Will climate change make Chinese people more comfortable? A scenario analysis based on the weather preference index

Our study area is China. According to the 'Geographic Atlas of China' compiled by Wang et al. (2009), we divided the country into 12 climatic zones based on the annual accumulated temperature (figure 1). These climatic zones consist of the cold temperate zone, middle temperate zone, warm temperate zone, north subtropical zone, middle subtropical zone, south subtropical zone, marginal tropical zone, plateau tropical zone, plateau subtropical zone, plateau temperate zone, plateau subfrigid zone and plateau frigid zone (Wang et al., 2009).

Zoom In Zoom Out Reset image size

The meteorological observation data were obtained from the China Meteorological Data Network (http://data.cma.cn/data/detail/dataCode). The data included the average daily maximum temperatures in January and July, the average daily relative humidity in July and precipitation data from 756 meteorological stations nationwide from 1971 to 2013. After obtaining the meteorological data from each station, we used the Kriging interpolation method to carry out the spatial interpolation of the station data and obtain historical meteorological data with a spatial resolution of 1 km. Specifically, we selected the spherical function model of ordinary Kriging and used the surrounding 16 meteorological stations as the search radius to carry out the interpolation. Finally, we calculated the mean values of these data using the county as the basic unit.

The climate simulation data included GCM data from 2025 to 2100 under the RCP2.6, RCP4.5 and RCP8.5 emission scenarios, which were derived from the World Climate Research Programme (WCRP) Coupled Model Intercomparison Project Phase Five (CMIP5) climate model data (https://esgf-node.llnl.gov/search/cmip5) published by the Program for Climate Model Diagnosis and Intercomparison (PCMDI). The GCM data used included 7 models, which are listed in table 1.

In terms of the climate model data, the RCP2.6 scenario indicates that greenhouse gas emissions will rise first, then fall, and finally reach a stable state in the next 100 years. In 2100, the radiative forcing will be less than 2.6 W m⁻² under this scenario. The RCP4.5 scenario indicates that the greenhouse gas emissions will rise first; then, the upward trend will gradually slow and finally stabilize in the next 100 years. In 2100, the radiative forcing will stabilize at 4.5 W m⁻² under this scenario. The RCP8.5 scenario indicates that greenhouse gas emissions will continue to rise over the next 100 years. In 2100, the radiative forcing will reach 8.5 W m⁻² under this scenario (Vuuren et al 2011).

The county-level population data were obtained from China's 2010 population census data (the census office of the state council in the national bureau of statistics of population and employment statistics, China, 2012).

Based on the methods in the study by Ai and Cheng (2015), we first resampled the data of the seven GCMs by using the nearest neighbor method to unify the resolution and then obtained the GCM simulation results with resolution of 1° by averaging the values of the simulation results of the seven models each year (figure 2). Specifically, we averaged the values of the five indicators in the simulation results of the seven models each year from 2025 to 2100, including the average daily maximum temperatures in January and July, the average daily relative humidity in July, annual precipitation and the number of days in which precipitation occurred annually.

Zoom In Zoom Out Reset image size

The WPI indicates the changes in the population in a region caused by weather conditions. According to Egan and Mullin (2016), the WPI is calculated based on five meteorological indicators, and the calculation process can be expressed as follows:

$$\begin{align} WP{I_{jt}} & = A \times JANMA{X_{jt}} + B \times JANMA{X_{jt}}^2 \nonumber\\ & \quad {{\text{ }} + C \times JULYH{I_{jt}} + D \times JULYH{I_{jt}}^2} \nonumber\\ & \quad { + E \times JULYR{H_{jt}} + F \times JULYR{H_{jt}}^2} \nonumber\\ & \quad { + G \times PRECIPI{N_{jt}} + H \times PRECIPI{N_{jt}}^2} \nonumber\\ & \quad { + I \times PRECIPDAY{S_{jt}} + J \times PRECIPDAY{S_{jt}}^2} \end{align} \tag{ 1 }$$

where WPI_jt represents the WPI of region j in year t. JANMAX_jt represents the average daily maximum temperature in January in this region in year t, JULYHI_jt represents the daily heat index in July in this region in year t (Stull and Ahrens 2000), JULYRH_jt represents the average daily relative humidity in July in this region in year t, PRECIPIN_jt represents the annual precipitation in this region in year t, and PRECIPDAYS_jt represents the number of days in which precipitation occurs annually in this region in year t. A-J are the coefficients before each meteorological indicator. According to Egan and Mullin (2016), these coefficients can be obtained through multiple regression analysis based on the historical time series of the population and climate data. However, China currently lacks such long-term population and climate data. Considering that people's perception of bioclimatic comfort is similar, we directly adopted the coefficients of various meteorological indicators obtained by Egan and Mullin (2016), which were based on the population and climate data in the United States from 1880 to 2000. The calculation formulas for the WPI after determining the coefficients can be expressed as:

