Trends in extreme high temperature at different altitudes of Southwest China during 1961–2014

XUE Yuting, CHEN Qunling, ZHANG Jiyu nd HUANG Ping

aCollege of Atmospheric Sciences, Chengdu University of Information Technology, Chengdu, China; bCenter for Monsoon System Research,Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China; cEarth System Modeling Center and Climate Dynamics Research Center, Nanjing University of Information Science and Technology, Nanjing, China; dState Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics, Chinese Academy of Sciences, Beijing, China

ABSTRACT This study analyzed the trends in extreme high temperature in Southwest China based on the observed daily maximum temperature and average temperature data from 410 Chinese stations recently released by the China Meteorological Administration. The authors found that the trends in extreme high temperature at different altitudes of Southwest China exhibit staged variations during a recent 50-year period (1961—2014). The trends in mean temperature and maximum temperature also exhibit phase variation. All temperature-related variables increase gently during the period 1975—94, whereas they increase dramatically during the recent period of 1995—2014,with a rate that is approximately two to ten times more than that during 1975—94. In addition, the trends in mean temperature, maximum temperature, and the frequency of extreme high temperature in the low altitudes transit from negative to positive in the two periods, while they increase dramatically in the mid- and high-altitude areas during 1995—2014, the well-known global warming hiatus period. In particular, the maximum temperature increases much faster than that of average temperature. This result implies that the regional temperature trend could be apparently different from the global mean temperature change.

KEYWORDS Southwest China; extreme high temperature; trend change; altitude

1. Introduction

In recent years, global warming induced by greenhouse gases has caused extreme weather events to occur more frequently, more intensively, and over a wider range worldwide (e.g. Easterling et al. 1997; Frich et al. 2002;Zhai and Pan 2003; Gong, Pan, and Wang 2004; IPCC 2007;Zhang et al. 2011; IPCC 2013; Song et al. 2014; Shen et al.2017; Cheng and Zhu 2018). Under global warming,Southwest China has suffered from extreme heat events for a long time (Liu et al. 2006; You et al. 2008a; Li et al.2012a; Ma et al. 2013; Peng 2014; Li 2015; Qin et al. 2015;Jiang et al. 2016; Wang 2018). For example, the extreme high temperature event in Sichuan Province in summer 2006 led to more than half of the local meteorological stations reaching a new high-temperature record, resulting in direct economic losses of 10 billion RMB (Li 2015).Moreover, the climate and terrain in Southwest China are quite diverse, including the Yunnan-Guizhou Plateau, the Hengduan Mountains, the Sichuan Basin, and other complex topographies (Giorgi et al. 1997; You et al. 2008a; You et al. 2008b; Li et al. 2012a; Hu, Huang, and Qu 2017).Thus, understanding the spatial and temporal variability of extreme weather events in Southwest China under global warming is crucial for adapting to global warming in this region.

Numerous studies have shown that there have been apparent regional temperature changes under global warming (Katz and Brown 1992; Easterling et al. 1997;Frich et al. 2002; Zhai and Pan 2003; Gong, Pan, and Wang 2004; Qian and Lin 2004; Alexander et al. 2006;Zhang et al. 2011; Zhou and Ren 2011; Li et al. 2012b;Wang, Ren, and Zhang 2014; Shen et al. 2017). Frich et al.(2002) found that extreme temperature over land globally shows an increasing trend during the late 20th century based on 10 extreme temperature indices calculated using daily maximum and minimum temperature.Alexander et al. (2006) found that every season during the 20th century has a significant warming phenomenon. Song et al. (2014) found that extreme temperature events probably associated with heat events occur mainly in Asia and Europe. The whole climate change trend of China is basically consistent with that of the world, but also shows obvious regional distinctions that are mainly caused by altitude (Sun, Wang, and Yuan 2008; You et al. 2008b; Li et al. 2012a, 2012b).

In recent decades, the number of summer high temperature extreme days has increased across China, with a strikingly abrupt increase in the mid-1990s that lagged behind the global warming trend (Wei and Chen 2011;Shen et al. 2017; Wu et al. 2017; Shi et al. 2018). Hu,Huang, and Qu (2017) based on absolute and percentile indices found that an increase in heat days and heatwave frequency took place in most areas of China during 1960—2013. Furthermore, this warming trend presents an obvious turning point, with faster warming in recent decades (Zhang, Gao, and Cui 2008). The causes of temporal and spatial changes in extreme climate in China are closely related to decadal changes in atmospheric circulation. Several studies have affirmed that extreme climate anomalies have major and direct influences mostly from anomalies of the western Pacific subtropical high (Sun, Wang, and Yuan 2008; Peng 2014). The weakening of the climatological meridional wind during recent decades and the rise in Eurasian temperatures have also contributed to the changes in extreme temperature in China (Li et al. 2011; Yang et al. 2012).

