Abstract

This study provides a first quantification of possible benefits of global warming mitigation through heat stress reduction over East Asia by comparing projection results between low-emission and high-emission scenarios, as well as between 1.5° and 2.0°C target temperature conditions. Future changes in summer heat stress over East Asia were examined based on the wet-bulb globe temperature (WBGT) using CMIP5 multimodel simulations. Changes in the intensity, frequency, and duration of heat stress were analyzed in terms of area fraction across RCP2.6, RCP4.5, and RCP8.5 scenarios and also between two selected model groups representing 1.5°- and 2.0°C-warmer worlds. Severe heat stress, exceeding the 50-yr return value of the present-day period, is expected to become very frequent, occurring every second year over the large part of East Asia by the 2040s, irrespective of RCP scenarios. The frequency of extreme daily heat stress events is predicted to increase in a similar speed of expansion, with signals emerging from the low latitudes. The WBGT signal emergence is found to be much faster than that of corresponding temperature alone due to the smaller variability in WBGT, supporting previous findings. The 1.5°C-warmer world would have about 20% reduction in areas experiencing severe heat stress over East Asia, compared to the 2.0°C-warmer world, with significant changes identified over the low latitudes. Further, compared to the transient world, the equilibrium world exhibits larger increases in heat stress over East Asia, likely due to the warmer ocean surface in the northwestern North Pacific. This suggests an important role of ocean warming patterns in the regional assessment of global warming mitigation.

1. Introduction

The probability of heat waves has been increasing over many parts of the world, in line with global warming, and the intensity and duration of heat waves have also increased (Meehl and Tebaldi 2004; Perkins et al. 2012). In addition, the percentage of land area with increased summer temperatures exceeding a certain threshold (e.g., 5 sigma) is projected to grow rapidly and include approximately 60% of the world by the late twenty-first century (Coumou and Robinson 2013). Extreme weather events, like heat stress, usually occur locally, with associated impacts and physical mechanisms being very different among regions (Seneviratne et al. 2012). Accordingly, regional-scale studies are fundamentally required to project future changes in extreme events and assess their impacts on natural and human systems. East Asia is one of the most vulnerable regions, experiencing an increasing number of heat waves (Christensen et al. 2007, 2013; Min et al. 2014; 2015), and it is also expected that this warming trend will continue in the twenty-first century (Yun et al. 2012; Baek et al. 2013; Collins et al. 2013; Ji and Kang 2015).

In addition to extreme high temperatures, several variables, such as humidity, solar radiation, and air pollution, affect heat-related illness and mortality (Fischer and Knutti 2013). Therefore, it is important to analyze not only temperature but also these other variables in order to comprehensively understand the effects of heat stress. Recent studies additionally considered humidity in assessing heat wave impacts on human health. Mora et al. (2017) identified a global threshold for deadly climatic conditions based on surface air temperature and relative humidity, finding that the percentage of the world’s population exposed to the deadly heat will increase from the current value of 30% up to 75% under high-emission scenarios. Russo et al. (2017) analyzed past and future humid heat waves at different global warming levels using the apparent temperature (Steadman 1979), showing that humidity increase can amplify heat wave magnitude and peak over the eastern United States and China. By applying high wet-blub temperatures that represent hot and humid conditions as a threshold for human adaptability (Sherwood and Huber 2010), Im et al. (2017) predicted future intensification of extreme heat waves in South Asia under high-emission scenarios. Other previous studies assessed the adverse effects of heat stress on health and labor productivity (Kjellstrom et al. 2009; Dunne et al. 2013), which is induced by the metabolic efficiency decrease of human beings resulting from high temperature and high humidity (Basu and Samet 2002; Sherwood and Huber 2010). Willett and Sherwood (2012) investigated the observed extent of summer heat stress over 15 regions during 1973–2003 and found overall increases in heat stress.

Recently, the United Nations Framework Convention on Climate Change (UNFCCC) proposed two long-term global goals in the 2015 Paris Agreement (UNFCCC 2015): “holding the increase in the global average temperature to well below 2.0°C above preindustrial levels and pursuing efforts to limit the temperature increase to 1.5°C above preindustrial levels.” Assessment of climate change through target temperatures can be a more efficient analysis in terms of regional assessment and impact (Seneviratne et al. 2016). Despite the need for effective and scientific information about the impacts of these target temperatures by policy makers, the assessment and comparative analysis of climate change impacts under these temperature-increase scenarios has been limited thus far. Therefore, this study aims to analyze the future changes in summer heat stress over East Asia using multimodel simulations integrated under different greenhouse gas emission scenarios and also to evaluate possible benefits of global warming mitigation under the two target temperatures. We focus on the spatial extent of heat stress by analyzing the changes in areas experiencing extreme heat stress over the East Asian land regions.

