Abstract

The characteristics of heat waves (HWs) in southern China in 1979–2010 are studied by using both reanalysis and station datasets. Guangdong Province of China (GDPC) is taken as an example. It is found that the westward movement of the western North Pacific subtropical high (WNPSH) is the primary factor for large-scale HWs occurring in GDPC. When an HW occurs, a hot and dry atmospheric column prevails over southern China. The region is overlaid by anomalous subsidence, which leads to warming, and clear sky, which causes greater solar heating. HWs are accompanied by an anomalous high pressure center and anticyclone near the surface, with anomalous land–sea northwesterly flow, thus reducing sea–land moisture transport and drying the atmosphere over land. The evolution of the high pressure anomaly and high temperature is associated with the westward displacement of WNPSH, with a prominent positive anomaly in 500-hPa height migrating westward. All these features associated with HWs in GDPC coincide with high-temperature extremes in the whole region of southern China and parts of Indochina. Significant increases in HW frequency (+0.19 events decade−1), HW days (+2.86 days decade−1), the duration of the longest event (+0.38 days decade−1), and the hottest temperature of the hottest event (+0.23°C decade−1) are also observed. These upward trends are more prominent in the Pearl River delta (PRD) region, and urbanization contributes to nearly 50% of the increase in HW frequency in PRD. It is also noticed that HWs are commencing earlier and ending later, and urbanization may advance the timing of the onset of HW events.

1. Introduction

Heat waves (HWs) exert notable influences on public health since they not only directly lead to fatalities because of heat stroke but also increase the risk of insomnia, fatigue, various respiratory and cardiovascular diseases, etc. (IPCC 2012; Meehl and Tebaldi 2004). The HW event that affected California for two weeks in 2006 led to at least 140 deaths (Knowlton et al. 2009). During the summer of 2013, a total of 5758 HW-related illness cases were reported in China (Gu et al. 2016). It is a great concern that the HWs are projected to become more intense, more frequent, and longer lasting in the coming decades (Cowan et al. 2014; Kunkel et al. 2010; Lau and Nath 2014). The increasing threat of HW calls for an improved understanding of the physical causes of HWs. Such studies are of much interest to the academic community and provide useful information for formulating mitigation and adaptation measures to cope with climate change.

HW events are often accompanied by notable changes in the ambient atmospheric circulation and precipitation fields, as well as in the conditions of the nearby land and ocean surfaces (Lau and Nath 2012, 2014). The spatiotemporal characteristics and associated mechanisms of HW have been a focus of investigation in recent studies. For example, long-term trends in the HWs observed over the United States have been reported by Gaffen and Ross (1998), Kunkel et al. (1999), and Peterson et al. (2013). Lau and Nath (2012) examined the synoptic characteristics of HWs in various regions of North America using a high-resolution model simulation. Attention has also been devoted to HW behavior in the European region (Della-Marta et al. 2007; Lau and Nath 2014). In particular, the HW events in 2003 and 2010 are noteworthy (Barriopedro et al. 2011; Black et al. 2004; Kovats et al. 2004). The 2003 and 2010 summers were likely the warmest on record over approximately 25% of Europe (Barriopedro et al. 2011). Approximately 35 000 people died as a result of the European HW event in 2003, the most severe heat-related mortality event in recorded history (Poumadere et al. 2005). The 2010 HW event exceeded the 2003 episode in terms of spatial extent and caused around 55 000 deaths in Russia (Barriopedro et al. 2011; Grumm 2011). Moreover, HWs have frequently been observed in Asian monsoonal regions and caused even more severe upheavals to human society and ecosystems, because of Asia’s denser population and rapid socioeconomic development (Chen and Lu 2015; Ding et al. 2010; Kothawale et al. 2010; Revadekar et al. 2013).

Based on station observations, Ding et al. (2010) suggested that strengthening of the western North Pacific subtropical high (WNPSH) results in anomalously high frequencies of HWs in eastern China. The interannual variability of HWs in southeastern China is shown to be modulated by the El Niño–Southern Oscillation (ENSO)–monsoon coupled system, which involves the East Asian monsoon, the WNPSH, the East Asian jet stream and the South Asia high (W. Wang et al. 2013). A recent case study revealed that the poleward displacement of the East Asian jet stream, associated with a prominent westward extension of the WNSPH, is the primary cause of the HW events over the southwestern and southeastern parts of China in the summers of 2003 and 2006 (Wang et al. 2016). However, a single case may not uncover all the relevant factors and therefore cannot provide a comprehensive description of various HW characteristics.

As one of the most densely urbanized and populated regions in the world, Guangdong Province of China along with the nearby areas such as Macau and Hong Kong (hereafter collectively referred as GDPC) is one of the main hubs of China’s fast economic growth and has a population of more than 116 million people (Liu et al. 2014; National Bureau of Statistics of China 2015). Heat extremes have caused much damage to GDPC society in the past and are projected to have even more severe impacts in the future (ARRCCSC 2013). GDPC is located in the subtropics along the south coast of China (Fig. 1). Its climate bears a strong relationship with the East Asian monsoon, and it is influenced by the evolution of the WNPSH. Because of the particular socioeconomic and weather and climate setting of GDPC, it is likely that, compared with other regions, HWs in GDPC may exhibit different characteristics, synoptic behavior, and responses to urbanization.

Fig. 1.

A map of the study regions. Green dots indicate the locations of weather stations. Pink box denotes the PRD region.

Fig. 1.

A map of the study regions. Green dots indicate the locations of weather stations. Pink box denotes the PRD region.

The focus of the present study is to analyze synoptic features and circulation patterns accompanying the occurrence of HWs in GDPC and to examine the long-term changes in HW characteristics. Moreover, the possible contributions of urbanization to HW activity are estimated by comparing the HW characteristics in the Pearl River delta (PRD) and non-PRD regions of GDPC. More than half of the population of GDPC lives within the PRD with a population density of 1190 persons km−2 (National Bureau of Statistics of China 2015). This study is expected to enhance our understanding of HWs in southern China and to provide guidance for enhancing our skill in forecasting the occurrence of HWs in that region.

The remainder of this paper is organized as follows. Section 2 introduces the datasets used in this research. Results on the identification of HW events in GDPC are presented in section 3. Section 4 provides a full discussion of the synoptic development and large-scale atmospheric circulation associated with HWs. Section 5 examines the long-term change of HW characteristics and assesses the influence of urbanization. The main conclusions and discussion are offered in section 6.

