An Air Stagnation Index to Qualify Extreme Haze Events in Northern China
1. Introduction
Northern China (NC; Fig. 1) has been experiencing severe air pollution during recent years because of high quantities of emissions accompanied by rapid economic development and urbanization (Hu et al. 2014; IPCC 2013; Mijling et al. 2013; Tie et al. 2009; Yuxuan Wang et al. 2013; Zhang et al. 2012). Aerosol concentrations over NC are highly sensitive to meteorological parameters and show strong variabilities at daily (Quan et al. 2014; Guo et al. 2017), monthly (Zhang et al. 2012; Guo et al. 2017), and annual time scales (Mu and Liao 2014; Feng et al. 2016) since the meteorological factors, including wind speed, planetary boundary layer height (PBLH), and precipitation, can modulate aerosol dilution, transport, deposition processes, and even chemical reactions (Mu and Liao 2014; Seinfeld and Pandis 2006; Zhao et al. 2012). Hence, meteorology plays an important role in the occurrence of extreme haze events, characterized by extremely high concentrations of particulate matter smaller than 2.5 μm in diameter (PM2.5) (Quan et al. 2011; Zhao et al. 2013; R. Zhang et al. 2014; Guo et al. 2015; Wang et al. 2015; Gao et al. 2016). Under favorable meteorological conditions, PM2.5 can rapidly accumulate to very high levels. For example, Quan et al. (2014) reported a wave of haze events over the North China Plain (NCP) with a maximal observed PM2.5 concentration reaching 600 μg m−3 and a sustained haze event even persisted up to 6 days in Beijing during December 2012. They indicated that stagnation weather, generally characterized by low wind speeds and decreased PBLH, was the dominant factor in these haze events.
(a) Geographic distribution of the observation sites (diamonds) for the observed PM2.5 concentrations in NC (32°–47°N, 105°–132.5°E, including north China, northeast China, and their near areas in the manuscript). The polygon represents the BTH region (36°–42°N, 114°–118.5°E). (b) Time-averaged PM2.5 concentrations in 0.625° × 0.5° grids obtained via interpolation from 1 Apr 2013 to 31 Mar 2016. (c),(d) The geographic distributions of correlation coefficients between the monthly mean ASITS values and PM2.5 concentrations and ASINOAA values and PM2.5 concentrations, respectively, in NC from 1 Apr 2013 to 31 Mar 2016. The dots in (c) and (d) denote areas statistically passing the 99% confidence level. The observed PM2.5 data are from the China National EMC published by the MEP.
Citation: Journal of the Atmospheric Sciences 75, 10; 10.1175/JAS-D-17-0354.1
It is desired to establish a single “meteorological index” that could qualify the relationship between variations in meteorology and aerosol concentrations. It could be useful to understand the impacts of variations in weather and climate on the air quality, and to predict the occurrences of haze events. However, it is a very difficult issue to make a uniform index to trace the evolution of aerosol concentrations in any region and time scale because the interactions between aerosols and multimeteorological factors are extremely complex, even with many unclear mechanics, especially on the diurnal time scale (Z. Li et al. 2017). Hence, a meteorological index should have its focused spatial and temporal scope. These meteorological indices that were applied to NC in previous studies are presented in Table 1. In summary, two methods were used to establish these indices. The first and most common one is to establish an empirical–statistical model using PM2.5 concentrations and multivariable meteorological data by multilinear regression (MLR; Tai et al. 2010, 2012; Feng et al. 2016) or through other statistical methods, such as the summation of multinormalized atmospheric factors (Cai et al. 2017). An empirical–statistical meteorological index is efficient for studying variations at a specific place and time scale but it lacks a sufficient physical background. The application scopes of these indices are narrow. The second method is to establish an index that can physically characterize the air stagnation weather, which is strongly related to the variability in aerosol concentrations (Jacob and Winner 2009; Sun et al. 2013; Ye et al. 2016).
Meteorological indices used for qualifying variations in the aerosol concentrations or air quality from previous studies in NC.
The air stagnation index (ASI) was issued by the National Oceanic and Atmospheric Administration (NOAA) in the United States (https://www.ncdc.noaa.gov/societal-impacts/air-stagnation/overview). Here we refer to it as the ASINOAA, which was also referred to as NCDC ASI in Horton et al. (2012, 2014). ASINOAA is defined as the percentage of stagnation days in a period of time (e.g., 1 month). Stagnation days herein refer to days in which the daily sea level geostrophic wind speed is less than 8 m s−1 (or when the surface wind speeds are less than 4 m s−1, or the wind speed at 10 m is less than 3.2 m s−1), on which the daily wind speed at 500 hPa is less than 13 m s−1, and on which no precipitation occurs and the daily temperature is inverted within the PBL (using the layer between the surface and 850 hPa in practice). By the definition, the ASINOAA scheme mainly involves three key processes related to variations in aerosol concentrations: 1) local horizontal ventilation in the atmospheric column, that is, ventilation potency, 2) local vertical diffusion potency in the PBL, and 3) wet deposition potency. The ASINOAA has been widely used in many studies in many regions of the world (Leibensperger et al. 2008; Horton et al. 2012, 2014), including China (Huang et al. 2017). And it is reportedly able to characterize the seasonal and interannual variations in aerosol pollution (Mamtimin and Meixner 2011). However, it is worthy to note that the ASINOAA is semiempirical based on observations in the United States (Korshover 1976; Korshover and Angell 1982; Wang and Angell 1999), and thus it could not be suitable for capturing the haze event effectively in other regions. For example, the criterion of temperature inverse in the definition of ASINOAA, which only uses two pressure levels, could not be applicable in China because the temperature profiles in the low troposphere, including the PBL, seem extremely complex in China because of the strong positive feedback induced by high aerosol concentrations (Yuan Wang et al. 2013; Guo et al. 2016; Z. Li et al. 2017). According to Cai et al. (2017), only 28% of the severe haze events in Beijing were captured using the ASINOAA. In addition, the ASINOAA cannot reflect short-term variations in aerosol concentrations because of its representation of stagnation days as percentages in a given period of time.
To fill these gaps, considering the basic aerosol processes related to meteorological fields, we introduced a new ASI scheme (hereafter referred to as ASITS) that is applicable to daily and monthly variations in aerosol concentrations in NC. This paper is organized as follows. An illustration of the ASITS scheme, the meteorological data, and the observed PM2.5 concentrations are provided in section 2. In section 3, we demonstrate the geographic distribution and probability distribution function (PDF) of ASITS and the performance of ASITS in capturing the monthly variations in PM2.5 concentrations over NC, which are accompanied by a comparison of ASITS with the ASINOAA. Then we apply the ASITS to the daily PM2.5 concentration variations and prediction of extreme haze occurrence in NC. We also discuss the uncertainties in the relationship between ASITS and the PM2.5 concentrations induced by the scheme parameters of ASITS. Finally, in section 4, we draw our conclusions and discuss the potential applications of ASITS.