2. Data and methods

In this study, wet-bulb globe temperature (WBGT) was used as a heat stress index following previous studies (Fischer and Knutti 2013; Willett and Sherwood 2012; Knutson and Ploshay 2016). WBGT is an index designed by Yaglou and Minard (1957) to measure the heat stress of a soldier at an American military training center and is standardized by ISO 7243 (ISO 1989). Different equations are used to calculate WBGT, depending on the presence of solar radiation. Equation (1) is used to evaluate the heat stress in the outdoors where solar radiation is present, and Eq. (2) is used to evaluate the heat stress in the indoors or outdoors without solar radiation:

 
formula
 
formula

where Tw is wet-bulb temperature (°C), Ta is air temperature (°C), and Tg is black-globe temperature (°C). Given the difficulty to obtain black-globe temperature (Lemke and Kjellstrom 2012), a simple formula, as in Eq. (3) below, was devised by the American College of Sports Medicine (ACSM 1984), and many previous studies on long-term changes in heat stress used this method (Fischer and Knutti 2013; Willett and Sherwood 2012; Knutson and Ploshay 2016):

 
formula
 
formula

where e is vapor pressure (hPa) and RH is relative humidity (%). Here, we calculate vapor pressure from relative humidity and air temperature using Eq. (4). This “simplified WBGT” formula ignores the effects of wind and radiation on thermal stress. However, wind and radiation are unlikely to have a significant effect on the trend of heat stress (Willett and Sherwood 2012).

We defined the East Asian domain as the area bounded by 20°–50°N, 100°–150°E, and focused on the summer season (JJA). To examine changes in the characteristics of heat stress, summer mean WBGT () and the summer frequency () and maximum duration () of daily extreme WBGT were analyzed (Table 1). For each variable, we calculated the fraction of the area over East Asia, exceeding present-day thresholds for each year during the future period (2011–99). The present-day thresholds for were estimated as follows. We first calculated the JJA mean WBGT for each year during the 60-yr historical period (1951–2010) using monthly data on each grid. Then, the 50-yr return value of was obtained from the historical period and used as the threshold for future exceedance. The present-day thresholds for and were estimated similarly, but considering daily WBGT extremes. First, in order to eliminate the seasonality influence, the daily 95th percentile of WBGT was calculated from the historical data (1951–2010), in which 5-day moving averages were applied to smooth interdaily variations. After that, was calculated by counting the number of days with daily WBGT exceeding the 95th percentile during each summer of the historical period, while was obtained by finding the maximum consecutive days out of the heat stress events. Finally, the 50-yr return values of and were obtained to be used as thresholds for future exceedance (Table 1).

Table 1.

Descriptions of the WBGT-based indices for heat stress intensity, frequency, and duration and their present-day thresholds for the analysis of future emergence.

Descriptions of the WBGT-based indices for heat stress intensity, frequency, and duration and their present-day thresholds for the analysis of future emergence.
Descriptions of the WBGT-based indices for heat stress intensity, frequency, and duration and their present-day thresholds for the analysis of future emergence.

Monthly and daily data of surface air temperature (TAS) and near-surface relative humidity (RH) from multimodel simulations for phase 5 of the Coupled Model Intercomparison Project (CMIP5; Taylor et al. 2012) were used to analyze the future changes in summer heat stress over East Asia and its characteristics. We combined historical simulations integrated with anthropogenic (greenhouse gases and aerosols) plus natural (solar and volcanic activities) forcings (1951–2005) with the representative concentration pathway 4.5 (RCP4.5) scenario simulations (2006–10), producing a 60-yr-long time series for a present-day condition including historical human influences. Future projections are based on RCP2.6, RCP4.5, and RCP8.5 scenario simulations for the period of 2011–99 (Table 2).

Table 2.

CMIP5 model simulations used in this study. All models provide historical (1951–2005) and RCP scenario runs (2006–99). Ensemble member from each model is provided such that, for example, r1i1p1 represents a first realization of the first version of the perturbed physics model. Selected models for 1.5° and 2.0°C warming in the equilibrium world are indicated, with global mean temperature changes relative to preindustrial level (each 1861–1900 mean) being provided in parentheses. Ensemble member marked by an asterisk indicates runs that provide data for both equilibrium worlds.

CMIP5 model simulations used in this study. All models provide historical (1951–2005) and RCP scenario runs (2006–99). Ensemble member from each model is provided such that, for example, r1i1p1 represents a first realization of the first version of the perturbed physics model. Selected models for 1.5° and 2.0°C warming in the equilibrium world are indicated, with global mean temperature changes relative to preindustrial level (each 1861–1900 mean) being provided in parentheses. Ensemble member marked by an asterisk indicates runs that provide data for both equilibrium worlds.
CMIP5 model simulations used in this study. All models provide historical (1951–2005) and RCP scenario runs (2006–99). Ensemble member from each model is provided such that, for example, r1i1p1 represents a first realization of the first version of the perturbed physics model. Selected models for 1.5° and 2.0°C warming in the equilibrium world are indicated, with global mean temperature changes relative to preindustrial level (each 1861–1900 mean) being provided in parentheses. Ensemble member marked by an asterisk indicates runs that provide data for both equilibrium worlds.

To analyze differences in the East Asian heat stress under different target temperatures (1.5° vs 2.0°C), we first calculated the future changes in global mean (land plus ocean) temperature changes during the late twenty-first century (2070–99) with respect to the preindustrial level (1861–1900) from all available models. Here, only RCP2.6 and RCP4.5 scenario simulations were considered because they are scenarios that stabilize radiative forcing at 2.6 and 4.5 W m−2 by 2100 (van Vuuren et al. 2011; Thomson et al. 2011). Model runs with global mean temperature changes (future minus preindustrial) ranging from +1.3° to +1.7°C were selected to represent a 1.5°C-warmer world, while models with changes ranging from +1.8° to +2.2°C were classified into the model group for the 2.0°C-warmer world (see Table 2 for selected model runs). The heat stress changes were then compared between two selected model groups, representing 1.5°- and 2.0°C-warmer conditions, respectively, in the “equilibrium world” assumption. When checking the time series of global mean temperatures, as well as East Asian mean temperature and WBGT, near-stable climate conditions, indeed, occurred during the last 30 yr of the twenty-first century (not shown). When using an increased number of model runs, selected by extending the range of each target temperature from +1.25° to +1.75°C and from +1.75° to +2.25°C, similar results were obtained (not shown), indicating that heat stress responses to different target temperatures in East Asia are generally insensitive to model samples.

