Long-Term Changes in Summer Extreme Wet Bulb Globe Temperature over China

中国夏季湿球温度长期变化特征分析

+ Author Affiliations + Find other works by these authors
  • Corresponding author: Ying SUN, sunying@cma.gov.cn
  • Funds:

    Supported by the National Key Research and Development Program of China (2018YFC1507702) and National Natural Science Foundation of China (42025503)

  • doi: 10.1007/s13351-021-1080-4
  • Note: This paper will appear in the forthcoming issue. It is not the finalized version yet. Please use with caution.

PDF

  • The wet bulb globe temperature (WBGT) has important implication for human health. Previous studies widely use the monthly data but rarely investigate the extreme WBGT because of data limitation. In this study, we use 6-h station data to analyze the changes in the WBGT and three categories (intensity indices, absolute threshold indices, and frequency indices) of extreme WBGT indices in summers of 1961–2017. It is found that the spatial distributions of long-term trends in summer mean WBGT are consistent with those in mean temperature. The trend value of WBGT is smaller than the mean temperature, because of the decrease of relative humidity. For the extreme indices, the intensity and frequency of WBGT and fixed threshold indices have changed. The increase of intensity indices and warm WBGT days and nights, and decrease of cold WBGT days and nights have been observed in most China, especially over northwestern China. The number of days with daily maximum WBGT exceeding 31.4°C (WXge31) and minimum WBGT exceeding 27.9°C (WNge27) over southeastern China have increased since 1961. The spread of probability distributions of WXge31, WNge27, and warm WBGT days and nights is becoming wider, reflecting increased variability of extreme indices. In addition, urbanization effects on the WBGT are investigated. The impacts of urbanization on most of extreme WBGT indices are not detected, except for absolute thresholds indices. This may be due to the decrease of relative humidity in urban stations, which is almost two times larger than that in rural stations. However, we also note that the homogenization issue of humidity data may affect our conclusions.
    利用6小时分辨率台站观测数据,对1961–2017年中国夏季湿球温度(wet bulb globe temperature,WBGT)的变化及3类极端WBGT指数(强度、绝对阈值和频率)的长期变化特征进行分析。夏季WBGT长期变化趋势的空间分布与平均气温基本一致,但趋势值较小,相对湿度的降低是造成这种差异的主要原因。对于极端WBGT指数,中国大部分地区都观测到强度指数的上升,尤其是西北地区,这与中国的变暖和西北地区的暖湿化是一致的。1961年以来,中国东南部地区,特别是长江中下游地区日最高气温超过31.4℃ (WXge31)和最低气温超过27.9℃ (WNge29)的天数呈增加趋势。与1961–1990年相比,1991–2017年绝对阈值指数和暖昼(WX90p)、暖夜(WN90p)的概率密度分布变宽,反映了极端指数变率的增大。强度和频率指数与中国夏季平均气温呈近似线性关系,绝对阈值指数随平均气温的升高而加速增加。城市化对极端WBGT长期变化影响的分析表明,除绝对阈值指数外,其他大部分指数均没有检测到城市化的影响,这可能是由于城市地区的“干岛效应”(即城市站点相对湿度的下降趋势接近乡村站点的两倍)所致。
  • 加载中
  • Fig. 1.  Summer mean temperature (TEM), wet bulb globe temperature (WBGT), relative humidity (RHU), and wet bulb globe temperature calculated based on monthly mean temperature (WBGT_mon) and humidity long-term trends and anomalies (relative to 1981–2010 mean). The upper panel shows linear trends at individual stations during 1961–2017, and color scales indicate the magnitude of the trends. The lower panel shows time series. The blue lines in Figs. 1f, h represent time series of WBGT and RHU based on homogenized daily TEM and RH dataset of China National Surface Weather Station (V1.0).

    Fig. 2.  As in Fig. 1, but for daily maximum and minimum temperaure (TEM_max and TEM_min), wet bulb globe temperature (WBGT_max and WBGT_min), specific humidity (SHU_max and SHU_min), and relative humidity (RHU_max and RHU_min) during 1961–2017.

    Fig. 3.  Linear trends of summer extreme WBGT indices at individual stations during 1961−2017. Color scales indicate the magnitude of the trends.

    Fig. 4.  Time series of summer extreme WBGT index anomalies (relative to 1981−2010 mean) for China national mean.

    Fig. 5.  Histograms and Gaussian-fit probability density functions of summer extreme WBGT index anomalies (relative to 1981−2010 mean) from 1961−1990 (light blue bars and blue line) and 1991−2017 (orange bars and red line). (Units: same as Fig.4).

    Fig. 6.  Scatter plots of the summer extreme WBGT index anomalies (relative to 1981−2010 mean) and China national mean temperature anomalies (x-axis). Correlation coefficients are shown in the bottom right-hand corner of each panel.

    Fig. 7.  Time series of summer mean daily maximum and minimum WBGT(WBGT_max, WBGT_min, units: °C), relative humidity (RHU_Max, RHU_min, units: %) and summer extreme WBGT index anomalies (Units: same as Fig.4) (relative to 1981-2010 mean) for urban stations (red line), rural stations(blue line) and their difference (black line).