$$\begin{align}\small WP{I_{jt}} & = 0.0488 \times JANMA{X_{jt}} + 0.0013 \times JANMA{X_{jt}}^2 \nonumber\\ & \quad - 0.0170 \times JULYH{I_{jt}} - 0.0008 \times JULYH{I_{jt}}^2 \nonumber\\ & \quad - 0.0385 \times JULYR{H_{jt}} - 0.0003 \times JULYR{H_{jt}}^2 \nonumber\\ & \quad - 0.0048 \times PRECIPI{N_{jt}} + 0.0002 \nonumber\\ & \quad \times PRECIPI{N_{jt}}^2 + 0.0065 \times PRECIPDAY{S_{jt}} \nonumber\\ & \quad - 0.0002 \times PRECIPDAY{S_{jt}}^2 \end{align} \tag{ 2 }$$

Using the above formula, we first calculated the WPI of each county in China from 1971 to 2013 based on meteorological observation data. Then, the WPI of each county in China from 2025 to 2100 was calculated under different climate change scenarios using GCM data. To ensure the comparability between historical and future WPI values, we followed the methods in Egan and Mullin (2016) and used the WPI from 1971 to 2005 based on climate model data and the WPI from the observation data in the same period for regression analysis. The regression analysis results were used to correct the WPI from 2025 to 2100. The calibration process can be expressed as follows:

$$\begin{equation}WP{I_{ch}} = 0.35 \times WP{I_{si}} + 1.17\end{equation} \tag{ 3 }$$

where WPI_ch represents the corrected WPI, and WPI_si represents the WPI before correction. Finally, we followed Egan and Mullin (2016) and calculated the population-weighted mean WPI for the regions of China and climatic zones from 1971 to 2013 and the population-weighted mean WPI from 2025 to 2100 under different climate scenarios. The calculation of the population-weighted mean WPI can be expressed as:

$$\begin{equation}WP{I_{kt}} = \frac{{\mathop \sum \nolimits_1^m WP{I_{jt}} \cdot {P_j}}}{{\mathop \sum \nolimits_1^m {P_j}}}\end{equation} \tag{ 4 }$$

where WPI_kt represents the population-weighted mean WPI of region k in year t, P_j represents the population of county j, and m is the total number of counties in region k. According to Egan and Mullin (2016), we used imperial units of measure to calculate the WPI and then transformed results into SI units to analyze the trend of WPI.

Trend analysis is a common method used to analyze climate change (Shi et al 2014). Therefore, based on the study of Shi et al. (2014), we utilized trend analysis to evaluate the changes in the WPI from 1971 to 2013 and the trend of the WPI from 2025 to 2100. The trend analysis can be expressed as:

$$\begin{equation}WP{I_{jt}} = {a_j} + {b_j}t\end{equation} \tag{ 5 }$$

where${\text{ }}WP{I_{jt}}$represents the WPI of region j in year t, a_j is the regression constant, and b_j is the trend value, which is obtained by the least squares method:

$$\begin{array}{l} {a_j} = \frac{1}{n}\sum\limits_{t = {t_i}}^{{t_i} + n - 1} W P{I_{jt}} - {b_j}\frac{1}{n}\sum\limits_{t = {t_i}}^{{t_i} + n - 1} t ,\\ {b_j} = \;\frac{{\sum\nolimits_{t = {t_i}}^{{t_i} + n - 1} W P{I_{jt}}t - \frac{1}{n}\left( {\sum\nolimits_{t = {t_i}}^{{t_i} + n - 1} W P{I_{jt}}} \right)\left( {\sum\nolimits_{t = {t_i}}^{{t_i} + n - 1} t } \right)}}{{\sum\nolimits_{t = {t_i}}^{{t_i} + n - 1} {{t^2}} - \frac{1}{n}{{\left( {\sum\nolimits_{t = {t_i}}^{{t_i} + n - 1} t } \right)}^2}}} \end{array} \tag{ 6 }$$

where t_i is the starting year and n is the length of time in years. When b_j is greater than 0, the WPI in region j increases over time. When b_j is less than 0, the WPI in region j decreases over time. The size of b_j reflects the rate of WPI change in region j.