It is noted that previous studies evaluated the effects of climate change at the 1.5° and 2.0°C target temperatures in the “transient world” through time slices centered around a specific level of warming analysis (Fischer and Knutti 2015; Knutti et al. 2016; Schleussner et al. 2016a,b). This method, however, is known to have a disadvantage in that it neglects the influence of time-lagged systems (Schleussner et al. 2016a). In this respect, we also calculated the difference in heat stress changes at the 1.5° and 2.0°C target temperatures under the transient world and compared the results with those from the equilibrium world. For the transient world estimation, we used the RCP8.5 scenarios of those selected models in the equilibrium world (see Table 2) and found 30-yr periods during which global mean temperature changes become +1.5° and +2.0°C relative to the preindustrial level (1861–1900). These 30-yr time slices were then analyzed to obtain the changes in heat stress intensity, frequency, and duration indices (see above) in the transient world at the target temperatures of 1.5° and 2.0°C, respectively.

3. Results

a. Future changes in heat stress

Before examining future changes, we have evaluated CMIP5 models in terms of WBGT climatology in comparison with the ERA-Interim and NCEP2 reanalyses. Bias, root mean squared error (RMSE), and the Taylor diagram analysis were used to evaluate model skills in terms of summer mean climatology of daily mean WBGT and its 95th percentile (Figs. S1 and S2, in online supplementary material, file JCLI-D-17-0449s1). Results show that the summer mean WBGT over East Asia is well simulated by CMIP5 models, with some underestimation over the ocean and overestimation in some parts of the land (Fig. S1). MME bias is −1.50°C, and RMSE is 2.07°C, with some differences across models. Note that these are not large when considering the uncertainty in reanalysis data by comparing ERA-Interim with the NCEP2 reanalysis (bias of −0.58°C and RMSE of 0.81°C). The Taylor diagram shows that the spatial correlation coefficients of all models are high (r > 0.8), and the standard deviation of spatial variability is also similar to the reanalysis (Fig. S1). Similar results were obtained for summer mean patterns of the 95th percentile of daily mean WBGT (Fig. S2).

Figure 1 displays time series in the area fraction of , , and over 1951–2099, which represent changes in percentage area affected by extreme heat stress over the land regions of East Asia. All scenario runs project that the fractional area of extreme heat stress starts to increase from the 2000s and will have exponential increases by the 2040s. This is consistent with the results of previous studies showing that the frequency of heat waves above a certain threshold increases nonlinearly as the temperature rises (Fischer and Knutti 2015; Knutti et al. 2016; Schleussner et al. 2016a). The future projections of the RCP2.6 and RCP 4.5 scenarios show that the area fraction for gradually stabilizes after the 2040s and 2060s, respectively. In the late twenty-first century, severe heat stress events exceeding the 50-yr return value are expected to occur in approximately 70% and 90% of land regions of East Asia under the two scenarios. For the RCP8.5 scenario, severe heat stress is expected to occur in most of the East Asian lands after the 2060s. In the case of extreme daily WBGT, the increasing patterns of the frequency () are very similar to those of (Fig. 1b), but weaker trends are seen in the duration () with a larger intermodel uncertainty (Fig. 1c).

Fig. 1.

Time series of area fractions of East Asia land for (a) summer mean WBGT, (b) summer extreme WBGT frequency, and (c) summer maximum duration, exceeding the 50-yr return values (see Table 1 for definition), obtained from CMIP5 multimodel simulations. Thick lines represent multimodel means, and thin lines depict individual runs.

Fig. 1.

Time series of area fractions of East Asia land for (a) summer mean WBGT, (b) summer extreme WBGT frequency, and (c) summer maximum duration, exceeding the 50-yr return values (see Table 1 for definition), obtained from CMIP5 multimodel simulations. Thick lines represent multimodel means, and thin lines depict individual runs.

The spatial distributions of future changes of and are compared for each decade to examine the spatial pattern of the heat stress expansion over East Asia (Fig. 2 for RCP4.5 scenario). In the 2040s, severe heat stresses exceeding the 50-yr return value occur every second year (more than 5 yr decade−1) on average, and in the 2090s, most of the land areas are expected to experience severe heat stresses almost every year. Both and show an extension of severe heat stress areas from the southwestern part of the land area to the north with time. This is consistent with previous studies, showing that the faster emergence of heat waves in the low latitudes is due to the relatively low year-to-year natural variability, compared to high latitudes (Diffenbaugh and Scherer 2011; Sillmann et al. 2013). The good agreement between and results indicates that mechanisms behind the changes in extreme daily heat stress are consistent with those deriving the mean changes (cf. Griffiths et al. 2005; Min et al. 2013). Changes in maximum duration of extreme heat stress exhibit similar spatial patterns but with weaker amplitudes (not shown).

Fig. 2.