2. Datasets

The primary datasets used in this study are obtained from the Climate Forecast System Reanalysis (CFSR) products (http://cfs.ncep.noaa.gov/cfsr) generated by the National Centers for Environmental Prediction (NCEP). The product has a fine spatial resolution of 38 km, and its utility for studying various regional details of the observed climate system has been demonstrated (Lau and Nath 2014). This global reanalysis dataset covers the period of 1979–2010. More details of the data input and procedures used for generating the CFSR products are documented by Saha et al. (2010). The observed patterns presented in this study are obtained from daily grids of maximum near-surface temperature Tmax, relative humidity, sea level pressure (SLP), near-surface wind, 500-hPa geopotential height, and horizontal wind at the 850- and 250-hPa levels. In our study, HWs are identified by daily time series of Tmax since it is in accordance with common climatological practice in many regions such as China (Xu et al. 2009), the Czech Republic (Huth et al. 2000), Denmark (Christensen 2006), Sweden (Åström et al. 2014), and the Netherlands (Huynen et al. 2001). The Tmax-derived HW has also been shown to be closely associated with substantial societal impacts on human health and economic activities (Beniston 2004; Meehl and Tebaldi 2004).

In addition to the reanalysis data, station-based observations are also examined. These daily rain gauge precipitation data (SURF_CLI_CHN_MUL_DAY_V3.0 dataset) are provided by the China Meteorological Data Center (http://data.cma.gov.cn/). Daily observations at 35 ground stations in the GDPC region are collected for the 1979–2010 period. A map showing the locations of the 35 stations is presented in Fig. 1.

To examine the correspondence between Tmax in the reanalysis products and station-observed Tmax, we linearly interpolated the gridded values of Tmax in the CFSR dataset to sites of the individual weather stations, and the results are then compared with those observed at weather stations. Figure 2a shows the scatterplot of CFSR-interpolated Tmax versus station-observed Tmax. There is a high correlation (ρ = 0.82; p < 0.001) between the two sets of summer Tmax, showing a high degree of consistency. Also, Fig. 2b depicts the two time series of Tmax at the Guangzhou station, the center and the largest city of GDPC, based on the CFSR and station datasets. It is evident that these two time series bear a strong correspondence with each other. These results indicate that CFSR can accurately capture the comprehensive characteristics of Tmax at the station scale. Hence, we focused mainly on the results derived from CFSR in the following analysis.

Fig. 2.

Comparison between Tmax observed at weather stations and the CFSR-interpolated Tmax (°C). (a) Scatterplot of station-observed vs CFSR-interpolated Tmax at all available stations in the summers (from 1 May to 30 Sep) of 1979–2010, with a straight red line showing their linear regression model; (b) daily time series of station-observed (blue) and CFSR-interpolated (red) Tmax at the Guangzhou station in the summers of 1979 and 1980. Dashed black line in (a) denotes the 1:1 line. Dashed vertical line in (b) denotes the separation of the summers of 1978 and 1980.

Fig. 2.

Comparison between Tmax observed at weather stations and the CFSR-interpolated Tmax (°C). (a) Scatterplot of station-observed vs CFSR-interpolated Tmax at all available stations in the summers (from 1 May to 30 Sep) of 1979–2010, with a straight red line showing their linear regression model; (b) daily time series of station-observed (blue) and CFSR-interpolated (red) Tmax at the Guangzhou station in the summers of 1979 and 1980. Dashed black line in (a) denotes the 1:1 line. Dashed vertical line in (b) denotes the separation of the summers of 1978 and 1980.

3. Identification of heat waves

In this study, we focused on the May–September (MJJAS) season in the years 1979–2010. We considered the entire GDPC region as a whole and identified HW events over that region (Fig. 1). HW events are identified by daily time series of the spatial average of Tmax. The data at land points only (or stations) are used in computing this average. The 90th and 75th percentile values (hereafter referred to as T1 and T2, respectively) of this population of daily data in all available years are noted. An HW event is identified in the region when the spatial mean of Tmax is higher than T1 for three consecutive days or more, the event-averaged value of Tmax is higher than T1, and Tmax is higher than T2 throughout the event (Lau and Nath 2012, 2014).

An empirical orthogonal functions (EOF) analysis is conducted of the Tmax anomalies over the wider southern China area (i.e., 20°–27°N, 108°–120°E) to examine the spatial structure of the Tmax field in this region. Anomalies are obtained by removing the climatological seasonal cycle, as obtained by computing the multiyear averages for individual calendar days and then performing a 31-day running mean, from the daily time series for various years. The leading EOF mode explains 58% of the total variance. It is observed in Fig. 3 that the first EOF mode exhibits a strong degree of spatial uniformity over southern China, implying that the spatial average over the region is a good measure of the dominant mode of HW variability in that area. This deduction is reinforced by the high correlation between the temporal coefficient of the first EOF mode and the spatial mean of Tmax over GDPC (0.89; significant at the 99.9% confidence level). These results indicate that it is appropriate to use the spatial average of Tmax to identify the HW events occurring in that region.

Fig. 3.

Regression chart of Tmax in MJJAS 1979–2010 vs the normalized temporal coefficients of the first leading EOF mode of Tmax. The number in parentheses in the title indicates the fraction of variance explained by the EOF mode.

Fig. 3.

Regression chart of Tmax in MJJAS 1979–2010 vs the normalized temporal coefficients of the first leading EOF mode of Tmax. The number in parentheses in the title indicates the fraction of variance explained by the EOF mode.

Based on the spatial average of Tmax, the reanalysis and station datasets yielded altogether 66 and 60 HW events, respectively. Figure 4 depicts the occurrence of the identified HWs in the summers of various years. Out of the 60 HWs derived from the station dataset, 46 events (i.e., 77%) and 309 HW days (i.e., 63%) overlap with those derived from the CFSR dataset, indicating that the identification of HWs is robust across different datasets. It is noteworthy from Fig. 4 that the occurrence of HWs exhibits significant decadal changes. For both reanalysis and station datasets, HWs mainly occurred in late July and early August in the 1980s. However, since the 1990s they tend to occur more frequently in mid- and late August, early September, and late June.

Fig. 4.

Occurrences of regional HWs in GDPC during 1979–2010, as derived from (a) CFSR and (b) weather station observations. Shading indicates the spatial average of Tmax. Shaded strips from dark red to light red denote from strong to weak Niño winters, respectively.

Fig. 4.

Occurrences of regional HWs in GDPC during 1979–2010, as derived from (a) CFSR and (b) weather station observations. Shading indicates the spatial average of Tmax. Shaded strips from dark red to light red denote from strong to weak Niño winters, respectively.

As indicated in Fig. 4, the El Niño events of the 1982/83 and 1997/98 winters were followed by significantly prolonged HW activities. When comparing HW statistics during El Niño and non–El Niño years, we estimate that the average number of HW days, the average duration of the yearly longest HWs, and the average duration of all HWs during El Niño years are 19, 13.9, and 10.3 days, respectively. These values in El Niño years are significantly higher than those in non–El Niño years (i.e., 13.5, 9.4, and 7.7 days, respectively). The differences in the HW frequency and severity between El Niño and non–El Niño years are marginal.

