, while Zhang et al. (2020a) used
43   large ensemble simulations to show that anthropogenic forcing can lead to a robust slowdown, particularly
44   outside of the tropics at higher latitudes. Yamaguchi and Maeda (2020b) found a significant slowdown in the
45   western North Pacific over the past 40 years and attributed the slowdown to a combination of natural
46   variability and global warming. The slowing trend since 1900 over the U.S. is robust and significant after
47   removing multidecadal variability from the time series (Kossin, 2019). Among the hypotheses discussed is
48   the physical linkage between warming and slowing circulation (Held and Soden 2006, see also Section
49   8.2.2.2), with expectations of Arctic amplification and weakening circulation patterns through weakening
50   meridional temperature gradients (Coumou et al., 2018; see also Cross-Chapter Box 10.1), or through
51   changes in planetary wave dynamics (Mann et al., 2017). The tropics expansion and the poleward shift of the
52   mid-latitude westerlies associated with warming is also suggested for the reason of the slowdown (Zhang et
53   al., 2020a). However, the connection of these mechanisms to the slowdown has not been robustly shown yet.
54   Furthermore, slowing trends have not been unambiguously observed in circulation patterns that steer TCs
55   such as the Walker and Hadley circulations (Section 2.3.1.4), although these circulations generally slow
     Do Not Cite, Quote or Distribute                     11-92                                      Total pages: 345
     Final Government Distribution                     Chapter11                                    IPCC AR6 WGI

 1   down in numerical simulations under global warming (Sections 4.5.1.6 and 8.4.2.2).
 2
 3   The observed poleward trend in western North Pacific TCs remains significant after accounting for the
 4   known modes of dominant interannual to decadal variability in the region (Kossin et al., 2016a), and is also
 5   found in CMIP5 model-simulated TCs (in the recent historical period 1980–2005), although it is weaker than
 6   observed and is not statistically significant (Kossin et al., 2016a). However, the trend is significant in 21st
 7   century CMIP5 projections under the RCP8.5 scenario, with a similar spatial pattern and magnitude to the
 8   past observed changes in that basin over the period 1945–2016, supporting a possible anthropogenic GHG
 9   contribution to the observed trends (Knutson et al., 2019; Kossin et al., 2016a).
10
11   The recent active TC seasons in some basins have been studied to determine whether there is anthropogenic
12   influence. For 2015, Murakami et al. (2017) explored the unusually high TC frequency near Hawaii and in
13   the eastern Pacific basin. Zhang et al. (2016) considered unusually high Accumulated Cyclone Energy
14   (ACE) in the western North Pacific. Yang et al. (2018) and Yamada et al. (2019) looked at TC intensification
15   in the western North Pacific. These studies suggest that the anomalous TC activity in 2015 was not solely
16   explained by the effect of an extreme El Niño (see BOX 11.3), that there was also an anthropogenic
17   contribution, mainly through the effects of SSTs in subtropical regions. In the post-monsoon seasons of 2014
18   and 2015, tropical storms with lifetime maximum winds greater than 46 m s−1 were first observed over the
19   Arabian Sea, and Murakami et al. (2017b) showed that the probability of late-season severe tropical storms is
20   increased by anthropogenic forcing compared to the preindustrial era. Murakami et al. (2018) concluded that
21   the active 2017 Atlantic hurricane season was mainly caused by pronounced SSTs in the tropical North
22   Atlantic and that these types of seasonal events will intensify with projected anthropogenic forcing. The
23   trans-basin SST change, which might be driven by anthropogenic aerosol forcing, also affects TC activity.
24   Takahashi et al. (2017) suggested that a decrease in sulfate aerosol emissions caused about half of the
25   observed decreasing trends in TC genesis frequency in the south-eastern region of the western North Pacific
26   during 1992–2011.
27
28   Event attribution is used in case studies of TCs to test whether the severities of recent intense TCs are
29   explained without anthropogenic effects. In a case study of Hurricane Sandy (2012), Lackmann (2015) found
30   no statistically significant impact of anthropogenic climate change on storm intensity, while projections in a
31   warmer world showed significant strengthening. On the other hand, Magnusson et al. (2014) found that in
32   ECMWF simulations, the simulated cyclone depth and intensity, as well as precipitation, were larger when
33   the model was driven by the warmer actual SSTs than the climatological average SSTs. In super typhoon
34   Haiyan, which struck the Philippines on 8 November 2013, Takayabu et al. (2015) took an event attribution
35   approach with cloud system-resolving (~1km) downscaling ensemble experiments to evaluate the
36   anthropogenic effect on typhoons, and showed that the intensity of the simulated worst case storm in the
37   actual conditions was stronger than that in a hypothetical condition without historical anthropogenic forcing
38   in the model. However, in a similar approach with two coarser parameterized convection models, Wehner et
39   al. (2018) found conflicting human influences on Haiyan’s intensity. Patricola and Wehner (2018) found
40   little evidence of an attributable change in intensity of hurricanes Katrina (2005), Irma (2017), and Maria
41   (2017) using a regional climate model configured between 3 and 4.5 km resolution. They did, however, find
42   attributable increases in heavy precipitation totals. These results imply that higher resolution, such as in a
43   convective permitting 5-km or less mesh model, is required to obtain a robust anthropogenic intensification
44   of a strong TC by simulating realistic rapid intensification of a TC (Kanada and Wada, 2016; Kanada et al.,
45   2017a), and that whether the intensification of TCs can be attributed to the recent warming depends on the
46   case.
47
48   The dominant factor in the extreme rainfall amounts during Hurricane Harvey’s passage onto the U.S. in
49   2017 was its slow translation speed. But studies published after the event have argued that anthropogenic
50   climate change contributed to an increase in rain rate, which compounded the extreme local rainfall caused
51   by the slow translation. Emanuel (2017) used a large set of synthetically-generated storms and concluded
52   that the occurrence of extreme rainfall as observed in Harvey was substantially enhanced by anthropogenic
53   changes to the larger-scale ocean and atmosphere characteristics. Trenberth et al. (2018) linked Harvey’s
54   rainfall totals to the anomalously large ocean heat content from the Gulf of Mexico. van Oldenborgh et al.
55   (2017) and Risser and Wehner (2017) applied extreme value analysis to extreme rainfall records in the
     Do Not Cite, Quote or Distribute                     11-93                                     Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   Houston, Texas region and both attributed large increases to climate change. Large precipitation increases
 2   during Harvey due to global warming were also found using climate models (van Oldenborgh et al., 2017;
 3   Wang et al., 2018b). Harvey precipitation totals were estimated in these papers to be 3 to 10 times more
 4   probable due to climate change. A best estimate from a regional climate and flood model is that urbanization
 5   increased the risk of the Harvey flooding by a factor of 21 (Zhang et al., 2018c). Anthropogenic effects on
 6   precipitation increases were also predicted in advance from a forecast model for Hurricane Florence in 2018
 7   (Reed et al., 2020).
 8
 9   In summary, it is very likely that the recent active TC seasons in the North Atlantic, the North Pacific, and
10   Arabian basins cannot be explained without an anthropogenic influence. The anthropogenic influence on
11   these changes is principally associated to aerosol forcing, with stronger contributions to the response in the
12   North Atlantic. It is more likely than not that the slowdown of TC translation speed over the USA has
13   contributions from anthropogenic forcing. It is likely that the poleward migration of TCs in the western
14   North Pacific and the global increase in TC intensity rates cannot be explained entirely by natural variability.
15   Event attribution studies of specific strong TCs provide limited evidence for anthropogenic effects on TC
16   intensifications so far, but high confidence for increases in TC heavy precipitation. There is high confidence
17   that anthropogenic climate change contributed to extreme rainfall amounts during Hurricane Harvey (2017)
18   and other intense TCs.
19
20
21   11.7.1.5    Projections
22
23   A summary of studies on TC projections for the late 21st century, particularly studies since AR5, is given by
24   Knutson et al. (2020), which is an assessment report mandated by the World Meteorological Organization
25   (WMO). Studies subsequent to Knutson et al. (2020) are generally consistent and the confidence assessments
26   here closely follow theirs (Cha et al., 2020), although there are some differences due to the different
27   confidence calibrations between the IPCC and WMO reports.
28
29   There is not an established theory for the drivers of future changes in the frequency of TCs. Most, but not all,
30   high-resolution global simulations project significant reductions in the total number of TCs, with the bulk of
31   the reduction at the weaker end of the intensity spectrum as the climate warms (Knutson et al., 2020). Recent
32   exceptions based on high-resolution coupled model results are noted in Bhatia et al. (2018) and Vecchi et al.
33   (2019). Vecchi et al. (2019) showed that the representation of synoptic-scale seeds for TC genesis in their
34   high-resolution model causes different projections of global TC frequency, and there is evidence for a
35   decrease in seeds in some projected TC simulations (Sugi et al., 2020). However, other research indicates
36   that TC seeds are not an independent control on climatological TC frequency, rather the seeds covary with
37   the large-scale controls on TCs (Patricola et al., 2018). While empirical genesis indices derived from
38   observations and reanalysis describe well the observed subseasonal and interannual variability of current TC
39   frequency (Camargo et al., 2007, 2009; Tippett et al., 2011; Menkes et al., 2012), they fail to predict the
40   decreased TC frequency found in most high-resolution model simulations (Zhang et al., 2010; Camargo,
41   2013; Wehner et al., 2015), as they generally project an increase as the climate warms. This suggests a
42   limitation of the use of the empirical genesis indices for projections of TC genesis, in particular due to their
43   sensitivity to the humidity variable considered in the genesis index for these projections (Camargo et al.,
44   2014). In a different approach, a statistical-dynamical downscaling framework assuming a constant seeding
45   rate with warming (Emanuel, 2013, 2021) exhibits increases in TC frequency consistent with genesis indices
46   based projections, while downscaling with a different model leads to two different scenarios depending on
47   the humidity variable considered (Lee et al., 2020a). This disparity in the sign of the projected change in
48   global TC frequency and the difficulty in explaining the mechanisms behind the different signed responses
49   further emphasizes the lack of process understanding of future changes in tropical cyclogenesis (Walsh et al.,
50   2015; Hoogewind et al., 2020). Even within a single model, uncertainty in the pattern of future SST changes
51   leads to large uncertainties (including the sign) in the projected change in TC frequency in individual ocean
52   basins, although global TCs would appear to be less sensitive (Yoshida et al., 2017; Bacmeister et al., 2018).
53
54   Changes in SST and atmospheric temperature and moisture play a role in tropical cyclogenesis (Walsh et al.,
55   2015). Reductions in vertical convective mass flux due to increased tropical stability have been associated
     Do Not Cite, Quote or Distribute                     11-94                                      Total pages: 345
     Final Government Distribution                     Chapter11                                    IPCC AR6 WGI

 1   with a reduction in cyclogenesis (Held and Zhao, 2011; Sugi et al., 2012). Satoh et al. (2015) further posits
 2   that the robust simulated increase in the number of intense TCs, and hence increased vertical mass flux
 3   associated with intense TCs, must lead to a decrease in overall TC frequency because of this association. The
 4   Genesis Potential Index can be modified to mimic the TC frequency decreases of a model by altering the
 5   treatment of humidity (Camargo et al., 2014), supporting the idea that increased mid-tropospheric saturation
 6   deficit (Emanuel et al., 2008) controls TC frequency, but the approach remains empirical. Other possible
 7   controlling factors, such as a decline in the number of seeds (held constant in Emanuel’s downscaling
 8   approach, or dependent on the genesis index formulation in the approach proposed by Lee et al., 2020)
 9   caused by increased atmospheric stability have been proposed, but questioned as an important factor
10   (Patricola et al., 2018). The resolution of atmospheric models affects the number of seeds, hence TC genesis
11   frequency (Vecchi et al., 2019; Sugi et al., 2020; Yamada et al., 2021). The diverse and sometimes
12   inconsistent projected changes in global TC frequency by high-resolution models indicate that better process
13   understanding and improvement of the models are needed to raise confidence in these changes.
14
15   Most TC-permitting model simulations (10-60 km or finer grid spacing) are consistent in their projection of
16   increases in the proportion of intense TCs (Category 4-5), as well as an increase in the intensity of the
17   strongest TCs defined by maximum wind speed or central pressure fall (Murakami et al., 2012; Tsuboki et
18   al., 2015; Wehner et al., 2018a; Knutson et al., 2020). The general reduction in the total number of TCs,
19   which is concentrated in storms weaker than or equal to Category 1, contributes to this increase. The models
20   are somewhat less consistent in projecting an increase in the frequency of Category 4-5 TCs (Wehner et al.,
21   2018a). The projected increase in the intensity of the strongest TCs is consistent with theoretical
22   understanding (e.g., Emanuel, 1987) and observations (e.g., Kossin et al., 2020). For a 2°C global warming,
23   the median proportion of Category 4–5 TCs increases by 13%, while the median global TC frequency
24   decreases by 14%, which implies that the median of the global Category 4–5 TC frequency is slightly
25   reduced by 1% or almost unchanged (Knutson et al., 2020). Murakami et al. (2020) projected a decrease in
26   TC frequency over the coming century in the North Atlantic due to greenhouse warming, as consistent with
27   Dunstone et al. (2013), and a reduction in TC frequency almost everywhere in the tropics in response to +1%
28   CO2 forcing; exceptions include the central North Pacific (Hawaii region), east of the Philippines in the
29   North Pacific, and two relatively small regions in the northern Arabian Sea and Bay of Bengal. These
30   projections can vary substantially between ocean basins, possibly due to differences in regional SST
31   warming and warming patterns (Sugi et al., 2017; Yoshida et al., 2017; Bacmeister et al., 2018). A summary
32   of projections of TC characteristics is schematically shown by Figure 11.20.
33
34   The increase in global TC maximum surface wind speeds is about 5% for a 2°C global warming across a
35   number of high-resolution multi-decadal studies (Knutson et al., 2020). This indicates the deepening in
36   global TC minimum surface pressure under the global warming conditions. A regional cloud-permitting
37   model study shows that the strongest TC in the western North Pacific can be as strong as 857 hPa in
38   minimum surface pressure with a wind speed of 88 m s-1 under warming conditions in 2074-2087 (Tsuboki
39   et al., 2015). TCs are also measured by quantities such as Accumulated Cyclone Energy (ACE) and the
40   power dissipation index (PDI), which conflate TC intensity, frequency, and duration (Murakami et al., 2014).
41   Several TC modeling studies (Yamada et al., 2010; Kim et al., 2014a; Knutson et al., 2015) project little
42   change or decreases in the globally-accumulated value of PDI or ACE, which is due to the decrease in the
43   total number of TCs.
44
45   A projected increase in global average TC rainfall rates of about 12% for a 2°C global warming is consistent
46   with the Clausius-Clapeyron scaling of saturation specific humidity (Knutson et al., 2020). Increases
47   substantially greater than Clausius-Clapeyron scaling are projected in some regions, which is caused by
48   increased low-level moisture convergence due to projected TC intensity increases in those regions (Knutson
49   et al., 2015; Phibbs and Toumi, 2016; Patricola and Wehner, 2018; Liu et al., 2019c). Projections of TC
50   precipitation using large-ensemble experiments (Kitoh and Endo, 2019) show that the annual maximum 1-
51   day precipitation total is projected to increase, except for the western North Pacific where there is only a
52   small change or even a reduction is projected, mainly due to a projected decrease of TC frequency. They also
53   show that the 10-year return value of extreme Rx1day associated with TCs will greatly increase in a region
54   extending from Hawaii to the south of Japan. TC tracks and the location of topography relative to TCs
55   significantly affect precipitation, thus in general, areas on the eastern and southern faces of mountains have
     Do Not Cite, Quote or Distribute                    11-95                                     Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   more impacts of TC precipitation changes (Hatsuzuka et al., 2020). Projection studies using variable-
 2   resolution models in the North Atlantic (Stansfield et al., 2020) indicate that TC-related precipitation rates
 3   within North Atlantic TCs and the amount of hourly precipitation due to TC are projected to increase by the
 4   end of the century compared to a historical simulation. However, the annual average TC-related Rx5day over
 5   the eastern United States is projected to decrease because of a decrease in landfalling TCs. RCM studies with
 6   around 25-50 km grid spacing are used to study projected changes in TCs. The projected changes of TCs in
 7   Southeast Asia simulated by RCMs are consistent with those of most global climate models, showing a
 8   decrease in TC frequency and an increase in the amount of TC-associated precipitation or an increase in the
 9   frequency of intense TCs (Redmond et al., 2015; Gallo et al., 2019).
10
11   Projected changes in TC tracks or TC areas of occurrence in the late 21st century vary considerably among
12   available studies, although there is better agreement in the western North Pacific. Several studies project
13   either poleward or eastward expansion of TC occurrence over the western North Pacific region, and more TC
14   occurrence in the central North Pacific (Yamada et al., 2017; Yoshida et al., 2017; Wehner et al., 2018a;
15   Roberts et al., 2020b). The observed poleward expansion of the latitude of maximum TC intensity in the
16   western North Pacific is consistently reproduced by the CMIP5 models and downscaled models and these
17   models show further poleward expansion in the future; the projected mean migration rate of the mean
18   latitude where TCs reach their lifetime-maximum intensity is 0.2±0.1° from CMIP5 model results, while it is
19   0.13±0.04° from downscaled models in the western North Pacific (Kossin et al., 2014, 2016a). In the North
20   Atlantic, while the location of TC maximum intensity does not show clear poleward migration
21   observationally (Kossin et al., 2014), it tends to migrate poleward in projections (Garner et al., 2017). The
22   poleward migration is less robust among models and observations in the Indian Ocean, eastern North Pacific,
23   and South Pacific (e.g., Tauvale and Tsuboki, 2019; Ramsay et al. 2018; Cattiaux et al. 2020). There is
24   presently no clear consensus in projected changes in TC translation speed (Knutson et al., 2020), although
25   recent studies suggest a slowdown outside of the tropics (Kossin, 2019; Yamaguchi et al., 2020; Zhang et al.,
26   2020a), but regionally there can even be an acceleration of the storms (Hassanzadeh et al., 2020).
27
28   The spatial extent, or “size”, of the TC wind-field is an important determinant of storm surge and damage.
29   No detectable anthropogenic influences on TC size have been identified to date, because TCs in observations
30   vary in size substantially (Chan and Chan, 2015) and there is no definite theory on what controls TC size,
31   although this is an area of active research (Chavas and Emanuel, 2014; Chan and Chan, 2018). However,
32   projections by high-resolution models indicate future broadening of TC wind-fields when compared to TCs
33   of the same categories (Yamada et al., 2017), while Knutson et al. (2015) simulates a reasonable interbasin
34   distribution of TC size climatology, but projects no statistically significant change in global average TC size.
35   A plausible mechanism is that as the tropopause height becomes higher with global warming, the eye wall
36   areas become wider because the eye walls are inclined outward with height to the tropopause. This effect is
37   only reproduced in high-resolution convection-permitting models capturing eye walls, and such modeling
38   studies are not common. Moreover, the projected TC size changes are generally on the order of 10% or less,
39   and these size changes are still highly variable between basins and studies. Thus, the projected change in
40   both magnitude and sign of TC size is uncertain.
41
42   The coastal effects of TCs depend on TC intensity, size, track, and translation speed. Projected increases in
43   sea level, average TC intensity, and TC rainfall rates each generally act to further elevate future storm surge
44   and fresh-water flooding (see Section 9.6.4.2). Changes in TC frequency could contribute toward increasing
45   or decreasing future storm surge risk, depending on the net effects of changes in weaker vs stronger storms.
46   Several studies (McInnes et al., 2014, 2016; Little et al., 2015; Garner et al., 2017; Timmermans et al., 2017,
47   2018) have explored future projections of storm surge in the context of anthropogenic climate change with
48   the influence of both sea level rise and future TC changes. Garner et al. (2017) investigated the near future
49   changes in the New York City coastal flood hazard, and suggested a small change in storm-surge height
50   because effects of TC intensification are compensated by the offshore shifts in TC tracks, but concluded that
51   the overall effect due to the rising sea levels would increase the flood hazard. Future projection studies of
52   storm surge in East Asia, including China, Japan and Korea, also indicate that storm surge due to TCs
53   become more severe (Yang et al., 2018b; Mori et al., 2019, 2021; Chen et al., 2020c). For the Pacific islands,
54   McInnes et al. (2014) found that the future projected increase in storm surge in Fiji is dominated by sea level
55   rise, and projected TC changes make only a minor contribution. Among various storm surge factors, there is
     Do Not Cite, Quote or Distribute                     11-96                                      Total pages: 345
     Final Government Distribution                       Chapter11                                    IPCC AR6 WGI

 1   high confidence that sea level rise will lead to a higher possibility of extreme coastal water levels in most
 2   regions, with all other factors assumed equal.
 3
 4   In the North Atlantic, vertical wind shear, which inhibits TC genesis and intensification, varies in a quasi-
 5   dipole pattern with one center of action in the tropics and another along the U.S. southeast coast (Vimont and
 6   Kossin, 2007). This pattern of variability creates a protective barrier of high shear along the U.S. coast
 7   during periods of heightened TC activity in the tropics (Kossin, 2017), and appears to be a natural part of the
 8   Atlantic ocean-atmosphere climate system (Ting et al., 2019). Greenhouse gas forcing in CMIP5 and the
 9   Community Earth System Model Large Ensemble (CESM-LE; Kay et al., 2015) simulations, however,
10   erodes the pattern and degrades the natural shear barrier along the U.S. coast. Following the Representative
11   Concentration Pathway 8.5 (RCP8.5) emission scenario, the magnitude of the erosion of the barrier equals
12   the amplitude of past natural variability (time of emergence) by the mid-twenty-first century (Ting et al.,
13   2019). The projected reduction of shear along the U.S. East Coast with warming is consistent among studies
14   (e.g., Vecchi and Soden, 2007).
15
16   In summary, average peak TC wind speeds and the proportion of Category 4-5 TCs will very likely increase
17   globally with warming. It is likely that the frequency of Category 4-5 TCs will increase in limited regions
18   over the western North Pacific. It is very likely that average TC rain-rates will increase with warming, and
19   likely that the peak rain-rates will increase at greater than the Clausius-Clapeyron scaling rate of 7% per °C
20   of warming in some regions due to increased low-level moisture convergence caused by regional increases in
21   TC wind-intensity. It is likely that the average location where TCs reach their peak wind-intensity will
22   migrate poleward in the western North Pacific Ocean as the tropics expand with warming, and that the global
23   frequency of TCs over all categories will decrease or remain unchanged.
24
25
26   11.7.2 Extratropical storms
27
28   This section focuses on extratropical cyclones (ETCs) that are either classified as strong or extreme by using
29   some measure of their intensity, or by being associated with the occurrence of extremes in variables such as
30   precipitation or near-surface wind speed (Seneviratne et al., 2012). Since AR5, the high relevance of ETCs
31   for extreme precipitation events has been well established (Pfahl and Wernli, 2012; Catto and Pfahl, 2013;
32   Utsumi et al., 2017), with 80% or more of hourly and daily precipitation extremes being associated with
33   either ETCs or fronts over oceanic mid-latitude regions, and somewhat smaller, but still very large,
34   proportions of events over mid-latitude land regions (Utsumi et al., 2017). The emphasis in this section is on
35   individual ETCs that have been identified using some detection and tracking algorithms. Mid-latitude
36   atmospheric rivers are assessed in Section 8.3.2.8.
37
38
39   11.7.2.1 Observed trends
40
41   Chapter 2 (Section 2.3.1.4.3) concluded that there is overall low confidence in recent changes in the total
42   number of ETCs over both hemispheres and that there is medium confidence in a poleward shift of the storm
43   tracks over both hemispheres since the 1980s. Overall, there is also low confidence in past-century trends in
44   the number and intensity of the strongest ETCs due to the large interannual and decadal variability (Feser et
45   al., 2015; Reboita et al., 2015; Wang et al., 2016; Varino et al., 2018) and due to temporal and spatial
46   heterogeneities in the number and type of assimilated data in reanalyses, particularly before the satellite era
47   (Krueger et al., 2013; Tilinina et al., 2013; Befort et al., 2016; Chang and Yau, 2016; Wang et al., 2016).
48   There is medium confidence that the agreement among reanalyses and among detection and tracking
49   algorithms is higher when considering stronger cyclones (Neu et al., 2013; Pepler et al., 2015; Wang et al.,
50   2016). Over the Southern Hemisphere, there is high confidence that the total number of ETCs with low
51   central pressures (<980 hPa) has increased between 1979 and 2009, with all eight reanalyses considered by
52   Wang et al. (2016), showing positive trends and five of them showing statistically significant trends. Similar
53   results were found by (Reboita et al., 2015) using a different detection and tracking algorithm and a single
54   reanalysis product. Over the Northern Hemisphere, there is high agreement among reanalyses that the
55   number of cyclones with low central pressures (<970 hPa) has decreased in summer and winter during the
     Do Not Cite, Quote or Distribute                      11-97                                      Total pages: 345
     Final Government Distribution                       Chapter11                                      IPCC AR6 WGI

 1   period 1979-2010 (Tilinina et al., 2013; Chang et al., 2016). However, changes exhibit substantial decadal
 2   variability and do not show monotonic trends since the 1980s. For example, over the Arctic and North
 3   Atlantic, Tilinina et al. (2013) showed the number of cyclones with very low central pressure (<960 hPa)
 4   increased from 1979 to 1990 and then declined until 2010 in all five reanalyses considered. Over the North
 5   Pacific, the number of cyclones with very low central pressure reached a peak around 2000 and then
 6   decreased until 2010 in the five reanalyses considered (Tilinina et al., 2013).
 7
 8   Overall, however, it should be noted that characterising trends in the dynamical intensity of ETCs (e.g., wind
 9   speeds) using the absolute central pressure is problematic because the central pressure depends on the
10   background mean sea level pressure, which varies seasonally and regionally (e.g., Befort et al., 2016).
11
12
13   11.7.2.2 Model evaluation
14
15   There is high confidence that coarse-resolution climate models (e.g., CMIP5 and CMIP6) underestimate the
16   dynamical intensity of ETCs, including the strongest ETCs, as measured using a variety of metrics, including
17   mean pressure gradient, mean vorticity and near-surface winds, over most regions (Colle et al., 2013; Zappa
18   et al., 2013a; Govekar et al., 2014; Di Luca et al., 2016; Trzeciak et al., 2016; Seiler et al., 2018; Priestley et
19   al., 2020). There is also high confidence that most current climate models underestimate the number of
20   explosive systems (i.e., systems showing a decrease in mean sea level pressure of at least 24 hPa in 24 hours)
21   over both hemispheres (Seiler and Zwiers, 2016a; Gao et al., 2020; Priestley et al., 2020). There is high
22   confidence that the underestimation of the intensity of ETCs is associated with the coarse horizontal
23   resolution of climate models, with higher horizontal resolution models, including models from HighResMIP
24   and CORDEX, usually showing better performance (Colle et al., 2013; Zappa et al., 2013a; Di Luca et al.,
25   2016; Trzeciak et al., 2016; Seiler et al., 2018; Gao et al., 2020; Priestley et al., 2020). The improvement by
26   higher-resolution models is found even when comparing models and reanalyses after postprocessing data to a
27   common resolution (Zappa et al., 2013a; Di Luca et al., 2016; Priestley et al., 2020). The systematic bias in
28   the intensity of ETCs has also been linked to the inability of current climate models to well resolve diabatic
29   processes, particularly those related to the release of latent heat (Willison et al., 2013; Trzeciak et al., 2016)
30   and the formation of clouds (Govekar et al., 2014). There is medium confidence that climate models simulate
31   well the spatial distribution of precipitation associated with ETCs over the Northern Hemisphere, together
32   with some of the main features of the ETC life cycle, including the maximum in precipitation occurring just
33   before the peak in dynamical intensity (e.g., vorticity) as observed in a reanalysis and observations
34   (Hawcroft et al., 2018). There is, however, large observational uncertainty in ETC-associated precipitation
35   (Hawcroft et al., 2018) and limitations in the simulation of frontal precipitation, including too low rainfall
36   intensity over mid-latitude oceanic areas in both hemispheres (Catto et al., 2015).
37
38
39   11.7.2.3 Detection and attribution, event attribution
40
41   Chapter 3 (Section 3.3.3.3) concluded that there is low confidence in the attribution of observed changes in
42   the number of ETCs in the Northern Hemisphere and that there is high confidence that the poleward shift of
43   storm tracks in the Southern Hemisphere is linked to human activity, mostly due to emissions of ozone-
44   depleting substances. Specific studies attributing changes in the most extreme ETCs are not available. The
45   human influence on individual extreme ETC events has been considered only a few times and there is overall
46   low confidence in the attribution of these changes (NASEM, 2016; Vautard et al., 2019).
47
48
49   11.7.2.4 Projections
50
51   The frequency of ETCs is expected to change primarily following a poleward shift of the storm tracks as
52   discussed in Chapters 4 (Section 4.5.1.6, see also Figure 4.31) and 8 (Section 8.4.2.8). There is medium
53   confidence that changes in the dynamical intensity (e.g., wind speeds) of ETCs will be small, although
54   changes in the location of storm tracks can lead to substantial changes in local extreme wind speeds (Zappa
55   et al., 2013b; Chang, 2014; Li et al., 2014; Seiler and Zwiers, 2016b; Yettella and Kay, 2017; Barcikowska
     Do Not Cite, Quote or Distribute                      11-98                                       Total pages: 345
     Final Government Distribution                     Chapter11                                    IPCC AR6 WGI

 1   et al., 2018; Kar-Man Chang, 2018). Yettella and Kay (2017) detected and tracked ETCs over both
 2   hemispheres in an ensemble of 30 CESM-LE simulations, differing only in their initial conditions, and found
 3   that changes in mean wind speeds around ETC centres are often negligible between present (1986-2005) and
 4   future (2081-2100) periods. Using 19 CMIP5 models, Zappa et al. (2013b) found an overall reduction in the
 5   number of cyclones associated with low-troposphere (850-hPa) wind speeds larger than 25 m s-1 over the
 6   North Atlantic and Europe with the number of the 10% strongest cyclones decreasing by about 8% and 6%
 7   in DJF and JJA according to the RCP4.5 scenario (2070-2099 vs. 1976-2005). Over the North Pacific, Chang
 8   (2014) showed that CMIP5 models project a decrease in the frequency of ETCs with the largest central
 9   pressure perturbation (i.e., the depth, strongly related with low-level wind speeds) by the end of the century
10   according to simulations using the RCP8.5 scenario. Using projections from CMIP5 GCMs under the
11   RCP8.5 scenario (1981-2000 to 2081-2100), Seiler and Zwiers (2016b) projected a northward shift in the
12   number of explosive ETCs in the northern Pacific, with fewer and weaker events south, and more frequent
13   and stronger events north of 45°N. Using 19 CMIP5 GCMs under the RCP8.5 scenario, (Kar-Man Chang,
14   2018) found a significant decrease in the number of ETCs associated with extreme wind speeds (2081–2100
15   vs. 1980–99) over the Northern Hemisphere (average decrease of 17%) and over some smaller regions,
16   including the Pacific and Atlantic regions.
17
18   Over the Southern Hemisphere, future changes (RCP8.5 scenario; 1980-1999 to 2081-2100) in extreme
19   ETCs were studied by Chang (2017) using 26 CMIP5 models and a variety of intensity metrics (850-hPa
20   vorticity, 850-hPa wind speed, mean sea level pressure and near-surface wind speed). They found that the
21   number of extreme cyclones is projected to increase by at least 20% and as much as 50%, depending on the
22   specific metric used to define extreme ETCs. Increases in the number of strong cyclones appear to be robust
23   across models and for most seasons, although they show strong regional variations with increases occurring
24   mostly over the southern flank of the storm track, consistent with a shift and intensification of the storm
25   track. Overall, there is medium confidence that projected changes in the dynamical intensity of ETCs depend
26   on the resolution and formulation (e.g., explicit or implicit representation of convection) of climate models
27   (Booth et al., 2013; Michaelis et al., 2017; Zhang and Colle, 2017).
28
29   As reported in AR5 and in Chapter 8 (Section 8.4.2.8), despite small changes in the dynamical intensity of
30   ETCs, there is high confidence that the precipitation associated with ETCs will increase in the future (Zappa
31   et al., 2013b; Marciano et al., 2015; Pepler et al., 2016; Zhang and Colle, 2017; Michaelis et al., 2017;
32   Yettella and Kay, 2017; Barcikowska et al., 2018; Zarzycki, 2018; Hawcroft et al., 2018; Kodama et al.,
33   2019; Bevacqua et al., 2020c; Reboita et al., 2020). There is high confidence that increases in precipitation
34   will follow increases in low-level water vapour (i.e., about 7% per degree of surface warming; Box 11.1) and
35   will be largest for higher warming levels (Zhang and Colle, 2017). There is medium confidence that
36   precipitation changes will show regional and seasonal differences due to distinct changes in atmospheric
37   humidity and dynamical conditions (Zappa et al., 2015; Hawcroft et al., 2018), with even decreases in some
38   specific regions such as the Mediterranean (Zappa et al., 2015; Barcikowska et al., 2018). There is high
39   confidence that snowfall associated with wintertime ETCs will decrease in the future, because increases in
40   tropospheric temperatures lead to a lower proportion of precipitation falling as snow (O’Gorman, 2014;
41   Rhoades et al., 2018; Zarzycki, 2018). However, there is medium confidence that extreme snowfall events
42   associated with wintertime ETCs will change little in regions where snowfall will be supported in the future
43   (O’Gorman, 2014; Zarzycki, 2018).
44
45   In summary, there is low confidence in past changes in the dynamical intensity (e.g., maximum wind speeds)
46   of ETCs and medium confidence that in the future these changes will be small, although changes in the
47   location of storm tracks could lead to substantial changes in local extreme wind speeds. There is high
48   confidence that average and maximum ETC precipitation-rates will increase with warming, with the
49   magnitude of the increases associated with increases in atmospheric water vapour. There is medium
50   confidence that projected changes in the intensity of ETCs, including wind speeds and precipitation, depend
51   on the resolution and formulation of climate models.
52
53
54   11.7.3 Severe convective storms
55
     Do Not Cite, Quote or Distribute                    11-99                                     Total pages: 345
     Final Government Distribution                     Chapter11                                    IPCC AR6 WGI

 1   Severe convective storms are convective systems that are associated with extreme phenomena such as
 2   tornadoes, hail, heavy precipitation (rain or snow), strong winds, and lightning. The assessment of changes in
 3   severe convective storms in SREX (Chapter 3, Seneviratne et al., 2012) and AR5 (Chapter 12, Collins et al.,
 4   2013) is limited and focused mainly on tornadoes and hail storms. SREX assessed that there is low
 5   confidence in observed trends in tornadoes and hail because of data inhomogeneities and inadequacies in
 6   monitoring systems. Subsequent literature assessed in the Climate Science Special Report (Kossin et al.,
 7   2017) led to the assessment of the observed tornado activity over the 2000s in the United States with a
 8   decrease in the number of days per year with tornadoes and an increase in the number of tornadoes on these
 9   days (medium confidence). However, there is low confidence in past trends for hail and severe thunderstorm
10   winds. Climate models consistently project environmental changes that would support an increase in the
11   frequency and intensity of severe thunderstorms that combine tornadoes, hail, and winds (high confidence),
12   but there is low confidence in the details of the projected increase. Regional aspects of severe convective
13   storms and details of the assessment of tornadoes and hail are also assessed in Chapter 12 (Section 12.3.3.2
14   for tornadoes; Section 12.3.4.5 for hail; Section 12.4.5.3 for Europe, Section 12.4.6.3 for North America, and
15   Section 12.7.2 for regional gaps and uncertainties).
16
17
18   11.7.3.1 Mechanisms and drivers
19
20   Severe convective storms are sometimes embedded in synoptic-scale weather systems, such as TCs, ETCs,
21   and fronts (Kunkel et al., 2013). They are also generated as individual events as mesoscale convective
22   systems (MCSs) and mesoscale convective complexes (MCCs) (a special type of a large,organized and long-
23   lived MCS), without being clearly embedded within larger-scale weather systems. In addition to the general
24   vigorousness of precipitation, hail, and winds associated with MCSs, characteristics of MCSs are viewed in
25   new perspectives in recent years, probably because of both the development of dense mesoscale observing
26   networks and advances in high-resolution mesoscale modelling (Sections 11.7.3.2 and 11.7.3.3). The
27   horizontal scale of MCSs is discussed with their organization of the convective structure and it is examined
28   with a concept of "convective aggregation" in recent years (Holloway et al., 2017). MCSs sometimes take a
29   linear shape and stay almost stationary with successive production of cumulonimbus on the upstream side
30   (back-building type convection), and cause heavy rainfall (Schumacher and Johnson, 2005). Many of the
31   recent severe rainfall events in Japan are associated with band-shaped precipitation systems (Kunii et al.,
32   2016; Oizumi et al., 2018; Tsuguti et al., 2018; Kato, 2020), suggesting common characteristics of severe
33   precipitation, at least in East Asia. The convective modes of severe storms in the United States can be
34   classified into rotating or linear modes and preferable environmental conditions for these modes, such as
35   vertical shear, have been identified (Trapp et al., 2005; Smith et al., 2013; Allen, 2018). Cloud microphysics
36   characteristics of MCSs were examined and the roles of warm rain processes on extreme precipitation were
37   emphasized recently (Sohn et al., 2013; Hamada et al., 2015; Hamada and Takayabu, 2018). Idealized
38   studies also suggest the importance of ice and mixed-phase processes of cloud microphysics on extreme
39   precipitation (Sandvik et al., 2018; Bao and Sherwood, 2019). However, it is unknown whether the types of
40   MCSs are changing in recent periods or observed ubiquitously all over the world.
41
42   Severe convective storms occur under conditions preferable for deep convection, that is, conditionally
43   unstable stratification, sufficient moisture both in lower and middle levels of the atmosphere, and a strong
44   vertical shear. These large-scale environmental conditions are viewed as necessary conditions for the
45   occurrence of severe convective systems, or the resulting tornadoes and lightning, and the relevance of these
46   factors strongly depends on the region (e.g., Antonescu et al., 2016a; Allen, 2018; Tochimoto and Niino,
47   2018). Frequently used metrics are atmospheric static stability, moisture content, convective available
48   potential energy (CAPE) and convective inhibition (CIN), wind shear or helicity, including storm-relative
49   environmental helicity (SREH) (Tochimoto and Niino, 2018; Elsner et al., 2019). These metrics, largely
50   controlled by large-scale atmospheric circulations or synoptic weather systems, such as TCs and ETCs, are
51   then generally used to examine severe convective systems. In particular, there is high confidence that CAPE
52   in the tropics and the subtropics increases in response to global warming (Singh et al., 2017a), as supported
53   by theoretical studies (Singh and O’Gorman, 2013; Seeley and Romps, 2015; Romps, 2016; Agard and
54   Emanuel, 2017). The uncertainty, however, arises from the balance between factors affecting severe storm
55   occurrence. For example, the warming of mid-tropospheric temperatures leads to an increase in the freezing
     Do Not Cite, Quote or Distribute                    11-100                                     Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   level, which leads to increased melting of smaller hailstones, while there may be some offset by stronger
 2   updrafts driven by increasing CAPE, which would favour the growth of larger hailstones, leading to less
 3   melting when falling (Allen, 2018; Mahoney, 2020).
 4
 5   There are few studies on relations between changes in severe convective storms and those of the large-scale
 6   circulation patterns. Tornado outbreaks in the United States are usually associated with ETCs with their
 7   frontal systems and TCs (Fuhrmann et al., 2014; Tochimoto and Niino, 2016). In early June in East Asia,
 8   associated with the Baiu/Changma/Mei-yu, severe precipitation events are frequently caused by MCSs.
 9   Severe precipitation events are also caused by remote effects of TCs, known as predecessor rain events
10   (PREs) (Galarneau et al., 2010). Atmospheric rivers and other coherent types of enhanced water vapour flux
11   also have the potential to induce severe convective systems (Kamae et al., 2017; Ralph et al., 2018; Waliser
12   and Guan, 2017; see Section 8.3.2.8.1). Combined with the above drivers, topographic effects also enhance
13   the intensity and duration of severe convective systems and the associated precipitation (Ducrocq et al.,
14   2008; Piaget et al., 2015). However, the changes in these drivers are not generally significant, so their
15   relations to severe convective storms are unclear.
16
17   In summary, severe convective storms are sometimes embedded in synoptic-scale weather systems, such as
18   TCs, ETCs, and fronts, and modulated by large-scale atmospheric circulation patterns. The occurrence of
19   severe convective storms and the associated severe events, including tornadoes, hail, and lightning, is
20   affected by environmental conditions of the atmosphere, such as CAPE and vertical shear. The uncertainty,
21   however, arises from the balance between these environmental factors affecting severe storm occurrence.
22
23
24   11.7.3.2 Observed trends
25
26   Observed trends in severe convective storms or MCSs are not well documented, but the climatology of
27   MCSs has been analysed in specific regions (North America, South America, Europe, Asia; regional aspects
28   of convective storms are separately assessed in Chapter 12). As the definition of severe convective storms
29   varies depending on the literature, it is not straightforward to make a synthesizing view of observed trends in
30   severe convective storms in different regions. However, analysis using satellite observations provides a
31   global view of MCSs (Kossin et al., 2017). The global distribution of thunderstorms is captured (Zipser et
32   al., 2006; Liu and Zipser, 2015) by using the satellite precipitation measurements by the Tropical Rainfall
33   Measuring Mission (TRMM) and Global Precipitation Mission (GPM) (Hou et al., 2014). The climatological
34   characteristics of MCSs are provided by satellite analyses in South America (Durkee and Mote, 2010;
35   Rasmussen and Houze, 2011; Rehbein et al., 2018) and those of MCC in the Maritime Continent by
36   Trismidianto and Satyawardhana (2018). Analysis of the environmental conditions favourable for severe
37   convective events indirectly indicates the climatology and trends of severe convective events (Allen et al.,
38   2018; Taszarek et al., 2018, 2019), though favourable conditions depend on the location, such as the
39   difference for tornadoes associated with ETCs between the United States and Japan (Tochimoto and Niino,
40   2018).
41
42   Observed trends in severe convective storms are highly regionally dependent. In the United States, it is
43   indicated that there is no significant increase in convective storms, and hail and severe thunderstorms
44   (Kossin et al., 2017; Kunkel et al., 2013). There is an upward trend in the frequency and intensity of extreme
45   precipitation events in the United States (high confidence) (Kunkel et al., 2013; Easterling et al., 2017), and
46   MCSs have increased in occurrence and precipitation amounts since 1979 (limited evidence) (Feng et al.,
47   2016). Significant interannual variability of hailstone occurrences is found in the Southern Great Plains of
48   the United States (Jeong et al., 2020). The mean annual number of tornadoes has remained relatively
49   constant, but their variability of occurrence has increased since the 1970s, particularly over the 2000s, with a
50   decrease in the number of days per year, but an increase in the number of tornadoes on these days (Brooks et
51   al., 2014; Elsner et al., 2015, 2019; Kossin et al., 2017; Allen, 2018). There has been a shift in the
52   distribution of tornadoes, with increases in tornado occurrence in the mid-south of the US and decreases over
53   the High Plains (Gensini and Brooks, 2018). Trends in MCSs are relatively more visible for particular
54   aspects of MCSs, such as lengthening of active seasons and dependency on duration. MCSs have increased
55   in occurrence and precipitation amounts since 1979 (Easterling et al., 2017). Feng et al. (2016) analysed that
     Do Not Cite, Quote or Distribute                    11-101                                      Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   the observed increases in springtime total and extreme rainfall in the central United States are dominated by
 2   MCSs, with increased frequency and intensity of long-lasting MCSs.
 3
 4   Studies on trends in severe convective storms and their ingredients outside of the United States are limited.
 5   Westra et al. (2014) found that there is an increase in the intensity of short-duration convective events
 6   (minutes to hours) over many regions of the world, except eastern China. In Europe, a climatology of
 7   tornadoes shows an increase in detected tornadoes between 1800 to 2014, but this trend might be affected by
 8   the density of observations (Antonescu et al., 2016b, 2016a). An increase in the trend in extreme daily
 9   rainfall is found in southeastern France, where MCSs play a key role in this type of event (Blanchet et al.,
10   2018; Ribes et al., 2019). Trend analysis of the mean annual number of days with thunderstorms since 1979
11   in Europe indicates an increase over the Alps and central, southeastern, and eastern Europe, with a decrease
12   over the southwest (Taszarek et al., 2019). In the Sahelian region, Taylor et al. (2017) analysed MCSs using
13   satellite observations since 1982 and showed an increase in the frequency of extreme storms. In Bangladesh,
14   the annual number of propagating MCSs decreased significantly during 1998-2015 based on TRMM
15   precipitation data (Habib et al., 2019). Prein and Holland (2018) estimated the hail hazard from large-scale
16   environmental conditions using a statistical approach and showed increasing trends in the United States,
17   Europe, and Australia. However, trends in hail on regional scales are difficult to validate because of an
18   insufficient length of observations and inhomogeneous records (Allen, 2018). The high spatial variability of
19   hail suggests it is reasonable that there would be local signals of both positive and negative trends and the
20   trends that are occurring in hail globally are uncertain. In China, the total number of days that have either a
21   thunderstorm or hail have decreased by about 50% from 1961 to 2010, and the reduction in these severe
22   weather occurrences correlates strongly with the weakening of the East Asian summer monsoon (Zhang et
23   al., 2017b). More regional aspects of severe convective storms are detailed in Chapter 12.
24
25   In summary, because the definition of severe convective storms varies depending on the literature and the
26   region, it is not straightforward to make a synthesizing view of observed trends in severe convective storms
27   in different regions. In particular, observational trends in tornadoes, hail, and lightning associated with
28   severe convective storms are not robustly detected due to insufficient coverage of the long-term
29   observations. There is medium confidence that the mean annual number of tornadoes in the United States has
30   remained relatively constant, but their variability of occurrence has increased since the 1970s, particularly
31   over the 2000s, with a decrease in the number of days per year and an increase in the number of tornadoes on
32   these days (high confidence). Detected tornadoes have also increased in Europe, but the trend depends on the
33   density of observations.
34
35
36   11.7.3.3 Model evaluation
37
38   The explicit representation of severe convective storms requires non-hydrostatic models with horizontal grid
39   spacings below 5 km, denoted as convection-permitting models or storm-resolving models (Section 10.3.1).
40   Convection-permitting models are becoming available to run over a wide domain, such as a continental scale
41   or even over the global area, and show realistic climatological characteristics of MCSs (Prein et al., 2015;
42   Guichard and Couvreux, 2017; Satoh et al., 2019). Such high-resolution simulations are computationally too
43   expensive to perform at the larger domain and for long periods and alternative methods by using an RCM
44   with dynamical downscaling are generally used (Section 10.3.1). Convection-permitting models are used as
45   the flagship project of CORDEX to particularly study projections of thunderstorms (Section 10.3.3).
46   Simulations of North American MCSs by a convection-permitting model conducted by Prein et al. (2017a)
47   were able to capture the main characteristics of the observed MCSs, such as their size, precipitation rate,
48   propagation speed, and lifetime. Cloud-permitting model simulations in Europe also showed sub-daily
49   precipitation realistically (Ban et al., 2014; Kendon et al., 2014). Evaluation of precipitation conducted using
50   convection-permitting simulations around Japan showed that finer resolution improves intense precipitation
51   (Murata et al., 2017). MCSs over Africa simulated using convection-permitting models showed better
52   extreme rainfall (Kendon et al., 2019) and diurnal cycles and convective rainfall over land than the coarser-
53   resolution RCMs or GCMs (Stratton et al., 2018; Crook et al., 2019).
54
55   The other modeling approach is the analysis of the environmental conditions that control characteristics of
     Do Not Cite, Quote or Distribute                     11-102                                     Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   severe convective storms using the typical climate model results in CMIP5/6 (Allen, 2018). Severe
 2   convective storms are generally formed in environments with large CAPE and tornadic storms are, in
 3   particular, formed with a combination of large CAPE and strong vertical wind shear. As the processes
 4   associated with severe convective storms occur over a wide range of spatial and temporal scales, some of
 5   which are poorly understood and are inadequately sampled by observational networks, the model calibration
 6   approaches are in general difficult and insufficiently validated. Therefore, model simulations and their
 7   interpretations should be done with much caution.
 8
 9   In summary, there are typically two kinds of modeling approaches for studying changes in severe convective
10   storms. One is to use convection-permitting models in wider regions or the global domain in time-sliced
11   downscaling methods to directly simulate severe convective storms. The other is the analysis of the
12   environmental conditions that control characteristics of severe convective storms by using coarse-resolution
13   GCMs. Even in finer-resolution convection-permitting models, it is difficult to directly simulate tornadoes,
14   hail storms, and lightning, so modeling studies of these changes are limited.
15
16
17   11.7.3.4 Detection and attribution, event attribution
18
19   It is extremely difficult to detect differences in time and space of severe convective storms (Kunkel et al.,
20   2013). Although some ingredients that are favourable for severe thunderstorms have increased over the
21   years, others have not; thus, overall, changes in the frequency of environments favourable for severe
22   thunderstorms have not been statistically significant. Event attribution studies on severe convective events
23   have now been undertaken for some cases. For the case of the July 2018 heavy rainfall event in Japan (BOX
24   11.3), Kawase et al. (2019) took a storyline approach to show that the rainfall during this event in Japan was
25   increased by approximately 7% due to the recent rapid warming around Japan. For the case of the December
26   2015 extreme rainfall event in Chennai, India, the extremity of the event was equally caused by the warming
27   trend in the Bay of Bengal SSTs and the strong El Niño conditions (van Oldenborgh et al., 2016; Boyaj et al.,
28   2018). For hailstorms, such as those that caused disasters in the United States in 2018, detection of the role
29   of climate change in changing hail storms is more difficult, because hail storms are not, in general, directly
30   simulated by convection-permitting models and not adequately represented by the environmental parameters
31   of coarse-resolution GCMs (Mahoney, 2020).
32
33   In summary, it is extremely difficult to detect and attribute changes in severe convective storms, except for
34   case study approaches by event attribution. There is limited evidence that extreme precipitation associated
35   with severe convective storms has increased in some cases.
36
37
38   11.7.3.5 Projections
39
40   Future projections of severe convective storms are usually studied either by analysing the environmental
41   conditions simulated by climate models or by a time slice approach with higher-resolution convection-
42   permitting models by comparing simulations downscaled with climate model results under historical
43   conditions and those under hypothesized future conditions (Kendon et al., 2017; Allen, 2018). Up to now,
44   individual studies using convection-permitting models gave projections of extreme events associated with
45   severe convective storms in local regions, and it is not generally possible to obtain global or general views of
46   projected changes of severe convective storms. Prein et al. (2017b) investigated future projections of North
47   American MCS simulations and showed an increase in MCS frequency and an increase in total MCS
48   precipitation volume by the combined effect of increases in maximum precipitation rates associated with
49   MCSs and increases in their size. Rasmussen et al. (2017) investigated future changes in the diurnal cycle of
50   precipitation by capturing organized and propagating convection and showed that weak to moderate
51   convection will decrease and strong convection will increase in frequency in the future. Ban et al. (2015)
52   found the day-long and hour-long precipitation events in summer intensify in the European region covering
53   the Alps. Kendon et al. (2019) showed future increases in extreme 3-hourly precipitation in Africa. Murata et
54   al. (2015) investigated future projections of precipitation around Japan and showed a decrease in monthly
55   mean precipitation in the eastern Japan Sea region in December, suggesting convective clouds become
     Do Not Cite, Quote or Distribute                    11-103                                      Total pages: 345
     Final Government Distribution                      Chapter11                                   IPCC AR6 WGI

