Abstract

Southern China, located in the tropical–subtropical East Asian monsoonal region, presents a unique anticyclonic–cyclonic circulation pattern during extreme heat (EH), obviously different from the typical anticyclone responsible for EH in many other regions. Associated with the evolution of EH in southern China, the anticyclonic–cyclonic anomalies propagate northwestward over the Philippines and southern China. Before the EH onsets, the anticyclonic anomaly dominates southern China, resulting in stronger subsidence over southern China and stronger southerly (southwesterly) flow over the western (northern) margins of southern China. The southerly (southwesterly) flow transports more water vapor to the north of southern China, thus, together with the local stronger subsidence, resulting in drier air condition and accordingly favoring the occurrence of EH. Conversely, after the EH onsets, the cyclonic component approaches southern China and offsets the high temperature.

The oscillations of temperature and circulation anomalies over southern China exhibit a periodicity of about 10 days and indicate the influence of a quasi-biweekly oscillation, which originates from the tropical western Pacific and propagates northwestward. Therefore, the 5–25-day-filtered data are extracted to further analyze the quasi-biweekly oscillation. It turns out that the evolution of the filtered circulation remarkably resembles the original anomalies with comparable amplitudes, indicating that the quasi-biweekly oscillation is critical for the occurrence of EH in southern China. The quasi-biweekly oscillation could explain more than 50% of the intraseasonal variance of daily maximum temperature Tmax and vorticity over southern China and 80% of the warming amplitude of EH onsets. The close relationship between the circulation of the quasi-biweekly oscillation and the EH occurrence indicates the possibility of medium-range forecasting for high temperature in southern China.

1. Introduction

Under the background of global warming, extreme heat (EH) has become more frequent on a global scale and has attracted greater attention in the scientific literature (e.g., Lau and Nath 2012; Corobovet al. 2013; IPCC 2013; Weaver et al. 2014; Lu and Chen 2016). EH has a significant impact on human health, causing heat-related illnesses and increasing mortality, especially in the elderly (Kilbourne 1997; Díaz et al. 2002; Luber and McGeehin 2008). In fact, among heat, cold, drought, storm, and flood events, EH is reported to be the most prominent cause of weather-related human mortality (CDC 2004; WMO 2013). Moreover, EH threatens social and economic activities, increasing the consumption of electricity and water and inducing forest fires and crop losses (Valor et al. 2001; Zhang and Wang 2002; Peng et al. 2004; Coumou and Rahmstorf 2012). Therefore, EH has become an important public concern and detailed investigation into the causes of EH occurrence is an urgent demand.

Southern China, which is located in the tropical–subtropical region, experiences EH frequently in summer, with top-rank climatological frequency (~13 days yr−1) and increasing trend (~4.5 days decade−1) among the cities in eastern China (Gao et al. 2008; Wei and Chen 2009). In addition, southern China has a dense population, appearing more vulnerable to EH. For instance, the population of Guangdong Province (a metropolis in southern China) has exceeded 0.1 billion and is growing at a speed of 1.9% yr−1. Therefore, understanding the causes of EH occurrence in southern China is of great social significance.

The atmospheric general circulation responsible for EH occurrence is typically characterized by an anticyclone, which would induce subsidence and concurrent adiabatic heating, reduce cloud cover, and increase solar radiation at surface, favoring higher air temperature (e.g., Meehl and Tebaldi 2004; Wei et al. 2004; Maheras et al. 2006; Gershunov et al. 2009). However, spatial differences are likely to be significant for the circulation associated with EH in different regions. For instance, Loikith and Broccoli (2012) investigated the circulation patterns responsible for EH over North America and indicated that the pattern for the EH over the southern domain is distinct from the typical anticyclone for other regions. A similar discrepancy takes place in China. Chen and Lu (2015) systematically compared the composite synoptic circulation anomalies responsible for the EH in different regions of eastern China and found that the circulation associated with EH in southern China shows a unique pattern, which is manifested by a pair of anticyclonic and cyclonic anomalies (Fig. 8 of Chen and Lu 2015). The unique pattern implies that the influencing factors and processes for the EH in southern China may differ from the general situation and are scientifically worthy of specific study.

Most of the previous research concerning the synoptic circulation associated with EH events in southern China was in the form case studies, which reported various results. For instance, Ji et al. (2005), Lin et al. (2005), and Wang et al. (2016) analyzed the persistent EH over southern China in the summer of 2003 and indicated that the persistent dominance of subtropical high over southern China leads to the extreme high temperature. In contrast, Huang et al. (2005) showed that, for the EH event in Guangdong Province in July 2004, the circulation featured a high pressure system to the northwest and a low pressure system to the southeast of Guangdong. The discrepancy indicates great variation of circulation among different EH cases, and thus a composite analysis based on a large number of cases is necessary to obtain general and reliable features of circulation responsible for the EH in southern China. Thereby, investigating the evolution of circulation pattern associated with EH in southern China from a climatological statistical perspective forms the main motivation of the present study.