4. Synoptic behavior and atmospheric circulations

In this section, we use the identified HW events to construct composite patterns so as to illustrate the typical synoptic patterns associated with this phenomenon. The synoptic characteristics are detected by examining the composite anomalies of the pertinent variables for the HW events. In the composite analysis, an average is first taken over the duration of each HW event; these means over the 66 cases on the CFSR data are then averaged (by giving equal weight to each event) to yield the composite values.

To find out if there is contamination between the events occurring close to each other, we compare the results for the cases for which there are no heat waves in the preceding 15 days (Teng et al. 2016). With this criteria, 44 (out of 66) HW events for the CFSR dataset are identified. Then the composite analyses are performed for these 44 HW events, and we find that the results are virtually identical to that for all identified 66 HWs. Moreover, composite analyses are also performed on different categories of HWs including the earliest, longest, and strongest HWs in a given year, respectively. The results on the earliest, longest, and strongest HWs are nearly the same as those on all identified 66 HWs. These findings suggest the robustness of our analysis results, and we, therefore, present the composite patterns for all HWs in the following subsections.

a. Near-surface patterns

The distributions of surface anomaly patterns, including Tmax, precipitation, SLP, near-surface horizontal wind vector, and total cloud cover are shown in Fig. 5. Positive temperature anomalies are seen to extend much beyond the GDPC region and cover all of southern China as well as parts of Indochina (Fig. 5a). The region and the surrounding area experience suppressed precipitation (Fig. 5a). HWs are also accompanied by higher than normal SLP in the vicinity of GDPC (Fig. 5b). These regions are also under the influences of an anomalous anticyclonic flow near the surface (Fig. 5c). The presence of this anticyclone induces an anomalous northwesterly flow from the coastal regions of southern China to the South China Sea, thus implying a reduction in the moisture transport from the sea to southern China. Such decreased moisture transport is likely to be responsible for the hot and dry conditions over southern China. However, the southwesterly flow on the northwest side of the anticyclone can transport more moisture to the northern regions such as the middle and lower Yangtze River basin (YRB) regions, thus leading to increased precipitation over these regions (see Fig. 5a). As shown in Fig. 5c, the southwesterly flow also favors an anomalous convergence over the middle and lower YRB regions, increasing precipitation there. Moreover, the warm and dry anomalies over southern China are accompanied by decreased cloud cover (Fig. 5d). Less cloud cover is consistent with increased solar radiation arriving at the surface (not shown), which would contribute to warming the surface in this region.

Fig. 5.

Composite charts of anomalies of (a) precipitation (shading) and Tmax (contours, interval: 0.3°C), (b) SLP, (c) near-surface wind, and (d) total cloud cover for HWs occurring in GDPC. The border of GDPC is indicated by the thick green outline. Orography is indicated in (c) using color shading. Dashed contours denote significance at the 95% confidence level.

Fig. 5.

Composite charts of anomalies of (a) precipitation (shading) and Tmax (contours, interval: 0.3°C), (b) SLP, (c) near-surface wind, and (d) total cloud cover for HWs occurring in GDPC. The border of GDPC is indicated by the thick green outline. Orography is indicated in (c) using color shading. Dashed contours denote significance at the 95% confidence level.

It is also noteworthy that when HWs occur in GDPC, there are high-temperature anomalies in eastern Indochina (Fig. 5a). Such hot extremes are accompanied by suppressed precipitation over that region (Fig. 5a). As shown in Fig. 5c, the anomalous anticyclone centered over Hainan is accompanied by anomalously strong easterly winds over the eastern coastal regions of Indochina. The upslope motion on the western side of the coast mountain range in Indochina is weakened as a result of the presence of anomalous easterly flow over the region (see the orography depicted in Fig. 5c), and the southwest monsoon flow over Indochina is weakened. These anomalies lead to warm and dry (see Fig. 5a) conditions on the western side of that mountain range.

Anomalies in different meteorological variables associated with HWs in GDPC have also been computed using the station dataset. To compare the surface conditions during HWs to the CFSR, we use more stations (i.e., 81 stations) across a wider region and calculate the composite pattern from observations at these stations in the wider region. Figure 6 shows the composite charts of anomalous Tmax, precipitation, sunshine duration, and evaporation at individual stations. These patterns are consistent with those derived from the CFSR dataset. When HWs occur in GDPC, the entire southern China region experiences high-temperature extremes (Fig. 6a) and deficient precipitation (Fig. 6b). HWs coincide with prolonged sunshine duration (Fig. 6c), and this pattern is linked to the reduced cloud cover and enhanced solar radiation (Fig. 5d). Figure 6d shows that enhanced evaporation accompanies HWs in GDPC.

Fig. 6.

Composite charts of anomalies of (a) Tmax, (b) precipitation, (c) sunshine duration, and (d) evaporation for station-derived HWs occurring in GDPC. Pink box denotes the PRD region.

Fig. 6.

Composite charts of anomalies of (a) Tmax, (b) precipitation, (c) sunshine duration, and (d) evaporation for station-derived HWs occurring in GDPC. Pink box denotes the PRD region.

b. Vertical structure

HWs in GDPC are not only characterized by hot and dry air near the surface but also coincident with a warm and dry air column above that region. Figure 7 shows the latitude–altitude cross sections of temperature and relative humidity within the zone between 110° and 120°E. In the case of temperature (Fig. 7a), at the low level, high temperatures are found around 25°N, corresponding to the extreme temperature pattern near the surface (see Fig. 7a). This high-temperature anomaly extends to the upper troposphere. In the case of humidity (Fig. 7b), HWs are associated with a dry air column above the region. The largest reduction in relative humidity occurs between 500 and 300 hPa, and the largest specific humidity reduction appears between 700 and 500 hPa.

Fig. 7.

Composite vertical–meridional cross sections of anomalies of (a) temperature and (b) specific (contours, interval: 0.1 g kg−3) and relative humidity (shading) of the 110°E–120°E zone for HWs occurring in GDPC.

Fig. 7.

Composite vertical–meridional cross sections of anomalies of (a) temperature and (b) specific (contours, interval: 0.1 g kg−3) and relative humidity (shading) of the 110°E–120°E zone for HWs occurring in GDPC.

Figure 7a indicates that the warm temperature anomaly tilts northward with increasing height and a warming center appears at around the 300-hPa level at 35°N. This upper-level warming is likely related to the anticyclone at the upper atmosphere (see Fig. 8c), which is associated with strong westerly (easterly) anomalies to the north (south) of around 35°N. This pattern suggests a northward displacement of the East Asian jet stream (EAJS), which is associated with the westward extension of WNPSH. As shown in Fig. 9a, the westward extension of WNPSH covers parts of northern China and reaches as far north as 35°N.

Fig. 8.

Composite charts of anomalies of (a),(c) wind and (b),(d) velocity potential (shading) and divergent wind at (a),(b) 850- and (c),(d) 250-hPa levels for HWs occurring in GDPC. Gray shading in (a),(c) denotes significance at the 95% confidence level.

Fig. 8.

