Air pollution is substantially modulated by meteorological conditions, and especially their diurnal variations may play a key role in air quality evolution. However, the behaviors of temperature diurnal cycles along with the associated atmospheric condition and their effects on air quality in China remain poorly understood. Here, for the first time, we examine the diurnal cycles of day-to-day temperature change and reveal their impacts on winter air quality forecasting in mountain-basin areas. Three different diurnal cycles of the preceding day-to-day temperature change are identified and exhibit notably distinct effects on the day-to-day changes in atmospheric-dispersion conditions and air quality. The diurnal cycle with increasing temperature obviously enhances the atmospheric stability in the lower troposphere and suppresses the development of the planetary boundary layer, thus deteriorating the air quality on the following day. By contrast, the diurnal cycle with decreasing temperature in the morning is accompanied by a worse dispersion condition with more stable atmosphere stratification and weaker surface wind speed, thereby substantially worsening the air quality. Conversely, the diurnal cycle with decreasing temperature in the afternoon seems to improve air quality on the following day by enhancing the atmospheric-dispersion conditions on the following day. The findings reported here are critical to improve the understanding of air pollution in mountain-basin areas and exhibit promising potential for air quality forecasting.
Air pollution is not only affected by anthropogenic emissions (Streets et al., 2001; Zhang et al., 2009; Kelly and Zhu, 2016) but also controlled by atmospheric-dispersion conditions (Wei et al., 2011; Li et al., 2015; Ye et al., 2016; Zhang et al., 2020). Stagnant meteorological conditions significantly contribute to the formation and maintenance of heavy air pollution as they play important roles in regulating the increment of air pollutant concentrations (Deng et al., 2014; Bei et al., 2016; Zhang et al., 2016; Wang et al., 2018). It is noted that atmospheric-dispersion capacity is substantially modulated by synoptic patterns, and hence the evolutions of large-scale synoptic systems can lead to the improvement or deterioration of air quality (Yarnal, 1993; Miao et al., 2017; Ning et al., 2019, 2020; Dong et al., 2020). In China, high anthropogenic emissions from coal-fired heating (Xiao et al., 2015), frequent temperature inversion (Xu et al., 2019; Feng et al., 2020; Guo et al., 2020), and shallow planetary boundary layer (PBL) structure (Li et al., 2017; Miao et al., 2018; Su et al., 2020) result in frequent occurrence of heavy-air-pollution events in winter. These factors highlight the significance of further revealing the physical mechanism of atmospheric-dispersion evolutions.
The behaviors of diurnal cycles of atmospheric-dispersion conditions and their effects on air quality remain poorly understood, although air pollution significantly modulated by atmospheric-dispersion conditions has been well demonstrated. For instance, as a typical synoptic process occurring in winter in China, the cooling process could cause rapid changes in meteorological and environmental conditions. Cooling processes induce significant day-to-day temperature variations and thus result in substantial changes in air quality (Hu et al., 2018; Ning et al., 2018b; Kang et al., 2019). Many previous studies revealed that cooling processes could remove air pollutants by invading lots of cold fresh airflows (Kalkstein and Corrigan, 1986; Gimson, 1994; Hu et al., 2018; Ning et al., 2018b) or exacerbate air pollution by transporting air pollutants (Fu et al., 2008; Ding et al., 2013; Luo et al., 2018; Kang et al., 2019). Nevertheless, most of these studies did not consider the influences of diurnal cycles of cooling processes on air quality. Are the influences of cooling processes occurring during daytime and nighttime on air quality similar or different? There are two key questions. The first one is what the behaviors of the diurnal cycles of atmospheric-dispersion conditions are, and the second one is how these behaviors affect air quality, especially how the diurnal cycles of day-to-day temperature change affect air pollution. Exploring the answers to these questions is critical for fully understanding winter air pollution and is also urgently needed for improving air quality forecasting in China.
Map of the Sichuan Basin (SCB) in Southwest China.