2. Data and methods
a. The scheme of ASITS
The ASITS scheme owns several favorable characteristics when compared with other previous indices: 1) the ASITS scheme does not include statistic–empirical parameters to represent the effects of meteorological variables and characterize the stagnation weather; 2) the ASITS scheme can be applied to the variations in PM2.5 concentrations on daily to monthly time scales (see section 3); and 3) ASITS is calculated with the same frequency of meteorological variables.
b. PM2.5 data
The observed surface PM2.5 concentrations in NC are published by the China Ministry of Environmental Protection Environmental Monitoring Center (MEP/EMC) (http://113.108.142.147:20035/emcpublish/). The datasets became available in January 2013 when the MEP began to release real-time PM2.5 data. This study used daily and monthly averaged PM2.5 concentrations from 1 April 2013 to 31 March 2016 at approximately 472 sites in NC (Fig. 1a; 32°–47°N, 105°–132.5°E). In addition, we obtained the gridded PM2.5 data by spatially interpolating the original MEP observation sites into the meteorological data grid (0.625° × 0.5°; see section 2c) to avoid a heterogeneous distribution of observation sites and missing values at many sites. Such an approach was used in previous studies that analyzed the influences of meteorological factors on aerosol concentrations (Tai et al. 2010, 2012; T. Zhang et al. 2016). It should be noted that interpolation errors exist in areas where the PM2.5 observation sites are scarce, including northern Inner Mongolia and some mountainous areas in southwestern NC. Thus, analysis hereinafter will avoid these areas. Figure 1b shows the geographic distribution of the time-averaged PM2.5 concentrations by interpolating the observed data over NC from 1 April 2013 to 31 March 2016. The PM2.5 concentrations are the highest in the central and southern parts of the Beijing–Tianjin–Hebei region (BTH), up to 110–120 μg m−3, which is corresponding to the previous studies (Li et al. 2016; Zou et al. 2017). Some other areas in NC including the western Shandong Province, central Henan Province, central Shanxi Province, and the vicinity of Harbin had the mean PM2.5 concentrations of 90–100, 90–100, 70–80, and 70–80 μg m−3, respectively, which were also reported in previous studies (Quan et al. 2011; Zhang et al. 2012).
c. Meteorological data
The meteorological data, including the geopotential height, PBLH, and precipitation, come from the Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2), dataset, which is available from the Goddard Earth Sciences Data and Information Services Center (GES DISC, https://disc.sci.gsfc.nasa.gov). The highest-resolution datasets (0.625° × 0.5°) in MERRA-2 are used to retrieve accurate interpolation results. The MERRA-2 meteorological datasets assimilated conventional and satellite observations detailed in Koster et al. (2016) and Gelaro et al. (2017), and were validated in Bosilovich et al. (2015). Compared with the National Centers of Environmental Prediction–Department of Energy AMIP-II reanalysis (NCEP-2) data, the normalized mean bias of geopotential height and wind speed in the low troposphere in NC (lower than 500 hPa) were 0.1% and 7%, respectively. Reichle et al. (2017) assessed the seasonal biases and time series correlation of the MERRA-2 dataset versus the Global Precipitation Climatology Project, version 2.2 (GPCPv2.2), precipitation dataset over 1980–2015 and showed that the MERRA-2 precipitation data have no apparent bias over NC. MERRA-2 data have been used in many studies on the structure and the dynamics of atmospheric circulation systems (Coy et al. 2016; Bosilovich et al. 2017) and the global atmospheric water balance and variability (Bosilovich et al. 2017). In this manuscript, the ASITS is calculated using the MERRA-2 data from 1 April 2013 to 31 March 2016. Besides the original gridded MERRA-2 data, we also use the sites’ data that are obtained via bilinear interpolation from the corresponding gridded data into the PM2.5 monitoring network of the MEP/EMC.
3. Results and analysis
a. The monthly variations of ASITS in NC
ASITS had a robust relationship with the monthly variations in PM2.5 concentrations in NC. Figure 1c is the spatial distribution of correlation coefficients between the monthly ASITS and PM2.5 concentrations in the same grid. It is shown that the monthly variations in ASITS are correlated positively with PM2.5. The correlation coefficients in 99.1% of the areas in NC passed the t test at a 99% confidence level. The correlation coefficients in BTH, Shanxi, and the northern Anhui Province were large and reached up to 0.9–1.0, 0.8–0.9, and 0.8–0.9, respectively. Since these regions were also the severe aerosol-polluted regions in NC (Fig. 1b), it is suggested that the high aerosol pollution levels in these areas should be largely attributed to the air stagnation conditions. In addition, we also show the geographic distribution of correlation coefficients between the monthly ASINOAA values and PM2.5 concentrations in NC (Fig. 1d). It is clear that the ASITS shows a more significant relationship with PM2.5 concentrations than the ASINOAA among the grid cells throughout NC, indicating that the ASITS is more suitable for analyzing monthly variations of PM2.5 concentrations in NC than the ASINOAA.
To conduct a more in-depth analysis of these relationships, the PDFs of the ASITS, ASINOAA, and the PM2.5 concentrations are shown in Fig. 2. The PDF of the monthly PM2.5 concentrations in NC was a smooth lognormal distribution (Fig. 2a), with the highest probability density at approximately 41 μg m−3. The lognormal PDF distribution of PM2.5 concentrations is similar to the results from previous studies (Šakalys et al. 2004; Ovadnevaitė et al. 2007; Lu 2002). We calculate the normalized mean bias [NMB;
The PDFs (circles) and their lognormal fitting curves (solid lines) for the monthly averaged (a) PM2.5, (b) ASITS, and (c) ASINOAA values in NC, obtained from the MEP PM2.5 observations from April 2013 to March 2016.
Citation: Journal of the Atmospheric Sciences 75, 10; 10.1175/JAS-D-17-0354.1
Considering that the BTH is the most polluted area in NC (Fig. 1b), we particularly compare the monthly variations in the regional-mean ASITS, ASINOAA, and PM2.5 in BTH from April 2013 to March 2016 (Fig. 3). It is clear that the ASITS scheme can capture the peaks and troughs of the monthly PM2.5 concentration series, with a significant correlation coefficient of 0.86, which passes the t test at 99% confidence level. The ASITS had a clear seasonal fluctuation with high values from December to February (1.6–2.4) and low values from May to September (0.3–1.0). The simultaneous monthly mean PM2.5 concentrations in BTH were 100–140 μg m−3 from December to February and 30–50 μg m−3 from May to September.
The monthly mean PM2.5 concentrations (black line; μg m−3), and ASITS (red line) and ASINOAA (blue line) values in BTH from April 2013 to March 2016. The correlation coefficients between the PM2.5 concentrations and the ASITS and ASINOAA are 0.86 and −0.37, respectively.
Citation: Journal of the Atmospheric Sciences 75, 10; 10.1175/JAS-D-17-0354.1
b. The daily variations of ASITS in NC
The ASITS scheme can also be applied to daily time-scale variations, which is highly needed in analyzing haze events. Figure 4a shows the geographic distribution of correlation coefficients between the daily ASITS and PM2.5 concentrations in NC from 1 April 2013 to 31 March 2016. Among almost all of the NC grids, the daily ASITS and PM2.5 concentrations showed a significant correlation. The correlation coefficients were 0.6–0.7 in BTH and 0.5–0.6 in the other areas with high PM2.5 concentrations (as shown in Fig. 1b). In addition, similar to the monthly analyses, we also calculate the PDFs of the daily PM2.5 concentrations and the daily ASITS (Figs. 4b and 4c, respectively). The PDFs of PM2.5 and ASITS were lognormal distributions, with similar NMBs of 16.9% and 9.7%, respectively. These results indicate that the ASITS can also reflect the daily PM2.5 variations in NC.