 1   shallower in the future in the winter over the Japan Sea.
 2
 3   The other approach is the projection of the environmental conditions that control characteristics of severe
 4   convective storms by analysing climate model results. There is high confidence that CAPE, particularly
 5   summertime mean CAPE and high percentiles of the CAPE in the tropics and subtropics, increases in
 6   response to global warming in an ensemble of climate models including those of CMIP5, mainly from
 7   increased low-level specific humidity (Sobel and Camargo, 2011; Singh et al., 2017a; Chen et al., 2020b).
 8   CIN becomes stronger over most land areas under global warming, resulting mainly from reduced low-level
 9   relative humidity over land (Chen et al., 2020b). However, there are large differences within the CMIP5
10   ensemble for environmental conditions, which contribute to some degree of uncertainty (Allen, 2018).
11   Because the relation between simulated environments in models and the occurrence of severe convective
12   storms are in general insufficiently validated, the confidence level of the projection of severe convective
13   storms with the approach of the environmental conditions is generally low.
14
15   In the United States, projected changes in the environmental conditions show an increase in CAPE and no
16   changes or decreases in the vertical wind shear, suggesting favourable conditions for an increase in severe
17   convective storms in the future, but the interpretation of how tornadoes or hail will change is an open
18   question because of the strong dependence on shear (Brooks, 2013). Diffenbaugh et al. (2013) showed robust
19   increases in the occurrence of the favourable environments for severe convective storms with increased
20   CAPE and stronger low-level wind shear in response to future global warming. A downscaling approach
21   showed that the variability of the occurrence of severe convective storms increases in spring in late 21st
22   century simulations (Gensini and Mote, 2015). Future changes in hail occurrence in the United States
23   examined through convection-permitting dynamical downscaling suggested that the hail season may begin
24   earlier in the year and exhibit more interannual variability with increases in the frequency of large hail in
25   broad areas over the United States (Trapp et al., 2019). There is medium confidence that the frequency and
26   variability of the favourable environments for severe convective storms will increase in spring, and low
27   confidence for summer and autumn (Diffenbaugh et al., 2013; Gensini and Mote, 2015; Hoogewind et al.,
28   2017). The occurrence of hail events in Colorado in the United States was examined by comparing both
29   present-day and projected future climates using high-resolution model simulations capable of resolving
30   hailstorms (Mahoney et al., 2012), which showed hail is almost eliminated at the surface in the future in
31   most of the simulations, despite more intense future storms and significantly larger amounts of hail generated
32   in-cloud.
33
34   Future changes in severe convection environments show enhancement of instability with less robust changes
35   in the frequency of strong vertical wind shear in Europe (Púčik et al. 2017) and in Japan (Muramatsu et al.
36   2016). In Japan, the frequency of conditions favourable for strong tornadoes increases in spring and partly in
37   summer.
38
39   In summary, the average and maximum rain rates associated with severe convective storms increase in a
40   warming world in some regions including the USA (high confidence). There is high confidence from climate
41   models that CAPE increases in response to global warming in the tropics and subtropics, suggesting more
42   favourable environments for severe convective storms. The frequency of springtime severe convective
43   storms is projected to increase in the USA leading to a lengthening of the severe convective storm season
44   (medium confidence), evidence in other regions is limited. There is significant uncertainty about projected
45   regional changes in tornadoes, hail, and lightning due to limited analysis of simulations using convection-
46   permitting models (high confidence).
47
48
49   11.7.4 Extreme winds
50
51   Extreme winds are defined here in terms of the strongest near-surface wind speeds that are generally
52   associated with extreme storms, such as TCs, ETCs, and severe convective storms. In previous IPCC reports,
53   near-surface wind speed (including extremes), has not been assessed as a variable in its own right, but rather
54   in the context of other extreme atmospheric or oceanic phenomena. The exception was the SREX report
55   (Seneviratne et al., 2012), which specifically examined past changes and projections of mean and extreme
     Do Not Cite, Quote or Distribute                    11-104                                     Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   near-surface wind speeds. A strong decline in extreme winds compared to mean winds was reported for the
 2   continental northern mid-latitudes. Due to the small number of studies and uncertainties in terrestrial-based
 3   surface wind measurements, the findings were assigned low confidence in the SREX. AR5 reported a
 4   weakening of mean and maximum winds from the 1960s or 1970s to the early 2000s in the tropics and mid-
 5   latitudes and increases in high latitudes, but with low confidence in changes in the observed surface winds
 6   over land (Hartmann et al., 2013). Observed trends in mean wind speed over land and the ocean are assessed
 7   in Section 2.3.1.4.4. Aspects of climate impact-drivers for winds are addressed in Section 12.3.3 and 12.5.2.3
 8   and their regional changes are assessed in Section 12.4.
 9
10   Observationally, although not specifically addressing extreme wind speed changes, negative surface wind
11   speed trends (stilling) were found in the tropics and mid-latitudes of both hemispheres of -0.014 m s-1 year-1,
12   while positive trends were reported at high latitudes poleward of 70 degrees, based on a review of 148
13   studies (McVicar et al., 2012a). An earlier study attributed the stilling to both changes in atmospheric
14   circulation and an increase in surface roughness due to an overall increase in vegetation cover (Vautard et
15   al., 2010). Since then, a number of additional studies have mostly confirmed these general negative mean-
16   wind trends based on anemometer data for Spain (Azorin-Molina et al., 2017), Turkey, (Dadaser-Celik and
17   Cengiz, 2014), the Netherlands, (Wever, 2012), Saudi Arabia, (Rehman, 2013), Romania, (Marin et al.,
18   2014), and China (Chen et al., 2013). Lin et al. (2013) note that wind speed variability over China is greater
19   at high elevation locations compared to those closer to mean sea level. Hande et al. (2012), using radiosonde
20   data, found an increase in surface wind speed on Macquarie Island.
21
22   A number of new studies have examined surface wind speeds over the ocean based on ship-based
23   measurements, satellite altimeters, and Special Sensor Microwave/Imagers (SSM/I) (Tokinaga and Xie,
24   2011; Zieger et al., 2014). It has been noted that wind speed trends tend to be stronger in altimeter
25   measurements, although the spatial patterns of change are qualitatively similar in both instruments (Zieger et
26   al., 2014). Liu et al. (2016) found positive trends in surface wind speeds over the Arctic Ocean in 20 years of
27   satellite observations. Small positive trends in mean wind speed were found in 33 years of satellite data,
28   together with larger trends in the 90th percentile values over global oceans (Ribal and Young, 2019). These
29   results were consistent with an earlier study that found a positive trend in 1-in-100 year wind speeds (Young
30   et al., 2012). A positive change in mean wind speeds was found for the Arabian Sea and the Bay of Bengal
31   (Shanas and Kumar, 2015) and Zheng et al. (2017) found that positive wind speed trends over the ocean
32   were larger during winter seasons than summer seasons.
33
34   Changes in extreme winds are associated with changes in the characteristics (locations, frequencies, and
35   intensities) of extreme storms, including TCs, ETCs, and severe convective storms. For TCs, as assessed in
36   Section 11.7.1.5, it is projected that the average peak TC wind speeds will increase globally with warming,
37   while the global frequency of TCs over all categories will decrease or remain unchanged; the average
38   location where TCs reach their peak wind-intensity will migrate poleward in the western North Pacific
39   Ocean as the tropics expand with warming. Frequency, intensities, and geographical distributions of extreme
40   wind events associated with TCs will change according to these TC changes. For ETCs, by the end of the
41   century, CMIP5 models show the number of ETCs associated with extreme winds will significantly decrease
42   in the mid- and high latitudes of the Northern Hemisphere in winter, with the projected decrease being larger
43   over the Atlantic (Kar-Man Chang, 2018), while it will significantly increase irrespective of the season in the
44   Southern Hemisphere (Chang, 2017)(Section 11.7.2.4). Over the ocean in the subtropics, a large ensemble of
45   60-km global model simulations indicated that extreme winds associated with storm surges will intensify
46   over 15–35°N in the Northern Hemisphere (Mori et al., 2019). On the other hand, extreme surface wind
47   speeds will mostly decrease due to decreases in the number and intensity of TCs over most tropical areas of
48   the Southern Hemisphere (Mori et al., 2019). The projected changes in the frequency of extreme winds are
49   associated with the future changes in TCs and ETCs.
50
51   Extreme cyclonic windstorms that share some characteristics with both TCs and ETCs occur regularly over
52   the Mediterranean Sea and are often referred to as “medicanes” (Ragone et al., 2018; Miglietta and Rotunno,
53   2019; Ragone et al., 2018; Miglietta and Rotunno, 2019; Zhang et al., 2020e). Medicanes pose substantial
54   threats to regional islands and coastal zones. A growing body of literature consistently found that the
55   frequency of medicanes decreases under warming, while the strongest medicanes become stronger
     Do Not Cite, Quote or Distribute                    11-105                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   (González-Alemán et al., 2019; Tous et al., 2016; Romero and Emanuel, 2017; Romera et al., 2017;
 2   Cavicchia et al., 2014; Romero and Emanuel, 2013; Gaertner et al., 2007). This is also consistent with
 3   expected global changes in TCs under warming (11.7.1). Based on the consistency of these studies, it is
 4   likely that medicanes will decrease in frequency, while the strongest medicanes become stronger under
 5   warming scenario projections (medium confidence).
 6
 7   In summary, the observed intensity of extreme winds is becoming less severe in the lower to mid-latitudes,
 8   while becoming more severe in higher latitudes poleward of 60 degrees (low confidence). Projected changes
 9   in the frequency and intensity of extreme winds are associated with projected changes in the frequency and
10   intensity of TCs and ETCs (medium confidence).
11
12
13   11.8 Compound events
14
15   The IPCC SREX (SREX Ch3) first defined compound events as “(1) two or more extreme events occurring
16   simultaneously or successively, (2) combinations of extreme events with underlying conditions that amplify
17   the impact of the events, or (3) combinations of events that are not themselves extremes but lead to an
18   extreme event or impact when combined”. Further definitions of compound events have emerged since the
19   SREX. Zscheischler et al. (2018) defined compound events broadly as “the combination of multiple drivers
20   and/or hazards that contributes to societal or environmental risk”. This definition is used in the present
21   assessment, because of its clear focus on the risk framework established by the IPCC, and also highlighting
22   that compound events may not necessarily result from dependent drivers. Compound events have been
23   classified into preconditioned events, where a weather-driven or climate-driven precondition aggravates the
24   impacts of a hazard; multivariate events, where multiple drivers and/or hazards lead to an impact; temporally
25   compounding events, where a succession of hazards leads to an impact; and spatially compounding events,
26   where hazards in multiple connected locations cause an aggregated impact (Zscheischler et al., 2020).
27   Drivers include processes, variables, and phenomena in the climate and weather domain that may span over
28   multiple spatial and temporal scales. Hazards (such as floods, heat waves, wildfires) are usually the
29   immediate physical precursors to negative impacts, but can occasionally have positive outcomes (Flach et
30   al., 2018).
31
32
33   11.8.1 Overview
34
35   The combination of two or more – not necessarily extreme – weather or climate events that occur i) at the
36   same time, ii) in close succession, or iii) concurrently in different regions, can lead to extreme impacts that
37   are much larger than the sum of the impacts due to the occurrence of individual extremes alone. This is
38   because multiple stressors can exceed the coping capacity of a system more quickly. The contributing events
39   can be of similar types (clustered multiple events) or of different types (Zscheischler et al., 2020). Many
40   major weather- and climate-related catastrophes are inherently of a compound nature (Zscheischler et al.,
41   2018). This has been highlighted for a broad range of hazards, such as droughts, heat waves, wildfires,
42   coastal extremes, and floods (Westra et al., 2016; AghaKouchak et al., 2020; Ridder et al., 2020). Co-
43   occurring extreme precipitation and extreme winds can result in infrastructural damage (Martius et al.,
44   2016); the compounding of storm surge and precipitation extremes can cause coastal floods (Wahl et al.,
45   2015); the combination of drought and heat can lead to tree mortality (Allen et al., 2015)(see also Section
46   11.6); wildfires increase occurrences of hailstorms and lightning (Zhang et al., 2019e). Compound storm
47   types consisting of co-located cyclone, front and thunderstorm systems have a higher chance of causing
48   extreme rainfall and extreme winds than individual storm types (Dowdy and Catto, 2017). Extremes may
49   occur at similar times at different locations (De Luca et al., 2020a,b) but affect the same system, for instance,
50   spatially-concurrent climate extremes affecting crop yields and food prices (Anderson et al., 2019; Singh et
51   al., 2018). Studies also show an increasing risk for breadbasket regions to be concurrently affected by
52   climate extremes with increasing global warming, even between 1.5°C and 2°C of global warming (Gaupp et
53   al., 2019) (Box 11.2). Concomitant extreme conditions at different locations become more probable as
54   changes in climate extremes are emerging over an increasing fraction of the land area (Sections 11.2.3,
55   11.2.4, 11.8.2, 11.8.3; Box 11.4).
     Do Not Cite, Quote or Distribute                     11-106                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1
 2   Finally, impacts may occur because of large multivariate anomalies in the climate drivers, if systems are
 3   adapted to historical multivariate climate variability (Flach et al., 2017). For instance, ecosystems are
 4   typically adapted to the local covariability of temperature and precipitation such that a bivariate anomaly
 5   may have a large impact even though neither temperature nor precipitation may be extreme based on a
 6   univariate assessment (Mahony and Cannon, 2018). Given that almost all systems are affected by weather
 7   and climate phenomena at multiple space-time scales (Raymond et al., 2020), it is natural to consider
 8   extremes in a compound event framework. It should be noted, however, that multi-hazard dependencies can
 9   also decrease risk, for instance when hazards are negatively correlated (Hillier et al., 2020). Despite this
10   recognition, the literature on past and future changes in compound events has been limited, but is growing.
11   This section assesses examples of types of compound events in available literature.
12
13   In summary, compound events include the combination of two or more – not necessarily extreme – weather
14   or climate events that occur i) at the same time, ii) in close succession, or iii) concurrently in different
15   regions. The land area affected by concurrent extremes has increased (high confidence). Concurrent extreme
16   events at different locations, but possibly affecting similar sectors (e.g., breadbaskets) in different regions,
17   will become more frequent with increasing global warming, in particular above +2°C of global warming
18   (high confidence).
19
20
21   11.8.2 Concurrent extremes in coastal and estuarine regions
22
23   Coastal and estuarine zones are prone to a number of meteorological extreme events and also to concurrent
24   extremes. A major climati-impact driver in coastal regions around the world is floods (Chapter 12), and flood
25   occurrence may be influenced by the dependence between storm surge, extreme rainfall, river flow, but also
26   by sea level rise, waves and tides, as well as groundwater for estuaries. Floods with multiple drivers are often
27   referred to as “compound floods” (Wahl et al., 2015; Moftakhari et al., 2017; Bevacqua et al., 2020b).
28
29   At US coasts, the probability of co-occurring storm surge and heavy precipitation is higher for the
30   Atlantic/Gulf coast relative to the Pacific coast (Wahl et al., 2015). Furthermore, six studied locations on the
31   US coast with long overlapping time series show an increase in the dependence between heavy precipitation
32   and storm surge over the last century, leading to more frequent co-occurring storm surge and heavy
33   precipitation events at the present day (Wahl et al., 2015). Storm surge and extreme rainfall are also
34   dependent in most locations on the Australian coasts (Zheng et al., 2013) and in Europe along the Dutch
35   coasts (Ridder et al., 2018), along the Mediterranean Sea, the Atlantic coast and the North Sea (Bevacqua et
36   al., 2019). The probability of flood occurrence can be assessed via the dependence between storm surge and
37   river flow (Bevacqua et al., 2020a, 2020b). For instance, the occurrence of a North Sea storm surge in close
38   succession with an extreme Rhine or Meuse river discharge is much more probable due to their dependence,
39   compared to if both events would be independent (Kew et al., 2013; Klerk et al., 2015). Significant
40   dependence between high sea levels and high river discharge are found for more than half of the available
41   station observations, which are mostly located around the coasts of North America, Europe, Australia, and
42   Japan (Ward et al., 2018). Combining global river discharge with a global storm surge model, hotspots of
43   compound flooding have been discovered that are not well covered by observations, including Madagascar,
44   Northern Morocco, Vietnam, and Taiwan (Couasnon et al., 2020). In the Dutch Noorderzijlvest area, there is
45   more than a two-fold increase in the frequency of exceeding the highest warning level compared to the case
46   if storm surge and heavy precipitation were independent (van den Hurk et al., 2015). In other regions and
47   seasons, the dependence can be insignificant (Wu et al., 2018b) and there can be significant seasonal and
48   regional differences in the storm surge-heavy precipitation relationship. Assessments of flood probabilities
49   are often not based on actual flood measurements and instead are estimated from its main drivers including
50   astronomical tides, storm surge, heavy precipitation, and high streamflow. Such single driver analyses might
51   underestimate flood probabilities if multiple correlated drivers contribute to flood occurrence (e.g., van den
52   Hurk et al., 2015).
53
54   Many coastal areas are also prone to the occurrence of compound precipitation and wind extremes, which
55   can cause damage, including to infrastructure and natural environments. A high percentage of co-occurring
     Do Not Cite, Quote or Distribute                     11-107                                     Total pages: 345
     Final Government Distribution                     Chapter11                                    IPCC AR6 WGI

 1   wind and precipitation extremes are found in coastal regions and in areas with frequent tropical cyclones.
 2   Finally, the combination of extreme wave height and duration is also shown to influence coastal erosion
 3   processes (Corbella and Stretch, 2012).
 4
 5   Aspects of concurrent extremes in coastal and estuarine environments have increased in frequency and/or
 6   magnitude over the last century in some regions. These include an increase in the dependence between heavy
 7   precipitation and storm surge over the last century, leading to more frequent co-occurring storm surge and
 8   heavy precipitation events in the present day along US coastlines (Wahl et al., 2015). In Europe, the
 9   probability of compound flooding occurrence increases most strongly along the Atlantic coast and the North
10   Sea under strong warming. This increase is mostly driven by an intensification of precipitation extremes and
11   aggravated flooding probability due to sea level rise (Bevacqua et al., 2019). At the global scale and under a
12   high emissions scenario, the concurrence probability of meteorological conditions driving compound
13   flooding would increase by more than 25% on average along coastlines worldwide by 2100, compared to the
14   present (Bevacqua et al., 2020b). Sea level extremes and their physical impacts in the coastal zone arise from
15   a complex set of atmospheric, oceanic, and terrestrial processes that interact on a range of spatial and
16   temporal scales and will be modified by a changing climate, including sea level rise (McInnes et al., 2016).
17   Interactions between sea level rise and storm surges (Little et al., 2015), and sea level and fluvial flooding
18   (Moftakhari et al., 2017) are projected to lead to more frequent and more intense compound coastal flooding
19   events as sea levels continue to rise.
20
21   In summary, there is medium confidence that over the last century the probability of compound flooding has
22   increased in some locations, including along the US coastline. There is medium confidence that the
23   occurrence and magnitude of compound flooding in coastal regions will increase in the future due to both sea
24   level rise and increases in heavy precipitation.
25
26
27   11.8.3 Concurrent droughts and heat waves
28
29   Concurrent droughts and heat waves have a number of negative impacts on human society and natural
30   ecosystems. Studies since SREX and AR5 show several occurrences of observed combinations of drought
31   and heat waves in various regions.
32
33   Over most land regions, temperature and precipitation are strongly negatively correlated during summer
34   (Zscheischler and Seneviratne, 2017), mostly due to land-atmosphere feedbacks (Sections 11.1.6, 11.3.2),
35   but also because synoptic-scale weather systems favourable for extreme heat are also unfavourable for rain
36   (Berg et al., 2015). This leads to a strong correlation between droughts and heat waves (Zscheischler and
37   Seneviratne, 2017). Drought events characterized by low precipitation and extreme high temperatures have
38   occurred, for example, in California (AghaKouchak et al., 2014), inland eastern Australia (King et al., 2014),
39   and large parts of Europe (Orth et al., 2016b). The 2018 growing season was both record-breaking dry and
40   hot in Germany (Zscheischler and Fischer, 2020).
41
42   The probability of co-occurring meteorological droughts and heat waves has increased in the observational
43   period in many regions and will continue to do so under unabated warming (Herrera-Estrada and Sheffield,
44   2017; Zscheischler and Seneviratne, 2017; Hao et al., 2018; Sarhadi et al., 2018; Alizadeh et al., 2020; Wu et
45   al., 2021). Overall, projections of increases in co-occurring drought and heat waves are reported in northern
46   Eurasia (Schubert et al., 2014), Europe ; Sedlmeier et al., 2018), southeast Australia (Kirono et al., 2017),
47   multiple regions of the United States (Diffenbaugh et al., 2015; Herrera-Estrada and Sheffield 2017),
48   northwest China (Li et al., 2019c; Kong et al., 2020) and India (Sharma and Mujumdar, 2017). The dominant
49   signal is related to the increase in heat wave occurrence, which has been attributed to anthropogenic forcing
50   (11.3.4). This means that even if drought occurrence is unaffected, compound hot and dry events will be
51   more frequent (Sarhadi et al., 2018; Yu and Zhai, 2020).
52
53   Drought and heat waves are also associated with fire weather, related through high temperatures, low soil
54   moisture, and low humidity. Fire weather refers to weather conditions conducive to triggering and sustaining
55   wildfires, which generally include temperature, soil moisture, humidity, and wind (Chapter 12). Concurrent
     Do Not Cite, Quote or Distribute                    11-108                                    Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   hot and dry conditions amplify conditions that promote wildfires (Schubert et al., 2014; Littell et al., 2016;
 2   Hope et al., 2019, Dowdy, 2018). Burnt area extent in western US forests (Abatzoglou and Williams, 2016)
 3   and particularly in California (Williams et al., 2019) has been linked to anthropogenic climate change via a
 4   significant increase in vapour pressure deficit, a primary driver of wildfires. A study of the western US
 5   examined the correlation between historical water-balance deficits and annual area burned, across a range of
 6   vegetation types from temperate rainforest to desert (McKenzie and Littell, 2017). The relationship between
 7   temperature and dryness, and wildfire, varied with ecosystem type, and the fire-climate relationship was both
 8   nonstationary and vegetation-dependent. In many fire-prone regions, such as the Mediterranean and China’s
 9   Daxing’anling region, projections for increased severity of future drought and heat waves may lead to an
10   increased frequency of wildfires relative to observed (Ruffault et al., 2018; Tian et al., 2017). Observations
11   show a long-term trend towards more dangerous weather conditions for bushfires in many regions of
12   Australia, which is attributable at least in part to anthropogenic climate change (Dowdy, 2018). There is
13   emerging evidence that recent regional surges in wildland fires are being driven by changing weather
14   extremes (SRCCL Ch2, Cross-Chapter Box 3; Jia et al., 2019). Between 1979 and 2013, the global burnable
15   area affected by long fire-weather seasons doubled, and the mean length of the fire-weather season increased
16   by 19% (Jolly et al., 2015). However, at the global scale, the total burned area has been decreasing between
17   1998 and 2015 due to human activities mostly related to changes in land use (Andela et al., 2017). Given the
18   projected high confidence increase in compound hot and dry conditions, there is high confidence that fire
19   weather conditions will become more frequent at higher levels of global warming in some regions. This
20   assessment is also consistent with assessments of Chapter 12 for regional projected changes in fire weather.
21   The SRCCL Ch2 assessed with high confidence that future climate variability is expected to enhance the risk
22   and severity of wildfires in many biomes such as tropical rainforests.
23
24   In summary, there is high confidence that concurrent heat waves and droughts have increased in frequency
25   over the last century at the global scale due to human influence. There is medium confidence that weather
26   conditions that promote wildfires (fire weather) have become more probable in southern Europe, northern
27   Eurasia, the US, and Australia over the last century. There is high confidence that compound hot and dry
28   conditions become more probable in nearly all land regions as global mean temperature increases. There is
29   high confidence that fire weather conditions will become more frequent at higher levels of global warming in
30   some regions.
31
32
33   [START BOX 11.4 HERE]
34
35
36   BOX 11.4: Case study: Global-scale concurrent climate anomalies at the example of the 2015-2016
37             extreme El Niño and the 2018 boreal spring/summer extremes
38
39   Occurrence of concurrent or near-concurrent extremes in different parts of a region, or in different locations
40   around the world challenges adaptation and risk management capacity. This can occur as a result of natural
41   climate variability, as climates in different parts of the world are inter-connected through teleconnections. In
42   addition, in a warming climate, the probability of having several locations being affected simultaneously by
43   e.g. hot extremes and heat waves increases strongly as a function of global warming, with detectable changes
44   even for changes as small as +0.5°C of additional global warming (Sections 11.2.5 and 11.3, Cross-chapter
45   Box 11.1). Recent articles have highlighted the risks associated with concurrent extremes over large spatial
46   scales (e.g. Lehner and Stocker, 2015; Boers et al., 2019; Gaupp et al., 2019). There is evidence that such
47   global-scale extremes associated with hot temperature extremes are increasing in occurrence (Sippel et al.,
48   2015; Vogel et al., 2019). Hereafter, the focus is on two recent global-scale events that featured concurrent
49   extremes in several regions across the world. The first focuses on concurrent extremes driven by variability
50   in tropical Pacific SSTs associated with the 2015-2016 extreme El Niño, while the second is a case study of
51   the impacts of global warming combined with abnormal atmospheric circulation patterns in the 2018 boreal
52   spring/summer.
53
54
55   [START BOX 11.4, FIGURE 1 HERE]
56
     Do Not Cite, Quote or Distribute                    11-109                                      Total pages: 345
     Final Government Distribution                          Chapter11                                        IPCC AR6 WGI

 1   Box 11.4, Figure 1: Analysis of the percentage of land area affected by temperature extremes larger than two (orange)
 2                       or three (blue) standard deviations in June-July-August (JJA) between 30°N and 80°N using a
 3                       normalization. The more appropriate estimate is the corrected normalization. These panels show
 4                       for both estimates a substantial increase in the overall land area affected by very high hot extremes
 5                       since 1990 onward. Adapted from Sippel et al. (2015)
 6
 7   [END BOX 11.4, FIGURE 1 HERE]
 8
 9
10   The extreme El Niño in 2015-2016
11
12   El Niño-Southern Oscillation (ENSO) is one of the phenomena that have the ability to bring multitudes of
13   extremes in different parts of the world, especially in the extreme cases of El Niño (Annex VI.4).
14   Additionally, the background climate warming associated with greenhouse gas forcing can significantly
15   exacerbate extremes in parts of the world even under normal El Niño conditions. The 2015-2016 El Niño
16   event was one of the three extreme El Nino events since 1980s since the availability of satellite rainfall
17   observations. According to some measures, it was the strongest El Niño over the past 145 years (Barnard et
18   al., 2017). The 2015-2016 warmth was unprecedented at the central equatorial Pacific (Niño4: 5°N–5°S,
19   150°E–150°W) and this exceptional warmth was unlikely to have occurred entirely naturally, appearing to
20   reflect an anthropogenically forced trend (Newman et al., 2018)). In particular, its signal was seen in very
21   high monthly Global Mean Surface Temperature (GMST) values in late 2015 and early 2016, contributing to
22   the highest record of GMST in 2016 (Section 2.3.1.1). Both the ENSO amplitude and the frequency of high-
23   magnitude events since 1950 is higher than over the pre-industrial period (medium confidence; Section
24   2.4.2), suggesting that global extremes similar to those associated with the 2015-2016 El Niño would occur
25   more frequently under further increases in global warming. Hereafter, the 2015-2016 El Niño event is
26   referred to as “the 2015-2016 extreme El Niño” (Annex VI.4.1). A brief summary of extreme events that
27   happened in 2015-2016 is provided in Section 6.2.2, 6.5.1.1 of the Special Report on the Ocean and
28   Cryosphere in a Changing Climate (SROCC’s). We provide some highlights illustrating extremes that
29   occurred in different parts of the world during the 2015-2016 extreme El Niño in BOX11.4-Table 1, as well
30   as a short summary hereafter.
31
32   Several regions were strongly affected by droughts in 2015, including Indonesia, Australia, the Amazon
33   region, Ethiopia, Southern Africa, and Europe. As a result, global measurements of land water anomalies
34   were particularly low in that year (Humphrey et al., 2018). In 2015, Indonesia experienced a severe drought
35   and forest fire causing pronounced impact on economy, ecology and human health due to haze crisis (Field
36   et al., 2016; Huijnen et al., 2016; Patra et al., 2017; Hartmann et al., 2018). The northern part of Australia
37   experienced high temperatures and low precipitation between late 2015 and early 2016, and the extensive
38   mangrove trees were damaged along the Gulf of Carpentaria in northern Australia (Duke et al., 2017). The
39   Amazon region experienced the most intense droughts of this century in 2015-2016. This drought was more
40   severe than the previous major droughts that occurred in the Amazon in 2005 and 2010 (Lewis et al., 2011;
41   Erfanian et al., 2017; Panisset et al., 2018). The 2015-2016 Amazon drought impacted the entirety of South
42   America north of 20°S during the austral spring and summer (Erfanian et al., 2017). It also increased forest
43   fire incidence by 36% compared to the preceding 12 years (Aragão et al., 2018) and as a consequence,
44   increased the biomass burning outbreaks and the carbon monoxide (CO) concentration in the area, affecting
45   air quality (Ribeiro et al., 2018). This out-of-season drought affected the water availability for human
46   consumption and agricultural irrigation and it also left rivers with very low water levels, without conditions
47   of ship transportation, due to large sandbanks, preventing the arrival of food, medicines, and fuels (INMET,
48   2017). Eastern African countries were impacted by drought in 2015. It was found that the drought in
49   Ethiopia, which was the worst in several decades, was associated with the 2015-2016 extreme El Niño that
50   developed early in the year (Blunden and Arndt, 2016; Philip et al., 2018b). It was suggested that
51   anthropogenic warming contributed to the 2015 Ethiopian and southern African droughts by increasing SSTs
52   and local air temperatures (Funk et al., 2016, 2018b; Yuan et al., 2018a). It has also been suggested that the
53   2015-2016 extreme El Niño affected circulation patterns in Europe during the 2015-2016 winter (Geng et al.,
54   2017; Scaife et al., 2017).
55

     Do Not Cite, Quote or Distribute                         11-110                                         Total pages: 345
     Final Government Distribution                           Chapter11                                            IPCC AR6 WGI

 1   It was identified that 2015 was a year of a particularly high CO2 growth rate, possibly related to some of the
 2   mentioned droughts, in particular in Indonesia and the Amazon region, leading to higher CO2 release in
 3   combination with less CO2 uptake from land areas (Humphrey et al., 2018). The impact of the 2015-2016
 4   extreme El Niño on vegetation systems via drought was also shown from satellite data (Kogan and Guo,
 5   2017). Overall, tropical forests were a carbon source to the atmosphere during the 2015–2016 El Niño–
 6   related drought, with some estimates suggesting that up to 2.3 PgC were released (Brando et al., 2019).
 7
 8   The 2015-2016 extreme El Niño has induced extreme precipitation in some regions. Severe rainfall events
 9   were observed in Chennai city in India in Devember 2015 and Yangtze river region in China in June-July
10   2016, and it was shown that these rainfall events are partly attributed to the 2015-2016 extreme El Niño (van
11   Oldenborgh et al., 2016; Boyaj et al., 2018; Sun and Miao, 2018; Yuan et al., 2018b; Zhou et al., 2018).
12
13   In 2015, the activity of tropical cyclones was notably high in the North Pacific (Blunden and Arndt, 2016).
14   Over the western North Pacific, the number of category 4 and 5 Tropical Cyclones (TCs) was 13, which is
15   more than twice its typical annual value of 6.3 (Zhang et al., 2016a). Similarly, a record-breaking number of
16   TCs was observed in the eastern North Pacific, particularly in the western part of that domain (Collins et al.,
17   2016; Murakami et al., 2017a). These extraordinary TC activities were related to the average SST anomaly
18   during that year, which were associated with the 2015-2016 extreme El Niño and the positive phase of the
19   Pacific Meridional Mode (PMM) (Murakami et al., 2017a; Hong et al., 2018; Yamada et al., 2019).
20   However, it has been suggested that the intense TC activities in both the western and the eastern North
21   Pacific in 2015 were not only due to the El Niño, but also to a contribution of anthropogenic forcing
22   (Murakami et al., 2017a; Yang et al., 2018d). The impact of the Indian Ocean SST also was suggested to
23   contribute to the extreme TC activity in the western North Pacific in 2015 (Zhan et al., 2018). In contrast, in
24   Australia, it was the least active TC season since satellite records began in 1969-70 (Blunden and Arndt,
25   2017).
26
27
28   [START BOX 11.4, TABLE 1 HERE]
29
30   Box 11.4, Table 1: List of events related to the 2015-2016 Extreme El Niño in the literature.
31
      Region               Period                Events                          References
      Indonesia            July 2015 to June     droughts, forest fire           (Field et al., 2016; Huijnen et al., 2016; Patra et
                           2016                                                  al., 2017; Hartmann et al., 2018)
      Northern Australia   Between late 2015     high temperature and drought    (Duke et al., 2017)
                           and early 2016
      Amazon               September 2015 to     droughts, forest fire           (Jiménez-Muñoz et al., 2016; Erfanian et al.,
                           May 2016                                              2017; Aragão et al., 2018; Panisset et al., 2018;
                                                                                 Ribeiro et al., 2018)
      The entirety of      Austral spring and    droughts                        (Erfanian et al., 2017)
      South America        2015-2016 summer
      north of 20°S
      Ethiopia             February-September    droughts                        (Blunden and Arndt, 2016; Philip et al., 2018b)
                           2015
      Southern Africa      November 2015–        droughts                        (Funk et al., 2016, 2018a; Blamey et al., 2018;
                           April 2016                                            Yuan et al., 2018a)
      Europe               Boreal 2015-2016      effects on of circulation       (Geng et al., 2017; Scaife et al., 2017)
                           winter                patterns
      India                May 2016              high temperature                (van Oldenborgh et al., 2018)
      India                December 2015         extreme rainfall                (van Oldenborgh et al., 2016; Boyaj et al.,
                                                                                 2018)
      China                June-July 2016        extreme rainfall                (Sun and Miao, 2018; Yuan et al., 2018b; Zhou
                                                                                 et al., 2018)
      Western North        Boreal summer 2015    the large number (13) of        (Blunden and Arndt, 2016; Mueller et al.,
      Pacific                                    category 4 and 5 tropical       2016a; Zhang et al., 2016b; Hong et al., 2018;
                                                 cyclones                        Yamada et al., 2019)
      Eastern North        Boreal summer 2015    a record-breaking number of     (Collins et al., 2016; Murakami et al., 2017a)
      Pacific                                    tropical cyclones
      Global               2015-2016 El Niño     high CO2 release to the         (Humphrey et al., 2018; Brando et al., 2019)
                                                 atmosphere associated with
     Do Not Cite, Quote or Distribute                          11-111                                            Total pages: 345
     Final Government Distribution                           Chapter11                                        IPCC AR6 WGI

                                                  droughts and fires in several
                                                  affected regions
 1
 2
 3   [END BOX 11.4, TABLE 1 HERE]
 4
 5
 6   Global-scale temperature extremes and concurrent precipitation extremes in boreal 2018 spring and
 7   summer
 8
 9   In the 2018 boreal spring-summer season (May-August), wide areas of the mid-latitudes in the Northern
10   Hemisphere experienced heat extremes and in part enhanced drought (Kornhuber et al., 2019; Vogel et al.,
11   2019; Box 11.3, Figure 2). The reported impacts included the following (Vogel et al., 2019): 90 deaths from
12   heat strokes in Quebec (Canada), 1469 deaths from heat strokes in Japan (Shimpo et al., 2019a), 48 heat-
13   related deaths in South Korea (Min et al., 2020), heat warning affecting 90,000 students in the USA, fires in
14   numerous countries (Canada (British Columbia), USA (California), Lapland, Latvia), crop losses in the UK,
15   Germany and Switzerland (Vogel et al., 2019) and overall in central and northern Europe (leading to yield
16   reductions of up to 50% for the main crops; Toreti et al., 2019), fish deaths in Switzerland, and melting of
17   roads in the Netherlands and the UK, among others. In addition to the numerous hot and dry extremes, an
18   extremely heavy rainfall event occurred over wide areas of Japan from 28 June to 8 July 2018 (Tsuguti et al.,
19   2018), which was followed by a heat wave (Shimpo et al., 2019b). The heavy precipitation event caused
20   more than 230 deaths in Japan, and was named as “the Heavy Rain Event of July 2018”.
21
22   The heavy precipitation event was characterized by unusually widespread and persistent rainfall and locally
23   anomalous total precipitation led by band-shaped precipitation systems, which are frequently associated with
24   heavy precipitation events in East Asia (Kato, 2020; Section 11.7.3). The extreme rainfall in Japan was
25   caused by anomalous moisture transport with a combination of abnormal jet condition (Takemi and Unuma,
26   2019; Takemura et al., 2019; Tsuji et al., 2019; Yokoyama et al., 2020), which can be viewed as an
27   atmospheric river (Yatagai et al., 2019; Sections 8.2.2.8, 11.7.2) caused by intensified inflow velocity and
28   high SST around Japan (Kawase et al., 2019; Sekizawa et al., 2019).
29
30   This precipitation event and the subsequent heat wave are related to abnormal condition of the jet and North
31   Pacific Subtropical High in this month (Shimpo et al., 2019a; Ren et al., 2020), which caused extreme
32   conditions from Europe, Eurasia, and North America (Kornhuber et al., 2019; Box 11.4, Figure 2). A role of
33   Atlantic SST anomaly on the meandering jets and the subtropical high have been suggested (Liu et al.,
34   2019a). These dynamic and thermodynamic components generally have substantial influence on extreme
35   rainfall in East Asia (Oh et al., 2018), but it is under investigation whether these factors were due to
36   anthropogenic forcing.
37
38
39   [START BOX 11.4, FIGURE 2 HERE]

40   Box 11.4, Figure 2: Meteorological conditions in July 2018. The color shading shows the monthly mean near-
41                       surface air temperature anomaly with respect to 1981 to 2010. Contour lines indicate the
42                       geopotential height in m, highlighted are the isolines on 12'000 m and 12'300 m, which indicate
43                       the approximate positions of the polar-front jet and subtropical jet, respectively. The light blue-
44                       green ellipse shows the approximate extent of the strong precipitation event that occurred at the
45                       beginning of July in the region of Japan and Korea. All data is from the global ECMWF
46                       Reanalysis v5 (ERA5, Hersbach et al., 2020).