The Sichuan Basin (SCB) is one of the areas with the heaviest air pollution in China (Zhang et al., 2012; Ning et al., 2018a). With a high population density in the SCB, its heavy air pollution thus poses serious health hazards to residents (Liao et al., 2017; Qiu et al., 2018; Zhu et al., 2018; Zhao et al., 2018). It is noted that the SCB has a unique topography, with Qinling-Daba and Wu mountains in the north and east and with Qinghai–Tibet Plateau and Yunnan–Guizhou Plateau in the west and south of the basin (Fig. 1). The combination of this complex topography results in unique weather and climate, like the southwest vortex, the Huaxi autumn rain season, etc. The southwest vortex, southern branch, and Qinghai–Tibet high pressure are often formed over the SCB or the Tibetan Plateau, and the complex synoptic systems significantly affect atmospheric-dispersion conditions (Wang et al., 1993; Wei et al., 2014; Feng et al., 2016; Yu et al., 2016; Ning et al., 2019, 2020). Therefore, both the physical mechanism of atmospheric conditions' effects on air pollution and the air quality forecasting in the SCB are more complicated than these in the eastern plain regions of China (Chen and Xie, 2012; Wang et al., 2014; Ning et al., 2019; Zhang et al., 2019). To better understand the formation mechanism of air pollution and improve air quality forecasting in mountain-basin areas, the effects of diurnal variations in atmospheric-dispersion conditions on winter air quality in the SCB call for urgent examinations.
The scientific goals of this study are to first cluster the typical diurnal cycles of day-to-day temperature change in the SCB during wintertime and then to examine the mechanisms underlying the effects of the identified typical diurnal cycles on the following day-to-day air quality changes. We expect our study to help in better understanding the physical mechanism of air quality evolutions and improve air pollution forecasting in mountain-basin areas. The rest of this paper is organized as below. Data and methodology are introduced in Sect. 2. Section 3 describes the results of our study. Discussion related to our findings is given in Sect. 4. Our conclusions are summarized in Sect. 5.
Hourly concentrations of surface PM
Hourly winter surface temperature data observed at 105 meteorological
stations in the SCB (Fig. 1) from December 2006 to February 2020 are
also collected. Their regional averages are used to determine the diurnal
cycles of day-to-day temperature change. Additionally, daily mean surface
wind speed in the 18 cities of the SCB is also collected. To explore the
thermodynamic structure of the lower troposphere, daily potential
temperature profiles at 20:00 Beijing time (BJT; UTC
To obtain winter lower-troposphere stability and reveal the possible
mechanism of the formation of diurnal cycles of day-to-day temperature
change, 700 hPa temperature, air pressure and air temperature at 2 m above
the ground, total cloud cover,
The lower-troposphere stability (LTS) is defined as the differences in potential temperature between 700 hPa and the surface (Slingo, 1987). LTS can describe the thermal state of the lower troposphere and thus can evaluate the vertical mixing of air pollutants in the lower troposphere (Guo et al., 2016a, b). A larger LTS indicates a stronger stability in the lower troposphere and a weaker vertical mixing of air pollutants.
Changes in Calinski–Harabasz values with different numbers of identified clusters.
The day-to-day temperature change for each hour of a given day is defined by
the hourly temperature differences between two neighboring days
(Karl et al., 1995):
Three identified diurnal cycles of day-to-day temperature
change based on the
To investigate the effects of diurnal cycles of day-to-day temperature
change on air quality, we also calculate the day-to-day changes in air
pollutant concentrations and atmospheric-dispersion conditions following
the temperature change within 1 d. The following day-to-day changes in
air pollutant concentrations (or atmospheric-dispersion conditions) are
defined by the differences in air pollutant concentrations (or
meteorological conditions) between the next day and the current day:
The nighttime
Vertical west–east cross-sections of the day-to-day
changes in wind vectors (synthesized by
Clustering methods divide the objects into specific groups, with the goal
that all data objects assigned to the same cluster have common
characteristics, while different clusters have distinct characteristics
(Darby, 2005). The clustering methods have been widely used in
climate and environmental research (Bardossy et al., 1995; Cavazos,
2000; Luo and Lau, 2017; Bernier et al., 2019). In this study, the regional
average values of day-to-day temperature change in the SCB and the
Spatial distribution of the day-to-day changes in surface
PM
The selection of an optimal number of clusters is illustrated in Fig. 2, which shows Calinski–Harabasz values associated with the numbers of clusters ranging from 2 to 10. The Calinski–Harabasz value with three clusters reaches the highest value, indicating that the optimal number of clustering is three. Three dominant diurnal cycles of day-to-day temperature change are therefore identified in the SCB. The three typical diurnal cycles of day-to-day temperature change are depicted in Fig. 3. The days for Cluster 1, Cluster 2, and Cluster 3 are 455 (accounting for 36.9 % of total days), 413 (33.5 %), and 365 d (29.6 %), respectively, indicating that the differences in the occurrence frequency among the three diurnal cycles are not noticeable. However, the diurnal cycles of day-to-day temperature change among the three clusters exhibit obvious differences.
