Editing IPCC:AR6/WGI/Chapter-11 (section)

== 11.3 Temperature Extremes ==

<div id="h1-4-siblings" class="h1-siblings"></div>

This section assesses changes in temperature extremes at global, continental and regional scales. The main focus is on the changes in the magnitude and frequency of moderate extreme temperatures (those that occur several times a year) to very extreme temperatures (those that occur once in 10 or more years) of time scales from a day to a season, though there is a strong emphasis on the daily scale where literature is most concentrated.

<div id="11.3.1" class="h2-container"></div>

<span id="mechanisms-and-drivers"></span>
=== 11.3.1 Mechanisms and Drivers ===

<div id="h2-24-siblings" class="h2-siblings"></div>

The SREX (IPCC, 2012) and AR5 (IPCC, 2014) concluded that greenhouse gas forcing is the dominant factor for the increases in the intensity, frequency, and duration of warm extremes and the decrease in those of cold extremes. This general global-scale warming is modulated by large-scale atmospheric circulation patterns, as well as by feedbacks such as soil moisture-evapotranspiration–temperature and snow/ice-albedo–temperature feedbacks, and local forcings such as land-use change or changes in aerosol concentrations at the regional and local scales (Sections 11.1.5 and 11.1.6, and Box 11.1). Therefore, changes in temperature extremes at regional and local scales can have heterogeneous spatial distributions. Changes in the magnitudes (or intensities) of extreme temperatures are often larger than changes in global surface temperature, because of larger warming on land than on the ocean surface ( [[IPCC:Wg1:Chapter:Chapter-2#2.3.1.1|Section 2.3.1.1]] ), and because of feedbacks, though they are of similar magnitude to changes in the local mean temperature (Figure 11.2).

Extreme temperature events are associated with large-scale meteorological patterns ( [[#Grotjahn--2016|Grotjahn et al., 2016]] ). Quasi-stationary anticyclonic circulation anomalies or atmospheric blocking events are linked to temperature extremes in many regions, such as in Australia ( [[#Parker--2014|Parker et al., 2014]] ; [[#Perkins-Kirkpatrick--2016|Perkins-Kirkpatrick et al., 2016]] ), Europe ( [[#Brunner--2017|Brunner et al., 2017]] , 2018; [[#Schaller--2018|Schaller et al., 2018]] ), Eurasia ( [[#Yao--2017|Yao et al., 2017]] ), Asia ( [[#Chen--2016|Chen et al., 2016]] ; [[#Ratnam--2016|Ratnam et al., 2016]] ; [[#Rohini--2016|Rohini et al., 2016]] ), and North America ( [[#Yu--2018|Yu et al., 2018]] , 2019; [[#Zhang--2019|Zhang and Luo, 2019]] ). Mid-latitude planetary wave modulations affect short-duration temperature extremes such as heatwaves ( [[#Perkins--2015|Perkins, 2015]] ; [[#Kornhuber--2020|Kornhuber et al., 2020]] ). The large-scale modes of variability (Annex IV) affect the strength, frequency and persistence of these meteorological patterns and, hence, temperature extremes. For example, cold and warm extremes in the mid-latitudes are associated with atmospheric circulation patterns such as the Pacific-North American (PNA) pattern, as well as atmosphere–ocean coupled modes such as Pacific Decadal Variability (PDV), the North Atlantic Oscillation (NAO), and Atlantic Multi-decadal Variability (AMV) ( [[#11.1.5|Section 11.1.5]] ; [[#Kamae--2014|Kamae et al., 2014]] ; [[#Johnson--2018|Johnson et al., 2018]] ; [[#Ruprich-Robert--2018|Ruprich-Robert et al., 2018]] ; [[#Yu--2018|Yu et al., 2018]] , 2020; [[#Müller--2020|Müller et al., 2020]] ; [[#Qasmi--2021|Qasmi et al., 2021]] ). Changes in the modes of variability in response to warming would therefore affect temperature extremes ( [[#Clark--2013|Clark and Brown, 2013]] ; [[#Horton--2015|Horton et al., 2015]] ). The level of confidence in those changes varies, both in the observations and in future projections, affecting the level of confidence in changes in temperature extremes in different regions. As highlighted in Chapters 2 to 4 of this Report, it is ''likely'' that there have been observational changes in the extratropical jets and mid-latitude jet meandering ( [[IPCC:Wg1:Chapter:Chapter-2#2.3.1.4.3|Section 2.3.1.4.3]] and Cross-Chapter Box 10.1). There is ''low confidence'' in possible effects of Arctic warming on mid-latitude temperature extremes (Cross-Chapter Box 10.1). A large portion of the multi-decadal changes in extreme temperature remains after the removal of the effect of these modes of variability, and can be attributed to human influence ( [[#Kamae--2017b|Kamae et al., 2017b]] ; [[#Wan--2019|Wan et al., 2019]] ). Thus, global warming dominates changes in temperature extremes at the regional scale and it is ''very unlikely'' that dynamic responses to greenhouse-gas induced warming would alter the direction of these changes.

Land–atmosphere feedbacks strongly modulate regional- and local-scale changes in temperature extremes ( ''high confidence'' ) ( [[#11.1.6|Section 11.1.6]] ; [[#Seneviratne--2013|Seneviratne et al., 2013]] ; [[#Lemordant--2016|Lemordant et al., 2016]] ; [[#Donat--2017|Donat et al., 2017]] ; [[#Sillmann--2017b|Sillmann et al., 2017b]] ; [[#Hirsch--2019|Hirsch et al., 2019]] ). This effect is particularly notable in mid-latitude regions where the drying of soil moisture amplifies high temperatures, especially through increases in sensible heat flux ( [[#Whan--2015|Whan et al., 2015]] ; [[#Douville--2016|Douville et al., 2016]] ; [[#Vogel--2017|Vogel et al., 2017]] ). Land–atmosphere feedbacks amplifying temperature extremes also include boundary-layer feedbacks and effects on atmospheric circulation ( [[#Miralles--2014a|Miralles et al., 2014a]] ; [[#Schumacher--2019|Schumacher et al., 2019]] ). Soil-moisture–temperature feedbacks affect past and present-day heatwaves in observations and model simulations, both locally ( [[#Miralles--2014a|Miralles et al., 2014a]] ; [[#Cowan--2016|Cowan et al., 2016]] , 2020; [[#Hauser--2016|Hauser et al., 2016]] ; [[#Meehl--2016|Meehl et al., 2016]] ; [[#Wehrli--2019|Wehrli et al., 2019]] ) and beyond the regions of feedback occurrence through changes in regional circulation patterns ( [[#Stéfanon--2014|Stéfanon et al., 2014]] ; [[#Koster--2016|Koster et al., 2016]] ; [[#Sato--2019|Sato and Nakamura, 2019]] ). The uncertainty due to the representation of land–atmosphere feedbacks in ESMs is a cause of discrepancy between observations and simulations ( [[#Clark--2006|Clark et al., 2006]] ; [[#Mueller--2014|Mueller and Seneviratne, 2014]] ; [[#Meehl--2016|Meehl et al., 2016]] ). The decrease of plant transpiration or the increase of stomata resistance under enhanced CO <sub>2</sub> concentrations is a direct CO <sub>2</sub> forcing of land temperatures (warming due to reduced evaporative cooling), which contributes to higher warming on land ( [[#Lemordant--2016|Lemordant et al., 2016]] ; [[#Vicente-Serrano--2020b|Vicente-Serrano et al., 2020b]] ). The snow/ice-albedo feedback plays an important role in amplifying temperature variability in the high latitudes ( [[#Diro--2018|Diro et al., 2018]] ) and can be the largest contributor to the rapid warming of cold extremes in the mid- and high latitudes of the Northern Hemisphere ( [[#Gross--2020|Gross et al., 2020]] ).

Regional external forcings, including land-use changes and emissions of anthropogenic aerosols, play an important role in the changes of temperature extremes in some regions ( ''high confidence'' ) ( [[#11.1.6|Section 11.1.6]] ). Deforestation may have contributed to about one third of the warming of hot extremes in some mid-latitude regions since the pre-industrial time ( [[#Lejeune--2018|Lejeune et al., 2018]] ). Aspects of agricultural practice, including no-till farming, irrigation, and overall cropland intensification, may cool hot temperature extremes ( [[#Davin--2014|Davin et al., 2014]] ; N.D. [[#Mueller--2016|]] [[#Mueller--2016|Mueller et al., 2016]] ). For instance, cropland intensification has been suggested to be responsible for a cooling of the highest temperature percentiles in Midwest USA (N.D. [[#Mueller--2016|]] [[#Mueller--2016|Mueller et al., 2016]] ). Irrigation has been shown to be responsible for a cooling of hot temperature extremes of up to 1°C–2°C in many mid-latitude regions in the present climate (Thieryet al., 2017, 2020), a process not represented in most of state-of-the-art ESMs (CMIP5, CMIP6). Double cropping may have led to increased hot extremes in the inter-cropping season in part of China ( [[#Jeong--2014|Jeong et al., 2014]] ). Rapid increases in summer warming in western Europe and north-east Asia since the 1980s are linked to a reduction in anthropogenic aerosol precursor emissions over Europe ( [[#Nabat--2014|Nabat et al., 2014]] ; [[#Dong--2016b|Dong et al., 2016b]] , 2017), in addition to the effect of increased greenhouse gas forcing (see also [[IPCC:Wg1:Chapter:Chapter-10#10.1.3.1|Section 10.1.3.1]] ). This effect of aerosols on temperature-related extremes is also noted for declines in short-lived anthropogenic aerosol emissions over North America ( [[#Mascioli--2016|Mascioli et al., 2016]] ). On the local scale, the urban heat island (UHI) effect results in higher temperatures in urban areas than in their surrounding regions, and contributes to warming in regions of rapid urbanization, in particular for nighttime temperature extremes (Box 10.3; [[#Phelan--2015|Phelan et al., 2015]] ; [[#Chapman--2017|Chapman et al., 2017]] ; Y. [[#Sun--2019|]] [[#Sun--2019|]] [[#Sun--2019|]] [[#Sun--2019|Sun et al., 2019]] ). But these local and regional forcings are generally not or not well represented in the CMIP5 and CMIP6 simulations (see also [[#11.3.3|Section 11.3.3]] ), contributing to uncertainty in model simulated changes.