Composite charts of anomalies of (a),(c) wind and (b),(d) velocity potential (shading) and divergent wind at (a),(b) 850- and (c),(d) 250-hPa levels for HWs occurring in GDPC. Gray shading in (a),(c) denotes significance at the 95% confidence level.

Fig. 9.

Composite charts of anomalous (a) 500-hPa geopotential height and (b) altitude–longitude cross section of geopotential height along 20°–27.5°N for HWs occurring in GDPC. Dashed contours denote significance at the 95% confidence level.

Fig. 9.

Composite charts of anomalous (a) 500-hPa geopotential height and (b) altitude–longitude cross section of geopotential height along 20°–27.5°N for HWs occurring in GDPC. Dashed contours denote significance at the 95% confidence level.

The above synoptic characteristics are related to changes in large-scale circulations. Figure 8 shows the composite anomalies of low-level (i.e., 850 hPa) and high-level (i.e., 250 hPa) atmospheric circulation associated with HWs. At the low level, as expected, an anomalous anticyclone appears over southern China (Fig. 8a), corresponding to the anticyclone appearing near the surface (see Fig. 5c). On the other hand, an anomalous cyclone appears over the western North Pacific (WNP) region. Anomalous northerlies prevail over southeastern China and the South China Sea. These winds reduce the moisture transport from ocean to land, thus leading to suppressed precipitation over southern China. In the upper troposphere (i.e., 250 hPa), an anticyclone dominates the circulation over the East Asian continent. A weak cyclone anomaly is discernible over northern Indochina. Furthermore, a prominent convergence center is evident in the upper atmosphere over southern China (Fig. 8d) and a divergence center appears at the low level (Fig. 8b). These two centers are associated with downward air motion over southern China. This subsiding flow also contributes to the warm and dry condition in that region. At the same time, a relatively weaker upward air motion, characterized by a low-level convergence center and a high-level divergence center, is noticeable over the WNP region. These anomalies are likely linked to the westward shift of the WNPSH, as discussed in the following subsection.

c. Western North Pacific subtropical high

The changes in the position of WNPSH associated with HWs are depicted in Fig. 9. Figure 9 shows the horizontal distribution and longitude–altitude cross section of the composite anomalies of the geopotential height field. The 500-hPa geopotential height anomaly field exhibits a high center over the land, covering most parts of southern China and southeastern China, and a low center over the WNP region. These patterns imply that a westward movement of the WNPSH accompanies HWs in GDPC. This circulation change is partially responsible for the anomalous subsidence over southern China. The westward displacement of the WNPSH is even more remarkable between 700 and 500 hPa, as shown in Fig. 9b. HWs are accompanied by a high anomaly on the west side of the WNPSH and a low anomaly on its east side. These findings suggest that the westward penetration of the WNPSH may play a crucial role in the development of HWs in GDPC.

We proceed to examine the evolution of the WNPSH at both the incipient and termination stages of HWs (Fig. 10). Figures 10a,c,e,g show that the westward movement of the WNPSH corresponds well to the initiation of HWs. Three days before the beginning of a typical HW (day −3), the edge of the WNPSH reaches the southeastern coastal regions of China. The WNPSH continues to extend westward into southern China. On the day before the HW starts (day −1), the WNPSH dominates most parts of southern China. On the day when the HW begins (day 0), the WNPSH reaches the northeastern coastline of Indochina.

Fig. 10.

Composite chart of 500-hPa geopotential height (a),(c),(e),(g) before and (b),(d),(f),(h) after the HWs occurred in GDPC. Day 0 in (g),(h) denotes the day when HW begins (or ends), day −1 denotes the day before day 0, etc.

Fig. 10.

Composite chart of 500-hPa geopotential height (a),(c),(e),(g) before and (b),(d),(f),(h) after the HWs occurred in GDPC. Day 0 in (g),(h) denotes the day when HW begins (or ends), day −1 denotes the day before day 0, etc.

As shown in Figs. 10b,d,f,h, the end of HWs is also associated with change in the WNPSH. On the third day before the HW ends (i.e., day −3), the WNPSH dominates the entire southern China region, and it begins to weaken and to retreat from southern China on day −2. The area of southern China covered by the WNPSH continues to decrease on day −1, and the WNPSH almost disappears from the region of GDPC on the day when the HW ends (i.e., day 0).

Additional insights into the role of the WNSPH in HW development can be gained by studying the evolution of the synoptic characteristics associated with HWs (Fig. 11). Composite charts in Fig. 11 are arranged from top to bottom for the time from −3 to 0 days at a 1-day interval. These patterns portray the temporal sequence of evolution of the 500-mb height (1 mb = 1 hPa), SLP, surface winds, Tmax, and precipitation associated with HWs in GDPC. As shown in Fig. 11, the positive anomaly centers of SLP and the 500-mb geopotential height appear near 130°E three days before the HW onset (day −3). This positive center gradually intensifies and moves westward, and it dominates southern China one day before the HW starts (day −1). High-temperature, anticyclone, and deficient precipitation anomalies also appear over the southeastern coastal region of China about three days before the onset of HWs (day −3). These anomalies then gradually intensify, extend southwestward, and eventually cover most parts of southeastern China on day −1 (see second to fourth columns of Fig. 11). The strong correspondence between temporal sequence of various meteorological fields and the development of the WNPSH confirms that the westward displacement of the WNPSH is critical for the formation of HWs over the GDPC region. Previous examination of changes in the WNPSH found a westward extension since the late 1970s, which is partly due to the atmospheric response to the warming in the Indian Ocean–western Pacific (Zhou et al. 2009). This interdecadal change of the WNPSH may also play a significant role in inducing more frequent HWs in GDPC.

Fig. 11.

Composite charts of the evolution of (left) anomalous 500-mb geopotential height, (left center) SLP and near-surface wind, (right center) Tmax, and (right) precipitation prior to the occurrence of HW in GDPC. Day 0 denotes the day when HW begins, day −1 denotes the day before day 0, etc. Green boundary denotes the location of GDPC.

Fig. 11.

Composite charts of the evolution of (left) anomalous 500-mb geopotential height, (left center) SLP and near-surface wind, (right center) Tmax, and (right) precipitation prior to the occurrence of HW in GDPC. Day 0 denotes the day when HW begins, day −1 denotes the day before day 0, etc. Green boundary denotes the location of GDPC.

Figure 12 shows the evolution of various fields in the termination stage of HWs. As in Fig. 11, composite charts in Fig. 12 are arranged from top to bottom with increasing time lags, from −3 to +2 days at 1-day intervals. About two days before the end of the HWs (i.e., day −2), the positive anomaly of 500-mb geopotential height over eastern China becomes weaker (Fig. 12, left). This feature continues to diminish in the following days, as it migrates toward the continental interior. Simultaneously, the negative geopotential height anomaly over the WNP intensifies and moves gradually westward. On day 0, the negative anomaly covers the whole GDPC region. In addition, the high SLP anomaly and the associated anticyclonic circulation centering over Hainan (i.e., in the western part of GDPC) weaken and migrate westward at this termination stage (Fig. 12, left center). On the day at the end of the HW, negative SLP and cyclonic wind anomalies appear over the GDPC region. Moreover, the negative precipitation anomaly over southern China becomes weaker, and it vanishes from GDPC on the end day of the HW (Fig. 12, right). High-temperature anomalies over southern China move westward (Fig. 12, right center). High temperatures extend to southwestern China and parts of northern Indochina on days +1 and +2.