When comparing the climate trends of nine regions in China, it is found that the temperature change in Southwest China is different from that in most parts of the country (Sun et al. 2008; Wu et al. 2017). Previous studies on the spatial and temporal climate-change trend in Southwest China have documented the existence of obvious inconsistencies in the spatial distributions of meteorological variables, with apparent interannual and interdecadal variations (Ma et al. 2013;Li 2015). There was an increase in the temperature trend until the mid-1980s and a continuing warming with greater magnitude during a recent 50-year period (Li et al. 2012b). Furthermore, the warming in highaltitude areas has been found to have occurred earlier and faster than that in low-altitude areas of Southwest China (You et al. 2008b; You et al. 2010; Li et al. 2012a;Song et al. 2014; Hu, Huang, and Qu 2017; You et al.2019), as reflected in the results of other studies around the world (Giorgi et al. 1997; Rangwala and Miller 2012;Revadekar et al. 2013).

Although much research has been conducted on temperature changes in Southwest China, the duration and sampling of the datasets used in previous studies is a limitation. In the present study, we used the data of 2400 high-resolution meteorological stations, newly released by the China Meteorological Administration(CMA), with 410 stations in Southwest China, to analyze the trend in the spatial variation of extreme high temperature in the region. The rest of the paper is organized as follows: section 2 describes the data and methods used in the study; results are presented in section 3;and section 4 summarizes our findings.

2. Data and methods

2.1 Data

This study used the newly released daily temperature datasets containing 2400 stations in China by the National Climate Center of the CMA. In total, there are 668 stations in the research region, i.e. Southwest China,including the provinces of Sichuan, Guizhou and Yunnan, and Chongqing Municipality, as shown in Figure 1. The stations with discontinuous data in the summer (June-July-August, JJA) during the period of 1961 to 2014 were excluded, and so ultimately the datasets from 410 stations were chosen in the present study.

2.2 Methods

There have been various previous definitions of extreme high temperature, such as those based on the standard fixed-threshold method using 35°C, the percentile index method, and so on (Wei and Chen 2011; Hu, Huang, and Qu 2017). Considering the range of altitudes in Southwest China (Figure 1), we adopted the percentile index method to define the frequency of extreme high temperature events in the region, as in Li et al. (2017) and Wei and Chen (2011). In this study, the daily highest temperatures of the same day during JJA from 1961 to 2014 were arranged in ascending order to obtain the data of the 90th percentile. Then, the 92 highest temperatures were selected as the thresholds of the extreme high temperature events. When the daily maximum temperature exceeded the threshold of that day, this day was chosen as an extreme high temperature event.

Figure 1. Location of Southwest China, along with the altitude (units: m) and the regional division applied in this study.

Under global warming, the temperature in Southwest China has shown a general increasing trend during the last five decades. However, due to the influence of altitude and other factors, the temperature trends in different areas are distinct (Ge et al. 2011; Hu, Huang,and Qu 2017), with faster increases over the high altitudes (Li 2015). Thus, we divided Southwest China into three regions according to altitude, as shown in Figure 1.Region A is the low-altitude region (28°—35°N, 103°—111°E),including the Sichuan Basin and the Chongqing Mountains; region B is the middle-altitude region (20°—28°N, 96°—111°E), including the Yunnan-Guizhou Plateau and southern Sichuan Province; and region C is the highaltitude region (28°—35°N, 96°—103°E), including the western Sichuan Plateau and northern Yunnan Plateau (Ma et al. 2013).

Apart from the regional difference, the temperature variation can also differ over different time periods (Ge et al. 2011). Therefore, we selected two periods (1975—94 and 1995—2014) from 1961 to 2014. In this study, the least-squares regression method was used to analyze the long-term trend, which is a widely used method to reflect long-term changes. Here, the situation of 92 days in a summer was used to represent the climate state in the year (Ge et al. 2011; Qin et al. 2015). The sample sizes of the two periods (1975—94 and 1995—2014) were both 20, and the r-test was used to assess the significance of the trends in extreme high temperature and associated large-scale circulation.