Spatial distribution of exceedance frequency (years) over East Asia for (a) summer mean WBGT and (b) summer extreme WBGT frequency for each decade from 2010 to 2099, obtained from CMIP5 multimodel means of the RCP4.5 scenario simulations.

Fig. 2.

Spatial distribution of exceedance frequency (years) over East Asia for (a) summer mean WBGT and (b) summer extreme WBGT frequency for each decade from 2010 to 2099, obtained from CMIP5 multimodel means of the RCP4.5 scenario simulations.

b. Comparison with temperature changes

To understand the role of temperature increase in the heat stress expansion, the area fractions of temperature changes were compared with those of heat stress changes over the East Asian lands for 2070–99 (Fig. 3). Parameters , , and were calculated using the same method as in WBGT (see above). Results show that the area increases in WBGT were, overall, greater than those in temperature for all heat stress indices of mean intensity (Fig. 3a), extreme frequency (Fig. 3b), and maximum duration (Fig. 3c). In addition, the median area fraction indicates that the WBGT has a greater change than the temperature in most of the models, even though the uncertainty range is similar. Note that for the RCP8.5 scenario, the uncertainty ranges are much smaller than other scenarios, which is because the area fraction converges close to 100% in the latter half of the twenty-first century (Fig. 2). As a whole, models suggest that the area experiencing extreme heat stress will expand faster than would be suggested by the increased temperature alone, representing the larger impacts when considering humidity changes in addition to temperature changes (Fischer and Knutti 2013; Knutson and Ploshay 2016; Russo et al. 2017). When using an increased number of CMIP5 models that provide monthly data only, results remain unchanged, except for a slight increase in the uncertainty range of RCP2.6 (not shown), which indicates the robustness of our results.

Fig. 3.

The 2070–99 mean area fractions over East Asia for (a) summer mean, (b) summer extreme frequency, and (c) summer maximum duration of (left) WBGT (left) and (right) TAS, exceeding the corresponding 50-yr return value, calculated from CMIP5 RCP scenarios.

Fig. 3.

The 2070–99 mean area fractions over East Asia for (a) summer mean, (b) summer extreme frequency, and (c) summer maximum duration of (left) WBGT (left) and (right) TAS, exceeding the corresponding 50-yr return value, calculated from CMIP5 RCP scenarios.

To further understand the reason why the area of extreme WBGT expands faster than that of extreme temperature, we examined spatial patterns of the signal and the noise for WBGT in comparison with those for temperature (Fig. 4). Here, the signal was defined as the mean changes for the late twenty-first century (2070–99) estimated from the RCP8.5 scenario relative to the historical period (1951–2010). The noise, representing the spread of historical variability, is defined as the 50-yr return value minus the mean of the historical period. Results show that the future change in (5.20°C) is slightly larger than (5.03°C) because the vapor pressure is greatly increased at low latitudes and along the coastal area (see Fig. 7 below). In contrast, the land mean of noise (1.50°C) is smaller than that of noise (1.87°C). This implies that high area expands faster than high area due mainly to weaker noise in than in . The same results were found for and (not shown), confirming the important role of signal-to-noise ratio. WBGT is generally less variable than temperature over land because of the cancelation effect between temperature and vapor, leading to a smaller variability in WBGT than a given temperature variability. This is consistent with previous studies (Willett and Sherwood 2012; Knutson and Ploshay 2016) that found that the spread of changes in WBGT was substantially smaller than that in temperature from both observations and models.

Fig. 4.

Spatial distribution of (upper) signal and (lower) noise in summer mean (a),(c) WBGT and (b),(d) TAS over East Asia.

Fig. 4.

Spatial distribution of (upper) signal and (lower) noise in summer mean (a),(c) WBGT and (b),(d) TAS over East Asia.

c. Heat stress changes under different target temperatures

Figure 5 shows the area fractions of , , and under different global mean temperature changes averaged over the period 2070–99, which were obtained from individual model simulations under different RCP scenarios. When the global mean temperature increases by more than 3°C relative to the preindustrial level (1861–1900), such as all RCP8.5 scenario runs and many RCP4.5 runs, the area of mean and extreme heat stress would expand over most of the land areas of East Asia. Most of the RCP2.6 scenario runs and some of the RCP4.5 runs show weaker global warming. To evaluate the differences in heat stress expansion between different target temperatures, we selected six models (eight runs) representing a 1.5°C-warmer world, with global mean temperature changes ranging from +1.3° to +1.7°C, and eight models (nine runs) representing a 2.0°C-warmer world, with global mean temperature changes ranging from +1.8° to +2.2°C (see Table 2 for model runs). Comparing results between the two selected model groups indicates that the land areas of severe heat stress in East Asia would be reduced by approximately 20% (: 20.4%; : 19.1%; and : 19.3%) if the global warming is reduced from 2.0° to 1.5°C (Fig. 5).

Fig. 5.

Fraction of areas with (a) summer mean, (b) summer extreme frequency, and (c) summer extreme maximum duration of WBGT, exceeding 50-yr return values according to global mean temperature changes in the late twenty-first century (2070–99), relative to preindustrial levels. Diamonds represent multimodel means of each RCP scenario.

Fig. 5.

Fraction of areas with (a) summer mean, (b) summer extreme frequency, and (c) summer extreme maximum duration of WBGT, exceeding 50-yr return values according to global mean temperature changes in the late twenty-first century (2070–99), relative to preindustrial levels. Diamonds represent multimodel means of each RCP scenario.