47   [END BOX 11.4, FIGURE 2 HERE]
48
49
50   Regarding the hot extremes that occurred across the Northern Hemisphere in the 2018 boreal May-July time
51   period, Vogel et al. (2019) found that the event was unprecedented in terms of the total area affected by hot
52   extremes (on average about 22% of populated and agricultural areas in the Northern Hemisphere) for that

     Do Not Cite, Quote or Distribute                          11-112                                         Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   period, but was consistent with a +1°C climate which was the estimated global mean temperature anomaly
 2   around that time (for 2017; SR1.5). This study also found that events similar to the 2018 May-July
 3   temperature extremes would approximately occur 2 out of 3 years under +1.5°C of global warming, and
 4   every year under +2°C of global warming. Imada et al. (2019) also suggests that the mean annual occurrence
 5   of extremely hot days in Japan will be expected to increase by 1.8 times under a global warming level of 2°C
 6   above pre-industrial levels. Kawase et al. (2019) showed that the extreme rainfall in Japan during this event
 7   was increased by approximately 7% due to recent rapid warming around Japan. Hence, it is virtually certain
 8   that these 2018 concurrent events would not have occurred without human-induced global warming.
 9   Concurrent events of this type are also projected to happen more frequently under higher levels of global
10   warming. On the other hand, there is currently low confidence in projected changes in the frequency or
11   strength of the anomalous circulation patterns leading to concurrent extremes (e.g. Cross-Chapter Box 10.1).
12
13   The case studies presented in this Box illustrate the current state of knowledge regarding the contribution of
14   human-induced climate change to recent concurrent extremes in the global domain. Recent years have seen a
15   more frequent occurrence of such events. The heat wave in Europe in the 2019 boreal summer and its
16   coverage in the global domain is an additional example (Vautard et al., 2020a). However, there are still very
17   few studies investigating which types of concurrent extreme events could occur under increasing global
18   warming. It has been noted that such events could also be of particular risk for concurrent impacts in the
19   world’s breadbaskets (Zampieri et al., 2017; Kornhuber et al., 2020).
20
21   In summary, the 2015-2016 extreme El Niño and the 2018 boreal spring/summer extremes were two
22   examples of recent concurrent extremes. The El Niño event in 2015-2016 was one of the three extreme El
23   Niño events since 1980s and there are many extreme events concurrently observed in this period including
24   droughts, heavy precipitation, and more frequent intense tropical cyclones. Both the ENSO amplitude and
25   the frequency of high-magnitude events since 1950 is higher than over the pre-industrial period (medium
26   confidence), suggesting that global extremes similar to those associated with the 2015-2016 El Niño would
27   occur more frequently under further increases in global warming. The 2018 boreal spring/summer extremes
28   were characterized by heat extremes and enhanced droughts in wide areas of the mid-latitudes in the
29   Northern Hemisphere and extremely heavy rainfall in East Asia. These concurrent events were generally
30   related to abnormal condition of the jet and North Pacific Subtropical High, but also amplified by
31   background global warming. It is virtually certain that these 2018 concurrent extreme events would not have
32   occurred without human-induced global warming. Recent years have seen a more frequent occurrence of
33   such concurrent events. However, it is still unknown which types of concurrent extreme events could occur
34   under increasing global warming.
35
36
37   [END BOX 11.4 HERE]
38
39
40   11.9 Regional information on extremes
41
42   This section complements the assessments of changes in temperature extremes (Section 11.3), heavy
43   precipitation (Section 11.4), and droughts (Section 11.6), by providing additional regional details. Owing to
44   the large number of regions and space limitations, the regional assessment for each of the AR6 reference
45   regions (see Section 1.5.2.2 for a description) is presented here in a set of tables. The tables are organized
46   according to types of extremes (temperature, heavy precipitation, droughts) for Africa (Tables 11.4-11.6),
47   Asia (Table 11.7-11.9), Australasia (Tables 11.10-11.12), Central and South America (Tables 11.13-11.15),
48   Europe (Tables 11.16-11.18), and North America (Tables 11.19-11.21). Each table contains regional
49   assessments for observed changes, the human contribution to the observed changes, and projections of
50   changes in these extremes at 1.5°C, 2°C and 4°C of global warming. Expanded versions of the tables with
51   full evidence and rationale for assessments are provided in the Chapter Appendix (Tables 11.A.4-11.A.21).
52
53
54   11.9.1 Overview
55
56   Sections 11.9.2, 11.9.3., and 11.9.4 provide brief summaries of the underlying evidence used to derive the
     Do Not Cite, Quote or Distribute                    11-113                                      Total pages: 345
     Final Government Distribution                      Chapter11                                    IPCC AR6 WGI

 1   regional assessments for temperature extremes, heavy precipitation events, and droughts, respectively. The
 2   assessments take into account evidence from studies based on global datasets (global studies), as well as
 3   regional studies. Global studies include analyses for all continents and AR6 regions with sufficient data
 4   coverage, and provide an important basis for cross-region consistency, as the same data and methods are
 5   used for all regions. However, individual regional studies may include additional information that is missed
 6   in global studies and thus provide an important regional calibration for the assessment.
 7
 8   The assessments are presented using the calibrated confidence and likelihood language (Box 1.1). Low
 9   confidence is assessed when there is limited evidence, either because of a lack of available data in the region
10   and/or a lack of relevant studies. Low confidence is also assessed when there is a lack of agreement on the
11   evidence of a change, which may be due to large variability or inconsistent changes depending on the
12   considered subregions, time frame, models, assessed metrics, or studies. In cases when the evidence is
13   strongly contradictory, for example with substantial regional changes of opposite sign, “mixed signal” is
14   indicated. With an assessment of low confidence, the direction of change is not indicated in the tables. A
15   direction of change (increase or decrease) is provided with an assessment of medium confidence, high
16   confidence, likely, or higher likelihood levels. Likelihood assessments are only provided in the case of high
17   confidence. In some cases, there may be confidence in a small or no change.
18
19   For projections, changes are assessed at three global warming levels (GWLs, CC-Box 11.1): 1.5°C, 2°C and
20   4°C. Literature based both on GWL projections and on scenario-based projections is used for the
21   assessments. In the case of literature on scenario-based projections, a mapping between scenarios/time
22   frames and GWLs was performed as documented in CC-Box 11.1. Projections of changes in temperature and
23   precipitation extremes are assessed relative to two different baselines: the recent past (1995-2014) and pre-
24   industrial (1850-1900). With smaller changes relative to the variability, in particular because droughts
25   happen on longer timescales compared to extremes of daily temperature and precipitation, it is more difficult
26   to distinguish changes in drought relative to the recent past. As such, changes in droughts are assessed
27   relative to the pre-industrial baseline, unless indicated otherwise.
28
29
30   11.9.2 Temperature extremes
31
32   Tables 11.4, 11.7, 11.10, 11.13, 11.16, and 11.19 include assessments for past chn temperature extremes and
33   their attribution, as well as future projections. The evidence is mostly drawn from changes in metrics based
34   on daily maximum and minimum temperatures, similar to those used in Section 11.3. The regional
35   assessments start from global studies that used consistent analyses for all regions globally with sufficient
36   data. This includes Dunn et al. (2020) for observed changes and Li et al. (2020) and the Chapter 11
37   Supplementary Material (11.SM) for projections with the CMIP6 multi-model ensemble. Evidence from
38   regional studies, and those based on the CMIP5 multi-model ensemble or CORDEX simulations, are then
39   used to refine the confidence assessments. For attribution, Seong et al. (2020) provide a consistent analysis
40   for AR6 regions and Wang et al. (2017) for SREX regions. Additional regional studies, including event
41   attribution analyses (Section 11.2), are used when available. In some regions that were not analysed in Seong
42   et al. (2020) and with no known event attribution studies, medium confidence of a human contribution is
43   assessed when there is strong evidence of changes from observations that are in the direction of model
44   projected changes for the future, the magnitude of projected changes increases with global warming, and
45   there is no other evidence to the contrary. Understanding of how temperature extremes change with the mean
46   temperature and overwhelming evidence of a human contribution to the observed larger-scale changes in the
47   mean temperature and temperature extremes further support this assessment.
48
49
50   11.9.3 Heavy precipitation
51




     Do Not Cite, Quote or Distribute                    11-114                                      Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   Tables 11.5, 11.8, 11.11, 11.14, 11.17, and 11.20 include assessments for past changes in heavy precipitation
 2   events and their attribution, as well as future projections. The evidence is mostly drawn from changes in
 3   metrics based on one-day or five-day precipitation amounts, as addressed in Section 11.4. Similar to
 4   temperature extremes, the assessment of changes in heavy precipitation uses global studies, including Dunn
 5   et al. (2020) and Sun et al. (2020) for observed changes, and Li et al. (2020) and the Chapter 11
 6   Supplementary Material (11.SM) for projected changes using the CMIP6 multi-model ensemble. For
 7   attribution, Paik et al. (2020) provided continental analyses where data coverage was sufficient, but no
 8   attribution studies based on global data are available for the regional scale. For each region, regional studies,
 9   and studies based on the CMIP5 multi-model ensemble or CORDEX simulations, are also considered in the
10   assessments for past changes, attribution, and projections.
11
12
13   11.9.4 Droughts
14
15   Tables 11.6, 11.9, 11.12, 11.15, 11.18, and 11.21 provide regional tables on past, attributed and projected
16   changes in droughts. The assessment is subdivided in three drought categories corresponding to four drought
17   types: i) meteorological droughts, ii) agricultural and ecological droughts, and iii) hydrological droughts (see
18   Section 11.6). A list of metrics and global studies used for the assessments is provided below. The evidence
19   from global studies is complemented in each continent with evidence from regional studies. An overview of
20   studies considered for the assessments in projections is provided in Table 11.3.
21
22   Meteorological droughts are assessed based on observed and projected changes in precipitation-only metrics
23   such as the Standardized Precipitation Index (SPI) and Consecutive Dry Days (CDD). Observed changes are
24   assessed based on two global studies, Dunn et al. (2020) for CDD and Spinoni et al. (2019) for SPI. For
25   projections, evidence for changes at 1.5°C and 2°C of global warming is drawn from Xu et al. (2019) and
26   Touma et al. (2015) (based on RCP8.5 for 2010-2054 compared to 1961-2005) for SPI (CMIP5) and the
27   Chapter 11 Supplementary Material (11.SM) for CDD (CMIP6). For projections at 4°C of global warming,
28   evidence is drawn from several sources, including Touma et al. (2015) and Spinoni et al. (2020) for SPI
29   (from CMIP5 and CORDEX, respectively), and the Chapter 11 Supplementary Material (11.SM) for CDD
30   (CMIP6). No global-scale studies are available for the attribution of meteorological drought, and thus this
31   assessment is based on regional detection and attribution or event attribution studies.
32
33   Agricultural and ecological droughts are assessed based on observed and projected changes in total column
34   soil moisture, complemented by evidence on changes in surface soil moisture, water-balance (precipitation
35   minus evapotranspiration (ET)) and metrics driven by precipitation and atmospheric evaporative demand
36   (AED) such as the SPEI and PDSI (Section 11.6). In the case of the latter, only studies including estimates
37   based on the Penman-Monteith equation (SPEI-PM and PDSI-PM) are considered because of biases
38   associated with temperature-only approaches (Section 11.6). In arid regions in which AED-based metrics can
39   increase strongly in projections, more weight is given to soil moisture projections. For observed changes,
40   evidence is drawn from several sources: Padrón et al. (2020) for changes in precipitation minus ET, as well
41   as soil moisture from the multi-model Land Surface Snow and Soil Moisture Model Intercomparison Project
42   within CMIP6 (LS3MIP, Van Den Hurk et al., 2016; Chapter 11 Supplementary Material (11.SM)); Greve et
43   al. (2014) for changes in precipitation minus ET, and precipitation minus AED; Spinoni et al. (2019) for
44   changes in SPEI-PM; and Dai and Zhao (2017) for changes in PDSI-PM. For projections at 1.5°C of global
45   warming, evidence is drawn from Xu et al. (2019) based on CMIP5 and the Chapter 11 Supplementary
46   Material (11.SM) based on CMIP6 for changes in total column and surface soil moisture, and from Naumann
47   et al. (2018) for changes in SPEI-PM, based on EC-Earth simulations driven with SSTs from seven CMIP5
48   ESMs. For projections at 2°C of global warming, evidence is drawn from Xu et al. (2019) based on CMIP5,
49   and Cook et al. (2020) (SSP1-2.6, 2071-2100 compared to pre-industrial) and the Chapter 11 Supplementary
50   Material (11.SM) based on CMIP6, for changes in total column and surface soil moisture; evidence is also
51   drawn from Naumann et al. (2018) for changes in SPEI-PM. For projections at 4°C of global warming,
52   evidence is mostly drawn from Cook et al. (2020) (SSP3-7.0, 2071-2100) and the Chapter 11 Supplementary
53   Material (11.SM) based on CMIP6 for changes in total column and surface soil moisture, and from Vicente-
54   Serrano et al. (2020) for changes in SPEI-PM based on CMIP5. No global-scale studies with regional-scale
     Do Not Cite, Quote or Distribute                     11-115                                      Total pages: 345
     Final Government Distribution                                     Chapter11                                            IPCC AR6 WGI

 1   information are available for the attribution of agricultural and ecological droughts, and thus this assessment
 2   is based on regional detection and attribution or event attribution studies.
 3
 4   Hydrological droughts are assessed based on observed and projected changes in low flows, complemented
 5   by information on changes in mean runoff. For observed changes, evidence is drawn from three studies (Dai
 6   and Zhao, 2017; Gudmundsson et al., 2019, 2021). For projected changes at 1.5°C of global warming,
 7   evidence is drawn from Touma et al. (2015) based on analyses of the Standardized Runoff Index (SRI)
 8   (CMIP5, based on 2010-2054 compared to 1961-2005), complemented with regional studies when available.
 9   For projected changes at 2°C of global warming, evidence is also drawn from Cook et al. (2020) for changes
10   in runoff in CMIP6 (Scenario SSP1-2.6, 2071-2100), and from Zhai et al. (2020) for changes in low flows
11   based on simulations with a single model. For projected changes at 4°C of global warming, evidence is
12   drawn from Touma et al. (2015) based on CMIP5 analyses of SRI, Cook et al. (2020) for changes in surface
13   and total runoff based on CMIP6, and Giuntoli et al. (2015) for changes in low flows based on the Inter-
14   Sectoral Impact Model Intercomparison Project (ISI-MIP) based on six Global Hydrological Models
15   (GHMs) and five GCMs, including an analysis of inter-model signal-to-noise ratio. One global-scale study
16   with regional-scale information is available for the attribution of hydrological droughts (Gudmundsson et al.,
17   2021), but only in a few AR6 regions. This information was complemented with evidence from regional
18   detection and attribution, and event attribution studies when available.
19
20
21   [START TABLE 11.3 HERE]
22
23   Table 11.3: Global analyses considered for the assessments of drought projections. “MET” refers to meteorological
24               droughts, “AGR/ECOL” to agricultural and ecological droughts, and “HYDR” to hydrological droughts
25
      Reference               Model data          Index                      Drought type   Projection horizon(s)                Baseline
      Chapter 11 Suppl.       CMIP6               CDD, Soil moisture         MET            1.5°C, 2°C, 4°C                      1850-1900
      Material (11.SM)                            (total, surface)
      Cook et al. (2020)      CMIP6               Soil moisture (total,      AGR/ECOL,      2071-2011, SSP1-2.6 (~2°C, CC-Box    1850-1900
                                                  surface), Runoff (total,   HYDR           11.1; Chapter 4, Table 4.2)
                                                  surface)                                  2071-2011, SSP3-7-3 (~4°C, CC-Box
                                                                                            11.1; Chapter 4, Table 4.2)
      Xu et al. (2019)        CMIP5               SPI, Soil moisture         MET,           1.5°C, 2°C                           1971-2000
                                                  (total, surface)           AGR/ECOL
      Touma et al. (2015)     CMIP5               SPI, SRI                   MET, HYDR      2010-2054, RCP8.5 (~1.5°C; CC-Box    1961-2005
                                                                                            11.1, 11.SM.1)
                                                                                            2055-2099, RCP8.5 (~3.5°C, CC-Box
                                                                                            11.1, 11.SM.1)
      Spinoni et al. (2020)   CORDEX              SPI                        MET            2071-2100, RCP4.5 (~2.5°C, CC-Box    1981-2010
                              (CMIP5 driving                                                11.1, 11.SM.1)
                              GCMs, RCMs)                                                   2071-2100, RCP8.5 (~4.5°C, CC-Box
                                                                                            11.1, 11.SM.1)
      Naumann et al.          1 GCM (EC-          SPEI-PM                    AGR/ECOL       1.5°C, 2°C, (3°C)                    0.6°C
      (2018)                  EARTH3-HR
                              v3.1) driven with
                              SST fields from 7
                              CMIP5 GCMs
      Vicente-Serrano et      CMIP5               SPEI-PM                    AGR/ECOL       2070-2100, RCP8.5 (~4.5°C, CC-Box    1970-2000
      al. (2020)                                                                            11.1, 11.SM.1)
      Giuntoli et al.         ISI-MIP (6 GHMs     Low-flows days             HYDR           2066-2099, RCP8-5 (~4°C, CC-Box      1972-2005
      (2015)                  and 5 CMIP5                                                   11.1, 11.SM.1)
                              GCMs)
      Zhai et al. (2020)      1 GHM (VIC)         Extreme low runoff         HYDR           1.5°C, 2°C                           2006-2015
                              driven by 4
                              CMIP5 GCMs
26
27
28   [END TABLE 11.3 HERE]
29
30

     Do Not Cite, Quote or Distribute                                    11-116                                             Total pages: 345
     Final Government Distribution                         Chapter11                                      IPCC AR6 WGI

 1   Frequently Asked Questions
 2
 3   FAQ 11.1: How do changes in climate extremes compare with changes in climate averages?
 4   Human-caused climate change alters the frequency and intensity of climate variables (e.g., surface
 5   temperature) and phenomena (e.g., tropical cyclones) in a variety of ways. We now know that the ways in
 6   which average and extreme conditions have changed (and will continue to change) depend on the variable
 7   and the phenomenon being considered. Changes in local surface temperature extremes follow closely the
 8   corresponding changes in local average surface temperatures. On the contrary, changes in precipitation
 9   extremes (heavy precipitation) generally do not follow those in average precipitation and can even move in
10   the opposite direction (e.g., with average precipitation decreasing but extreme precipitation increasing).
11   Climate change will manifest very differently depending on which region, which season and which variable
12   we are interested in. For example, over some parts of the Arctic, temperatures will warm at rates about 3-4
13   times higher during winter compared to summer months. And in summer, most of northern Europe will
14   experience larger temperatures increases than most places in Southeast South America and Australasia, with
15   differences that can be larger than 1°C depending on the level of global warming. In general, differences
16   across regions and seasons arise because the underlying physical processes differ drastically across regions
17   and seasons.
18   Climate change will also manifest differently for different weather regimes and can lead to contrasting
19   changes in average and extreme conditions. Observations of the recent past and climate model projections
20   show that, in most places, changes in daily temperatures are dominated by a general warming in which both
21   the climatological average and extreme values are shifted towards higher temperatures, making warm
22   extremes more frequent and cold extremes less frequent. The top panels in FAQ 11.1, Figure 1 show
23   projected changes in surface temperature for long-term average conditions (left) and for extreme hot days
24   (right) during the warm season (summer in mid- to high-latitudes). Projected increases in long-term average
25   temperature differ substantially in different places, varying from less than 3°C in some places in central
26   South Asia and southern South America to over 7°C in some places in North America, north Africa and the
27   Middle East. Changes in extreme hot days follow changes in average conditions quite closely, although in
28   some places the warming rates for extremes can be intensified (e.g., southern Europe and the Amazon basin)
29   or weakened (e.g., northern Asia and Greenland) compared to average values.
30   Recent observations and global and regional climate model projections point to changes in precipitation
31   extremes (including both rainfall and snowfall extremes) differing drastically from those in average
32   precipitation. The bottom panels in FAQ 11.1, Figure 1 show projected changes in the long-term average
33   precipitation (left) and in heavy precipitation (right). Averaged precipitation changes show striking regional
34   differences, with substantial drying in places such as southern Europe and northern South America and
35   wetting in places such as Middle East and southern South America. Changes in extreme heavy precipitation
36   are much more uniform, with systematic increases over nearly all land regions. The physical reasons behind
37   the different response of averaged and extreme precipitation are now well understood. The intensification of
38   extreme precipitation is driven by the increase in atmospheric water vapour (about 7% per 1°C of warming
39   near the surface), although this is modulated by various dynamical changes. In contrast, changes in average
40   precipitation are driven not only by moisture increases but also by slower processes that constrain future
41   changes to on be only about 2–3% per 1°C of warming near the surface.
42   In summary, the specific relationship between changes in average and extreme conditions strongly depends
43   on the variable or phenomenon being considered. At the local scale, average and extreme surface
44   temperature changes are strongly related, while average and extreme precipitation changes are often weakly
45   related. For both variables, the changes in average and extreme conditions vary strongly across different
46   places due to the effect of local and regional processes.
47
48   [START FAQ 11.1, FIGURE 1 HERE]
     FAQ 11.1, Figure 1: Global maps of future changes in surface temperature (top panels) and precipitation (bottom
                        panels) for long-term average (left) and extreme conditions (right). All changes were estimated
                        using the CMIP6 ensemble mean for a scenario with a global warming of 4°C relative to 1850-
     Do Not Cite, Quote or Distribute                       11-117                                        Total pages: 345
    Final Government Distribution                       Chapter11                                       IPCC AR6 WGI

                       1900 temperatures. Average surface temperatures refer to the warmest three-month season
                       (summer in mid- to high-latitudes) and extreme temperature refer to the hottest day in a year.
                       Precipitation changes, which can include both rainfall and snowfall changes, are normalized by
                       1850-1900 values and shown in percentage; extreme precipitation refers to the largest daily
                       rainfall in a year.


1   [END FAQ 11.1, FIGURE 1 HERE]
2
3   [END FAQ 11.1 HERE]
4




    Do Not Cite, Quote or Distribute                      11-118                                        Total pages: 345
     Final Government Distribution                        Chapter11                                      IPCC AR6 WGI

 1   [START FAQ 11.2 HERE]
 2
 3   FAQ 11.2: Will unprecedented extremes occur as a result of human-induced climate change?
 4
 5   Climate change has already increased the magnitude and frequency of extreme hot events and decreased the
 6   magnitude and frequency of extreme cold events, and, in some regions, intensified extreme precipitation
 7   events. As the climate moves away from its past and current states, we will experience extreme events that
 8   are unprecedented, either in magnitude, frequency, timing or location. The frequency of these unprecedented
 9   extreme events will increase with increasing global warming. Additionally, the combined occurrence of
10   multiple unprecedented extremes may result in large and unprecedented impacts.
11
12   Human-induced climate change has already affected many aspects of the climate system. In addition to the
13   increase in global surface temperature, many types of weather and climate extremes have changed. In most
14   regions, the frequency and intensity of hot extremes have increased and those of cold extremes have
15   decreased. The frequency and intensity of heavy precipitation events have increased at a global scale and
16   over a majority of land regions. Although extreme events such as land and marine heatwaves, heavy
17   precipitation, drought, tropical cyclones, and associated wildfires and coastal flooding have occurred in the
18   past and will continue to occur in the future, they often come with different magnitudes or frequencies in a
19   warmer world. For example, future heatwaves will last longer and have higher temperatures, and future
20   extreme precipitation events will be more intense in several regions. Certain extremes, such as extreme cold,
21   will be less intense and less frequent with increasing warming.
22
23   Unprecedented extremes – that is, events not experienced in the past – will occur in the future in five
24   different ways (FAQ 11.2, Figure 1). First, events that are considered to be extreme in the current climate
25   will occur in the future with unprecedented magnitudes. Second, future extreme events will also occur with
26   unprecedented frequency. Third, certain types of extremes may occur in regions that have not previously
27   encountered those types of events. For example, as the sea level rises, coastal flooding may occur in new
28   locations, and wildfires are already occurring in areas, such as parts of the Arctic, where the probability of
29   such events was previously low. Fourth, extreme events may also be unprecedented in their timing. For
30   example, extremely hot temperatures may occur either earlier or later in the year than they have in the past.
31
32   Finally, compound events, where multiple extreme events of either different or similar types occur
33   simultaneously and/or in succession, may be more probable or severe in the future. These compound events
34   can often impact ecosystems and societies more strongly than when such events occur in isolation. For
35   example, a drought along with extreme heat will increase the risk of wildfires and agriculture damages or
36   losses. As individual extreme events become more severe as a result of climate change, the combined
37   occurrence of these events will create unprecedented compound events. This could exacerbate the intensity
38   and associated impacts of these extreme events.
39
40   Unprecedented extremes have already occurred in recent years, relative to the 20th century climate. Some
41   recent extreme hot events would have had very little chance of occurring without human influence on the
42   climate (see FAQ 11.3). In the future, unprecedented extremes will occur as the climate continues to warm.
43   Those extremes will happen with larger magnitudes and at higher frequencies than previously experienced.
44   Extreme events may also appear in new locations, at new times of the year, or as unprecedented compound
45   events. Moreover, unprecedented events will become more frequent with higher levels of warming, for
46   example at 3°C of global warming compared to 2°C of global warming.
47
48   [START FAQ 11.2, FIGURE 1 HERE]
49
50   FAQ 11.2, Figure 1: New types of unprecedented extremes that will occur as a result of climate change.
51
52   [END FAQ 11.2, FIGURE 1 HERE]
53
54   [END FAQ 11.2 HERE]
55
     Do Not Cite, Quote or Distribute                       11-119                                       Total pages: 345
     Final Government Distribution                      Chapter11                                     IPCC AR6 WGI

 1   [START FAQ 11.3 HERE]
 2
 3   FAQ 11.3: Did climate change cause that recent extreme event in my country?
 4
 5   While it is difficult to identify the exact causes of a particular extreme event, the relatively new science of
 6   event attribution is able to quantify the role of climate change in altering the probability and magnitude of
 7   some types of weather and climate extremes. There is strong evidence that characteristics of many individual
 8   extreme events have already changed because of human-driven changes to the climate system. Some types of
 9   highly impactful extreme weather events have occurred more often and have become more severe due to
10   these human influences. As the climate continues to warm, the observed changes in the probability and/or
11   magnitude of some extreme weather events will continue as the human influences on these events increase.
12
13   It is common to question whether human-caused climate change caused a major weather- and climate-related
14   disaster. When extreme weather and climate events do occur, both exposure and vulnerability play an
15   important role in determining the magnitude and impacts of the resulting disaster. As such, it is difficult to
16   attribute a specific disaster directly to climate change. However, the relatively new science of event
17   attribution enables scientists to attribute aspects of specific extreme weather and climate events to certain
18   causes. Scientists cannot answer directly whether a particular event was caused by climate change, as
19   extremes do occur naturally and any specific weather and climate event is the result of a complex mix of
20   human and natural factors. Instead, scientists quantify the relative importance of human and natural
21   influences on the magnitude and/or probability of specific extreme weather events. Such information is
22   important for disaster risk reduction planning, because improved knowledge about changes in the probability
23   and magnitude of relevant extreme events enables better quantification of disaster risks.
24
25   On a case-by-case basis, scientists can now quantify the contribution of human influences to the magnitude
26   and probability of many extreme events. This is done by estimating and comparing the probability or
27   magnitude of the same type of event between the current climate – including the increases in greenhouse gas
28   concentrations and other human influences – and an alternate world where the atmospheric greenhouse gases
29   remained at pre-industrial levels. FAQ 11.3 Figure 1 illustrates this approach using differences in
30   temperature and probability between the two scenarios as an example. Both the pre-industrial (blue) and
31   current (red) climates experience hot extremes, but with different probabilities and magnitudes. Hot extremes
32   of a given temperature have a higher probability of occurrence in the warmer current climate than in the
33   cooler pre-industrial climate. Additionally, an extreme hot event of a particular probability will be warmer in
34   the current climate than in the pre-industrial climate. Climate model simulations are often used to estimate
35   the occurrence of a specific event in both climates. The change in the magnitude and/or probability of the
36   extreme event in the current climate compared to the pre-industrial climate is attributed to the difference
37   between the two scenarios, which is the human influence.
38
39   Attributable increases in probability and magnitude have been identified consistently for many hot extremes.
40   Attributable increases have also been found for some extreme precipitation events, including hurricane
41   rainfall events, but these results can vary among events. In some cases, large natural variations in the climate
42   system prevent attributing changes in the probability or magnitude of a specific extreme to human influence.
43   Additionally, attribution of certain classes of extreme weather (e.g., tornadoes) is beyond current modelling
44   and theoretical capabilities. As the climate continues to warm, larger changes in probability and magnitude
45   are expected, and as a result it will be possible to attribute future temperature and precipitation extremes in
46   many locations to human influences. Attributable changes may emerge for other types of extremes as the
47   warming signal increases.
48
49   In conclusion, human-caused global warming has resulted in changes in a wide variety of recent extreme
50   weather events. Strong increases in probability and magnitude, attributable to human influence, have been
51   found for many heat waves and hot extremes around the world.
52
53
54   [START FAQ11.3 FIGURE 1 HERE]
55
     Do Not Cite, Quote or Distribute                     11-120                                     Total pages: 345
     Final Government Distribution                         Chapter11                                      IPCC AR6 WGI

 1   FAQ 11.3, Figure 1: Changes in climate result in changes in the magnitude and probability of extremes. Example of
 2                        how temperature extremes differ between a climate with pre-industrial greenhouse gases (shown
 3                        in blue) and the current climate (shown in orange) for a representative region. The horizontal
 4                        axis shows the range of extreme temperatures, while the vertical axis shows the annual chance of
 5                        each temperature event’s occurrence. Moving towards the right indicates increasingly hotter
 6                        extremes that are more rare (less probable). For hot extremes, an extreme event of a particular
 7                        temperature in the pre-industrial climate would be more probable (vertical arrow) in the current
 8                        climate. An event of a certain probability in the pre-industrial climate would be warmer
 9                        (horizontal arrow) in the current climate. While the climate under greenhouse gases at the pre-
10                        industrial level experiences a range of hot extremes, such events are hotter and more frequent in
11                        the current climate.
12
13   [END FAQ11.3 FIGURE 1 HERE]
14
15   [END FAQ11.3 HERE]
16
17




     Do Not Cite, Quote or Distribute                       11-121                                        Total pages: 345
                                      Final Government Distribution                                         Chapter11                                                  IPCC AR6 WGI

1   Large tables
2   Color scale for tables for changes in temperature extremes and heavy precipitation
                            Fact          Virtually       Extremely       Very           Likely         High            Medium         Low
                                          certain         likely          likely                        confidence      confidence     confidence
     Increasing hot
     extremes, decreasing
     cold extremes
     Decreasing hot
     extremes, increasing
     cold extremes
     Inconsistent sign
3
4   Color scale for tables for changes in droughts
                            Fact          Virtually       Extremely       Very           Likely         High            Medium         Low
                                          certain         likely          likely                        confidence      confidence     confidence
     Increasing drought
     Decreasing drought
     Inconsistent sign
5
6   [START TABLE 11.4 HERE]
7
8   Table 11.4: Observed trends, human contribution to observed trends, and projected changes at 1.5°C, 2°C and 4°C of global warming for temperature extremes in Africa,
9               subdivided by AR6 regions. See Sections 11.9.1 and 11.9.2 for details
     All Africa                                                        Detection and attribution;                                               Projections
                                          Observed trends
                                                                           event attribution                        1.5 °C                           2 °C                             4 °C
                                   Insufficient data for the          Limited evidence for the      CMIP6 models project a robust      CMIP6 models project a           CMIP6 models project a
                                   continent, but there is high       continent, but there is       increase in the intensity and      robust increase in the           robust increase in the
                                   confidence of an increase in       medium confidence in a        frequency of TXx events and a      intensity and frequency of       intensity and frequency of
                                   the intensity and frequency of     human contribution to the     robust decrease in the intensity   TXx events and a robust          TXx events and a robust
                                   hot extremes and decrease in       observed increase in the      and frequency of TNn events (Li    decrease in the intensity and    decrease in the intensity and
                                   the intensity and frequency of     intensity and frequency of    et al., 2020; Annex). Median       frequency of TNn events (Li      frequency of TNn events (Li
                                   cold extremes in all               hot extremes and decrease     increase of more than 0.5°C in     et al., 2020; Annex). Median     et al., 2020; Annex). Median
                                   subregions with sufficient         in the intensity and          the 50-year TXx and TNn events     increase of more than 1°C in     increase of more than 3°C in
                                   data                               frequency of cold             compared to the 1°C warming        the 50-year TXx and TNn          the 50-year TXx and TNn
                                                                      extremes for all              level (Li et al., 2020)            events compared to the 1°C       events compared to the 1°C
                                                                      subregions with sufficient                                       warming level (Li et al.,        warming level ((Li et al.,
                                                                      data                                                             2020)                            2020)
                                   Medium confidence in the           Medium confidence in a        Increase in the intensity and      Increase in the intensity and    Increase in the intensity and
                                   increase in the intensity and      human contribution to the     frequency of hot extremes:         frequency of hot extremes:       frequency of hot extremes:
                                   frequency of hot extremes          observed increase in the      Very likely (compared with the     Extremely likely (compared       Virtually certain (compared
                                   and decrease in the intensity      intensity and frequency of    recent past (1995-2014))           with the recent past (1995-      with the recent past (1995-
                                   and frequency of cold              hot extremes and decrease     Extremely likely (compared with    2014))                           2014))
                                   extremes.                          in the intensity and          pre-industrial)
    Do Not Cite, Quote or Distribute                                   11-122                                                    Total pages: 345
                                 Final Government Distribution                                      Chapter11                                                       IPCC AR6 WGI
                                                               frequency of cold                                                 Virtually certain (compared         Virtually certain (compared
                                                               extremes.                    Decrease in the intensity and        with pre-industrial)                with pre-industrial)
                                                                                            frequency of cold extremes:
                                                                                            Very likely (compared with the       Decrease in the intensity and       Decrease in the intensity and
                                                                                            recent past (1995-2014))             frequency of cold extremes:         frequency of cold extremes:
                                                                                            Extremely likely (compared with      Extremely likely (compared          Virtually certain (compared
                                                                                            pre-industrial)                      with the recent past (1995-         with the recent past (1995-
                                                                                                                                 2014))                              2014))
                                                                                                                                 Virtually certain (compared         Virtually certain (compared
                                                                                                                                 with pre-industrial)                with pre-industrial)
                          2   Significant ncreases in the      Robust evidence of a       CMIP6 models project a robust         CMIP6 models project a robust       CMIP6 models project a robust
    Mediterranean (MED)
                              intensity and frequency of hot   human contribution to the increase in the intensity and          increase in the intensity and       increase in the intensity and
                              extremes and significant         observed increase in the   frequency of TXx events and a         frequency of TXx events and a       frequency of TXx events and a
                              decreases in the intensity and   intensity and frequency of robust decrease in the intensity and robust decrease in the intensity     robust decrease in the intensity
                              frequency of cold extremes       hot extremes and decrease frequency of TNn events (Li et al., and frequency of TNn events            and frequency of TNn events
                              (Peña-Angulo et al., 2020; El    in the intensity and       2020; Annex). Median increase of      (Li et al., 2020; Annex).           (Li et al., 2020; Annex).
                              Kenawy et al., 2013; Acero et                               more than 0.5°C in the 50-year TXx Median increase of more than           Median increase of more than
                                                               frequency of cold
                              al., 2014; Fioravanti et al.,                               and TNn events compared to the 1°C 1°C in the 50-year TXx and             3.5°C in the 50-year TXx and
                                                               extremes (Seong et al.,
                              2016; Ruml et al., 2017;                                    warming level (Li et al., 2020) and TNn events compared to the            TNn events compared to the
                              Türkeş and Erlat, 2018; Donat    2020; Wang et al., 2017; more than 2°C in annual TXx and         1°C warming level (Li et al.,       1°C warming level (Li et al.,
                              et al., 2013, 2014, 2016;        Sippel and Otto, 2014;     TNn compared to pre-industrial        2020) and more than 2.5°C in        2020) and more than 5°C in
                              Filahi et al., 2016; Driouech    Wilcox et al., 2018)       (Annex).                              annual TXx and TNn compred          annual TXx and TNn compared
                              et al., 2021; Dunn et al.,                                                                        to pre-industrial (Annex).          to pre-industrial (Annex).
                              2020)                                                       Additional evidence from CMIP5
                                                                                          and RCM simulations for an increase Additional evidence from              Additional evidence from
                                                                                          in the intensity and frequency of hot CMIP5 and RCM simulations           CMIP5 and RCM simulations
                                                                                                                                for an increase in the intensity    for an increase in the intensity
                                                                                          extremes and decrease in the
                                                                                                                                                                    and frequency of hot extremes
                                                                                          intensity and frequency of cold       and frequency of hot extremes
                                                                                                                                                                    and decrease in the intensity
                                                                                          extremes (Cardoso et al., 2019; Zollo and decrease in the intensity       and frequency of cold extremes
                                                                                          et al., 2016; Weber et al., 2018)     and frequency of cold extremes      (Cardoso et al., 2019; Nastos
                                                                                                                                (Cardoso et al., 2019; Tomozeiu     and Kapsomenakis, 2015;
                                                                                                                                et al., 2014; Abaurrea et al.,      Tomozeiu et al., 2014; Cardell
                                                                                                                                2018; Nastos and                    et al., 2020; Zollo et al., 2016;
                                                                                                                                Kapsomenakis, 2015; Cardell         Giorgi et al., 2014; Driouech et
                                                                                                                                                                    al., 2020; Coppola et al., 2021a;
                                                                                                                                et al., 2020; Zollo et al., 2016;
                                                                                                                                                                    Engelbrecht et al., 2015)
                                                                                                                                Weber et al., 2018; Coppola et
                                                                                                                                al., 2021a)




2
 This region includes both northern Africa and southern Europe
Do Not Cite, Quote or Distribute                       11-123                                                             Total pages: 345
                             Final Government Distribution                                       Chapter11                                                        IPCC AR6 WGI

                          Very likely increase in the      Human influence likely       Increase in the intensity and          Increase in the intensity and      Increase in the intensity and
                          intensity and frequency of       contributed to the           frequency of hot extremes:             frequency of hot extremes:         frequency of hot extremes:
                          hot extremes and decrease        observed increase in         Likely (compared with the recent       Very likely (compared with the     Virtually certain (compared
                          in the intensity and             the intensity and            past (1995-2014))                      recent past (1995-2014))           with the recent past (1995-
                          frequency of cold extremes       frequency of hot             Very likely (compared with pre-        Extremely likely (compared         2014))
                                                           extremes and decrease        industrial)                            with pre-industrial)               Virtually certain (compared
                                                                                                                                                                  with pre-industrial)
                                                           in the intensity and
                                                                                        Decrease in the intensity and          Decrease in the intensity and
                                                           frequency of cold                                                   frequency of cold extremes:        Decrease in the intensity and
                                                                                        frequency of cold extremes:
                                                           extremes                     Likely (compared with the recent       Very likely (compared with the     frequency of cold extremes:
                                                                                        past (1995-2014))                      recent past (1995-2014))           Virtually certain (compared
                                                                                        Very likely (compared with pre-        Extremely likely (compared         with the recent past (1995-
                                                                                        industrial)                            with pre-industrial)               2014))
                                                                                                                                                                  Virtually certain (compared
                                                                                                                                                                  with pre-industrial)
 Sahara (SAH)             Significant increases in the     Strong evidence of           CMIP6 models project a robust          CMIP6 models project a robust      CMIP6 models project a robust
                          intensity and frequency of hot   changes from                 increase in the intensity and          increase in the intensity and      increase in the intensity and
                          extremes and significant         observations that are in     frequency of TXx events and a          frequency of TXx events and a      frequency of TXx events and a
                          decreases in the intensity and   the direction of model       robust decrease in the intensity and   robust decrease in the intensity   robust decrease in the intensity
                          frequency of cold extremes       projected changes for the    frequency of TNn events (Li et al.,    and frequency of TNn events        and frequency of TNn events
                          (Donat et al., 2014a; Moron et   future. The magnitude of     2020; Annex). Median increase of       (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                          al., 2016; Dunn et al., 2020)    projected changes            more than 0.5°C in the 50-year TXx     Median increase of more than       Median increase of more than
                                                           increases with global        and TNn events compared to the 1°C     1°C in the 50-year TXx and         3.5°C in the 50-year TXx and
                                                           warming.                     warming level (Li et al., 2020) and    TNn events compared to the         TNn events compared to the
                                                                                        more than 2°C in annual TXx and        1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                                                        TNn compared to pre-industrial         2020 ) and more than 2.5°C in      2020) and more than 5°C in
                                                                                        (Annex).                               annual TXx and TNn compared        annual TXx and TNn compared
                                                                                                                               to pre-industrial (Annex).         to pre-industrial (Annex).
                                                                                        Additional evidence from CMIP5
                                                                                        and CORDEX simulations for an          Additional evidence from           Additional evidence from
                                                                                        increase in the intensity and          CMIP5 and CORDEX                   CMIP5/CMIP3 and CORDEX
                                                                                        frequency of hot extremes (Weber et    simulations for an increase in     simulations for an increase in
                                                                                        al., 2018)                             the intensity and frequency of     the intensity and frequency of
                                                                                                                               hot extremes (Weber et al.,        hot extremes (Coppola et al.,
                                                                                                                               2018; Coppola et al., 2021a)       2021a; Engelbrecht et al., 2015;
                                                                                                                                                                  Giorgi et al., 2014)
                          Likely increase in the           Medium confidence in a       Increase in the intensity and          Increase in the intensity and      Increase in the intensity and
                          intensity and frequency of hot   human contribution to the    frequency of hot extremes:             frequency of hot extremes:         frequency of hot extremes:
                          extremes and decrease in the     observed increase in the     Likely (compared with the recent       Very likely (compared with the     Virtually certain (compared
                          intensity and frequency of       intensity and frequency of   past (1995-2014))                      recent past (1995-2014))           with the recent past (1995-
                          cold extremes                    hot extremes and decrease    Very likely (compared with pre-        Extremely likely (compared         2014))
                                                           in the intensity and         industrial)                            with pre-industrial)               Virtually certain (compared
                                                           frequency of cold                                                                                      with pre-industrial)
                                                           extremes.                    Decrease in the intensity and          Decrease in the intensity and
                                                                                        frequency of cold extremes:            frequency of cold extremes:        Decrease in the intensity and
                                                                                        Likely (compared with the recent       Very likely (compared with the     frequency of cold extremes:
                                                                                        past (1995-2014))                      recent past (1995-2014))           Virtually certain (compared

Do Not Cite, Quote or Distribute                            11-124                                                     Total pages: 345
                              Final Government Distribution                                        Chapter11                                                        IPCC AR6 WGI

                                                                                          Very likely (compared with pre-        Extremely likely (compared         with the recent past (1995-
                                                                                          industrial).                           with pre-industrial)               2014))
                                                                                                                                                                    Virtually certain (compared
                                                                                                                                                                    with pre-industrial)
 Western Africa (WAF)      Significant increases in the      Strong evidence of           CMIP6 models project a robust          CMIP6 models project a robust      CMIP6 models project a robust
                           intensity and frequency of hot    changes from                 increase in the intensity and          increase in the intensity and      increase in the intensity and
                           extremes and significant          observations that are in     frequency of TXx events and a          frequency of TXx events and a      frequency of TXx events and a
                           decreases in the intensity and    the direction of model       robust decrease in the intensity and   robust decrease in the intensity   robust decrease in the intensity
                           frequency of cold extremes        projected changes for the    frequency of TNn events (Li et al.,    and frequency of TNn events        and frequency of TNn events
                           (Barry et al., 2018; Chaney et    future. The magnitude of     2020; Annex). Median increase of       (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                           al., 2014; Dunn et al., 2020;     projected changes            more than 0.5°C in the 50-year TXx     Median increase of more than       Median increase of more than
                           Mouhamed et al., 2013;            increases with global        and TNn events compared to the         1°C in the 50-year TXx and         3°C in the 50-year TXx and
                           Perkins-Kirkpatrick and           warming.                     1°C warming level (Li et al., 2020)    TNn events compared to the         TNn events compared to the
                           Lewis, 2020)                                                   and more than 1.5°C in annual TXx      1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                                                          and TNn compared to pre-industrial     2020) and more than 2°C in         2020) and more than 4.5°C in
                                                                                          (Annex).                               annual TXx and TNn compared        annual TXx and TNn compared
                                                                                                                                 to pre-industrial (Annex).         to pre-industrial (Annex).
                                                                                          Additional evidence from CMIP5
                                                                                          and CORDEX simulations for an          Additional evidence from           Additional evidence from
                                                                                          increase in the intensity and          CMIP5 and CORDEX                   CMIP5/CMIP3 and CORDEX
                                                                                          frequency of hot extremes (Weber       simulations for an increase in     simulations for an increase in
                                                                                          et al., 2018)                          the intensity and frequency of     the intensity and frequency of
                                                                                                                                 hot extremes (Weber et al.,        hot extremes (Coppola et al.,
                                                                                                                                 2018; Coppola et al., 2021a)       2021a; Engelbrecht et al., 2015;
                                                                                                                                                                    Giorgi et al., 2014)
                           Likely increase in the            Medium confidence in a       Increase in the intensity and          Increase in the intensity and      Increase in the intensity and
                           intensity and frequency of hot    human contribution to the    frequency of hot extremes:             frequency of hot extremes:         frequency of hot extremes:
                           extremes and decrease in the      observed increase in the     Likely (compared with the recent       Very likely (compared with the     Virtually certain (compared
                           intensity and frequency of        intensity and frequency of   past (1995-2014))                      recent past (1995-2014))           with the recent past (1995-
                           cold extremes                     hot extremes and decrease    Very likely (compared with pre-        Extremely likely (compared         2014))
                                                             in the intensity and         industrial)                            with pre-industrial)               Virtually certain (compared
                                                             frequency of cold                                                                                      with pre-industrial)
                                                             extremes.                    Decrease in the intensity and          Decrease in the intensity and
                                                                                          frequency of cold extremes:            frequency of cold extremes:        Decrease in the intensity and
                                                                                          Likely (compared with the recent       Very likely (compared with the     frequency of cold extremes:
                                                                                          past (1995-2014))                      recent past (1995-2014))           Virtually certain (compared
                                                                                          Very likely (compared with pre-        Extremely likely (compared         with the recent past (1995-
                                                                                          industrial).                           with pre-industrial)               2014))
                                                                                                                                                                    Virtually certain (compared
                                                                                                                                                                    with pre-industrial)
 Northern Eastern Africa   Increases in the intensity and    Evidence of a human           CMIP6 models project a robust         CMIP6 models project a robust      CMIP6 models project a robust
 (NEAF)                    frequency of hot extremes         contribution to the           increase in the intensity and         increase in the intensity and      increase in the intensity and
                           and decreases in the intensity    observed increase in the      frequency of TXx events and a         frequency of TXx events and a      frequency of TXx events and a
                           and frequency of cold             intensity and frequency of    robust decrease in the intensity      robust decrease in the intensity   robust decrease in the intensity
                           extremes (Perkins-Kirkpatrick     hot extremes and decrease     and frequency of TNn events (Li       and frequency of TNn events        and frequency of TNn events
                           and Lewis, 2020; Chaney et        in the intensity and          et al., 2020; Annex). Median          (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                           al., 2014; Gebrechorkos et al.,   frequency of cold             increase of more than 0.5°C in        Median increase of more than       Median increase of more than

Do Not Cite, Quote or Distribute                              11-125                                                        Total pages: 345
                             Final Government Distribution                                     Chapter11                                                     IPCC AR6 WGI

                          2018; Omondi et al., 2014;      extremes (Otto et al.,       the 50-year TXx and TNn events     1°C in the 50-year TXx and         2.5°C in the 50-year TXx and
                          Dunn et al., 2020)              2015; Philip et al., 2020;   compared to the 1°C warming        TNn events compared to the         TNn events compared to the
                                                          Marthews et al., 2015;       level (Li et al., 2020) and more   1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                          Kew et al., 2021; Funk et    than 1.5°C in annual TXx and       2020) and more than 2°C in         2020) and more than 4°C in
                                                          al., 2015)                   TNn compared to pre-industrial     annual TXx and TNn compared        annual TXx and TNn compared
                                                                                       (Annex).                           to pre-industrial (Annex).         to pre-industrial (Annex).