3. Results

3.1 Trends in summer extreme high temperature at different altitudes

The trends in summer temperature over the three regions of different altitudes during the two time periods are obviously different (Figure 2). For the whole of Southwest China, the warming rate in 1961—2014 is 0.09°C/10 yr, with 0.01°C/10 yr in 1975—94 and 0.27°C/10 yr in 1995—2014 (Figure 2(a)). The trend is much larger in the recent two decades (1995—2014) than in other periods. In the low-altitude region (region A), the warming rate in 1961—2014 is 0.03°C/10 yr, with -0.10°C/10 yr in 1975—94 and a much larger trend of 0.27°C/10 yr in 1995—2014 (Figure 2(b)). In the mid- and high-altitude regions, the warming rates in 1961—2014 are 0.12°C/10 yr and 0.15°C/10 yr, with 0.07°C/10 yr and 0.11°C/10 yr in 1975—94, and 0.33°C/10 yr and 0.4°C/10 yr in 1995—2014, respectively (Figure 2(c,d)). The warming rates in 1995—2014 in the mid- and high-altitude regions are both greater than that of the whole period. In short,regardless of the altitude, we find that the warming rate of the summer temperature in Southwest China accelerates significantly during the recent two decades compared with other periods, especially in the high-altitude region.

Figure 2. Variation in summer average temperature in Southwest China in different time periods from 1961 to 2014: (a) whole of Southwest China; (b) low-altitude region; (c) mid-altitude region; (d) high-altitude region. The black solid line denotes the trend in 1961—2014; the black dotted line denotes the trend in 1975—1994; and the black long dotted line denotes the trend in 1995—2014. The linear regression coefficients denoting the warming rate are shown in the top-left corner (units: °C yr-1).

Previous studies have shown that changes in maximum temperature have a greater impact on extreme temperature than average temperature (Katz and Brown 1992;Easterling et al. 1997; Gong, Pan, and Wang 2004; Fang et al. 2016; Sein, Chidthaisong, and Oo 2018). Thus, the variation in maximum temperature in Southwest China was analyzed. In the whole of Southwest China and the low-altitude regions, the warming rate in 1961—2014 is about 0.10°C/10 yr, with a decreasing trend in 1975—94 and a dramatic increasing trend in 1995—2014 (Figure 3(a,b)). Meanwhile, in the mid- and high-altitude regions, the warming rate in 1961—2014 is slightly larger than that in Southwest China and the low-altitude region, with a weak increasing trend in 1975—94 and significant increasing trend in 1995—2014 (Figure 3(c,d)). In general, the warming rate of the summer maximum temperature in 1995—2014 is about 10 times that in 1975—94, and the faster increase is more apparent in the mid- and high-altitude regions.

Previous studies have found that the temperature change is closely related to the change in the frequency of extreme high temperature (Zhang, Gao, and Cui 2008;Ma et al. 2013). Figure 4 shows that the trend in extreme high temperature frequency in 1975—94 is relatively gentle at all altitudes, with a growth rate of less than once per 10 years, which is largely related to the trends in average and maximum temperatures. However, in 1995—2014, the extreme high temperature frequency trends at all altitudes increase significantly, resembling those of average and maximum temperature. The rate in the mid-altitude region is 5.2 times per 10 years (Figure 4(c)), whilst that in the highaltitude region is 3.4 times per 10 years and that in the lowaltitude region is 2.8 times per 10 years. The trend in summer extreme high temperature frequency increases across the whole of Southwest China, with a strikingly abrupt increase in 1995—2014.

Figure 3. Variation in summer maximum temperature (Tmax) in Southwest China in different time periods from 1961 to 2014: (a) whole of Southwest China; (b) low-altitude region; (c) mid-altitude region; (d) high-altitude region. The black solid line denotes the trend in 1961—2014; the black dotted line denotes the trend in 1975—1994; and the black long dotted line denotes the trend in 1995—2014. The linear regression coefficients denoting the warming rate are shown in the top-left corner (units: °C yr-1).

3.2 Spatial distributions of the trend changes in summer extreme high temperature

The spatial distribution of the linear trend in average temperature is further investigated. Figure 5(a) shows that the large trend in average temperature during 1975—94 is situated mainly over the mid- and high-altitude regions,whilst the low-altitude area shows a slight decreasing trend. Compared with the trend in 1975—94, the trend of average temperature in 1995—2014 increases signifciantly,with maxima over the mid- and high-altitude areas, where the maximum warming rate reaches 2.8°C/10 yr. The trends in the low-altitude areas turn from negative to positive(Figure 5(b)). Most of the average temperature trend shown in Figure 5 passes the r-test (i.e. is statistically signifciant) with a confdience level of 90%.

Figure 5(c,d) shows the spatial distribution of the summer maximum temperature trends during the two periods. In 1975—94, the highest temperature shows a weak increasing trend in the mid- and high-altitude regions, while the major part of the low-altitude region shows a decreasing trend. The greatest decreasing trend is 0.7°C/10 yr in the low-altitude region, and the increasing trend in the mid-altitude areas reaches 0.9°C/10 yr(Figure 5(c)). The highest temperature trend in 1995—2014 is more remarkable than that in 1975—94(Figure 5(d)). The faster increasing trend during the recent decades is more obvious in the mid- and highaltitude regions, where the maximum warming rate reaches 4°C/10 yr. This result is consistent with the result from Figure 3. The spatial distributions of the average temperature and the maximum temperature trends show significant acceleration during 1995—2014. Most of the maximum temperature trend shown in Figure 5(c,d) passes the r-test (i.e. is statistically significant) with a confidence level of 90%. The trend in maximum temperature in Southwest China is obviously faster than the trend in average temperature.