Fig. 12.

As in Fig. 11, but for during the end of HW in GDPC. Day −1 (+1) denotes the day before (after) the end of the HW, etc.

Fig. 12.

As in Fig. 11, but for during the end of HW in GDPC. Day −1 (+1) denotes the day before (after) the end of the HW, etc.

d. Upper-troposphere pattern

The preceding analyses show that HWs in GDPC and southern China are attributable to the westward movement of the WNPSH and are accompanied by anomalously high pressure, anticyclonic flows, and deficient precipitation. These patterns are also associated with upper-troposphere changes.

Figure 13 shows the composite chart of 200-hPa wind and geopotential height. The height and wind anomalies exhibit a wavelike pattern with centers of alternating polarities spreading across the entire midlatitude belt. This result suggests that HWs may be linked with planetary-scale circulation features. Moreover, these individual centers exhibit no significant spatial displacements throughout the onset stage of the HWs (not shown). Similar wavelike patterns associated with heat waves have also been found in the European, Russian, and North American regions (Lau and Nath 2014; Schubert et al. 2011; Teng et al. 2013).

Fig. 13.

Composite chart of anomalous 250-hPa wind (vectors) and geopotential height (shading) one day before the occurrence of HW in GDPC.

Fig. 13.

Composite chart of anomalous 250-hPa wind (vectors) and geopotential height (shading) one day before the occurrence of HW in GDPC.

5. Long-term changes in HW characteristics and effects of urbanization

In the previous section, we have identified the synoptic features and large-scale circulations associated with HWs occurring in GDPC and southern China. Now we proceed to examine the long-term changes in various measures of HW activity. These measures include the yearly number of HW events (HWN), yearly sum of HW days (HWF), length of the longest yearly event (HWD), hottest day of hottest yearly event (HWA; i.e., amplitude), average length of all yearly events (HWL), average magnitude of all yearly events (HWM), the onset date of the first event of the year (HWT1), and the termination date of the last event of the year (HWT2). Time series of the HW measures as computed using CFSR data (weather station observation) are depicted in Fig. 14 (Fig. 15). Their trends in 1979–2010 are estimated by the Mann–Kendall trend analysis as described by Sen (1968), and their significance is obtained via Student’s t test (Table 1).

Fig. 14.

Time series of (a) MJJAS Tmax, (b) HWN, (c) HWF, (d) HWD, (e) HWA, (f) HWL, (g) HWM, and (h) HWT1 and HWT2 as computed for CFSR. Straight lines indicate their corresponding linear trends.

Fig. 14.

Time series of (a) MJJAS Tmax, (b) HWN, (c) HWF, (d) HWD, (e) HWA, (f) HWL, (g) HWM, and (h) HWT1 and HWT2 as computed for CFSR. Straight lines indicate their corresponding linear trends.

Fig. 15.

As in Fig. 14, but for weather station observations.

Fig. 15.

As in Fig. 14, but for weather station observations.

Table 1.

Sen’s trends (decade−1) of heat wave characteristics in GDPC, PRD, and non-PRD regions in 1979–2010. Boldface values denote significance at the 95% level.

Sen’s trends (decade−1) of heat wave characteristics in GDPC, PRD, and non-PRD regions in 1979–2010. Boldface values denote significance at the 95% level.
Sen’s trends (decade−1) of heat wave characteristics in GDPC, PRD, and non-PRD regions in 1979–2010. Boldface values denote significance at the 95% level.

a. Long-term changes

As shown in Fig. 14, all HW measures except for HWL and HWT1 exhibit prominent upward trends. HWN has increased significantly by 0.19 events decade−1 since 1979. HWF bears a trend of +2.86 days decade−1. HWD exhibits a slight prolonging trend of 0.38 days decade−1, suggesting that the longest HW of the year is becoming longer. Regarding the HW severity, both HWA and HWM show increasing trends (i.e., 0.23° and 0.03°C decade−1, respectively), implying that the HW amplitude is becoming stronger. Besides the upward trends in HW measures, the Tmax averaged throughout the MJJAS season also shows a significant warming trend of 0.54°C decade−1 (see Fig. 14a and Table 1).

The upward trends in HW measures based on station data (Fig. 15) are about twice as strong as those based on CFSR data. For station observations, HWN and HWF exhibit increasing trends of 0.48 events decade−1 and 5.00 days decade−1, respectively. HWD exhibits a slight prolonging trend of 2.43 days decade−1. HWA and HWM show increasing trends of 0.21° and 0.14°C decade−1, respectively. For both the CFSR and station datasets, HWL (i.e., the average length of all yearly events) does not show a significant increasing or decreasing trend for the whole study period. However, our analysis reveals a downward trend in the HW duration before the late 1990s but a statistically significant lengthening of HW duration after the 1990s (i.e., 3.33 and 2.86 days decade−1 for CFSR and station observations, respectively). A possible reason for this shift is the distinct increase in precipitation in the southeastern part of China around the mid-1990s (Kwon et al. 2007). Wu et al. (2010) suggested that the pronounced increase in precipitation in southern China around the early 1990s is likely to be induced by an increase in the Tibetan Plateau snow cover and an increase in sea surface temperature (SST) in the equatorial Indian Ocean. Recent studies also revealed such influence of SST in the tropical Indian Ocean on precipitation in southern China (Chen et al. 2016; Li et al. 2016).

Besides the duration and severity of HWs, we also examine the trends of the timing of HWs during the 1979–2010 period. Figures 14h and 15h depict the time series of HWT1 (i.e., the onset date of the first HW event) and HWT2 (i.e., the ending date of the last event) for the CFSR and station datasets, respectively. The onset date (i.e., HWT1) exhibits a downward trend of 1.64 (5.24) days decade−1 for CFSR (station observations), while the demise date (i.e., HWT2) exhibits an upward trend of 12.0 (5.37) days decade−1 for CFSR (station observations). These results suggest that HWs are commencing earlier and ending later. The downward trend for the onset date is even stronger since 1998 [i.e., 15.67 (29.29) days decade−1 for CFSR (station observations)].

b. Effects of urbanization

The positive trends of various HW measures are attributable to not only greenhouse gas warming but also urbanization at the regional scale. To gain an appreciation of the possible influences of urbanization on HWs, HWs are identified in PRD and non-PRD regions separately. PRD is the metropolitan center of GDPC (see Fig. 1) and has been experiencing the most rapid urbanization in GDPC. Yearly time series of HW characteristics in PRD and non-PRD regions are presented in Fig. 16.