    Table 1.  Extreme WBGT indices used in this study

    Short nameLong nameDefinitionUnit
    WXxMax WBGT_XMonthly maximum value of daily maximum WBGT°C
    WNnMin WBGT_NMonthly minimum value of daily minimum WBGT°C
    WNxMax WBGT_NMonthly maximum value of daily minimum WBGT°C
    WXnMin WBGT_XMonthly minimum value of daily maximum WBGT°C
    WXge27WBGT_X of at least 27°CMonthly count of days when WBGT_X ≥ 27.9°Cday
    WXge29WBGT_X of at least 29°CMonthly count of days when WBGT_X ≥ 29.4°Cday
    WXge31WBGT_X of at least 31°CMonthly count of days when WBGT_X ≥ 31.4°Cday
    WNge27WBGT_N of at least 27°CMonthly count of days when WBGT_N ≥ 27.9°Cday
    WX10pAmount of cool daysPercentage of days when WBGT_X < 10th percentile%
    WX90pAmount of hot daysPercentage of days when WBGT_X > 90th percentile%
    WN10pAmount of cold nightsPercentage of days when WBGT_N < 10th percentile%
    WN90pAmount of warm nightsPercentage of days when WBGT_N > 90th percentile%
    Download: Download as CSV

    Table 2.  Linear trends of summer extreme WBGT indices (unit: same as in Fig.3) over China during 1961−2017. NW: Northwestern, NE: northeastern, SW: southwestern, and SE: southeastern

    NWNESWSE
    WXx0.0100.0080.0130.005
    WNn0.0520.0440.0250.019
    WNx0.0290.0190.0140.006
    WXn0.0350.0290.0110.011
    WXge270.0030.0230.0160.372
    WXge290.0010.0060.0040.136
    WXge310.0000.0000.0000.004
    WNge270.0000.0000.0000.016
    WX10p-0.208-0.116-0.155-0.070
    WX90p0.1130.1050.1470.132
    WN10p-0.341-0.193-0.167-0.159
    WN90p0.2240.1590.2010.149
    Download: Download as CSV
  • [1]

    USARIEM: U.S. Army Research Institute of Environmental Medicine. [Available online at http://www.usariem.army.mil/heatill/precenhi.htm]. (查阅所有网上资料, 未找到本条文献信息且网址打不开, 请联系作者确认).
    [2]

    ACGIH, 1996: Threshold Limit Values for Chemical Substances and Physical Agents and Biological Exposure Indices (1996–1997). American Conference of Governmental Industrial Health, Cincinnati, OH, USA. (查阅所有网上资料, 未找到本条文献信息, 请联系作者确认).
    [3]

    Budd, G. M., 2008: Wet-bulb globe temperature (WBGT)—its history and its limitations. J. Sci. Med. Sport, 11, 20–32. doi: 10.1016/j.jsams.2007.07.003.
    [4]

    Buzan, J. R., K. Oleson, and M. Huber, 2015: Implementation and comparison of a suite of heat stress metrics within the Community Land Model version 4.5. Geosci. Model Dev., 8, 151–170. doi: 10.5194/gmd-8-151-2015.
    [5]

    Chu, Z. Y., and G. Y. Ren, 2005: Change in urban heat island magnitude and its effect on mean air temperature record in Beijing region. Acta Meteor. Sinica, 63, 534–540. doi: 10.11676/qxxb2005.053. (in Chinese)
    [6]

    Dole, R., M. Hoerling, J. Perlwitz, et al., 2011: Was there a basis for anticipating the 2010 Russian heat wave? Geophys. Res. Lett., 38, L06702. doi: 10.1029/2010GL046582.
    [7]

    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. doi: 10.1038/nclimate1827.
    [8]

    Fouillet, A., G. Rey, V. Wagner, et al., 2008: Has the impact of heat waves on mortality changed in France since the European heat wave of summer 2003? A study of the 2006 heat wave. Int. J. Epidemiol., 37, 309–317. doi: 10.1093/ije/dym253.
    [9]

    ISO, 1989: ISO 7243: 1989 Hot environments – Estimation of the heat stress on working man, based on the WBGT-index (wet bulb globe temperature). Geneva: International Standards Organization.
    [10]

    Knutson, T. R., and J. J. Ploshay, 2016: Detection of anthropogenic influence on a summertime heat stress index. Climatic Change, 138, 25–39. doi: 10.1007/s10584-016-1708-z.
    [11]

    Li, C., X. B. Zhang, F. Zwiers, et al., 2017: Recent very hot summers in Northern Hemispheric land areas measured by wet bulb globe temperature will be the norm within 20 years. Earth’s Future, 5, 1203–1216. doi: 10.1002/2017EF000639.
    [12]

    Li, C., Y. Sun, F. Zwiers, et al., 2020: Rapid warming in summer wet bulb globe temperature in China with human-induced climate change. J. Climate, 33, 5697–5711. doi: 10.1175/jcli-d-19-0492.1.
    [13]

    Liang, C. Z., G. Z. Zheng, N. Zhu, et al., 2011: A new environmental heat stress index for indoor hot and humid environments based on Cox regression. Build. Environ., 46, 2472–2479. doi: 10.1016/j.buildenv.2011.06.013.
    [14]

    Liljegren, J. C., R. A. Carhart, P. Lawday, et al., 2008: Modeling the wet bulb globe temperature using standard meteorological measurements. J. Occup. Environ. Hyg., 5, 645–655. doi: 10.1080/15459620802310770.
    [15]

    National Meteorological Information Center, 2019: Daily Meteorological Dataset of Basic Meteorological Elements of China National Surface Weather Station (V3.0)(1951–2010). A Big Earth Data Platform for Three Poles. Available online at http://poles.tpdc.ac.cn/zh-hans/data/52c77e9c-df4a-4e27-8e97-d363fdfce10a/. (in Chinese)
    [16]

    Parsons, K., 2006: Heat stress standard ISO 7243 and its global application. Ind. Health, 44, 368–379. doi: 10.2486/indhealth.44.368.
    [17]

    Qian, C., 2016: On trend estimation and significance testing for non-Gaussian and serially dependent data: Quantifying the urbanization effect on trends in hot extremes in the megacity of Shanghai. Climate Dyn., 47, 329–344. doi: 10.1007/s00382-015-2838-0.
    [18]

    Qian, C., X. B. Zhang, and Z. Li, 2019: Linear trends in temperature extremes in China, with an emphasis on non-Gaussian and serially dependent characteristics. Climate Dyn., 53, 533–550. doi: 10.1007/s00382-018-4600-x.
    [19]