A primitive analysis of the evolution of the anticyclonic–cyclonic anomaly during EH events in southern China demonstrates that a periodicity of about 10 days is manifested in the anticyclone and cyclone successively occurring over southern China (shown later in Fig. 4), which originate from the tropical western Pacific and propagate northwestwards. This indicates that the circulation anomaly is associated with the quasi-biweekly oscillation of tropical atmosphere. The quasi-biweekly oscillation is an important intraseasonal component in the tropical atmosphere, with kinetic energy even stronger than the 30–60-day oscillation (Li and Zhou 1995). The quasi-biweekly oscillation over the western North Pacific emerges from the equatorial region and propagates northwestward (Wen and Zhang 2005; Kikuchi and Wang 2009; Chen and Sui 2010). It has been well documented that the quasi-biweekly oscillation exerts a prominent influence on the East Asian summer monsoon and the precipitation over East Asia (e.g., Krishnamurti and Ardanuy 1980; Jia and Yang 2013; Li and Zhou 2013; Yang et al. 2014; Xu and Lu 2015). For example, Chen et al. (2015) revealed that during the wet phases of the quasi-biweekly oscillation (i.e., the phases when more precipitation occurs over southern China) there is an anomalous anticyclone over the South China Sea and western North Pacific, enhancing the northward water vapor transport to southern China. However, the impact of the quasi-biweekly oscillation on the EH in southern China has not been documented in the literature. Therefore, investigating the quasi-biweekly oscillation of the anomalous anticyclonic–cyclonic pair and its impact on EH in southern China is the second motivation of the present study.

The rest of the paper is organized as follows: Data and definitions are described in section 2. In section 3, the evolution of the original circulation anomalies associated with extreme heat events in southern China are analyzed, and the results indicate that the quasi-biweekly oscillations of tropical atmospheric circulation play an important role. Therefore, the quasi-biweekly oscillations associated with EH in southern China are further analyzed in section 4, before presenting the conclusions of the study in section 5.

2. Data and definitions

a. Data

Temperature data used in the current study are the homogenized daily mean, maximum, and minimum surface air temperature series of 753 national standard stations in China, which were recently updated to include the years to 2013 (Li et al. 2016). The homogenization of temperature data efficiently revises the inhomogeneity in temperature data caused by frequent changes of observing locations and protocols (Li and Yan 2009). For example, the homogenized data show warming trends almost everywhere in China, whereas the inhomogenized data exhibit some cooling trends, especially in the southern parts.

Circulation data are the 6-hourly data from the European Centre for Medium-Range Weather Forecasts (ECMWF) interim reanalysis (ERA-Interim; Dee et al. 2011). The data at 0600 UTC [i.e., 1400 local standard time (LST)] are used to analyze the EH-related anomalies. The horizontal resolution is 1.5° × 1.5° and the vertical layers include 23 pressure levels from 1000 to 200 hPa.

The data periods span 35 summers, from July to August (JA) between 1979 and 2013. The selection of season is based on the temporal distribution of daily maximum temperature Tmax and EH frequency over southern China (shown later in Fig. 2), which is explained in detail in section 3.

b. Definitions

The definition of southern China (Fig. 1) is similar to that of Chen and Lu (2015), including stations in the southeastern domain of China that exhibit EH-related circulation patterns that are poorly correlated with the average anticyclonic pattern of eastern China. However, the western boundary of the domain is extended to 105°E and the northern boundary is extended to 28°N [cf. 110°E and 26°N in Chen and Lu (2015)]. This modification is based on the spatial distribution of the climatological Tmax averaged during JA. The climatological summer Tmax is more than 31°C for most of the stations in this domain, except the northwestern region, and the central domain exhibits Tmax more than 33°C. We repeated the composite analysis of Chen and Lu (2015) for circulation anomalies on the EH days of southern China according to the new definition and gained a similar anticyclonic–cyclonic pattern (not shown explicitly herein, but can be partially found in Fig. 4), indicating the rationality of the redefined southern China domain in the current study. Accordingly, there are totally 149 stations in southern China, as indicated by the dots in Fig. 1.

Fig. 1.

The climatological Tmax averaged for JA during 1979–2013 (contours, °C; contour interval of 1°C). The dashed rectangle denotes the domain of southern China defined for the current study. The dots denote the observational stations in southern China, and the circles denote the observational stations outside southern China.

Fig. 1.

The climatological Tmax averaged for JA during 1979–2013 (contours, °C; contour interval of 1°C). The dashed rectangle denotes the domain of southern China defined for the current study. The dots denote the observational stations in southern China, and the circles denote the observational stations outside southern China.

In the present study, an EH event is defined as days on which more than one-third of the 149 stations in southern China (i.e., over 50 stations) present Tmax exceeding 35°C. The threshold of 35°C is employed by the China Meteorological Administration to define EH and widely used by previous studies (e.g., Wei and Sun 2007; Wei and Chen 2009; Chen and Lu 2014a,b). Based on this definition, there are totally 436 EH days in southern China, accounting for 20% of the entire study period (2170 days).

The evolution of anomalies during EH events are analyzed in the present study. To better obtain the preceding signals of EH, EH days with gaps of no more than 5 days were merged into one EH event. In this way, the composite analyses from 5 days before (−5d) to one day before (−1d) the EH events could eliminate the interferences of the anomalies on EH days and clearly illustrate the preceding signals inducing EH. According to this definition, 96 EH events are gained, and the average duration of each event is 4.5 days.

Lanczos temporal filtering is performed to extract the quasi-biweekly oscillation signals. To highlight the quasi-biweekly oscillation, the annual cycle of the series is removed before the Lanczos filter is applied, through subtracting the climatological mean (from 1979 to 2013) of the identical days from the raw data. Composite analyses are performed to investigate the synoptic anomalies associated with EH events. These anomalies are compared to the whole summertime series for significance estimation, using the Student’s t test and a significance level of 95%. Effective sample lengths are used because of the persistence in the analysis series (Wilks 2006).

3. Evolution of the original circulation anomalies associated with extreme heat events in southern China

Figure 2 illustrates the daily evolution of 35-yr mean Tmax and accumulated EH occurrence from May to October. The value of Tmax gradually increases from May to June, then stays at a relatively stable level of more than 31°C in July and August, and then obviously decreases in September and October. The maximum value for Tmax occurs between mid-July and early August, with amplitude of over 32°C. The occurrence of EH is concentrated in July and August, accounting for 49% (241 days) and 40% (195 days), respectively. The proportions in June and September are much smaller, accounting for 4.7% (23 days) and 5.7% (28 days), respectively. There is only one EH day in May and one in October during 35 years. Therefore, July and August are taken as the summer season of southern China and the present study analyzes the EH events in JA, totally accounting for 89% of all the EH events.