The spatial patterns of changes are compared between 2.0° and 1.5°C target temperature increases (Fig. 6). As shown in Fig. 5, a 1.5°C world would have the area where exceeds 50-yr return values reduced by about 20.4%, compared to a 2.0°C world. The reduction appeared mainly in the inland areas of northern China (Figs. 6a,d). The difference in the East Asian mean between the 2.0° and 1.5°C targets was approximately 0.57°C in the equilibrium world (Fig. 6g) and approximately 0.65°C in the transient world (Fig. 6h). This means that when the target temperature is reduced from 2.0° to 1.5°C, land mean WBGT can be reduced by approximately 0.5°–0.6°C over East Asia, providing a benefit of warming mitigation by the weakened intensity of heat stresses on average. The spatial pattern of the 2.0°–1.5°C difference in from the equilibrium world (Fig. 6g) is characterized by a relatively large difference in the low-latitude ocean areas. This stronger change in low latitudes seems to be largely due to the stronger increase in humidity over the subtropical ocean at the 2.0°C target than at the 1.5°C target (Fig. 7). When comparing changes in the temperature term (0.567Ta) and the vapor pressure term (0.393e) of WBGT [see Eq. (3)], differences in the temperature term between the 2.0° and 1.5°C worlds are larger over land (Figs. 7a,b), while differences in the vapor pressure term are larger over the oceans (Figs. 7b,d).

Fig. 6.

Spatial distribution of changes (relative to 1951–2010 mean) for the (left) equilibrium world, (middle) the transient world, and (right) their differences over East Asia with (top) 1.5°C world, (middle) 2.0°C world, and (bottom) their differences obtained from selected models (Table 2). (a),(b),(d),(e) Hatching denotes areas with changes exceeding 50-yr return values. (c),(f),(g),(h) Dots represent statistically significant difference at 5% level based on a t test.

Fig. 6.

Spatial distribution of changes (relative to 1951–2010 mean) for the (left) equilibrium world, (middle) the transient world, and (right) their differences over East Asia with (top) 1.5°C world, (middle) 2.0°C world, and (bottom) their differences obtained from selected models (Table 2). (a),(b),(d),(e) Hatching denotes areas with changes exceeding 50-yr return values. (c),(f),(g),(h) Dots represent statistically significant difference at 5% level based on a t test.

Fig. 7.

Spatial distribution of changes in the temperature term (0.567Ta, W_TAS) and the vapor pressure term (0.393e, W_VP) of Eq. (3) in 2070–99 relative to 1951–2010 for the equilibrium world over East Asia for (top) 1.5°C world and (bottom) 2.0°C world, obtained from selected models.

Fig. 7.

Spatial distribution of changes in the temperature term (0.567Ta, W_TAS) and the vapor pressure term (0.393e, W_VP) of Eq. (3) in 2070–99 relative to 1951–2010 for the equilibrium world over East Asia for (top) 1.5°C world and (bottom) 2.0°C world, obtained from selected models.

On the other hand, in the transient world, a larger difference in WBGT between the two target temperatures is centered over the Korean Peninsula (Figs. 6b,e Accordingly, when the target temperature is reduced from 2.0° to 1.5°C, a relatively large reduction in heat stress intensity is expected over central-eastern China, Korea, and Japan (Fig. 6h). Figures 6c and 6f further compare the difference in heat stress changes between the equilibrium world and the transient world at the 1.5°- and 2.0°C-warmer worlds, respectively. It is found that the equilibrium world models predict a larger increase in over East Asia, compared to the transient world models, particularly at the 1.5°C target (Fig. 6c). The difference is particularly large along the coastal region, which resembles the WBGT trend pattern (Fig. 6h). This large difference is likely associated with slower warming of the local ocean in the transient world than in the equilibrium world, representing the influence of changes in time-lagged systems (Schleussner et al. 2016a). The relatively small difference in WBGT changes between the two worlds at the 2.0°C target supports this interpretation (the details of which need further investigation). The implication is that transient simulations need to be used with care when assessing mitigation effects on a regional scale like East Asia, where the ocean role is known to be very important, because they could underestimate the future changes in heat stress with slower warming in the regional ocean. Similar results are also found for and (not shown).

4. Discussion

There are currently only a few CMIP5 models available for comparing the two equilibrium worlds according to the target temperatures (Table 2), and a concern might be raised of whether the differences in heat stress response patterns between the two target temperatures, as well as between the equilibrium and transient worlds (shown in Fig. 6), may be in part caused by the differences in the selected models used in the analyses. To check this, we have conducted a sensitivity test to the use of the same subset of models that provide data for both equilibrium worlds (1.5° and 2.0°C) as well as transient worlds (model runs marked by asterisks in Table 2). Overall, results are found to be similar despite using a smaller number of model runs (not shown). However, some subregional differences are noticeable, indicating a stronger influence of model uncertainty on smaller-scale responses.