4. Summary and discussion

Figure 4. Variation in summer extreme temperature frequency (Frequency; units:d) in Southwest China in different time periods from 1961 to 2014: (a) whole of Southwest China; (b) low-altitude region; (c) mid-altitude region; (d) high-altitude region. The black solid line denotes the trend in 1961—2014; the black dotted line denotes the trend in 1975—1994; and the black long dotted line denotes the trend in 1995—2014. The linear regression coefficients denoting the warming rate are shown in the top-left corner (units: °C yr-1).

This study used the percentile method to investigate the spatial and temporal variations in summer extreme temperature over Southwest China during the period 1961—2014, based on high-resolution daily average and maximum temperature data from 410 stations over Southwest China. From 1961 to 2014, the summer average and maximum temperature in Southwest China shows an obvious increasing trend, but with remarkable differences in two periods. In recent decades, 1995—2014,the average and maximum temperatures both show an accelerated increase relative to that in 1975—94. The trend in average temperature in 1995—2014 is about 2—4 times that in 1975—1994, and the maximum temperature shows a higher rate of acceleration of about 4 to 11 times in 1995—2014 relative to 1975—94.

In 1961—2014, the trend in extreme high temperature frequency in Southwest China also shows an apparent phase variation. In 1975—94, the trend in extreme high temperature frequency is as gentle as the trend in average temperature, whereas in 1995—2014 the trend in extreme high temperature frequency increases at a rate of about 3 times every 10 years, which is about 7 to 13 times that in 1975—94. Analysis of the spatial distribution showed that the accelerated increase mainly happens in the mid- and high-altitude areas. Although the spatial distribution of the warming rates of maximum temperature is highly similar to that of average temperature, it warms faster than the average temperature. The trend in maximum temperature in the low-altitude areas transits from negative in 1975—94 to positive in 1995—2014, whereas the rest of the region shows a remarkable acceleration in 1995—2014, especially in the mid-altitude region of Yunnan at a rate of about 1.5°C/10 yr.

The faster warming increase in the high-altitude regions is consistent with the effect of elevation revealed in previous studies (Giorgi et al. 1997; Liu et al. 2006; Li et al. 2012a; Rangwala and Miller 2012; You et al. 2019).You et al. (2008b) showed that the elevation dependency of the warming rate is not found in research on the temperature increase in the eastern and central Tibetan Plateau over the period 1961—2005. The different result between the present conclusion and that in You et al. (2008b) could be explained by the different research period, 1961—2005, in You et al. (2008b) and the present study, 1961—2014. The faster increase in the high-altitude regions is mainly contributed by the variation in the recent decade (Figure 2). This result further supports the perspective reviewed in Rangwala and Miller (2012) that it is still uncertain whether the highaltitude regions generally are warming faster than the other regions due to the inadequacies in observations and the climatic internal variability.

Figure 5. Spatial distribution of the summer (a, b) average temperature (rate of Tmean; units: °C yr-1) and (c, d) maximum temperature(rate of Tmax; units: °C yr-1) warming rate in (a, c) 1975—1994 and (b, d) 1995—2014. The dots show the sites at which the trend in mean temperature is statistically significant with a confidence of 90% according to the r-test.

A previous study by Sun et al. (2008b), which investigated the temperature trend over China during 1951—2001, showed that the warming trend in Southwest China is much weaker than that in North China. In Wu et al.(2017), using datasets from 1961 to 2011, the rate of temperature increase over Southwest China is comparable to the rate of increase over other Chinese regions. In contrast,based on the extended period to 2014 here, the results show that the warming trend during the recent two decades in Southwest China increases much faster, even than the rate of high-latitude regions.

The acceleration of average temperature and extreme high temperature in Southwest China during 1995—2014 differs from the warming hiatus in global-mean surface temperature (Kosaka and Xie 2013; Huang et al. 2017; Ma et al. 2017). The extreme high temperature in Southwest China could be greatly affected by the large-scale circulation (Guan et al. 2015; Li 2015; Ma et al. 2017). Ma et al.(2017) suggested that the consistent increase in radiativeforced temperature at the high-altitude regions could be the reason for the non-appearance of a warming hiatus in the Tibetan Plateau, and thus in Southwest China.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by the National Natural Science Foundation of China [grant numbers 41722504 and 41975116] and the Youth Innovation Promotion Association of the Chinese Academy of Sciences [grant number 2016074].