(a) Geographic distribution of correlation coefficients between the daily ASITS values and the PM2.5 concentrations in NC from 1 Apr 2013 to 31 Mar 2016. The dots denote areas statistically passing the 99% confidence level. The selected 10 aerosol-polluted cities with severe aerosol pollutions in NC are marked by black diamonds. (b),(c) The PDFs (circles) and their lognormal fitting profiles (black lines) for the daily PM2.5 concentrations and ASITS values, respectively, in NC.
Citation: Journal of the Atmospheric Sciences 75, 10; 10.1175/JAS-D-17-0354.1
Furthermore, 10 urban observation sites at aerosol-polluted cities, including Beijing, Tianjin, Shijiazhuang, Xi’an, Zhengzhou, Yinchuan, Ji’nan, Taiyuan, Harbin, and Hohhot (marked in Fig. 4a), are selected to examine the relationship between the local ASITS and PM2.5 concentrations. All those cities have been suffering heavy aerosol pollution reported by previous studies (Huang et al. 2011; Sun et al. 2013; Han et al. 2014; Mao et al. 2014; Wang et al. 2014; Q. Zhang et al. 2015; Gui et al. 2016; J. Li et al. 2017; Wang et al. 2017). The correlation coefficients between the daily ASITS and PM2.5 at the 10 polluted cities range from 0.4 to 0.7 and pass the t test at a 99% confidence level, further indicating that the ASITS can be used to assess the PM2.5 variations in polluted cities of NC.
To present the relationship between high ASITS and extreme haze, we define an extreme haze day (EHD) as the day with the 5% highest daily PM2.5 concentrations during the period of the analysis (from 1 April 2015 to 31 March 2016 in Fig. 5). The high ASITS values essentially companied the occurrence of EHDs for the 10 cities (Fig. 5). Taking Beijing as an example, EHDs occurred mainly from October 2015 to March 2016 with maximum daily mean PM2.5 concentrations of up to 200–500 μg m−3, which is similar to the findings in previous studies (Gui et al. 2016; J. Li et al. 2017). The high ASITS values in the period also generally companied the EHDs, with the peak values ranging from 1.8 to 7.0. Two other BTH cities, Tianjin and Shijiazhuang, had a similar EHD occurrence as Beijing during the same period. The local ASITS values in the two cities ranged from 3.0 to 10.0 and also effectively captured the aerosol pollution episodes. In Xi’an, Yinchuan, Ji’nan, Taiyuan, and Hohhot, the EHDs in the wintertime were always accompanied by high ASITS values. It should be noted that the relationship between the daily PM2.5 concentrations and ASITS may be interfered by other factors, including strong nonlocal aerosol mass transportation, emergency emission control measures, and accidental emissions from biomass burning in NC (C. Chen et al. 2015; Z. Chen et al. 2015; Yang et al. 2015; Long et al. 2016). For example, the ASITS scheme does not favorably track the two extreme events in Zhengzhou during January 2016, possibly because they were largely contributed by transportation from other places (Wang et al. 2017).
PM2.5 concentrations (μg m−3) on non-EHDs (black lines) and EHDs (blue lines), and the ASITS values (red lines) for the selected 10 aerosol-polluted cities from Fig. 4a between 1 April 2015 and 31 Mar 2016.
Citation: Journal of the Atmospheric Sciences 75, 10; 10.1175/JAS-D-17-0354.1
To present the close relationship between high ASITS and EHDs, Fig. 6 illustrates the ASITS PDFs for EHDs (daily PM2.5 concentration ≥142 μg m−3) and non-EHDs (daily PM2.5 concentration <142 μg m−3) in NC from 1 April 2013 to 31 March 2016. It is clear that the ASITS on EHDs and non-EHDs had substantially different values. Although these PDFs were lognormal (similar to those shown in Figs. 4b and 4c), the frequency of air stagnation weather in NC increased markedly on EHDs compared with non-EHDs, with a significant shift in ASITS toward higher values. The ASITS of highest probability density and mean ASITS values of EHDs were 1.3 and 2.5, respectively, greater than the values of non-EHDs (0.3 and 1.2, respectively). These results suggest that ASITS can qualify the occurrence of EHDs in NC.
ASITS PDFs on EHDs (red thick line; defined as days with daily PM2.5 concentrations >142 μg m−3) and non-EHDs (blue thick line; defined as days with daily PM2.5 concentrations ≤142 μg m−3) between 1 Apr 2013 and 31 Mar 2016 in NC. The red (blue) thin line represents the ASITS values of the most likely ASITS on EHDs (non-EHDs). The red (blue) dashed line represents the mean ASITS values on EHDs (non-EHDs).
Citation: Journal of the Atmospheric Sciences 75, 10; 10.1175/JAS-D-17-0354.1
c. Application of ASITS as a predictor of extreme haze events
Because air stagnation weather provides a favorable atmospheric environment for the accumulation of aerosol masses (Wang and Angell 1999), it is suggested that the effects of air stagnation weather on the PM2.5 concentrations are not immediate but rather precede variations in the PM2.5 concentrations. Hence, an ASI should have the ability to be a leading predictor of EHDs. To verify this assumption, Fig. 7 presents the lead–lag correlation coefficients between the PM2.5 concentrations and ASITS in the same 10 cities as shown in Fig. 5. It shows that the largest correlation coefficients between ASITS and the PM2.5 concentrations were not simultaneous. Rather, ASITS led PM2.5 concentrations by 1 day for all of the 10 cities, with the maximal significant lead–lag correlation coefficients ranging from 0.6 to 0.8. Significant lead correlation coefficients were also maintained for at least 4 days in each of the 10 cities as air stagnation days as defined by Wang and Angell (1999). Moreover, the correlation coefficients of ASITS leading PM2.5 concentrations by 1 day were significant in all of the areas in NC and were larger than the simultaneous correlation coefficients in approximately 98% of the grid cells in NC (Fig. 8). In the severely polluted regions mentioned in section 2b, such as the southern BTH, the lead correlation coefficients reached up to 0.6–0.8. These results indicate that ASITS can reflect the aerosol mass accumulation and diffusion processes in NC.
Lead–lag correlation coefficients between the PM2.5 concentrations and ASITS in the 10 aerosol-polluted cities in Figs. 4a and 5 between 1 Apr 2015 and 31 Mar 2016. The positive (negative) x-axis values denote the ASITS lead corresponding to the PM2.5 concentrations ranging from −5 to 5 days. In particular, a value of 0 on the x axis denotes a simultaneous correlation. The dashed curves denote the t test at a 99% confidence level.
Citation: Journal of the Atmospheric Sciences 75, 10; 10.1175/JAS-D-17-0354.1
As in Fig. 4a, but for the leading correlation coefficients between ASITS and the PM2.5 concentrations at a lead of 1 day. The hatched areas denote the areas in which the lead correlation coefficients are greater than the simultaneous correlation coefficients shown in Fig. 4a.
Citation: Journal of the Atmospheric Sciences 75, 10; 10.1175/JAS-D-17-0354.1
Lower limit of daily PM2.5 concentrations on EHDs (PM2.5-EHD), corresponding lower limits of ASITS on the previous day (ASITS-PD) calculated using data from 1 Apr 2013 to 31 Mar 2014, and prediction accuracies for EHDs using ASITS from 1 Apr 2015 to 31 Mar 2016 for the 10 cities in NC.