We used daily mean temperature and daily mean relative humidity to calculate daily mean WBGT, from which we calculated summer mean intensity, summer extreme frequency, and summer maximum duration of heat stress (as described in Table 1). However, some previous studies estimated daily maximum WBGT from daily maximum temperature and relative humidity (Fischer and Schär 2010; Russo et al. 2017), where they used daily minimum relative humidity rather than daily mean relative humidity, assuming that the relative humidity is very low at the time of daily maximum temperature (Fischer and Schär 2010). Here, we checked the robustness of our results to the different definition of daily WBGT. Using available models (same as in Table 2, but without HadGEM2-ES and NorESM1-M), we calculated the daily maximum WBGT from daily maximum temperature and daily minimum relative humidity and repeated our analysis of heat stress responses to 1.5° and 2.0°C target temperatures. Results are displayed in Fig. 8, which shows that overall, amplitudes and spatial patterns of heat stress changes obtained from using daily maximum WBGT are very similar to those based on daily mean WBGT. This represents the negligible influence of the definition of daily WBGT on the relative changes in heat stress at different target temperatures and their differences.

Fig. 8.

As in Fig. 6, but using daily maximum WBGT calculated from daily maximum temperature and daily minimum relative humidity.

Fig. 8.

As in Fig. 6, but using daily maximum WBGT calculated from daily maximum temperature and daily minimum relative humidity.

Considering the importance of assessing absolute magnitude of heat stress in the future (Im et al. 2017; Russo et al. 2017), here, we briefly examined the absolute values of summer mean and summer maximum WBGT in the 1.5°- and 2.0°C-warmer worlds (Fig. S3). For the summer mean, WBGT higher than 28°C, a threshold for high risk to health (Willett and Sherwood 2012), is found mainly over the low-latitude ocean areas and eastern China in ERA-Interim. The 28°C WBGT in ERA-Interim is equivalent to about 26°C in the CMIP5 historical simulations due to the model underestimation over East Asia (see also Fig. S1), which can be seen from Russo et al. (2017). Results from the 1.5° and 2°C equilibrium warming indicate that a large part of land areas will experience high heat stress intensity in the future, with 26°C lines shifted northward, covering Korea and Japan. A similar expansion of heat stress area is found for the summer maximum WBGT, and eastern China is expected to have WBGT higher than 35°C in the future, a threshold for extreme risk to health (Willett and Sherwood 2012), which could be higher when considering larger model bias in summer maximum WBGT than summer mean (Fig. S3). These changes in magnitude and peak of heat stress are, overall, in line with previous studies based on the different heat wave indices and associated thresholds (Mora et al. 2017; Im et al. 2017; Russo et al. 2017). However, given the large model bias and associated projection uncertainties, absolute changes in heat stress need to be further investigated in the future work.

In addition to changes in 30-yr statistics of heat stress intensity, frequency, and duration, changes in the interannual variability of heat stress can be important. For this, we compared the standard deviations of summer mean WBGT between the present period and future equilibrium worlds. Detrended data were used for the historical simulations. Results suggested no significant differences in the interannual variability of WBGT between historical and future simulations, as well as between 1.5° and 2.0°C equilibrium worlds (not shown), implying that interannual fluctuations of heat stress will not change much in the future.

5. Conclusions

This study provides a quantitative assessment of the future changes in summer heat stress over East Asia using CMIP5 multimodel simulations based on the RCP scenarios (2011–99). Changes in summer mean heat stress intensity () and daily extreme heat stresses frequency () and maximum duration () were analyzed relative to the historical period (1951–2010). Particularly, the space–time emergence of heat stress characteristics was examined at the 1.5° and 2.0°C target temperatures of the 2015 Paris Agreement. To estimate the extent of future heat stress expansion, we calculated the area fraction experiencing severe heat stress (exceeding the 50-yr return values of the historical period). The extreme heat stress area is shown to increase exponentially from the 2000s. In the 2040s, severe heat stress is projected to occur over most of the East Asian land areas for more than 5 yr decade−1, and in the 2090s, almost every year. The spatial distribution shows that severe heat stresses occur faster from the low latitudes, which is also seen in the pattern of extreme frequency , indicating identical mechanisms at work for mean and extreme changes in heat stress. In addition, it is shown that the area of extreme heat stress expands more rapidly than the area of extreme temperature because the WBGT has less variability than the temperature, supporting previous studies.

Heat stress extent over East Asia under the two target temperatures was examined using selected model simulations under two assumptions. One is under the equilibrium condition (based on RCP2.6 and RCP4.5 scenario simulations), and the other is based on a transient world (based on RCP8.5 scenario). In the equilibrium world, when the target temperature is reduced from 2.0° to 1.5°C, the area with severe heat stress over East Asia is reduced by approximately 20%. Relatively large reduction of heat stress by warming mitigation is found over the low-latitude oceanic areas. The benefit of warming mitigation through heat wave reduction is more pronounced in the transient world, which was attributed to the slow response in the oceanic region, suggesting that one needs to carefully consider the effect of land–sea contrast. Also, this indicates an important role of ocean surface warming patterns in the regional assessment of global warming mitigation.

Although our results are based on limited samples of available model simulations, our initial findings provide an important implication for assessing possible avoided risks in extreme events, such as heat stress, under different target temperatures. Physical mechanisms for deriving the heat stress strengthening and expansion need to be investigated in the future work, including the quantification of the thermodynamic and dynamic effects with careful considerations of the land–ocean contrast and their interactions.

Acknowledgments

We thank three anonymous reviewers for their constructive comments. This work was supported by the Korea Meteorological Administration Research and Development Program under Grant KMIPA 2015-2082. We acknowledge the World Climate Research Programme’s Working Group on Coupled Modelling, which is responsible for CMIP, and we thank the climate modeling groups (listed in Table 2 of this paper) for producing and making available their model output.