Fig. 2.

Daily evolution of 35-yr mean Tmax (solid contour, °C) and accumulated EH occurrence (bars, days) from May to October.

Fig. 2.

Daily evolution of 35-yr mean Tmax (solid contour, °C) and accumulated EH occurrence (bars, days) from May to October.

Before giving the circulation anomalies associated with EH in southern China, the climatological summer wind filed in the lower troposphere is shown in Fig. 3. Obvious southwesterly summer monsoon prevails over the South China Sea and southern China, and the southwesterly turns into a westerly north of 30°N in eastern China. The southwesterly would transport water vapor from the South China Sea to southern China and farther north, affecting the moisture conditions and relevant radiation processes, and thus air temperature.

Fig. 3.

Climatological 700-hPa wind field averaged during JA from 1979 to 2013 (m s−1). The blue dots denote the stations in southern China. The abbreviation SCS refers to the location of the South China Sea.

Fig. 3.

Climatological 700-hPa wind field averaged during JA from 1979 to 2013 (m s−1). The blue dots denote the stations in southern China. The abbreviation SCS refers to the location of the South China Sea.

It should be mentioned that the summer climate over southern China differs before and after late July, because of the seasonal variations of western Pacific subtropical high. After late July, the subtropical high shifts northward and modifies the distribution of circulation and moisture. To check the influence of mean state on EH, we also analyzed the circulation anomalies associated with EH events in early summer (from 1 to 20 July; 39 events in total) and late summer (from 21 July to 31 August; 57 events in total), respectively. The circulation anomalies for the EH in both early and late summer turn out to be similar to the whole season (not shown). Therefore, the present study focuses on all the EH from July to August, which represents the general situation in the target summer season.

Figure 4 shows the evolution of the composite anomalies of lower-tropospheric winds (vectors), specific humidity (shadings) and midtropospheric vertical velocity (contours), from 3 days before (−3d) to 2 days after (2d) the EH onsets in southern China. The wind anomalies show an anticyclonic–cyclonic pattern propagating northwestward from the western Pacific to southern China. At −3d, there is already a significantly anomalous anticyclone to the east of southern China and a relatively weaker cyclone to the east of the Philippines (Fig. 4a). The anticyclonic–cyclonic anomaly gradually gets stronger and moves northwestward, with a strong anticyclonic anomaly over southern China at −1d and 0d (Figs. 4c and 4d, respectively). Then the anticyclone gets weaker while the cyclone to the southeast of southern China continues to develop. The anticyclone disappears at 2d and the cyclone moves to east of southern China. The circulation anomaly over southern China at 2d is nearly opposite to the anomaly at −3d, showing a half cycle of approximately 5 days and indicating the signal of quasi-biweekly oscillation. Apart from the significant anticyclonic–cyclonic pair located over southern China and Philippines, an anomalous anticyclone to the southeast of the pair is discernible north of the equator from −1d, forming an anticyclonic–cyclonic–anticyclonic wave train. This wave train originating from the equator is similar to the result of Chen and Sui (2010), who indicated that the wave train is closely associated with the equatorial Rossby wave.

Fig. 4.

Evolution of the composite anomalies of 700-hPa wind (vectors, m s−1), 700-hPa specific humidity (shading, g kg−1; interval of 0.4 g kg−1), and 500-hPa vertical velocity (dp/dt; contours, 10−4 hPa s−1; contour interval of 3 × 10−4 hPa s−1; the red contours denote subsidence, and the green contours denote ascendance) for the EH onsets in southern China (a)–(f) from −3d to 2d. Only the contours statistically significant at the 95% confidence level are plotted, and the significant vectors are in black. In (a), the blue line from 25°N, 110°E to 35°N, 120°E denotes the trace used to define the 700-hPa southwesterly northwest of southern China, which will be used later in Fig. 7c. In (c), the blue line from 0°, 145°E to 30°N, 110°E denotes the trace used in Fig. 6 to analyze the evolution of vorticity. In (d), the blue dashed box (20°–28°N, 105°–120°E) denotes the area used to define the average 700-hPa vorticity, 500-hPa vertical velocity, and 700-hPa specific humidity over southern China, used later in Figs. 7a,b.

Fig. 4.

Evolution of the composite anomalies of 700-hPa wind (vectors, m s−1), 700-hPa specific humidity (shading, g kg−1; interval of 0.4 g kg−1), and 500-hPa vertical velocity (dp/dt; contours, 10−4 hPa s−1; contour interval of 3 × 10−4 hPa s−1; the red contours denote subsidence, and the green contours denote ascendance) for the EH onsets in southern China (a)–(f) from −3d to 2d. Only the contours statistically significant at the 95% confidence level are plotted, and the significant vectors are in black. In (a), the blue line from 25°N, 110°E to 35°N, 120°E denotes the trace used to define the 700-hPa southwesterly northwest of southern China, which will be used later in Fig. 7c. In (c), the blue line from 0°, 145°E to 30°N, 110°E denotes the trace used in Fig. 6 to analyze the evolution of vorticity. In (d), the blue dashed box (20°–28°N, 105°–120°E) denotes the area used to define the average 700-hPa vorticity, 500-hPa vertical velocity, and 700-hPa specific humidity over southern China, used later in Figs. 7a,b.