Fig. 16.

Time series of (left)–(right) MJJAS Tmax, HWN, HWF, HWD, HWA, HWT1, and HWT2 as computed by averaging over PRD sites (red) and non-PRD sites (blue) in (top) CFSR and (bottom) weather station observations. Dashed lines indicate their corresponding linear trends, and only trends significant at the 95% confidence level are shown.

Fig. 16.

Time series of (left)–(right) MJJAS Tmax, HWN, HWF, HWD, HWA, HWT1, and HWT2 as computed by averaging over PRD sites (red) and non-PRD sites (blue) in (top) CFSR and (bottom) weather station observations. Dashed lines indicate their corresponding linear trends, and only trends significant at the 95% confidence level are shown.

The increasing trend of MJJAS mean Tmax in PRD is stronger than that in the non-PRD region [i.e., 0.61°C (0.41°C) and 0.54°C (0.36°C) decade−1 in PRD and non-PRD, respectively, for the CFSR (station observation) dataset]. These results suggest that the most urbanized area (i.e., PRD) exhibits a more prominent warming trend than the other regions (i.e., non-PRD), implying that urbanization plays an important role in regional warming. The CFSR-derived (station derived) HW frequency (i.e., HWN) increases by 0.44 (1.00) decade−1 in PRD and 0.23 (0.51) in non-PRD. Compared with the non-PRD region, HW frequency in PRD is higher by 47.8% for CFSR and 49% for station observations. The HW days (i.e., HWF) increase by 3.00 (7.29) days decade−1 in PRD and 2.86 (5.00) days decade−1 in non-PRD for CFSR (station observations), suggesting 1.3% (31.4%) more HW days in PRD than non-PRD regions. Moreover, by using the ANOVA test, we compare the HW trends at individual stations and find that the increasing trends in HW frequency and HW days at PRD stations are significantly larger than those at non-PRD stations. The average trends in HWN, HWF, and HWD at PRD stations are 0.73 ± 0.19 events decade−1 (where ± denotes the comparison interval at the 95% confidence level) and 6.33 ± 1.51 and 1.69 ± 0.56 days decade−1, respectively, whereas those at non-PRD stations are 0.52 ± 0.11 events decade−1 and 4.15 ± 0.87 and 0.52 ± 0.33 days decade−1, respectively. It is noteworthy that the difference is mainly due to urbanization at the regional scale since the overall impact of greenhouse gas warming should be similar at the regional scale. Therefore, these results imply that local urbanization plays an important role (i.e., nearly 50%) in increasing the occurrence of HWs. Moreover, the hottest day of the hottest events (i.e., HWA) increases by 0.21°C (0.26°C) decade−1 in PRD and 0.04°C (0.14°C) decade−1 in non-PRD for CFSR (station observations). Nevertheless, the differences in the trends of HWL in PRD and non-PRD are marginal.

Another effect of urbanization on HWs is the timing of HWs. As shown in the last two columns of Fig. 16, the onset time of HW (i.e., HWT1) in both PRD and non-PRD exhibits an advancing trend while the ending date (i.e., HWT2) shows a delaying trend. HWT1 in PRD and non-PRD advances by 4.81 (6.25) and 0.17 (4.32) days decade−1, respectively, for CFSR (weather station observations). This result indicates that HWs in PRD are emerging earlier than those in the non-PRD region. Considering that the overall impact of climate change should be similar at such a regional scale, the difference between the trends of HWT1 in PRD and non-PRD is mainly induced by urbanization, which is suggested to advance the occurrence of HWs in the summer season. On the other hand, the trend of ending time of HWs (i.e., HWT2) in PRD is lower than non-PRD. HWT2 is delayed by 9.43 (5.45) days decade−1 in PRD and 12.1 (8.71) days decade−1 in non-PRD for CFSR (weather station observations), implying that urbanization may advance the ending date of HWs. The advancing role of urbanization on both the emerging and ending date of HWs is likely due to the fact that the urban region is heated and cooled at a faster rate than the nonurban region at the beginning and end of summer, respectively. A recent study by Qian et al. (2016) also revealed that urbanization plays an important role in advancing the timing of the season cycle.

6. Conclusions and discussion

In this study, we examined the characteristics of heat waves over the GDPC region. Examination of the synoptic behavior of HWs indicates that high surface temperatures in the HW region are overlaid by a hot and dry air column. HWs are also accompanied by anomalous positive surface pressure and anticyclonic circulation. The prevalent anomalous offshore, northwesterly flow reduces the moisture transport from sea to land, thus leading to dry conditions over the southern China region. The region is covered by warming and dominated by anomalously sinking air column and clear sky, which prolongs sunshine duration and enhances the solar radiation arriving at the surface, thus leading to greater solar heating. The evolution of the SLP and temperature anomalies with time is associated with the westward displacement of the WNPSH. These findings suggest that the westward movement of the WNPSH is partially responsible for the occurrence of HWs over the GDPC region. In addition, we also showed that HWs are linked to a planetary-scale circumglobal wave chain in the midlatitude zone of the Northern Hemisphere.

An analysis of the trends in HW measures derived from CFSR data reveals significant upward trends in the frequency of HW occurrence, the number of HW days per year, and HW severity in the entire 1979–2010 study period. It is noted that the increases in HW duration, frequency, and the number of HW days per year have accelerated after the 1990s. The upward trends in HW measures based on station observations are even stronger. Our analysis also reveals that the secular increases in HW measures are stronger for a more densely urbanized subregion (i.e., PRD). It is estimated that urbanization accounts for nearly 50% of the increasing trend of HW frequency. We also noticed that HWs are commencing earlier and ending later, and urbanization may advance the timing of the HWs.

Besides reanalysis and station datasets, some previous studies used climate models to identify the mechanisms underlying HWs, such as HWs in North America (Kunkel et al. 2010; Lau and Nath 2012), Europe (Beniston 2004; Lau and Nath 2014), and parts of Asia (Altinsoy et al. 2013). It is interesting to compare the results reported in this study with those derived from model simulations. Examination of the output from model simulation will broaden our understanding of the role of the processes mentioned above in the development of HWs, as well as the effects of projected climate changes on these associations.

As suggested by previous studies (Stefanon et al. 2012; W. Wang et al. 2013; Wang et al. 2016), extreme high temperature and its variation can be affected by various processes. For instance, W. Wang et al. (2013) reported that the “exit region” and the “tail” of the East Asian jet stream are associated with variation of extreme high temperatures in southeastern China. Their follow-up study revealed that high temperature extremes over southeastern China may be bonded to a coupled system between ENSO and East Asian summer monsoon (Wang et al. 2014). Other factors such as urbanization and land-use and/or land-cover changes have also been emphasized (Avila et al. 2012; Oleson et al. 2015; M. Wang et al. 2013). In the present study, we showed that densely urbanized areas exhibit more prominent increasing trends in HW measures, and the difference between PRD and non-PRD is more noticeable in CFSR than weather station observations. However, the relative influences of urbanization, the related land-use and/or land-cover changes, and other individual mechanisms in the development of present-day and future heat waves call for further quantitative study.