    Ren, G. Y., and Y. Q. Zhou, 2014: Urbanization effect on trends of extreme temperature indices of national stations over mainland China, 1961–2008. J. Climate, 27, 2340–2360. doi: 10.1175/JCLI-D-13-00393.1.
    [20]

    Ren, G. Y., J. Guo, M. Z. Xu, et al., 2005: Climate changes of China’s mainland over the past half century. Acta Meteor. Sinica, 63, 942–956. doi: 10.11676/qxxb2005.090. (in Chinese)
    [21]

    Ren, G. Y., Y. H. Ding, Z. C. Zhao, et al., 2012: Recent progress in studies of climate change in China. Adv. Atmos. Sci., 29, 958–977. doi: 10.1007/s00376-012-1200-2.
    [22]

    Ren, G. Y., Y. J. Yuan, Y. J. Liu, et al., 2016: Changes in precipitation over Northwest China. Arid Zone Res., 33, 1–19. doi: 10.13866/j.azr.2016.01.01. (in Chinese)
    [23]

    Ren, L. W., D. Q. Wang, N. An, et al., 2020: Anthropogenic influences on the persistent night-time heat wave in summer 2018 over Northeast China. Bull. Amer. Meteor. Soc., 101, S83–S88. doi: 10.1175/BAMS-D-19-0152.1.
    [24]

    Stull, R., 2011: Wet-bulb temperature from relative humidity and air temperature. J. Appl. Meteor. Climatol., 50, 2267–2269. doi: 10.1175/JAMC-D-11-0143.1.
    [25]

    Sun, Y., X. B. Zhang, F. W. Zwiers, et al., 2014: Rapid increase in the risk of extreme summer heat in Eastern China. Nat. Climate Change, 4, 1082–1085. doi: 10.1038/nclimate2410.
    [26]

    Sun, Y., T. Hu, X. B. Zhang, et al., 2019: Contribution of global warming and urbanization to changes in temperature extremes in Eastern China. Geophys. Res. Lett., 46, 11,426–11,434. doi: 10.1029/2019GL084281.
    [27]

    Tao, B., G. Y. Ren, X. Zhao, et al., 2020: Half-century urban drying in Shijiazhuang City. Environ. Res. Commun., 2, 075006. doi: 10.1088/2015-7620/aba10f.
    [28]

    Wang, Y. J., Y. Sun, T. Hu, et al., 2018: Attribution of temperature changes in Western China. Int. J. Climatol., 38, 742–750. doi: 10.1002/joc.5206.
    [29]

    Welton, G., 2011: The Impact of Russia’s 2010 Grain Export Ban. Oxfam Research Reports, 32 pp. (查阅所有网上资料, 未找到对应的出版信息, 请联系作者确认).
    [30]

    Yang, X. C., Y. L. Hou, and B. D. Chen, 2011: Observed surface warming induced by urbanization in east China. J. Geophys. Res. Atmos., 116, D14113. doi: 10.1029/2010JD015452.
    [31]

    Yang, X. C., L. R. Leung, N. Z. Zhao, et al., 2017: Contribution of urbanization to the increase of extreme heat events in an urban agglomeration in east China. Geophys. Res. Lett., 44, 6940–6950. doi: 10.1002/2017GL074084.
    [32]

    Zhu, Y. N., L. J. Cao, G. L. Tang, et al., 2015: Homogenization of surface relative humidity over China. Progr. Inquisit. Mutat. Climatis, 11, 379–386. doi: 10.3969/j.issn.1673-1719.2015.06.001. (in Chinese)
  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Long-Term Changes in Summer Extreme Wet Bulb Globe Temperature over China

    Corresponding author: Ying SUN, sunying@cma.gov.cn
  • 1. Laboratory for Climate Studies, National Climate Center, China Meteorological Administration, Beijing 100081
  • 2. Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters, Nanjing University of Information Science & Technology, Nanjing 210044
Funds: Supported by the National Key Research and Development Program of China (2018YFC1507702) and National Natural Science Foundation of China (42025503)

Abstract: The wet bulb globe temperature (WBGT) has important implication for human health. Previous studies widely use the monthly data but rarely investigate the extreme WBGT because of data limitation. In this study, we use 6-h station data to analyze the changes in the WBGT and three categories (intensity indices, absolute threshold indices, and frequency indices) of extreme WBGT indices in summers of 1961–2017. It is found that the spatial distributions of long-term trends in summer mean WBGT are consistent with those in mean temperature. The trend value of WBGT is smaller than the mean temperature, because of the decrease of relative humidity. For the extreme indices, the intensity and frequency of WBGT and fixed threshold indices have changed. The increase of intensity indices and warm WBGT days and nights, and decrease of cold WBGT days and nights have been observed in most China, especially over northwestern China. The number of days with daily maximum WBGT exceeding 31.4°C (WXge31) and minimum WBGT exceeding 27.9°C (WNge27) over southeastern China have increased since 1961. The spread of probability distributions of WXge31, WNge27, and warm WBGT days and nights is becoming wider, reflecting increased variability of extreme indices. In addition, urbanization effects on the WBGT are investigated. The impacts of urbanization on most of extreme WBGT indices are not detected, except for absolute thresholds indices. This may be due to the decrease of relative humidity in urban stations, which is almost two times larger than that in rural stations. However, we also note that the homogenization issue of humidity data may affect our conclusions.