The evolution of the western North Pacific subtropical high has also been analyzed. The subtropical high anomalously extends to the west before the EH onsets, and then retreats to the climatological position since the EH onsets (not show). The result suggests that the anticyclonic component of the anticyclonic–cyclonic pair is related to the western extension of the subtropical high over western North Pacific, and the anticyclonic–cyclonic pair shows some unique features. We have examined the evolution of circulation anomalies associated with EH events in southern China case by case (not shown). It turns out that 84% of all events are associated with an anomalous anticyclone over southern China, and 58% of these cases feature an anticyclonic–cyclonic pair, indicating the important role of the anticyclonic–cyclonic pattern in the EH occurrence.

There is anomalous vertical motion associated with the anticyclonic–cyclonic pattern. Before the EH onsets, subsidence anomaly associated with the anticyclonic component gradually strengthens and significant subsidence anomaly occurs over southern China from −2d to 0d (Figs. 4b–d). The stronger subsidence would increase the adiabatic heating, reduce the cloud cover, and increase the solar radiation at surface, favoring the occurrence of EH. Composite analyses demonstrate that there is significantly greater adiabatic heating, less total cloud cover, and more net shortwave radiative flux at surface over southern China on EH onset days (not shown), verifying the abovementioned heating processes. After the EH onsets, the subsidence anomaly weakens while the ascendance anomaly associated with the cyclonic component gradually develops and moves toward southern China (Figs. 4e–f), inhibiting the occurrence of EH.

From the perspective of lower-tropospheric horizontal winds, southerly and southwesterly anomalies associated with the anticyclone occur over southern China and the area to the north at −3d (Fig. 4a) and then move to the northwest of southern China from −2d to 0d (Figs. 4b–d). Accompanied by the approaching of anomalous cyclone, an obvious northeasterly anomaly appears over southern China from 1d to 2d (Figs. 4e,f). Overlapping with the climatological winds (Fig. 3), the southwesterly (northeasterly) anomaly to the north of southern China would increase (decrease) the northward transport of water vapor from southern China to the northern regions and result in lower (higher) humidity over southern China. Consequently, the specific humidity over southern China significantly reduces before the EH onsets (Figs. 4c,d), and the drier air would favor more solar radiation at surface and thus further enhance the surface air temperature. On the contrary, the drier air condition weakens after the EH onsets (Figs. 4e,f) and so as the higher temperature. These results are consistent with Chen and Lu (2015): The averaged composite from −2d to 0d shows an anomalous anticyclonic–cyclonic pattern, and stronger subsidence and lower humidity over southern China.

Therefore, the anticyclonic–cyclonic pattern would influence EH occurrence through inducing anomalous vertical motion and humidity condition: an anomalous anticyclone and the associated subsidence and dry air condition would favor higher temperature through increasing solar radiation at surface, and the anomalous cyclone would have the opposite effect on temperature. The composite evolution of net shortwave radiative flux at the surface is consistent with the anticyclonic–cyclonic pattern: Before the EH onsets, there is a significant increase of local shortwave radiation associated with the anticyclonic anomaly, which first occurs over the western Pacific and then enhances and propagates northwestwards to southern China, and the positive shortwave radiative flux anomaly gradually diminishes after onsets with the approaching of the cyclonic anomaly (not shown). The increase of local net shortwave radiation contributes to the occurrence of EH in southern China. Figure 5 shows the evolution of the composite anomalies of observed Tmax. Significant higher temperature occurs over southern China at −1d and then grows to the strongest at 0d, coinciding with the onsets of EH events. The warming amplitudes are about 2°C over southern China at 0d and 1d, and then the amplitudes gradually weaken afterward.

Fig. 5.

Evolution of the composite Tmax anomalies for the EH onsets in southern China (a)–(f) from −3d to 2d [contours, °C; contour interval of 1°C; the solid (dashed) contours denote positive (negative) anomalies]. Yellow-shaded areas are statistically significant at the 95% confidence level according to the Student’s t test. The blue dots denote the stations in southern China.

Fig. 5.

Evolution of the composite Tmax anomalies for the EH onsets in southern China (a)–(f) from −3d to 2d [contours, °C; contour interval of 1°C; the solid (dashed) contours denote positive (negative) anomalies]. Yellow-shaded areas are statistically significant at the 95% confidence level according to the Student’s t test. The blue dots denote the stations in southern China.

It is shown that the circulation anomalies exhibit obvious oscillations during the EH events in southern China, which can be more clearly illustrated by the propagation of lower-tropospheric vorticity along the southeast–northwest trace crossing the anticyclonic–cyclonic anomaly (Fig. 6). The trace is from 0°, 145°E to 30°N, 110°E, as denoted by the blue line in Fig. 4c. Figure 6 shows that the vorticity obviously originates from the equatorial western Pacific and propagates northwestward. Systematic positive vorticity and negative vorticity rank alternately along the trace from the equator to around 30°N, with a horizontal scale of about 1000 km for each vorticity, forming a wave train which is also shown in Fig. 4. Southern China is located at about 23°–30°N, 119°–110°E. Negative vorticity strengthens and propagates toward southern China before EH onsets, and then the negative vorticity gradually weakens and positive vorticity dominates instead, with each one persisting for about 5 days over southern China (i.e., quasi-biweekly oscillation for a complete cycle). The strongest negative vorticity appears over southern China around the EH onsets, indicating that the warming associated with the anticyclone greatly contributes to the occurrence of EH events. It is well demonstrated that the northwestward-propagating quasi-biweekly oscillation from the tropical Pacific plays an important role in inducing the EH events in southern China.

Fig. 6.

Propagation of the composite 700-hPa vorticity anomaly (shading, 10−6 s−1) within 0°–30°N, 145°–110°E from 10 days before (−10d) to 10 days after (+10d) the EH onsets in southern China. Areas statistically significant at the 95% confidence level are marked by black dots. The trace is denoted by the blue line in Fig. 4c.

Fig. 6.