Human comfort and wellness are affected not only by high temperature extremes but also by humidity, the radiation regime, and wind speed (Houghton 1985). Public health is more severely impacted by simultaneously high temperature and humidity (Pal and Eltahir 2015; Robinson 2001). In the present study, an HW event is defined on the basis of temperature only. To address more directly the impacts of extreme weather events on human society, it is of great interest to extend our investigation by combining high temperature and high humidity extremes. These analyses should be performed in future studies.

Acknowledgments

This study is partially supported by the National Natural Science Foundation of China (Grant 41401052). The appointment of NCL at The Chinese University of Hong Kong is partially supported by the AXA Research Fund. The appointment of ML at The Chinese University of Hong Kong is supported by the Postdoctoral Fellowship Scheme of the Faculty of Social Science (Grant 3132220) and the Focused Innovations Scheme (Grant 1907001).

REFERENCES

REFERENCES
Altinsoy
,
H.
,
T.
Ozturk
,
M.
Turkes
, and
M.
Kurnaz
,
2013
: Simulating the climatology of extreme events for the central Asia domain using the RegCM 4.0 regional climate model. Advances in Meteorology, Climatology and Atmospheric Physics, C. G. Helmis and P. T. Nastos, Eds., Springer, 365–370.
ARRCCSC
,
2013
: Assess report on region climate change in south China 2012: Summary for policymakers. China Meteorological Press, 150 pp.
Åström
,
C.
,
K. L.
Ebi
,
J.
Langner
, and
B.
Forsberg
,
2014
:
Developing a heatwave early warning system for Sweden: Evaluating sensitivity of different epidemiological modelling approaches to forecast temperatures
.
Int. J. Environ. Res. Public Health
,
12
,
254
267
, doi:.
Avila
,
F.
,
A.
Pitman
,
M.
Donat
,
L.
Alexander
, and
G.
Abramowitz
,
2012
:
Climate model simulated changes in temperature extremes due to land cover change
.
J. Geophys. Res.
,
117
,
D04108
, doi:.
Barriopedro
,
D.
,
E. M.
Fischer
,
J.
Luterbacher
,
R. M.
Trigo
, and
R.
García-Herrera
,
2011
:
The hot summer of 2010: Redrawing the temperature record map of Europe
.
Science
,
332
,
220
224
, doi:.
Beniston
,
M.
,
2004
:
The 2003 heat wave in Europe: A shape of things to come? An analysis based on Swiss climatological data and model simulations
.
Geophys. Res. Lett.
,
31
,
L02202
, doi:.
Black
,
E.
,
M.
Blackburn
,
G.
Harrison
,
B.
Hoskins
, and
J.
Methven
,
2004
:
Factors contributing to the summer 2003 European heatwave
.
Weather
,
59
,
217
223
, doi:.
Chen
,
J.
,
Z.
Wen
,
R.
Wu
,
X.
Wang
,
C.
He
, and
Z.
Chen
,
2016
:
An interdecadal change in the intensity of interannual variability in summer rainfall over southern China around early 1990s
.
Climate Dyn.
, doi:.
Chen
,
R.
, and
R.
Lu
,
2015
:
Comparisons of the circulation anomalies associated with extreme heat in different regions of eastern China
.
J. Climate
,
28
,
5830
5844
, doi:.
Christensen
,
O. B.
,
2006
: Regional climate change in Denmark according to a global 2-degree-warming scenario. Danish Climate Centre Rep. 06-02, 17 pp.
Cowan
,
T.
,
A.
Purich
,
S.
Perkins
,
A.
Pezza
,
G.
Boschat
, and
K.
Sadler
,
2014
:
More frequent, longer, and hotter heat waves for Australia in the twenty-first century
.
J. Climate
,
27
,
5851
5871
, doi:.
Della-Marta
,
P. M.
,
J.
Luterbacher
,
H.
von Weissenfluh
,
E.
Xoplaki
,
M.
Brunet
, and
H.
Wanner
,
2007
:
Summer heat waves over western Europe 1880–2003, their relationship to large-scale forcings and predictability
.
Climate Dyn.
,
29
,
251
275
, doi:.
Ding
,
T.
,
W.
Qian
, and
Z.
Yan
,
2010
:
Changes in hot days and heat waves in China during 1961–2007
.
Int. J. Climatol.
,
30
,
1452
1462
, doi:.
Gaffen
,
D. J.
, and
R. J.
Ross
,
1998
:
Increased summertime heat stress in the US
.
Nature
,
396
,
529
530
, doi:.
Grumm
,
R. H.
,
2011
:
The central European and Russian heat event of July–August 2010
.
Bull. Amer. Meteor. Soc.
,
92
,
1285
1296
, doi:.
Gu
,
S.
,
C.
Huang
,
L.
Bai
,
C.
Chu
, and
Q.
Liu
,
2016
:
Heat-related illness in China, summer of 2013
.
Int. J. Biometeor.
,
60
,
131
137
, doi:.
Houghton
,
D. D.
,
1985
: Handbook of Applied Meteorology. John Wiley & Sons, 1461 pp.
Huth
,
R.
,
J.
Kyselý
, and
L.
Pokorná
,
2000
:
A GCM simulation of heat waves, dry spells, and their relationships to circulation
.
Climatic Change
,
46
,
29
60
, doi:.
Huynen
,
M. M.
,
P.
Martens
,
D.
Schram
,
M. P.
Weijenberg
, and
A. E.
Kunst
,
2001
:
The impact of heat waves and cold spells on mortality rates in the Dutch population
.
Environ. Health Perspect.
,
109
,
463
470
, doi:.
IPCC
,
2012
: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation: Special Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, 582 pp.
Knowlton
,
K.
, and Coauthors
,
2009
:
The 2006 California heat wave: Impacts on hospitalizations and emergency department visits
.
Environ. Health Perspect.
,
117
,
61
67
, doi:.
Kothawale
,
D.
,
J.
Revadekar
, and
K. R.
Kumar
,
2010
:
Recent trends in pre-monsoon daily temperature extremes over India
.
J. Earth Syst. Sci.
,
119
,
51
65
, doi:.
Kovats
,
S.
,
T.
Wolf
, and
B.
Menne
,
2004
: Heatwave of August 2003 in Europe: Provisional estimates of the impact on mortality. Eurosurveillance, 8, 2409. [Available online at http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=2409.]
Kunkel
,
K. E.
,
R. A.
Pielke