                                                                                       Additional evidence from CMIP5     Additional evidence from           Additional evidence from
                                                                                       and CORDEX simulations for an      CMIP5 and CORDEX                   CMIP5/CMIP3 and CORDEX
                                                                                       increase in the intensity and      simulations for an increase in     simulations for an increase in
                                                                                       frequency of hot extremes          the intensity and frequency of     the intensity and frequency of
                                                                                       (Weber et al., 2018)               hot extremes (Weber et al.,        hot extremes Coppola et al.,
                                                                                                                          2018; Coppola et al., 2021a)       2021a; Engelbrecht et al., 2015;
                                                                                                                                                             Giorgi et al., 2014)
                          Medium confidence in the        Medium confidence in a       Increase in the intensity and      Increase in the intensity and      Increase in the intensity and
                          increase in the intensity and   human contribution to the    frequency of hot extremes:         frequency of hot extremes:         frequency of hot extremes:
                          frequency of hot extremes       observed increase in the     Likely (compared with the recent   Very likely (compared with the     Virtually certain (compared
                                                          intensity and frequency of   past (1995-2014))                  recent past (1995-2014))           with the recent past (1995-
                                                          hot extremes                 Very likely (compared with pre-    Extremely likely (compared         2014))
                                                                                       industrial)                        with pre-industrial)               Virtually certain (compared
                                                                                                                                                             with pre-industrial)
                                                                                       Decrease in the intensity and      Decrease in the intensity and
                                                                                       frequency of cold extremes:        frequency of cold extremes:        Decrease in the intensity and
                                                                                       Likely (compared with the recent   Very likely (compared with the     frequency of cold extremes:
                                                                                       past (1995-2014))                  recent past (1995-2014))           Virtually certain (compared
                                                                                       Very likely (compared with pre-    Extremely likely (compared         with the recent past (1995-
                                                                                       industrial).                       with pre-industrial)               2014))
                                                                                                                                                             Virtually certain (compared
                                                                                                                                                             with pre-industrial)
 Central Africa (CAF)     Insufficient data to assess     Limited evidence             CMIP6 models project a robust      CMIP6 models project a robust      CMIP6 models project a robust
                          trends (Dunn et al., 2020)                                   increase in the intensity and      increase in the intensity and      increase in the intensity and
                                                                                       frequency of TXx events and a      frequency of TXx events and a      frequency of TXx events and a
                                                                                       robust decrease in the intensity   robust decrease in the intensity   robust decrease in the intensity
                                                                                       and frequency of TNn events (Li    and frequency of TNn events        and frequency of TNn events
                                                                                       et al., 2020; Annex). Median       (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                                                                                       increase of more than 0C in the    Median increase of more than       Median increase of more than
                                                                                       50-year TXx and TNn events         1°C in the 50-year TXx and         3°C in the 50-year TXx and
                                                                                       compared to the 1°C warming        TNn events compared to the         TNn events compared to the
                                                                                       level (Li et al., 2020) and more   1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                                                       than 1.5°C in annual TXx and       2020) and more than 2°C in         2020) and more than 4.5°C in
                                                                                       TNn compared to pre-industrial     annual TXx and TNn compared        annual TXx and TNn compared
                                                                                       (Annex).                           to pre-industrial (Annex).         to pre-industrial (Annex).

                                                                                       Additional evidence from CMIP5     Additional evidence from           Additional evidence from
                                                                                       and CORDEX simulations for an      CMIP5 and CORDEX                   CMIP5/CMIP3 and CORDEX
                                                                                       increase in the intensity and      simulations for an increase in     simulations for an increase in
                                                                                                                          the intensity and frequency of     the intensity and frequency of

Do Not Cite, Quote or Distribute                           11-126                                                   Total pages: 345
                                  Final Government Distribution                                      Chapter11                                                     IPCC AR6 WGI

                                                                                             frequency of hot extremes          hot extremes (Weber et al.,        hot extremes (Coppola et al.,
                                                                                             (Weber et al., 2018)               2018; Coppola et al., 2021a)       2021a; Engelbrecht et al., 2015;
                                                                                                                                                                   Giorgi et al., 2014)
                               Low confidence                   Low confidence               Increase in the intensity and      Increase in the intensity and      Increase in the intensity and
                                                                                             frequency of hot extremes:         frequency of hot extremes:         frequency of hot extremes:
                                                                                             Likely (compared with the recent   Very likely (compared with the     Virtually certain (compared
                                                                                             past (1995-2014))                  recent past (1995-2014))           with the recent past (1995-
                                                                                             Very likely (compared with pre-    Extremely likely (compared         2014))
                                                                                             industrial)                        with pre-industrial)               Virtually certain (compared
                                                                                                                                                                   with pre-industrial)
                                                                                             Decrease in the intensity and      Decrease in the intensity and
                                                                                             frequency of cold extremes:        frequency of cold extremes:        Decrease in the intensity and
                                                                                             Likely (compared with the recent   Very likely (compared with the     frequency of cold extremes:
                                                                                             past (1995-2014))                  recent past (1995-2014))           Virtually certain (compared
                                                                                             Very likely (compared with pre-    Extremely likely (compared         with the recent past (1995-
                                                                                             industrial).                       with pre-industrial)               2014))
                                                                                                                                                                   Virtually certain (compared
                                                                                                                                                                   with pre-industrial)
 South Eastern Africa (SEAF)   Increases in the intensity and   Evidence of a human          CMIP6 models project a robust      CMIP6 models project a robust      CMIP6 models project a robust
                               frequency of hot extremes        contribution to the          increase in the intensity and      increase in the intensity and      increase in the intensity and
                               and decreases in the intensity   observed increase in the     frequency of TXx events and a      frequency of TXx events and a      frequency of TXx events and a
                               and frequency of cold            intensity and frequency of   robust decrease in the intensity   robust decrease in the intensity   robust decrease in the intensity
                               extremes (Perkins-Kirkpatrick    hot extremes and decrease    and frequency of TNn events (Li    and frequency of TNn events        and frequency of TNn events
                               and Lewis, 2020;                 in the intensity and         et al., 2020; Annex). Median       (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                               Gebrechorkos et al., 2018;       frequency of cold            increase of more than 0.5°C in     Median increase of more than       Median increase of more than
                               Omondi et al., 2014; Chaney      extremes (Otto et al.,       the 50-year TXx and TNn events     1°C in the 50-year TXx and         2.5°C in the 50-year TXx and
                               et al., 2014)                    2015; Philip et al., 2020;   compared to the 1°C warming        TNn events compared to the         TNn events compared to the
                                                                Marthews et al., 2015;       level (Li et al., 2020) and more   1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                                Kew et al., 2021; Funk et    than 1.5°C in annual TXx and       2020) and more than 2°C in         2020) and more than 4°C in
                                                                al., 2015)                   TNn compared to pre-industrial     annual TXx and TNn compared        annual TXx and TNn compared
                                                                                             (Annex).                           to pre-industrial (Annex).         to pre-industrial (Annex).

                                                                                             Additional evidence from CMIP5     Additional evidence from           Additional evidence from
                                                                                             and CORDEX simulations for an      CMIP5 and CORDEX                   CMIP5/CMIP3 and CORDEX
                                                                                             increase in the intensity and      simulations for an increase in     simulations for an increase in
                                                                                             frequency of hot extremes          the intensity and frequency of     the intensity and frequency of
                                                                                             (Weber et al., 2018)               hot extremes (Weber et al.,        hot extremes (Coppola et al.,
                                                                                                                                2018; Coppola et al., 2021a)       2021a; Engelbrecht et al., 2015;
                                                                                                                                                                   Giorgi et al., 2014)
                               Medium confidence in the         Medium confidence in a       Increase in the intensity and      Increase in the intensity and      Increase in the intensity and
                               increase in the intensity and    human contribution to the    frequency of hot extremes:         frequency of hot extremes:         frequency of hot extremes:
                               frequency of hot extremes        observed increase in the     Likely (compared with the recent   Very likely (compared with the     Virtually certain (compared
                                                                intensity and frequency of   past (1995-2014))                  recent past (1995-2014))           with the recent past (1995-
                                                                hot extremes                 Very likely (compared with pre-    Extremely likely (compared         2014))
                                                                                             industrial)                        with pre-industrial)               Virtually certain (compared
                                                                                                                                                                   with pre-industrial)


Do Not Cite, Quote or Distribute                                 11-127                                                   Total pages: 345
                               Final Government Distribution                                      Chapter11                                                     IPCC AR6 WGI

                                                                                          Decrease in the intensity and      Decrease in the intensity and      Decrease in the intensity and
                                                                                          frequency of cold extremes:        frequency of cold extremes:        frequency of cold extremes:
                                                                                          Likely (compared with the recent   Very likely (compared with the     Virtually certain (compared
                                                                                          past (1995-2014))                  recent past (1995-2014))           with the recent past (1995-
                                                                                          Very likely (compared with pre-    Extremely likely (compared         2014))
                                                                                          industrial).                       with pre-industrial)               Virtually certain (compared
                                                                                                                                                                with pre-industrial)
 Western Southern Africa    Significant increases in the     Robust evidence of a         CMIP6 models project a robust      CMIP6 models project a robust      CMIP6 models project a robust
 (WSAF)                     intensity and frequency of hot   human contribution to the    increase in the intensity and      increase in the intensity and      increase in the intensity and
                            extremes and significant         observed increase in the     frequency of TXx events and a      frequency of TXx events and a      frequency of TXx events and a
                            decreases in the intensity and   intensity and frequency of   robust decrease in the intensity   robust decrease in the intensity   robust decrease in the intensity
                            frequency of cold extremes       hot extremes and decrease    and frequency of TNn events (Li    and frequency of TNn events        and frequency of TNn events
                            (Russo et al., 2016; Perkins-    in the intensity and         et al., 2020; Annex). Median       (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                            Kirkpatrick and Lewis, 2020;     frequency of cold            increase of more than 0.5°C in     Median increase of more than       Median increase of more than
                            Kruger and Nxumalo, 2017;        extremes (Seong et al.,      the 50-year TXx and TNn events     1°C in the 50-year TXx and         2.5°C in the 50-year TXx and
                            Mbokodo et al., 2020; Dunn       2020; Wang et al., 2017)     compared to the 1°C warming        TNn events compared to the         TNn events compared to the
                            et al., 2020)                                                 level (Li et al., 2020) and more   1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                                                          than 1.5°C in annual TXx and       2020) and more than 2°C in         2020) and more than 4.5°C in
                                                                                          TNn compared to pre-industrial     annual TXx and TNn compared        annual TXx and TNn compared
                                                                                          (Annex).                           to pre-industrial (Annex).         to pre-industrial (Annex).

                                                                                          Additional evidence from CMIP5     Additional evidence from           Additional evidence from
                                                                                          and CORDEX simulations for an      CMIP5 and CORDEX                   CMIP5/CMIP3 and CORDEX
                                                                                          increase in the intensity and      simulations for an increase in     simulations for an increase in
                                                                                          frequency of hot extremes          the intensity and frequency of     the intensity and frequency of
                                                                                          (Weber et al., 2018)               hot extremes (Weber et al.,        hot extremes (Coppola et al.,
                                                                                                                             2018; Coppola et al., 2021a)       2021a; Engelbrecht et al., 2015;
                                                                                                                                                                Giorgi et al., 2014)
                            Likely increase in the           Human influence likely       Increase in the intensity and      Increase in the intensity and      Increase in the intensity and
                            intensity and frequency of hot   contributed to the           frequency of hot extremes:         frequency of hot extremes:         frequency of hot extremes:
                            extremes and decrease in the     observed increase in the     Likely (compared with the recent   Very likely (compared with the     Virtually certain (compared
                            intensity and frequency of       intensity and frequency of   past (1995-2014))                  recent past (1995-2014))           with the recent past (1995-
                            cold extremes                    hot extremes and decrease    Very likely (compared with pre-    Extremely likely (compared         2014))
                                                             in the intensity and         industrial)                        with pre-industrial)               Virtually certain (compared
                                                             frequency of cold                                                                                  with pre-industrial)
                                                             extremes                     Decrease in the intensity and      Decrease in the intensity and
                                                                                          frequency of cold extremes:        frequency of cold extremes:        Decrease in the intensity and
                                                                                          Likely (compared with the recent   Very likely (compared with the     frequency of cold extremes:
                                                                                          past (1995-2014))                  recent past (1995-2014))           Virtually certain (compared
                                                                                          Very likely (compared with pre-    Extremely likely (compared         with the recent past (1995-
                                                                                          industrial).                       with pre-industrial)               2014))
                                                                                                                                                                Virtually certain (compared
                                                                                                                                                                with pre-industrial)
 Eastearn Southern Africa   Significant increases in the     Robust evidence of a         CMIP6 models project a robust      CMIP6 models project a robust      CMIP6 models project a robust
 (ESAF)                     intensity and frequency of hot   human contribution to the    increase in the intensity and      increase in the intensity and      increase in the intensity and
                            extremes and significant         observed increase in the     frequency of TXx events and a      frequency of TXx events and a      frequency of TXx events and a
                            decreases in the intensity and   intensity and frequency of   robust decrease in the intensity   robust decrease in the intensity   robust decrease in the intensity
Do Not Cite, Quote or Distribute                              11-128                                                    Total pages: 345
                             Final Government Distribution                                      Chapter11                                                     IPCC AR6 WGI

                          frequency of cold extremes       hot extremes and decrease    and frequency of TNn events (Li    and frequency of TNn events        and frequency of TNn events
                          (Dunn et al., 2020; Russo et     in the intensity and         et al., 2020; Annex). Median       (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                          al., 2016; Perkins-Kirkpatrick   frequency of cold            increase of more than 0.5°C in     Median increase of more than       Median increase of more than
                          and Lewis, 2020; Kruger and      extremes (Seong et al.,      the 50-year TXx and TNn events     0.5°C in the 50-year TXx and       2.5°C in the 50-year TXx and
                          Nxumalo, 2017; Mbokodo et        2020; Wang et al., 2017)     compared to the 1°C warming        TNn events compared to the         TNn events compared to the
                          al., 2020)                                                    level (Li et al., 2020) and more   1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                                                        than 1.5°C in annual TXx and       2020) and more than 2°C in         2020) and more than 4°C in
                                                                                        TNn compared to pre-industrial     annual TXx and TNn compared        annual TXx and TNn compared
                                                                                        (Annex).                           to pre-industrial (Annex).         to pre-industrial (Annex).

                                                                                        Additional evidence from CMIP5     Additional evidence from           Additional evidence from
                                                                                        and CORDEX simulations for an      CMIP5 and CORDEX                   CMIP5/CMIP3 and CORDEX
                                                                                        increase in the intensity and      simulations for an increase in     simulations for an increase in
                                                                                        frequency of hot extremes          the intensity and frequency of     the intensity and frequency of
                                                                                        (Weber et al., 2018)               hot extremes (Weber et al.,        hot extremes (Coppola et al.,
                                                                                                                           2018; Coppola et al., 2021a)       2021a; Engelbrecht et al., 2015;
                                                                                                                                                              Giorgi et al., 2014)
                          Likely increase in the           High confidence in a         Increase in the intensity and      Increase in the intensity and      Increase in the intensity and
                          intensity and frequency of hot   human contribution to the    frequency of hot extremes:         frequency of hot extremes:         frequency of hot extremes:
                          extremes and decrease in the     observed increase in the     Likely (compared with the recent   Very likely (compared with the     Virtually certain (compared
                          intensity and frequency of       intensity and frequency of   past (1995-2014))                  recent past (1995-2014))           with the recent past (1995-
                          cold extremes                    hot extremes and decrease    Very likely (compared with pre-    Extremely likely (compared         2014))
                                                           in the intensity and         industrial)                        with pre-industrial)               Virtually certain (compared
                                                           frequency of cold                                                                                  with pre-industrial)
                                                           extremes                     Decrease in the intensity and      Decrease in the intensity and
                                                                                        frequency of cold extremes:        frequency of cold extremes:        Decrease in the intensity and
                                                                                        Likely (compared with the recent   Very likely (compared with the     frequency of cold extremes:
                                                                                        past (1995-2014))                  recent past (1995-2014))           Virtually certain (compared
                                                                                        Very likely (compared with pre-    Extremely likely (compared         with the recent past (1995-
                                                                                        industrial).                       with pre-industrial)               2014))
                                                                                                                                                              Virtually certain (compared
                                                                                                                                                              with pre-industrial)
 Madagascar (MDG)         Increases in the intensity and   Limited evidence             CMIP6 models project a robust      CMIP6 models project a robust      CMIP6 models project a robust
                          frequency of hot extremes                                     increase in the intensity and      increase in the intensity and      increase in the intensity and
                          and decreases in the intensity                                frequency of TXx events and a      frequency of TXx events and a      frequency of TXx events and a
                          and frequency of cold                                         robust decrease in the intensity   robust decrease in the intensity   robust decrease in the intensity
                          extremes (Vincent et al.,                                     and frequency of TNn events (Li    and frequency of TNn events        and frequency of TNn events
                          2011; Donat et al., 2013)                                     et al., 2020; Annex). Median       (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                                                                                        increase of more than 0.5°C in     Median increase of more than       Median increase of more than
                                                                                        the 50-year TXx and TNn events     0.5°C in the 50-year TXx and       2°C in the 50-year TXx and
                                                                                        compared to the 1°C warming        TNn events compared to the         TNn events compared to the
                                                                                        level (Li et al., 2020) and more   1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                                                        than 1.5°C in annual TXx and       2020) and more than 2°C in         2020) and more than 3.5°C in
                                                                                        TNn compared to pre-industrial     annual TXx and TNn compared        annual TXx and TNn compared
                                                                                        (Annex).                           to pre-industrial (Annex).         to pre-industrial (Annex).




Do Not Cite, Quote or Distribute                            11-129                                                   Total pages: 345
                                  Final Government Distribution                                            Chapter11                                                     IPCC AR6 WGI

                                                                                                   Additional evidence from CMIP5       Additional evidence from         Additional evidence from
                                                                                                   and CORDEX simulations for an        CMIP5 and CORDEX                 CMIP5/CMIP3 and CORDEX
                                                                                                   increase in the intensity and        simulations for an increase in   simulations for an increase in
                                                                                                   frequency of hot extremes            the intensity and frequency of   the intensity and frequency of
                                                                                                   (Weber et al., 2018)                 hot extremes (Weber et al.,      hot extremes (Coppola et al.,
                                                                                                                                        2018; Coppola et al., 2021a)     2021a; Engelbrecht et al., 2015;
                                                                                                                                                                         Giorgi et al., 2014)
                               Medium confidence in the             Low confidence                 Increase in the intensity and        Increase in the intensity and    Increase in the intensity and
                               increase in the intensity and                                       frequency of hot extremes:           frequency of hot extremes:       frequency of hot extremes:
                               frequency of hot extremes                                           Likely (compared with the recent     Very likely (compared with the   Virtually certain (compared
                               and decrease in the intensity                                       past (1995-2014))                    recent past (1995-2014))         with the recent past (1995-
                               and frequency of cold                                               Very likely (compared with pre-      Extremely likely (compared       2014))
                               extremes                                                            industrial)                          with pre-industrial)             Virtually certain (compared
                                                                                                                                                                         with pre-industrial)
                                                                                                   Decrease in the intensity and        Decrease in the intensity and
                                                                                                   frequency of cold extremes:          frequency of cold extremes:      Decrease in the intensity and
                                                                                                   Likely (compared with the recent     Very likely (compared with the   frequency of cold extremes:
                                                                                                   past (1995-2014))                    recent past (1995-2014))         Virtually certain (compared
                                                                                                   Very likely (compared with pre-      Extremely likely (compared       with the recent past (1995-
                                                                                                   industrial).                         with pre-industrial)             2014))
                                                                                                                                                                         Virtually certain (compared
                                                                                                                                                                         with pre-industrial)
1
2   [END TABLE 11.4 HERE]
3
4   [START TABLE 11.5 HERE]
5
6   Table 11.5: Observed trends, human contribution to observed trends, and projected changes at 1.5°C, 2°C and 4°C of global warming for heavy precipitation in Africa, subdivided
7               by AR6 regions. See Sections 11.9.1 and 11.9.3 for details.
     Region                    Observed trends                        Detection and attribution;       Projections
                                                                      event attribution                1.5 °C                                     2 °C                                               4 °C
     All Africa                Insufficient data to assess trends     Limited evidence                 CMIP6 models project an increase in        CMIP6 models project a robust increase in the      CMIP6 models project a
                                                                                                       the intensity and frequency of heavy       intensity and frequency of heavy precipitation     robust increase in the
                                                                                                       precipitation (Li et al., 2020a).          (Li et al., 2020a). Median increase of more than   intensity and frequency of
                                                                                                       Median increase of more than 2% in         6% in the 50-year Rx1day and Rx5day events         heavy precipitation (Li et al.,
                                                                                                       the 50-year Rx1day and Rx5day              compared to the 1°C warming level (Li et al.,      2020a). Median increase of
                                                                                                       events compared to the 1°C warming         2020a)                                             more than 20% in the 50-
                                                                                                       level (Li et al., 2020a)                                                                      year Rx1day and Rx5day
                                                                                                                                                                                                     events compared to the 1°C
                                                                                                                                                                                                     warming level (Li et al.,
                                                                                                                                                                                                     2020a)
                               Low confidence                         Low confidence                   Intensification of heavy precipitation:    Intensification of heavy precipitation:            Intensification of heavy
                                                                                                       High confidence (compared with the         Likely (compared with the recent past (1995-       precipitation:
                                                                                                       recent past (1995-2014))                   2014))
                                                                                                       Likely (compared with pre-industrial)      Very likely (compared with pre-industrial)
    Do Not Cite, Quote or Distribute                                 11-130                                                      Total pages: 345
                                 Final Government Distribution                                        Chapter11                                                    IPCC AR6 WGI
                                                                                                                                                                                                 Extremely likely (compared
                                                                                                                                                                                                 with the recent past (1995-
                                                                                                                                                                                                 2014))
                                                                                                                                                                                                 Virtually certain (compared
                                                                                                                                                                                                 with pre-industrial)
                          3   Lack of agreement on the              Limited evidence (Añel et     CMIP6 models, CMIP5 models, and           CMIP6 models project a robust increase in the       CMIP6 models project a
    Mediterranean (MED)
                              evidence of trends (Sun et al.,       al., 2014; U.S. Department    RCMs project inconsistent changes in      intensity and frequency of heavy precipitation (Li robust increase in the intensity
                              2020; Casanueva et al., 2014; de      of Agriculture Economic       the region (Li et al., 2020; Cardell et   et al., 2020; Annex). Median increase of more       and frequency of heavy
                              Lima et al., 2015; Gajić-Čapka et     Research Service, 2016)       al., 2020; Zollo et al., 2016; Samuels    than 2% in the 50-year Rx1day and Rx5day            precipitation (Li et al., 2020;
                              al., 2015; Ribes et al., 2019;                                      et al., 2018)                             events compared to the 1°C warming level (Li et Annex). Median increase of
                              Peña-Angulo et al., 2020;                                                                                     al., 2020a) and more than 0% in annual Rx1day       more than 8% in the 50-year
                              Rajczak and Schär, 2017;                                                                                      and Rx5day and less than -2% in annual Rx30day Rx1day and Rx5day events
                              Jacob et al., 2018; Coppola et al.,                                                                           compared to pre-industrial (Annex).                 compared to the 1°C warming
                              2021a; Donat et al., 2014;                                                                                                                                        level (Li et al., 2020a) and
                              Mathbout et al., 2018; Dunn et                                                                                Additional evidence from CMIP5 and RCM              more than 2% in annual
                              al., 2020)                                                                                                    simulations for an increase in the intensity of     Rx1day and Rx5day and less
                                                                                                                                            heavy precipitation (Cardell et al., 2020; Zollo et than -2% in annual Rx30day
                                                                                                                                            al., 2016; Samuels et al., 2018)                    compared to pre-industrial
                                                                                                                                                                                                (Annex).

                                                                                                                                                                                                 Additional evidence from
                                                                                                                                                                                                 CMIP5 and RCM simulations
                                                                                                                                                                                                 for an increase in the intensity
                                                                                                                                                                                                 of heavy precipitation (Cardell
                                                                                                                                                                                                 et al., 2020; Tramblay and
                                                                                                                                                                                                 Somot, 2018; Zollo et al.,
                                                                                                                                                                                                 2016; Samuels et al., 2018;
                                                                                                                                                                                                 Monjo et al., 2016; Rajczak et
                                                                                                                                                                                                 al., 2013; Coppola et al.,
                                                                                                                                                                                                 2021b; Driouech et al., 2020)
                              Low confidence                        Low confidence                Intensification of heavy precipitation:   Intensification of heavy precipitation:              Intensification of heavy
                                                                                                  Low confidence (compared with the         Medium confidence (compared with the recent          precipitation:
                                                                                                  recent past (1995-2014))                  past (1995-2014))                                    High confidence (compared
                                                                                                  Medium confidence (compared with          High confidence (compared with pre-industrial)       with the recent past (1995-
                                                                                                  pre-industrial)                                                                                2014))
                                                                                                                                                                                                 High confidence (compared
                                                                                                                                                                                                 with pre-industrial)

    Sahara (SAH)              Insufficient data to assess trends    Limited evidence             CMIP6 models project an increase in the    CMIP6 models project a robust increase in the        CMIP6 models project a
                              (Sun et al., 2020; Dunn et al.,                                    intensity and frequency of heavy           intensity and frequency of heavy precipitation (Li   robust increase in the intensity
                              2020)                                                              precipitation (Li et al., 2020; Annex).    et al., 2020; Annex). Median increase of more        and frequency of heavy
                                                                                                 Median increase of more than 4% in the     than 8% in the 50-year Rx1day and Rx5day             precipitation (Li et al., 2020;
                                                                                                 50-year Rx1day and Rx5day events           events compared to the 1°C warming level (Li et      Annex). Median increase of
                                                                                                 compared to the 1°C warming level (Li      al., 2020a) and more than 20% in annual Rx1day,      more than 30% in the 50-year
                                                                                                 et al., 2020a) and more than 15% in                                                             Rx1day and Rx5day events


3
 This region includes both northern Africa and southern Europe
Do Not Cite, Quote or Distribute                       11-131                                                               Total pages: 345
                                  Final Government Distribution                                      Chapter11                                                   IPCC AR6 WGI

                                                                                                annual Rx1day, Rx5day, and Rx30day        Rx5day, and Rx30day compared to pre-industrial compared to the 1°C warming
                                                                                                compared to pre-industrial (Annex).       (Annex).                                       level (Li et al., 2020a) and
                                                                                                                                                                                         more than 40% in annual
                                                                                                                                                                                         Rx1day, Rx5day, and
                                                                                                                                                                                         Rx30day compared to pre-
                                                                                                                                                                                         industrial (Annex).
                               Low confidence                        Low confidence             Intensification of heavy precipitation:   Intensification of heavy precipitation:              Intensification of heavy
                                                                                                High confidence (compared with the        Likely (compared with the recent past (1995-         precipitation:
                                                                                                recent past (1995-2014))                  2014))                                               Extremely likely (compared
                                                                                                Likely (compared with pre-industrial)     Very likely (compared with pre-industrial)           with the recent past (1995-
                                                                                                                                                                                               2014))
                                                                                                                                                                                               Virtually certain (compared
                                                                                                                                                                                               with pre-industrial)
 Western Africa (WAF)          Insufficient data and a lack of       Limited evidence (Parker   CMIP6 models project an increase in the   CMIP6 models project a robust increase in the   CMIP6 models project a
                               agreement on the evidence of          et al., 2017)              intensity and frequency of heavy          intensity and frequency of heavy precipitation (Li
                                                                                                                                                                                          robust increase in the intensity
                               trends (Mouhamed et al.,                                         precipitation (Li et al., 2020; Annex).   et al., 2020; Annex). Median increase of more   and frequency of heavy
                               2013; Chaney et al., 2014;                                       Median increase of more than 4% in the    than 8% in the 50-year Rx1day and Rx5day        precipitation (Li et al., 2020;
                               Sanogo et al., 2015; Zittis,                                     50-year Rx1day and Rx5day events          events compared to the 1°C warming level (Li et Annex). Median increase of
                               2017; Barry et al., 2018; Sun et                                 compared to the 1°C warming level (Li     al., 2020a) and more than 15% in annual Rx1day  more than 25% in the 50-year
                               al., 2020; Dunn et al., 2020)                                    et al., 2020a) and more than 10% in       and Rx5day and 10% in annual Rx30day            Rx1day and Rx5day events
                                                                                                annual Rx1day and Rx5day and 8% in        compared to pre-industrial (Annex).             compared to the 1°C warming
                                                                                                annual Rx30day compared to pre-                                                           level (Li et al., 2020a) and
                                                                                                industrial (Annex).                   Additional evidence from CMIP5 and CORDEX more than 30% in annual
                                                                                                                                      simulations for an increase in the intensity of     Rx1day and Rx5day and 15%
                                                                                                Additional evidence from CMIP5 and    heavy precipitation (Nikulin et al., 2018; Déqué et in annual Rx30day compared
                                                                                                CORDEX simulations for an increase in al., 2017)                                          to pre-industrial (Annex).
                                                                                                the intensity of heavy precipitation
                                                                                                (Nikulin et al., 2018)                                                                    Additional evidence from
                                                                                                                                                                                          CMIP5 and CORDEX
                                                                                                                                                                                          simulations for an increase in
                                                                                                                                                                                          the intensity of heavy
                                                                                                                                                                                          precipitation (Giorgi et al.,
                                                                                                                                                                                          2014; Dosio et al., 2019;
                                                                                                                                                                                          Akinsanola and Zhou, 2018;
                                                                                                                                                                                          Coppola et al., 2021b)
                               Low confidence                        Low confidence             Intensification of heavy precipitation:   Intensification of heavy precipitation:              Intensification of heavy
                                                                                                High confidence (compared with the        Likely (compared with the recent past (1995-         precipitation:
                                                                                                recent past (1995-2014))                  2014))                                               Extremely likely (compared
                                                                                                Likely (compared with pre-industrial)     Very likely (compared with pre-industrial)           with the recent past (1995-
                                                                                                                                                                                               2014))
                                                                                                                                                                                               Virtually certain (compared
                                                                                                                                                                                               with pre-industrial)
 North Eastern Africa (NEAF)   Insufficient data to assess trends    Limited evidence           CMIP6 models project an increase in the   CMIP6 models project a robust increase in the        CMIP6 models project a
                               (Sun et al., 2020; Dunn et al.,                                  intensity and frequency of heavy          intensity and frequency of heavy precipitation (Li   robust increase in the intensity
                               2020)                                                            precipitation (Li et al., 2020; Annex).   et al., 2020; Annex). Median increase of more        and frequency of heavy
                                                                                                Median increase of more than 4% in the    than 8% in the 50-year Rx1day and Rx5day             precipitation (Li et al., 2020;

Do Not Cite, Quote or Distribute                                    11-132                                                 Total pages: 345
                             Final Government Distribution                                       Chapter11                                                   IPCC AR6 WGI

                                                                                            50-year Rx1day and Rx5day events          events compared to the 1°C warming level (Li et      Annex). Median increase of
                                                                                            compared to the 1°C warming level (Li     al., 2020a) and more than 10% in annual Rx1day,      more than 25% in the 50-year
                                                                                            et al., 2020a) and more than 8% in        Rx5day, and Rx30day compared to pre-industrial       Rx1day and Rx5day events
                                                                                            annual Rx1day and Rx5day and 6% in        (Annex).                                             compared to the 1°C warming
                                                                                            annual Rx30day compared to pre-                                                                level (Li et al., 2020a) and
                                                                                            industrial (Annex).                                                                            more than 35% in annual
                                                                                                                                                                                           Rx1day and Rx5day and 30%
                                                                                                                                                                                           in annual Rx30day compared
                                                                                                                                                                                           to pre-industrial (Annex).
                          Low confidence                        Low confidence              Intensification of heavy precipitation:   Intensification of heavy precipitation:              Intensification of heavy
                                                                                            High confidence (compared with the        Likely (compared with the recent past (1995-         precipitation:
                                                                                            recent past (1995-2014))                  2014))                                               Extremely likely (compared
                                                                                            Likely (compared with pre-industrial)     Very likely (compared with pre-industrial)           with the recent past (1995-
                                                                                                                                                                                           2014))
                                                                                                                                                                                           Virtually certain (compared
                                                                                                                                                                                           with pre-industrial)
 Central Africa (CAF)     Insufficient data to assess trends    Limited evidence (Otto et   CMIP6 models project an increase in the   CMIP6 models project a robust increase in the   CMIP6 models project a
                          (Sun et al., 2020; Dunn et al.,       al., 2013)                  intensity and frequency of heavy          intensity and frequency of heavy precipitation (Li
                                                                                                                                                                                      robust increase in the intensity
                          2020)                                                             precipitation (Li et al., 2020; Annex).   et al., 2020; Annex). Median increase of more   and frequency of heavy
                                                                                            Median increase of more than 2% in the    than 6% in the 50-year Rx1day and Rx5day        precipitation (Li et al., 2020;
                                                                                            50-year Rx1day and Rx5day events          events compared to the 1°C warming level (Li et Annex). Median increase of
                                                                                            compared to the 1°C warming level (Li     al., 2020a) and more than 10% in annual Rx1day, more than 20% in the 50-year
                                                                                            et al., 2020a) and more than 10% in       Rx5day, and Rx30day compared to pre-industrial  Rx1day and Rx5day events
                                                                                            annual Rx1day and Rx5day and 8% in        (Annex).                                        compared to the 1°C warming
                                                                                            annual Rx30day compared to pre-                                                           level (Li et al., 2020a) and
                                                                                            industrial (Annex).                   Additional evidence from CMIP5 and CORDEX more than 30% in annual
                                                                                                                                  simulations for an increase in the intensity of     Rx1day and Rx5day and 20%
                                                                                            Additional evidence from CMIP5 and    heavy precipitation (Nikulin et al., 2018; Déqué et in annual Rx30day compared
                                                                                            CORDEX simulations for an increase in al., 2017; Coppola et al., 2021b)                   to pre-industrial (Annex).
                                                                                            the intensity of heavy precipitation
                                                                                            (Nikulin et al., 2018)                                                                    Additional evidence from
                                                                                                                                                                                      CMIP5 and CORDEX
                                                                                                                                                                                      simulations for an increase in
                                                                                                                                                                                      the intensity of heavy
                                                                                                                                                                                      precipitation (Diedhiou et al.
                                                                                                                                                                                      2018; Fotso-Nguemo et al.
                                                                                                                                                                                      2018; Sonkoué et al. 2019;
                                                                                                                                                                                      Coppola et al., 2021b)
                          Low confidence                        Low confidence              Intensification of heavy precipitation:   Intensification of heavy precipitation:              Intensification of heavy
                                                                                            High confidence (compared with the        Likely (compared with the recent past (1995-         precipitation:
                                                                                            recent past (1995-2014))                  2014))                                               Extremely likely (compared
                                                                                            Likely (compared with pre-industrial)     Very likely (compared with pre-industrial)           with the recent past (1995-
                                                                                                                                                                                           2014))
                                                                                                                                                                                           Virtually certain (compared
                                                                                                                                                                                           with pre-industrial)



Do Not Cite, Quote or Distribute                               11-133                                                  Total pages: 345
                                  Final Government Distribution                              Chapter11                                                   IPCC AR6 WGI

 South Eastern Africa (SEAF)   Insufficient data to assess trends    Limited evidence   CMIP6 models project an increase in the   CMIP6 models project a robust increase in the        CMIP6 models project a
                               (Sun et al., 2020; Dunn et al.,                          intensity and frequency of heavy          intensity and frequency of heavy precipitation (Li   robust increase in the intensity
                               2020)                                                    precipitation (Li et al., 2020; Annex).   et al., 2020; Annex). Median increase of more        and frequency of heavy
                                                                                        Median increase of more than 2% in the    than 4% in the 50-year Rx1day and Rx5day             precipitation (Li et al., 2020;
                                                                                        50-year Rx1day and Rx5day events          events compared to the 1°C warming level (Li et      Annex). Median increase of
                                                                                        compared to the 1°C warming level (Li     al., 2020a) and more than 8% in annual Rx1day        more than 15% in the 50-year
                                                                                        et al., 2020a) and more than 6% in        and Rx5day and 6% in annual Rx30day compared         Rx1day and Rx5day events
                                                                                        annual Rx1day and Rx5day and 4% in        to pre-industrial (Annex).                           compared to the 1°C warming
                                                                                        annual Rx30day compared to pre-                                                                level (Li et al., 2020a) and
                                                                                        industrial (Annex).                                                                            more than 25% in annual
                                                                                                                                                                                       Rx1day and Rx5day and 15%
                                                                                                                                                                                       in annual Rx30day compared
                                                                                                                                                                                       to pre-industrial (Annex).
                               Low confidence                        Low confidence     Intensification of heavy precipitation:   Intensification of heavy precipitation:              Intensification of heavy
                                                                                        High confidence (compared with the        Likely (compared with the recent past (1995-         precipitation:
                                                                                        recent past (1995-2014))                  2014))                                               Extremely likely (compared
                                                                                        Likely (compared with pre-industrial)     Very likely (compared with pre-industrial)           with the recent past (1995-
                                                                                                                                                                                       2014))
                                                                                                                                                                                       Virtually certain (compared
                                                                                                                                                                                       with pre-industrial)
 West Southern Africa          Intensification of heavy              Limited evidence   CMIP6 models project inconsistent         CMIP6 models project inconsistent changes in the CMIP6 models project an
 (WSAF)                        precipitation (Sun et al., 2020;                         changes in the region (Li et al., 2020,   region (Li et al., 2020, Annex)                  increase in the intensity and
                               Donat et al., 2013)                                      Annex)                                                                                     frequency of heavy
                                                                                                                                                                                   precipitation (Li et al., 2020;
                                                                                                                                                                                   Annex). Median increase of
                                                                                                                                                                                   more than 10% in the 50-year
                                                                                                                                                                                   Rx1day and Rx5day events
                                                                                                                                                                                   compared to the 1°C warming
                                                                                                                                                                                   level (Li et al., 2020a) and
                                                                                                                                                                                   more than 4% in annual
                                                                                                                                                                                   Rx1day and Rx5day and 0% in
                                                                                                                                                                                   annual Rx30day compared to
                                                                                                                                                                                   pre-industrial (Annex).

                                                                                                                                                                                       Additional evidence from
                                                                                                                                                                                       CMIP5 and RCM simulations
                                                                                                                                                                                       for an increase in the intensity
                                                                                                                                                                                       of heavy precipitation (Pinto et
                                                                                                                                                                                       al., 2016; Dosio et al., 2019;
                                                                                                                                                                                       Giorgi et al., 2014; Coppola et
                                                                                                                                                                                       al., 2021b)
                               Medium confidence in the              Low confidence     Intensification of heavy precipitation:   Intensification of heavy precipitation:              Intensification of heavy
                               intensification of heavy                                 Low confidence (compared with the         Low confidence (compared with the recent past        precipitation:
                               precipitation.                                           recent past (1995-2014))                  (1995-2014))                                         High confidence (compared
                                                                                        Low confidence (compared with pre-        Medium confidence (compared with pre-                with the recent past (1995-
                                                                                        industrial)                               industrial)                                          2014))


Do Not Cite, Quote or Distribute                                    11-134                                          Total pages: 345
                                  Final Government Distribution                              Chapter11                                                     IPCC AR6 WGI

                                                                                                                                                                                       Likely (compared with pre-
                                                                                                                                                                                       industrial)
 East Southern Africa (ESAF)   Intensification of heavy              Limited evidence   CMIP6 models project an increase in the    CMIP6 models project an increase in the intensity   CMIP6 models project a
                               precipitation (Sun et al., 2020;                         intensity and frequency of heavy           and frequency of heavy precipitation (Li et al.,    robust increase in the intensity
                               Donat et al., 2013)                                      precipitation (Li et al., 2020; Annex).    2020; Annex). Median increase of more than 2%       and frequency of heavy
                                                                                        Median increase of more than 2% in the     in the 50-year Rx1day and Rx5day events             precipitation (Li et al., 2020;
                                                                                        50-year Rx1day and Rx5day events           compared to the 1°C warming level (Li et al.,       Annex). Median increase of
                                                                                        compared to the 1°C warming level (Li      2020a) and more than 6% in annual Rx1day and        more than 15% in the 50-year
                                                                                        et al., 2020a) and more than 4% in         Rx5day and 2% in annual Rx30day compared to         Rx1day and Rx5day events
                                                                                        annual Rx1day and Rx5day and 0% in         pre-industrial (Annex).                             compared to the 1°C warming
                                                                                        annual Rx30day compared to pre-                                                                level (Li et al., 2020a) and
                                                                                        industrial (Annex).                                                                            more than 15% in annual
                                                                                                                                                                                       Rx1day and Rx5day and 8% in
                                                                                                                                                                                       annual Rx30day compared to
                                                                                                                                                                                       pre-industrial (Annex).