中国夏季湿球温度长期变化特征分析

利用6小时分辨率台站观测数据,对1961–2017年中国夏季湿球温度(wet bulb globe temperature,WBGT)的变化及3类极端WBGT指数(强度、绝对阈值和频率)的长期变化特征进行分析。夏季WBGT长期变化趋势的空间分布与平均气温基本一致,但趋势值较小,相对湿度的降低是造成这种差异的主要原因。对于极端WBGT指数,中国大部分地区都观测到强度指数的上升,尤其是西北地区,这与中国的变暖和西北地区的暖湿化是一致的。1961年以来,中国东南部地区,特别是长江中下游地区日最高气温超过31.4℃ (WXge31)和最低气温超过27.9℃ (WNge29)的天数呈增加趋势。与1961–1990年相比,1991–2017年绝对阈值指数和暖昼(WX90p)、暖夜(WN90p)的概率密度分布变宽,反映了极端指数变率的增大。强度和频率指数与中国夏季平均气温呈近似线性关系,绝对阈值指数随平均气温的升高而加速增加。城市化对极端WBGT长期变化影响的分析表明,除绝对阈值指数外,其他大部分指数均没有检测到城市化的影响,这可能是由于城市地区的“干岛效应”(即城市站点相对湿度的下降趋势接近乡村站点的两倍)所致。
    • Since the 21st century, heat wave events have occurred frequently all over the world, causing a large number of economic losses and casualties. Extreme heat wave that hit Europe in 2003 killed tens of thousands of people. In 2010, the Russian heat wave caused the global food price to rise. In 2018, the persistent night-time heat waves in northeastern China also had a serious impact on human health (Dole et al., 2011; Welton, 2011; Ren et al., 2020). Recent attribution studies have shown that human activities have caused the increase in the frequency of extreme heat wave. In the future, such extreme high-temperature events will be likely to occur with global warming under different emission scenarios (Ren et al., 2012; Sun et al., 2014; Wang et al., 2018).

      Usually, temperature and consecutive high temperature days are used to describe the intensity of extreme heat wave events. On the one hand, temperature is the most important indicator to evaluate the effect of human heat stress. On the other hand, global, long-term, and high-quality temperature observation data are easy to obtain. However, with the further increase of global temperature, the use of temperature indicators alone will gradually become inadequate. If the air humidity is low, even at higher temperatures, human body can dissipate heat by evaporative cooling. However, in a humid and hot environment, the evaporative cooling effect caused by perspiration will be significantly weakened, which will lead to the rise of core temperature of human body. Therefore, compared with temperature indicators solitary, the comprehensive consideration of humidity and temperature indicators can reflect the impact of high temperature heat wave events on human beings more intuitively.

      Wet bulb globe temperature (WBGT) is an ISO standard for quantifying thermal comfort (ISO, 1989). As an indicator with temperature and humidity, the application of WBGT has a long history in military, sports, and workplace safety (ACGIH, 1996; Parsons, 2006; Budd, 2008). The research on heat-acclimated of soldiers by the US military (ACGIH, 1996) shows that when WBGT is lower than 28°C, military personnel can work for 45 min and rest for 15 min to undertake moderate workload. However, when WBGT rises to 31.1°C, the guidance is much more stringent: 15-min work and 45-min rest. USARIEM recommends that if the WBGT exceeds 32°C, all physical exercise and strenuous exercise should be suspended. Recent heat waves with WBGT between 29 and 31°C have caused thousands of deaths (Fouillet et al., 2008) and experimental evidence suggests that at the condition of above 32°C WBGT, most physical labors become unsafe (Liang et al., 2011; Buzan et al., 2015).

      Eastern China is one of the regions with the highest extreme WBGT and the most densely populated area in the world. Previous studies have shown that the WBGT in almost all regions of China has experienced significant warming since 1960, and there is a much faster warming after the late 1990s (Li et al., 2017, 2020). The warming may lead to a significant increase in the probability of extreme heat waves, which are rare in the mid-20th century. Most of previous studies are based on the monthly data and little study investigates the extreme WBGT because of data limitation. Since extreme high WBGT has a direct impact on the health of such a large number of people, as well as other impacts such as labor capacity and air conditioning energy demand, it is very important to focus on the variation of extreme WBGT events. Eastern China is also the area with the fastest urbanization process and urban population growth. The urban heat island effect has been proved to have a significant contribution to the increase of temperature and extreme temperature events (Ren and Zhou, 2014; Sun et al., 2019). Qian (2016) estimated that urbanization contribution about 38% and 37%−76% of the linear trend of temperature and hot extremes in Shanghai. Sun et al. (2014) reported about 24% of the trend in summer temperature in eastern China could be attributed to urbanization. Yet little has been done in exploring the impact of urbanization on humidity. It is still unclear whether or to what extent urbanization contributes to the changes of extreme WBGT events. The city is densely populated with a large number of outdoor workers (e.g., construction, delivery, etc.), which exposed to high temperatures. Therefore, the study on urbanization contribution is of great practical significance.

      In previous studies, extreme heat waves based on WBGT are rarely involved. One of the important reasons is the lack of high-quality long-term WBGT observation with high spatial and temporal resolution. Fortunately, near surface temperature and relative humidity data set provided by the National Meteorological Information Center (2019) of China Meteorological Administration provides the basis for detailed analysis of these indicators. In this paper, 6-h observation station data is used to calculate highest and lowest WBGT of the day. Then extreme WBGT indices are calculated based on daily maximum and minimum WBGT. The long-term changes of extreme WBGT indices and the relationships between WBGT indices and mean temperature are analyzed, and then the effects of urbanization are assessed.