Day-to-day changes in the PT vertical profiles at 20:00 BJT following the three identified diurnal cycles within 1 d at four
sounding stations. Chengdu
In particular, in Cluster 1 (diurnal cycle with increasing temperature), all the
temperature changes are positive for 24 h throughout all days, indicating
that temperature increases during the past 24 h and exhibits a maximum
change approaching 1.5
To reveal the underlying mechanism of the formation of the above three diurnal cycles of day-to-day temperature change, we also investigate the nighttime and daytime day-to-day changes in total cloud cover that could play a key role in temperature changes by modulating atmospheric radiations. Figure 4 shows the nighttime and daytime day-to-day changes in total cloud cover associated with the three diurnal cycles. Corresponding to the diurnal cycle with increasing temperature (Cluster 1), the total cloud exhibits a slight increase in the eastern SCB and a decrease in the western SCB (Fig. 4a). The dipole spatial distribution could result in a weak change in the regional average temperature across the SCB during nighttime (Fig. 3). During daytime, negative changes in total cloud cover are observed in the entire basin (Fig. 4d) that are beneficial to the obvious increase in temperature in the afternoon (Fig. 3). In contrast, both the nighttime and daytime changes in total cloud cover are positive in the entire basin for Cluster 2 (Fig. 4b and e), which could induce the increasing temperature during nighttime and decreasing temperature during afternoon (Fig. 3). Corresponding to the diurnal cycle with decreasing temperature in the morning (Cluster 3), obvious decreases in the total cloud cover are observed in the entire basin during nighttime (Fig. 4c) that are beneficial to the temperature decrease.
Moreover, the SCB is located in the eastern Tibetan Plateau, and the complex
topography could play a key role in modulating the temperature changes
over the SCB (Ning et al., 2018b, 2019). Therefore, the vertical
west–east cross-sections of the day-to-day changes in wind vectors
(synthesized by
Heavy air pollution during winter in the SCB is mainly caused by high
concentrations of particulate matter (PM
In contrast, negative changes in PM
The contributions of gaseous pollutants in the SCB to winter air pollution are
also very important as the SCB has a large number of motor vehicles and
industries (Ning et al., 2018a). Therefore, the following
day-to-day changes in three major gaseous (SO
Figures 6 and S1 collectively indicate that the air quality in the SCB corresponding to Cluster 1 and Cluster 3 will deteriorate on the following day, while the air quality corresponding to Cluster 2 will improve. These results suggest that the modulations of the following day-to-day changes in winter air quality by diurnal cycles of day-to-day temperature change are obvious and important. Thus, the diurnal cycles of day-to-day temperature change exhibit promising potential for winter air quality forecasting on the following day in the SCB.
To reveal the potential influence mechanism of the diurnal cycles of day-to-day temperature change on the following day-to-day changes in air quality, the atmospheric-dispersion conditions corresponding to the three identified diurnal cycles are investigated. Firstly, the following day-to-day changes in PT vertical profiles at four sounding stations in the SCB (Fig. 7) are examined to explore the thermodynamic structure in the lower troposphere. Then, the following day-to-day changes in the three meteorological parameters related to atmospheric-dispersion conditions, including LTS (Fig. 8a–c), PBLH (Fig. 8d–f), and WS (Fig. 8g–i), are also investigated to evaluate the evolutions of atmospheric-dispersion capacity.
Spatial distribution of the day-to-day changes in LTS
Under the diurnal cycle with increasing temperature (Cluster 1), three sounding stations (Yibin, Dazhou, and Chongqing) experience increases in PT between 950 and 800 hPa on the following day (Fig. 7d,g, j). In Chengdu, decreased PT is observed below 900 hPa, while increased PT appears between 900 and 750 hPa (Fig. 7a). All the PT profiles over the four sounding stations show higher temperature change in the higher level (800–850 hPa) than the lower level (900–950 hPa), which could enhance the atmospheric stability in the lower troposphere. As shown in Fig. 8a, increased LTS is observed in most of the cities in the SCB, indicating that the atmospheric stratification in the lower troposphere becomes more stable. The stable atmospheric stratification inhibits the vertical mixing of the atmosphere and suppresses the development of the PBL (Karppinen et al., 2001; Bei et al., 2016). As shown in Fig. 8d, obviously decreased PBLH is observed in all 18 cities of the SCB.