                                                                                                                                                                                       Additional evidence from
                                                                                                                                                                                       CMIP5 and RCM simulations
                                                                                                                                                                                       for an increase in the intensity
                                                                                                                                                                                       of heavy precipitation (Pinto et
                                                                                                                                                                                       al., 2016; Dosio et al., 2019;
                                                                                                                                                                                       Giorgi et al., 2014; Coppola et
                                                                                                                                                                                       al., 2021b)
                               Medium confidence in the              Low confidence     Intensification of heavy precipitation:    Intensification of heavy precipitation:             Intensification of heavy
                               intensification of heavy                                 Medium confidence (compared with the       High confidence (compared with the recent past      precipitation:
                               precipitation.                                           recent past (1995-2014))                   (1995-2014))                                        Very likely (compared with the
                                                                                        High confidence (compared with pre-        Likely (compared with pre-industrial)               recent past (1995-2014))
                                                                                        industrial)                                                                                    Extremely likely (compared
                                                                                                                                                                                       with pre-industrial)




 Madagascar (MDG)              Insufficient data to assess trends    Limited evidence    CMIP6 models project an increase in        CMIP6 models project an increase in the             CMIP6 models project a
                               and trends in available data are                          the intensity and frequency of heavy       intensity and frequency of heavy precipitation      robust increase in the
                               not significant (Sun et al., 2020;                        precipitation (Li et al., 2020; Annex).    (Li et al., 2020; Annex). Median increase of        intensity and frequency of
                               Dunn et al., 2020; Donat et al.,                          Median increase of more than 2% in         more than 2% in the 50-year Rx1day and              heavy precipitation (Li et al.,
                               2013; Vincent et al., 2011)                               the 50-year Rx1day and Rx5day              Rx5day events compared to the 1°C warming           2020; Annex). Median
                                                                                         events compared to the 1°C warming         level (Li et al., 2020a) and more than 4% in        increase of more than 15%
                                                                                         level (Li et al., 2020a) and more than     annual Rx1day and Rx5day and 2% in annual           in the 50-year Rx1day and
                                                                                         4% in annual Rx1day and Rx5day and         Rx30day compared to pre-industrial (Annex).         Rx5day events compared to
                                                                                         0% in annual Rx30day compared to                                                               the 1°C warming level (Li et
                                                                                         pre-industrial (Annex).                    Additional evidence from CMIP5 and                  al., 2020a) and more than
                                                                                                                                    CORDEX simulations for an increase in the           15% in annual Rx1day and
                                                                                         Additional evidence from CMIP5 and         intensity of heavy precipitation (Weber et al.,     Rx5day and 6% in annual
                                                                                         CORDEX simulations for an increase         2018)                                               Rx30day compared to pre-
                                                                                         in the intensity of heavy precipitation                                                        industrial (Annex).
                                                                                         (Weber et al., 2018)
Do Not Cite, Quote or Distribute                                    11-135                                         Total pages: 345
                                           Final Government Distribution                                    Chapter11                                                     IPCC AR6 WGI
                                                                                                                                                                                                        Additional evidence from
                                                                                                                                                                                                        CMIP5 and CORDEX
                                                                                                                                                                                                        simulations for an increase
                                                                                                                                                                                                        in the intensity of heavy
                                                                                                                                                                                                        precipitation (Weber et al.,
                                                                                                                                                                                                        2018)
                                        Low confidence                     Low confidence               Intensification of heavy precipitation:    Intensification of heavy precipitation:              Intensification of heavy
                                                                                                        Medium confidence (compared with           High confidence (compared with the recent past       precipitation:
                                                                                                        the recent past (1995-2014))               (1995-2014))                                         Very likely (compared with
                                                                                                        High confidence (compared with pre-        Likely (compared with pre-industrial)                the recent past (1995-2014))
                                                                                                        industrial)                                                                                     Extremely likely (compared
                                                                                                                                                                                                        with pre-industrial)

1
2   [END TABLE 11.5 HERE]
3
4
5   [START TABLE 11.6 HERE]
6
7   Table 11.6: Observed trends, human contribution to observed trends, and projected changes at 1.5°C, 2°C and 4°C of global warming for meteorological droughts (MET),
8               agricultural and ecological droughts (AGR/ECOL), and hydrological droughts (HYDR) in Africa, subdivided by AR6 regions. See Sections 11.9.1 and 11.9.4 for
9               details.
         Region and drought                                                                                                                            Projections
                                       Observed trends              Human contribution
                type                                                                                     1.5 °C                                            2 °C                                          4 °C
            4        MET         ENTRY IDENTICAL TO            ENTRY IDENTICAL TO           ENTRY IDENTICAL TO EU-MED                       ENTRY IDENTICAL TO EU-MED                       ENTRY IDENTICAL TO EU-MED
        MED
                                 EU-MED                        EU-MED
                     AGR         ENTRY IDENTICAL TO            ENTRY IDENTICAL TO           ENTRY IDENTICAL TO EU-MED                       ENTRY IDENTICAL TO EU-MED                       ENTRY IDENTICAL TO EU-MED
                     ECOL        EU-MED                        EU-MED
                     HYDR        ENTRY IDENTICAL TO            ENTRY IDENTICAL TO           ENTRY IDENTICAL TO EU-MED                       ENTRY IDENTICAL TO EU-MED                       ENTRY IDENTICAL TO EU-MED
                                 EU-MED                        EU-MED
        Sahara       MET         Low confidence: Limited       Low confidence: Limited      Low confidence: Mixed signals (seasonally        Low confidence: Mixed signals (seasonally      Low confidence: Mixed signals
        (SAH)                    evidence.                     evidence                     and geographically varying) and non-robust      and geographically varying) and non-robust      (seasonally and geographically varying)
                                                                                            changes (Cook et al., 2020). Slightly reduced   changes (Cook et al., 2020). Slightly reduced   and non-robust changes (Cook et al.,
                                                                                            drying based on CDD (Chapter 11                 drying based on CDD (Chapter 11                 2020). Reduced drying based on CDD
                                                                                            Supplementary Material (11.SM)).                Supplementary Material (11.SM)).                (Chapter 11 Supplementary Material
                                                                                                                                                                                            (11.SM)).
                     AGR         Low confidence; Limited       Low confidence; Limited      Low confidence: Limited evidence and            Low confidence: Limited evidence and            Low confidence: Limited evidence and
                     ECOL        evidence.                     evidence.                    inconsistent signals in CMIP6 (Chapter 11       inconsistent signals in CMIP6 (Chapter 11       inconsistent signals in CMIP6. (Cook et
                                                                                            Supplementary Material (11.SM)).                Supplementary Material (11.SM)).                al., 2020; Vicente-Serrano et al.,


    4
        This region includes both northern Africa and southern Europe
    Do Not Cite, Quote or Distribute                                      11-136                                                  Total pages: 345
                              Final Government Distribution                                       Chapter11                                                     IPCC AR6 WGI

                                                                                                                                                                                  2020a)(Chapter 11 Supplementary
                                                                                                                                                                                  Material (11.SM))

           HYDR     Low confidence: Limited        Low confidence: Limited        Low confidence: Limited evidence. One           Low confidence: Limited evidence. One           Low confidence: Inconsistent trends
                    evidence                       evidence                       study shows lack of signal (Touma et al.,       study shows lack of signal (Touma et al.,       (Touma et al., 2015; Cook et al., 2020)
                                                                                  2015)                                           2015)
 Western   MET      Medium confidence:             Low confidence: Mixed          Low confidence: Mixed signal. Mean              Low confidence: Mixed signal. Mean              Medium confidence: Increase in
 Africa             Increased drying based on      signals (Lawal et al., 2016;   increase of CDD over larger part of Guinea      increase of CDD over larger part of Guinea      meteorological droughts, mostly on
 (WAF)              CDD and SPI (Chaney et         Knutson and Zeng, 2018).       Coast in 25 CORDEX AFR runs, 1.5°C              Coast in 25 CORDEX AFR runs, 1.5°C              seasonal time scale. Seasonal CDD
                    al., 2014; Barry et al.,                                      minus 1971-2000 (Klutse et al., 2016); slight   minus 1971-2000 (Klutse et al., 2016); slight   increases in the region for MAM and JJA
                    2018; Spinoni et al.,          Drying attributable in
                    2019; Dunn et al., 2020)       fraction of region to climate  increase in SPI-based meteorological            increase in SPI-based meteorological            (Dosio et al., 2019), increase in SPI-based
                                                   change over 1901-2010 and      drought frequency and magntidue in the          drought frequency and magntidue in the          meteorological drought frequency and
                                                   1951-2010 time frames, but     Niger and Volta river basin in CORDEX           Niger and Volta river basin in CORDEX           magnitude in Niger and Volta river basins
                                                   trend reversal from 1981-      simulations (Oguntunde et al., 2020); but       simulations (Oguntunde et al., 2020); but       (Oguntunde et al., 2020); and slight
                                                   2010 (Knutson and Zeng,        inconsistent changes in CDD in CMIP6            inconsistent changes in CDD in CMIP6            increase in SPI-based meteorological
                                                   2018)                          GCMs (Diedhiou et al., 2018)(Chapter 11         GCMs (Diedhiou et al., 2018)(Chapter 11         drought for overall region (Spinoni et al.,
                                                                                  Supplementary Material (11.SM)), as well as     Supplementary Material (11.SM)), as well as     2020). Mixed signal in annual CDD
                                                   No evidence that late onset of
                                                   2015 wet season in Nigeria     in mean precipitation in CMIP6 GCMs             in mean precipitation in CMIP6 GCMs             (Akinsanola and Zhou, 2018; Dosio et al.,
                                                   was due to human               (Cook et al., 2020)                             (Cook et al., 2020)                             2019)(Chapter 11 Supplementary Material
                                                   contribution (Lawal et al.,                                                                                                    (11.SM)).
                                                   2016)
           AGR      Medium confidence:             Low confidence: Limited        Low confidence: Inconsistent signals            Low confidence: Inconsistent signals            Low confidence: Mixed signal.
           ECOL     Increased drying based on      evidence                       (geographical and inter-model variations) in    (geographical and inter-model variations) in    Inconsistent changes depending on
                    water-balance estimates                                       soil moisture and SPEI-PM (Naumann et al.,      soil moisture and SPEI-PM (Naumann et al.,      subregion, indices, and season (Naumann
                    and SPEI-PM, with
                                                                                  2018; Xu et al., 2019a)(Chapter 11              2018; Xu et al., 2019a; Cook et al.,            et al., 2018; Cook et al., 2020; Vicente-
                    stronger signals for SPEI-
                    PM (Greve et al., 2014;                                       Supplementary Material (11.SM))                 2020)(Chapter 11 Supplementary Material         Serrano et al., 2020a)(Chapter 11
                    Spinoni et al., 2019;                                                                                         (11.SM))                                        Supplementary Material (11.SM)). Most
                    Padrón et al., 2020)                                                                                                                                          projections show a drying in Western half
                                                                                                                                                                                  of domain.
           HYDR     Medium confidence:             Low confidence:                Low confidence: Limited evidence. One           Low confidence: Inconsistent signal             Low confidence: Inconsistent
                    Decrease in streamflow         LimitedLimited evidence        study shows lack of signal (Touma et al.,       (Touma et al., 2015; Cook et al., 2020)         projections and/or non-robust changes
                    (Dai and Zhao, 2017;                                          2015)                                                                                           (Giuntoli et al., 2015; Touma et al., 2015;
                    Tramblay et al., 2020).
                                                                                                                                                                                  Cook et al., 2020)
 North     MET      Low confidence: Mixed          Low confidence: Limited        Low confidence: Inconsistent trends.            Low confidence:Inconsistent trends.             Medium confidence: Decrease in
 Eastern            signals. Increasing drought    evidence on attribution of     Inconsistent and weak signals in SPI            Inconsistent changes in CDD (Chapter 11         meteorological drought (Sillmann et al.,
 Africa             in part of the region, in      long-term trends.              (Nguvava et al., 2019; Xu et al., 2019a),       Supplementary Material (11.SM)) and SPI         2013b; Dosio et al., 2019; Cook et al.,
 (NEAF)             particular in recent two
                                                                                  with high spatial variation (Nguvava et al.,    (Nguvava et al., 2019; Xu et al., 2019a); but   2020; Spinoni et al., 2020)
                    decades; but decreasing        Robust evidence that recent
                    drought trends in other part                                  2019); inconsistent signals in CDD in           tendency towards increase in mean
                                                   meteorological drought
                    of domain (NOTE: wetting       events (in 2016 and 2017)      CMIP6 (Chapter 11 Supplementary Material        precipitation (Cook et al., 2020).
                    trend in Horn of Africa in     are not attributable to        (11.SM)).                                                                                       Sillmann et al. (2013), (2081-2100)/1981-
                    Spinoni et al. 2019)(Funk      anthropogenic climate                                                                                                          2000, rcp8.5, CMIP3-CMIP5

Do Not Cite, Quote or Distribute                              11-137                                                    Total pages: 345
                              Final Government Distribution                                         Chapter11                                                     IPCC AR6 WGI

                    et al., 2015a; Nicholson,      change (Lott et al., 2013;       Nguvava et al. (2019): projections at 1.5°C     Nguvava et al. (2019): projections at 2°C      Decrease of CDD
                    2017; Spinoni et al., 2019)    Marthews et al., 2015; Uhe et    GWL in Cordex AFR data, compared to             GWL in Cordex AFR data, compared to
                    “No trends in observations     al., 2017; Funk et al., 2018b;   1971-2000: non significant changes in SPI-      1971-2000: non significant changes in SPI-     Dosio et al. (2019), (2070-2099/1981-
                    in Ethiopia”; “large           Otto et al., 2018a; Philip et    12-based meteorological drought frequency       12-based meteorological drought frequency
                                                                                                                                                                                   2010), rcp 8.5, 23 RCM: Decrease in CDD
                    variability” (Philip et al.,   al., 2018b; Kew et al., 2021)    and intensity.                                  and intensity.
                    2018b)

           AGR      Low confidence:                Low confidence: Limited          Low confidence: Inconsistent trends             Low confidence: Inconsistent trends, but       Medium confidence: Decrease in soil
           ECOL     Inconsistent trends            evidence because of lack of      (Naumann et al., 2018; Xu et al.,               tendency to wetting (Naumann et al., 2018;     moisture-based drought (Cook et al., 2020;
                    (Greve et al., 2014; Dai       studies                          2019a)(Chapter 11 Supplementary Material        Xu et al., 2019a; Cook et al., 2020)(Chapter   Vicente-Serrano et al., 2020a)(Chapter 11
                    and Zhao, 2017; Spinoni                                         (11.SM))                                        11 Supplementary Material (11.SM))             Supplementary Material (11.SM))
                    et al., 2019; Padrón et
                    al., 2020)

           HYDR     Low confidence: Limited        Low confidence: Limited          Low confidence: Limited evidence. One           Low confidence: Limited evidence due to       Medium confidence: Decrease in
                    evidence                       evidence on attribution of       study shows lack of signal (Touma et al.,       lack of studies;inconsistent trends (Touma et hydrological drought compared to pre-
                                                   long-term trends (Fenta et       2015)                                           al., 2015; Cook et al., 2020)                 industrial conditions and recent past
                                                   al., 2017)
                                                                                                                                                                                  (Giuntoli et al., 2015; Cook et al., 2020)
                                                                                                                                                                                  but some inconsistent signals (Touma et
                                                                                                                                                                                  al., 2015)
 Central   MET      Medium confidence              Low confidence:                  Low confidence; Mixed signal. Drying            Low confidence; Mixed signal. Robust           Low confidence: Mixed signal,
 Africa             Decrease in SPI (Spinoni       Inconsistent signal in           tendency (increasing CDD) in CORDEX             drying tendency (increasing CDD) in            depending on multi-model experiment and
 (CAF)              et al., 2019) and mean         observations vs models for       AFR simulations compared to 1971-2000           CORDEX AFR simulations compared to             considered index (Fotso-Nguemo et al.,
                    rainfall. (Aguilar et al.,     1951-2010 trends (Knutson
                                                                                    (Mba et al., 2018); but tendency towards less   1971-2000 (Mba et al., 2018); but              2018; Dosio et al., 2019; Sonkoué et
                    2009; Hua et al., 2016; Dai    and Zeng, 2018); no signal in
                    and Zhao, 2017)                single-model based study         drying (CDD decrease) in CMIP6 GCMs             inconsistent signal in CMIP6 GCMs (with        al., 2019; Spinoni et al., 2020)(Chapter
                                                   (Otto et al., 2013)              (Chapter 11 Supplementary Material              tendency towards CDD decrease (Chapter 11      11 Supplementary Material (11.SM)).
                                                                                    (11.SM)), consistent with increase in           Supplementary Material (11.SM));
                                                                                    precipitation at higher warming levels (Cook    consistent with projected increase in mean     Increase in mean precipitation in CMIP6
                                                                                    et al., 2020). Inconsistent signals in SPI in   precipitation (Cook et al., 2020));            GCMs (Cook et al., 2020). Increase in
                                                                                    CMIP5 GCMs (Xu et al., 2019a)                   inconsistent signals in CDD in CMIP5           CDD (increase in meteorological drought)
                                                                                                                                    GCMs (Sonkoué et al., 2019). Decrease          in CORDEX AFR simulations (Dosio et
                                                                                                                                    frequency of SPI-based droughts in CMIP5       al., 2019; Fotso-Nguemo et al., 2019) but
                                                                                                                                    (Xu et al., 2019a).                            inconsistent CDD signals in CMIP6 (with
                                                                                                                                                                                   tendency towards CDD decrease; Chapter
                                                                                                                                                                                   11 Supplementary Material (11.SM)) and
                                                                                                                                                                                   CMIP5 GCMs (Sonkoué et al., 2019).
                                                                                                                                                                                   Increase in SPI (less drying) in CMIP5
                                                                                                                                                                                   GCMs (Spinoni et al., 2020).
           AGR      Medium confidence              Low confidence: Limited          Low confidence: Inconsistent signals.           Low confidence: Inconsistent signals.          Low confidence. Inconsistent signals.
           ECOL     Decrease in water-balance      evidence due to lack of          Slight tendency towards soil moisture           Inconsistent trends in duration vs             Tendency towards wetting in CMIP6 soil
                    availability or SPEI, but      studies                          wetting in CMIP5 (Xu et al., 2019a) and         frequency of soil moisture-based               moisture (Cook et al., 2020)(Chapter 11
                    some regional variability
                                                                                    CMIP6 (Chapter 11 Supplementary Material        drought events in CMIP5 (Xu et al., 2019a);    Supplementary Material (11.SM));
                    and index dependency of
                    trends (Greve et al., 2014;                                                                                     slight mean soil moisture wetting in CMIP6

Do Not Cite, Quote or Distribute                              11-138                                                     Total pages: 345
                               Final Government Distribution                                       Chapter11                                                    IPCC AR6 WGI

                    Dai and Zhao, 2017;                                             (11.SM)); and slight increase (less drying in (Chapter 11 Supplementary Material            inconsistent signals in SPEI-PM (Vicente-
                    Spinoni et al., 2019;                                           SPEI-PM (Naumann et al., 2018)                (11.SM)); slight wetting of SPEI-PM based     Serrano et al., 2020a)
                    Padrón et al., 2020)                                                                                          events (Naumann et al., 2018).


           HYDR     Low confidence: Limited         Low confidence: Limited         Low confidence: Limited evidence. One        Low confidence: Limited evidence and           Low confidence: Inconsistent projections
                    evidence. Decrease in           evidence                        study shows lack of signal (Touma et al.,    inconsistent trends in mean runoff in two      and/or non-robust changes (Giuntoli et al.,
                    streamflow from 1950-                                           2015)                                        studies (Touma et al., 2015; Cook et al.,      2015; Touma et al., 2015; Cook et al.,
                    2012 in southern part of
                                                                                                                                 2020)                                          2020)
                    domain (Dai and Zhao,
                    2017)
 South     MET      Low confidence:                 Low confidence: Limited         Low confidence: Inconsistent changes         Low confidence: Inconsistent changes           Low confidence: Inconsistent trends
 Eastern            Inconsistent trends in          evidence on attribution of      (Osima et al., 2018; Xu et al.,              (Osima et al., 2018; Xu et al.,                between studies and subregions (Sillmann
 Africa             SPI (Spinoni et al., 2019)      long-term trends.               2019a)(Chapter 11 Supplementary Material     2019a)(Chapter 11 Supplementary Material       et al., 2013b; Dosio et al., 2019; Vicente-
 (SEAF)             but occurrence of strong
                                                                                    (11.SM)) and lack of signal (Nangombe et     (11.SM)) and lack of signal (Nangombe et       Serrano et al., 2020a)(Chapter 11
                    drought events in recent        Robust evidence that recent
                    years (Funk et al., 2015a;                                      al., 2018)                                   al., 2018)                                     Supplementary Material (11.SM)).
                                                    drought events are not
                    Nicholson, 2017)                attributable to anthropogenic
                                                    climate change (Uhe et al.,     Xu et al. (2019): Inconsistent or weak trends Xu et al. (2019): Inconsistent or weak trends Inconsistent or no changes in SPI
                                                    2017; Funk et al., 2018b)       in SPI                                        in SPI                                        (Vicente-Serrano et al., 2020a)

                                                                                    Osima et al. (2018): Cordex AFR data,CTL     Osima et al. (2018): Cordex AFR data,CTL       Sillmann et al. (2013), (2081-2100)/1981-
                                                                                    1971-2000, RCP8.5,                           1971-2000, RCP8.5, Robust increase of          2000, rcp8.5, CMIP3-CMIP5:
                                                                                    consistent increase of CDD over southern     CDD over southern part.                        Decrease of CDD
                                                                                    part
                                                                                                                                 Chapter 11 Supplementary Material              Dosio et al. (2019), (2070-2099/1981-
                                                                                    Chapter 11 Supplementary Material            (11.SM): inconsistent changes in CDD           2010), rcp 8.5, 23 RCM: Decrease in CDD
                                                                                    (11.SM): Inconsistent changes in CDD
                                                                                                                                                                                Inconsistent trends in CDD in CMIP6
                                                                                                                                                                                (Chapter 11 Supplementary Material
                                                                                                                                                                                (11.SM))
           AGR      Low confidence:                 Low confidence: Limited         Low confidence: Inconsistent trends (Xu et Low confidence; Inconsistent trends (Xu et Low confidence: Inconsistent trends
           ECOL     Inconsistent trends (Greve      evidence due to lack of         al., 2019a)(Chapter 11 Supplementary       al., 2019a; Cook et al., 2020) (Chapter 11 (Cook et al., 2020; Vicente-Serrano et al.,
                    et al., 2014; Spinoni et al.,   studies                         Material (11.SM))                          Supplementary Material (11.SM))            2020a)(Chapter 11 Supplementary
                    2019; Padrón et al., 2020)
                                                                                                                                                                          Material (11.SM))



           HYDR     Low confidence:                 Low confidence: Limited         Low confidence; Limited evidence. One        Low confidence: Limited evidence;              Low confidence: Inconsistent trends.
                    Inconsistent trends (Dai        evidence                        study shows lack of signal (Touma et al.,    inconsistent trends in runoff in two studies   Increase in runoff in a study based on
                    and Zhao, 2017)                                                 2015)                                        (Touma et al., 2015; Cook et al., 2020)        CMIP6 (Cook et al., 2020) but
                                                                                                                                                                                inconsistent or non-robust trends in studies
                                                                                                                                                                                based on ISIMIP and CMIP5 ensembles
                                                                                                                                                                                (Giuntoli et al., 2015; Touma et al., 2015)

Do Not Cite, Quote or Distribute                               11-139                                                    Total pages: 345
                              Final Government Distribution                                        Chapter11                                                     IPCC AR6 WGI

 Western    MET     Low confidence:               Low confidence:                  Medium confidence: Increase. Increases in       High confidence: Increases in dryness          Likely: Increase (CDD amd SPI)
 Southern           Inconsistent trends           Limited evidence and             dryness (CDD) (Maúre et al., 2018)(Chapter      (CDD, DF, NDD) (Maúre et al., 2018;            (Sillmann et al., 2013b; Giorgi et al.,
 Africa             (Spinoni et al., 2019; Dunn   inconsistent observed            11 Supplementary Material (11.SM)) both         Coppola et al., 2021b) (Chapter 11             2014; Touma et al., 2015; Pinto et al.,
 (WSAF)             et al., 2020)                 trends.                                                                          Supplementary Material (11.SM)); slight but
                                                                                   compared to pre-industrial climate and                                                         2016; Abiodun et al., 2019; Dosio et al.,
                                                                                                                                   weaker increase in SPI (Abiodun et al.,
                    Dunn et al. (2020):           But recent meteorological        recent past. Increase in CDD for changes of     2019; Naik and Abiodun, 2019; Xu et al.,       2019; Naik and Abiodun, 2020; Spinoni et
                    Conflicting trends in CDD     drought attributable to          +0.5°C in global warming based on CMIP5         2019)                                          al., 2020; Coppola et al., 2021b)
                    depending on time frame       anthropogenic climate            for overall SREX/AR5 South Africa region
                                                  change (Bellprat et al., 2015)   (Wartenburger et al., 2017), but only weak      Maúre et al. (2018): 25 Cordex AFR             Using Cordex , CTL :1981-2010,RCP 8.5
                                                                                   shift in mean precipitation in large-ensemble   run ,CTL 1971-2000, RCP8.5,                    2071-2100 (Spinoni et al., 2020)
                                                  Recent meteorological            single-model experiment for +0.5°C of           -Increase of CDD                               Robust increase of drought frequency and
                                                  drought (2015/2016 drought                                                                                                      severity (SPI-12 )
                                                                                   global warming (Nangombe et al., 2018).
                                                  in southern Africa)                                                              Coppola et al. (2021b), (2041-2060)/1995-
                                                  attributable to anthropogenic    Slight but weaker increase in SPI compared      2014, rcp 8.5, CMIP5-CORDEX-CMIP6
                                                                                   to CDD (Abiodun et al., 2019; Xu et al.,                                                       Based on Giorgi et al., 2014,
                                                  climate change (Otto et al.,                                                     Increase in DF (drought frequency) and
                                                                                                                                                                                  5GCM/1RCM, CTL: 1976-2005, rcp 8.5,
                                                  2018b; Funk et al., 2018a;       2019a; Naik and Abiodun, 2020)                  NDD (number of dry days )
                                                                                                                                                                                  2071-2100:
                                                  Yuan et al., 2018; Pascale et
                                                                                                                                                                                  Increase of CDD
                                                  al., 2020)                       Maúre et al. (2018): 25 Cordex AFR run,         NB: Weaker signals in SPI (Xu et al., 2019a)
                                                                                   CTL 1971-2000, RCP8.5,                                                                         Sillmann et al. (2013), (2081-2100)/1981-
                                                                                                                                Cordex AFR data, CTL 1971-2000, RCP8.5,
                                                                                   -Increase of CDD                                                                               2000, rcp8.5, CMIP3-CMIP5
                                                                                                                                pre-industrial reference period (1861-1890) )
                                                                                                                                (Abiodun et al., 2019; Naik and Abiodun,          Increase of CDD
                                                                                   NB: Weaker signals in SPI (Xu et al., 2019a) 2020): Non-significant increase in SPI-based
                                                                                                                                drought frequency and intensity.                  Coppola et al. (2021b), (2080-2099)/1995-
                                                                                   Cordex AFR data, CTL 1971-2000, RCP8.5,                                                        2014, rcp 8.5, CMIP5-CORDEX-CMIP6
                                                                                   pre-industrial reference period (1861-1890)
                                                                                   (Abiodun et al., 2019; Naik and Abiodun,                                                       Increase in DF (drought frequency) and
                                                                                   2020)                                                                                          NDD (number of dry days )
                                                                                   Non-significant increase in SPI-based
                                                                                   drought frequency and intensity
                                                                                                                                                                                  Dosio et al. (2019) (2070-2099/1981-
                                                                                                                                                                                  2010), rcp 8.5, 23 RCM: Increase in CDD

                                                                                                                                                                                  Pinto et al. (2016): (2069-2098/1976-
                                                                                                                                                                                  2005), rcp 8.5,4 GCM/2RCM: Increase in
                                                                                                                                                                                  CDD.
            AGR     Medium confidence:            Low confidence: Limited          Medium confidence; Drought increase.            High confidence: Drought increase.             Likely: Drought increase.
            ECOL    Drought increase based        evidence:                        Decrease in SM both compared to recent          Decrease in SM (Xu et al., 2019) (Chapter      Decrease in SM (Chapter 11
                    on water-balance estimates    Given small number of            past (Xu et al., 2019) and pre-industrial       11 Supplementary Material (11.SM))             Supplementary Material (11.SM))
                    and SPEI                      studies based on soil
                                                                                   (Chapter 11 Supplementary Material              (Cook et al., 2020); but conflicting changes   (Cook et al., 2020) and SPEI-PM
                    (Greve et al., 2014;          moisture (Yuan et al., 2018a)
                                                                                   (11.SM)) baselines; butut conflicting           of drought magnitude based on SPEI-PM          (Vicente-Serrano et al., 2020a)
                    Spinoni et al., 2019;         and atmospheric drought
                                                                                   changes of drought magnitude based on           (Naumann et al., 2018)
                    Padrón et al., 2020)          indices (Nangombe et al.,
                                                  2020)                            SPEI-PM compared to 0.6°C baseline
                                                                                   (Naumann et al., 2018)




Do Not Cite, Quote or Distribute                             11-140                                                      Total pages: 345
                             Final Government Distribution                                       Chapter11                                                     IPCC AR6 WGI

            HYDR    Low confidence: Limited      Low confidence: Limited         Low confidence: Limited evidence. One           Medium confidence; Increased drying             Medium confidence: Increased drying
                    evidence. Decrease in        evidence                        study shows lack of signal (Touma et al.,       (Touma et al., 2015; Cook et al., 2020;         (Giuntoli et al., 2015; Touma et al., 2015;
                    runoff in larger AR5                                         2015)                                           Zhai et al., 2020a)                             Cook et al., 2020)
                    “Southern Africa” region,
                    but weaker signal
                    depending on time frame
                    (Gudmundsson et al., 2019,
                    2021); non significant
                    drying tendency (Dai and
                    Zhao, 2017)
 Eastern    MET     Medium confidence:           Low confidence: Limited         Medium confidence: Increases in                 High confidence: Increase in                    Likely: Increase in meteorological
 Southern           Dominant increase in         evidence on attribution of      meteorological drought based on CDD             meteorological drought based on                 drought (CDD amd SPI) (Sillmann et al.,
 Africa             meteorological drought in    long-term trends.               (Maúre et al., 2018)(Chapter 11                 (CDD,DF,NDD) (Maúre et al., 2018;               2013b; Giorgi et al., 2014; Touma et al.,
 (ESAF)             SPI and CDD (Spinoni et
                                                                                 Supplementary Material (11.SM)) both            Coppola et al., 2021b)(Chapter 11               2015; Pinto et al., 2016; Dosio et al., 2019;
                    al., 2019; Dunn et al.,      Medium confidence that
                    2020)                        human-influence has             compared to pre-industrial climate and          Supplementary Material (11.SM)) and SPI         Spinoni et al., 2020; Coppola et al.,
                                                 contributed to stronger         recent past. Non-significant increase in SPI-   (Abiodun et al., 2019; Xu et al., 2019a) both   2021b)(Chapter 11 Supplementary
                                                 recent meteorological           based drought (Abiodun et al., 2017); lack of   compared to recent past and pre-industrial      Material (11.SM))
                                                 drought.(Bellprat et al., 2015; signal in SPI compared to recent past (1970-    period.                                         Using Cordex, CTL :1981-2010,RCP 8.5,
                                                 Funk et al., 2018a; Yuan et     2000) (Xu et al., 2019a). Increase in CDD                                                       2071-2100 (Spinoni et al., 2020)
                                                 al., 2018a)                     for changes of +0.5°C in global warming         Maúre et al. (2018): 25 Cordex AFR run          Robust increase of drought frequency and
                                                                                 based on CMIP5 for overall SREX/AR5             ,CTL 1971-2000, RCP8.5: Increase of CDD         severity (SPI-12,SPEI-12)
                                                                                 South Africa region (Wartenburger et al.,
                                                                                 2017), but only weak shift in mean              (Coppola et al., 2021b), (2041-2060)/1995-      Based on Giorgi et al. (2014),
                                                                                 precipitation in large-ensemble single-model    2014, rcp 8.5, CMIP5-CORDEX-CMIP6:              5GCM/1RCM, CTL: 1976-2005, rcp 8.5,
                                                                                 experiment for +0.5°C of global warming         Increase in DF (drought frequency) and          2071-2100:
                                                                                                                                                                                 Increase of CDD
                                                                                 (Nangombe et al., 2018).                        NDD (number of dry days )

                                                                                                                                                                                 Sillmann et al. (2013), (2081-2100)/1981-
                                                                                 Maúre et al. (2018): 25 Cordex AFR run Abiodun et al. (2019): Cordex AFR
                                                                                                                             data,CTL 1971-2000, RCP8.5, pre-industrial          2000, rcp8.5, CMIP3-CMIP5
                                                                                 ,CTL 1971-2000, RCP8.5,
                                                                                                                             reference period (1861-1890): increase in           Increase of CDD
                                                                                 -Increase of CDD
                                                                                                                             SPI-based meteorological drought frequency
                                                                                                                             and intensity.                                      (Coppola et al., 2021b), (2080-
                                                                                 Cordex AFR data,CTL 1971-2000, RCP8.5,
                                                                                                                                                                                 2099)/1995-2014, rcp 8.5, CMIP5-
                                                                                 pre-industrial reference period (1861-1890) Xu et al. (2019): Drying in SPI at 2°C
                                                                                                                                                                                 CORDEX-CMIP6
                                                                                 (Abiodun et al., 2019)
                                                                                                                             compared to 1970-2000 conditions.                   Increase in DF (drought frequency) and
                                                                                 SPI non-significant drought frequency &
                                                                                                                                                                                 NDD (number of dry days )
                                                                                 intensity increase

                                                                                                                                                                                 Dosio et al. (2019), (2070-2099/1981-
                                                                                                                                                                                 2010), rcp 8.5, 23 RCM
                                                                                                                                                                                 Increase in CDD

                                                                                                                                                                                 Pinto et al. (2016): (2069-2098/1976-
                                                                                                                                                                                 2005), rcp 8.5,4 GCM/2RCM: Increase in
                                                                                                                                                                                 CDD

Do Not Cite, Quote or Distribute                            11-141                                                    Total pages: 345
                              Final Government Distribution                                 Chapter11                                                     IPCC AR6 WGI

           AGR      Medium confidence              Low confidence: Limited   Medium confidence: Increase in drought.        Medium confidence: Increase in                  High confidence: Increase in drought:
           ECOL     Increase, based on water-      evidence                  Decrease in SM both compared to recent         drought;decrease in SM (Xu et al., 2019a;       decrease in SM (Chapter 11
                    balances estimates, PDSI       (Yuan et al., 2018a)      past (Xu et al., 2019) and pre-industrial      Cook et al., 2020) (Chapter 11                  Supplementary Material (11.SM))
                    and SPEI-PM (Greve et al.,
                                                                             (Chapter 11 Supplementary Material             Supplementary Material (11.SM)); but            (Cook et al., 2020) and SPEI-PM
                    2014; Dai and Zhao, 2017;
                    Spinoni et al., 2019;                                    (11.SM)) baselines; but inconsistent changes   inconsistent changes of in drought              (Vicente-Serrano et al., 2020a)
                    Padrón et al., 2020)                                     of drought magnitude based on SPEI-PM          magnitude based on SPEI-PM (Naumann et
                                                                             compared to +0.6°C baseline (Naumann et        al., 2018)
                                                                             al., 2018)
           HYDR     Low confidence: Limited        Low confidence: Limited   Low confidence: Limited evidence. One          Medium confidence; Increased drying             Medium confidence: Increased drying
                    evidence.                      evidence                  study shows lack of signal (Touma et al.,      (Touma et al., 2015; Cook et al., 2020;         (Giuntoli et al., 2015; Touma et al., 2015;
                    Decrease in runoff in larger                             2015)                                          Zhai et al., 2020a).                            Cook et al., 2020)
                    AR5 “Southern Africa”
                    region, but weaker signal
                    depending on time frame
                    (Gudmundsson et al., 2019,
                    2021); non significant
                    drying tendency (Dai and
                    Zhao, 2017)
 Mada-     MET      Low confidence:                Low confidence: Limited   Medium confidence: Increase in                 High confidence: Increase in                    Likely: Increase in meteorological
 gascar             Inconsistent trends            evidence                  meteorological drought based on SPI            meteoroloigcal drought based on several         drought based on CDD and SPI
 (MDG)              (Vincent et al., 2011;                                   compared to recent past (Abiodun et al.,       metrics, including SPI (Abiodun et al., 2019;   (Sillmann et al., 2013b; Giorgi et al.,
                    Spinoni et al., 2019)
                                                                             2019; Xu et al., 2019a) and CDD compared       Xu et al., 2019a), CDD (Chapter 11              2014; Touma et al., 2015; Pinto et al.,
                                                                             to pre-industrial baseline (Chapter 11         Supplementary Material (11.SM)), and DF         2016; Dosio et al., 2019; Spinoni et al.,
                                                                             Supplementary Material (11.SM)).               (drought frequency) and NDD (number of          2020; Coppola et al., 2021b)
                                                                                                                            dry days) (Coppola et al., 2021b)
                                                                                                                                                                            Sillmann et al. (2013), (2081-2100)/1981-
                                                                                                                          (Coppola et al., 2021b), (2041-2060)/1995-        2000, rcp8.5, CMIP3-CMIP5
                                                                                                                          2014, rcp 8.5, CMIP5-CORDEX-CMIP6                 Increase of CDD
                                                                             Abiodun et al. (2019): Cordex AFR            Increase in DF (drought frequency) and
                                                                             data,CTL 1971-2000, RCP8.5, pre-industrial NDD (number of dry days )                           Spinoni et al. (2020): Using CORDEX,
                                                                             reference period (1861-1890)                                                                   CTL:1981-2010,RCP 8.5, 2071-2100
                                                                             SPI (drought frequency & intensity increase) Abiodun et al. (2019): Cordex AFR                 Robust increase of drought frequency and
                                                                                                                          data,CTL 1971-2000, RCP8.5, pre-industrial        severity (SPI-12)
                                                                                                                          reference period (1861-1890): Increase in
                                                                                                                          SPI-based drought frequency and intensity.        (Coppola et al., 2021b), (2080-
                                                                                                                                                                            2099)/1995-2014, rcp 8.5, CMIP5-
                                                                                                                                                                            CORDEX-CMIP6
                                                                                                                                                                            Increase in DF (drought frequency) and
                                                                                                                                                                            NDD (number of dry days )

                                                                                                                                                                            Dosio et al. (2019), (2070-2099/1981-
                                                                                                                                                                            2010), rcp 8.5, 23 RCM: Increase in CDD




Do Not Cite, Quote or Distribute                             11-142                                               Total pages: 345
                                  Final Government Distribution                                        Chapter11                                                      IPCC AR6 WGI

                  AGR    Low confidence:                Low confidence: Limited         Low confidence: Inconsistent or weak          Medium confidence: Increase in drought.        High confidence: Increase in drought.
                  ECOL   Inconsistent trends based      evidence                        trends (Xu et al., 2019) (Chapter 11          Decrease in SM (Chapter 11 Supplementary       Robust decrease in SM (Chapter 11
                         on water-balance                                               Supplementary Material (11.SM))               Material (11.SM); (Cook et al., 2020) and in   Supplementary Material (11.SM))
                         estimates, PDSI and SPEI
                                                                                        (Naumann et al., 2018)                        SPEI-PM (Naumann et al., 2018)                 (Cook et al., 2020) and SPEI-PM
                         (Greve et al., 2014; Dai
                         and Zhao, 2017; Spinoni et                                                                                                                                  (Vicente-Serrano et al., 2020a)
                         al., 2019; Padrón et al.,
                         2020)
                  HYDR   Low confidence: Limited        Low confidence: Limited          Low confidence: Limited evidence. One         Low confidence: Inconsistent trends.           Medium confidence: Increase in
                         evidence. Inconsistent         evidence                         study shows lack of signal (Touma et al.,     Inconsistent trends (Cook et al., 2020) or     drought based on two studies based on
                         trends in one study (Dai                                        2015)                                         weak drying (Touma et al., 2015; Zhai et       CMIP5 (Giuntoli et al., 2015; Touma et
                         and Zhao, 2017)                                                                                               al., 2020b)
                                                                                                                                                                                      al., 2015), but some inconsistent trends
                                                                                                                                                                                      in CMIP6 mean runoff trends (Cook et
                                                                                                                                                                                      al., 2020)
1
2   [END TABLE 11.6 HERE]
3
4
5   [START TABLE 11.7 HERE]
6
7   Table 11.7: Observed trends, human contribution to observed trends, and projected changes at 1.5°C, 2°C and 4°C of global warming for temperature extremes in Asia, subdivided
8               by AR6 regions. See Sections 11.9.1 and 11.9.2 for details
                                                                     Detection and attribution;                                                 Projections
                Region                Observed trends
                                                                         event attribution                      1.5 °C                              2 °C                              4 °C
     All Asia                  Most subregions show a very        Robust evidence of a human       CMIP6 models project a            CMIP6 models project a            CMIP6 models project a
                               likely increase in the intensity   contribution to the observed     robust increase in the            robust increase in the            robust increase in the
                               and frequency of hot               increase in the intensity and    intensity and frequency of        intensity and frequency of        intensity and frequency of
                               extremes and decrease in the       frequency of hot extremes        TXx events and a robust           TXx events and a robust           TXx events and a robust
                               intensity and frequency of         and decrease in the intensity    decrease in the intensity and     decrease in the intensity and     decrease in the intensity and
                               cold extremes                      and frequency of cold            frequency of TNn events (Li       frequency of TNn events (Li       frequency of TNn events (Li
                                                                  extremes (Hu et al., 2020;       et al., 2020). Median increase    et al., 2020). Median increase    et al., 2020). Median increase
                                                                  Seong et al., 2020)              of more than 0.5°C in the 50-     of more than 1°C in the 50-       of more than 4.5°C in the 50-
                                                                                                   year TXx and TNn events           year TXx and TNn events           year TXx and TNn events
                                                                                                   compared to the 1°C warming       compared to the 1°C warming       compared to the 1°C warming
                                                                                                   level (Li et al., 2020)           level (Li et al., 2020)           level (Li et al., 2020)
                               Very likely increase in the        Human influence very likely      Increase in the intensity and     Increase in the intensity and     Increase in the intensity and
                               intensity and frequency of hot     contributed to the observed      frequency of hot extremes :       frequency of hot extremes:        frequency of hot extremes:
                               extremes and decrease in the       increase in the intensity and    Very likely (compared with        Extremely likely (compared        Virtually certain (compared
                               intensity and frequency of         frequency of hot extremes        the recent past (1995-2014))      with the recent past (1995-       with the recent past (1995-
                               cold extremes                      and decrease in the intensity    Extremely likely (compared        2014))                            2014))
                                                                  and frequency of cold            with pre-industrial)              Virtually certain (compared       Virtually certain (compared
                                                                                                                                     with pre-industrial)              with pre-industrial)
                                                                  extremes
                                                                                                   Decrease in the intensity and
                                                                                                   frequency of cold extremes:       Decrease in the intensity and     Decrease in the intensity and

    Do Not Cite, Quote or Distribute                               11-143                                                    Total pages: 345
                              Final Government Distribution                                        Chapter11                                                        IPCC AR6 WGI
                                                                                              Very likely (compared with    frequency of cold extremes:              frequency of cold extremes:
                                                                                              the recent past (1995-2014))  Extremely likely (compared               Virtually certain (compared
                                                                                              Extremely likely (compared    with the recent past (1995-              with the recent past (1995-
                                                                                              with pre-industrial)          2014))                                   2014))
                                                                                                                            Virtually certain (compared              Virtually certain (compared
                                                                                                                            with pre-industrial)                     with pre-industrial)
 Russian Arctic (RAR)      Significant increases in the     Evidence of a human           CMIP6 models project a robust CMIP6 models project a robust               CMIP6 models project a robust
                           intensity and frequency of hot   contribution to the observed  increase in the intensity and    increase in the intensity and            increase in the intensity and
                           extremes and significant         increase in the intensity and frequency of TXx events and a frequency of TXx events and a               frequency of TXx events and a
                           decreases in the intensity and   frequency of hot extremes     robust decrease in the intensity robust decrease in the intensity         robust decrease in the intensity
                           frequency of cold extremes       and decrease in the intensity and frequency of TNn events (Li and frequency of TNn events               and frequency of TNn events
                           (Donat et al., 2016a; Sui et     and frequency of cold         et al., 2020; Annex). Median     (Li et al., 2020; Annex).                (Li et al., 2020; Annex).
                           al., 2017; Dunn et al., 2020)    extremes (Wang et al., 2017c) increase of more than 0.5°C in Median increase of more than               Median increase of more than
                                                                                          the 50-year TXx and TNn events 1.5°C in the 50-year TXx and               4.5°C in the 50-year TXx and
                                                                                          compared to the 1°C warming      TNn events compared to the               TNn events compared to the
                                                                                          level (Li et al., 2020) and more 1°C warming level (Li et al.,            1°C warming level (Li et al.,
                                                                                          than 1.5°C in annual TXx and     2020) and more than 2.5°C in             2020) and more than 5.5°C in
                                                                                          TNn compared to pre-industrial annual TXx and TNn compared                annual TXx and TNn compared
                                                                                          (Annex).                         to pre-industrial (Annex).               to pre-industrial (Annex).