Taking Beijing for example, an EHD is defined by a daily PM2.5 concentration ≥217.4 μg m−3. Correspondingly, the previous day’s ASITS value of an EHD would be greater than 1.9. Using such predictions by ASITS in the previous day, 72.7% (16 of 22 days) of the EHD can be captured in Beijing. In the other cities, the PM2.5-EHD values were 103.9–300.3 μg m−3, with the ASITS-PD values ranging from 1.7 to 3.9. The EHD prediction accuracies using ASITS-PD were 44.4%–100.0% in the 10 cities. The highest prediction accuracy was located in Shijiazhuang, which was also the most polluted among the 10 cities, with a PM2.5-EHD of 300.3 μg m−3. With this approach, although the ASITS scheme has no explicit consideration of emission, the ASITS prediction should also have an adjustment to the emission change to some degree, because the observed PM2.5 concentrations in the recent 2 years are always used to build the statistical prediction model. Hence, if the emissions change in recent years, the PM2.5 concentrations and the ASITS-PD can also have a corresponding adjustment. To summarize, we can predict all the occurrences of EHDs in Shijiazhuang over the last year using data from the previous 2 years only. The ASITS represents an easily used and computationally efficient method for EHD predictions, which could be a supplement to the CTM predictions in NC (see Table S2 in the online supplementary material). At last, it should be noted that this approach would not be favorable when the emissions substantially change in a very short term (such as several days), which is beyond the consideration of the statistical prediction model.
d. ASITS with low- and high-frequency signals of PM2.5 concentrations
Many previous studies emphasized that changes in synoptic patterns are important drivers for variations in the PM2.5 concentrations and the occurrences of aerosol pollution events in NC (Ye et al. 2016; Bei et al. 2016; Zheng et al. 2015; Y. Zhang et al. 2016). Modulating by such synoptic activity, the PM2.5 concentrations have a strong cycle with a period of 4–7 days in NC (Guo et al. 2014; Quan et al. 2014, 2015). Synoptic activities can impact the variations in PM2.5 concentrations via controlling the air stagnation conditions (Ye et al. 2016). Driven by favorable synoptic system activities, a stagnation weather pattern could last for several days (Wang and Angell 1999). Hence, in addition to monthly and daily time scales, it is necessary to study the relationship between ASITS and PM2.5 concentration variations over a period of several days.
In this manuscript, we define low-frequency (high frequency) variations as a period greater than or equal to (less than) 5 days. Figure 9 shows the low-pass and high-pass PM2.5 concentrations and the ASITS values for Beijing. The low- and high-pass signals are calculated using a Lanczos filter (Duchon 1979). It is shown that the relationship between ASITS and the PM2.5 concentrations was more significant in the low-frequency variations than in the high-frequency variations. The low-frequency ASITS series was able to capture the crests and troughs of the PM2.5 concentrations, especially for the episodes with high PM2.5 concentrations in the wintertime. The simultaneous and largest lead–lag correlation coefficients in the low-frequency variations were 0.61 and 0.68, respectively (the largest lead–lag correlation was also observed when the ASITS led the PM2.5 concentrations by 1 day, as shown in Fig. 7). By comparison, the high-frequency ASITS series was not clearly related to the variations in the PM2.5 concentrations. The simultaneous correlation in the high-frequency series was not as significant as that in the low-frequency series. Although the lead–lag correlation coefficient for ASITS that led the PM2.5 concentrations by 1 day passed the t test with 99% confidence, the correlation value was only 0.33, which was far smaller than the corresponding value of low-frequency series. The same results can also be seen in other cities in NC (see Figs. S1–S9 in the supplementary material). The results indicate that the influence of ASITS on the variations in the PM2.5 concentrations mainly reflects the impacts of cycles of activity in synoptic systems, which has been reported in previous studies (Guo et al. 2014; Quan et al. 2014, 2015). In addition, since the ASITS mainly reflects the variations in the PM2.5 concentrations on synoptic time scales, it could also be used to assess the extreme haze events persisting for several days (see Table S1 in the supplementary material).
(a) Low-pass and (c) high-pass PM2.5 concentrations (black lines) and ASITS (red lines) and the corresponding lead–lag correlation coefficients between the (b) low-pass PM2.5 concentrations and ASITS and the (d) high-pass PM2.5 concentrations and ASITS in Beijing between 1 Mar 2015 and 31 Mar 2016. The positive and negative x-axis values for (b) and (d) denote the ASITS lead and lag correlation coefficients, respectively, corresponding to the PM2.5 concentrations. The dashed curves denote the t test at a 99% confidence level.
Citation: Journal of the Atmospheric Sciences 75, 10; 10.1175/JAS-D-17-0354.1
e. Sensitivity of the ASITS scheme to its parameters
Although the abovementioned results demonstrate that the new ASITS scheme in this manuscript is strongly related to the PM2.5 concentrations in NC, the power-law exponent 
As in Fig. 1c, but for the parameters α = 1, 1/2, 1/3, and 1/5 in (8).
Citation: Journal of the Atmospheric Sciences 75, 10; 10.1175/JAS-D-17-0354.1
Another noteworthy parameter in the ASITS is the PBLH since the PBLH could have much uncertainty due to various diagnosis methods and data (Von Engeln and Teixeira 2013; W. Zhang et al. 2016; Y. Zhang et al. 2014; Seidel et al. 2010; McGrath-Spangler and Molod 2014). To estimate the impacts of the uncertainty in PBLH, we calculated the daily ASITS when PBLH have random relative error less than 25% and 50%, respectively (Table 3). It is found that the ASITS with 25% and 50% uncertainty of PBLH were significantly correlated with the ASITS used in the manuscript, with the correlation coefficients of 0.98 and 0.91, respectively, and with the small normalized mean biases of 13% and 29%, respectively. Additionally, the PBLH directly from the reanalysis datasets should not be suitable for the real-time diagnosis of ASITS and haze prediction. So in this case, PBLH derived from other methods such as critical bulk Richardson number (BRN; McGrath-Spangler and Molod 2014; Guo et al. 2016; Y. Zhang et al. 2014) could also be used to calculate ASITS (see Fig. S10 in the supplementary material).
The correlation coefficients and NMB between the daily ASITS in the manuscript and the ASITS calculated by PBLH with uncertainties of 25% and 50% in NC from 1 Apr 2013 to 31 Mar 2016.
At last, we analyzed the uncertainty of ASITS induced by precipitation threshold by comparing the ASITS with the counterparts using the thresholds of 2 and 5 mm day−1, respectively. It is also shown that the threshold of precipitation could not affect the ASITS values so much (Table 4).
The correlation coefficients and NMB between the daily ASITS used in the manuscript and the ASITS calculated by the thresholds of 2 and 5 mm day−1 precipitation in NC from 1 Apr 2013 to 31 Mar 2016.
4. Conclusions and discussion
This study introduced a new ASITS scheme to quantify the air stagnation conditions and investigate the relationship between ASITS and observed PM2.5 concentrations on monthly and daily time scales. The new ASITS scheme considers air diffusion potency via advection, turbulence potency in the PBL, and the effect of wet deposition. The ASITS scheme was then applied to analyze and predict the occurrence of extreme haze events in NC between 1 April 2013 and 31 March 2016 using MERRA-2 meteorological fields and MEP/EMC PM2.5 concentrations in NC.