    2.   Data and methods
    • There are two types of method for the calculations of WBGT, representing WBGT for indoor conditions and outdoor conditions, respectively. WBGT for outdoor conditions needs to consider variables such as wind speed and solar radiation. Due to the lack of reliable wind and sunlight data, WBGT is calculated as the weighted sum of wet bulb temperature (WBT) and air temperature (TA), i.e., WBGT = 0.7 × WBT + 0.3 × TA (Liljegren et al., 2008). The WBGT estimated by this method represents a lower limit for WBGT in well shaded or indoor conditions (Dunne et al., 2013; Knutson and Ploshay, 2016). WBT is calculated based on TA and relative humidity (RH) according to the method of Stull (2011), which is reliable for TA between −20 and 50°C and relative humidity of 5%−100%. In this study, quality-controlled 6-h observations of near-surface temperature and relative humidity at 2419 meteorological stations for period 1961–2017 from meteorological dataset of basic meteorological elements of China National Surface Weather Station (V3.0) (National Meteorological Information Center, 2019), which can reflect the temperature and humidity changes of China in the past 60 years, are used to calculate WBGT.

    • We define 12 extreme WBGT indices inspired by the definition of the extreme indices developed by the ETCCDI (see http://etccdi.pacificclimate.org/list_27_indices.shtml). The extreme indices include three categories: the intensity indices that are the monthly/daily maximum or minimum WBGT (WXx, WXn, WNn, and WNx), the frequency indices that are the percentage of days exceeding the percentile thresholds relative to local climate (WX10p, WX90p, WN10p, and WN90p), and the threshold indices that are the number of days exceeding the absolute thresholds 27.9, 29.4, and 31.4°C (WXge27, WXge29, WXge31, and WNge27). These three thresholds relate to the cycle of 50% work and 50% rest to conduct heavy (350−500 kcal h−1), moderate (200−350 kcal h−1), and light (< 200 kcal h−1) labor (ACGIH, 1996), respectively. Note that light labor mentioned here would be equivalent to walking and heavy labor would be far less exertive than marathon running.

      Short nameLong nameDefinitionUnit
      WXxMax WBGT_XMonthly maximum value of daily maximum WBGT°C
      WNnMin WBGT_NMonthly minimum value of daily minimum WBGT°C
      WNxMax WBGT_NMonthly maximum value of daily minimum WBGT°C
      WXnMin WBGT_XMonthly minimum value of daily maximum WBGT°C
      WXge27WBGT_X of at least 27°CMonthly count of days when WBGT_X ≥ 27.9°Cday
      WXge29WBGT_X of at least 29°CMonthly count of days when WBGT_X ≥ 29.4°Cday
      WXge31WBGT_X of at least 31°CMonthly count of days when WBGT_X ≥ 31.4°Cday
      WNge27WBGT_N of at least 27°CMonthly count of days when WBGT_N ≥ 27.9°Cday
      WX10pAmount of cool daysPercentage of days when WBGT_X < 10th percentile%
      WX90pAmount of hot daysPercentage of days when WBGT_X > 90th percentile%
      WN10pAmount of cold nightsPercentage of days when WBGT_N < 10th percentile%
      WN90pAmount of warm nightsPercentage of days when WBGT_N > 90th percentile%

      Table 1.  Extreme WBGT indices used in this study

    • To assess the effect of urbanization on the change in extreme WBGT, the rural and non-rural stations need to be classified. Many classification methods have been developed in previous studies, such as dynamic classification used by Yang et al. (2011, 2017) and fixed classification used by Ren and Zhou (2014), in which times and distance of relocations, population in the nearby cities, and the percentage of the built-up areas nearby the station are considered during the procedure. According to fixed classification (Ren and Zhou, 2014), 142 stations among 2400 stations are chosen as rural station in this study. These rural stations have good spatial distribution and density, with only few of 5º × 5º grid boxes (will be described in the next section) having no rural stations, which are located along the coastal lines or national boundaries. The length and quality of rural station records also meet the needs of this study. Previous studies (e.g., Sun et al., 2019) have shown that these data can be used to analyze the difference between the rural and non-rural stations and urbanization impacts on the changes in temperature extremes. Following the methods using in the study of Chu and Ren (2005), Cu is calculated to measure urbanization contribution as follows:

      $$ {C}_{\rm u}=\left|\frac{{T}_{\rm u}-{T}_{\rm r}}{{T}_{\rm u}}\right|\times 100\text%, $$ (1)

      where Tu is the linear trend of index in urban station and Tr is trend in rural station.

    • WBGT is calculated based on observed data at 2419 stations in mainland China. The temporal resolution of observation station data is 6 h, so we can obtain the WBGT at 0200, 0800, 1400, and 2000 LST (local standard time). The highest and lowest WBGT of the day are chosen as daily maximum and minimum WBGT. The WBGT of 1400 and 0200 LST are not used to represent the daily maximum and minimum WBGT directly, although they usually occur at these two times. The monthly anomalies of extreme WBGT indices are computed relative to 1981–2010 baseline period, and summer (June–July–August, JJA) mean indices are calculated by indies of three summer months. For the intensity indices, e.g., WXx, the maximum of extreme index in JJA is selected as the value for the summer index. For the absolute threshold index, e.g., WXge27, the sum of extreme index in JJA is calculated as the value for the summer. For the frequency index, e.g., WX10p, the average of extreme index in JJA is calculated as the value for the summer. Also, because of many missing values found before 1970, the period of 1981–2010 is chosen as the baseline period. The resulting indices anomalies are aggregated to 5° × 5° grid by averaging all available station anomalies within a predefined grid cell. The similar data processing procedure has been used in previous studies to investigate near surface temperature changes in China (Sun et al., 2014; Wang et al., 2018). The spatial average can be estimated relatively accurate according to the non-uniform station observations.

    3.   Results
    • Figure 1 shows the long-term trends and temporal variations of summer mean temperture and WBGT during 1961–2017. Summer WBGT has increased over most part of China with increased temperature. The changes in WBGT in northern China are more obvious than those in southern China, and more obvious in western China than in eastern China. The most pronounced increases occur in northwestern China, with over 0.03°C yr−1 warming oberseved in majority area of Xinjiang Region and Qinghai Province. This is consistent with the rapid warming and moistening of northwestern China (Ren et al., 2005, 2016). Compared the trends of temperature and WBGT, it can be found that the spatial distribution characteristics of both are consistent, but the trend value of WBGT is smaller than temperature, which is the result of the effect of relative humidity on WBGT. As shown in Fig. 1d, significant drops in relative humidity can be found in most China, especially in areas with obvious increased temperature. The decrease in relative humidity partly offsets the increase in WBGT caused by the increase in temperature.