                                                                                             Additional evidence from           Additional evidence from            Additional evidence from
                                                                                             CMIP5 and RCM simulations for      CMIP5 and RCM simulations           CMIP5 and RCM simulations
                                                                                             an increase in the intensity and   for an increase in the intensity    for an increase in the intensity
                                                                                             frequency of hot extremes and      and frequency of hot extremes       and frequency of hot extremes
                                                                                             decrease in the intensity and      and decrease in the intensity       and decrease in the intensity
                                                                                             frequency of cold extremes (Xu     and frequency of cold extremes      and frequency of cold extremes
                                                                                             et al. 2017; Han et al. 2018;      (Xu et al. 2017; Han et al. 2018;   (Xu et al. 2017; Han et al. 2018;
                                                                                             Khlebnikova et al. 2019)           Khlebnikova et al. 2019)            Khlebnikova et al. 2019)
                           Very likely increase in the      Medium confidence in a           Increase in the intensity and      Increase in the intensity and       Increase in the intensity and
                           intensity and frequency of hot   human contribution to the        frequency of hot extremes:         frequency of hot extremes:          frequency of hot extremes:
                           extremes and decrease in the     observed increase in the         Likely (compared with the recent   Very likely (compared with the      Virtually certain (compared
                           intensity and frequency of       intensity and frequency of hot   past (1995-2014))                  recent past (1995-2014))            with the recent past (1995-
                           cold extremes                    extremes and decrease in the     Very likely (compared with pre-    Extremely likely (compared          2014))
                                                            intensity and frequency of       industrial)                        with pre-industrial)                Virtually certain (compared
                                                            cold extremes.                                                                                          with pre-industrial)
                                                                                             Decrease in the intensity and      Decrease in the intensity and
                                                                                             frequency of cold extremes:        frequency of cold extremes:         Decrease in the intensity and
                                                                                             Likely (compared with the recent   Very likely (compared with the      frequency of cold extremes:
                                                                                             past (1995-2014))                  recent past (1995-2014))            Virtually certain (compared
                                                                                             Very likely (compared with pre-    Extremely likely (compared          with the recent past (1995-
                                                                                             industrial).                       with pre-industrial)                2014))
                                                                                                                                                                    Virtually certain (compared
                                                                                                                                                                    with pre-industrial)
 Arabian Peninsula (ARP)   Significant increases in the     Strong evidence of changes       CMIP6 models project a robust      CMIP6 models project a robust       CMIP6 models project a robust
                           intensity and frequency of hot   from observations that are in    increase in the intensity and      increase in the intensity and       increase in the intensity and
                           extremes and significant         the direction of model           frequency of TXx events and a      frequency of TXx events and a       frequency of TXx events and a
                           decreases in the intensity and   projected changes for the        robust decrease in the intensity   robust decrease in the intensity    robust decrease in the intensity

Do Not Cite, Quote or Distribute                             11-144                                                      Total pages: 345
                              Final Government Distribution                                        Chapter11                                                        IPCC AR6 WGI

                           frequency of cold extremes        future. The magnitude of         and frequency of TNn events (Li    and frequency of TNn events        and frequency of TNn events
                           (Dunn et al., 2020; Almazroui     projected changes increases      et al., 2020; Annex). Median       (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                           et al., 2014; Barlow et al.,      with global warming.             increase of more than 0.5°C in     Median increase of more than       Median increase of more than
                           2016; Donat et al., 2014;                                          the 50-year TXx and TNn events     1.5°C in the 50-year TXx and       3.5°C in the 50-year TXx and
                           Nazrul Islam et al., 2015;                                         compared to the 1°C warming        TNn events compared to the         TNn events compared to the
                           Rahimi and Hejabi, 2018;                                           level (Li et al., 2020) and more   1°C warming level (Li et al.,      1°C warming level (Li et al.,
                           Donat et al., 2014; Rahimi et                                      than 2°C in annual TXx and TNn     2020) and more than 2.5°C in       2020) and more than 5.5°C in
                           al., 2018)                                                         compared to pre-industrial         annual TXx and TNn compared        annual TXx and TNn compared
                                                                                              (Annex).                           to pre-industrial (Annex).         to pre-industrial (Annex).

                                                                                              Additional evidence from           Additional evidence from           Additional evidence from
                                                                                              CMIP5 and RCM simulations for      CMIP5 and RCM simulations          CMIP5 and RCM simulations
                                                                                              an increase in the intensity and   for an increase in the intensity   for an increase in the intensity
                                                                                              frequency of hot extremes and      and frequency of hot extremes      and frequency of hot extremes
                                                                                              decrease in the intensity and      and decrease in the intensity      and decrease in the intensity
                                                                                              frequency of cold extremes         and frequency of cold extremes     and frequency of cold extremes
                                                                                              (Almazroui, 2019b)                 (Almazroui, 2019b)                 (Almazroui, 2019b)
                           Very likely increase in the       Medium confidence in a           Increase in the intensity and      Increase in the intensity and      Increase in the intensity and
                           intensity and frequency of hot    human contribution to the        frequency of hot extremes:         frequency of hot extremes:         frequency of hot extremes:
                           extremes and decrease in the      observed increase in the         Likely (compared with the recent   Very likely (compared with the     Virtually certain (compared
                           intensity and frequency of        intensity and frequency of hot   past (1995-2014))                  recent past (1995-2014))           with the recent past (1995-
                           cold extremes                     extremes and decrease in the     Very likely (compared with pre-    Extremely likely (compared         2014))
                                                             intensity and frequency of       industrial)                        with pre-industrial)               Virtually certain (compared
                                                             cold extremes.                                                                                         with pre-industrial)
                                                                                              Decrease in the intensity and      Decrease in the intensity and
                                                                                              frequency of cold extremes:        frequency of cold extremes:        Decrease in the intensity and
                                                                                              Likely (compared with the recent   Very likely (compared with the     frequency of cold extremes:
                                                                                              past (1995-2014))                  recent past (1995-2014))           Virtually certain (compared
                                                                                              Very likely (compared with pre-    Extremely likely (compared         with the recent past (1995-
                                                                                              industrial).                       with pre-industrial)               2014))
                                                                                                                                                                    Virtually certain (compared
                                                                                                                                                                    with pre-industrial)
 West Central Asia (WCA)   Significant increases in the      Robust evidence of a human       CMIP6 models project a robust      CMIP6 models project a robust      CMIP6 models project a robust
                           intensity and frequency of hot    contribution to the observed     increase in the intensity and      increase in the intensity and      increase in the intensity and
                           extremes and significant          increase in the intensity and    frequency of TXx events and a      frequency of TXx events and a      frequency of TXx events and a
                           decreases in the intensity and    frequency of hot extremes        robust decrease in the intensity   robust decrease in the intensity   robust decrease in the intensity
                           frequency of cold extremes        and decrease in the intensity    and frequency of TNn events (Li    and frequency of TNn events        and frequency of TNn events
                           (Hu et al., 2016; Jiang et al.,   and frequency of cold            et al., 2020; Annex). Median       (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                           2013; Dunn et al., 2020)          extremes (Seong et al., 2020;    increase of more than 0.5°C in     Median increase of more than       Median increase of more than
                                                             Wang et al., 2017; Dong et       the 50-year TXx and TNn events     1.5°C in the 50-year TXx and       5°C in the 50-year TXx and
                                                             al., 2018; Kim et al., 2019)     compared to the 1°C warming        TNn events compared to the         TNn events compared to the
                                                                                              level (Li et al., 2020) and more   1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                                                              than 2°C in annual TXx and TNn     2020) and more than 3°C in         2020) and more than 6°C in
                                                                                              compared to pre-industrial         annual TXx and TNn compared        annual TXx and TNn compared
                                                                                              (Annex).                           to pre-industrial (Annex).         to pre-industrial (Annex).

                                                                                              Additional evidence from           Additional evidence from           Additional evidence from

Do Not Cite, Quote or Distribute                              11-145                                                     Total pages: 345
                             Final Government Distribution                                      Chapter11                                                         IPCC AR6 WGI

                                                                                           CMIP5 simulations for an           CMIP5 simulations for an            CMIP5 simulations for an
                                                                                           increase in the intensity and      increase in the intensity and       increase in the intensity and
                                                                                           frequency of hot extremes and      frequency of hot extremes and       frequency of hot extremes and
                                                                                           decrease in the intensity and      decrease in the intensity and       decrease in the intensity and
                                                                                           frequency of cold extremes (Han    frequency of cold extremes          frequency of cold extremes
                                                                                           et al., 2018)                      (Han et al., 2018)                  (Han et al., 2018)

                          Very likely increase in the      High confidence in a human      Increase in the intensity and      Increase in the intensity and       Increase in the intensity and
                          intensity and frequency of hot   contribution to the observed    frequency of hot extremes:         frequency of hot extremes:          frequency of hot extremes:
                          extremes and decrease in the     increase in the intensity and   Likely (compared with the recent   Very likely (compared with the      Virtually certain (compared
                          intensity and frequency of       frequency of hot extremes       past (1995-2014))                  recent past (1995-2014))            with the recent past (1995-
                          cold extremes                    and decrease in the intensity   Very likely (compared with pre-    Extremely likely (compared          2014))
                                                           and frequency of cold           industrial)                        with pre-industrial)                Virtually certain (compared
                                                           extremes.                                                                                              with pre-industrial)
                                                                                           Decrease in the intensity and      Decrease in the intensity and
                                                                                           frequency of cold extremes:        frequency of cold extremes:         Decrease in the intensity and
                                                                                           Likely (compared with the recent   Very likely (compared with the      frequency of cold extremes:
                                                                                           past (1995-2014))                  recent past (1995-2014))            Virtually certain (compared
                                                                                           Very likely (compared with pre-    Extremely likely (compared          with the recent past (1995-
                                                                                           industrial).                       with pre-industrial)                2014))
                                                                                                                                                                  Virtually certain (compared
                                                                                                                                                                  with pre-industrial)
 West Siberia (WSB)       Significant increases in the     Robust evidence of a human      CMIP6 models project a robust      CMIP6 models project a robust       CMIP6 models project a robust
                          intensity and frequency of hot   contribution to the observed    increase in the intensity and      increase in the intensity and       increase in the intensity and
                          extremes and significant         increase in the intensity and   frequency of TXx events and a      frequency of TXx events and a       frequency of TXx events and a
                          decreases in the intensity and   frequency of hot extremes       robust decrease in the intensity   robust decrease in the intensity    robust decrease in the intensity
                          frequency of cold extremes       and decrease in the intensity   and frequency of TNn events (Li    and frequency of TNn events         and frequency of TNn events
                          (Degefie et al., 2014;           and frequency of cold           et al., 2020; Annex). Median       (Li et al., 2020; Annex).           (Li et al., 2020; Annex).
                          Salnikov et al., 2015; Donat     extremes (Wang et al., 2017;    increase of more than 0.5°C in     Median increase of more than        Median increase of more than
                          et al., 2016a; Zhang et al.,     Seong et al., 2020; Dong et     the 50-year TXx and TNn events     1°C in the 50-year TXx and          4°C in the 50-year TXx and
                          2019c, 2019b; Dunn et al.,       al., 2018)                      compared to the 1°C warming        TNn events compared to the          TNn events compared to the
                          2020)                                                            level (Li et al., 2020) and more   1°C warming level (Li et al.,       1°C warming level (Li et al.,
                                                                                           than 2°C in annual TXx and TNn     2020) and more than 2.5°C in        2020) and more than 5°C in
                                                                                           compared to pre-industrial         annual TXx and TNn compared         annual TXx and TNn compared
                                                                                           (Annex).                           to pre-industrial (Annex).          to pre-industrial (Annex).

                                                                                           Additional evidence from           Additional evidence from            Additional evidence from
                                                                                           CMIP5 and RCM simulations for      CMIP5 and RCM simulations           CMIP5 and RCM simulations
                                                                                           an increase in the intensity and   for an increase in the intensity    for an increase in the intensity
                                                                                           frequency of hot extremes and      and frequency of hot extremes       and frequency of hot extremes
                                                                                           decrease in the intensity and      and decrease in the intensity       and decrease in the intensity
                                                                                           frequency of cold extremes (Xu     and frequency of cold extremes      and frequency of cold extremes
                                                                                           et al. 2017; Han et al. 2018;      (Xu et al. 2017; Han et al. 2018;   (Xu et al. 2017; Han et al. 2018;
                                                                                           Khlebnikova et al. 2019)           Khlebnikova et al. 2019)            Khlebnikova et al. 2019)
                          Very likely increase in the      High confidence in a human      Increase in the intensity and    Increase in the intensity and         Increase in the intensity and
                          intensity and frequency of hot   contribution to the observed    frequency of hot extremes:       frequency of hot extremes:            frequency of hot extremes:
                          extremes and decrease in the     increase in the intensity and   Likely (compared with the recent Very likely (compared with the        Virtually certain (compared

Do Not Cite, Quote or Distribute                            11-146                                                    Total pages: 345
                             Final Government Distribution                                      Chapter11                                                         IPCC AR6 WGI

                          intensity and frequency of       frequency of hot extremes       past (1995-2014))                  recent past (1995-2014))            with the recent past (1995-
                          cold extremes                    and decrease in the intensity   Very likely (compared with pre-    Extremely likely (compared          2014))
                                                           and frequency of cold           industrial)                        with pre-industrial)                Virtually certain (compared
                                                           extremes.                                                                                              with pre-industrial)
                                                                                           Decrease in the intensity and      Decrease in the intensity and
                                                                                           frequency of cold extremes:        frequency of cold extremes:         Decrease in the intensity and
                                                                                           Likely (compared with the recent   Very likely (compared with the      frequency of cold extremes:
                                                                                           past (1995-2014))                  recent past (1995-2014))            Virtually certain (compared
                                                                                           Very likely (compared with pre-    Extremely likely (compared          with the recent past (1995-
                                                                                           industrial).                       with pre-industrial)                2014))
                                                                                                                                                                  Virtually certain (compared
                                                                                                                                                                  with pre-industrial)
 East Siberia (ESB)       Significant increases in the     Robust evidence of a human      CMIP6 models project a robust      CMIP6 models project a robust       CMIP6 models project a robust
                          intensity and frequency of hot   contribution to the observed    increase in the intensity and      increase in the intensity and       increase in the intensity and
                          extremes and significant         increase in the intensity and   frequency of TXx events and a      frequency of TXx events and a       frequency of TXx events and a
                          decreases in the intensity and   frequency of hot extremes       robust decrease in the intensity   robust decrease in the intensity    robust decrease in the intensity
                          frequency of cold extremes       and decrease in the intensity   and frequency of TNn events (Li    and frequency of TNn events         and frequency of TNn events
                          (Dashkhuu et al., 2015; Donat    and frequency of cold           et al., 2020; Annex). Median       (Li et al., 2020; Annex).           (Li et al., 2020; Annex).
                          et al., 2016a; Zhang et al.,     extremes (Wang et al., 2017;    increase of more than 0.5°C in     Median increase of more than        Median increase of more than
                          2019c; Dunn et al., 2020)        Seong et al., 2020; Dong et     the 50-year TXx and TNn events     1.5°C in the 50-year TXx and        4.5°C in the 50-year TXx and
                                                           al., 2018)                      compared to the 1°C warming        TNn events compared to the          TNn events compared to the
                                                                                           level (Li et al., 2020) and more   1°C warming level (Li et al.,       1°C warming level (Li et al.,
                                                                                           than 2°C in annual TXx and TNn     2020) and more than 2.5°C in        2020) and more than 5.5°C in
                                                                                           compared to pre-industrial         annual TXx and TNn compared         annual TXx and TNn compared
                                                                                           (Annex).                           to pre-industrial (Annex).          to pre-industrial (Annex).

                                                                                           Additional evidence from           Additional evidence from            Additional evidence from
                                                                                           CMIP5 and RCM simulations for      CMIP5 and RCM simulations           CMIP5 and RCM simulations
                                                                                           an increase in the intensity and   for an increase in the intensity    for an increase in the intensity
                                                                                           frequency of hot extremes and      and frequency of hot extremes       and frequency of hot extremes
                                                                                           decrease in the intensity and      and decrease in the intensity       and decrease in the intensity
                                                                                           frequency of cold extremes (Xu     and frequency of cold extremes      and frequency of cold extremes
                                                                                           et al. 2017; Han et al. 2018;      (Xu et al. 2017; Han et al. 2018;   (Xu et al. 2017; Han et al. 2018;
                                                                                           Khlebnikova et al. 2019)           Khlebnikova et al. 2019)            Khlebnikova et al. 2019)
                          Very likely increase in the      High confidence in a human      Increase in the intensity and      Increase in the intensity and       Increase in the intensity and
                          intensity and frequency of hot   contribution to the observed    frequency of hot extremes:         frequency of hot extremes:          frequency of hot extremes:
                          extremes and decrease in the     increase in the intensity and   Likely (compared with the recent   Very likely (compared with the      Virtually certain (compared
                          intensity and frequency of       frequency of hot extremes       past (1995-2014))                  recent past (1995-2014))            with the recent past (1995-
                          cold extremes                    and decrease in the intensity   Very likely (compared with pre-    Extremely likely (compared          2014))
                                                           and frequency of cold           industrial)                        with pre-industrial)                Virtually certain (compared
                                                           extremes.                                                                                              with pre-industrial)
                                                                                           Decrease in the intensity and      Decrease in the intensity and
                                                                                           frequency of cold extremes:        frequency of cold extremes:         Decrease in the intensity and
                                                                                           Likely (compared with the recent   Very likely (compared with the      frequency of cold extremes:
                                                                                           past (1995-2014))                  recent past (1995-2014))            Virtually certain (compared
                                                                                           Very likely (compared with pre-    Extremely likely (compared          with the recent past (1995-
                                                                                           industrial).                       with pre-industrial)                2014))

Do Not Cite, Quote or Distribute                            11-147                                                    Total pages: 345
                             Final Government Distribution                                        Chapter11                                                        IPCC AR6 WGI

                                                                                                                                                                   Virtually certain (compared
                                                                                                                                                                   with pre-industrial)
 Russian Far East (RFE)   Significant increases in the      Robust evidence of a human      CMIP6 models project a robust      CMIP6 models project a robust       CMIP6 models project a robust
                          intensity and frequency of hot    contribution to the observed    increase in the intensity and      increase in the intensity and       increase in the intensity and
                          extremes and significant          increase in the intensity and   frequency of TXx events and a      frequency of TXx events and a       frequency of TXx events and a
                          decreases in the intensity and    frequency of hot extremes       robust decrease in the intensity   robust decrease in the intensity    robust decrease in the intensity
                          frequency of cold extremes        and decrease in the intensity   and frequency of TNn events (Li    and frequency of TNn events         and frequency of TNn events
                          (Donat et al., 2016; Dunn et      and frequency of cold           et al., 2020; Annex). Median       (Li et al., 2020; Annex).           (Li et al., 2020; Annex).
                          al., 2020; Zhang et al., 2019b)   extremes (Seong et al., 2020;   increase of more than 0.5°C in     Median increase of more than        Median increase of more than
                                                            Wang et al., 2017; Dong et      the 50-year TXx and TNn events     1°C in the 50-year TXx and          4.5°C in the 50-year TXx and
                                                            al., 2018)                      compared to the 1°C warming        TNn events compared to the          TNn events compared to the
                                                                                            level (Li et al., 2020) and more   1°C warming level (Li et al.,       1°C warming level (Li et al.,
                                                                                            than 1.5°C in annual TXx and       2020) and more than 2.5°C in        2020) and more than 5°C in
                                                                                            TNn compared to pre-industrial     annual TXx and TNn compared         annual TXx and TNn compared
                                                                                            (Annex).                           to pre-industrial (Annex).          to pre-industrial (Annex).

                                                                                            Additional evidence from           Additional evidence from            Additional evidence from
                                                                                            CMIP5 and RCM simulations for      CMIP5 and RCM simulations           CMIP5 and RCM simulations
                                                                                            an increase in the intensity and   for an increase in the intensity    for an increase in the intensity
                                                                                            frequency of hot extremes and      and frequency of hot extremes       and frequency of hot extremes
                                                                                            decrease in the intensity and      and decrease in the intensity       and decrease in the intensity
                                                                                            frequency of cold extremes (Xu     and frequency of cold extremes      and frequency of cold extremes
                                                                                            et al. 2017; Han et al. 2018;      (Xu et al. 2017; Han et al. 2018;   (Xu et al. 2017; Han et al. 2018;
                                                                                            Khlebnikova et al. 2019).          Khlebnikova et al. 2019).           Khlebnikova et al. 2019).
                          Very likely increase in the       High confidence in a human      Increase in the intensity and      Increase in the intensity and       Increase in the intensity and
                          intensity and frequency of hot    contribution to the observed    frequency of hot extremes:         frequency of hot extremes:          frequency of hot extremes:
                          extremes and decrease in the      increase in the intensity and   Likely (compared with the recent   Very likely (compared with the      Virtually certain (compared
                          intensity and frequency of        frequency of hot extremes       past (1995-2014))                  recent past (1995-2014))            with the recent past (1995-
                          cold extremes                     and decrease in the intensity   Very likely (compared with pre-    Extremely likely (compared          2014))
                                                            and frequency of cold           industrial)                        with pre-industrial)                Virtually certain (compared
                                                            extremes.                                                                                              with pre-industrial)
                                                                                            Decrease in the intensity and      Decrease in the intensity and
                                                                                            frequency of cold extremes:        frequency of cold extremes:         Decrease in the intensity and
                                                                                            Likely (compared with the recent   Very likely (compared with the      frequency of cold extremes:
                                                                                            past (1995-2014))                  recent past (1995-2014))            Virtually certain (compared
                                                                                            Very likely (compared with pre-    Extremely likely (compared          with the recent past (1995-
                                                                                            industrial).                       with pre-industrial)                2014))
                                                                                                                                                                   Virtually certain (compared
                                                                                                                                                                   with pre-industrial)
 East Asia (EAS)          Significant increases in the      Robust evidence of a human      CMIP6 models project a robust      CMIP6 models project a robust       CMIP6 models project a robust
                          intensity and frequency of hot    contribution to the observed    increase in the intensity and      increase in the intensity and       increase in the intensity and
                          extremes and significant          increase in the intensity and   frequency of TXx events and a      frequency of TXx events and a       frequency of TXx events and a
                          decreases in the intensity and    frequency of hot extremes       robust decrease in the intensity   robust decrease in the intensity    robust decrease in the intensity
                          frequency of cold extremes        and decrease in the intensity   and frequency of TNn events (Li    and frequency of TNn events         and frequency of TNn events
                          (Lin et al., 2017; Lu et al.,     and frequency of cold           et al., 2020; Annex). Median       (Li et al., 2020; Annex).           (Li et al., 2020; Annex).
                          2016, 2018; Wang et al.,          extremes (Seong et al., 2020;   increase of more than 0.5°C in     Median increase of more than        Median increase of more than

Do Not Cite, Quote or Distribute                             11-148                                                    Total pages: 345
                              Final Government Distribution                                       Chapter11                                                           IPCC AR6 WGI

                           2013a; Yin et al., 2017; Zhou    Wang et al., 2017; Imada et     the 50-year TXx and TNn events        1°C in the 50-year TXx and          4°C in the 50-year TXx and
                           et al., 2016; Dunn et al.,       al., 2014, 2019; Kim et al.,    compared to the 1°C warming           TNn events compared to the          TNn events compared to the
                           2020)                            2018; Lu et al., 2016, 2018;    level (Li et al., 2020) and more      1°C warming level (Li et al.,       1°C warming level (Li et al.,
                                                            Takahashi et al., 2016; Ye      than 1.5°C in annual TXx and          2020) and more than 2°C in          2020) and more than 4.5°C in
                                                            and Li, 2017; Zhou et al.,      TNn compared to pre-industrial        annual TXx and TNn compared         annual TXx and TNn compared
                                                            2016)                           (Annex).                              to pre-industrial (Annex).          to pre-industrial (Annex).

                                                                                            Additional evidence from              Additional evidence from            Additional evidence from
                                                                                            CMIP5 and RCM simulations for         CMIP5 and RCM simulations           CMIP5 and RCM simulations
                                                                                            an increase in the intensity and      for an increase in the intensity    for an increase in the intensity
                                                                                                                                  and frequency of hot extremes       and frequency of hot extremes
                                                                                            frequency of hot extremes and
                                                                                                                                  and decrease in the intensity       and decrease in the intensity
                                                                                            decrease in the intensity and         and frequency of cold extremes      and frequency of cold extremes
                                                                                            frequency of cold extremes (Guo       (Guo et al., 2018; Imada et al.,    (Guo et al., 2018; Imada et al.,
                                                                                            et al., 2018; Imada et al., 2019;     2019; Li et al., 2018c; Seo et      2019; Li et al., 2018c; Seo et
                                                                                            Li et al., 2018c; Seo et al., 2014;   al., 2014; Sui et al., 2018; Wang   al., 2014; Sui et al., 2018; Wang
                                                                                            Sui et al., 2018; Wang et al.,        et al., 2017a, 2017c; Xu et al.,    et al., 2017a, 2017c; Xu et al.,
                                                                                            2017a, 2017c; Xu et al., 2016a;       2016a; Zhou et al., 2014; Shi et    2016a; Zhou et al., 2014; Shi et
                                                                                                                                  al., 2018; Sun et al., 2019a)       al., 2018; Sun et al., 2019a)
                                                                                            Zhou et al., 2014; Shi et al.,
                                                                                            2018; Sun et al., 2019a)
                           Very likely increase in the      Human influence likely          Increase in the intensity and         Increase in the intensity and       Increase in the intensity and
                           intensity and frequency of hot   contributed to the observed     frequency of hot extremes:            frequency of hot extremes:          frequency of hot extremes:
                           extremes and decrease in the     increase in the intensity and   Likely (compared with the recent      Very likely (compared with the      Virtually certain (compared
                           intensity and frequency of       frequency of hot extremes       past (1995-2014))                     recent past (1995-2014))            with the recent past (1995-
                           cold extremes                    and decrease in the intensity   Very likely (compared with pre-       Extremely likely (compared          2014))
                                                            and frequency of cold           industrial)                           with pre-industrial)                Virtually certain (compared
                                                            extremes                                                                                                  with pre-industrial)
                                                                                            Decrease in the intensity and         Decrease in the intensity and
                                                                                            frequency of cold extremes:           frequency of cold extremes:         Decrease in the intensity and
                                                                                            Likely (compared with the recent      Very likely (compared with the      frequency of cold extremes:
                                                                                            past (1995-2014))                     recent past (1995-2014))            Virtually certain (compared
                                                                                            Very likely (compared with pre-       Extremely likely (compared          with the recent past (1995-
                                                                                            industrial).                          with pre-industrial)                2014))
                                                                                                                                                                      Virtually certain (compared
                                                                                                                                                                      with pre-industrial)
 East Central Asia (ECA)   Significant increases in the     Robust evidence of a human      CMIP6 models project a robust         CMIP6 models project a robust       CMIP6 models project a robust
                           intensity and frequency of hot   contribution to the observed    increase in the intensity and         increase in the intensity and       increase in the intensity and
                           extremes and significant         increase in the intensity and   frequency of TXx events and a         frequency of TXx events and a       frequency of TXx events and a
                           decreases in the intensity and   frequency of hot extremes       robust decrease in the intensity      robust decrease in the intensity    robust decrease in the intensity
                           frequency of cold extremes       and decrease in the intensity   and frequency of TNn events (Li       and frequency of TNn events         and frequency of TNn events
                           (Dunn et al., 2020)              and frequency of cold           et al., 2020; Annex). Median          (Li et al., 2020; Annex).           (Li et al., 2020; Annex).
                                                            extremes (Seong et al., 2020;   increase of more than 0.5°C in        Median increase of more than        Median increase of more than
                                                            Wang et al., 2017; Dong et      the 50-year TXx and TNn events        1°C in the 50-year TXx and          3.5°C in the 50-year TXx and
                                                            al., 2018; Kim et al., 2019)    compared to the 1°C warming           TNn events compared to the          TNn events compared to the
                                                                                            level (Li et al., 2020) and more      1°C warming level (Li et al.,       1°C warming level (Li et al.,
                                                                                            than 2°C in annual TXx and TNn        2020) and more than 2.5°C in        2020) and more than 5.5°C in

Do Not Cite, Quote or Distribute                             11-149                                                      Total pages: 345
                             Final Government Distribution                                          Chapter11                                                       IPCC AR6 WGI

                                                                                              compared to pre-industrial         annual TXx and TNn compared annual TXx and TNn compared
                                                                                              (Annex).                           to pre-industrial (Annex).  to pre-industrial (Annex).

                                                                                              Additional evidence from           Additional evidence from           Additional evidence from
                                                                                              CMIP5 simulations for an           CMIP5 simulations for an           CMIP5 simulations for an
                                                                                              increase in the intensity and      increase in the intensity and      increase in the intensity and
                                                                                              frequency of hot extremes and      frequency of hot extremes and      frequency of hot extremes and
                                                                                              decrease in the intensity and      decrease in the intensity and      decrease in the intensity and
                                                                                              frequency of cold extremes (Han    frequency of cold extremes         frequency of cold extremes
                                                                                              et al., 2018)                      (Han et al., 2018)                 (Han et al., 2018)
                          Very likely increase in the        High confidence in a human       Increase in the intensity and      Increase in the intensity and      Increase in the intensity and
                          intensity and frequency of hot     contribution to the observed     frequency of hot extremes:         frequency of hot extremes:         frequency of hot extremes:
                          extremes and decrease in the       increase in the intensity and    Likely (compared with the recent   Very likely (compared with the     Virtually certain (compared
                          intensity and frequency of         frequency of hot extremes        past (1995-2014))                  recent past (1995-2014))           with the recent past (1995-
                          cold extremes                      and decrease in the intensity    Very likely (compared with pre-    Extremely likely (compared         2014))
                                                             and frequency of cold            industrial)                        with pre-industrial)               Virtually certain (compared
                                                             extremes.                                                                                              with pre-industrial)
                                                                                              Decrease in the intensity and      Decrease in the intensity and
                                                                                              frequency of cold extremes:        frequency of cold extremes:        Decrease in the intensity and
                                                                                              Likely (compared with the recent   Very likely (compared with the     frequency of cold extremes:
                                                                                              past (1995-2014))                  recent past (1995-2014))           Virtually certain (compared
                                                                                              Very likely (compared with pre-    Extremely likely (compared         with the recent past (1995-
                                                                                              industrial).                       with pre-industrial)               2014))
                                                                                                                                                                    Virtually certain (compared
                                                                                                                                                                    with pre-industrial)
 Tibetan Plateau (TIB)    Significant increases in the       Robust evidence of a human       CMIP6 models project a robust      CMIP6 models project a robust      CMIP6 models project a robust
                          intensity and frequency of hot     contribution to the observed     increase in the intensity and      increase in the intensity and      increase in the intensity and
                          extremes and significant           increase in the intensity and    frequency of TXx events and a      frequency of TXx events and a      frequency of TXx events and a
                          decreases in the intensity and     frequency of hot extremes        robust decrease in the intensity   robust decrease in the intensity   robust decrease in the intensity
                          frequency of cold extremes         and decrease in the intensity    and frequency of TNn events (Li    and frequency of TNn events        and frequency of TNn events
                          (Donat et al., 2016a; Hu et al.,   and frequency of cold            et al., 2020; Annex). Median       (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                          2016; Sun et al., 2017; Yin et     extremes (Seong et al., 2020;    increase of more than 0.5°C in     Median increase of more than       Median increase of more than
                          al., 2019; Zhang et al., 2019c;    Wang et al., 2017; Yin et al.,   the 50-year TXx and TNn events     1°C in the 50-year TXx and         4°C in the 50-year TXx and
                          Dunn et al., 2020)                 2019)                            compared to the 1°C warming        TNn events compared to the         TNn events compared to the
                                                                                              level (Li et al., 2020) and more   1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                                                              than 1.5°C in annual TXx and       2020) and more than 2°C in         2020) and more than 4.5°C in
                                                                                              TNn compared to pre-industrial     annual TXx and TNn compared        annual TXx and TNn compared
                                                                                              (Annex).                           to pre-industrial (Annex).         to pre-industrial (Annex).

                                                                                              Additional evidence from           Additional evidence from           Additional evidence from
                                                                                              CMIP5 and RCM simulations for      CMIP5 and RCM simulations          CMIP5 and RCM simulations
                                                                                              an increase in the intensity and   for an increase in the intensity   for an increase in the intensity
                                                                                              frequency of hot extremes and      and frequency of hot extremes      and frequency of hot extremes
                                                                                              decrease in the intensity and      and decrease in the intensity      and decrease in the intensity
                                                                                              frequency of cold extremes         and frequency of cold extremes     and frequency of cold extremes
                                                                                              (Zhou et al., 2014; Singh and      (Zhou et al., 2014; Singh and      (Zhou et al., 2014; Singh and
                                                                                              Goyal, 2016; Zhang et al.,         Goyal, 2016; Zhang et al.,         Goyal, 2016; Zhang et al.,

Do Not Cite, Quote or Distribute                              11-150                                                       Total pages: 345
                             Final Government Distribution                                       Chapter11                                                          IPCC AR6 WGI

                                                                                           2016a; Xu et al., 2017; Han et       2016a; Xu et al., 2017; Han et      2016a; Xu et al., 2017; Han et
                                                                                           al., 2018; Li et al., 2018a)         al., 2018; Li et al., 2018a)        al., 2018; Li et al., 2018a)
                          Very likely increase in the      High confidence in a human      Increase in the intensity and        Increase in the intensity and       Increase in the intensity and
                          intensity and frequency of hot   contribution to the observed    frequency of hot extremes:           frequency of hot extremes:          frequency of hot extremes:
                          extremes and decrease in the     increase in the intensity and   Likely (compared with the recent     Very likely (compared with the      Virtually certain (compared
                          intensity and frequency of       frequency of hot extremes       past (1995-2014))                    recent past (1995-2014))            with the recent past (1995-
                          cold extremes                    and decrease in the intensity   Very likely (compared with pre-      Extremely likely (compared          2014))
                                                           and frequency of cold           industrial)                          with pre-industrial)                Virtually certain (compared
                                                           extremes.                                                                                                with pre-industrial)
                                                                                           Decrease in the intensity and        Decrease in the intensity and
                                                                                           frequency of cold extremes:          frequency of cold extremes:         Decrease in the intensity and
                                                                                           Likely (compared with the recent     Very likely (compared with the      frequency of cold extremes:
                                                                                           past (1995-2014))                    recent past (1995-2014))            Virtually certain (compared
                                                                                           Very likely (compared with pre-      Extremely likely (compared          with the recent past (1995-
                                                                                           industrial).                         with pre-industrial)                2014))
                                                                                                                                                                    Virtually certain (compared
                                                                                                                                                                    with pre-industrial)
 South Asia (SAS)         Significant increases in the     Robust evidence of a human      CMIP6 models project a robust        CMIP6 models project a robust       CMIP6 models project a robust
                          intensity and frequency of hot   contribution to the observed    increase in the intensity and        increase in the intensity and       increase in the intensity and
                          extremes and significant         increase in the intensity and   frequency of TXx events and a        frequency of TXx events and a       frequency of TXx events and a
                          decreases in the intensity and   frequency of hot extremes       robust decrease in the intensity     robust decrease in the intensity    robust decrease in the intensity
                          frequency of cold extremes       and decrease in the intensity   and frequency of TNn events (Li      and frequency of TNn events         and frequency of TNn events
                          (Chakraborty et al., 2018;       and frequency of cold           et al., 2020; Annex). Median         (Li et al., 2020; Annex).           (Li et al., 2020; Annex).
                          Dimri, 2019; Donat et al.,       extremes (Seong et al., 2020;   increase of more than 0C in the      Median increase of more than        Median increase of more than
                          2016; Dunn et al., 2020; Roy,    Wang et al., 2017; Wehner et    50-year TXx and TNn events           1°C in the 50-year TXx and          3.5°C in the 50-year TXx and
                          2019; Sheikh et al., 2015;       al., 2016; Kumar, 2017; van     compared to the 1°C warming          TNn events compared to the          TNn events compared to the
                          Rohini et al., 2016; Zahid and   Oldenborgh et al., 2018)        level (Li et al., 2020) and more     1°C warming level (Li et al.,       1°C warming level (Li et al.,
                          Rasul, 2012)                                                     than 1°C in annual TXx and TNn       2020) and more than 1.5°C in        2020) and more than 4°C in
                                                                                           compared to pre-industrial           annual TXx and TNn compared         annual TXx and TNn compared
                                                                                           (Annex).                             to pre-industrial (Annex).          to pre-industrial (Annex).

                                                                                           Additional evidence from             Additional evidence from            Additional evidence from
                                                                                           CMIP5 and RCM simulations for        CMIP5 and RCM simulations           CMIP5 and RCM simulations
                                                                                           an increase in the intensity and     for an increase in the intensity    for an increase in the intensity
                                                                                           frequency of hot extremes and        and frequency of hot extremes       and frequency of hot extremes
                                                                                           decrease in the intensity and        and decrease in the intensity       and decrease in the intensity
                                                                                           frequency of cold extremes (Ali      and frequency of cold extremes      and frequency of cold extremes
                                                                                           et al., 2019; Han et al., 2018;      (Ali et al., 2019; Han et al.,      (Ali et al., 2019; Han et al.,
                                                                                           Kharin et al., 2018; Sillmann et     2018; Kharin et al., 2018;          2018; Kharin et al., 2018;
                                                                                           al., 2013; Xu et al., 2017; Murari   Sillmann et al., 2013; Xu et al.,   Sillmann et al., 2013; Xu et al.,
                                                                                           et al., 2015; Nasim et al., 2018)    2017; Murari et al., 2015;          2017; Murari et al., 2015;
                                                                                                                                Nasim et al., 2018)                 Nasim et al., 2018)
                          High confidence in the           High confidence in a human      Increase in the intensity and        Increase in the intensity and       Increase in the intensity and
                          increase in the intensity and    contribution to the observed    frequency of hot extremes:           frequency of hot extremes:          frequency of hot extremes:
                          frequency of hot extremes        increase in the intensity and   Likely (compared with the recent     Very likely (compared with the      Virtually certain (compared
                          and decrease in the intensity    frequency of hot extremes       past (1995-2014))                    recent past (1995-2014))            with the recent past (1995-
                                                           and decrease in the intensity   Very likely (compared with pre-      Extremely likely (compared          2014))
Do Not Cite, Quote or Distribute                            11-151                                                     Total pages: 345
                             Final Government Distribution                                          Chapter11                                                         IPCC AR6 WGI

                          and frequency of cold             and frequency of cold             industrial)                          with pre-industrial)               Virtually certain (compared
                          extremes                          extremes.                                                                                                 with pre-industrial)
                                                                                              Decrease in the intensity and        Decrease in the intensity and
                                                                                              frequency of cold extremes:          frequency of cold extremes:        Decrease in the intensity and
                                                                                              Likely (compared with the recent     Very likely (compared with the     frequency of cold extremes:
                                                                                              past (1995-2014))                    recent past (1995-2014))           Virtually certain (compared
                                                                                              Very likely (compared with pre-      Extremely likely (compared         with the recent past (1995-
                                                                                              industrial).                         with pre-industrial)               2014))
                                                                                                                                                                      Virtually certain (compared
                                                                                                                                                                      with pre-industrial)
 Southeast Asia (SEA)     Significant increases in the      Robust evidence of a human        CMIP6 models project a robust        CMIP6 models project a robust      CMIP6 models project a robust
                          intensity and frequency of hot    contribution to the observed      increase in the intensity and        increase in the intensity and      increase in the intensity and
                          extremes and significant          increase in the intensity and     frequency of TXx events and a        frequency of TXx events and a      frequency of TXx events and a
                          decreases in the intensity and    frequency of hot extremes         robust decrease in the intensity     robust decrease in the intensity   robust decrease in the intensity
                          frequency of cold extremes        and decrease in the intensity     and frequency of TNn events (Li      and frequency of TNn events        and frequency of TNn events
                          (Donat et al., 2016a; Supari et   and frequency of cold             et al., 2020; Annex). Median         (Li et al., 2020; Annex).          (Li et al., 2020; Annex).
                          al., 2017; Cheong et al., 2018;   extremes (Seong et al., 2020;     increase of more than 0C in the      Median increase of more than       Median increase of more than
                          Zhang et al., 2019c; Dunn et      Wang et al., 2017; King et al.,   50-year TXx and TNn events           0.5°C in the 50-year TXx and       2.5°C in the 50-year TXx and
                          al., 2020)                        2016; Min et al., 2020)           compared to the 1°C warming          TNn events compared to the         TNn events compared to the
                                                                                              level (Li et al., 2020) and more     1°C warming level (Li et al.,      1°C warming level (Li et al.,
                                                                                              than 1°C in annual TXx and TNn       2020) and more than 1.5°C in       2020) and more than 4°C in
                                                                                              compared to pre-industrial           annual TXx and TNn compared        annual TXx and TNn compared
                                                                                              (Annex).                             to pre-industrial (Annex).         to pre-industrial (Annex)

                                                                                              Additional evidence from             Additional evidence from           Additional evidence from
                                                                                              CMIP5 simulations for an             CMIP5 simulations for an           CMIP5 simulations for an
                                                                                              increase in the intensity and        increase in the intensity and      increase in the intensity and
                                                                                                                                   frequency of hot extremes and      frequency of hot extremes and
                                                                                              frequency of hot extremes and
                                                                                                                                   decrease in the intensity and      decrease in the intensity and
                                                                                              decrease in the intensity and        frequency of cold extremes         frequency of cold extremes
                                                                                              frequency of cold extremes (Han      (Kharin et al., 2018; Xu et al.,   (Kharin et al., 2018; Xu et al.,
                                                                                              et al., 2018; Kharin et al., 2018;   2017).                             2017)..
                                                                                              Xu et al., 2017)
                          High confidence in the            High confidence in a human        Increase in the intensity and        Increase in the intensity and      Increase in the intensity and
                          increase in the intensity and     contribution to the observed      frequency of hot extremes:           frequency of hot extremes:         frequency of hot extremes:
                          frequency of hot extremes         increase in the intensity and     Likely (compared with the recent     Very likely (compared with the     Virtually certain (compared
                          and decrease in the intensity     frequency of hot extremes         past (1995-2014))                    recent past (1995-2014))           with the recent past (1995-
                          and frequency of cold             and decrease in the intensity     Very likely (compared with pre-      Extremely likely (compared         2014))
                          extremes                          and frequency of cold             industrial)                          with pre-industrial)               Virtually certain (compared
                                                            extremes.                                                                                                 with pre-industrial)
                                                                                              Decrease in the intensity and        Decrease in the intensity and
                                                                                              frequency of cold extremes:          frequency of cold extremes:        Decrease in the intensity and
                                                                                              Likely (compared with the recent     Very likely (compared with the     frequency of cold extremes:
                                                                                              past (1995-2014))                    recent past (1995-2014))           Virtually certain (compared
                                                                                              Very likely (compared with pre-      Extremely likely (compared         with the recent past (1995-
                                                                                              industrial).                         with pre-industrial)               2014))
                                                                                                                                                                      Virtually certain (compared

Do Not Cite, Quote or Distribute                             11-152                                                        Total pages: 345
                                  Final Government Distribution                                      Chapter11                                                        IPCC AR6 WGI

                                                                                                                                                                      with pre-industrial)

1   [END TABLE 11.7 HERE]
2
3
4   [START TABLE 11.8 HERE]
5
6   Table 11.8: Observed trends, human contribution to observed trends, and projected changes at 1.5°C, 2°C and 4°C of global warming for heavy precipitation in Asia, subdivided
7               by AR6 regions. See Sections 11.9.1 and 11.9.3 for details
                                                                   Detection and attribution;
                Region                Observed trends                                                                                        Projections
                                                                       event attribution
                                                                                                              1.5 °C                              2 °C                              4 °C
     All Asia                  Significant intensification of    Robust evidence of a human      CMIP6 models project a             CMIP6 models project a           CMIP6 models project a
                               heavy precipitation (Sun et       contribution to the observed    robust increase in the             robust increase in the           robust increase in the
                               al., 2020)                        intensification of heavy        intensity and frequency of         intensity and frequency of       intensity and frequency of
                                                                 precipitation                   heavy precipitation (Li et al.,    heavy precipitation (Li et al.,  heavy precipitation ((Li et al.,
                                                                                                 2020a). Median increase of         2020a). Median increase of       2020a). Median increase of
                                                                                                 more than 2% in the 50-year        more than 6% in the 50-year      more than 15% in the 50-year
                                                                                                 Rx1day and Rx5day events           Rx1day and Rx5day events         Rx1day and Rx5day events
                                                                                                 compared to the 1°C warming        compared to the 1°C warming      compared to the 1°C warming
                                                                                                 level (Li et al., 2020a)           level (Li et al., 2020a)         level (Li et al., 2020a)
                               Likely intensification of heavy   Human influence likely          Intensification of heavy           Intensification of heavy         Intensification of heavy
                               precipitation                     contributed to the observed     precipitation:                     precipitation:                   precipitation:
                                                                 intensification of heavy        Likely (compared with the          Very likely (compared with       Virtually certain (compared
                                                                 precipitation                   recent past (1995-2014))           the recent past (1995-2014))     with the recent past (1995-
                                                                                                 Very likely (compared with         Extremely likely (compared       2014))
                                                                                                 pre-industrial)                    with pre-industrial)             Virtually certain (compared
                                                                                                                                                                     with pre-industrial)
     Russian Arctic (RAR)      Insufficient data and a lack of   Limited evidence               CMIP6 models project a robust CMIP6 models project a robust CMIP6 models project a robust
                               agreement on the evidence of                                     increase in the intensity and     increase in the intensity and     increase in the intensity and
                               trends (Sun et al., 2020; Dunn                                   frequency of heavy                frequency of heavy                frequency of heavy
                               et al., 2020)                                                    precipitation (Li et al., 2020;   precipitation (Li et al., 2020;   precipitation (Li et al., 2020;
                                                                                                Annex). Median increase of        Annex). Median increase of        Annex). Median increase of
                                                                                                more than 4% in the 50-year       more than 8% in the 50-year       more than 25% in the 50-year
                                                                                                Rx1day and Rx5day events          Rx1day and Rx5day events          Rx1day and Rx5day events
                                                                                                compared to the 1°C warming compared to the 1°C warming compared to the 1°C warming
                                                                                                level (Li et al., 2020a) and more level (Li et al., 2020a) and more level (Li et al., 2020a) and more
                                                                                                than 10% in annual Rx1day and than 10% in annual Rx1day,            than 25% in annual Rx1day and
                                                                                                Rx5day and 8% in annual           Rx5day, and Rx30day               Rx5day and 20% in annual
                                                                                                Rx30day compared to pre-          compared to pre-industrial        Rx30day compared to pre-
                                                                                                industrial (Annex).               (Annex).                          industrial (Annex).