ASITS showed a close relationship with the monthly variations and PDF of the PM2.5 concentrations in NC. The correlation coefficients among 99.1% of the areas in NC passed the t test at a 99% confidence level. In many aerosol-polluted areas of NC, the correlation coefficients between ASITS and the PM2.5 concentrations reached 0.8–1.0. The PDFs of the monthly ASITS values and PM2.5 concentrations in NC all showed lognormal distributions with similar NMBs. On the daily time scale, ASITS is applicable for analyzing the variations in the PM2.5 concentrations and extreme haze events in NC. The daily ASITS and PM2.5 in NC were significantly correlated from 1 April 2013 to 31 March 2016. The ASITS scheme was able to track the variations in the PM2.5 concentrations in the 10 polluted cities in NC. In addition, because the correlation coefficients for ASITS leading the PM2.5 concentrations by 1 day were greater than the simultaneous correlation coefficients for approximately 98% of the grid cells of NC, ASITS is a useful leading predictor of the extreme haze events in NC. The prediction accuracies of extreme haze days using the ASITS scheme were favorably 44.4%–100.0% for the 10 cities. It is also found that the air stagnation conditions represented by ASITS and the PM2.5 concentrations mainly depended on low-frequency variations (signal period ≥5 days) compared with high-frequency variations (signal period <5 days). These results indicate that the influence of ASITS on the variations in PM2.5 concentrations mainly reflects the impacts of activities in synoptic systems, as was reported in previous studies.
It should be noted that this study only applied the ASITS scheme to NC. But we find that ASITS does not have a comparably strong relationship with the variations in the PM2.5 concentrations over southern China (SC) as it does for NC. This could be possibly attributed to two reasons: 1) some parameters of ASITS are designed mainly with regard to the magnitudes of the variables for NC, and 2) other meteorological parameters, including higher humidity, temperature, and volatile organic compound emissions, are quite different in SC (Fu and Liao 2012) and are not related to the air stagnation conditions but could markedly impact the aerosol concentrations over SC (Li et al. 2014; Jiang et al. 2012; Y. W. Zhang et al. 2015; Chan et al. 2017; Qin et al. 2016; Jiang et al. 2008). Hence, a modified meteorology index involving additional physical and chemical processes is required to qualify EHDs and variations in the aerosol concentrations for SC.
Because of the good performance in illustrating the variations in the PM2.5 concentrations driven by air stagnation weather, the ASITS scheme may be a useful tool for analyzing variations in the aerosol pollution in NC that are driven by variations in meteorological conditions on the synoptic scale and month-to-month time scale. Because of the significant leading correlations shown between the daily ASITS values and the PM2.5 concentrations and simple calculation requirements, the ASITS scheme can be used to predict the occurrences of EHDs in NC. In section 3c, we simply used an empirical model with a constant ASITS-PD by the data of the first 2 years to predict the occurrences of EHDs in the third year. In practice, the ASITS and PM2.5 data in the previous year can be used to establish a dynamically updated ASITS-PD and then to make a more favorable prediction. We think the ASITS-based prediction is just a supplementary tool for extreme haze event prediction and analysis in NC. For example, the ASITS could be a useful analysis tool to understand the causes of haze event occurrence in NC. By comparing with the ASI and CTM prediction, more could be understood about the effects of emissions and meteorology, respectively, in an extreme haze event. The ASITS prediction could also combine with the regional meteorological models and regional air quality models to forecast the extreme haze in the subsequent few days. Finally, because of the good relationship between monthly ASITS and PM2.5 concentrations and the simple expressions of ASITS, the ASITS could be helpful for understanding the impacts of climate change and general atmospheric circulation on long-term variations in haze events over NC, which is one of the frontier study issues (Feng et al. 2016; Cai et al. 2017; Chen and Wang 2015; Ding and Liu 2014).
Acknowledgments
This work was supported by the National Natural Science Foundation of China (Grants 41705135 and 41575010), the Major Research Plan of the National Natural Science Foundation of China (Grant 91544219), and the Ministry of Science and Technology of China (Grant IUMKY201620).
REFERENCES
Bei, N., and Coauthors, 2016: Typical synoptic situations and their impacts on the wintertime air pollution in the Guanzhong basin, China. Atmos. Chem. Phys., 16, 7373–7387, https://doi.org/10.5194/acp-16-7373-2016.
Bosilovich, M. G., and Coauthors, 2015: MERRA-2 : Initial evaluation of the climate. NASA Tech. Memo. NASA/TM-2015-104606/Vol. 43, 145 pp., https://gmao.gsfc.nasa.gov/pubs/docs/Bosilovich803.pdf.
Bosilovich, M. G., F. R. Robertson, L. Takacs, A. Molod, and D. Mocko, 2017: Atmospheric water balance and variability in the MERRA-2 reanalysis. J. Climate, 30, 1177–1196, https://doi.org/10.1175/JCLI-D-16-0338.1.
Cai, W., K. Li, H. Liao, H. Wang, and L. Wu, 2017: Weather conditions conducive to Beijing severe haze more frequent under climate change. Nat. Climate Change, 7, 257–262, https://doi.org/10.1038/nclimate3249.
Chan, K. L., S. Wang, C. Liu, B. Zhou, M. O. Wenig, and A. Saiz-Lopez, 2017: On the summertime air quality and related photochemical processes in the megacity Shanghai, China. Sci. Total Environ., 580, 974–983, https://doi.org/10.1016/j.scitotenv.2016.12.052.
Chen, C., and Coauthors, 2015: Characteristics and sources of submicron aerosols above the urban canopy (260 m) in Beijing, China, during the 2014 APEC summit. Atmos. Chem. Phys., 15, 12 879–12 895, https://doi.org/10.5194/acp-15-12879-2015.
Chen, D., Z. Liu, J. Fast, and J. Ban, 2016: Simulations of sulfate–nitrate–ammonium (SNA) aerosols during the extreme haze events over northern China in October 2014. Atmos. Chem. Phys., 16, 10 707–10 724, https://doi.org/10.5194/acp-16-10707-2016.
Chen, H., and H. Wang, 2015: Haze days in north China and the associated atmospheric circulations based on daily visibility data from 1960 to 2012. J. Geophys. Res. Atmos., 120, 5895–5909, https://doi.org/10.1002/2015JD023225.
Chen, J., and Coauthors, 2012: A parameterization of low visibilities for hazy days in the North China Plain. Atmos. Chem. Phys., 12, 4935–4950, https://doi.org/10.5194/acp-12-4935-2012.
Chen, Z., J. Zhang, T. Zhang, W. Liu, and J. Liu, 2015: Haze observations by simultaneous lidar and WPS in Beijing before and during APEC, 2014. Sci. China Chem., 58, 1385–1392, https://doi.org/10.1007/s11426-015-5467-x.
Cobourn, W. G., 2010: An enhanced PM2.5 air quality forecast model based on nonlinear regression and back-trajectory concentrations. Atmos. Environ., 44, 3015–3023, https://doi.org/10.1016/j.atmosenv.2010.05.009.
Coy, L., K. Wargan, A. M. Molod, W. R. McCarty, and S. Pawson, 2016: Structure and dynamics of the quasi-biennial oscillation in MERRA-2. J. Climate, 29, 5339–5354, https://doi.org/10.1175/JCLI-D-15-0809.1.
Ding, Y., and Y. Liu, 2014: Analysis of long-term variations of fog and haze in China in recent 50 years and their relations with atmospheric humidity. Sci. China Earth Sci., 57, 36–46, https://doi.org/10.1007/s11430-013-4792-1.