In summary, greenhouse gas forcing is the dominant driver leading to the warming of temperature extremes. At regional scales, changes in temperature extremes are modulated by changes in large-scale patterns and modes of variability, feedbacks including soil-moisture–evapotranspiration–temperature or snow/ice–albedo–temperature feedbacks, and local and regional forcings such as land-use and land-cover changes, or aerosol concentrations, and decadal and multi-decadal natural variability. This leads to heterogeneity in regional changes and their associated uncertainties ( ''high confidence'' ). Changes in anthropogenic aerosol concentrations have ''likely'' affected trends in hot extremes in some regions. Irrigation and crop expansion have attenuated increases in summer hot extremes in some regions, such as the Midwestern USA ( ''medium confidence'' ). Urbanization has ''likely'' exacerbated the effects of global warming in cities, in particular for nighttime temperature extremes.

<div id="11.3.2" class="h2-container"></div>

<span id="observed-trends"></span>
=== 11.3.2 Observed Trends ===

<div id="h2-25-siblings" class="h2-siblings"></div>

The SREX (IPCC, 2012) reported a ''very likely'' decrease in the number of cold days and nights and increase in the number of warm days and nights at the global scale. Confidence in trends was assessed as regionally variable ( ''low to medium confidence'' ) due to either a lack of observations or varying signals in sub-regions.

Since SREX (IPCC, 2012) and AR5 (IPCC, 2014), many regional-scale studies have examined trends in temperature extremes using different metrics that are based on daily temperatures, such as the Commission for Climatology/World Climate Research Program/Commission for Oceanography and Marine Meteorology joint Expert Team on Climate Change Detection and Indices (ETCCDI) indices ( [[#Dunn--2020|Dunn et al., 2020]] ). The additional observational records, along with a stronger warming signal, show very clearly that changes observed at the time of AR5 (IPCC, 2014) continued, providing strengthened evidence of an increase in the intensity and frequency of hot extremes and decrease in the intensity and frequency of cold extremes. While the magnitude of the observed trends in temperature-related extremes varies depending on the region, spatial and temporal scales, and metric assessed, evidence of a warming effect is overwhelming, robust, and consistent. In particular, an increase in the intensity and frequency of hot extremes is almost always associated with an increase in the hottest temperatures and in the number of heatwave days. It is also the case for changes (decreases) in cold extremes. For this reason, and to simplify the presentation, the phrase ‘increase in the intensity and frequency of hot extremes’ is used to represent, collectively, an increase in the magnitude of extreme day and/or night temperatures, in the number of warm days and/or nights, and in the number of heatwave days. Changes in cold extremes are assessed similarly.

On the global scale, evidence of an increase in the number of warm days and nights and a decrease in the number of cold days and nights, and an increase in the coldest and hottest extreme temperatures is very robust and consistent among all variables. Figure 11.2 displays time series of globally averaged TXx and TNn on land. Warming of land mean TXx is similar to the mean temperature warming on land, which is about 45% higher than global warming ( [[IPCC:Wg1:Chapter:Chapter-2#2.3.1|Section 2.3.1]] ). Warming of land mean TNn is even higher, with about 3°C of warming since 1960 (Figure 11.2). Figure 11.9 shows maps of linear trends over 1960–2018 in TXx, TNn, and frequency of warm days (TX90p). The maps for TXx and TNn show trends consistent with overall warming in most regions, with a particularly high warming of TXx in Europe and north-western South America, and a particularly high warming of TNn in the Arctic. Consistent with the observed warming in global surface temperature ( [[IPCC:Wg1:Chapter:Chapter-2#2.3.1.2|Section 2.3.1.2]] ) and the observed trends in TXx and TNn, the frequency of TX90p has increased, while that of cold nights (TN10p) has decreased since the 1950s: Nearly all land regions showed statistically significant decreases in TN10p ( [[#Alexander--2016|Alexander, 2016]] ; [[#Dunn--2020|Dunn et al., 2020]] ), though trends in TX90p are variable with some decreases in Southern South America, mainly during austral summer ( [[#Rusticucci--2017|Rusticucci et al., 2017]] ). A decrease in the number of cold spell days is also observed over nearly all land surface areas ( [[#Easterling--2016|Easterling et al., 2016]] ) and in the northern mid-latitudes in particular ( [[#van%20Oldenborgh--2019|van Oldenborgh et al., 2019]] ). These observed changes are also consistent when a new global land surface daily air temperature dataset is analysed (P. [[#Zhang--2019|]] [[#Zhang--2019|]] [[#Zhang--2019|]] [[#Zhang--2019|Zhang et al., 2019]] ). Warming trends in temperature extremes globally, and in most land areas, over the path century are also found to be consistent in a range of observation-based datasets ( [[#Fischer--2014|Fischer and Knutti, 2014]] ; [[#Donat--2016a|Donat et al., 2016a]] ; [[#Dunn--2020|Dunn et al., 2020]] ), with the extremes related to daily minimum temperatures changing faster than those related to daily maximum temperatures ( [[#Dunn--2020|Dunn et al., 2020]] ; see Figure 11.2). Seasonal variations in trends in temperature-related extremes have been demonstrated. A warming in warm-season temperature extremes is detected, even during the ‘slower surface global warming’ period from the late 1990s to early 2010s (Cross-Chapter Box 3.1; [[#Kamae--2014|Kamae et al., 2014]] ; [[#Seneviratne--2014|Seneviratne et al., 2014]] ; [[#Imada--2017|Imada et al., 2017]] ). Many studies of past changes in temperature extremes for particular regions or countries show trends consistent with this global picture, as summarized below and in Tables 11.4, 11.7, 11.10, 11.13, 11.16 and 11.19.

<div id="_idContainer047" class="Basic-Text-Frame"></div>

[[File:9f3091f4b853f00de2588be1834e43f8 IPCC_AR6_WGI_Figure_11_9.png]]

'''Figure 11.9 |''' '''Linear trends over 1960–2018 for three temperature extreme indices: (a)''' the annual maximum daily maximum temperature (TXx), '''(b)''' the annual minimum daily minimum temperature (TNn), and '''(c)''' the annual number of days when daily maximum temperature exceeds its 90th percentile from a base period of 1961–1990 (TX90p); based on the HadEX3 dataset (Dunn et al. , 20 20). Linear trends are calculated only for grid points with at least 66% of the annual values over the period and which extend to at least 2009. Areas without sufficient data are shown in grey. No overlay indicates regions where the trends are significant at the p = 0.1 level. Crosses indicate regions where trends are not significant. Further details on data sources and processing are available in the chapter data table (Table 11.SM.9).

In Africa (Table 11.4), while it is difficult to assess changes in temperature extremes in parts of the continent because of a lack of data, evidence of an increase in the intensity and frequency of hot extremes and decrease in the intensity and frequency of cold extremes is clear and robust in regions where data are available. These include an increase in the frequency of warm days and nights and a decrease in the frequency of cold days and nights with ''high confidence'' ( [[#Donat--2013a|Donat et al., 2013a]] , 2014b; [[#Kruger--2013|Kruger and Sekele, 2013]] ; [[#Chaney--2014|Chaney et al., 2014]] ; [[#Filahi--2016|Filahi et al., 2016]] ; [[#Moron--2016|Moron et al., 2016]] ; [[#Ringard--2016|Ringard et al., 2016]] ; [[#Barry--2018|Barry et al., 2018]] ; [[#Gebrechorkos--2018|Gebrechorkos et al., 2018]] ) and an increase in heatwaves ( [[#Russo--2016|Russo et al., 2016]] ; [[#Ceccherini--2017|Ceccherini et al., 2017]] ). The increase in TNn is more notable than in TXx (Figure 11.9). Cold spells occasionally strike subtropical areas, but are ''likely'' to have decreased in frequency ( [[#Barry--2018|Barry et al., 2018]] ). The frequency of cold events has ''likely'' decreased in South Africa ( [[#Song--2014|Song et al., 2014]] ; [[#Kruger--2017|Kruger and Nxumalo, 2017]] ), North Africa ( [[#Filahi--2016|Filahi et al., 2016]] ; [[#Driouech--2021|Driouech et al., 2021]] ), and the Sahara ( [[#Donat--2016a|Donat et al., 2016a]] ). Over the whole continent, there is ''medium confidence'' in an increase in the intensity and frequency of hot extremes and decrease in the intensity and frequency of cold extremes; it is ''likely'' that similar changes have also occurred in areas with poor data coverage, as warming is widespread and as projected future changes are similar over all regions ( [[#11.3.5|Section 11.3.5]] ).