Propagation of the composite 700-hPa vorticity anomaly (shading, 10−6 s−1) within 0°–30°N, 145°–110°E from 10 days before (−10d) to 10 days after (+10d) the EH onsets in southern China. Areas statistically significant at the 95% confidence level are marked by black dots. The trace is denoted by the blue line in Fig. 4c.

To better illustrate the relationship between the oscillations of circulation and temperature anomalies over southern China, four meteorological variables are selected to represent the anticyclonic–cyclonic pattern. The four selected variables are the 700-hPa vorticity (vor700), 500-hPa vertical velocity (ome500), 700-hPa specific humidity (q700) averaged over 20°–28°N, 105°–120°E covering the center of anomalies over southern China, and the 700-hPa southwesterly (SW700) averaged along the trace from 25°N, 110°E to 35°N, 120°E, which lies to the northwest of southern China as denoted by the blue line in Fig. 4a. The relationships among the four selected variables and the Tmax (Fig. 7) are consistent with the results presented earlier: Before the EH onset, the approaching and strengthening of the anomalous anticyclonic component since −3d induces anomalous subsidence over southern China and southwesterly flow to the northwest, and the southwesterly flow leads to lower specific humidity over southern China, and all these meteorological conditions are favorable for the increase of surface air temperature. Conversely, after the EH onsets, the negative vor700 weakens and turns into the opposite phase on 2d, and the higher temperature weakens subsequently. All the four selected variables exhibit clear quasi-biweekly oscillations over southern China and contribute to the temperature oscillation during the EH events. Moreover, the evolution of circulation appears to lead the Tmax evolution by about 1 day: The strongest Tmax anomaly occurs from 0d to 1d, whereas vor700 and SW700 are strongest at −1d. The leading occurrence of atmospheric circulation compared to temperature anomalies suggests that the circulation anomalies would affect simultaneously the temperature tendency and then temperatures.

Fig. 7.

Evolution of the composite anomalies of Tmax averaged over southern China stations (solid lines, right y axis) and the variables representing the anticyclonic–cyclonic pattern (dashed lines, left y axis): (a) vor700 (10−6 s−1), (b) ome500 (10−4 hPa s−1), and (c) q700 (g kg−1; dashed line with dots) averaged over 20°–28°N, 105°–120°E, denoted by the blue dashed box in Fig. 4d and SW700 (m s−1; dashed line with crosses) averaged along the trace from 25°N, 110°E to 35°N, 120°E, denoted by the blue line in Fig. 4a. [Note that all values use the left y axis in (c).]

Fig. 7.

Evolution of the composite anomalies of Tmax averaged over southern China stations (solid lines, right y axis) and the variables representing the anticyclonic–cyclonic pattern (dashed lines, left y axis): (a) vor700 (10−6 s−1), (b) ome500 (10−4 hPa s−1), and (c) q700 (g kg−1; dashed line with dots) averaged over 20°–28°N, 105°–120°E, denoted by the blue dashed box in Fig. 4d and SW700 (m s−1; dashed line with crosses) averaged along the trace from 25°N, 110°E to 35°N, 120°E, denoted by the blue line in Fig. 4a. [Note that all values use the left y axis in (c).]

It should be mentioned that the EH in southern China could last for a long period, but the composite analysis shows that the temperature anomaly obviously decreases after +1d (Fig. 7) and thus exhibits a clear quasi-biweekly oscillation. This is because that a large proportion of the selected EH events are short-lived EH, with 19 events lasting for only one day, 20 lasting for two days, and 21 lasting for three days, accounting for 73% of the events together. Thus, the composite results mainly manifest the short-lived EH. However, there are also some long-lived EH events, which could last for more than one month (e.g., the EH events in 2003). The anomalies associated with long-lived EH would be different, which could be partially deduced from the temperature evolution. Detailed comparisons between the short- and long-lived EH events will be performed in the future.

4. Evolution of the circulation of quasi-biweekly oscillations associated with extreme heat events in southern China

The evolution of the original circulation anomalies for EH events in southern China show an obvious oscillation of vorticity over southern China with a periodicity of about 10 days, originating from the tropical western Pacific and propagating northwestward. Thus it is indicated that the EH in southern China may be influenced by the quasi-biweekly oscillations of tropical atmospheric circulation, which is investigated in detail in this section.

Before analyzing the oscillations of atmospheric circulation during EH events, we first examine the frequency spectrum of Tmax to identify the predominant periodicity of EH. Spectral analysis is performed for the Tmax series of each summer and then averaged over the 35 years from 1979 to 2013. Figure 8a shows the averaged spectrum and the significance level. The first significant periodicity peak occurs around 10 days, manifesting the signal of quasi-biweekly oscillation. Based on the frequency spectrum of Tmax, we defined 5–25-day oscillation as the quasi-biweekly oscillation. The average intraseasonal variance of Tmax for the 5–25-day component is 1.2 K2, accounting for 55% of the variance for the original series (2.2 K2). We further examined the frequency spectrum year by year, and found that 33 out of 35 years show periodicity peak of 5–25 days significant at a 95% level, and the highest variance proportion explained by the 5–25-day component reaches 84% in the summer of 1996. Therefore, the quasi-biweekly oscillation turns out to be an important component for the Tmax oscillation over southern China. The same spectral analysis is applied to the 700-hPa vorticity averaged over southern China (Fig. 8b) and has similar features: the first significant periodicity peak for vor700 also occurs around 10 days, manifesting the signal of quasi-biweekly oscillation, and the average intraseasonal variance of vor700 for the 5–25-day component (5.7 × 10−11 s−2) accounts for 55% of the variance for the original series (10.3 × 10−11 s−2). Thereby, it is suggested that the quasi-biweekly oscillation plays an important role in the EH over southern China.

Fig. 8.