China's weather conditions improved from 1971 to 2013 (figure 3). Specifically, China's WPI increased from 1.34 to 1.59 with a change rate of 0.03 per decade (0.03/10 a) (table 2). There were 1567 counties experiencing an increase in WPI, which accounted for 68.40% of the total number of counties in China. These counties had a total area of 7.17 million km² (74.50% of the land area in China) and a total population of 0.94 billion (70.36% of the total population in China). The increase in maximum daily January temperature (0.04 °C/10 a) and the decrease in July relative humidity (−1.31%/10 a) were the main factors related to the improvement in China's weather conditions (table 2). Among the 12 climatic zones, the plateau frigid zone had the most obvious improvement in weather conditions. The WPI of the plateau frigid zone was 0.19 in 1971 and 0.53 in 2013, with a change rate of 0.24/10 a. In the plateau frigid zone, the entire region experienced improved weather conditions (figure 3(d)).

Zoom In Zoom Out Reset image size

In 2013, the WPI was high in the southwest and low in the northeast. Among the climatic zones, the marginal tropical zone had the highest WPI, with the WPI reaching 6.15, which was 2.87 higher than the national average. The cold temperate zone had the lowest WPI, with the WPI values of −1.51, which was 1.95 lower than the national average (figure 3(c)).

(a-the change in WPI in China, b-the spatial pattern of WPI in 1971, c-the spatial pattern of WPI in 2013, d-the spatial pattern of change rate of WPI from 1971 to 2013)

Under the RCP2.6 scenario, the weather conditions in China will slightly deteriorate from 2025 to 2100 (figure 4). The WPI in China will decrease from 1.38 to 1.33 with a change rate of −0.01/10 a. Specifically, 1522 counties (66.43% of the total number of counties in China), which cover 6.53 million km² (67.85% of the total area of China), are projected to experience deteriorated weather conditions. In these counties, 0.85 billion people will experience deteriorated weather conditions, which account for 64.15% of the total population in China. The increase in maximum daily July heat index (0.14 °C/10 a) will be the main cause of the deteriorated weather conditions (table 3). Among the 12 climatic zones, 7 climatic zones will have more than 50% of the counties experiencing deteriorated weather conditions, accounting for 58.33% of the total number of climatic zones (figure 5(a)). Specifically, of the climatic zones, the warm temperate zone will have the greatest amount of people that will experience deteriorated weather conditions. In this climatic zone, 0.41 billion people (97.48% of the regional population) will experience deteriorated weather conditions, and the middle subtropical zone (0.17 billion people, 61.89% of the regional population), the north subtropical zone (0.16 billion people, 58.35% of the regional population), and the middle temperate zone (0.10 billion people, 66.29% of the regional population) have the second, third, and fourth greatest amount of people, respectively, that will experience deteriorated weather conditions.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

(a-comparison of the changes in WPI in the past and future; b-change in the WPI in the future)

(a-RCP2.6; b-RCP4.5; c-RCP8.2)

Under the RCP4.5 scenario, the weather conditions in China will also slightly deteriorate from 2025 to 2100 (figure 4). The WPI in China will decrease from 1.39 to 1.28 with a change rate of −0.02/10 a. There will be 0.71 billion people experiencing deteriorated weather conditions, which accounts for 53.28% of the total population in China. The increase in the maximum daily July heat index (0.53 °C/10 a) will be the main cause of the deteriorated weather conditions (table 3). Among the 12 climatic zones, the warm temperate zone will have the greatest amount of people that experience deteriorated weather conditions (figure 5(b)). Specifically, 0.35 billion people or 84.58% of the total population in the warm temperate will experience deteriorated weather conditions, and the north subtropical zone (0.16 billion people, 58.49% of the total population) and the middle temperate zone (0.12 billion people, 79.01% of the total population) will have the second and third greatest amount of people, respectively, that experience deteriorated weather conditions.

Under the RCP8.5 scenario, the weather conditions in China will clearly deteriorate from 2025 to 2100 (figure 4). The WPI in China will decrease from 1.37 to 0.45 with a change rate of−0.19/10 a. A total of 1.22 billion people or 91.58% of the total population in China will experience deteriorated weather conditions. The increase in the maximum daily July heat index (1.72 °C/10 a) will be the main cause of the deteriorated weather conditions (table 3). Among the 12 climatic zones, in 9 climatic zones, more than half of their populations will experience deteriorated weather conditions, accounting for 75.00% of the total number of climatic zones (figure 5(c)). Specifically, the warm temperate zone will have the greatest amount of people (0.39 billion people, 94.78% of the total population) experiencing deteriorated weather conditions, and the north subtropical zone (0.25 billion people, 94.55% of the total population) and the middle subtropical zone (0.23 billion people, 84.85% of the total population) will have the second and third greatest amount of people, respectively, experiencing deteriorated weather conditions.