In Asia (Table 11.7), there is very ''robust evidence'' for a ''very likely'' increase in the intensity and frequency of hot extremes and decrease in the intensity and frequency of cold extremes in recent decades. This is clear in global studies (e.g., [[#Alexander--2016|Alexander, 2016]] ; [[#Dunn--2020|Dunn et al., 2020]] ), as well as in numerous regional studies (Table 11.7). The area fraction with extreme warmth in Asia increased during 1951–2016 ( [[#Imada--2018|Imada et al., 2018]] ). The frequency of warm extremes increased and the frequency of cold extremes decreased in East Asia ( [[#Zhou--2016|]] [[#Zhou--2016|B. Zhou et al., 2016]] ; [[#Chen--2017|Chen and Zhai, 2017]] ; [[#Yin--2017|Yin et al., 2017]] ; W. [[#Lee--2018|]] [[#Lee--2018|]] [[#Lee--2018|Lee et al., 2018]] ; [[#Qian--2019|Qian et al., 2019]] ) and west Asia (Acar Deniz and Gönençgil, 2015; [[#Erlat--2016|Erlat and Türkeş, 2016]] ; [[#Rahimi--2018|Rahimi and Hejabi, 2018]] ; [[#Rahimi--2018|Rahimi et al., 2018]] ) with ''high confidence'' . The duration of heat extremes has also lengthened in some regions, for example, in southern China ( [[#Luo--2016|Luo and Lau, 2016]] ), but there is ''medium confidence'' of heat extremes increasing in frequency in South Asia ( [[#AlSarmi--2014|AlSarmi and Washington, 2014]] ; [[#Sheikh--2015|Sheikh et al., 2015]] ; [[#Mazdiyasni--2017|Mazdiyasni et al., 2017]] ; [[#Zahid--2017|Zahid et al., 2017]] ; [[#Nasim--2018|Nasim et al., 2018]] ; [[#Khan--2019|Khan et al., 2019]] ; [[#Sen%20Roy--2019|Sen Roy, 2019]] ). Warming trends in daily temperature extremes indices have also been observed in central Asia ( [[#Hu--2016|Hu et al., 2016]] ; [[#Feng--2018|Feng et al., 2018]] ), the Hindu Kush Himalaya ( [[#Sun--2017|Sun et al., 2017]] ), and South East Asia ( [[#Supari--2017|Supari et al., 2017]] ; [[#Cheong--2018|Cheong et al., 2018]] ). The intensity and frequency of cold spells in all Asian regions have been decreasing since the beginning of the 20th century ( ''high confidence'' ) ( [[#Sheikh--2015|Sheikh et al., 2015]] ; [[#Donat--2016a|Donat et al., 2016a]] ; [[#Dong--2018|Dong et al., 2018]] ; [[#van%20Oldenborgh--2019|van Oldenborgh et al., 2019]] ).

In Australasia (Table 11.10), there is very ''robust evidence'' for ''very likely'' increases in the number of warm days and warm nights and decreases in the number of cold days and cold nights since 1950 ( [[#Lewis--2015|Lewis and King, 2015]] ; [[#Jakob--2016|Jakob and Walland, 2016]] ; [[#Alexander--2017|Alexander and Arblaster, 2017]] ). The increase in extreme minimum temperatures occurs in all seasons over most of Australia and typically exceeds the increase in extreme maximum temperatures (X.L. [[#Wang--2013|Wang et al., 2013]] b; [[#Jakob--2016|Jakob and Walland, 2016]] ). However, some parts of Southern Australia have shown stable or increased numbers of frost days since the 1980s ( [[#Dittus--2014|Dittus et al., 2014]] ) (see also [[#11.3.4|Section 11.3.4]] ). Similar positive trends in extreme minimum and maximum temperatures have been observed in New Zealand, in particular in the autumn and winter seasons, although they generally show higher spatial variability ( [[#Caloiero--2017|Caloiero, 2017]] ). In the tropical Western Pacific region, spatially coherent warming trends in maximum and minimum temperature extremes have been reported for the period 1951–2011 ( [[#Whan--2014|Whan et al., 2014]] ; [[#McGree--2019|McGree et al., 2019]] ).

In Central and South America (Table 11.13), there is ''high confidence'' that observed hot extremes (TN90p, TX90p) have increased, and cold extremes (TN10p, TX10p) have decreased over recent decades, though trends vary among different extremes types, datasets, and regions ( [[#Skansi--2013|Skansi et al., 2013]] ; [[#Dittus--2016|Dittus et al., 2016]] ; [[#Rusticucci--2017|Rusticucci et al., 2017]] ; [[#Meseguer-Ruiz--2018|Meseguer-Ruiz et al., 2018]] ; [[#Salvador--2018|Salvador and de Brito, 2018]] ; [[#Dereczynski--2020|Dereczynski et al., 2020]] ; [[#Dunn--2020|Dunn et al., 2020]] ; [[#Olmo--2020|Olmo et al., 2020]] ). An increase in the intensity and frequency of heatwave events was also observed between 1961 and 2014 in an area covering most of South America ( [[#Ceccherini--2016|Ceccherini et al., 2016]] ; [[#Geirinhas--2018|Geirinhas et al., 2018]] ). However, there is ''medium confidence'' that warm extremes (TXx and TX90p) have decreased in the last decades over the central region of South-Eastern South America (SES) during austral summer ( [[#Tencer--2012|Tencer and Rusticucci, 2012]] ; [[#Skansi--2013|Skansi et al., 2013]] ; [[#Rusticucci--2017|Rusticucci et al., 2017]] ; [[#Wu--2017|Wu and Polvani, 2017]] ). There is ''medium confidence'' that TNn extremes are warming faster than TXx extremes, with the largest warming rates observed over North-East Brazil (NEB) and Northern South America (NSA) for cold nights ( [[#Skansi--2013|Skansi et al., 2013]] ).

In Europe (Table 11.16), there is very ''robust evidence'' for a ''very likely'' increase in maximum temperatures and the frequency of heatwaves. The increase in the magnitude and frequency of high maximum temperatures has been observed consistently across regions, including in central Europe ( [[#Twardosz--2013|Twardosz and Kossowska-Cezak, 2013]] ; [[#Christidis--2015|Christidis et al., 2015]] ; [[#Lorenz--2019|Lorenz et al., 2019]] ) and southern Europe ( [[#Croitoru--2013|Croitoru and Piticar, 2013]] ; [[#El%20Kenawy--2013|El Kenawy et al., 2013]] ; [[#Christidis--2015|Christidis et al., 2015]] ; [[#Nastos--2015|Nastos and Kapsomenakis, 2015]] ; [[#Fioravanti--2016|Fioravanti et al., 2016]] ; [[#Ruml--2017|Ruml et al., 2017]] ). In Northern Europe, a strong increase in extreme winter warming events has been observed ( [[#Matthes--2015|Matthes et al., 2015]] ; [[#Vikhamar-Schuler--2016|Vikhamar-Schuler et al., 2016]] ). Temperature observations for winter cold spells show a long-term decreasing frequency in Europe (Brunner et al., 2018; [[#van%20Oldenborgh--2019|van Oldenborgh et al., 2019]] ), and typical cold spells, such as that observed during the 2009–2010 winter, had an occurrence probability two times smaller currently than if climate change had not occurred ( [[#Christiansen--2018|Christiansen et al., 2018]] ).

In North America (Table 11.19), there is very ''robust evidence'' for a ''very likely'' increase in the intensity and frequency of hot extremes and decrease in the intensity and frequency of cold extremes for the whole continent, though there are substantial spatial and seasonal variations in the trends. Minimum temperatures display warming consistently across the continent, while there are more contrasting trends in the annual maximum daily temperatures in parts of the USA (Figure 11.9; [[#Lee--2014|Lee et al., 2014]] ; [[#van%20Oldenborgh--2019|van Oldenborgh et al., 2019]] ; [[#Dunn--2020|Dunn et al., 2020]] ). In Canada, there is a clear increase in the intensity and frequency of hot extremes and decrease in the intensity and frequency of cold extremes ( [[#Vincent--2018|Vincent et al., 2018]] ). In Mexico, a clear warming trend in TNn was found, particularly in the northern arid region ( [[#Montero-Martínez--2018|Montero-Martínez et al., 2018]] ). The number of warm days has increased and the number of cold days has decreased ( [[#García-Cueto--2019|García-Cueto et al., 2019]] ). Cold spells have undergone a reduction in magnitude and intensity in all regions of North America ( [[#Bennett--2015|Bennett and Walsh, 2015]] ; [[#Donat--2016a|Donat et al., 2016a]] ; [[#Grotjahn--2016|Grotjahn et al., 2016]] ; [[#Vose--2017|Vose et al., 2017]] ; [[#García-Cueto--2019|García-Cueto et al., 2019]] ; [[#van%20Oldenborgh--2019|van Oldenborgh et al., 2019]] ).

Extreme heat events have increased around the Arctic since 1979, particularly over Arctic North America and Greenland ( [[#Matthes--2015|Matthes et al., 2015]] ; [[#Dobricic--2020|Dobricic et al., 2020]] ), which is consistent with summer melt ( [[IPCC:Wg1:Chapter:Chapter-9#9.4.1|Section 9.4.1]] ). Observations north of 60˚N show increases in winter warm days and nights over 1979–2015, while cold days and nights declined ( [[#Sui--2017|Sui et al., 2017]] ). Extreme heat days are particularly strong in winter, with observations showing the warmest mid-winter temperatures at the North Pole rising at twice the rate of mean temperature ( [[#Moore--2016|Moore, 2016]] ), as well as increases in Arctic winter warm days ( [[#Vikhamar-Schuler--2016|Vikhamar-Schuler et al., 2016]] ; [[#Graham--2017|Graham et al., 2017]] ). Arctic annual minimum temperatures have increased at about three times the rate of global surface temperature since the 1960s (Figures 11.2 and 11.9), consistent with the observed mean cold season (October–May) warming of 3.1°C in the region ( [[IPCC:Wg1:Chapter:Atlas|Atlas]] 11.2).