The 35-yr mean frequency spectrum of the summertime (a) Tmax and (b) vor700 averaged over southern China during JA from 1979 to 2013 (solid line). The dashed line is the red noise spectrum, and the dotted line is the spectrum of 95% confidence level. The black stars mark the first significant peaks. (Note that the x axis is logarithmic for the period.)

Fig. 8.

The 35-yr mean frequency spectrum of the summertime (a) Tmax and (b) vor700 averaged over southern China during JA from 1979 to 2013 (solid line). The dashed line is the red noise spectrum, and the dotted line is the spectrum of 95% confidence level. The black stars mark the first significant peaks. (Note that the x axis is logarithmic for the period.)

To better illustrate the quasi-biweekly oscillations associated with EH, Lanczos filtering is applied to all the original series to extract the 5–25-day components for analysis. Figure 9 shows the evolution of the composite 5–25-day-filtered lower-tropospheric winds, specific humidity, and midtropospheric vertical velocity. It is impressive that the evolution of the filtered anomalies (Fig. 9) is quite similar to the original anomalies (Fig. 4) with comparable amplitudes, indicating the quasi-biweekly oscillation is critical for the occurrence of EH in southern China. As a result, the evolution of the composite 5–25-day-filtered Tmax (Fig. 10) also resembles the original anomalies (Fig. 5). The warming amplitudes at 0d and 1d are about 1.5°C, only approximately 0.5°C cooler than the original temperature anomalies. In addition, the propagation of lower-tropospheric vorticity along the northwestward trace from the tropical Pacific to southern China for the 5–25-day-filtered anomalies (Fig. 11) is also quite similar to the original anomalies (Fig. 6), confirming that the quasi-biweekly oscillation of tropical atmospheric circulation is vital for the occurrence of EH in southern China.

Fig. 9.

As in Fig. 4, but for the 5–25-day-filtered anomalies. Contour interval is 0.3 g kg−1 for q700 (shading) and 2 × 10−4 hPa s−1 for ome500 (contours).

Fig. 9.

As in Fig. 4, but for the 5–25-day-filtered anomalies. Contour interval is 0.3 g kg−1 for q700 (shading) and 2 × 10−4 hPa s−1 for ome500 (contours).

Fig. 10.

As in Fig. 5, but for the 5–25-day-filtered Tmax. Contour interval is 0.5°C.

Fig. 10.

As in Fig. 5, but for the 5–25-day-filtered Tmax. Contour interval is 0.5°C.

Fig. 11.

As in Fig. 6, but for the 5–25-day-filtered vor700.

Fig. 11.

As in Fig. 6, but for the 5–25-day-filtered vor700.

The four selected meteorological variables associated with the anticyclonic–cyclonic pattern in the previous section are also used to represent the circulation of quasi-biweekly oscillation, and the relationships between the evolution of these variables and the Tmax anomaly averaged over southern China stations are analyzed (Fig. 12). Comparing Figs. 12 and 7, the relationships among the four selected variables and the Tmax are consistent, indicating that the same mechanism is responsible for the increase of temperature for both the original anomalies and the quasi-biweekly oscillation. The 5–25-day-filtered anomalies around the EH onsets account for about 75% of the original anomalies. For instance, the amplitudes of temperature anomalies are 0.6°, 1.2°, and 1.2°C from −1d to 1d for the filtered series, accounting for 75%, 80%, and 75% of the original anomalies (0.8°, 1.5°, and 1.6°C, respectively); that is, the warming ratio related to quasi-biweekly oscillation could be about 80% on EH onsets.

Fig. 12.

As in Fig. 7, but for the 5–25-day-filtered anomalies.

Fig. 12.

As in Fig. 7, but for the 5–25-day-filtered anomalies.

5. Conclusions

Using homogenized daily temperature data for China and 6-hourly reanalysis data from ERA-Interim for the summers (from July to August) during 1979 to 2013, the evolution of circulation anomalies associated with EH events over southern China are investigated in this study. The results show that the southern China EH event is influenced by a unique anticyclonic–cyclonic anomaly in the lower troposphere. The anticyclonic and cyclonic components dominate southern China in succession and affect the temperature anomalies. Before the EH onsets, the anticyclonic component approaches southern China gradually, and the associated subsidence anomaly occurs over southern China, with a southerly (southwesterly) anomaly occurring to the west (north) of southern China. The stronger subsidence would increase the air temperature through inducing adiabatic heating, reducing the cloud cover, and enhancing the solar radiation at surface. Meanwhile, the anomalous southerly and southwesterly flow, overlapped with the climatological southwesterly summer monsoon, would enhance the northward water vapor transport and result in lower humidity over southern China. The drier air would further favor higher surface air temperature by allowing more incoming solar radiation. In contrast, after the EH onsets, the cyclonic component approaches southern China gradually and offsets the high temperature. The EH onsets occur later than the strongest anticyclone, and the evolution of circulation anomalies tend to lead the temperature anomalies by about 1 day.

It is noticeable that the anticyclonic–cyclonic anomaly originates from the tropical Pacific and propagates northwestward to southern China, and the vorticity over southern China exhibits an oscillating periodicity of about 10 days, suggesting the influence of the quasi-biweekly oscillations of tropical atmospheric circulation on the EH in southern China. In fact, the southern China summertime Tmax and lower-tropospheric vorticity both exhibit a predominant periodicity of 5–25 days, explaining for more than 50% of the intraseasonal variance on average. Thereby, the 5–25-day oscillation components of the original series are extracted to illustrate the quasi-biweekly oscillations associated with EH over southern China. The warming ratio associated with the 5–25-day-filtered anomalies could be about 80% on EH onsets, and the circulation and temperature evolution around the EH onsets are quite similar to the original anomalies, indicating that the physical processes are alike and confirming the importance of the quasi-biweekly oscillations of tropical atmospheric circulation on the EH in southern China.