      Figure 1.  Summer mean temperature (TEM), wet bulb globe temperature (WBGT), relative humidity (RHU), and wet bulb globe temperature calculated based on monthly mean temperature (WBGT_mon) and humidity long-term trends and anomalies (relative to 1981–2010 mean). The upper panel shows linear trends at individual stations during 1961–2017, and color scales indicate the magnitude of the trends. The lower panel shows time series. The blue lines in Figs. 1f, h represent time series of WBGT and RHU based on homogenized daily TEM and RH dataset of China National Surface Weather Station (V1.0).

      We note that the relative humidity experienced a clear decrease around the early 2000s. Considering that the 6-h data used in this study are not homogenized, the homogenized daily data are used to test whether this decrease is caused by inhomogeneity. Based on homogenized daily relative humidity dataset of China National Surface Weather Station (V1.0) produced by the NMIC (Zhu et al., 2015), summer mean relative humidity is calculated, which is shown in blue line in Fig. 1h. The sudden drop around the 2000s is exhibited by both datasets, with more obvious drop appeared in non-homogenized data series. This indicates that the implementation of the new observing system (mostly during 2004–2007) clearly affects the changes in relative humidity, but the decrease of relative humidity also exists in the past decades. We then compare the summer mean WBGT using the homogenized and non-homogenized relative humidity dataset (blue and red lines in Fig. 1f). We found that the WBGT changes based on two datasets are quite similar, except that the WBGT calculated by inhomogenized humidity slightly underestimate the increasing trend after 2000. Therefore, it is reasonable to believe that the 6-h dataset used in this study can reflect the major characteristics of mean and extreme WBGT changes.

      In order to understand the difference between changes in near-surface temperature and WBGT, the long-term trends and temporal variations of near-surface specific humidity and relative humidity are examined (Fig. 2). The maximum and minimum specific humidity and relative humidity here correspond to the times when the daily maximum and minimum WBGT occur (generally 1400 and 0200 LST), reflecting the difference of daytime and nighttime humidity variations. The results show that the relative humidity in most parts of China is decreasing (Figs. 2g, h). This indicates that the increasing of surface water vapor can not keep up with temperature rising. The saturated water vapor pressure increases with the increasing of temperature, while the “absolute” surface water vapor does not increase to the same extent, and the atmosphere becomes relatively less saturated. Compared the variations of WBGT in daytime and nighttime, it can be found that the warming trend of WBGT in nighttime is larger than that in daytime in most areas of northern China, especially in northwestern China. On the one hand, it is due to the increase of nighttime temperature in northwestern China. On the other hand, it also arises from the more obvious increase of specific humidity at nighttime. This result confirms the warming and wetting in northwestern China, and suggests that the warming and wetting mainly occurred at nighttime. In terms of temporal variation, the specific humidity shows fluctuating changes before the 21st century and low value after the late 1990s. The contrast is more evident in the variation of relative humidity (Figs. 2m−p). This partly explains the difference between interannual variation of temperature and WBGT, in which temperature continues to rise after 1980, while WBGT experiences a warming period in the 1980s and 1990s and then becomes fluctuating after the late 1990s. Moreover, since the decline of relative humidity in the daytime is greater than that in the nighttime after the late 1990s, the difference between temperature and WBGT is more pronounced in the daytime.

      Figure 2.  As in Fig. 1, but for daily maximum and minimum temperaure (TEM_max and TEM_min), wet bulb globe temperature (WBGT_max and WBGT_min), specific humidity (SHU_max and SHU_min), and relative humidity (RHU_max and RHU_min) during 1961–2017.

    • Figures 3 and 4 show the long-term trends and temporal variations of 12 extreme WBGT indices in summer. For the intensity indices of WBGT (WXx, WNn, WNx, and WXn), increasing trends are observed in most China, consistent with warming in maximum and minimum WBGT. The long-term trends of four indices are larger in northern China than in southern China, and larger in western China than in eastern China. The linear trends of extreme indices in northwestern China are 0.010, 0.052, 0.029, and 0.035°C yr−1 (Table 2), respectively, which are over two times larger than those in southeastern China. The increasing trend of WNn is obvious, which is consistent with the significant increase of nighttime temperature and specific humidity. For the temporal changes, the seasonal maximum related indices (WXx and WNx) are in better agreement with the summer mean WBGT than the seasonal minimum related indices (WXn and WNn), while the daily maximum WBGT indices (WXx and WXn) are in better agreement than the daily minimum WBGT indices (WNx and WNn). WXx shows upward trend before the 1990s and then fluctuating, which is consistent with the change in summer mean WBGT, while the WNn shows consistent increase since the 1960s (Figs.4a−d).

      Figure 3.  Linear trends of summer extreme WBGT indices at individual stations during 1961−2017. Color scales indicate the magnitude of the trends.

      Figure 4.  Time series of summer extreme WBGT index anomalies (relative to 1981−2010 mean) for China national mean.