Trends in some measures of heatwaves are also observed at the global scale. Globally averaged heatwave intensity, heatwave duration, and the number of heatwave days have significantly increased from 1950–2011 ( [[#Perkins--2015|Perkins, 2015]] ). There are some regional differences in trends in characteristics of heatwaves, with significant increases reported in Europe ( [[#Russo--2015|Russo et al., 2015]] ; [[#Forzieri--2016|Forzieri et al., 2016]] ; [[#Sánchez-Benítez--2020|Sánchez-Benítez et al., 2020]] ) and Australia (CSIRO and BOM, 2016; [[#Alexander--2017|Alexander and Arblaster, 2017]] ). In Africa, there is ''medium confidence'' that heatwaves, regardless of the definition, have been becoming more frequent, longer-lasting, and hotter over more than three decades ( [[#Fontaine--2013|Fontaine et al., 2013]] ; [[#Mouhamed--2013|Mouhamed et al., 2013]] ; [[#Ceccherini--2016|Ceccherini et al., 2016]] , 2017; [[#Forzieri--2016|Forzieri et al., 2016]] ; [[#Moron--2016|Moron et al., 2016]] ; [[#Russo--2016|Russo et al., 2016]] ). The majority of heatwave characteristics examined in China between 1961 and 2014 show increases in heatwave days, consistent with warming ( [[#You--2017|You et al., 2017]] ; [[#Xie--2020|Xie et al., 2020]] ). Increases in the frequency and duration of heatwaves are also observed in Mongolia ( [[#Erdenebat--2016|Erdenebat and Sato, 2016]] ) and India ( [[#Ratnam--2016|Ratnam et al., 2016]] ; [[#Rohini--2016|Rohini et al., 2016]] ). In the UK, the lengths of short heatwaves have increased since the 1970s, while the lengths of long heatwaves (more than 10 days) have decreased over some stations in the south-east of England (M. [[#Sanderson--2017|]] [[#Sanderson--2017|Sanderson et al., 2017]] ). In Central and South America, there are increases in the frequency of heatwaves ( [[#Barros--2015|Barros et al., 2015]] ; [[#Bitencourt--2016|Bitencourt et al., 2016]] ; [[#Ceccherini--2016|Ceccherini et al., 2016]] ; [[#Piticar--2018|Piticar, 2018]] ), although decreases in Excess Heat Factor (EHF), which is a metric for heatwave intensity, are observed in South America in data derived from HadGHCND ( [[#Cavanaugh--2015|Cavanaugh and Shen, 2015]] ).

In summary, it is ''virtually certain'' that there has been an increase in the number of warm days and nights and a decrease in the number of cold days and nights on the global scale since 1950. Both the coldest extremes and hottest extremes display increasing temperatures. It is ''very likely'' that these changes have also occurred at the regional scale in Europe, Australasia, Asia, and North America. It is ''virtually certain'' that there has been increases in the intensity and duration of heatwaves and in the number of heatwave days at the global scale. These trends ''likely'' occur in Europe, Asia, and Australia. There is ''medium confidence'' in similar changes in temperature extremes in Africa and ''high confidence'' in South America; the lower confidence is due to reduced data availability and fewer studies. Annual minimum temperatures on land have increased about three times more than global surface temperature since the 1960s, with particularly strong warming in the Arctic ( ''hi'' ''gh confidence'' ).

<div id="11.3.3" class="h2-container"></div>

<span id="model-evaluation"></span>
=== 11.3.3 Model Evaluation ===

<div id="h2-26-siblings" class="h2-siblings"></div>

The AR5 assessed that CMIP3 and CMIP5 models generally captured the observed spatial distributions of the mean state and that the inter-model range of simulated temperature extremes was similar to the spread estimated from different observational datasets; the models generally captured trends in the second half of the 20th century for indices of extreme temperature, although they tended to overestimate trends in hot extremes and underestimate trends in cold extremes ( [[#Flato--2013|Flato et al., 2013]] ). Post-AR5 studies on the CMIP5 models’ performance in simulating mean and changes in temperature extremes continue to support the AR5 assessment ( [[#Fischer--2014|Fischer and Knutti, 2014]] ; [[#Sillmann--2014|Sillmann et al., 2014]] ; [[#Ringard--2016|Ringard et al., 2016]] ; [[#Borodina--2017b|Borodina et al., 2017b]] ; [[#Donat--2017|Donat et al., 2017]] ; [[#Di%20Luca--2020b|Di Luca et al., 2020b]] ). Over Africa, the observed warming in temperature extremes is captured by CMIP5 models, although it is underestimated in Western and Central Africa ( [[#Sherwood--2014|Sherwood et al., 2014]] ; [[#Diedhiou--2018|Diedhiou et al., 2018]] ). Over East Asia, the CMIP5 ensemble performs well in reproducing the observed trend in temperature extremes averaged over China ( [[#Dong--2015|Dong et al., 2015]] ). Over Australia, the multi-model mean performs better than individual models in capturing observed trends in gridded station-based ETCCDI temperature indices ( [[#Alexander--2017|Alexander and Arblaster, 2017]] ).

Initial analyses of CMIP6 simulations (H. [[#Chen--2020|]] [[#Chen--2020|Chen et al., 2020]] ; [[#Di%20Luca--2020a|Di Luca et al., 2020a]] ; [[#Kim--2020|Kim et al., 2020]] ; [[#Thorarinsdottir--2020|Thorarinsdottir et al., 2020]] ; [[#Wehner--2020|Wehner et al., 2020]] ; [[#Li--2021|Li et al., 2021]] ) indicate that the CMIP6 models perform similarly to the CMIP5 models regarding biases in hot and cold extremes. In general, CMIP5 and CMIP6 historical simulations are similar in their performance in simulating the observed climatology of extreme temperatures ( ''high confidence'' ) ''.'' The general warm bias in hot extremes and cold bias in cold extremes reported for CMIP5 models ( [[#Kharin--2013|Kharin et al., 2013]] ; [[#Sillmann--2013a|Sillmann et al., 2013a]] ) remain in CMIP6 models ( [[#Di%20Luca--2020a|Di Luca et al., 2020a]] ). However, there is some evidence that CMIP6 models better represent some of the underlying processes leading to extreme temperatures, such as seasonal and diurnal variability and synoptic-scale variability ( [[#Di%20Luca--2020a|Di Luca et al., 2020a]] ). Whether these improvements are sufficient to enhance our understanding of past changes, or to reduce uncertainties in future projections, remains unclear. The relative error estimates in the simulation of various indices of temperature extremes in the available CMIP6 models show that no single model performs the best on all indices, and the multi-model ensemble seems to outperform any individual model due to its reduction in systematic bias ( [[#Kim--2020|Kim et al., 2020]] ). Figure 11.10 show errors in the 1979–2014 average annual TXx and annual TNn simulated by available CMIP6 models in comparison with HadEX3 and ERA5 ( [[#Kim--2020|Kim et al., 2020]] ; [[#Wehner--2020|Wehner et al., 2020]] ; [[#Li--2021|Li et al., 2021]] ). While the magnitude of the model error depends on the reference dataset, the model evaluations drawn from different reference datasets are quite similar. In general, models reproduce the spatial patterns and magnitudes of both cold and hot temperature extremes quite well. There are also systematic biases. Hot extremes tend to be too cool in mountainous and high-latitude regions, but too warm in the eastern USA and South America. For cold extremes, CMIP6 models are too cool, except in north-eastern Eurasia and the southern mid-latitudes. Errors in seasonal mean temperatures are uncorrelated with errors in extreme temperatures and are often of opposite sign ( [[#Wehner--2020|Wehner et al., 2020]] ).

<div id="_idContainer049" class="Basic-Text-Frame"></div>

[[File:77e630c252e7fc746ad65fc1d186f932 IPCC_AR6_WGI_Figure_11_10.png]]

'''Figure 11.10 |''' '''Multi-model mean bias in temperature extremes (°C) for the period 1979–2014, calculated as the difference between the Coupled Model Intercomparison Project Phase 6 (CMIP6) multi-model mean and the average of observations from the values avail''' ab '''le''' in HadEX3. ''(a)'' The annual hottest temperatu '''re''' (TXx); and ''(b)'' the annual coldest temperature (TNn). Areas without sufficient data are shown in grey. Adapted from [[#Wehner--2020|Wehner et al. (2020)]] under the terms of the Creative Commons Attribution licence. Further details on data sources and processing are available in the chapter data table (Table 11.SM.9).

Atmospheric Model Intercomparison Project (AMIP) simulations are often used in event attribution studies to assess the influence of global warming on observed temperature-related extremes. These simulations typically capture the observed trends in temperature extremes, though some regional features, such as the lack of warming in daytime warm temperature extremes over South America and parts of North America, are not reproduced in the model simulations ( [[#Dittus--2018|Dittus et al., 2018]] ), possibly due to internal variability, deficiencies in local surface processes, or forcings that are not represented in the sea surface temperatures (SSTs). Additionally, the AMIP models assessed tend to produce overly persistent heatwave events. This bias in the duration of the events does not impact on the reliability of the models’ positive trends ( [[#Freychet--2018|Freychet et al., 2018]] ).

Several regional climate models (RCMs) have also been evaluated in terms of their performance in simulating the climatology of extremes in various regions of the Coordinated Regional Downscaling Experiment (CORDEX) ( [[#Giorgi--2009|Giorgi et al., 2009]] ), especially in East Asia ( [[#Ji--2015|Ji and Kang, 2015]] ; [[#Yu--2015|Yu et al., 2015]] ; [[#Park--2016|Park et al., 2016]] ; [[#Bucchignani--2017|Bucchignani et al., 2017]] ; [[#Gao--2017a|Gao et al., 2017a]] ; [[#Niu--2018|Niu et al., 2018]] ; Y. [[#Sun--2018b|]] [[#Sun--2018|Sun et al., 2018]] b ; [[#Wang--2019|Wang et al., 2019]] ), Europe ( [[#Vautard--2013|Vautard et al., 2013]] , 2021; [[#Smiatek--2016|Smiatek et al., 2016]] ; [[#Gaertner--2018|Gaertner et al., 2018]] ; [[#Cardoso--2019|Cardoso et al., 2019]] ; [[#Lorenz--2019|Lorenz et al., 2019]] ; [[#Jacob--2020|Jacob et al., 2020]] ; [[#Kim--2020|Kim et al., 2020]] ), and Africa (J. [[#Kim--2014|]] [[#Kim--2014|Kim et al., 2014]] ; [[#Diallo--2015|Diallo et al., 2015]] ; [[#Dosio--2017|Dosio, 2017]] ; [[#Samouly--2018|Samouly et al., 2018]] ; [[#Mostafa--2019|Mostafa et al., 2019]] ). Compared to GCMs, RCM simulations show an added value in simulating temperature-related extremes, though this depends on topographical complexity and the parameters employed (see [[IPCC:Wg1:Chapter:Chapter-10#10.3.3|Section 10.3.3]] ). The improvement with resolution is noted in East Asia ( [[#Park--2016|Park et al., 2016]] ; W. [[#Zhou--2016|]] [[#Zhou--2016|Zhou et al., 2016]] ; [[#Shi--2017|Shi et al., 2017]] ; [[#Hui--2018|Hui et al., 2018]] ). However, in the European CORDEX ensemble, different aerosol climatologies with various degrees of complexity were used in projections ( [[#Bartók--2017|Bartók et al., 2017]] ; [[#Lorenz--2019|Lorenz et al., 2019]] ) and the land surface models used in the RCMs do not account for physiological CO <sub>2</sub> effects on photosynthesis leading to enhanced water-use efficiency and decreased evapotranspiration ( [[#Schwingshackl--2019|Schwingshackl et al., 2019]] ), which could lead to biases in the representation of temperature extremes in these projections ( [[#Boé--2020|Boé et al., 2020]] ). In addition, there are key cold biases in temperature extremes over areas with complex topography ( [[#Niu--2018|Niu et al., 2018]] ). Over North America, 12 RCMs were evaluated over the ARCTIC-CORDEX region ( [[#Diaconescu--2018|Diaconescu et al., 2018]] ). Models performed well at simulating climate indices related to mean air temperature and hot extremes over most of the Canadian Arctic, with the exception of the Yukon region where models displayed the largest biases related to topographic effects. Two RCMs were evaluated against observed extremes indices over North America over the period 1989–2009, with a cool bias in minimum temperature extremes shown in both RCMs ( [[#Whan--2016|Whan and Zwiers, 2016]] ). The most significant biases are found in TXx and TNn, with fewer differences in the simulation of annual minimum daily maximum temperature (TXn) and annual maximum daily minimum temperature (TNx) in Central and Western North America. Over Central and South America, maximum temperatures from the Eta RCM are generally underestimated, although hot days, warm nights, and heatwaves are increasing in the period 1961–1990, in agreement with observations ( [[#Chou--2014b|Chou et al., 2014b]] ; [[#Tencer--2016|Tencer et al., 2016]] ; [[#Bozkurt--2019|Bozkurt et al., 2019]] ).