The close relationship between the circulation of the quasi-biweekly oscillation and the occurrence of EH in southern China would be helpful for the medium-range forecast of high-temperature weather events, serving as a scientific reference for the government decision making on disaster prevention. However, it requires detailed investigations as to whether and how well the circulation of the quasi-biweekly oscillations can predict the temperature anomalies and EH in southern China. On the other hand, the role of the anticyclonic–cyclonic pattern in affecting the evolution of EH is presented in this paper, but the mechanism for the formation and maintenance of this pattern is still not clear, although it has been found that this pattern is related to quasi-biweekly oscillations. Also of concern is that the present study focuses on the composite circulation anomalies inducing EH (i.e., the general situation), whereas distinct circulation patterns might occur on EH days. For instance, some case studies have emphasized an anomalous anticyclone over southern China during EH events (Ji et al. 2005; Lin et al. 2005; Wang et al. 2016), rather than the anticyclonic–cyclonic pattern. Therefore, further work should be performed to compare the different circulation patterns, in order to comprehensively understand the cause of EH in southern China. Furthermore, it has been underlined that the changes of regional circulation patterns remarkably affect the regional trends of EH occurrence under global warming (Horton et al. 2015). Therefore, it would be expected that a reliable evaluation and projection on the change of EH under global warming would be enormously challenging for southern China, where the associated circulation system is complicated.

Acknowledgments

The authors greatly appreciate the comments from two anonymous reviewers. This work is jointly supported by the National Natural Science Foundation of China (Grant 41320104007), National Key Basic Research and Development Projects of China (2014CB953901), State Key Laboratory of Severe Weather opening project, and Sun Yat-sen University start-up funding project for young teachers.