      NWNESWSE
      WXx0.0100.0080.0130.005
      WNn0.0520.0440.0250.019
      WNx0.0290.0190.0140.006
      WXn0.0350.0290.0110.011
      WXge270.0030.0230.0160.372
      WXge290.0010.0060.0040.136
      WXge310.0000.0000.0000.004
      WNge270.0000.0000.0000.016
      WX10p-0.208-0.116-0.155-0.070
      WX90p0.1130.1050.1470.132
      WN10p-0.341-0.193-0.167-0.159
      WN90p0.2240.1590.2010.149

      Table 2.  Linear trends of summer extreme WBGT indices (unit: same as in Fig.3) over China during 1961−2017. NW: Northwestern, NE: northeastern, SW: southwestern, and SE: southeastern

      For the absolute threshold indices of WBGT, the increases in extreme WBGT with daily maximum WBGT exceeding 27.9 and 29.4°C, and daily minimum WBGT exceeding 27.9°C mainly occur over southeastern China, and rarely happen over northern and southwestern China. As expected, higher temperature and higher relative humidity lead to higher summer mean WBGT over southeastern China, which is favorable to the occurrence of extreme high WBGT. It is worth noting that, although the middle and lower reaches of the Yangtze River are the areas with the least obvious upward trend of summer mean WBGT and intensity indices, the increases of absolute threshold indices are the most significant. This reflects the nonlinear relationship between the absolute threshold indices and summer WBGT. Among the four indices, WXge27 shows the most remarkable rising trend, 0.372 day yr−1, followed by WXge29, 0.136 day yr−1. The number of days with WBGT above 27.9 and 29.4°C has consistently increased since 1961, which means the heat-stress burden on society continues to mount. Under the condition of daily maximum WBGT higher than 27.9 (29.4)°C, people need at least 50% rest to conduct heavy (moderate) labor, which will seriously affect the efficiency of outdoor labor. In addition, we also note the occurrence of extreme high WBGT with daily maximum exceeding 31.4°C and minimum WBGT exceeding 27.9°C over southeastern China after the 21th century. The former is even not suitable for long-term walking outside, while extreme high WBGT night is not conducive to the recovery of the heat-stress and causes stress on the body to accumulate.

      For the frequency indices, the number of warm WBGT days and nights (WX90p and WN90p) has increased and the number of cold WBGT days and nights (WX10p and WN10p) has decreased. The long-term trends in warm and cold WBGT days show similar pattern but with opposite value, as well as in warm and cold WBGT nights. The most remarkable changes are observed in northwestern China. Similar conclusions have been obtained from long-term trends of temperature frequency indices by previous studies (Qian et al., 2019), except that the most remarkable changes occur in southeastern China.

    • As discussed in last section, the extreme WBGT indices (i.e., WXx and WXn) show changes around the 1990s, so we separate the period of 1961−2017 into the periods of 1961−1990 and 1991−2017 to investigate the probability distributions of WBGT, respectively. For the intensity indices of extreme WBGT, the probability distributions of WXx, WNn, WNx, and WXn of 1991−2017 shift to warmer regimes relative to 1961–1990, which are mainly due to the increase of daily maximum and minimum WBGT (Fig. 5). The rightward shift of WXx is the least obvious among the four intensity indices, but the probability distribution during 1991−2017 becomes obviously wider than that during 1961−1990, indicating the increased variability of WXx. Combined with the previous results, although the long-term trend of WXx is small, the possibility of extreme high WBGT has increased. For the absolute threshold indices, both the shape and position of probability distributions have changed. The variability has increased for four indices, especially WNge27. Notice that extreme high WBGT rarely happen during 1961−1990, the characteristics of variability evaluation showed in Figs. 5f−h may be spurious.

      Figure 5.  Histograms and Gaussian-fit probability density functions of summer extreme WBGT index anomalies (relative to 1981−2010 mean) from 1961−1990 (light blue bars and blue line) and 1991−2017 (orange bars and red line). (Units: same as Fig.4).

      For frequency indices, probability distributions of cold WBGT days and nights shift leftward and the probability distributions of warm WBGT days and nights become wider. It is consistent with the characteristics reflected by the temporal variations of WBGT indices. The probability of cold WBGT days and nights shows a persistently decrease after 1961, while warm WBGT days and nights fluctuated before 1990 and then increased. The long-term trend contributes to the increase of variability of extreme warm WBGT. Relationships between extreme WBGT indices and summer mean WBGT anomaly are also investigated and the linear-fitting slopes of 1961−1990 and 1991−2017 are calculated, respectively (figures omitted). The dispersion of frequency indices is not large, indicating a close correlation between extreme indices and summer mean WBGT. The slopes of cold WBGT days and nights (WX10p and WN10p) in 1961−1990 are larger than those in 1991−2017. In the period of 1961−1990, the cold WBGT days and nights decreased rapidly with the slight increase of summer WBGT, while in the period of 1991−2017, the changes in the cold WBGT days and nights with the change of WBGT were only 70% of that of 1961−1990, which lead to similar variability of cold WBGT days and nights of both periods.

    • Figure 6 shows relationships between anomalies of extreme WBGT indices and summer mean temperature over China. An approximately linear relationship can be found between mean temperature and WNn, WNx, WXn, and the frequency indices. The increase in summer mean temperature can lead to warming of WNn, WNx, WXn, more frequent warm WBGT events (WX90p and WN90p), and less frequent cold WBGT events (WX10p and WN10p). Generally, changes in the intensity indices and frequency indices track well with China mean temperature change. A 1°C increase in summer mean temperature corresponds to the increase of 0.46, 0.91, 0.74, and 1.19°C in WXx, WXn, WNx, and WNn, respectively. For frequency indices, the change is 5.63%, −7.01%, 7.52%, and −8.91% in WX90p, WX10p, WN90p, and WN10p, respectively. It is clear that the changes in extreme indices related to daily minimum WBGT are larger than those related to daily maximum WBGT, and extreme indices related to percentage of days, which lower than 10th percentile of the season are larger than those higher than 90th percentile.

      Figure 6.  Scatter plots of the summer extreme WBGT index anomalies (relative to 1981−2010 mean) and China national mean temperature anomalies (x-axis). Correlation coefficients are shown in the bottom right-hand corner of each panel.