Some land forcings are not well represented in climate models. As highlighted in the Special Report on Climate Change and Land (SRCCL) Chapter 2, there is ''high agreement'' that temperate deforestation leads to summer warming and winter cooling ( [[#Anderson--2011|Anderson et al., 2011]] ; [[#Gálos--2011|Gálos et al., 2011]] , 2013; [[#Anderson-Teixeira--2012|Anderson-Teixeira et al., 2012]] ; [[#Chen--2012|Chen et al., 2012]] ; [[#Wickham--2013|Wickham et al., 2013]] ; [[#Zhao--2014|Zhao and Jackson, 2014]] ; [[#Ahlswede--2017|Ahlswede and Thomas, 2017]] ; [[#Bright--2017|Bright et al., 2017]] ; [[#Strandberg--2019|Strandberg and Kjellström, 2019]] ), which has substantially contributed to the warming of hot extremes in the northern mid-latitudes over the course of the 20th century ( [[#Lejeune--2018|Lejeune et al., 2018]] ) and in recent years ( [[#Strandberg--2019|Strandberg and Kjellström, 2019]] ). However, observed forest effects on the seasonal and diurnal cycle of temperature are not well-captured in several ESMs: while observations show a cooling effect of forest cover compared to non-forest vegetation during daytime ( [[#Li--2015|Li et al., 2015]] ), in particular in arid, temperate, and tropical regions ( [[#Alkama--2016|Alkama and Cescatti, 2016]] ), several ESMs simulate a warming of daytime temperatures for regions with forest versus non-forest cover ( [[#Lejeune--2017|Lejeune et al., 2017]] ). Also irrigation effects, which can lead to regional cooling of temperature extremes, are generally not integrated in current generations of ESMs ( [[#11.3.1|Section 11.3.1]] ).

In summary, there is ''high confidence'' that climate models can reproduce the mean state and overall warming of temperature extremes observed globally and in most regions, although the magnitude of the trends may differ. The ability of models to capture observed trends in temperature-related extremes depends on the metric evaluated, the way indices are calculated, and the time periods and spatial scales considered. Regional climate models add value in simulating temperature-related extremes over GCMs in some regions. Some land forcings on temperature extremes are not well-captured (effects of deforestation) or generally not representated (irrigation) in ESMs.

<div id="11.3.4" class="h2-container"></div>

<span id="detection-and-attribution-event-attribution"></span>
=== 11.3.4 Detection and Attribution, Event Attribution ===

<div id="h2-27-siblings" class="h2-siblings"></div>

The SREX (IPCC, 2012) assessed that it is ''likely'' anthropogenic influences have led to the warming of extreme daily minimum and maximum temperatures at the global scale. The AR5 concluded that human influence has ''very likely'' contributed to the observed changes in the intensity and frequency of daily temperature extremes on the global scale in the second half of the 20th century (IPCC, 2014). With regard to individual, or regionally or locally specific events, AR5 concluded that it is ''likely'' human influence has substantially increased the probability of occurrence of heatwaves in some locations.

Studies since AR5 continue to attribute the observed increase in the frequency or intensity of hot extremes and the observed decrease in the frequency or intensity of cold extremes to human influence, dominated by anthropogenic greenhouse gas emissions, on global and continental scales, and for many AR6 regions. These include attribution of changes in the magnitude of annual TXx, TNx, TXn, and TNn, based on different observational datasets including, HadEX2 and HadEX3, CMIP5 and CMIP6 simulations, and different statistical methods ( [[#Kim--2016|Kim et al., 2016]] ; Z. [[#Wang--2017|]] [[#Wang--2017|Wang et al., 2017]] a; [[#Seong--2021|Seong et al., 2021]] ). As is the case for an increase in mean temperature ( [[IPCC:Wg1:Chapter:Chapter-3#3.3.1|Section 3.3.1]] ), an increase in extreme temperature is mostly due to greenhouse gas forcing, offset by aerosol forcing. The aerosols’ cooling effect is clearly detectable over Europe and Asia ( [[#Seong--2021|Seong et al., 2021]] ). As much as 75% of the moderate daily hot extremes (above 99.9th percentile) over land are due to anthropogenic warming ( [[#Fischer--2015|Fischer and Knutti, 2015]] ). New results are found to be more robust due to the extended period that improves the signal-to-noise ratio. The effect of anthropogenic forcing is clearly detectable and attributable in the observed changes in these indicators of temperature extremes, even at country and sub-country scales, such as in Canada ( [[#Wan--2019|Wan et al., 2019]] ). Changes in the number of warm nights, warm days, cold nights, and cold days, and other indicators such as the Warm Spell Duration Index (WSDI), are also attributed to anthropogenic influence ( [[#Christidis--2016|Christidis and Stott, 2016]] ; [[#Hu--2020|Hu et al., 2020]] ).

Regional studies, including for Asia ( [[#Dong--2018|Dong et al., 2018]] ; [[#Lu--2018|Lu et al., 2018]] ), Australia ( [[#Alexander--2017|Alexander and Arblaster, 2017]] ), and Europe ( [[#Christidis--2016|Christidis and Stott, 2016]] ), found similar results. A clear anthropogenic signal is also found in the trends in the Combined Extreme Index (CEI) for North America, Asia, Australia, and Europe ( [[#Dittus--2016|Dittus et al., 2016]] ). While various studies have described increasing trends in several heatwave metrics (heatwave duration, the number of heatwave days, etc.) in different regions (e.g., [[#Cowan--2014|Cowan et al., 2014]] ; [[#Bandyopadhyay--2016|Bandyopadhyay et al., 2016]] ; M. [[#Sanderson--2017|]] [[#Sanderson--2017|Sanderson et al., 2017]] ), few recent studies have explicitly attributed these changes to causes; most of them stated that observed trends are consistent with anthropogenic warming. The detected anthropogenic signals are clearly separable from the response to natural forcing, and the results are generally insensitive to the use of different model samples, as well as different data availability, indicating robust attribution. Studies of monthly, seasonal, and annual records in various regions ( [[#Kendon--2014|Kendon, 2014]] ; [[#Lewis--2015|Lewis and King, 2015]] ; [[#Bador--2016|Bador et al., 2016]] ; [[#Meehl--2016|Meehl et al., 2016]] ; [[#Zhou--2019|]] [[#Zhou--2019|C. Zhou et al., 2019]] ) and globally ( [[#King--2017|King, 2017]] ) show an increase in the breaking of hot records and a decrease in the breaking of cold records ( [[#King--2017|King, 2017]] ). Changes in anthropogenically attributablerecord-breaking rates are noted to be largest over the Northern Hemisphere land areas ( [[#Shiogama--2016|Shiogama et al., 2016]] ). Yin and Sun (2018) found clear evidence of an anthropogenic signal in the changes in the number of frost and ice days, when multiple model simulations were used. In some key wheat-producing regions of Southern Australia, increases in frost days or frost season length have been reported ( [[#Dittus--2014|Dittus et al., 2014]] ; [[#Crimp--2016|Crimp et al., 2016]] ); these changes are linked to decreases in rainfall, cloud-cover, and subtropical ridge strength, despite an overall increase in regional mean temperatures ( [[#Dittus--2014|Dittus et al., 2014]] ; [[#Pepler--2018|Pepler et al., 2018]] ).