REFERENCES

REFERENCES
CDC
,
2004
: Extreme heat: A prevention guide to promote your personal health and safety. Centers for Disease Control Rep., 8 pp. [Available online at www.ct.gov/agingservices/lib/agingservices/pdf/extremeheatguide.pdf.]
Chen
,
G.
, and
C.-H.
Sui
,
2010
:
Characteristics and origin of quasibiweekly oscillation over the western North Pacific during boreal summer
.
J. Geophys. Res.
,
115
,
D14113
, doi:.
Chen
,
J.
,
Z.
Wen
,
R.
Wu
,
Z.
Chen
, and
P.
Zhao
,
2015
:
Influences of northward propagating 25–90-day and quasi-biweekly oscillations on eastern China summer rainfall
.
Climate Dyn.
,
45
,
105
124
, doi:.
Chen
,
R.
, and
R.
Lu
,
2014a
:
Large-scale circulation anomalies associated with ‘tropical night’ weather in Beijing
.
Int. J. Climatol.
,
34
,
1980
1989
, doi:.
Chen
,
R.
, and
R.
Lu
,
2014b
:
Dry tropical nights and wet extreme heat in Beijing: Atypical configurations between high temperature and humidity
.
Mon. Wea. Rev.
,
142
,
1792
1802
, doi:.
Chen
,
R.
, and
R.
Lu
,
2015
:
Comparisons of the circulation anomalies associated with extreme heat weather in different regions in eastern China
.
J. Climate
,
28
,
5830
5844
, doi:.
Corobov
,
R.
,
S.
Sheridan
,
N.
Opopol
, and
K.
Ebi
,
2013
:
Heat-related mortality in Moldova: The summer of 2007
.
Int. J. Climatol.
,
33
,
2551
2560
, doi:.
Coumou
,
D.
, and
S.
Rahmstorf
,
2012
:
A decade of weather extremes
.
Nat. Climate Change
,
2
,
491
496
.
Dee
,
D. P.
, and Coauthors
,
2011
:
The ERA-Interim reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Roy. Meteor. Soc.
,
137
,
553
597
, doi:.
Díaz
,
J.
,
R.
García
,
F. V.
de Castro
,
E.
Hernandez
,
C.
Lopez
, and
A.
Otero
,
2002
:
Effects of extremely hot days on people older than 65 years in Seville (Spain) from 1986 to 1997
.
Int. J. Biometeor.
,
46
,
145
149
, doi:.
Gao
,
R.
,
L.
Wang
, and
G.
Gao
,
2008
:
The trend of variation in high temperature days during 1956–2006 in China
.
Adv. Climate Change Res.
,
4
,
177
181
.
Gershunov
,
A.
,
D. R.
Cayan
, and
S. F.
Iacobellis
,
2009
:
The great 2006 heat wave over California and Nevada: Signal of an increasing trend
.
J. Climate
,
22
,
6181
6203
, doi:.
Horton
,
D. E.
,
N. C.
Johnson
,
D.
Singh
,
D. L.
Swain
,
B.
Rajaratnam
, and
N. S.
Diffenbaugh
,
2015
:
Contribution of changes in atmospheric circulation patterns to extreme temperature trends
.
Nature
,
522
,
465
469
, doi:.
Huang
,
Z.
,
Y.
Xiong
, and
L.
Lin
,
2005
:
Analysis of synoptic characteristic and element related to torrid weather of 37°C or more in Guangzhou (in Chinese)
.
Meteorology
,
31
,
24
27
.
IPCC
,
2013
: Summary for policymakers. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 3–29.
Ji
,
Z.
,
G.
Lin
,
X.
Li
, and
Y.
Xiong
,
2005
:
High temperature anomalies in Guangdong in summer 2003 and its climatic background (in Chinese)
.
J. Trop. Meteor.
,
21
,
207
216
.
Jia
,
X.
, and
S.
Yang
,
2013
:
Impact of the quasi-biweekly oscillation over the western North Pacific on East Asian subtropical monsoon during early summer
.
J. Geophys. Res. Atmos.
,
118
,
4421
4434
, doi:.
Kikuchi
,
K.
, and
B.
Wang
,
2009
:
Global perspective of the quasi-biweekly oscillation
.
J. Climate
,
22
,
1340
1359
, doi:.
Kilbourne
,
E. M.
,
1997
: Heat waves and hot environments. The Public Health Consequences of Disasters, E. K. Noji, Ed., Oxford University Press, 245–269.
Krishnamurti
,
T. N.
, and
P.
Ardanuy
,
1980
:
The 10 to 20-day westward propagating mode and “breaks in the monsoons.”
Tellus
,
32
,
15
26
, 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:.
Li
,
C.
, and
Y.
Zhou
,
1995
:
On quasi-two-week (10–20 day) oscillation in the tropical atmosphere (in Chinese)
.
Sci. Atmos. Sin.
,
19
,
435
444
.
Li
,
C.
, and
W.
Zhou
,
2013
:
Modulation of western North Pacific tropical cyclone activities by the ISO. Part I: Genesis and intensity
.
J. Climate
,
26
,
2904
2918
, doi:.
Li
,
Z.
, and
Z.
Yan
,
2009
:
Homogenized daily mean/maximum/minimum temperature series for China from 1960–2008
.
Atmos. Oceanic Sci. Lett.
,
2
,
237
243
, doi:.
Li
,
Z.
,
L.
Cao
,
Y.
Zhu
, and
Z.
Yan
,
2016
:
Comparing two homogenized datasets of daily maximum/mean/minimum temperatures in China during 1960–2013
.
J. Meteor. Res.
,
30
,
53
66
, doi:.
Lin
,
J.
,
B.
Bi
, and
J.
He
,
2005
:
Physical mechanism responsible for western Pacific subtropical high variation and hot wave in southern China in July 2003 (in Chinese)
.
Chin. J. Atmos. Sci.
,
29
,
594
599
.
Loikith
,
P. C.
, and
A. J.
Broccoli
,
2012
:
Characteristics of observed atmospheric circulation patterns associated with temperature extremes over North America
.
J. Climate
,
25
,
7266
7281
, doi:.
Lu
,
R.
, and
R.
Chen
,
2016
:
A review of recent studies on extreme heat in China
.
Atmos. Oceanic Sci. Lett.
,
9
,
114
121
, doi:.
Luber
,
G.
, and
M.
McGeehin
,
2008
:
Climate change and extreme heat events
.
Amer. J. Prev. Med.
,
35
,
429
435
, doi:.
Maheras
,
P.
,
H.
Flocas
,
K.
Tolika
,
C.
Anagnostopoulou
, and
M.
Vafiadis
,
2006
:
Circulation types and extreme temperature changes in Greece
.
Climate Res.
,
30
,
161
174
, 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:.
Peng
,
S.
, and Coauthors
,
2004
:
Rice yields decline with higher night temperature from global warming
.
Proc. Natl. Acad. Sci. USA
,
101
,
9971
9975
, doi:.
Valor
,
E.
,
V.
Meneu
, and
V.
Caselles
,
2001
:
Daily air temperature and electricity load in Spain
.
J. Appl. Meteor.
,
40
,
1413
1421
, doi:.
Wang
,
W.
,
W.
Zhou
,
X.
Li
,
W.
Xin
, and
D.
Wang
,
2016
:
Synoptic-scale characteristics and atmospheric controls of summer heat waves in China
.
Climate Dyn.
,
46
,
2923
2941
, doi:.
Weaver
,
S. J.
,
A.
Kumar
, and
M.
Chen
,
2014
:
Recent increases in extreme temperature occurrence over land
.
Geophys. Res. Lett.
,
41
,
4669
4675
, doi:.
Wei
,
J.
, and
J. H.
Sun
,
2007
:
The analysis of summer heat wave and sultry weather in North China (in Chinese)
.
Climatic Environ. Res.
,
12
,
453
463
.
Wei
,
J.
,
H.
Yang
, and
S.
Sun
,
2004
:
Relationship between the anomaly longitudinal position of subtropical high in the western Pacific and severe hot weather in North China in summer (in Chinese)
.
Acta Meteor. Sin.
,
62
,
308
316
.
Wei
,
K.
, and
W.
Chen
,
2009
:
Climatology and trend of high temperature extremes across China in summer
.
Atmos. Oceanic Sci. Lett.
,
2
,
153
158
, doi:.
Wen
,
M.
, and
R.
Zhang
,
2005
:
Possible maintaining mechanism of climatological atmospheric quasi-biweekly oscillation around Sumatra
.
Chin. Sci. Bull.
,
50
,
1054
1056
.
Wilks
,
D. S.
,
2006
: Hypothesis testing. Statistical Methods in the Atmospheric Sciences, J. Hele, Ed., Academic Press, 131–177.
WMO
,
2013
: The global climate 2001–2010: A decade of climate extremes. WMO Tech. Rep. WMO-1103, 119 pp.
Xu
,
K.
, and
R.
Lu
,
2015
:
Break of the western North Pacific summer monsoon in early August
.
J. Climate
,
28
,
3420
3434
, doi:.
Yang
,
J.
,
Q.
Bao
,
B.
Wang
,
D.
Gong
,
H.
He
, and
M.
Gao
,
2014
:
Distinct quasi-biweekly features of the subtropical East Asian monsoon during early and late summers
.
Climate Dyn.
,
42
,
1469
1486
, doi:.
Zhang
,
X.
, and
Y.
Wang
,
2002
:
Relationship between power consumption and meteorological conditions during the summer in Beijing city and its forecast (in Chinese)
.
Meteor. Mon.
,
28
,
17
21
.