Jr.
, and
S. A.
Changnon
,
1999
:
Temporal fluctuations in weather and climate extremes that cause economic and human health impacts: A review
.
Bull. Amer. Meteor. Soc.
,
80
,
1077
1098
, doi:.
Kunkel
,
K. E.
,
X.-Z.
Liang
, and
J.
Zhu
,
2010
:
Regional climate model projections and uncertainties of U.S. summer heat waves
.
J. Climate
,
23
,
4447
4458
, doi:.
Kwon
,
M.
,
J. G.
Jhun
, and
K. J.
Ha
,
2007
:
Decadal change in East Asian summer monsoon circulation in the mid‐1990s
.
Geophys. Res. Lett.
,
34
,
L21706
, doi:.
Lau
,
N.-C.
, and
M. J.
Nath
,
2012
:
A model study of heat waves over North America: Meteorological aspects and projections for the twenty-first century
.
J. Climate
,
25
,
4761
4784
, doi:.
Lau
,
N.-C.
, and
M. J.
Nath
,
2014
:
Model simulation and projection of European heat waves in present-day and future climates
.
J. Climate
,
27
,
3713
3730
, doi:.
Li
,
C.
,
R.
Lu
, and
B.
Dong
,
2016
:
Interdecadal changes on the seasonal prediction of the western North Pacific summer climate around the late 1970s and early 1990s
.
Climate Dyn.
,
46
,
2435
2448
, doi:.
Liu
,
X.
,
L.
Ma
,
X.
Li
,
B.
Ai
,
S.
Li
, and
Z.
He
,
2014
:
Simulating urban growth by integrating landscape expansion index (LEI) and cellular automata
.
Int. J. Geogr. Inf. Sci.
,
28
,
148
163
, doi:.
Meehl
,
G. A.
, and
C.
Tebaldi
,
2004
:
More intense, more frequent, and longer lasting heat waves in the 21st century
.
Science
,
305
,
994
997
, doi:.
National Bureau of Statistics of China
,
2015
: Fourth one-percent national sample census. National Bureau of Statistics of China, 975 pp.
Oleson
,
K.
,
G.
Anderson
,
B.
Jones
,
S.
McGinnis
, and
B.
Sanderson
,
2016
:
Avoided climate impacts of urban and rural heat and cold waves over the U.S. using large climate model ensembles for RCP8.5 and RCP4.5
.
Climatic Change
, doi:.
Pal
,
J. S.
, and
E. A. B.
Eltahir
,
2015
:
Future temperature in southwest Asia projected to exceed a threshold for human adaptability
.
Nat. Climate Change
,
6
,
197
200
, doi:.
Peterson
,
T. C.
, and Coauthors
,
2013
:
Monitoring and understanding changes in heat waves, cold waves, floods, and droughts in the United States: State of knowledge
.
Bull. Amer. Meteor. Soc.
,
94
,
821
834
, doi:.
Poumadere
,
M.
,
C.
Mays
,
S.
Le Mer
, and
R.
Blong
,
2005
:
The 2003 heat wave in France: Dangerous climate change here and now
.
Risk Anal.
,
25
,
1483
1494
, doi:.
Qian
,
C.
,
G.
Ren
, and
Y.
Zhou
,
2016
:
Urbanization effects on climatic changes in 24 particular timings of the seasonal cycle in the middle and lower reaches of the Yellow River
.
Theor. Appl. Climatol.
,
124
,
781
791
, doi:.
Revadekar
,
J.
, and Coauthors
,
2013
:
Impact of altitude and latitude on changes in temperature extremes over South Asia during 1971–2000
.
Int. J. Climatol.
,
33
,
199
209
, doi:.
Robinson
,
P. J.
,
2001
:
On the definition of a heat wave
.
J. Appl. Meteor.
,
40
,
762
775
, doi:.
Saha
,
S.
, and Coauthors
,
2010
:
The NCEP Climate Forecast System Reanalysis
.
Bull. Amer. Meteor. Soc.
,
91
,
1015
1057
, doi:.
Schubert
,
S.
,
H.
Wang
, and
M.
Suarez
,
2011
:
Warm season subseasonal variability and climate extremes in the Northern Hemisphere: The role of stationary Rossby waves
.
J. Climate
,
24
,
4773
4792
, doi:.
Sen
,
P. K.
,
1968
:
Estimates of the regression coefficient based on Kendall’s tau
.
J. Amer. Stat. Assoc.
,
63
,
1379
1389
, doi:.
Stefanon
,
M.
,
F.
D’Andrea
, and
P.
Drobinski
,
2012
:
Heatwave classification over Europe and the Mediterranean region
.
Environ. Res. Lett.
,
7
,
014023
, doi:.
Teng
,
H.
,
G.
Branstator
,
H.
Wang
,
G. A.
Meehl
, and
W. M.
Washington
,
2013
:
Probability of US heat waves affected by a subseasonal planetary wave pattern
.
Nat. Geosci.
,
6
,
1056
1061
, doi:.
Teng
,
H.
,
G.
Branstator
,
G. A.
Meehl
, and
W. M.
Washington
,
2016
:
Projected intensification of subseasonal temperature variability and heat waves in the Great Plains
.
Geophys. Res. Lett.
,
43
,
2165
2173
, doi:.
Wang
,
M.
,
X.
Yan
,
J.
Liu
, and
X.
Zhang
,
2013
:
The contribution of urbanization to recent extreme heat events and a potential mitigation strategy in the Beijing–Tianjin–Hebei metropolitan area
.
Theor. Appl. Climatol.
,
114
,
407
416
, doi:.
Wang
,
W.
,
W.
Zhou
,
X.
Wang
,
S. K.
Fong
, and
K. C.
Leong
,
2013
:
Summer high temperature extremes in southeast China associated with the East Asian jet stream and circumglobal teleconnection
.
J. Geophys. Res. Atmos.
,
118
,
8306
8319
, doi:.
Wang
,
W.
,
W.
Zhou
, and
D.
Chen
,
2014
:
Summer high temperature extremes in southeast China: Bonding with the El Niño–Southern Oscillation and East Asian summer monsoon coupled system
.
J. Climate
,
27
,
4122
4138
, doi:.
Wang
,
W.
,
W.
Zhou
,
X.
Li
,
X.
Wang
, and
D.
Wang
,
2016
:
Synoptic-scale characteristics and atmospheric controls of summer heat waves in China
.
Climate Dyn.
,
46
,
2923
2941
, doi:.
Wu
,
R.
,
Z.
Wen
,
S.
Yang
, and
Y.
Li
,
2010
:
An interdecadal change in southern China summer rainfall around 1992/93
.
J. Climate
,
23
,
2389
2403
, doi:.
Xu
,
J.
,
Z.
Deng
, and
M.
Chen
,
2009
:
A summary of studying on characteristics of high temperature and heat wave damage in China (in Chinese)
.
J. Arid Meteor.
,
27
,
163
167
.
Zhou
,
T.
, and Coauthors
,
2009
:
Why the western Pacific subtropical high has extended westward since the late 1970s
.
J. Climate
,
22
,
2199
2215
, doi:.