Duchon, C. E., 1979: Lanczos filtering in one and two dimensions. J. Appl. Meteor., 18, 1016–1022, https://doi.org/10.1175/1520-0450(1979)018<1016:LFIOAT>2.0.CO;2.
Feng, J., H. Liao, and Y. Gu, 2016: A comparison of meteorology-driven interannual variations of surface aerosol concentrations in the eastern United States, eastern China, and Europe. SOLA, 12, 146–152, https://doi.org/10.2151/sola.2016-031.
Fu, Y., and H. Liao, 2012: Simulation of the interannual variations of biogenic emissions of volatile organic compounds in China: Impacts on tropospheric ozone and secondary organic aerosol. Atmos. Environ., 59, 170–185, https://doi.org/10.1016/j.atmosenv.2012.05.053.
Gao, M., G. R. Carmichael, P. E. Saide, Z. Lu, M. Yu, D. G. Streets, and Z. Wang, 2016: Response of winter fine particulate matter concentrations to emission and meteorology changes in north China. Atmos. Chem. Phys., 16, 11 837–11 851, https://doi.org/10.5194/acp-16-11837-2016.
Gelaro, R., and Coauthors, 2017: The Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). J. Climate, 30, 5419–5454, https://doi.org/10.1175/JCLI-D-16-0758.1.
Gui, K., and Coauthors, 2016: Aerosol optical properties based on ground and satellite retrievals during a serious haze episode in December 2015 over Beijing. Atmosphere, 7, 70, https://doi.org/10.3390/atmos7050070.
Guo, H., Y. Wang, and H. Zhang, 2017: Characterization of criteria air pollutants in Beijing during 2014–2015. Environ. Res., 154, 334–344, https://doi.org/10.1016/j.envres.2017.01.029.
Guo, J., and Coauthors, 2016: The climatology of planetary boundary layer height in China derived from radiosonde and reanalysis data. Atmos. Chem. Phys., 16, 13 309–13 319, https://doi.org/10.5194/acp-16-13309-2016.
Guo, L., X. Guo, C. Fang, and S. Zhu, 2015: Observation analysis on characteristics of formation, evolution and transition of a long-lasting severe fog and haze episode in north China. Sci. China Earth Sci., 58, 329–344, https://doi.org/10.1007/s11430-014-4924-2.
Guo, S., and Coauthors, 2014: Elucidating severe urban haze formation in China. Proc. Natl. Acad. Sci. USA, 111, 17 373–17 378, https://doi.org/10.1073/pnas.1419604111.
Han, S., and Coauthors, 2014: Characteristics and formation mechanism of a winter haze—Fog episode in Tianjin, China. Atmos. Environ., 98, 323–330, https://doi.org/10.1016/j.atmosenv.2014.08.078.
Hara, Y., and Coauthors, 2011: Seasonal characteristics of spherical aerosol distribution in eastern Asia: Integrated analysis using ground/space-based lidars and a chemical transport model. SOLA, 7, 121–124, https://doi.org/10.2151/sola.2011-031.
He, Q., C. Li, F. Geng, G. Zhou, W. Gao, W. Yu, Z. Li, and M. Du, 2016: A parameterization scheme of aerosol vertical distribution for surface-level visibility retrieval from satellite remote sensing. Remote Sens. Environ., 181, 1–13, https://doi.org/10.1016/j.rse.2016.03.016.
Horton, D. E., Harshvardhan, and N. S. Diffenbaugh, 2012: Response of air stagnation frequency to anthropogenically enhanced radiative forcing. Environ. Res. Lett., 7, 044034, https://doi.org/10.1088/1748-9326/7/4/044034.
Horton, D. E., C. B. Skinner, D. Singh, and N. S. Diffenbaugh, 2014: Occurrence and persistence of future atmospheric stagnation events. Nat. Climate Change, 4, 698–703, https://doi.org/10.1038/nclimate2272.
Hu, J., Y. Wang, Q. Ying, and H. Zhang, 2014: Spatial and temporal variability of PM2.5 and PM10 over the North China Plain and the Yangtze River delta, China. Atmos. Environ., 95, 598–609, https://doi.org/10.1016/j.atmosenv.2014.07.019.
Hu, J., J. Chen, Q. Ying, and H. Zhang, 2016: One-year simulation of ozone and particulate matter in China using WRF/CMAQ modeling system. Atmos. Chem. Phys., 16, 10 333–10 350, https://doi.org/10.5194/acp-16-10333-2016.
Huang, L., C.-S. Yuan, G. Wang, and K. Wang, 2011: Chemical characteristics and source apportionment of PM10 during a brown haze episode in Harbin, China. Particuology, 9, 32–38, https://doi.org/10.1016/j.partic.2010.07.022.
Huang, Q., X. Cai, Y. Song, and T. Zhu, 2017: Air stagnation in China (1985–2014): Climatological mean features and trends. Atmos. Chem. Phys., 17, 7793–7805, https://doi.org/10.5194/acp-17-7793-2017.
IPCC, 2013: Climate Change 2013: The Physical Science Basis. Cambridge University Press, 1535 pp., https://doi.org/10.1017/CBO9781107415324.
Jacob, D. J., and D. A. Winner, 2009: Effect of climate change on air quality. Atmos. Environ., 43, 51–63, https://doi.org/10.1016/j.atmosenv.2008.09.051.
Jiang, F., T. Wang, T. Wang, M. Xie, and H. Zhao, 2008: Numerical modeling of a continuous photochemical pollution episode in Hong Kong using WRF–Chem. Atmos. Environ., 42, 8717–8727, https://doi.org/10.1016/j.atmosenv.2008.08.034.
Jiang, F., Q. Liu, X. Huang, T. Wang, B. Zhuang, and M. Xie, 2012: Regional modeling of secondary organic aerosol over China using WRF/Chem. J. Aerosol Sci., 43, 57–73, https://doi.org/10.1016/j.jaerosci.2011.09.003.
Jiang, Y., R. Zhu, K. Zhu, and Z. Li, 2015: Numerical simulation on the air pollution potential in the severe air pollution episodes in Beijing–Tianjin–Hebei region (in Chinese). Acta Sci. Circumstantiae, 35, 2681–2692.
Korshover, J., 1976: Climatology of stagnating anticyclones east of the Rocky Mountains, 1936–1975. NOAA Tech. Memo. ERL ARL-55, 26 pp.
Korshover, J., and J. K. Angell, 1982: A review of air-stagnation cases in the eastern United States during 1981—Annual summary. Mon. Wea. Rev., 110, 1515–1518, https://doi.org/10.1175/1520-0493(1982)110<1515:AROASC>2.0.CO;2.
Koster, R. D., and Coauthors, 2016: MERRA-2 input observations: Summary and assessment. NASA/TM-2016-104606/Vol. 46, 51 pp., https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20160014544.pdf.
Leibensperger, E. M., L. J. Mickley, and D. J. Jacob, 2008: Sensitivity of US air quality to mid-latitude cyclone frequency and implications of 1980–2006 climate change. Atmos. Chem. Phys., 8, 7075–7086, https://doi.org/10.5194/acpd-8-12253-2008.
Li, J., and Z. Han, 2016: Aerosol vertical distribution over east China from RIEMS-Chem simulation in comparison with CALIPSO measurements. Atmos. Environ., 143, 177–189, https://doi.org/10.1016/j.atmosenv.2016.08.045.