A significant advance since AR5 has been a large number of studies focusing on extreme temperature events at monthly and seasonal scales, using various extreme event attribution methods. [[#Diffenbaugh--2017|Diffenbaugh et al. (2017)]] found that anthropogenic warming has increased the severity and probability of the hottest month by more than 80% of the available observational area on the global scale. [[#Christidis--2014|Christidis and Stott (2014)]] provide clear evidence that warm events have become more probable because of anthropogenic forcings. [[#Sun--2014|Sun et al. (2014)]] found that human influence has caused a more than 60-fold increase in the probability of the extreme warm 2013 summer in eastern China since the 1950s. Human influence is found to have increased the probability of the historically hottest summers in many regions of the world, both in terms of mean temperature ( [[#Mueller--2016|]] [[#Mueller--2016|B. Mueller et al., 2016]] ) and wet bulb globe temperature (WBGT; [[#Li--2017|]] [[#Li--2017|]] [[#Li--2017|]] [[#Li--2017|C. Li et al., 2017]] ). In most regions of the Northern Hemisphere, changes in the probability of extreme summer average WBGT were found to be about an order of magnitude larger than changes in the probability of extreme hot summers estimated by surface air temperature ( [[#Li--2017|]] [[#Li--2017|]] [[#Li--2017|]] [[#Li--2017|C. Li et al., 2017]] ). In addition to these generalized, global-scale approaches, extreme event studies have found an attributable increase in the probability of hot annual and seasonal temperatures in many locations, including Australia ( [[#Knutson--2014b|Knutson et al., 2014b]] ; [[#Lewis--2014|Lewis and Karoly, 2014]] ), China ( [[#Sun--2014|Sun et al., 2014]] ; [[#Sparrow--2018|Sparrow et al., 2018]] ; [[#Zhou--2020|Zhou et al., 2020]] ), Korea (Y.-H. [[#Kim--2018|]] [[#Kim--2018|]] [[#Kim--2018|Kim et al., 2018]] ) and Europe ( [[#King--2015b|King et al., 2015b]] ).

There have also been many extreme event attribution studies that examined short-duration temperature extremes, including daily temperatures, temperature indices, and heatwave metrics. Examples of these events from different regions are summarized in various annual Explaining Extreme Events supplements of the ''Bulletin of the American Meteorological Society'' ( [[#Peterson--2012|Peterson et al., 2012]] , 2013a; [[#Herring--2014|Herring et al., 2014]] , 2015, 2016, 2018, 2019, 2020), including a number of approaches to examine extreme events (described in [[#Easterling--2016|Easterling et al., 2016]] ; [[#Stott--2016|Stott et al., 2016]] ; [[#Otto--2017|Otto, 2017]] ). Several studies of recent events from 2016 onwards have determined an infinite risk ratio (a fraction of attributable risk, or FAR, of 1), indicating that the occurrence probability for such events is close to zero in model simulations without anthropogenic influences (see [[#Herring--2018|Herring et al., 2018]] , 2019, 2020; [[#Imada--2019|Imada et al., 2019]] ; [[#Vogel--2019|Vogel et al., 2019]] ). Though it is difficult to accurately estimate the lower bound of the uncertainty range of the FAR in these cases ( [[#Paciorek--2018|Paciorek et al., 2018]] ), the fact that those events are so far outside the envelop of the models with only natural forcing indicates that it is ''extremely unlikely'' for those events to occur without human influence.

Studies that focused on the attributable signal in observed cold extreme events show human influence reducing the probability of those events. Individual attribution studies on the extremely cold winter of 2011 in Europe ( [[#Peterson--2012|Peterson et al., 2012]] ), in the eastern USA during 2014 and 2015 ( [[#Trenary--2015|Trenary et al., 2015]] , 2016; [[#Wolter--2015|Wolter et al., 2015]] ; [[#Bellprat--2016|Bellprat et al., 2016]] ), in the cold spring of 2013 in the United Kingdom ( [[#Christidis--2014|Christidis et al., 2014]] ), and of 2016 in eastern China ( [[#Qian--2018|Qian et al., 2018]] ; Y. [[#Sun--2018b|]] [[#Sun--2018|Sun et al., 2018]] b ) all showed a reduced probability due to human influence on the climate. An exception is the study of [[#Grose--2018|Grose et al. (2018)]] , which found an increase in the probability of the severe western Australian frost of 2016 due to anthropogenically-driven changes in circulation patterns that drive cold outbreaks and frost probability.

Different event attribution studies can produce a wide range of changes in the probability of event occurrence because of different framing. The temperature event definition itself plays a crucial role in the attributable signal ( [[#Fischer--2015|Fischer and Knutti, 2015]] ; Kirchmeier‐Young et al., 2019). Large-scale, longer-duration events tend to have notably larger attributable risk ratios ( [[#Angélil--2014|Angélil et al., 2014]] , 2018; [[#Uhe--2016|Uhe et al., 2016]] ; [[#Harrington--2017|Harrington, 2017]] ; Kirchmeier‐Young et al., 2019), as natural variability is smaller. While uncertainty in the best estimates of the risk ratios may be large, their lower bounds can be quite insensitive to uncertainties in observations or model descriptions, thus increasing confidence in conservative attribution statements ( [[#Jeon--2016|Jeon et al., 2016]] ).

The relative strength of anthropogenic influences on temperature extremes is regionally variable, in part due to differences in changes in atmospheric circulation, land–surface feedbacks, and other external drivers such as aerosols. For example, in the Mediterranean and over western Europe, risk ratios on the order of 100 have been found ( [[#Kew--2019|Kew et al., 2019]] ; [[#Vautard--2020|Vautard et al., 2020]] ), whereas in the USA, changes are much less pronounced. This is probably a reflection of the land–surface feedback enhanced extreme 1930s temperatures that reduce the rarity of recent extremes, in addition to the definition of the events and framing of attribution analyses (e.g., spatial and temporal scales considered). Local forcing may mask or enhance the warming effect of greenhouse gases. In India, short-lived aerosols or an increase in irrigation may be masking the warming effect of greenhouse gases ( [[#Wehner--2018c|Wehner et al., 2018c]] ). Irrigation and crop intensification have been shown to lead to a cooling in some regions, in particular in North America, Europe, and India ( ''high confidence'' ) (N.D. [[#Mueller--2016|]] [[#Mueller--2016|Mueller et al., 2016]] ; [[#Thiery--2017|Thiery et al., 2017]] , 2020; [[#Chen--2019|Chen and Dirmeyer, 2019]] ). Deforestation has contributed about one third of the total warming of hot extremes in some mid-latitude regions since pre-industrial times ( [[#Lejeune--2018|Lejeune et al., 2018]] ). Despite all of these differences, and larger uncertainties at the regional scale, nearly all studies demonstrated that human influence has contributed to an increase in the frequency or intensity of hot extremes and to a decrease in the frequency or intensity of cold extremes.

In summary, long-term changes in various aspects of long- and short-duration extreme temperatures, including intensity, frequency, and duration have been detected in observations and attributed to human influence at global and continental scales. It is ''extremely likely'' that human influence is the main contributor to the observed increase in the intensity and frequency of hot extremes and the observed decrease in the intensity and frequency of cold extremes on the global scale. It is ''very likely'' that this applies on continental scales as well. Some specific recent hot extreme events would have been ''extremely unlikely'' to occur without human influence on the climate system. Changes in aerosol concentrations have affected trends in hot extremes in some regions, with the presence of aerosols leading to attenuated warming, in particular from 1950 to 1980. Crop intensification, irrigation and no-till farming have attenuated increases in summer hot extremes in some regions, such as Central North America ( ''medi'' ''um confidence'' ).

<div id="11.3.5" class="h2-container"></div>

<span id="projections"></span>
=== 11.3.5 Projections ===

<div id="h2-28-siblings" class="h2-siblings"></div>

The AR5 (Chapter12, [[#Collins--2013|Collins et al., 2013]] ) concluded that it is ''virtually certain'' there will be more frequent hot extremes and fewer cold extremes at the global scale and over most land areas in a future warmer climate, and it is ''very likely'' that heatwaves will occur with a higher frequency and longer duration.The SR1.5 (Chapter 3, [[#Hoegh-Guldberg--2018|Hoegh-Guldberg et al., 2018]] ) assessment on projected changes in hot extremes at 1.5°C and 2°C global warming is consistent with the AR5 assessment, concluding that it is ''very likely'' a global warming of 2°C, when compared with a 1.5°C warming, would lead to more frequent and more intense hot extremes on land, as well as to longer warm spells, affecting many densely inhabited regions. The SR1.5 also assessed it is ''very likely'' that the strongest increases in the frequency of hot extremes are projected for the rarest events, while cold extremes will become less intense and less frequent, and cold spells will be shorter.

New studies since AR5 and SR1.5 confirm these assessments. New literature since AR5 includes projections of temperature-related extremes in relation to changes in mean temperatures, projections based on CMIP6 simulations, projections based on stabilized global warming levels, and the use of new metrics. Constraints for the projected changes in hot extremes were also provided ( [[#Borodina--2017b|Borodina et al., 2017b]] ; [[#Sippel--2017b|Sippel et al., 2017b]] ; [[#Vogel--2017|Vogel et al., 2017]] ). Overall, projected changes in the magnitude of extreme temperatures over land are larger than changes in global mean temperature, over mid-latitude land regions in particular (Figures 11.3, 11.11; [[#Fischer--2014|Fischer et al., 2014]] ; [[#Seneviratne--2016|Seneviratne et al., 2016]] ; B.M. [[#Sanderson--2017|]] [[#Sanderson--2017|Sanderson et al., 2017]] ; [[#Wehner--2018b|Wehner et al., 2018b]] ; [[#Di%20Luca--2020b|Di Luca et al., 2020b]] ). Large warming in hot and cold extremes will occur, even at the 1.5°C GWL (Figure 11.11). At this level, widespread significant changes at the grid-box level occur for different temperature indices ( [[#Aerenson--2018|Aerenson et al., 2018]] ). In agreement with CMIP5 projections, CMIP6 simulations show that a 0.5°C increment in global warming will significantly increase the intensity and frequency of hot extremes, and decrease the intensity and frequency of cold extremes on the global scale (Figures 11.6, 11.8 and 11.12). It takes less than half of a degree for the changes in TXx to emerge above the level of natural variability (Figure 11.8) and the 66% ranges of the land medians of the 10-year or 50-year TXx events do not overlap between 1.0°C and 1.5°C in the CMIP6 multi-model ensemble simulations(Figure 11.6, [[#Li--2021|Li et al., 2021]] ).