The comfort of the human body (CHB) index is a widely used indicator to assess the CCI on weather conditions in China (Cao et al 2012, Guo et al 2015). The CHB index consists of three indicators: temperature, humidity and wind speed (Oliver 1973). Among them, the temperature is the main indicator of bioclimatic comfort, with humidity and wind speed acting as auxiliary indicators (Oliver 1973). Please refer to supporting information for the formula for calculating CHB index. Within a certain range, a higher CHB index represents a more comfortable climate (Guo et al 2015). To verify the reliability of the WPI, we used the CHB index to quantify bioclimatic comfort in China from 1971 to 2013. Then, we conducted a correlation analysis between the CHB index and WPI.

We found that the WPI can accurately reflect the CCI on weather conditions in China. The correlation coefficient between the WPI and the CHB index from 1971 to 2013 reached 0.91, passing the 0.001 level of significance (figure 6). The correlation coefficients between the WPI and CHB index in all climatic zones exceeded 0.33, passing the 0.05 level of significance (figure S1(https://stacks.iop.org/ERL/15/084028/mmedia)). Among the 12 climatic zones, the correlation coefficient between the WPI and CHB index was highest in the plateau frigid zone, reaching 0.86 and passing the 0.001 level of significance. Our results showed that the WPI exhibits a high consistency with the CHB index.

Zoom In Zoom Out Reset image size

Our results were consistent with the findings from existing studies. For example, Li et al. (2016) used a method based on the statistical bioclimatic comfort period and found that the national bioclimatic comfort showed an upward trend from 1961 to 2010, and in comparison to the first 25 years, the last 25 years had a total increase of 3.6 d in the bioclimatic comfort period.

The most important reason for the deterioration of China's future weather conditions is the rise in the July heat index. Therefore, the deterioration of weather conditions in the future is mainly manifested in the hotter summer. The continuous high temperature in summer directly affects human health, increases the incidence of cardiovascular, respiratory and digestive diseases, resulting in a large number of deaths (Stott et al 2004, Wang et al 2019). Due to the dense population, buildings and infrastructure, cities are more vulnerable to extreme weather events such as heat waves (Guerreiro et al 2018). Therefore, the CCI on weather conditions in urban areas has been an important part of climate change and urban sustainability research (Guerreiro et al 2018, 2018, Liu et al Zhang et al 2019a, 2019b, 2020). In 2012, more than half of the population in China lived in cities (Bai et al 2014). In 2014, the Chinese government began to implement the 'National Plan on New Urbanization', which promised people-oriented, urban-rural integrated, economic-intensive and harmonious development. This plan is expected to increase China's urbanization level to 60.0% by 2020, with an average annual increase of 1% (Bai et al 2014). In the context of future climate change, in comparison to other types of changes, changes in weather conditions in China's urban areas will have a more important impact on regional sustainable development (Cao et al 2018).

We found that nearly half of China's urban population will live in regions with deteriorated weather conditions by 2100. The total area of the regions with deteriorated weather conditions under the RCP2.6, RCP4.5 and RCP8.5 scenarios will be 2.34 million km² in China, accounting for 24.31% of the total area in China. These regions with deteriorated weather conditions will be mainly distributed in the middle temperate zone, the warm temperate zone, and the north subtropical zone (figure 7(a)). In particular, seven megacities (i.e. Beijing, Tianjin, Wuhan, Chengdu, Nanjing, Xi'an, and Shenyang) will be located in regions with deteriorated weather conditions. During the same period, 0.30 billion urban residents will live in regions with deteriorated weather conditions, accounting for 43.19% of the total urban population in China (figure 7(b)).

Zoom In Zoom Out Reset image size

Our findings were consistent with existing findings. For example, Zhou and Liu (2018) found that the frequency of extreme weather events in northern cities (e.g. Beijing, Tianjin, and Shenyang) will increase from 2021 to 2100. Li et al. (2018b) showed that in the future, Beijing, Shenyang and other northern cities would have a higher number of heat-related deaths. The continuous high temperature in summer in urban areas has already affected human health (Sun et al 2014, Wang et al 2019). Heat waves had killed nearly 4.5 million people in 31 major cities including Beijing, Tianjin and Shenyang from 2007 to 2013 (Yang et al 2019). The high temperature heat wave in Beijing in July 2010 led to 558 excess deaths. Among them, the excess deaths of cardiovascular and respiratory disease increased by 28% and 40%, respectively (Luan et al 2015). In the summer of 2010, the number of high-temperature deaths in Xi'an increased by 54% over the same period of the previous year (Xie et al 2015). To make matters worse, the heat-related mortality caused by high temperature will increase significantly in the future (IPCC 2018).