Li, J., H. Du, Z. Wang, Y. Sun, W. Yang, J. J. Li, X. Tang, and P. Fu, 2017: Rapid formation of a severe regional winter haze episode over a mega-city cluster on the North China Plain. Environ. Pollut., 223, 605–615, https://doi.org/10.1016/j.envpol.2017.01.063.
Li, L., and Coauthors, 2014: An integrated process rate analysis of a regional fine particulate matter episode over Yangtze River delta in 2010. Atmos. Environ., 91, 60–70, https://doi.org/10.1016/j.atmosenv.2014.03.053.
Li, Z., and Coauthors, 2016: Remote sensing of atmospheric particulate mass of dry PM2.5 near the ground: Method validation using ground-based measurements. Remote Sens. Environ., 173, 59–68, https://doi.org/10.1016/j.rse.2015.11.019.
Li, Z., and Coauthors, 2017: Aerosol and boundary-layer interactions and impact on air quality. Natl. Sci. Rev., 4, 810–833, https://doi.org/10.1093/nsr/nwx117.
Lin, J.-T., and M. B. McElroy, 2010: Impacts of boundary layer mixing on pollutant vertical profiles in the lower troposphere: Implications to satellite remote sensing. Atmos. Environ., 44, 1726–1739, https://doi.org/10.1016/j.atmosenv.2010.02.009.
Liu, J., D. Wu, and S. J. Fan, 2015: Distribution of regional pollution and the characteristics of vertical wind field in the Pearl River delta (in Chinese). Environ. Sci., 36, 3989–3998.
Liu, Y., M. Franklin, R. Kahn, and P. Koutrakis, 2007: Using aerosol optical thickness to predict ground-level PM2.5 concentrations in the St. Louis area: A comparison between MISR and MODIS. Remote Sens. Environ., 107, 33–44, https://doi.org/10.1016/j.rse.2006.05.022.
Long, X., and Coauthors, 2016: Impact of crop field burning and mountains on heavy haze in the North China Plain: A case study. Atmos. Chem. Phys., 16, 9675–9691, https://doi.org/10.5194/acp-16-9675-2016.
Lu, H.-C., 2002: The statistical characters of PM10 concentration in Taiwan area. Atmos. Environ., 36, 491–502, https://doi.org/10.1016/S1352-2310(01)00245-X.
Mamtimin, B., and F. X. Meixner, 2011: Air pollution and meteorological processes in the growing dryland city of Urumqi (Xinjiang, China). Sci. Total Environ., 409, 1277–1290, https://doi.org/10.1016/j.scitotenv.2010.12.010.
Mao, J., H. Sheng, H. Zhao, and C. Zhou, 2014: Observation study on the size distribution of sand dust aerosol particles over Yinchuan, China. Adv. Meteor., 2014, 157645, https://doi.org/10.1155/2014/157645.
McGrath-Spangler, E. L., and A. Molod, 2014: Comparison of GEOS-5 AGCM planetary boundary layer depths computed with various definitions. Atmos. Chem. Phys., 14, 6717–6727, https://doi.org/10.5194/acp-14-6717-2014.
Mijling, B., R. J. van der A, and Q. Zhang, 2013: Regional nitrogen oxides emission trends in East Asia observed from space. Atmos. Chem. Phys., 13, 12 003–12 012, https://doi.org/10.5194/acp-13-12003-2013.
Mu, Q., and H. Liao, 2014: Simulation of the interannual variations of aerosols in China: Role of variations in meteorological parameters. Atmos. Chem. Phys., 14, 11 177–11 219, https://doi.org/10.5194/acpd-14-11177-2014.
Ovadnevaitė, J., K. Kvietkus, and J. Šakalys, 2007: Evaluation of the impact of long-range transport and aerosol concentration temporal variations at the eastern coast of the Baltic Sea. Environ. Monit. Assess., 132, 365–375, https://doi.org/10.1007/s10661-006-9540-y.
Qin, Y. M., Y. J. Li, H. Wang, B. P. Y. L. Lee, D. D. Huang, and C. K. Chan, 2016: Particulate matter (PM) episodes at a suburban site in Hong Kong: Evolution of PM characteristics and role of photochemistry in secondary aerosol formation. Atmos. Chem. Phys., 16, 14 131–14 145, https://doi.org/10.5194/acp-16-14131-2016.
Quan, J., Q. Zhang, H. He, J. Liu, M. Huang, and H. Jin, 2011: Analysis of the formation of fog and haze in North China Plain (NCP). Atmos. Chem. Phys., 11, 8205–8214, https://doi.org/10.5194/acp-11-8205-2011.
Quan, J., X. Tie, Q. Zhang, Q. Liu, X. Li, Y. Gao, and D. Zhao, 2014: Characteristics of heavy aerosol pollution during the 2012–2013 winter in Beijing, China. Atmos. Environ., 88, 83–89, https://doi.org/10.1016/j.atmosenv.2014.01.058.
Quan, J., Q. Liu, X. Li, Y. Gao, X. Jia, J. Sheng, and Y. Liu, 2015: Effect of heterogeneous aqueous reactions on the secondary formation of inorganic aerosols during haze events. Atmos. Environ., 122, 306–312, https://doi.org/10.1016/j.atmosenv.2015.09.068.
Reichle, R. H., Q. Liu, R. D. Koster, C. S. Draper, S. P. P. Mahanama, and G. S. Partyka, 2017: Land surface precipitation in MERRA-2. J. Climate, 30, 1643–1664, https://doi.org/10.1175/JCLI-D-16-0570.1.
Šakalys, J., K. Kvietkus, D. Čeburnis, and D. Valiulis, 2004: The method of determination of heavy metals background concentration in the moss. Environ. Chem. Phys., 26, 109–117.
Seidel, D. J., C. O. Ao, and K. Li, 2010: Estimating climatological planetary boundary layer heights from radiosonde observations: Comparison of methods and uncertainty analysis. J. Geophys. Res., 115, D16113, https://doi.org/10.1029/2009JD013680.
Seinfeld, J. H., and S. N. Pandis, 2006: Atmospheric Chemistry and Physics: From Air Pollution to Climate Change. 2nd ed. Wiley-Interscience, 1232 pp.
Sun, X., Y. Yin, Y. Sun, Y. Sun, W. Liu, and Y. Han, 2013: Seasonal and vertical variations in aerosol distribution over Shijiazhuang, China. Atmos. Environ., 81, 245–252, https://doi.org/10.1016/j.atmosenv.2013.08.009.
Tai, A. P. K., L. J. Mickley, and D. J. Jacob, 2010: Correlations between fine particulate matter (PM2.5) and meteorological variables in the United States: Implications for the sensitivity of PM2.5 to climate change. Atmos. Environ., 44, 3976–3984, https://doi.org/10.1016/j.atmosenv.2010.06.060.
Tai, A. P. K., L. J. Mickley, D. J. Jacob, E. M. Leibensperger, L. Zhang, J. A. Fisher, and H. O. T. Pye, 2012: Meteorological modes of variability for fine particulate matter (PM2.5) air quality in the United States: Implications for PM2.5 sensitivity to climate change. Atmos. Chem. Phys., 12, 3131–3145, https://doi.org/10.5194/acp-12-3131-2012.
Tang, Y.-X., H. Liao, and J. Feng, 2017: Estimating emissions and concentrations of road dust aerosol over China using the GEOS-Chem model. Atmos. Ocean. Sci. Lett., 10, 298–305, https://doi.org/10.1080/16742834.2017.1320935.