<div id="_idContainer051" class="Basic-Text-Frame"></div>

[[File:f76a530dedf85907da4303bd377c6445 IPCC_AR6_WGI_Figure_11_11.png]]

'''Figure 11.11 |''' '''Projected changes in (a–c) annual maximum temperature (TXx) and (d–f) annual minimum temperature (TNn) at 1.5°C, 2°C, and 4°C of global warming compared to the 1850–1900 baseline.''' Results are based on simulations from the Coupled Model Intercomparison Project Phase 6 (CMIP6) multi-model ensemble under the Shared Socio-economic Pathways (SSPs) SSP1-1.9, SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5 scenarios. The numbers in the top right indicate the number of simulations included. Uncertainty is represented using the simple approach: no overlay indicates regions with high model agreement, where ≥80% of models agree on the sign of change; diagonal lines indicate regions with low model agreement, where &lt;80% of models agree on the sign of change. For more information on the simple approach, please refer to the Cross-Chapter Box ( [[IPCC:Wg1:Chapter:Atlas|Atlas]] 1. For details on the methods see Supplementary Material 11.SM.2. Changes in TXx and TNn are also displayed in the Interactive Atlas. Further details on data sources and processing are available in the chapter data table (Table 11.SM.9).

<div id="_idContainer053" class="Basic-Text-Frame"></div>

[[File:edd08452914b028e39967547a032c0ea IPCC_AR6_WGI_Figure_11_12.png]]

'''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 global warming levels relative to the 185''' ''0–1900 baseline.'' Extreme temperature events are defined as the daily maximum temperatures (TXx) that were exceeded on average once during a 10-year period (10-year event, blue) and once during a 50-year period (50-year event, orange) during the 1850–1900 base period. Results are shown for the global land. For each box plot, the horizontal line and the box represent the median and central 66% uncertainty range, respectively, of the intensity changes across the multi-model ensemble, and the ‘whiskers’ extend to the 90% uncertainty range. The results are based on the multi-model ensemble from simulations of global climate models contributing to the Coupled Model Intercomparison Project Phase 6 (CMIP6) under different Shared Socio-economic Pathway forcing scenarios. Adapted from [[#Li--2021|Li et al. (2021)]] . Further details on data sources and processing are available in the chapter data table (Table 11.SM.9).

Projected warming is larger for TNn and exhibits strong equator-to-pole amplification, similar to the warming of boreal winter mean temperatures. The warming of TXx is more uniform over land and does not exhibit this behaviour (Figure 11.11). The warming of temperature extremes on global and regional scales tends to scale linearly with global warming ( [[#11.1.4|Section 11.1.4]] ; Fischer et al., 2014; [[#Seneviratne--2016|Seneviratne et al., 2016]] ; [[#Wartenburger--2017|Wartenburger et al., 2017]] ; [[#Li--2021|Li et al., 2021]] ; see also SR1.5, Chapter 3). In the mid-latitudes, the rate of warming of hot extremes can be as large as twice the rate of global warming (Figure 11.11). In the Arctic winter, the rate of warming of the temperature of the coldest nights is about three times the rate of global warming (Appendix, Figure 11.A.1). Projected changes in temperature extremes can deviate from projected changes in annual mean warming in the same regions (Figures 11.3, 11.A.1 and 11.A.2; [[#Di%20Luca--2020b|Di Luca et al., 2020b]] ; [[#Wehner--2020|Wehner, 2020]] ) due to the additional processes that control the response of regional extremes, including, in particular, soil moisture–evapotranspiration–temperature feedbacks for hot extremes in the mid-latitudes and subtropical regions, and snow/ice–albedo–temperature feedbacks in high-latitude regions.

The probability of exceeding a certain hot extreme threshold will increase, while those for cold extreme will decrease with global warming ( [[#Mueller--2016|]] [[#Mueller--2016|B. Mueller et al., 2016]] ; [[#Lewis--2017b|Lewis et al., 2017b]] ; [[#Suarez-Gutierrez--2020b|Suarez-Gutierrez et al., 2020b]] ). The changes tend to scale nonlinearly with the level of global warming, with larger changes for more rare events ( [[#11.2.4|Section 11.2.4]] ; Cross-Chapter Box 11.11; Figures 11.6 and 11.12; [[#Fischer--2015|Fischer and Knutti, 2015]] ; [[#Kharin--2018|Kharin et al., 2018]] ; [[#Li--2021|Li et al., 2021]] ). For example, the CMIP5 ensemble projects the frequency of the present-day climate 20-year hottest daily temperature to increase by 80% at the 1.5°C GWL and by 180% at the 2.0°C GWL, and the frequency of the present-day climate 100-year hottest daily temperature to increase by 200% and more than 700% at the 1.5°C and 2.0°C warming levels, respectively ( [[#Kharin--2018|Kharin et al., 2018]] ). CMIP6 simulations project similar changes ( [[#Li--2021|Li et al., 2021]] ).

[[#Tebaldi--2018|Tebaldi and Wehner (2018)]] showed that, at the middle of the 21st century, 66% of the land surface area would experience the present-day 20-year return values of TXx and the running three-day average of the daily maximum temperature every other year, on average, under the Representative Concentration Pathway 8.5 (RCP8.5) scenario, as opposed to only 34% under RCP4.5. By the end of the century, these area fractions increase to 92% and 62%, respectively. Such nonlinearities in the characteristics of future regional extremes are shown, for instance, for Europe ( [[#Dosio--2018|Dosio and Fischer, 2018]] ; [[#Spinoni--2018b|Spinoni et al., 2018b]] ; [[#Lionello--2020|Lionello and Scarascia, 2020]] ), Asia ( [[#Guo--2017|Guo et al., 2017]] ; [[#Harrington--2018b|Harrington and Otto, 2018b]] ; [[#King--2018|King et al., 2018]] ), and Australia ( [[#Lewis--2017a|Lewis et al., 2017a]] ) under various global warming thresholds. The nonlinear increase in fixed-threshold indices (e.g., based on a percentile for a given reference period, or on an absolute threshold) as a function of global warming is consistent with a linear warming of the absolute temperature of the temperature extremes (e.g., [[#Whan--2015|Whan et al., 2015]] ). Compared to the historical climate, warming will result in strong increases in heatwave area, duration and magnitude ( [[#Vogel--2020b|Vogel et al., 2020b]] ). These changes are mostly due to the increase in mean seasonal temperature, rather than changes in temperature variability, though the latter can have an effect in some regions ( [[#Brown--2020|Brown, 2020]] ; [[#Di%20Luca--2020b|Di Luca et al., 2020b]] ; [[#Suarez-Gutierrez--2020a|Suarez-Gutierrez et al., 2020a]] ).

Projections of temperature-related extremes in RCMs in the CORDEX regions demonstrate robust increases under future scenarios and can provide information on finer spatial scales than GCMs (e.g., [[#Coppola--2021b|Coppola et al., 2021b]] ). Five RCMs in the CORDEX–East Asia region project increases in the 20-year return values of temperature extremes (summer maxima), with models that exhibit warm biases projecting stronger warming ( [[#Park--2019|Park and Min, 2019]] ). Similarly, in the African domain, future increases in TX90p and TN90p are projected ( [[#Dosio--2017|Dosio, 2017]] ; [[#Mostafa--2019|Mostafa et al., 2019]] ). This regional-scale analysis provides fine-scale information, such as distinguishing the increase in TX90p over sub-equatorial Africa (Democratic Republic of the Congo, Angola, and Zambia) with values over the Gulf of Guinea, Central African Republic, South Sudan, and Ethiopia. Empirical statistical downscaling has also been used to produce more robust estimates for future heatwaves compared to RCMs based on large multi-model ensembles ( [[#Furrer--2010|Furrer et al., 2010]] ; [[#Keellings--2014|Keellings and Waylen, 2014]] ; [[#Wang--2015|Wang et al., 2015]] ; [[#Benestad--2018|Benestad et al., 2018]] ).

In all continental regions, including Africa (Table 11.4), Asia (Table 11.7), Australasia (Table 11.10), Central and South America (Table 11.13), Europe (Table 11.16), North America (Table 11.19) and at the continental scale, it is ''very likely'' that the intensity and frequency of hot extremes will increase and the intensity and frequency of cold extremes will decrease compared with the 1995–2014 baseline, even under 1.5°C global warming. Those changes are ''virtually certain'' to occur under 4°C global warming. At the regional scale, and for almost all AR6 regions, it is ''likely'' that the intensity and frequency of hot extremes will increase and the intensity and frequency of cold extremes will decrease compared with the 1995–2014 baseline, even under 1.5°C global warming. Those changes are ''virtually certain'' to occur under 4°C global warming. Exceptions include lower confidence in the projected decrease in the intensity and frequency of cold extremes compared with the 1995–2014 baseline under 1.5°C of global warming ( ''medium confidence'' ) and 4°C of global warming ( ''very likely'' ) in Northern Central America, Central North America, and Western North America.

In Africa (Table 11.4), evidence includes increases in the intensity and frequency of hot extremes, such as warm days, warm nights, and heatwaves, and decreases in the intensity and frequency of cold extremes, such as cold days and cold nights over the continent, as projected by CMIP5, CMIP6, and CORDEX simulations (Giorgi et al., 2014; [[#Engelbrecht--2015|Engelbrecht et al., 2015]] ; [[#Lelieveld--2016|Lelieveld et al., 2016]] ; [[#Russo--2016|Russo et al., 2016]] ; [[#Dosio--2017|Dosio, 2017]] ; [[#Bathiany--2018|Bathiany et al., 2018]] ; [[#Mba--2018|Mba et al., 2018]] ; [[#Nangombe--2018|Nangombe et al., 2018]] ; [[#Weber--2018|Weber et al., 2018]] ; [[#Kruger--2019|Kruger et al., 2019]] ; [[#Coppola--2021b|Coppola et al., 2021b]] ; [[#Li--2021|Li et al., 2021]] ). Cold spells are projected to decrease under all RCPs, and even at low warming levels in Western and Central Africa ( [[#Diedhiou--2018|Diedhiou et al., 2018]] ). The number of cold days is projected to decrease in East Africa ( [[#Ongoma--2018b|Ongoma et al., 2018b]] ).