REFERENCES

REFERENCES
ACSM
,
1984
:
Prevention of thermal injuries during distance running: Position stand
.
Med. J. Aust.
,
141
,
876
879
.
Baek
,
H. J.
, and Coauthors
,
2013
:
Climate change in the 21st century simulated by HadGEM2-AO under representative concentration pathways
.
Asia-Pac. J. Atmos. Sci.
,
49
,
603
618
, https://doi.org/10.1007/s13143-013-0053-7.
Basu
,
R.
, and
J. M.
Samet
,
2002
:
Relation between elevated ambient temperature and mortality: A review of the epidemiologic evidence
.
Epidemiol. Rev.
,
24
,
190
202
, https://doi.org/10.1093/epirev/mxf007.
Christensen
,
J. H.
, and Coauthors
,
2007
:
Regional climate projections
. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 847–940.
Christensen
,
J. H.
, and Coauthors
,
2013
: Climate phenomena and their relevance for future regional climate change. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 1217–1308.
Collins
,
M.
, and Coauthors
,
2013
: Long-term climate change: Projections, commitments and irreversibility. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 1029–1136.
Coumou
,
D.
, and
A.
Robinson
,
2013
:
Historic and future increase in the global land area affected by monthly heat extremes
.
Environ. Res. Lett.
,
8
,
034018
, https://doi.org/10.1088/1748-9326/8/3/034018.
Diffenbaugh
,
N. S.
, and
M.
Scherer
,
2011
:
Observational and model evidence of global emergence of permanent, unprecedented heat in the 20th and 21st centuries
.
Climatic Change
,
107
,
615
624
, https://doi.org/10.1007/s10584-011-0112-y.
Dunne
,
J. P.
,
R. J.
Stouffer
, and
J. G.
John
,
2013
:
Reductions in labour capacity from heat stress under climate warming
.
Nat. Climate Change
,
3
,
563
566
, https://doi.org/10.1038/nclimate1827.
Fischer
,
E. M.
, and
C.
Schär
,
2010
:
Consistent geographical patterns of changes in high-impact European heatwaves
.
Nat. Geosci.
,
3
,
398
403
, https://doi.org/10.1038/ngeo866.
Fischer
,
E. M.
, and
R.
Knutti
,
2013
:
Robust projections of combined humidity and temperature extremes
.
Nat. Climate Change
,
3
,
126
130
, https://doi.org/10.1038/nclimate1682.
Fischer
,
E. M.
, and
R.
Knutti
,
2015
:
Anthropogenic contribution to global occurrence of heavy-precipitation and high-temperature extremes
.
Nat. Climate Change
,
5
,
560
564
, https://doi.org/10.1038/nclimate2617.
Griffiths
,
G. M.
, and Coauthors
,
2005
:
Change in mean temperature as a predictor of extreme temperature change in the Asia–Pacific region
.
Int. J. Climatol.
,
25
,
1301
1330
, https://doi.org/10.1002/joc.1194.
Im
,
E. S.
,
J. S.
Pal
, and
E. A. B.
Eltahir
,
2017
:
Deadly heat waves projected in the densely populated agricultural regions of South Asia
.
Sci. Adv.
,
3
,
e1603322
, https://doi.org/10.1126/sciadv.1603322.
ISO
,
1989
: Hot environments—Estimation of the heat stress on working man, based on the WBGT index. International Organization for Standardization 7243.
Ji
,
Z.
, and
S.
Kang
,
2015
:
Evaluation of extreme climate events using a regional climate model for China
.
Int. J. Climatol.
,
35
,
888
902
, https://doi.org/10.1002/joc.4024.
Kjellstrom
,
T.
,
R. S.
Kovats
,
S. J.
Lloyd
,
T.
Holt
, and
R. S. J.
Tol
,
2009
:
The direct impact of climate change on regional labor productivity
.
Arch. Environ. Occup. Health
,
64
,
217
227
, https://doi.org/10.1080/19338240903352776.
Knutson
,
T. R.
, and
J. J.
Ploshay
,
2016
:
Detection of anthropogenic influence on a summertime heat stress index
.
Climatic Change
,
138
,
25
39
, https://doi.org/10.1007/s10584-016-1708-z.
Knutti
,
R.
,
J.
Rogelj
,
J.
Sedláček
, and
E. M.
Fischer
,
2016
:
A scientific critique of the two-degree climate change target
.
Nat. Geosci.
,
9
,
13
18
, https://doi.org/10.1038/ngeo2595.
Lemke
,
B.
, and
T.
Kjellstrom
,
2012
:
Calculating workplace WBGT from meteorological data: A tool for climate change assessment
.
Ind. Health
,
50
,
267
278
, https://doi.org/10.2486/indhealth.MS1352.
Meehl
,
G. A.
, and
C.
Tebaldi
,
2004
:
More intense, more frequent, and longer lasting heat waves in the 21st century
.
Science
,
305
,
994
997
, https://doi.org/10.1126/science.1098704.
Min
,
S.-K.
,
W.
Cai
, and
P.
Whetton
,
2013
:
Influence of climate variability on seasonal extremes over Australia
.
J. Geophys. Res. Atmos.
,
118
,
643
654
, https://doi.org/10.1002/jgrd.50164.
Min
,
S.-K.
,
Y.-H.
Kim
,
M.-K.
Kim
, and