Tie, X., D. Wu, and G. Brasseur, 2009: Lung cancer mortality and exposure to atmospheric aerosol particles in Guangzhou, China. Atmos. Environ., 43, 2375–2377, https://doi.org/10.1016/j.atmosenv.2009.01.036.
Von Engeln, A., and J. Teixeira, 2013: A planetary boundary layer height climatology derived from ECMWF reanalysis data. J. Climate, 26, 6575–6590, https://doi.org/10.1175/JCLI-D-12-00385.1.
Wang, J. X. L., and J. K. Angell, 1999: Air stagnation climatology for the United States (1948-1998). NOAA/Air Resources Laboratory ATLAS 1, 76 pp., https://www.arl.noaa.gov/documents/reports/atlas.pdf.
Wang, S., S. Yu, P. Li, L. Wang, K. Mehmood, W. Liu, R. Yan, and X. Zheng, 2017: A study of characteristics and origins of haze pollution in Zhengzhou, China, based on observations and hybrid receptor models. Aerosol Air Qual. Res., 17, 513–528, https://doi.org/10.4209/aaqr.2016.06.0238.
Wang, X., and Coauthors, 2014: Severe haze episodes and seriously polluted fog water in Ji’nan, China. Sci. Total Environ., 493, 133–137, https://doi.org/10.1016/j.scitotenv.2014.05.135.
Wang, Y., A. Khalizov, M. Levy, and R. Zhang, 2013: New directions: Light absorbing aerosols and their atmospheric impacts. Atmos. Environ., 81, 713–715, https://doi.org/10.1016/j.atmosenv.2013.09.034.
Wang, Y., Q. Q. Zhang, K. He, Q. Zhang, and L. Chai, 2013: Sulfate–nitrate–ammonium aerosols over China: Response to 2000–2015 emission changes of sulfur dioxide, nitrogen oxides, and ammonia. Atmos. Chem. Phys., 13, 2635–2652, https://doi.org/10.5194/acp-13-2635-2013.
Wang, Y. H., Z. R. Liu, J. K. Zhang, B. Hu, D. S. Ji, Y. C. Yu, and Y. S. Wang, 2015: Aerosol physicochemical properties and implications for visibility during an intense haze episode during winter in Beijing. Atmos. Chem. Phys., 15, 3205–3215, https://doi.org/10.5194/acp-15-3205-2015.
Yang, Y. Q., and Coauthors, 2016: PLAM—A meteorological pollution index for air quality and its applications in fog-haze forecasts in north China. Atmos. Chem. Phys., 16, 1353–1364, https://doi.org/10.5194/acp-16-1353-2016.
Yang, Y. R., and Coauthors, 2015: Characteristics and formation mechanism of continuous hazes in China: A case study during the autumn of 2014 in the North China Plain. Atmos. Chem. Phys., 15, 8165–8178, https://doi.org/10.5194/acp-15-8165-2015.
Ye, X., Y. Song, X. Cai, and H. Zhang, 2016: Study on the synoptic flow patterns and boundary layer process of the severe haze events over the North China Plain in January 2013. Atmos. Environ., 124, 129–145, https://doi.org/10.1016/j.atmosenv.2015.06.011.
You, W., Z. Zang, L. Zhang, Y. Li, and W. Wang, 2016: Estimating national-scale ground-level PM25 concentration in China using geographically weighted regression based on MODIS and MISR AOD. Environ. Sci. Pollut. Res. Int., 23, 8327–8338, https://doi.org/10.1007/s11356-015-6027-9.
Zhang, Q., and Coauthors, 2015: Variations in PM2.5, TSP, BC, and trace gases (NO2, SO2, and O3) between haze and non-haze episodes in winter over Xi’an, China. Atmos. Environ., 112, 64–71, https://doi.org/10.1016/j.atmosenv.2015.04.033.
Zhang, R., Q. Li, and R. Zhang, 2014: Meteorological conditions for the persistent severe fog and haze event over eastern China in January 2013. Sci. China Earth Sci., 57, 26–35, https://doi.org/10.1007/s11430-013-4774-3.
Zhang, T., W. Gong, W. Wang, Y. Ji, Z. Zhu, and Y. Huang, 2016: Ground level PM2.5 estimates over China using satellite-based geographically weighted regression (GWR) models are improved by including NO2 and enhanced vegetation index (EVI). Int. J. Environ. Res. Public Health, 13, 1215, https://doi.org/10.3390/ijerph13121215.
Zhang, W., J. Guo, Y. Miao, H. Liu, Y. Zhang, Z. Li, and P. Zhai, 2016: Planetary boundary layer height from CALIOP compared to radiosonde over China. Atmos. Chem. Phys., 16, 9951–9963, https://doi.org/10.5194/acp-16-9951-2016.
Zhang, X. Y. C., Y. Q. Wang, T. Niu, S. L. Gong, Y. M. Zhang, and J. Y. Sun, 2012: Atmospheric aerosol compositions in China: Spatial/temporal variability, chemical signature, regional haze distribution and comparisons with global aerosols. Atmos. Chem. Phys., 12, 779–799, https://doi.org/10.5194/acp-12-779-2012.
Zhang, Y., and Z. Li, 2015: Remote sensing of atmospheric fine particulate matter (PM2.5) mass concentration near the ground from satellite observation. Remote Sens. Environ., 160, 252–262, https://doi.org/10.1016/j.rse.2015.02.005.
Zhang, Y., Z. Gao, D. Li, Y. Li, N. Zhang, X. Zhao, and J. Chen, 2014: On the computation of planetary boundary-layer height using the bulk Richardson number method. Geosci. Model Dev., 7, 2599–2611, https://doi.org/10.5194/gmd-7-2599-2014.
Zhang, Y., A. Ding, H. Mao, W. Nie, D. Zhou, L. Liu, X. Huang, and C. Fu, 2016: Impact of synoptic weather patterns and inter-decadal climate variability on air quality in the North China Plain during 1980–2013. Atmos. Environ., 124, 119–128, https://doi.org/10.1016/j.atmosenv.2015.05.063.
Zhang, Y. W., and Coauthors, 2015: Significant concentration changes of chemical components of PM1 in the Yangtze River delta area of China and the implications for the formation mechanism of heavy haze–fog pollution. Sci. Total Environ., 538, 7–15, https://doi.org/10.1016/j.scitotenv.2015.06.104.
Zhao, T. L., S. L. Gong, P. Huang, and D. Lavoué, 2012: Hemispheric transport and influence of meteorology on global aerosol climatology. Atmos. Chem. Phys., 12, 7609–7624, https://doi.org/10.5194/acp-12-7609-2012.
Zhao, X. J., P. S. Zhao, J. Xu, W. Meng, W. W. Pu, F. Dong, D. He, and Q. F. Shi, 2013: Analysis of a winter regional haze event and its formation mechanism in the North China Plain. Atmos. Chem. Phys., 13, 5685–5696, https://doi.org/10.5194/acp-13-5685-2013.
Zheng, G. J., and Coauthors, 2015: Exploring the severe winter haze in Beijing: The impact of synoptic weather, regional transport and heterogeneous reactions. Atmos. Chem. Phys., 15, 2969–2983, https://doi.org/10.5194/acp-15-2969-2015.
Zou, B., J. Chen, L. Zhai, X. Fang, and Z. Zheng, 2017: Satellite based mapping of ground PM2.5 concentration using generalized additive modeling. Remote Sens., 9, 1, https://doi.org/10.3390/rs9010001.