      The relationships between mean temperature and the absolute threshold indices are not as closely as other WBGT indices, the distributions of the points are more scattered, and the correlations are less significant, especially WXge31 and WNge27 (Fig. 6d). The occurrence of daily maximum WBGT over 31.4°C or daily minimum WBGT over 27.9°C increase with temperature at a faster rate when the temperature anomalies is above zero, suggesting that the extreme WBGT is accelerating at higher warming levels. This conclusion means that the occurrence frequency of extreme WBGT will increase more rapidly in feature with global warming. People will confront more risk of heat stress, in particular the children, elderly and people with ill health. To avoid the deadly impacts of future extreme WBGT, development of effective public health adaptation measures is also urgent.

    • Figure 7 shows the changes in extreme WBGT indices in urban and rural stations. Only grids with both urban and rural stations available are considered for the calculation of regional average. As described in Section 2.4, the WBGT averaged during the baseline period of 1981−2010 for urban and rural station is removed respectively before spatial average is conducted. For the mean WBGT and most extreme WBGT indices, the changes in rural and non-rural stations are quite similar and the differences are small. The variable Cu of the indices are also small, except for WXx. The upward trend of WXx in urban stations is smaller than that in rural stations, and Cu for WXx reach about 21%, showing the difference between the maximum temperature in urban and rural stations. This difference may be related to the aerosols influence on the WBGT in the urban stations but the reason is still unknown. Compared with previous studies, the urbanization contribution to the change of extreme WBGT is different from those to the extreme temperature. With the rapid development of urbanization in China, especially eastern China, the urban heat island effect could lead to enhanced extreme hight temperature events (Qian, 2016). Studies on extreme temperature events by Ren et al. (2014) 本条文献在文后文献中未体现and Sun et al. (2019) also confirmed the warming effects of urbanization on the change of extreme temperature, especially on the indices related to daily minimum temperature. Here, we note that the urbanization effect cannot be found in the changes of most extreme WBGT indices, which may be due to the drying trend in urban areas. As shown in Figs. 7c, d, there is difference between the changes of relative humidity in the urban and rural stations. The humidity in urban stations shows larger decreasing trend than rural stations, indicating a rapid drying in the urban stations. The linear trends of daily maximum and minimum relative humidity are −0.046 and −0.022% yr−1 in urban stations, almost two times larger then those in rural stations (−0.023 and −0.010% yr−1). The urbanization contributions to the maximum and minimum relative humidity decrease are over 100%. Tao et al. (2020) also showed that the urbanization contributes to the decreasing relative humidity in Shijiazhuang and Beijing about 78.7% and close to 100%, respectively, indicating the drying trend related to urbanization.

      Figure 7.  Time series of summer mean daily maximum and minimum WBGT(WBGT_max, WBGT_min, units: °C), relative humidity (RHU_Max, RHU_min, units: %) and summer extreme WBGT index anomalies (Units: same as Fig.4) (relative to 1981-2010 mean) for urban stations (red line), rural stations(blue line) and their difference (black line).

      On the other hand, slight differences can be found for four absolute thresholds indices between urban and rural stations (third row in Fig. 7). The urbanization contributions (Cu) for WXge27 and WXge29 reach 18.5% and 33.8%. One possibility is the extreme WBGT in the cities rise more than those in the rural regions so the difference between the urban and rural regions change from a negative to a positive value, with a very similar variation observed after the 1980s. However, we note the difference between the urban and rural stations mainly occurred in the early period before the 1980s, i.e., before the rapid development of cities in China. This reminds us that the difference may not be real but may arise from the data issue or other reasons.

    4.   Conclusions and Discussion
    • In this study, the extreme WBGT indices are analyzed based on 6-h observed data of 2419 stations in mainland of China. The changes in summer mean and extreme WBGT indices, and the impact of urbanization on the indices are examined. The major conclusions are as follows:

      1) The spatial distribution of change in WBGT is consistent with the mean temperature, but the trend value is smaller. The decrease of relative humidity is an important reason for the change in WBGT. The warming of daily minimum WBGT is larger than maximum WBGT, due to larger increase of temperature and less decrease of relative humidity in nighttime than those in daytime.

      2) Increases of WXx, WNn, WNx, and WXn are observed in most China, especially over northwestern China. This is consistent with the warming of China, and rapid warming and moistening of northwestern China. WXge31 and WNge29 over southeastern China have increased since 1961, especially over middle and lower reaches of the Yangtze River. Compared to 1961−1990 period, the probability distribution of absolute threshold indices and warm days (WX90p) and nights (WN90p) of 1991−2017 becomes wider, reflecting increased variability of extreme indices.

      3) The intensity and frequency indices show an approximately linear relationship with summer mean temperature over China, while the absolute threshold indices increase at an accelerated rate as the mean temperature rise.

      4) The urbanization effects cannot be found in the change of mean WBGT and most extreme WBGT indices, which may be due to drying trend in urban stations. The decrease of relative humidity in urban stations is almost two times larger than that in rural stations.

      In this study, we also find that the summer mean WBGT computed based on 6-h data and monthly data is quite similar in most regions of China, except that the monthly data slightly underestimate the cooling trend in central China. This indicates that the WBGT calculated based on monthly data reflect the long-term change of WBGT at the national scale, but it needs cautious when used at smaller spatial scale. Also, in our calculation, the daily maximum and minimum WBGT mainly occur at 1400 and 0200 LST, although 1400 and 0200 LST WBGTs are not directly selected. Usually, it is the case conforms to diurnal variations in temperature. While in extreme cases, the error of maximum and minimum WBGT based on 6-h data will be greater than that of temperature. The higher temporal resolution data (3-h or hourly) is still needed in further study.

      Acknowledgments. We thank the China National Meteorological Information Center for archiving the observational data (available at http://data.cma.cn/).

Reference (32)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return