Additionally, we also analyzed the following day-to-day changes in surface wind speed as the wind speed can represent the horizontal dispersion capacity of air pollutants (Lu et al., 2012; Deng et al., 2014). No noticeable decreases in wind speed appear in the SCB (Fig. 8g). These results suggest that the diurnal cycle with increasing temperature (Cluster 1) enhances atmospheric stability in the lower troposphere, which can weaken the vertical exchange of airflow and then suppress the development of the PBL, resulting in a small dispersion space of air pollutants and poor air quality in the SCB on the following day.
Compared with Cluster 1, an opposite vertical structure of PT changes (Fig. 7b, e, h, k) is observed for the diurnal cycle with decreasing temperature in the afternoon (Cluster 2), which could weaken the atmospheric stability in the lower troposphere. As shown in Fig. 8b, negative changes in LTS appear in all parts of the SCB, enhancing the vertical exchange of airflow and facilitating the development of the PBL. As a result, increased PBLH is observed in all parts of the SCB (Fig. 8e), and the regional average increment is up to 93.0 m. At the same time, the weakened atmospheric stability in the lower troposphere is also conducive to the development of surface wind speed. As shown in Fig. 8h, the surface wind speed in the entire SCB is strengthened obviously, indicating that the horizontal dispersion capacity of air pollutants is also improved. These results suggest that the diurnal cycle with decreasing temperature in the afternoon weakens atmospheric stability in the lower troposphere and creates good vertical mixing of airflow, which can promote the development of the PBL and surface wind speed, facilitating the improvement of air quality on the following day.
For the Cluster 3, the PT changes are not noticeable below 850 hPa over the four sounding stations. As shown in Fig. 7c, f, i, l, decreased PT is observed between 850 and 700 hPa, while obviously increased PT appears above 700 hPa. This vertical structure of PT changes suggests that the atmospheric stability is enhanced above the PBL over the SCB, which is demonstrated to play a key role in the formation of winter heavy-air-pollution events in the basin (Ning et al., 2018b, 2019). As shown in Fig. 8c, increased LTS appears in the entire SCB, and the increments of LTS are obviously larger than those for Cluster 1 (Fig. 8a), inhibiting the vertical mixing of atmosphere and suppressing the development of the PBL. As a result, decreased PBLH is observed in all parts of the SCB. Compared with Cluster 1, the enhanced atmospheric stability above the PBL also suppresses the development of surface wind speed. As shown in Fig. 8i, all parts of the SCB experience decreases in surface wind speed, weakening the horizontal dispersion capacity of air pollutants. These results suggest that both the vertical and horizontal dispersion capacity of air pollutants corresponding to Cluster 3 are worse than those corresponding to Cluster 1. The differences in the atmospheric-dispersion conditions between Cluster 3 and Cluster 1 can explain well that the air quality deterioration is more serious for Cluster 3 than Cluster 1 (Figs. 6 and S1).
It is worth noting that the following day-to-day air quality changes between Cluster 2 and Cluster 3 in mountain-basin areas are opposite, even though both of the two diurnal cycles are associated with cooling processes. In the cases of the cooling process mainly occurring in the afternoon (Cluster 2), the atmospheric-dispersion conditions are obviously improved, resulting in air quality improvement on the following day. In contrast, the atmospheric-dispersion conditions are obviously inhibited when the cooling process mainly appears in the morning (Cluster 3), resulting in air quality deterioration on the following day. These findings could improve our understanding of the effects of cooling processes on air quality (Kalkstein and Corrigan, 1986; Gimson, 1994; Hu et al., 2018; Ning et al., 2018b; Kang et al., 2019) and suggest that comprehensive investigations for the effects of diurnal cycles of atmospheric-dispersion conditions on air quality are urgently needed in the future to fully understand the physical mechanism of air quality evolutions.
Additionally, both Cluster 1 and Cluster 3 are associated with weakened atmospheric-dispersion conditions and lead to air quality deterioration on the following day. However, obvious differences in PT vertical profiles (Fig. 7) between Cluster 1 and Cluster 3 are observed. Especially for Cluster 3, decreased PT is observed between 850 and 700 hPa, while obviously increased PT appears above 700 hPa (Fig. 7c, f, i, l). This special vertical structure of PT is closely related to the foehn that is formed under the synergistic effects of cooling processes and the Tibetan Plateau (Ning et al., 2019), indicating that a stable layer exits above the PBL and acts as a lid covering the PBL (Ning et al., 2018b, 2019). The vertical structure of PT is demonstrated to play key roles in the formation of winter heavy-air-pollution events in mountain-basin areas by inhibiting the development of secondary circulation and the PBL (Ning et al., 2018b, 2019). These features suggest that the physical processes related to air pollution are more complex in mountain-basin areas than in the areas with flat terrain and urgently need to be further explored in the future.