In Asia (Table 11.7), evidence includes increases in the intensity and frequency of hot extremes, such as warm days, warm nights, and heatwaves, and decreases in the intensity and frequency of cold extremes, such as cold days and cold nights over the continent, as projected by CMIP5, CMIP6, and CORDEX simulations ( [[#Sillmann--2013b|Sillmann et al., 2013b]] ; [[#Zhou--2014|Zhou et al., 2014]] ; R. [[#Zhang--2015|Zhang et al., 2015]] ; [[#Zhao--2015|Zhao et al., 2015]] ; [[#Pal--2016|Pal and Eltahir, 2016]] ; [[#Singh--2016|Singh and Goyal, 2016]] ; [[#Xu--2017|Xu et al., 2017]] ; [[#Gao--2018|Gao et al., 2018]] ; [[#Han--2018|Han et al., 2018]] ; [[#Shin--2018|Shin et al., 2018]] ; [[#Sui--2018|Sui et al., 2018]] ; L. [[#Li--2019|]] [[#Li--2019|]] [[#Li--2019|Li et al., 2019]] ; [[#Zhu--2020|Zhu et al., 2020]] ). More intense heatwaves of longer durations and occurring at a higher frequency are projected over India ( [[#Murari--2015|Murari et al., 2015]] ; [[#Mishra--2017|Mishra et al., 2017]] ) and Pakistan ( [[#Nasim--2018|Nasim et al., 2018]] ). Future mid-latitude warm extremes, similar to those experienced during the 2010 event, are projected to become more extreme, with temperature extremes increasing potentially by 8.4°C (RCP8.5) over north-west Asia ( [[#van%20der%20Schrier--2018|van der Schrier et al., 2018]] ). Over West and East Siberia, and Russian Far East, an increase in extreme heat durations is expected in all scenarios ( [[#Sillmann--2013b|Sillmann et al., 2013b]] ; [[#Kattsov--2017|Kattsov et al., 2017]] ; [[#Reyer--2017|Reyer et al., 2017]] ). In the MENA regions (Arabian Peninsula and Western Central Asia), extreme temperatures could increase by almost 7°C by 2100 under RCP8.5 ( [[#Lelieveld--2016|Lelieveld et al., 2016]] ).

In Australasia (Table 11.10), evidence includes increases in the intensity and frequency of hot extremes, such as warm days, warm nights, and heatwaves, and decreases in the intensity and frequency of cold extremes, such as cold days and cold nights over the continent, as projected by CMIP5, CMIP6, and CORDEX simulations (CSIROand BOM, 2015; [[#Alexander--2017|Alexander and Arblaster, 2017]] ; [[#Lewis--2017a|Lewis et al., 2017a]] ; [[#Herold--2018|Herold et al., 2018]] ; [[#Coppola--2021b|Coppola et al., 2021b]] ; [[#Evans--2021|Evans et al., 2021]] ). Over most of Australia, increases in the intensity and frequency of hot extremes are projected to be predominantly driven by the long-term increase in mean temperatures ( [[#Di%20Luca--2020b|Di Luca et al., 2020b]] ). Future projections indicate a decrease in the number of frost days regardless of the region and season considered ( [[#Alexander--2017|Alexander and Arblaster, 2017]] ; [[#Herold--2018|Herold et al., 2018]] ).

In Central and South America (Table 11.13), evidence includes increases in the intensity and frequency of hot extremes, such as warm days, warm nights, and heatwaves, and decreases in the intensity and frequency of cold extremes, such as cold days and cold nights over the continent, as projected by CMIP5, CMIP6, and CORDEX simulations ( [[#Chou--2014a|Chou et al., 2014a]] ; [[#Cabré--2016|Cabré et al., 2016]] ; [[#López-Franca--2016|López-Franca et al., 2016]] ; [[#Stennett-Brown--2017|Stennett-Brown et al., 2017]] ; [[#Coppola--2021b|Coppola et al., 2021b]] ; [[#Li--2021|Li et al., 2021]] ; [[#Vichot-Llano--2021|Vichot-Llano et al., 2021]] ). Over South-Eastern South America during the austral summer, the increase in the frequency of TN90p is larger than that projected for TX90p, consistent with observed past changes ( [[#López-Franca--2016|López-Franca et al., 2016]] ). Under RCP8.5, the number of heatwave days are projected to increase for the intra-Americas region for the end of the 21st century ( [[#Angeles-Malaspina--2018|Angeles-Malaspina et al., 2018]] ). A general decrease in the frequency of cold spells and frost days is projected, as indicated by several indices based on minimum temperature ( [[#López-Franca--2016|López-Franca et al., 2016]] ).

In Europe (Table 11.16), evidence includes increases in the intensity and frequency of hot extremes, such as warm days, warm nights, and heatwaves, and decreases in the intensity and frequency of cold extremes, such as cold days and cold nights over the continent, as projected by CMIP5, CMIP6, and CORDEX simulations ( [[#Lau--2014|Lau and Nath, 2014]] ; [[#Ozturk--2015|Ozturk et al., 2015]] ; [[#Russo--2015|Russo et al., 2015]] ; [[#Schoetter--2015|Schoetter et al., 2015]] ; [[#Vogel--2017|Vogel et al., 2017]] ; [[#Winter--2017|Winter et al., 2017]] ; [[#Jacob--2018|Jacob et al., 2018]] ; [[#Lhotka--2018|Lhotka et al., 2018]] ; [[#Rasmijn--2018|Rasmijn et al., 2018]] ; [[#Suarez-Gutierrez--2018|Suarez-Gutierrez et al., 2018]] ; [[#Cardoso--2019|Cardoso et al., 2019]] ; [[#Lionello--2020|Lionello and Scarascia, 2020]] ; [[#Molina--2020|Molina et al., 2020]] ; [[#Coppola--2021b|Coppola et al., 2021b]] ; [[#Li--2021|Li et al., 2021]] ). Increases in heatwaves are greater over the southern Mediterranean and Scandinavia ( [[#Forzieri--2016|Forzieri et al., 2016]] ; [[#Abaurrea--2018|Abaurrea et al., 2018]] ; [[#Dosio--2018|Dosio and Fischer, 2018]] ; [[#Rohat--2019|Rohat et al., 2019]] ). Thebiggest increases in the number of heatwave days are expected for southern European cities ( [[#Guerreiro--2018a|Guerreiro et al., 2018a]] ; [[#Junk--2019|Junk et al., 2019]] ), and Central European cities will see the biggest increases in maximum heatwave temperatures ( [[#Guerreiro--2018a|Guerreiro et al., 2018a]] ).

In North America (Table 11.19), evidence includes increases in the intensity and frequency of hot extremes, such as warm days, warm nights, and heatwaves, and decreases in the intensity and frequency of cold extremes, such as cold days and cold nights over the continent, as projected by CMIP5, CMIP6, and CORDEX simulations ( [[#Grotjahn--2016|Grotjahn et al., 2016]] ; [[#Vose--2017|Vose et al., 2017]] ; [[#Alexandru--2018|Alexandru, 2018]] ; [[#Li--2018|]] [[#Li--2018|]] [[#Li--2018|]] [[#Li--2018|C. Li et al., 2018]] , 2021; [[#Yang--2018|]] [[#Yang--2018|]] [[#Yang--2018|]] [[#Yang--2018|]] [[#Yang--2018|C. Yang et al., 2018]] ; X. [[#Zhang--2019|]] [[#Zhang--2019|]] [[#Zhang--2019|]] [[#Zhang--2019|Zhang et al., 2019]] ; [[#Coppola--2021b|Coppola et al., 2021b]] ). Projections of temperature extremes for the end of the 21st century show that warm days and nights are ''very likely'' to increase, and cold days and nights are ''very likely'' to decrease in all regions. There is ''medium confidence'' in large increases in warm days and warm nights in summer, particularly over the USA, and in large decreases in cold days in Canada in autumn and winter ( [[#Grotjahn--2016|Grotjahn et al., 2016]] ; [[#Vose--2017|Vose et al., 2017]] ; [[#Alexandru--2018|Alexandru, 2018]] ; [[#Li--2018|]] [[#Li--2018|]] [[#Li--2018|]] [[#Li--2018|C. Li et al., 2018]] , 2021; [[#Yang--2018|]] [[#Yang--2018|]] [[#Yang--2018|]] [[#Yang--2018|]] [[#Yang--2018|C. Yang et al., 2018]] ; X. [[#Zhang--2019|]] [[#Zhang--2019|]] [[#Zhang--2019|]] [[#Zhang--2019|Zhang et al., 2019]] ; [[#Coppola--2021b|Coppola et al., 2021b]] ). Minimum winter temperatures are projected to rise faster than mean winter temperatures ( [[#Underwood--2017|Underwood et al., 2017]] ). Projections for the end of the century under RCP8.5 showed the four-day cold spell that happens on average once every five years is projected to warm by more than 10°C. CMIP5 models do not project current 1-in-20-year annual minimum temperature extremes to recur over much of the continent ( [[#Wuebbles--2014|Wuebbles et al., 2014]] ).

In summary, it is ''virtually certain'' that further increases in the intensity and frequency of hot extremes, and decreases in the intensity and frequency of cold extremes, will occur throughout the 21st century and around the world. It is ''virtually certain'' that the number of hot days and hot nights and the length, frequency, and/or intensity of warm spells or heatwaves compared to 1995–2014 will increase over most land areas. In most regions, changes in the magnitude of temperature extremes are proportional to global warming levels ( ''high confidence'' ). The highest increase of temperature of hottest days is projected in some mid-latitude and semi-arid regions, at about 1.5 times to twice the rate of global warming ( ''high confidence'' ). The highest increase of temperature of coldest days is projected in Arctic regions, at about three times the rate of global warming ( ''high confidence'' ). The probability of temperature extremes generally increases nonlinearly with increasing global warming levels ( ''high confidence'' ). Confidence in assessments depends on the spatial and temporal scales of the extreme in question, with ''high confidence'' in projections of temperature-related extremes at global and continental scales for daily to seasonal scales. There is ''high confidence'' that, on land, the magnitude of temperature extremes increases more strongly than global mean temperature.

<div id="11.4" class="h1-container"></div>

<span id="heavy-precipitation"></span>