                                                                                                Additional evidence from           Additional evidence from           Additional evidence from
                                                                                                CMIP5 simulations for an           CMIP5 simulations for an           CMIP5 simulations for an
                                                                                                increase in the intensity of       increase in the intensity of       increase in the intensity of
                                                                                                heavy precipitation (Sillmann et   heavy precipitation (Sillmann et   heavy precipitation (Sillmann et
    Do Not Cite, Quote or Distribute                              11-153                                                    Total pages: 345
                              Final Government Distribution                           Chapter11                                                          IPCC AR6 WGI

                                                                                 al., 2013b; Han et al., 2018;       al., 2013b; Han et al., 2018;       al., 2013b; Han et al., 2018;
                                                                                 Kharin et al., 2018;                Kharin et al., 2018;                Kharin et al., 2018;
                                                                                 Khlebnikova et al., 2019b)          Khlebnikova et al., 2019b)          Khlebnikova et al., 2019b)
                           Low confidence                     Low confidence     Intensification of heavy            Intensification of heavy            Intensification of heavy
                                                                                 precipitation:                      precipitation:                      precipitation:
                                                                                 Likely (compared with the           Very likely (compared with the      Virtually certain (compared
                                                                                 recent past (1995-2014))            recent past (1995-2014))            with the recent past (1995-
                                                                                 Very likely (compared with pre-     Extremely likely (compared          2014))
                                                                                 industrial)                         with pre-industrial)                Virtually certain (compared
                                                                                                                                                         with pre-industrial)
 Arabian Peninsula (ARP)   Insufficient data and a lack of    Limited evidence                                       CMIP6 models project an             CMIP6 models project a robust
                           agreement on the evidence of                                                              increase in the intensity and       increase in the intensity and
                           trends (Sun et al., 2020; Dunn                                                            frequency of heavy                  frequency of heavy
                           et al., 2020; Atif et al., 2020;                                                          precipitation (Li et al., 2020;     precipitation (Li et al., 2020;
                           Donat et al., 2014; Rahimi                                                                Annex). Median increase of          Annex). Median increase of
                           and Fatemi, 2019)                                                                         more than 8% in the 50-year         more than 20% in the 50-year
                                                                                                                     Rx1day and Rx5day events            Rx1day and Rx5day events
                                                                                                                     compared to the 1°C warming         compared to the 1°C warming
                                                                                                                     level (Li et al., 2020a) and more   level (Li et al., 2020a) and more
                                                                                                                     than 15% in annual Rx1day,          than 40% in annual Rx1day and
                                                                                                                     Rx5day, and Rx30day                 Rx5day and 45% in annual
                                                                                                                     compared to pre-industrial          Rx30day compared to pre-
                                                                                                                     (Annex).                            industrial (Annex).
                           Low confidence                     Low confidence     Intensification of heavy            Intensification of heavy            Intensification of heavy
                                                                                 precipitation:                      precipitation:                      precipitation:
                                                                                 Low confidence (compared with       Medium confidence (compared         Likely (compared with the
                                                                                 the recent past (1995-2014))        with the recent past (1995-         recent past (1995-2014))
                                                                                 Medium confidence (compared         2014))                              Very likely (compared with pre-
                                                                                 with pre-industrial)                High confidence (compared           industrial)
                                                                                                                     with pre-industrial)
 West Central Asia (WCA)   Intensification of heavy           Limited evidence   CMIP6 models project a robust       CMIP6 models project a robust       CMIP6 models project a robust
                           precipitation (Sun et al.,                            increase in the intensity and       increase in the intensity and       increase in the intensity and
                           2020; Hu et al., 2016; Zhang                          frequency of heavy                  frequency of heavy                  frequency of heavy
                           et al., 2017).                                        precipitation (Li et al., 2020;     precipitation (Li et al., 2020;     precipitation (Li et al., 2020;
                                                                                 Annex). Median increase of          Annex). Median increase of          Annex). Median increase of
                                                                                 more than 2% in the 50-year         more than 6% in the 50-year         more than 15% in the 50-year
                                                                                 Rx1day and Rx5day events            Rx1day and Rx5day events            Rx1day and Rx5day events
                                                                                 compared to the 1°C warming         compared to the 1°C warming         compared to the 1°C warming
                                                                                 level (Li et al., 2020a) and more   level (Li et al., 2020a) and more   level (Li et al., 2020a) and more
                                                                                 than 6% in annual Rx1day and        than 8% in annual Rx1day and        than 20% in annual Rx1day and
                                                                                 Rx5day and 4% in annual             Rx5day and 6% in annual             Rx5day and 15% in annual
                                                                                 Rx30day compared to pre-            Rx30day compared to pre-            Rx30day compared to pre-
                                                                                 industrial (Annex).                 industrial (Annex).                 industrial (Annex).

                                                                                 Additional evidence from            Additional evidence from            Additional evidence from
                                                                                 CMIP5 simulations for an            CMIP5 simulations for an            CMIP5 simulations for an
                                                                                 increase in the intensity of        increase in the intensity of        increase in the intensity of
Do Not Cite, Quote or Distribute                               11-154                                           Total pages: 345
                             Final Government Distribution                         Chapter11                                                            IPCC AR6 WGI

                                                                              heavy precipitation (Han et al.,     heavy precipitation (Han et al.,     heavy precipitation (Han et al.,
                                                                              2018)                                2018)                                2018)
                          Medium confidence in the         Low confidence     Intensification of heavy             Intensification of heavy             Intensification of heavy
                          intensitification of heavy                          precipitation:                       precipitation:                       precipitation:
                          precipitation                                       Likely (compared with the            Very likely (compared with the       Virtually certain (compared
                                                                              recent past (1995-2014))             recent past (1995-2014))             with the recent past (1995-
                                                                              Very likely (compared with pre-      Extremely likely (compared           2014))
                                                                              industrial)                          with pre-industrial)                 Virtually certain (compared
                                                                                                                                                        with pre-industrial)
 West Siberia (WSB)       Significant intensification of   Limited evidence   CMIP6 models project a robust        CMIP6 models project a robust        CMIP6 models project a robust
                          heavy precipitation (Sun et                         increase in the intensity and        increase in the intensity and        increase in the intensity and
                          al., 2020; Zhang et al., 2017)                      frequency of heavy                   frequency of heavy                   frequency of heavy
                                                                              precipitation (Li et al., 2020;      precipitation (Li et al., 2020;      precipitation (Li et al., 2020;
                                                                              Annex). Median increase of           Annex). Median increase of           Annex). Median increase of
                                                                              more than 2% in the 50-year          more than 4% in the 50-year          more than 15% in the 50-year
                                                                              Rx1day and Rx5day events             Rx1day and Rx5day events             Rx1day and Rx5day events
                                                                              compared to the 1°C warming          compared to the 1°C warming          compared to the 1°C warming
                                                                              level (Li et al., 2020a) and more    level (Li et al., 2020a) and more    level (Li et al., 2020a) and more
                                                                              than 6% in annual Rx1day,            than 8% in annual Rx1day and         than 15% in annual Rx1day,
                                                                              Rx5day, and Rx30day                  Rx5day and 6% in annual              Rx5day, and Rx30day
                                                                              compared to pre-industrial           Rx30day compared to pre-             compared to pre-industrial
                                                                              (Annex).                             industrial (Annex).                  (Annex).

                                                                              Additional evidence from             Additional evidence from             Additional evidence from
                                                                              CMIP5 simulations for an             CMIP5 simulations for an             CMIP5 simulations for an
                                                                              increase in the intensity of         increase in the intensity of         increase in the intensity of
                                                                              heavy precipitation (Sillmann et     heavy precipitation (Sillmann et     heavy precipitation (Sillmann et
                                                                              al., 2013b; Xu et al., 2017; Han     al., 2013b; Xu et al., 2017; Han     al., 2013b; Xu et al., 2017; Han
                                                                              et al., 2018; Kharin et al., 2018;   et al., 2018; Kharin et al., 2018;   et al., 2018; Kharin et al., 2018;
                                                                              Khlebnikova et al., 2019b)           Khlebnikova et al., 2019b)           Khlebnikova et al., 2019b)
                          High confidence in the           Low confidence     Intensification of heavy             Intensification of heavy             Intensification of heavy
                          intensitification of heavy                          precipitation:                       precipitation:                       precipitation:
                          precipitation                                       Likely (compared with the            Very likely (compared with the       Virtually certain (compared
                                                                              recent past (1995-2014))             recent past (1995-2014))             with the recent past (1995-
                                                                              Very likely (compared with pre-      Extremely likely (compared           2014))
                                                                              industrial)                          with pre-industrial)                 Virtually certain (compared
                                                                                                                                                        with pre-industrial)
 East Siberia (ESB)       Intensification of heavy         Limited evidence   CMIP6 models project a robust        CMIP6 models project a robust        CMIP6 models project a robust
                          precipitation (Knutson and                          increase in the intensity and        increase in the intensity and        increase in the intensity and
                          Zeng, 2018; Sun et al., 2020;                       frequency of heavy                   frequency of heavy                   frequency of heavy
                          Dunn et al., 2020)                                  precipitation (Li et al., 2020;      precipitation (Li et al., 2020;      precipitation (Li et al., 2020;
                                                                              Annex). Median increase of           Annex). Median increase of           Annex). Median increase of
                                                                              more than 2% in the 50-year          more than 4% in the 50-year          more than 20% in the 50-year
                                                                              Rx1day and Rx5day events             Rx1day and Rx5day events             Rx1day and Rx5day events
                                                                              compared to the 1°C warming          compared to the 1°C warming          compared to the 1°C warming
                                                                              level (Li et al., 2020a) and more    level (Li et al., 2020a) and more    level (Li et al., 2020a) and more
                                                                              than 6% in annual Rx1day and         than 8% in annual Rx1day and         than 20% in annual Rx1day and
Do Not Cite, Quote or Distribute                            11-155                                         Total pages: 345
                             Final Government Distribution                               Chapter11                                                            IPCC AR6 WGI

                                                                                    Rx5day and 4% in annual              Rx5day and 6% in annual              Rx5day and 15% in annual
                                                                                    Rx30day compared to pre-             Rx30day compared to pre-             Rx30day compared to pre-
                                                                                    industrial (Annex).                  industrial (Annex).                  industrial (Annex).

                                                                                    Additional evidence from             Additional evidence from             Additional evidence from
                                                                                    CMIP5 simulations for an             CMIP5 simulations for an             CMIP5 simulations for an
                                                                                    increase in the intensity of         increase in the intensity of         increase in the intensity of
                                                                                    heavy precipitation (Sillmann et     heavy precipitation (Sillmann et     heavy precipitation (Sillmann et
                                                                                    al., 2013b; Xu et al., 2017; Han     al., 2013b; Xu et al., 2017; Han     al., 2013b; Xu et al., 2017; Han
                                                                                    et al., 2018; Kharin et al., 2018;   et al., 2018; Kharin et al., 2018;   et al., 2018; Kharin et al., 2018;
                                                                                    Khlebnikova et al., 2019b)           Khlebnikova et al., 2019b)           Khlebnikova et al., 2019b)
                          Medium confidence in the     Low confidence               Intensification of heavy             Intensification of heavy             Intensification of heavy
                          intensitification of heavy                                precipitation:                       precipitation:                       precipitation:
                          precipitation                                             Likely (compared with the            Very likely (compared with the       Virtually certain (compared
                                                                                    recent past (1995-2014))             recent past (1995-2014))             with the recent past (1995-
                                                                                    Very likely (compared with pre-      Extremely likely (compared           2014))
                                                                                    industrial)                          with pre-industrial)                 Virtually certain (compared
                                                                                                                                                              with pre-industrial)
 Russian Far East (RFE)   Intensification of heavy     Limited evidence             CMIP6 models project a robust        CMIP6 models project a robust        CMIP6 models project a robust
                          precipitation (Sun et al.,                                increase in the intensity and        increase in the intensity and        increase in the intensity and
                          2020)                                                     frequency of heavy                   frequency of heavy                   frequency of heavy
                                                                                    precipitation (Li et al., 2020;      precipitation (Li et al., 2020;      precipitation (Li et al., 2020;
                                                                                    Annex). Median increase of           Annex). Median increase of           Annex). Median increase of
                                                                                    more than 4% in the 50-year          more than 8% in the 50-year          more than 25% in the 50-year
                                                                                    Rx1day and Rx5day events             Rx1day and Rx5day events             Rx1day and Rx5day events
                                                                                    compared to the 1°C warming          compared to the 1°C warming          compared to the 1°C warming
                                                                                    level (Li et al., 2020a) and more    level (Li et al., 2020a) and more    level (Li et al., 2020a) and more
                                                                                    than 8% in annual Rx1day and         than 10% in annual Rx1day,           than 25% in annual Rx1day and
                                                                                    Rx5day and 6% in annual              Rx5day, and Rx30day                  Rx5day and 20% in annual
                                                                                    Rx30day compared to pre-             compared to pre-industrial           Rx30day compared to pre-
                                                                                    industrial (Annex).                  (Annex).                             industrial (Annex).

                                                                                    Additional evidence from             Additional evidence from             Additional evidence from
                                                                                    CMIP5 simulations for an             CMIP5 simulations for an             CMIP5 simulations for an
                                                                                    increase in the intensity of         increase in the intensity of         increase in the intensity of
                                                                                    heavy precipitation (Sillmann et     heavy precipitation (Sillmann et     heavy precipitation (Sillmann et
                                                                                    al., 2013b; Xu et al., 2017; Han     al., 2013b; Xu et al., 2017; Han     al., 2013b; Xu et al., 2017; Han
                                                                                    et al., 2018; Kharin et al., 2018)   et al., 2018; Kharin et al., 2018)   et al., 2018; Kharin et al., 2018)
                          Medium confidence in the     Low confidence               Intensification of heavy             Intensification of heavy             Intensification of heavy
                          intensitification of heavy                                precipitation:                       precipitation:                       precipitation:
                          precipitation                                             Likely (compared with the            Very likely (compared with the       Virtually certain (compared
                                                                                    recent past (1995-2014))             recent past (1995-2014))             with the recent past (1995-
                                                                                    Very likely (compared with pre-      Extremely likely (compared           2014))
                                                                                    industrial)                          with pre-industrial)                 Virtually certain (compared
                                                                                                                                                              with pre-industrial)
 East Asia (EAS)          Intensification of heavy     Disagreement among studies   CMIP6 models project an              CMIP6 models project a robust CMIP6 models project a robust
                          precipitation (Sun et al.,   (Chen and Sun, 2017; Li et   increase in the intensity and        increase in the intensity and increase in the intensity and
Do Not Cite, Quote or Distribute                        11-156                                                   Total pages: 345
                              Final Government Distribution                                      Chapter11                                                          IPCC AR6 WGI

                           2020; Dunn et al., 2020; Baek   al., 2017; Burke et al., 2016;   frequency of heavy                  frequency of heavy                  frequency of heavy
                           et al., 2017; Nayak et al.,     Zhou et al., 2013; Ma et al.,    precipitation (Li et al., 2020;     precipitation (Li et al., 2020;     precipitation (Li et al., 2020;
                           2017; Ye and Li, 2017; Zhou     2017)                            Annex). Median increase of          Annex). Median increase of          Annex). Median increase of
                           et al., 2016)                                                    more than 2% in the 50-year         more than 6% in the 50-year         more than 20% in the 50-year
                                                                                            Rx1day and Rx5day events            Rx1day and Rx5day events            Rx1day and Rx5day events
                                                                                            compared to the 1°C warming         compared to the 1°C warming         compared to the 1°C warming
                                                                                            level (Li et al., 2020a) and more   level (Li et al., 2020a) and more   level (Li et al., 2020a) and more
                                                                                            than 4% in annual Rx1day and        than 6% in annual Rx1day and        than 20% in annual Rx1day and
                                                                                            Rx5day and 0% in annual             Rx5day and 2% in annual             Rx5day and 10% in annual
                                                                                            Rx30day compared to pre-            Rx30day compared to pre-            Rx30day compared to pre-
                                                                                            industrial (Annex).                 industrial (Annex).                 industrial (Annex).

                                                                                            Additional evidence from            Additional evidence from            Additional evidence from
                                                                                            CMIP5 simulations for an            CMIP5 simulations for an            CMIP5 simulations for an
                                                                                            increase in the intensity of        increase in the intensity of        increase in the intensity of
                                                                                            heavy precipitation (Ahn et al.,    heavy precipitation (Ahn et al.,    heavy precipitation (Ahn et al.,
                                                                                            2016; Guo et al., 2018;             2016; Guo et al., 2018;             2016; Guo et al., 2018;
                                                                                            Hatsuzuka et al., 2020; Kawase      Hatsuzuka et al., 2020; Kawase      Hatsuzuka et al., 2020; Kawase
                                                                                            et al., 2019; Kim et al., 2018;     et al., 2019; Kim et al., 2018;     et al., 2019; Kim et al., 2018;
                                                                                            Kusunoki, 2018; Kusunoki and        Kusunoki, 2018; Kusunoki and        Kusunoki, 2018; Kusunoki and
                                                                                            Mizuta, 2013; Li et al., 2018a;     Mizuta, 2013; Li et al., 2018a;     Mizuta, 2013; Li et al., 2018a;
                                                                                            Nayak and Dairaku, 2016; Ohba       Nayak and Dairaku, 2016; Ohba       Nayak and Dairaku, 2016; Ohba
                                                                                            and Sugimoto, 2020, 2019; Seo       and Sugimoto, 2020, 2019; Seo       and Sugimoto, 2020, 2019; Seo
                                                                                            et al., 2014; Wang et al., 2017a,   et al., 2014; Wang et al., 2017a,   et al., 2014; Wang et al., 2017a,
                                                                                            2017b; Zhou et al., 2014; Li et     2017b; Zhou et al., 2014; Li et     2017b; Zhou et al., 2014; Li et
                                                                                            al., 2018b)                         al., 2018b)                         al., 2018b)
                           Medium confidence in the        Low confidence                   Intensification of heavy            Intensification of heavy            Intensification of heavy
                           intensitification of heavy                                       precipitation:                      precipitation:                      precipitation:
                           precipitation                                                    High confidence (compared           Likely (compared with the           Extremely likely (compared
                                                                                            with the recent past (1995-         recent past (1995-2014))            with the recent past (1995-
                                                                                            2014))                              Very likely (compared with pre-     2014))
                                                                                            Likely (compared with pre-          industrial)                         Virtually certain (compared
                                                                                            industrial)                                                             with pre-industrial)
 East Central Asia (ECA)   Intensification of heavy        Limited evidence                 CMIP6 models project a robust       CMIP6 models project a robust       CMIP6 models project a robust
                           precipitation (Sun et al.,                                       increase in the intensity and       increase in the intensity and       increase in the intensity and
                           2020)                                                            frequency of heavy                  frequency of heavy                  frequency of heavy
                                                                                            precipitation (Li et al., 2020;     precipitation (Li et al., 2020;     precipitation (Li et al., 2020;
                                                                                            Annex). Median increase of          Annex). Median increase of          Annex). Median increase of
                                                                                            more than 4% in the 50-year         more than 6% in the 50-year         more than 20% in the 50-year
                                                                                            Rx1day and Rx5day events            Rx1day and Rx5day events            Rx1day and Rx5day events
                                                                                            compared to the 1°C warming         compared to the 1°C warming         compared to the 1°C warming
                                                                                            level (Li et al., 2020a) and more   level (Li et al., 2020a) and more   level (Li et al., 2020a) and more
                                                                                            than 8% in annual Rx1day and        than 10% in annual Rx1day,          than 25% in annual Rx1day,
                                                                                            Rx5day and 6% in annual             Rx5day, and Rx30day                 Rx5day, and Rx30day
                                                                                            Rx30day compared to pre-            compared to pre-industrial          compared to pre-industrial
                                                                                            industrial (Annex).                 (Annex).                            (Annex).



Do Not Cite, Quote or Distribute                            11-157                                                      Total pages: 345
                             Final Government Distribution                                     Chapter11                                                          IPCC AR6 WGI

                          Medium confidence in the           Low confidence               Intensification of heavy            Intensification of heavy            Intensification of heavy
                          intensitification of heavy                                      precipitation:                      precipitation:                      precipitation:
                          precipitation                                                   Likely (compared with the           Very likely (compared with the      Virtually certain (compared
                                                                                          recent past (1995-2014))            recent past (1995-2014))            with the recent past (1995-
                                                                                          Very likely (compared with pre-     Extremely likely (compared          2014))
                                                                                          industrial)                         with pre-industrial)                Virtually certain (compared
                                                                                                                                                                  with pre-industrial)
 Tibetan Plateau (TIB)    Intensification of heavy           Limited evidence             CMIP6 models project a robust       CMIP6 models project a robust       CMIP6 models project a robust
                          precipitation (Sun et al.,                                      increase in the intensity and       increase in the intensity and       increase in the intensity and
                          2020; Jiang et al., 2013; Hu et                                 frequency of heavy                  frequency of heavy                  frequency of heavy
                          al., 2016; Ge et al., 2017;                                     precipitation (Li et al., 2020;     precipitation (Li et al., 2020;     precipitation (Li et al., 2020;
                          Zhan et al., 2017; Liu et al.,                                  Annex). Median increase of          Annex). Median increase of          Annex). Median increase of
                          2019)                                                           more than 2% in the 50-year         more than 4% in the 50-year         more than 20% in the 50-year
                                                                                          Rx1day and Rx5day events            Rx1day and Rx5day events            Rx1day and Rx5day events
                                                                                          compared to the 1°C warming         compared to the 1°C warming         compared to the 1°C warming
                                                                                          level (Li et al., 2020a) and more   level (Li et al., 2020a) and more   level (Li et al., 2020a) and more
                                                                                          than 4% in annual Rx1day and        than 8% in annual Rx1day and        than 25% in annual Rx1day and
                                                                                          Rx5day and 2% in annual             Rx5day and 6% in annual             Rx5day and 20% in annual
                                                                                          Rx30day compared to pre-            Rx30day compared to pre-            Rx30day compared to pre-
                                                                                          industrial (Annex).                 industrial (Annex).                 industrial (Annex).

                                                                                          Additional evidence from            Additional evidence from            Additional evidence from
                                                                                          CMIP5 simulations for an            CMIP5 simulations for an            CMIP5 simulations for an
                                                                                          increase in the intensity of        increase in the intensity of        increase in the intensity of
                                                                                          heavy precipitation (Zhou et al.,   heavy precipitation (Zhou et al.,   heavy precipitation (Zhou et al.,
                                                                                          2014; Zhang et al., 2015c; Gao      2014; Zhang et al., 2015c; Gao      2014; Zhang et al., 2015c; Gao
                                                                                          et al., 2018; Han et al., 2018)     et al., 2018; Han et al., 2018)     et al., 2018; Han et al., 2018)
                          Medium confidence in the           Low confidence               Intensification of heavy            Intensification of heavy            Intensification of heavy
                          intensitification of heavy                                      precipitation:                      precipitation:                      precipitation:
                          precipitation                                                   Likely (compared with the           Very likely (compared with the      Virtually certain (compared
                                                                                          recent past (1995-2014))            recent past (1995-2014))            with the recent past (1995-
                                                                                          Very likely (compared with pre-     Extremely likely (compared          2014))
                                                                                          industrial)                         with pre-industrial)                Virtually certain (compared
                                                                                                                                                                  with pre-industrial)
 South Asia (SAS)         Significant intensification of     Disagreement among studies   CMIP6 models project an             CMIP6 models project an             CMIP6 models project a robust
                          heavy precipitation (Kim et        (Mukherjee et al., 2018a)    increase in the intensity and       increase in the intensity and       increase in the intensity and
                          al., 2019; Malik et al., 2016;     (Singh et al., 2014a; van    frequency of heavy                  frequency of heavy                  frequency of heavy
                          Pai et al., 2015; Rohini et al.,   Oldenborgh et al., 2016)     precipitation (Li et al., 2020;     precipitation (Li et al., 2020;     precipitation (Li et al., 2020;
                          2016; Roxy et al., 2017;                                        Annex). Median increase of          Annex). Median increase of          Annex). Median increase of
                          Sheikh et al., 2015; Singh et                                   more than 2% in the 50-year         more than 6% in the 50-year         more than 25% in the 50-year
                          al., 2014; Dunn et al., 2020;                                   Rx1day and Rx5day events            Rx1day and Rx5day events            Rx1day and Rx5day events
                          Hussain and Lee, 2013; Kim                                      compared to the 1°C warming         compared to the 1°C warming         compared to the 1°C warming
                          et al., 2019; Malik et al.,                                     level (Li et al., 2020a) and more   level (Li et al., 2020a) and more   level (Li et al., 2020a) and more
                          2016)                                                           than 8% in annual Rx1day and        than 10% in annual Rx1day and       than 30% in annual Rx1day and
                                                                                          Rx5day and 4% in annual             Rx5day and 8% in annual             Rx5day and 25% in annual
                                                                                          Rx30day compared to pre-            Rx30day compared to pre-            Rx30day compared to pre-
                                                                                          industrial (Annex).                 industrial (Annex).                 industrial (Annex).

Do Not Cite, Quote or Distribute                              11-158                                                  Total pages: 345
                             Final Government Distribution                                      Chapter11                                                        IPCC AR6 WGI


                                                                                           Additional evidence from           Additional evidence from           Additional evidence from
                                                                                           CMIP5 simulations for an           CMIP5 simulations for an           CMIP5 simulations for an
                                                                                           increase in the intensity of       increase in the intensity of       increase in the intensity of
                                                                                           heavy precipitation (Sillmann et   heavy precipitation (Sillmann et   heavy precipitation (Sillmann et
                                                                                           al., 2013b; Xu et al., 2017; Han   al., 2013b; Xu et al., 2017; Han   al., 2013b; Xu et al., 2017; Han
                                                                                           et al., 2018; Mukherjee et al.,    et al., 2018; Mukherjee et al.,    et al., 2018; Mukherjee et al.,
                                                                                           2018a; Ali et al., 2019b; Rai et   2018a; Ali et al., 2019b; Rai et   2018a; Ali et al., 2019b; Rai et
                                                                                           al., 2019)                         al., 2019)                         al., 2019)
                          High confidence in the            Low confidence                 Intensification of heavy           Intensification of heavy           Intensification of heavy
                          intensitification of heavy                                       precipitation:                     precipitation:                     precipitation:
                          precipitation                                                    Medium confidence (compared        High confidence (compared          Very likely (compared with the
                                                                                           with the recent past (1995-        with the recent past (1995-        recent past (1995-2014))
                                                                                           2014))                             2014))                             Extremely likely (compared
                                                                                           High confidence (compared          Likely (compared with pre-         with pre-industrial)
                                                                                           with pre-industrial)               industrial)
 Southeast Asia (SEA)     Intensification of heavy          Evidence of a human             CMIP6 models project an            CMIP6 models project an            CMIP6 models project a
                          precipitation (Sun et al.,        contribution for some events    increase in the intensity and      increase in the intensity and      robust increase in the
                          2020; Cheong et al., 2018; Li     (Otto et al., 2018a), but       frequency of heavy                 frequency of heavy                 intensity and frequency of
                          et al., 2018c; Siswanto et al.,   cannot be generalized           precipitation (Li et al., 2020;    precipitation (Li et al., 2020;    heavy precipitation (Li et al.,
                          2015; Supari et al., 2017;                                        Annex). Median increase of         Annex). Median increase of         2020; Annex). Median
                          Villafuerte and Matsumoto,                                        more than 0% in the 50-year        more than 4% in the 50-year        increase of more than 10% in
                          2015)                                                             Rx1day and Rx5day events           Rx1day and Rx5day events           the 50-year Rx1day and
                                                                                            compared to the 1°C warming        compared to the 1°C warming        Rx5day events compared to
                                                                                            level (Li et al., 2020a) and       level (Li et al., 2020a) and       the 1°C warming level (Li et
                                                                                            more than 4% in annual             more than 6% in annual             al., 2020a) and more than
                                                                                            Rx1day and Rx5day and 2%           Rx1day and Rx5day and 4%           20% in annual Rx1day and
                                                                                            in annual Rx30day compared         in annual Rx30day compared         Rx5day and 10% in annual
                                                                                            to pre-industrial (Annex).         to pre-industrial (Annex).         Rx30day compared to pre-
                                                                                                                                                                  industrial (Annex).
                                                                                            Additional evidence from           Additional evidence from
                                                                                            CMIP5 and CORDEX                   CMIP5 and CORDEX                   Additional evidence from
                                                                                            simulations for an increase in     simulations for an increase in     CMIP5 and CORDEX
                                                                                            the intensity of heavy             the intensity of heavy             simulations for an increase in
                                                                                            precipitation (Xu et al., 2017;    precipitation (Xu et al., 2017;    the intensity of heavy
                                                                                            Han et al., 2018; Tangang et       Han et al., 2018; Tangang et       precipitation (Xu et al., 2017;
                                                                                            al., 2018; Trinh-Tuan et al.,      al., 2018; Trinh-Tuan et al.,      Han et al., 2018; Tangang et
                                                                                            2019; Basconcillo et al.,          2019; Basconcillo et al.,          al., 2018; Trinh-Tuan et al.,
                                                                                            2016; Ge et al., 2017; Han et      2016; Ge et al., 2017; Han et      2019; Basconcillo et al.,
                                                                                            al., 2018; Marzin et al., 2015;    al., 2018; Marzin et al., 2015;    2016; Ge et al., 2017; Han et
                                                                                            Tangang et al., 2018; Trinh-       Tangang et al., 2018; Trinh-       al., 2018; Marzin et al., 2015;
                                                                                            Tuan et al., 2019; Xu et al.,      Tuan et al., 2019; Xu et al.,      Tangang et al., 2018; Trinh-
                                                                                            2017)                              2017)                              Tuan et al., 2019; Xu et al.,
                                                                                                                                                                  2017)
                          Medium confidence in the          Low confidence                  Intensification of heavy           Intensification of heavy           Intensification of heavy
                          intensitification of heavy                                        precipitation:                     precipitation:                     precipitation:
                          precipitation                                                     Medium confidence                  High confidence (compared          Very likely (compared with
                                                                                            (compared with the recent          with the recent past (1995-        the recent past (1995-2014))
Do Not Cite, Quote or Distribute                             11-159                                                     Total pages: 345
                                     Final Government Distribution                                  Chapter11                                                  IPCC AR6 WGI
                                                                                               past (1995-2014))                2014))                            Extremely likely (compared
                                                                                               High confidence (compared        Likely (compared with pre-        with pre-industrial)
                                                                                               with pre-industrial)             industrial)

1
2   [END TABLE 11.8 HERE]
3
4
5   [START TABLE 11.9 HERE]
6
7   Table 11.9: Observed trends, human contribution to observed trends, and projected changes at 1.5°C, 2°C and 4°C of global warming for meteorological droughts (MET),
8               agricultural and ecological droughts (AGR/ECOL), and hydrological droughts (HYDR) in Asia, subdivided by AR6 regions. See Sections 11.9.1 and 11.9.4 for details
      Region/ Drought   Observed trends                    Human contribution    Projections
           type                                                                  +1.5 °C                           +2 °C                                  +4 °C
      Russian    MET    Low confidence:                    Low confidence:      Low confidence:                   Low confidence: Limited evidence,      Medium confidence: Decrease in
      Arctic            Limited evidence. Tendency         Limited evidence     Limited evidence. Slight          but some evidence of decrease in dry   drought severity based on SPI
      (RAR)             towards decrease in CDD                                 decrease in CDD in CMIP6          spell duration (Khlebnikova et al.,    (Touma et al., 2015; Spinoni et al.,
                        (Dunn et al., 2020). Lack of                            (Chapter 11 Supplementary         2019b)(Chapter 11 Supplementary        2020) and CDD (Chapter 11
                        data in (Spinoni et al., 2019).                         Material (11.SM))                 Material (11.SM))                      Supplementary Material (11.SM)).
                AGR,    Low confidence:                    Low confidence:      Low confidence: Inconsistent      Low confidence: Inconsistent           Low confidence: Inconsistent trends.
                ECOL    Inconsistent trends (Greve et      Limited evidence     changes in soil moisture (Xu et   changes in soil moisture, variations   Inconsistent trends across models and
                        al., 2014; Padrón et al., 2020).                        al., 2019a)(Chapter 11            across subregions (Xu et al., 2019a)   subregions for surface and total soil
                                                                                Supplementary Material            (Chapter 11 Supplementary Material     moisture (Dai et al., 2018; Lu et al.,
                                                                                (11.SM)).                         (11.SM))                               2019; Cook et al., 2020)(Chapter 11
                                                                                                                                                         Supplementary Material (11.SM));
                                                                                                                                                         Slight drying in PDSI (Dai et al.,
                                                                                                                                                         2018); inconsistent trends or wetting in
                                                                                                                                                         SPEI-PM in CMIP5 (Cook et al.,
                                                                                                                                                         2014b; Vicente-Serrano et al., 2020a).




    Do Not Cite, Quote or Distribute                               11-160                                                  Total pages: 345
                                Final Government Distribution                                    Chapter11                                                    IPCC AR6 WGI

             HYDR   Low confidence: Limited            Low confidence:        Low confidence: Limited          Low confidence: Inconsistent              Low confidence: Mixed signals
                    evidence.                          Limited evidence.      evidence. One study shows lack   changes.. Increasing runoff in CMIP6      among studies (Prudhomme et al.,
                                                                              of signal (Touma et al., 2015)   (Cook et al., 2020) , inconsistent        2014; Giuntoli et al., 2015; Touma et
                                                                                                               signal in SRI depending on subregion      al., 2015; Cook et al., 2020)
                                                                                                               in CMIP5(Touma et al., 2015), or lack
                                                                                                               of signal (Zhai et al., 2020b) in
                                                                                                               available studies.

                                                                                                               (Cook et al., 2020): Increasing runoff
                                                                                                               in one study based on CMIP6 GCMs

                                                                                                               (Zhai et al., 2020b): Lack of signal in
                                                                                                               one study based on single
                                                                                                               hydrological model driven by HAPPI-
                                                                                                               MIP GCM simulations

                                                                                                               Touma et al. (2015): Inconsistent
                                                                                                               signal in SRI depending on subregion
                                                                                                               (CMIP5 GCMs)
  Arabian           Low confidence:                    Low confidence:         Low confidence: Limited          Low confidence: Limited                   Low confidence: Limited evidence
 Peninsula           Inconsistent or no signal         Limited evidence        evidence and inconsistent        evidence and inconsistent trends          and inconsistent trends (Touma et
  (ARP)             (Almazroui, 2019a; Almazroui       (Barlow and Hoell,      trends (Xu et al.,               (Xu et al., 2019a)(Chapter 11             al., 2015; Tabari and Willems,
                    and Islam, 2019).                  2015; Barlow et al.,    2019a)(Chapter 11                Supplementary Material (11.SM)).          2018)(Chapter 11 Supplementary
                                                       2016)                   Supplementary Material                                                     Material (11.SM)).
                    (Dunn et al., 2020): Wetting                               (11.SM)).
                    based on CDD in part of                                                                                                               (Touma et al., 2015): Inconsistant
             MET
                    domain, but missing data in                                                                                                           projections in CMIP5
                    large fraction of region.                                                                                                             (Tabari and Willems, 2018):
                                                                                                                                                          Dominant lack of signal
                    (Spinoni et al., 2019): Missing
                    data in this region.                                                                                                                  Chapter 11 Supplementary Material
                                                                                                                                                          (11.SM)): decreasing dryness based
                                                                                                                                                          on CDD
             AGR,   Low confidence: Limited            Low confidence:         Low confidence: Limited          Low confidence:                           Low confidence: Mixed signal
             ECOL   evidence. Drying in fraction of    Limited evidence        evidence and inconsistent        Limited evidence and inconsistent         between different metrics. including
                    region in one study, but                                   trends (Xu et al.,               trends (Xu et al., 2019a; Cook et         total and surface soil moisture
                    missing data in rest of region                             2019a)(Chapter 11                al., 2020)(Chapter 11                     (Chapter 11 Supplementary Material
                    (Greve et al., 2014).                                      Supplementary Material           Supplementary Material (11.SM))           (11.SM))(Rajsekhar and Gorelick,
                                                                               (11.SM))                                                                   2017; Dai et al., 2018; Lu et al.,
                    (Greve et al., 2014) : Drying in                                                                                                      2019; Cook et al., 2020), PDSI (Dai
                    part of region, but missing data                           (Naumann et al., 2018):                                                    et al., 2018) and SPEI-PM (Cook et
                    in large fraction of region.                               Missing data                                                               al., 2014b; Vicente-Serrano et al.,
                                                                                                                                                          2020a).
                    (Padrón et al., 2020) : Missing
                    data.

                    (Spinoni et al., 2019) : Missing
                    data.

Do Not Cite, Quote or Distribute                                11-161                                                Total pages: 345
                                Final Government Distribution                                        Chapter11                                                       IPCC AR6 WGI
            HYDR   Low confidence: Limited              Low confidence:           Low confidence: Limited             Low confidence: Limited evidence          Low confidence: Inconsistent
                   evidence Drying in one study         Limited evidence          evidence. One study shows           and inconsistent trends (Touma et         trends between models and studies
                   in northern part of region but                                 lack of signal (Touma et al.,       al., 2015; Cook et al., 2020; Zhai et     (Prudhomme et al., 2014; Giuntoli et
                   missing data in rest of region                                 2015).                              al., 2020b)                               al., 2015; Touma et al., 2015; Cook
                   (Dai and Zhao, 2017)                                                                                                                         et al., 2020)
   West     MET    Low confidence: Inconsistent         Low confidence:          Low confidence: t Limited           Low confidence: Inconsistent, weak        Low confidence: Mixed signals
  Central          trendsbetween subregions,            Limited evidence         evidence. Inconsistent or weak      and/or non-significant trends in SPI      between models and between regions
   Asia            based both on CDD and SPI                                     trends in available analyses (Xu    and CDD (Xu et al., 2019a; Spinoni et     (Touma et al., 2015; Han et al., 2018;
  (WCA)            (Spinoni et al., 2019; Dunn et                                et al., 2019a)(Chapter 11
                                                                                                                     al., 2020; Yao et al., 2020)(Chapter 11   Tabari and Willems, 2018; Spinoni et
                   al., 2020; Sharafati et al., 2020;                            Supplementary Material (11.SM))
                   Yao et al., 2020).                                                                                Supplementary Material (11.SM)).          al., 2020; Yao et al., 2020)(Chapter 11
                                                                                                                                                               Supplementary Material (11.SM))
            AGR,   Medium confidence: Increase          Low confidence:           Low confidence: Mixed               Low confidence: Mixed signals in          Medium confidence: Increased
            ECOL   in drought severity.                 Limited evidence.         signals in changes in drought       changes in drought severity,              drying in several metrics, but
                   Dominant signal shows drying                                   severity, depending on model        depending on model and index              substantial intermodel spread and
                   for soil moisture, water-balance     One study by Li et al.    and index (Naumann et al.,          (Naumann et al., 2018; Xu et al.,         lack of signal for total soil moisture
                   (precipitation-                      (2017) concluded that     2018; Xu et al., 2019a; Gu et       2019a; Cook et al., 2020; Gu et al.,      (Dai et al., 2018; Cook et al., 2020;
                   evapotranspiration), PDSI-PM         anthropogenic forcing     al., 2020)(Chapter 11               2020) (Chapter 11 Supplementary           Vicente-Serrano et al.,
                   and SPEI-PM, but with some           has increased AED         Supplementary Material              Material (11.SM)).                        2020a)(Chapter 11 Supplementary
                   differences between subregions       and contributed to        (11.SM)).                                                                     Material (11.SM)).
                   and studies (Greve et al., 2014;     drought severity over                                         Weak signals and inconsistent
                   Dai and Zhao, 2017; Li et al.,       the last decades.         Weak signals and inconsistent       trends between models and                  Increase in drought severity based
                   2017c; Spinoni et al., 2019;                                   trends between models for total     subregions for total and surface soil     on surface soil moisture (Dai et al.,
                   Padrón et al., 2020).                                          and surface soil moisture (Xu et    moisture (Xu et al., 2019a; Cook et       2018; Lu et al., 2019; Cook et al.,
                                                                                  al., 2019a)(Chapter 11              al., 2020)(Chapter 11                     2020). (Chapter 11 Supplementary
                                                                                  Supplementary Material              Supplementary Material (11.SM)),          Material (11.SM)): only median, not
                                                                                  (11.SM)), but increased drying      but increased drying based on SPEI-       83.5%ile), PDSI (Dai et al., 2018),
                                                                                  based on SPEI-PM (Naumann           PM (Naumann et al., 2018; Gu et           and SPEI-PM (Cook et al., 2014b;
                                                                                  et al., 2018; Gu et al., 2020).     al., 2020).                               Vicente-Serrano et al., 2020a); but
                                                                                                                                                                increase in median response and
                                                                                                                                                                substantial intermodel spread for
                                                                                                                                                                total soil moisture (Cook et al.,
                                                                                                                                                                2020)(Chapter 11 Supplementary
                                                                                                                                                                Material (11.SM))
            HYDR   Low confidence: Limited              Low confidence:           Low confidence: Limited             Low confidence: Inconsistent              Medium confidence: Increase of
                   evidence.                            Limited evidence          evidence. One study shows           trends in available studies (Touma        hydrological drought severity
                                                                                  lack of signal (Touma et al.,       et al., 2015; Cook et al., 2020; Zhai     (Prudhomme et al., 2014; Giuntoli et
                                                                                  2015).                              et al., 2020b)                            al., 2015; Touma et al., 2015; Cook
                                                                                                                                                                et al., 2020); but large intermodel
                                                                                                                                                                spread (only 2/3 of models showing
                                                                                                                                                                signal) (Touma et al., 2015) and
                                                                                                                                                                weak signal-to-noise ratio in eastern
                                                                                                                                                                half of domain (Giuntoli et al.,
                                                                                                                                                                2015).