Therefore, we need to pay attention to the CCI on weather conditions in urban areas, especially in cities located in regions with deteriorated weather conditions in Beijing, Tianjin and Shenyang. Additionally, we should implement efforts to improve the comprehensive prevention of heat wave risks during summer and extremely cold temperature risks in winter, as well as extreme drought and rain events, by integrating climate change adaptation into urban economic construction and social development planning. Specifically, we need to increase urban green space, e.g. parks and street trees (Hickman 2013, Wolch et al 2014, Georgescu et al 2014). Moreover, building shading design, green buildings design and block cooling design can also be adopted (Cao et al 2018, Chen and Tang 2019). These efforts could help improve public infrastructure and living conditions to improve resilience to climate change ( Zhai et al 2019).

(a-distribution of regions with deteriorated weather conditions, b-percentage of the area and urban population in region with deteriorated weather conditions in China)

Note: Only the area of the region with deteriorated weather conditions under all three RCPs are shown. The population data are from the History Database of the Global Environment (HYDE) for 2100 (Goldewijk et al 2010).

We found that as the emissions concentrations increase, the area of regions with severely deteriorated weather conditions (i.e. the WPI change rate is less than − 0.1/10 a) will gradually increase in the future. With the increase in emissions concentrations from RCP2.6 to RCP8.5, the area of regions with severely deteriorated weather conditions in China will increase from 0 under RCP2.6 to 0.06 million km² under RCP4.5 and to 3.27 million km² under RCP8.5 (figure 8(a), b). The number of people living in the regions with severely deteriorated weather conditions will increase from 0.03 billion under RCP4.5 to 0.87 billion under RCP8.5, representing a 29fold increase (figure 8(b)).

Zoom In Zoom Out Reset image size

Furthermore, we found that the weather conditions in the north subtropical zone and the warm temperate zone are more sensitive to changes in emissions concentrations. With the increase in emissions concentrations from RCP4.5 to RCP8.5, the area of the regions with severely deteriorated weather conditions will increase almost 100 fold, from 0.05 million km² to 0.49 million km² in the north subtropical zone (figure 8(c)). The number of people living in regions with severely deteriorated weather conditions will increase almost 10 fold, from 0.02 billion to 0.22 billion (figure 8(c)). In the warm temperate zone, the area of the regions with severely deteriorated weather conditions will increase from 0.01 million km² to 0.76 million km², representing 57 fold increase (figure 8(d)). The number of people living in the regions with severely deteriorated weather conditions will increase from 0.01 billion to 0.36 billion, representing a 34 fold increase (figure 8(d)).

If greenhouse gas emissions are not controlled, the regions with severely deteriorated weather conditions will further expand, and more residents will live in regions with severely deteriorated weather conditions. Therefore, we need to emphasize the great importance of mitigating regional climate change, especially in areas sensitive to emissions concentration changes such as the north subtropical zone and the warm temperate zone. In addition, mitigation is needed to optimize the regional industrial structure, reduce the proportion of fossil fuels used in the energy sector, and promote the utilization of clean energy (Haines et al 2007). However, regional carbon sequestration capacity can be enhanced to reduce the concentration of greenhouse gases and mitigate climate change through ecological restoration and afforestation (Harris et al 2006).

Our study used WPI for the first time to assess the CCI on weather conditions in China. The WPI was developed on the basis of the assumption that weather conditions have significant effects on both life quality and land productivity, which in turn affect the change rate of the local population (Rappaport 2007, Egan and Mullin 2016). Thus, the index can both measure the CCI on weather conditions and reflect people's perceptions of weather conditions. We found that the weather conditions will deteriorate over the next few decades and the deteriorated weather conditions will influence more than half of China's total population. Moreover, such situation will get worse if greenhouse gas emissions are not controlled. Our findings provide new insights into the CCI on weather conditions and references for policy makers to mitigate and adapt to climate change in the pursuit of sustainable development in China.

The climate model data used in this study are the GCM data with coarse resolution and considerable uncertainty. Therefore, the simulated results do not necessarily reflect the exact future weather conditions, and they are only projected weather conditions under the different scenarios. However, this scenario simulation provided a possible projection of the future CCI on weather conditions in China. In addition, due to the limitation of data availability, we used the formula proposed by Egan and Mullin (2016) to calculate the WPI, which may cause errors if used directly in China.

In a subsequent study, we will attempt to localize the WPI calculation formula and use the high-resolution Weather Research and Forecasting (WRF) model simulation data to assess the future CCI on weather conditions in China (Skamarock et al 2008).