Our study highlights that the following day-to-day air quality changes in mountain-basin areas are notably affected by the diurnal cycles of day-to-day temperature changes. We find that the identified diurnal cycles of day-to-day temperature variation in our study can explain well the evolutions of atmospheric-dispersion conditions and air quality on the following day and thus could be useful for air quality forecasting in mountain-basin areas. Currently, numerical models (including the WRF-Chem model and CMAQ model) (Grell et al., 2005; Byun and Ching, 1999) and statistical models (including statistical analysis, machine learning, the hybrid linear–nonlinear method, etc.) (Huang, 1992; Chelani and Devotta, 2006; Borse, 2020) are the two typical methods that have been widely used to forecast air quality by combining weather conditions and emission sources (Gidhagen et al., 2005). In the future, our findings should therefore be combined with numerical models or statistical models to improve air quality forecasting in mountain-basin areas.
Taking the SCB as an example, this study is the first examination of the behaviors of diurnal cycles of day-to-day temperature change using hourly temperature observations and their effects on the following day-to-day air quality changes in mountain-basin areas. Three diurnal cycles of day-to-day temperature change are identified, which notably affect the following day-to-day air quality changes. Among them, two diurnal cycles (i.e., Clusters 1 and 3) inhibit atmospheric-dispersion conditions by enhancing atmospheric stability, suppressing the PBL, and weakening surface wind speed, thus leading to air quality deterioration on the following day.
Compared with the diurnal cycle with increasing temperature (i.e., Cluster 1), the atmospheric-dispersion conditions are worse for the diurnal cycle with decreasing temperature in the morning (i.e., Cluster 3) and cause more serious deterioration of air quality. In contrast, the atmospheric-dispersion condition with weakened atmospheric stability, a deepened PBL, and enhanced surface wind speed is obviously improved for this type of diurnal cycle with decreasing temperature in the afternoon (i.e., Cluster 2), which improves the air quality on the following day. These results suggest that the identified diurnal cycles can explain well the evolutions of atmospheric-dispersion conditions and air quality on the following day. Our findings exhibit promising potential for air quality forecasting in mountain-basin areas.
The hourly air quality data were collected from the Ministry of Ecology and
Environment of the People's Republic of China (
The supplement related to this article is available online at:
DK performed data analysis, prepared the figures, and wrote original draft with contributions from all co-authors. GN designed the research and wrote the manuscript. SW, ML, XN, and MM provided interpretation and editing of the manuscript. JC performed data analysis and provided useful comments.
The authors declare that they have no conflict of interest.
Publisher’s note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This work was supported by the National Natural Science Foundation of China (grant nos. 91644226, 41830648, 41871029, and 41771453), the Major Scientific and Technological Projects in Sichuan Province (grant no. 2018SZDZX0023), the Applied Basic Research project of the Sichuan Science and Technology Department (grant no. 2020YJ0425), the Technology Innovation Research and Development project of the Chengdu Science and Technology Department (grant no. 2018-YF05-00219-SN), and the National Major Projects on High-Resolution Earth Observation Systems (grant no. 21-Y20B01-9001-19/22). The appointment of Ming Luo at Sun Yat-sen University is partially supported by the Pearl River Talent Recruitment Program of Guangdong Province, China (grant no. 2017GC010634). We would like to thank the following departments for the provided data: the Ministry of Ecology and Environment of the People's Republic of China, the China Meteorological Administration, and the European Centre for Medium-Range Weather Forecasts. The authors are thankful to the anonymous reviewers, who provided valuable comments and suggestions.
This research has been supported by the National Natural Science Foundation of China (grant nos. 91644226, 41830648, 41871029, and 41771453), the Major Scientific and Technological Projects in Sichuan Province (grant no. 2018SZDZX0023), the Applied Basic Research Project of the Sichuan Science and Technology Department (grant no. 2020YJ0425), the Technology Innovation Research and Development project of the Chengdu Science and Technology Department (grant no. 2018-YF05-00219-SN), and the National Major Projects on High-Resolution Earth Observation Systems (grant no. 21-Y20B01-9001-19/22). The appointment of Ming Luo at Sun Yat-sen University is partially supported by the Pearl River Talent Recruitment Program of Guangdong Province, China (grant no. 2017GC010634).
This paper was edited by Aijun Ding and reviewed by three anonymous referees.