Do Not Cite, Quote or Distribute                                 11-162                                                     Total pages: 345
                               Final Government Distribution                               Chapter11                                                     IPCC AR6 WGI

 Western   MET     Medium confidence: Decrease       Low confidence:    Low confidence: Inconsistent      Low confidence: Inconsistent trends      Low confidence: Inconsistent trends ,
 Siberia           in dryness based on SPI and       Limited evidence   evidence in CMIP5 (Xu et al.,     (Chapter 11 Supplementary Material       but slight decrease in some studies
 (WSB)             CDD, but some inconsistent                           2019a) and CMIP6 projections      (11.SM)) or slight decrease in drought   (Touma et al., 2015; Spinoni et al.,
                   trends in part of domain (Zhang                      (Chapter 11 Supplementary
                                                                                                          (Khlebnikova et al., 2019b; Xu et al.,   2020)(Chapter 11 Supplementary
                   et al., 2017a, 2019b;                                Material (11.SM)).
                   Khlebnikova et al., 2019b;                                                             2019a; Spinoni et al., 2020).            Material (11.SM)).
                   Spinoni et al., 2019; Dunn et
                   al., 2020).                                                                                                                     Spinoni et al. (2020): Slight decrease
                                                                                                          (Khlebnikova et al., 2019b): Mostly      Touma et al. (2015): Tendency
                   Khlebnikova et al. (2019): In                                                          decrease in CDD in a regional climate    towards decrease but partly lack of
                   part mixed signals within                                                              model driven by several CMIP5            model agreement.
                   domain
                                                                                                          models (RCP8.5, 2050-2059 relative
                   (Dunn et al., 2020): Mostly                                                                                                     Chapter 11 Supplementary Material
                   decreasing trend, including                                                            to 1990-1999)
                                                                                                                                                   (11.SM): Lack of model agreement
                   significant changes.
                   (Spinoni et al., 2019): Mostly                                                         Chapter 11 Supplementary Material
                   decreasing trends                                                                      (11.SM): Tendency towards decrease
                                                                                                          but lack of model agreement.
           AGR,    Low confidence: Inconsistent      Low confidence:    Low confidence: Inconsistent      Low confidence: Inconsistent trends      Low confidence: Mixed signals
           ECOL    trends according to subregions    Limited evidence   trends among different metrics    among different metrics. No signal       between different models and metrics,
                   or indices based on soil                             and models. Inconsistent soil     with total soil moisture (Chapter 11     including total and surface soil
                   moisture, PDSI-PM and SPEI-                          moisture projections in CMIP5     Supplementary Material (11.SM)) and      moisture in CMIP6 (Chapter 11
                   PM (Greve et al., 2014; Dai                          (Xu et al., 2019a) and CMIP6      SPEI-PM (Naumann et al., 2018; Gu        Supplementary Material
                   and Zhao, 2017; Li et al.,                           (Chapter 11 Supplementary         et al., 2020), and wetting trend with    (11.SM))(Cook et al., 2020), surface
                   2017c; Spinoni et al., 2019;                         Material (11.SM)), and decrease   surface soil moisture (Xu et al.,        soil moisture in CMIP5 (Dai et al.,
                   Padrón et al., 2020).                                in drought severity based on      2019a).                                  2018; Lu et al., 2019), PDSI (Dai et
                                                                        SPEI-PM (Naumann et al., 2018;                                             al., 2018) and SPEI-PM (Cook et al.,
                                                                        Gu et al., 2020).                                                          2014b; Vicente-Serrano et al., 2020a).

                                                                                                                                                   Difference in signal in CMIP6 vs
                                                                                                                                                   CMIP5: CMIP6 models show drying
                                                                                                                                                   in soil moisture, while CMIP5 models
                                                                                                                                                   show wetting (Cook et al., 2020)




Do Not Cite, Quote or Distribute                             11-163                                             Total pages: 345
                               Final Government Distribution                                Chapter11                                                     IPCC AR6 WGI

            HYDR   Low confidence: Limited            Low confidence:    Low confidence: Limited           Low confidence: Inconsistent             Low confidence: Inconsistent trends.
                   evidence. One study suggests       Limited evidence   evidence. One study shows         trends in available studies (Touma et    Mixed signal among studies and low
                   increasing weak (wetting) trend                       drying (Touma et al., 2015).      al., 2015; Cook et al., 2020; Zhai et    signal to noise ratio (Prudhomme et al.,
                   in runoff (Dai and Zhao, 2017).                                                         al., 2020b)                              2014; Giuntoli et al., 2015; Touma et
                   Some increase in runoff at                                                                                                       al., 2015; Cook et al., 2020)
                   stations from 1951-1990 and                                                             (Cook et al., 2020): Inconsistent
                   1961-2000 (Gudmundsson et                                                               trends including large seasonal
                   al., 2019)                                                                              variations

                                                                                                           (Zhai et al., 2020b): Inconsistent
                                                                                                           trends in study with single
                                                                                                           hydrological model driven with
                                                                                                           HAPPI-MIP GCM simulations

                                                                                                           (Touma et al., 2015): Increase in the
                                                                                                           frequency of hydrological droughts
                                                                                                           based on SRI in CMIP5
  Eastern   MET    Medium confidence: Decrease        Low confidence:    Low confidence: Limited           Medium confidence: Decrease in           Medium confidence: Decrease in
  Siberia          in the duration and frequency      Limited evidence   evidence. Tendency towards        frequency and severity of                meteorological drought severity
  (ESB)            of meteorological droughts                            decrease in SPI in CMIP5 (Xu et   meteorological droughts                  (Touma et al., 2015; Spinoni et al.,
                   (Khlebnikova et al., 2019b;                           al., 2019a) and CDD in CMIP6      (Khlebnikova et al., 2019b; Xu et al.,   2020)(Chapter 11 Supplementary
                   Spinoni et al., 2019; Dunn et                         (Chapter 11 Supplementary         2019a; Spinoni et al., 2020)(Chapter     Material (11.SM)).
                   al., 2020).                                           Material (11.SM)).                11 Supplementary Material (11.SM)).

                   (Khlebnikova et al., 2019b):
                   Decrease in fraction of dry days                                                        (Khlebnikova et al., 2019b):
                   and decrease in mean CDD, but                                                           Projections with a regional climate
                   inconsistent trends for                                                                 model driven with several CMIP5
                   maximum CDD, for 1991-2015                                                              GCMs (RCP8.5, 2050-2059 compared
                   compared to 1966-1990                                                                   with 1990-1999): Mostly decrease in
                                                                                                           CDD but increase in part of domain,
                   (Dunn et al., 2020): Significant                                                        in particular in the south
                   CDD decrease

                   (Spinoni et al., 2019): Mostly
                   decrease in SPI, but partly
                   mixed signals and inconsistent
                   trends




Do Not Cite, Quote or Distribute                              11-164                                              Total pages: 345
                               Final Government Distribution                                        Chapter11                                                  IPCC AR6 WGI

            AGR,   Low confidence: Inconsistent      Low confidence:           Low confidence: Mixed signal in Low confidence: Mixed signal in            Low confidence: Mixed signal in
            ECOL   trends depending on subregion     Limited evidence          changes in drought severity      changes in drought severity depending     drought changes depending on models
                   and index based on soil                                     depending on models and          on models and metrics.                    and metrics, including total and
                   moisture, PDSI-PM and SPEI-                                 metrics.                                                                   surface soil moisture in CMIP6
                   PM (Greve et al., 2014; Dai                                                                  Inconsistent trends in soil moisture      (Chapter 11 Supplementary Material
                   and Zhao, 2017; Spinoni et al.,                             Inconsistent trends in soil      (Xu et al., 2019a)(Chapter 11             (11.SM))(Cook et al., 2020), surface
                   2019; Padrón et al., 2020).                                 moisture (Xu et al.,             Supplementary Material (11.SM)) ,         soil moisture in CMIP5 (Dai et al.,
                                                                               2019a)(Chapter 11                but wetting tendency for SPEI-PM          2018; Lu et al., 2019), PDSI (Dai et
                                                                               Supplementary Material           (Naumann et al., 2018; Gu et al.,         al., 2018) and SPEI-PM (Cook et al.,
                                                                               (11.SM)), , but wetting tendency 2020).                                    2014b; Vicente-Serrano et al., 2020a).
                                                                               for SPEI-PM (Naumann et al.,
                                                                               2018; Gu et al., 2020).                                                    Difference in signal in CMIP6 vs
                                                                                                                                                          CMIP5: CMIP6 models show drying
                                                                                                                                                          in soil moisture, while CMIP5 models
                                                                                                                                                          show wetting (Cook et al., 2020)
            HYDR   Low confidence: Limited           Low confidence:           Low confidence: Limited           Low confidence: Inconsistent             Low confidence: Mixed signal among
                   evidence. One study suggests      Limited evidence          evidence. One study shows lack    trends in available studies (Touma et    studies (Prudhomme et al., 2014;
                   increasing (wetting) trend in                               of signal (Touma et al., 2015).   al., 2015; Cook et al., 2020; Zhai et    Giuntoli et al., 2015; Touma et al.,
                   runoff (Dai and Zhao, 2017).                                                                  al., 2020b)                              2015; Cook et al., 2020)
                   Some increase in runoff at
                   stations from 1951-1990 and                                                                   (Cook et al., 2020): Inconsistent
                   1961-2000 (Gudmundsson et                                                                     trends including large seasonal
                   al., 2019)                                                                                    variations

                                                                                                                 (Zhai et al., 2020b): Inconsistent
                                                                                                                 trends in one study based on single
                                                                                                                 hydrological model driven by HAPPI-
                                                                                                                 MIP GCM simulations

                                                                                                                 (Touma et al., 2015): Mixed signal.
 Russian    MET    Low confidence: Mixed             Low confidence:           Low confidence: Limited           Medium confidence: Decrease              Medium confidence: Decrease in
 Far East          signals between subregions and    Limited evidence.         evidence. Weak decrease in        (Khlebnikova et al., 2019b; Xu et al.,   drought severity (Touma et al., 2015;
  (RFE)            studies (Knutson and Zeng,        One study, Wilcox et      available analyses (Xu et al.,    2019a)(Chapter 11 Supplementary          Han et al., 2018; Spinoni et al.,
                   2018; Khlebnikova et al.,         al. in (Herring et al.,   2019a)(Chapter 11                 Material (11.SM)).
                                                                                                                                                          2020)(Chapter 11 Supplementary
                   2019b; Spinoni et al., 2019;      2015), but mostly         Supplementary Material            (Khlebnikova et al., 2019b): Regional
                   Dunn et al., 2020).               inconclusive.             (11.SM)).                         climate model driven by several          Material (11.SM)).
                                                                                                                 CMIP5 models (RCP8.5, 2050-2059
                                                                                                                 relative to 1990-1999): Mostly
                                                                                                                 decrease in CDD but also increase in
                                                                                                                 part of region (Kamtchatka
                                                                                                                 Peninsual).




Do Not Cite, Quote or Distribute                               11-165                                                   Total pages: 345
                                Final Government Distribution                                         Chapter11                                                    IPCC AR6 WGI

             AGR,   Low confidence: Inconsistent        Low confidence:          Low confidence: Inconsistent     Low confidence: Inconsistent trends Low confidence: Mixed signals
             ECOL   trends depending on subregion       Limited evidence         trends depending on model and    depending on model and index.         between different models and metrics,
                    based on soil moisture, PDSI-                                index.                                                                 including CMIP6 total and surface soil
                    PM and SPEI-PM (Greve et al.,                                                                 Inconsistent trends in CMIP6 total    moisture (Chapter 11 Supplementary
                    2014; Dai and Zhao, 2017;                                    Inconsistent trends in total and and surface soil moisture (Chapter 11 Material (11.SM))(Cook et al., 2020),
                    Spinoni et al., 2019; Padrón et                              surface soil moisture in CMIP6   Supplementary Material                and CMIP5-based surface soil
                    al., 2020).                                                  (Chapter 11 Supplementary        (11.SM))(Cook et al., 2020), but      moisture (Dai et al., 2018; Lu et al.,
                                                                                 Material (11.SM)), but wetting   wetting trends from CMIP5-based       2019), PDSI (Dai et al., 2018) and
                                                                                 trends from CMIP5-based surface surface soil moisture (Xu et al.,      SPEI-PM (Cook et al., 2014b; Vicente-
                                                                                 soil moisture (Xu et al., 2019a) 2019a) and SPEI-PM (Naumann et        Serrano et al., 2020a).
                                                                                 and SPEI-PM (Naumann et al.,     al., 2018; Gu et al., 2020).
                                                                                 2018; Gu et al., 2020).                                                Difference in signal in CMIP6 vs
                                                                                                                                                        CMIP5: CMIP6 models show drying
                                                                                                                                                        in soil moisture, while CMIP5 models
                                                                                 (Naumann et al., 2018) : EC-                                           show wetting (Cook et al., 2020).
                                                                                 Earth driven by SSTs from
                                                                                 several CMIP5 models.
             HYDR   Low confidence: Limited             Low confidence:          Low confidence: Limited           Low confidence: Inconsistent               Low confidence: Inconsistent signal
                    evidence. One study suggests        Limited evidence         evidence. One study shows lack    trends. Available studies show             among studies and metrics, with
                    decreasing (drying) trend in                                 of signal (Touma et al., 2015).   inconsistent signal with high seasonal     generally weak drying trend in summer
                    runoff (Dai and Zhao, 2017).                                                                   variations (Cook et al., 2020) or weak     season (Prudhomme et al., 2014;
                                                                                                                   signal (Touma et al., 2015; Zhai et al.,   Giuntoli et al., 2015; Touma et al.,
                                                                                                                   2020b).                                    2015; Cook et al., 2020)
 East Asia   MET    Low confidence: Lack of signal      Low confidence:           Low confidence: Limited           Low confidence: Inconsistent               Low confidence: Inconsistent
  (EAS)             and mixed trends between            Limited evidence          evidence. Inconsistent            trends depending on model, region          trends between different models and
                    subregions (Spinoni et al.,         (Qin et al., 2015a;       subregional trends (Xu et al.,    or index (Guo et al., 2018; Xu et al.,     important spatial variability (Zhou et
                    2019; Zhang et al., 2019a;          Herring et al., 2019).    2019a) or drying tendency         2019a; Spinoni et al.,                     al., 2014; Touma et al., 2015;
                    Dunn et al., 2020; Li et al.,                                 (Chapter 11 Supplementary         2020)(Chapter 11 Supplementary             Kusunoki, 2018a; Spinoni et al.,
                    2020b). Drying trends in                                      Material (11.SM)).                Material (11.SM)).                         2020)(Chapter 11 Supplementary
                    Southwestern China (Qin et al.,                                                                                                            Material (11.SM)).
                    2015a) and Northern China                                                                       (Spinoni et al., 2020): Tendency
                    (Qin et al., 2015b), but not for                                                                towards decreased in drought               (Zhou et al., 2014): Tendency
                    overall China (Li et al., 2020b).                                                               severity based on SPI.                     towards wetting in the north and
                                                                                                                                                               drying in the south based on CDD.
                                                                                                                    (Huang et al., 2018a): Important
                                                                                                                    subregional differences in SPI             (Kusunoki, 2018a): Increasing CDD
                                                                                                                    projections in a single GCM                (drying trend) over Japan based on
                                                                                                                                                               one GCM.
                                                                                                                    Chapter 11 Supplementary Material
                                                                                                                    (11.SM): Tendency towards drying
                                                                                                                    based on CDD (increasing CDD),
                                                                                                                    but inconsistent trends depending
                                                                                                                    on model.

                                                                                                                    (Xu et al., 2019a): Inconsistent
                                                                                                                    subregional trends based on SPI.



Do Not Cite, Quote or Distribute                                 11-166                                                   Total pages: 345
                                Final Government Distribution                                       Chapter11                                                   IPCC AR6 WGI
           AGR,                                        Low confidence:            Low confidence: Inconsistent      Low confidence: Mixed signals          Medium confidence: Increasing
           ECOL    Medium confidence: Increase         Limited evidence.          trends depending on model,        depending on model, subregion and      dryness as dominant signal in
                   in drying, especially since ca.                                subregion and index (Huang et     index (Gao et al., 2017b; Naumann      projections and over larger part of
                   1990; but wetting tendency          Zhang et al. (2020)        al., 2018a; Naumann et al.,       et al., 2018; Xu et al., 2019a; Cook   domain, but also inconsistent signal
                   beforehand and partly               concluded that             2018; Xu et al., 2019a; Gu et     et al., 2020; Gu et al.,               for some indices and part of the
                   inconsistent subregional trends.    anthropogenic forcing      al., 2020)(Chapter 11             2020)(Chapter 11 Supplementary         domain (Cook et al., 2014b, 2020;
                   Large-scale studies based on        contributed to 2018        Supplementary Material            Material (11.SM)).                     Cheng et al., 2015; Dai et al., 2018;
                   observed soil moisture,             drought, principally as    (11.SM)) .                                                               Naumann et al., 2018; Lu et al.,
                   modelled soil moisture or water     consequence of                                               (Gao et al., 2017b): Study for very    2019; Vicente-Serrano et al.,
                   balance driven by                   enhanced AED.              (Huang et al., 2018a):            small region (Loess Plateau).          2020a)(Chapter 11 Supplementary
                   meteorological observations,                                   Inconsistent projections in a                                            Material (11.SM)).
                   and SPEI-PM, show drying in         One study suggests         study with a single GCM for
                   northern part of domain             that soil moisture         the time frame 2016-2050 (for
                   (northern China, Russian part       drought conditions in      different scenarios) compared
                   of domain, Japan) as well as in     northern China have        to 1960-2005, i.e corresponding
                   Southwest China (east of            been intensified by        to 1.5°C projections compared
                   Tibetan Plateau), but there are     agriculture (Liu et al.,   to recent past.
                   some inconsistent trends in part    2015).
                   of region or some studies, as
                   well as for different time
                   frames (Greve et al., 2014;
                   Chen and Sun, 2015b; Cheng et
                   al., 2015; Qiu et al., 2016; Dai
                   and Zhao, 2017; Jia et al.,
                   2018; Spinoni et al., 2019; Li et
                   al., 2020b; Padrón et al., 2020).
                   Identified trends are also
                   confirmed by regional studies
                   (Liu et al., 2015; Qin et al.,
                   2015b; Liang et al., 2020;
                   Wang et al., 2020). Most of the
                   drying trend took place since
                   1990, with wetting trend
                   beforehand (Chen and Sun,
                   2015b; Wu et al., 2020b).




Do Not Cite, Quote or Distribute                                 11-167                                                   Total pages: 345
                                Final Government Distribution                                        Chapter11                                                 IPCC AR6 WGI
           HYDR    Medium confidence: Increase         Low confidence:            Low confidence: Limited         Low confidence: Limited                 Low confidence: Inconsistent trend
                   in hydrological drought in the      Limited evidence and       evidence. One study shows       evidence and inconsistent trends        between models and studies, and
                   region, in particular in northern   mixed signals.             lack of signal (Touma et al.,   in available studies (Touma et al.,     generally low signal-to-noise ratio
                   China; inconsistent trends in       Available evidence         2015).                          2015; Cook et al., 2020; Zhai et al.,   (Prudhomme et al., 2014; Giuntoli et
                   part of the region (Liu et al.,     suggests that a                                            2020b).                                 al., 2015; Touma et al., 2015; Cook
                   2015; Dai and Zhao, 2017;           combination of change                                                                              et al., 2020)
                   Zhang et al., 2018b).               in climatic drivers
                                                       (precipitation, Epot)                                                                              (Touma et al., 2015; Cook et al.,
                   Drying in large part of domain,     and human drivers                                                                                  2020): Generally inconsistent trends
                   in particular in northern China     (agriculture, water                                                                                between models, with low model
                   (Zhao and Dai, 2017)                management) are                                                                                    agreement.
                                                       responsible for trends
                   Increase of hydrological            (Liu et al., 2015;                                                                                 (Giuntoli et al., 2015): Trend
                   droughts in the Yangtze river       Zhang et al., 2018b).                                                                              towards drying but generally low
                   (Zhang et al., 2018b).                                                                                                                 signal-to-noise ratio except in small
                                                       Increasing                                                                                         subregion.
                                                       hydrological droughts
                                                       trends in the Yangtze
                                                       river are dominantly
                                                       driven by
                                                       precipitation, but
                                                       increases in potential
                                                       evaporation and
                                                       human activities also
                                                       play a role (Zhang et
                                                       al., 2018b). Drought
                                                       conditions in northern
                                                       China (soil moisture
                                                       and runoff) have been
                                                       intensified by
                                                       agriculture (Liu et al.,
                                                       2015).




Do Not Cite, Quote or Distribute                                 11-168                                                 Total pages: 345
                               Final Government Distribution                                Chapter11                                                  IPCC AR6 WGI
  Eastern   MET    Low confidence: Inconsistent      Low confidence:    Low confidence. Limited            Medium confidence: Decrease in          Medium confidence: Decrease in
  Central          trends between subregions,        Limited evidence   evidence; slight decrease in       drought severity, with weakly           drought severity (Touma et al., 2015;
   Asia            with overall tendency to                             meteorological drought in          inconsistent changes for some           Spinoni et al., 2020)(Chapter 11
  (ECA)            decrease (Spinoni et al., 2019;                      available analyses (Xu et al.,     indices (Xu et al., 2019a; Spinoni et   Supplementary Material (11.SM)).
                   Dunn et al., 2020).                                  2013) (Chapter 11                  al., 2020)(Chapter 11
                                                                        Supplementary Material             Supplementary Material (11.SM))
                                                                        (11.SM))
                                                                                                           (Spinoni et al., 2019): Strong
                                                                                                           decrease in drought for SPI-based
                                                                                                           metrics in RCP4.5 compared to
                                                                                                           1981-2010

                                                                                                           (Xu et al., 2019a): Decrease in
                                                                                                           frequency of SPI-based events but
                                                                                                           slight increase or inconsistent
                                                                                                           changes in duration of SPI-based
                                                                                                           events.

                                                                                                           Chapter 11 Supplementary Material
                                                                                                           (11.SM): substantial decrease in
                                                                                                           CDD
            AGR,   Medium confidence: Increase       Low confidence:    Low confidence: Mixed signal       Low confidence: Mixed signal in         Low confidence: Mixed trends
            ECOL   in drying, but some conflicting   Limited evidence   in changes in drought severity,    changes in drought severity.            between different models and
                   trends between drought metrics                       lack of signal based in total      Inconsistent trends in total and        drought metrics (Chapter 11
                   and sub-regions (Greve et al.,                       column soil moisture (Xu et al.,   surface soil moisture, with stronger    Supplementary Material
                   2014; Cheng et al., 2015; Dai                        2019a)(Chapter 11                  tendency to wetting, (Xu et al.,        (11.SM))(Cook et al., 2014b, 2020;
                   and Zhao, 2017; Li et al.,                           Supplementary Material             2019a; Cook et al., 2020)(Chapter       Dai et al., 2018; Lu et al., 2019;
                   2017c; Spinoni et al., 2019;                         (11.SM)) and SPEI-PM               11 Supplementary Material               Vicente-Serrano et al., 2020a).
                   Padrón et al., 2020; Zhang et                        (Naumann et al., 2018; Gu et       (11.SM)), and drying based on the
                   al., 2020c).                                         al., 2020).                        SPEI-PM (Naumann et al., 2018;
                                                                                                           Gu et al., 2020).
            HYDR   Low confidence: Limited           Low confidence:    Low confidence: Limited            Low confidence: Limited evidence        Low confidence: Mixed trends.
                   evidence. Mostly inconsistent     Limited evidence   evidence. One study shows          and inconsistent trends (Touma et       Model disagreement and inconsistent
                   trends in one study (Dai and                         lack of signal (Touma et al.,      al., 2015; Cook et al., 2020; Zhai et   changes among studies, seaons and
                   Zhao, 2017).                                         2015).                             al., 2020b)                             metrics, with overall low signal-to-
                                                                                                                                                   noise ratio (Prudhomme et al., 2014;
                                                                                                                                                   Giuntoli et al., 2015; Touma et al.,
                                                                                                                                                   2015; Cook et al., 2020).




Do Not Cite, Quote or Distribute                             11-169                                              Total pages: 345
                               Final Government Distribution                                 Chapter11                                                   IPCC AR6 WGI
  Tibetan   MET    Low confidence: Inconsistent        Low confidence:    Low confidence: Limited            Low confidence: Inconsistent            Low confidence: Inconsistent
  Plateau          trends (Jiang et al., 2013; Donat   Limited evidence   evidence. Weak or inconsistent     trends, but tendency towards            trends between models, but
   (TIB)           et al., 2016a; Hu et al., 2016;                        trends in available analyses (Xu   wetting (Xu et al., 2019a; Cook et      tendency towards wetting and
                   Dunn et al., 2020).                                    et al., 2019a)(Chapter 11          al., 2020)(Chapter 11                   decrease in drought (Zhou et al.,
                                                                          Supplementary Material             Supplementary Material (11.SM))         2014; Touma et al., 2015)(Chapter
                                                                          (11.SM))                                                                   11 Supplementary Material
                                                                                                             (Spinoni et al., 2019): No data in      (11.SM)).
                                                                                                             the region
                                                                                                             (Cook et al., 2020): Only analysis      (Zhou et al., 2014): Decrease of
                                                                                                             of mean precipitation but tendency      CDD is projected but there is large
                                                                                                             towards wetting in all seasons in the   uncertainty
                                                                                                             region
                                                                                                             (Xu et al., 2019a)(Chapter 11
                                                                                                             Supplementary Material (11.SM)):
                                                                                                             Weak trends but tendency towards
                                                                                                             wetting.
            AGR,   Low confidence: Inconsistent        Low confidence:    Low confidence: Inconsistent       Low confidence: Inconsistent            Low confidence: Inconsistent
            ECOL   trends. Spatially varying           Limited evidence   trends between models, indices     trends between models, indices and      trends between models, indices and
                   trends, with slight tendency to                        and subregions (Naumann et         subregions (Naumann et al., 2018;       subregions (Cook et al., 2014b,
                   overall wetting(Cheng et al.,                          al., 2018; Xu et al., 2019a; Gu    Xu et al., 2019a; Cook et al., 2020;    2020; Dai et al., 2018; Lu et al.,
                   2015; Dai and Zhao, 2017; Jia                          et al., 2020)(Chapter 11           Gu et al., 2020)(Chapter 11             2019; Vicente-Serrano et al.,
                   et al., 2018; Zhang et al.,                            Supplementary Material             Supplementary Material (11.SM)).        2020a)(Chapter 11 Supplementary
                   2018a; Li et al., 2020c; Wang                          (11.SM)).                                                                  Material (11.SM)).
                   et al., 2020).

                   (Greve et al., 2014; Spinoni et
                   al., 2019; Padrón et al., 2020):
                   Missing data in most of region.
            HYDR   Low confidence: Limited             Low confidence:    Low confidence: Limited            Low confidence: Limited                 Low confidence: Inconsistent
                   evidence.                           Limited evidence   evidence. One study shows          evidence and inconsistent trends        trends between models and studies,
                                                                          lack of signal (Touma et al.,      in available studies (Touma et al.,     and low signal-to-nois ratio
                                                                          2015).                             2015; Cook et al., 2020; Zhai et al.,   (Prudhomme et al., 2014; Giuntoli et
                                                                                                             2020b)                                  al., 2015; Touma et al., 2015; Cook
                                                                                                                                                     et al., 2020)




Do Not Cite, Quote or Distribute                               11-170                                              Total pages: 345
                               Final Government Distribution                                        Chapter11                                               IPCC AR6 WGI
  South    MET     Medium confidence: Increase         Low confidence:           Low confidence: Limited         Low confidence: Inconsistent           Low confidence: Inconsistent
   Asia            in meteorological drought.          Limited evidence          evidence and inconsistent       trends, with light tendency to         trends depending on model and
  (SAS)            Subregional differences but         (Fadnavis et al., 2019)   trends (Xu et al.,              decreased drying (Xu et al., 2019a;    subregion, with light tendency to
                   drying is dominant (Mishra et                                 2019a)(Chapter 11               Spinoni et al., 2020)(Chapter 11       decreases in meteorological drought
                   al., 2014b; Malik et al., 2016;                               Supplementary Material          Supplementary Material (11.SM))        in CMIP5 and CMIP6 (Mishra et al.,
                   Guhathakurta et al., 2017;                                    (11.SM)).                                                              2014b; Touma et al., 2015; Salvi and
                   Spinoni et al., 2019; Dunn et                                                                                                        Ghosh, 2016; Spinoni et al.,
                   al., 2020) (see also Section                                                                                                         2020)(Chapter 11 Supplementary
                   10.6.3)                                                                                                                              Material (11.SM)); light increased
                                                                                                                                                        drying in NDD in CORDEX-CORE
                                                                                                                                                        (Coppola et al., 2021b). Overall poor
                                                                                                                                                        climate model performance for
                                                                                                                                                        South Asia monsoon in CMIP5 and
                                                                                                                                                        CORDEX (Mishra et al., 2014a;
                                                                                                                                                        Saha et al., 2014; Sabeerali et al.,
                                                                                                                                                        2015; Singh et al., 2017b). See also
                                                                                                                                                        Section 10.6.3 for assessment for
                                                                                                                                                        changes in Indian summer monsoon
                                                                                                                                                        rainfall.
           AGR,    Low confidence: Lack of             Low confidence:           Low confidence: Inconsistent    Low confidence: Inconsistent           Medium confidence: Decreased
           ECOL    signal and inconsistent trends      Limited evidence          trends in drought between       trends in drought between              drying trend (Chapter 11
                   depending on subregion based                                  models and subregions           models, subregions and studies,        Supplementary Material
                   on soil moisture, PDSI-PM and                                 (Naumann et al., 2018; Xu et    but slight dominant tendency           (11.SM))(Cook et al., 2014b, 2020;
                   SPEI-PM (Greve et al., 2014;                                  al., 2019a; Gu et al.,          towards wetting(Naumann et al.,        Mishra et al., 2014b; Dai et al.,
                   Mishra et al., 2014b; Dai and                                 2020)(Chapter 11                2018; Xu et al., 2019a; Cook et al.,   2018; Lu et al., 2019; Vicente-
                   Zhao, 2017; Spinoni et al.,                                   Supplementary Material          2020; Gu et al.,                       Serrano et al., 2020a)(Chapter 11
                   2019; Padrón et al., 2020) and                                (11.SM))                        2020)(CMIP6.ANNEX-CH11)                Supplementary Material (11.SM))
                   decrease of the drying effect of
                   the atmospheric evaporative
                   demand (Jhajharia et al., 2015).

           HYDR    Low confidence: Limited             Low confidence:           Low confidence: Limited         Low confidence: Limited                Low confidence: Inconsistent
                   evidence. Inconsistent trends or    Limited evidence          evidence. One study shows       evidence. Lack of signal in CMIP5      trends between models and studies
                   limited data in available studies                             lack of signal (Touma et al.,   (Touma et al., 2015). Decrease in      (Prudhomme et al., 2014; Giuntoli et
                   (Zhao and Dai, 2017;                                          2015).                          dryness in CMIP6 (Cook et al.,         al., 2015; Touma et al., 2015; Cook
                   Gudmundsson et al., 2019,                                                                     2020); mostly inconsistent trends in   et al., 2020)
                   2021).                                                                                        HAPPI-MIP driven simulations
                                                                                                                 with one hydrological model (Zhai
                                                                                                                 et al., 2020b).




Do Not Cite, Quote or Distribute                                11-171                                                 Total pages: 345
                               Final Government Distribution                                 Chapter11                                                IPCC AR6 WGI
 Southeast   MET   Low confidence: Inconsistent     Low confidence:          Low confidence: Limited     Low confidence: Inconsistent            Medium confidence: Increase in
   Asia            trends between subregions        Limited evidence         evidence (Xu et al.,        trends between models, subregions       drying in CMIP6 and CORDEX
  (SEA)            (Spinoni et al., 2019; Dunn et   (Mcbride et al., 2015;   2019a)(Chapter 11           and studies (Tangang et al., 2018;      simulations (Cook et al., 2020;
                   al., 2020).                      King et al., 2016b)      Supplementary Material      Xu et al., 2019a; Spinoni et al.,       Supari et al., 2020; Coppola et al.,
                                                    although the the         (11.SM))                    2020)(Chapter 11 Supplementary          2021b) (Chapter 11 Supplementary
                                                    equatorial Asia                                      Material (11.SM)) but with overall      Material (11.SM)). but inconsistent
                                                    drought of 2015 has                                  drying in CMIP6 and CORDEX              trends or wetting in CMIP5-based
                                                    been attributed to                                   simulations (Tangang et al., 2018;      projections(Touma et al., 2015;
                                                    anthropogenic                                        Cook et al., 2020; Coppola et al.,      Cook et al., 2020; Spinoni et al.,
                                                    warming effects                                      2021b) (Chapter 11 Supplementary        2020; Supari et al., 2020)(
                                                    (Shiogama et al.,                                    Material (11.SM)).
                                                    2020).                                                                                       (Supari et al., 2020): Strong drying
                                                                                                         (Tangang et al., 2018): Projected       trend based on CDD in CORDEX
                                                                                                         drying based on CDD in CORDEX           simulations for Indonesia
                                                                                                         simulations for Indonesia
                                                                                                                                                 (Coppola et al., 2021b): Drying
                                                                                                         (Xu et al., 2019a): Inconsistent        based on number of dry days (NDD)
                                                                                                         trends across region based on SPI,      in CORDEX-CORE projects
                                                                                                         but with slight drying over
                                                                                                         Indonesia                               (Cook et al., 2020): Decreasing trend
                                                                                                                                                 in mean precipitation which is only
                                                                                                         (Spinoni et al., 2020): Wetting trend   found in CMIP6 and not in CMIP5.
                                                                                                         based on SPI
                                                                                                                                                 Chapter 11 Supplementary Material
                                                                                                         (Chapter 11 Supplementary               (11.SM): Strong projected drying
                                                                                                         Material (11.SM)): Drying trend         trend based on CDD in CMIP6
                                                                                                         based on CDD                            projections

                                                                                                                                                 (Touma et al., 2015): Inconsistent
                                                                                                                                                 trends in SPI in CMIP5 projections

                                                                                                                                                 (Spinoni et al., 2020): Wetting trend
                                                                                                                                                 based on SPI in CMIP5 projections.

                                                                                                                                                 (Cai et al., 2014a, 2015, 2018): An
                                                                                                                                                 increasing frequency of precipitation
                                                                                                                                                 deficits is projected as a
                                                                                                                                                 consequence of an increasing
                                                                                                                                                 frequency of extreme El Niño.




Do Not Cite, Quote or Distribute                             11-172                                            Total pages: 345
                                     Final Government Distribution                                     Chapter11                                                     IPCC AR6 WGI
                  AGR,   Low confidence: Inconsistent       Low confidence:          Low confidence: Inconsistent      Low confidence: Inconsistent             Low confidence: Mixed signal
                  ECOL   trends depending onsubregion       Limited evidence         trends depending on model,        trends depending on model,              depending on model and metric.
                         and index based on soil                                     subregion, index or study         subregion, index or study               Drying tendency based on CMIP6
                         moisture, PDSI-PM and SPEI-                                 (Naumann et al., 2018; Xu et      (Naumann et al., 2018; Xu et al.,       soil moisture projections (Cook et
                         PM (Greve et al., 2014; Dai                                 al., 2019a; Gu et al.,            2019a; Cook et al., 2020; Gu et al.,    al., 2020)(Chapter 11 Supplementary
                         and Zhao, 2017; Spinoni et al.,                             2020)(Chapter 11                  2020)(Chapter 11 Supplementary          Material (11.SM)), inconsistent
                         2019; Padrón et al., 2020).                                 Supplementary Material            Material (11.SM)).                      trends in CMIP5 surface soil
                                                                                     (11.SM))                                                                  moisture (Dai et al., 2018; Lu et al.,
                                                                                                                                                               2019), but wetting trends with PDSI
                                                                                                                                                               (Dai et al., 2018) and SPEI-PM
                                                                                                                                                               (Cook et al., 2014b; Vicente-Serrano
                                                                                                                                                               et al., 2020a) in studies driven with
                                                                                                                                                               CMIP5 data. ,

                                                                                                                                                               (Cook et al., 2020): Drying trend in
                                                                                                                                                               SEA in CMIP6), but not in CMIP5.
                  HYDR   Low confidence: Limited            Low confidence:          Low confidence: Limited           Low confidence: Limited                 Low confidence: Inconsistent trend
                         evidence. Regionally               Limited evidence         evidence. One study shows         evidence and inconsistent trends        between models and studies
                         inconsistent trends in one study                            decrease in hydrological          in available studies (Touma et al.,     (Prudhomme et al., 2014; Giuntoli et
                         (Dai and Zhao, 2017).                                       drought (Touma et al., 2015).     2015; Cook et al., 2020; Zhai et al.,   al., 2015; Touma et al., 2015; Cook
                                                                                                                       2020b)                                  et al., 2020)
1
2   [END TABLE 11.9 HERE]
3
4
5   [START TABLE 11.10 HERE]
6
7   Table 11.10: Observed trends, human contribution to observed trends, and projected changes at 1.5°C, 2°C and 4°C of global warming for temperature extremes in Australasia,
8                subdivided by AR6 regions. See Sections 11.9.1 and 11.9.2 for details.
                                                                      Detection and attribution;                                               Projections
                Region                   Observed trends
                                                                          event attribution                     1.5 °C                             2 °C                              4 °C
     All Australasia              Significant increases in the     Robust evidence of a human      CMIP6 models project a           CMIP6 models project a            CMIP6 models project a
                                  intensity and frequency of hot   contribution to the observed    robust increase in the           robust increase in the            robust increase in the
                                  extremes and decreases in the    increase in the intensity and   intensity and frequency of       intensity and frequency of        intensity and frequency of
                                  intensity and frequency of       frequency of hot extremes       TXx events and a robust          TXx events and a robust           TXx events and a robust
                                  cold extremes (CSIRO and         and decrease in the intensity   decrease in the intensity and    decrease in the intensity and     decrease in the intensity and
                                  BOM, 2015; Jakob and             and frequency of cold           frequency of TNn events (Li      frequency of TNn events (Li       frequency of TNn events (Li
                                  Walland, 2016; Alexander         extremes (Seong et al., 2020;   et al., 2020). Median increase   et al., 2020). Median increase    et al., 2020). Median increase
                                  and Arblaster, 2017)             Hu et al., 2020; Wang et al.,   of more than 0C in the 50-       of more than 0.5°C in the 50-     of more than 2.5°C in the 50-
                                                                   2017).                          year TXx and TNn events          year TXx and TNn events           year TXx and TNn events
                                                                                                   compared to the 1°C warming      compared to the 1°C warming       compared to the 1°C warming
                                                                                                   level (Li et al., 2020)          level (Li et al., 2020)           level (Li et al., 2020)

                                                                                                   Additional evidence from         Additional evidence from          Additional evidence from
                                                                                                   CMIP5 simulations for an         CMIP5 simulations for an          CMIP5 simulations for an

    Do Not Cite, Quote or Distribute                                11-173                                                    Total pages: 345
                               Final Government Distribution                                        Chapter11                                                        IPCC AR6 WGI
                                                                                               increase in the intensity and      increase in the intensity and       increase in the intensity and
                                                                                               frequency of hot extremes          frequency of hot extremes           frequency of hot extremes
                                                                                               and decrease in the intensity      and decrease in the intensity       and decrease in the intensity
                                                                                               and frequency of cold              and frequency of cold               and frequency of cold
                                                                                               extremes (Alexander and            extremes (Alexander and             extremes (Alexander and
                                                                                               Arblaster, 2017; Herold et al.,    Arblaster, 2017; Herold et al.,     Arblaster, 2017; Herold et al.,
                                                                                               2018; Evans et al., 2020;          2018; Evans et al., 2020;           2018; Evans et al., 2020;
                                                                                               Grose et al., 2020)                Grose et al., 2020)                 Grose et al., 2020)
                            Very likely increase in the      Human influence very likely       Increase in the intensity and      Increase in the intensity and       Increase in the intensity and
                            intensity and frequency of hot   contributed to the observed       frequency of hot extremes:         frequency of hot extremes:          frequency of hot extremes:
                            extremes and decrease in the     increase in the intensity and     Very likely (compared with         Extremely likely (compared          Virtually certain (compared
                            intensity and frequency of       frequency of hot extremes         the recent past (1995-2014))       with the recent past (1995-         with the recent past (1995-
                            cold extremes                    and decrease in the intensity     Extremely likely (compared         2014))                              2014))
                                                             and frequency of cold             with pre-industrial)               Virtually certain (compared         Virtually certain (compared
                                                                                                                                  with pre-industrial)                with pre-industrial)
                                                             extremes
                                                                                               Decrease in the intensity and
                                                                                               frequency of cold extremes:       Decrease in the intensity and        Decrease in the intensity and
                                                                                               Very likely (compared with        frequency of cold extremes:          frequency of cold extremes:
                                                                                               the recent past (1995-2014))      Extremely likely (compared           Virtually certain (compared
                                                                                               Extremely likely (compared        with the recent past (1995-          with the recent past (1995-
                                                                                               with pre-industrial)              2014))                               2014))
                                                                                                                                 Virtually certain (compared          Virtually certain (compared
                                                                                                                                 with pre-industrial)                 with pre-industrial)
 Northern Australia (NAU)   Significant increases in the     Robust evidence of a human      CMIP6 models project a robust CMIP6 models project a robust             CMIP6 models project a robust
                            intensity and frequency of hot   contribution to the observed    increase in the intensity and      increase in the intensity and        increase in the intensity and
                            extremes and significant         increase in the intensity and   frequency of TXx events and a frequency of TXx events and a             frequency of TXx events and a
                            decreases in the intensity and   frequency of hot extremes       robust decrease in the intensity robust decrease in the intensity       robust decrease in the intensity
                            frequency of cold extremes       and decrease in the intensity   and frequency of TNn events (Li and frequency of TNn events             and frequency of TNn events
                            (Perkins and Alexander,          and frequency of cold           et al., 2020; Annex). Median       (Li et al., 2020; Annex).            (Li et al., 2020; Annex).
                            2013; Wang et al., 2013c;        extremes (Wang et al., 2017; increase of more than 0C in the Median increase of more than               Median increase of more than
                            CSIRO and BOM, 2015;             Hu et al., 2020; Seong et al.,  50-year TXx and TNn events         0.5°C in the 50-year TXx and         3°C in the 50-year TXx and
                            Donat et al., 2016a;             2020; Knutson et al., 2014;     compared to the 1°C warming        TNn events compared to the           TNn events compared to the
                            Alexander and Arblaster,         Lewis and Karoly, 2014;         level (Li et al., 2020) and more 1°C warming level (Li et al.,          1°C warming level (Li et al.,
                            2017; Dunn et al., 2020)         Perkins et al., 2014; Arblaster than 1.5°C in annual TXx and       2020) and more than 2°C in           2020) and more than 3.5°C in
                                                             et al., 2014; Hope et al.,      TNn compared to pre-industrial annual TXx and TNn compared              annual TXx and TNn compared
                                                             2015, 2016; Perkins and         (Annex).                           to pre-industrial (Annex).           to pre-industrial (Annex).
                                                             Gibson, 2015)
                                                                                             Additional evidence from           Additional evidence from             Additional evidence from
                                                                                             CMIP5 simulations for an           CMIP5 simulations for an             CMIP5 simulations for an
                                                                                             increase in the intensity and      increase in the intensity and        increase in the intensity and
                                                                                             frequency of hot extremes and      frequency of hot extremes and        frequency of hot extremes and
                                                                                             decrease in the intensity and      decrease in the intensity and        decrease in the intensity and
                                                                                             frequency of cold extremes         frequency of cold extremes           frequency of cold extremes
                                                                                             (Alexander and Arblaster, 2017; (Alexander and Arblaster, 2017;         (Alexander and Arblaster, 2017;
                                                                                             Herold et al., 2018; Evans et al., Herold et al., 2018; Evans et al.,   Herold et al., 2018; Evans et al.,
                                                                                             2020; Grose et al., 2020)          2020; Grose et al., 2020)            2020; Grose et al., 2020)
                            High confidence in the           High confidence in a human      Increase in the intensity and       Increase in the intensity and       Increase in the intensity and
                            increase in the intensity and    contribution to the observed    frequency of hot extremes:          frequency of hot extremes:          frequency of hot extremes:

Do Not Cite, Quote or Distribute                              11-174                                                         Total pages: 345
                              Final Government Distribution                                         Chapter11                                                           IPCC AR6 WGI

                           frequency of hot extremes        increase in the intensity and     Likely (compared with the recent     Very likely (compared with the       Virtually certain (compared
                           and likely decrease in the       frequency of hot extremes         past (1995-2014))                    recent past (1995-2014))             with the recent past (1995-
                           intensity and frequency of       and decrease in the intensity     Very likely (co