C.
Park
,
2014
:
Assessing human contribution to the summer 2013 Korean heat wave [in “Explaining Extreme Events of 2013 from a Climate Perspective”]
.
Bull. Amer. Meteor. Soc.
,
95
(9),
S48
S51
.
Min
,
S.-K.
, and Coauthors
,
2015
:
Changes in weather and climate extremes over Korea and possible causes: A review
.
Asia-Pac. J. Atmos. Sci.
,
51
,
103
121
, https://doi.org/10.1007/s13143-015-0066-5.
Mora
,
C.
, and Coauthors
,
2017
:
Global risk of deadly heat
.
Nat. Climate Change
,
7
,
501
506
, https://doi.org/10.1038/nclimate3322.
Perkins
,
S. E.
,
L. V.
Alexander
, and
J. R.
Nairn
,
2012
:
Increasing frequency, intensity and duration of observed global heatwaves and warm spells
.
Geophys. Res. Lett.
,
39
,
L20714
, https://doi.org/10.1029/2012GL053361.
Russo
,
S.
,
J.
Sillmann
, and
A.
Sterl
,
2017
:
Humid heat waves at different warming levels
.
Sci. Rep.
,
7
,
7477
, https://doi.org/10.1038/s41598-017-07536-7.
Schleussner
,
C.-F.
, and Coauthors
,
2016a
:
Differential climate impacts for policy-relevant limits to global warming: The case of 1.5 °C and 2 °C
.
Earth Syst. Dyn.
,
7
,
327
351
, https://doi.org/10.5194/esd-7-327-2016.
Schleussner
,
C.-F.
, and Coauthors
,
2016b
:
Science and policy characteristics of the Paris Agreement temperature goal
.
Nat. Climate Change
,
6
,
827
835
, https://doi.org/10.1038/nclimate3096.
Seneviratne
,
S. I.
, and Coauthors
,
2012
: Changes in climate extremes and their impacts on the natural physical environment. Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation, C. B. Field et al., Eds., Cambridge University Press, 109–230.
Seneviratne
,
S. I.
,
M. G.
Donat
,
A. J.
Pitman
,
R.
Knutti
, and
R. L.
Wilby
,
2016
:
Allowable CO2 emissions based on regional and impact-related climate targets
.
Nature
,
529
,
477
483
, https://doi.org/10.1038/nature16542.
Sherwood
,
S. C.
, and
M.
Huber
,
2010
:
An adaptability limit to climate change due to heat stress
.
Proc. Natl. Acad. Sci. USA
,
107
,
9552
9555
, https://doi.org/10.1073/pnas.0913352107.
Sillmann
,
J.
,
V. V.
Kharin
,
F. W.
Zwiers
,
X.
Zhang
, and
D.
Bronaugh
,
2013
:
Climate extreme indices in the CMIP5 multimodel ensemble: Part 2. Future climate projections
.
J. Geophys. Res. Atmos.
,
118
,
2473
2493
, https://doi.org/10.1002/jgrd.50188.
Steadman
,
R.
,
1979
:
The assessment of sultriness. Part 1: A temperature-humidity index based on human physiology and clothing science
.
J. Appl. Meteor.
,
18
,
861
873
, https://doi.org/10.1175/1520-0450(1979)018<0861:TAOSPI>2.0.CO;2.
Taylor
,
K. E.
,
R. J.
Stouffer
, and
G. A.
Meehl
,
2012
:
An overview of CMIP5 and the experiment design
.
Bull. Amer. Meteor. Soc.
,
93
,
485
498
, https://doi.org/10.1175/BAMS-D-11-00094.1.
Thomson
,
A. M.
, and Coauthors
,
2011
:
RCP4.5: A pathway for stabilization of radiative forcing by 2100
.
Climatic Change
,
109
,
77
94
, https://doi.org/10.1007/s10584-011-0151-4.
UNFCCC
,
2015
: Report of the Conference of the Parties on its twenty-first session, held in Paris from 30 November to 13 December 2015. Part two: Action taken by the Conference of the Parties at its twenty-first session. United Nations Rep. FCCC/CP/2015/10/Add.1, 36 pp., http://unfccc.int/resource/docs/2015/cop21/eng/10a01.pdf.
van Vuuren
,
D. P.
, and Coauthors
,
2011
:
RCP2.6: Exploring the possibility to keep global mean temperature increase below 2°C
.
Climatic Change
,
109
,
95
116
, https://doi.org/10.1007/s10584-011-0152-3.
Willett
,
K. M.
, and
S.
Sherwood
,
2012
:
Exceedance of heat index thresholds for 15 regions under a warming climate using the wet-bulb globe temperature
.
Int. J. Climatol.
,
32
,
161
177
, https://doi.org/10.1002/joc.2257.
Yaglou
,
C. P.
, and
D.
Minard
,
1957
:
Control of heat casualties at military training centers
.
Amer. Med. Assoc. Arch. Ind. Health
,
16
,
302
316
.
Yun
,
K. S.
,
K. Y.
Heo
,
J. E.
Chu
,
K. J.
Ha
,
E. J.
Lee
,
Y.
Choi
, and
A.
Kitoh
,
2012
:
Changes in climate classification and extreme climate indices from a high-resolution future projection in Korea
.
Asia-Pac. J. Atmos. Sci.
,
48
,
213
226
, https://doi.org/10.1007/s13143-012-0022-6.

Footnotes

Supplemental information related to this paper is available at the Journals Online website: https://doi.org/10.1175/JCLI-D-17-0449.s1.

© 2018 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Supplemental Material