Widespread air pollutants of the North China Plain during the Asian summer monsoon season: a case study

During the Asian summer monsoon season, prevailing southeasterly–southwesterly winds are subject to delivering air pollutants from the North China Plain (NCP) to northeast and northwest China. In the present study, the WRF-CHEM model is used to evaluate contributions of trans-boundary transport of NCP emissions to the air quality in northeast and northwest China during a persistent air pollution episode from 22 to 28 May 2015. The WRF-CHEM model generally performs well in capturing the observed temporal variation and spatial distribution of fine particulate matter (PM_2.5), ozone (O₃), and NO₂. The simulated temporal variation of aerosol species is also in good agreement with measurements in Beijing during the episode. Model simulations show that NCP emissions contribute substantially to the PM_2.5 level in Liaoning and Shanxi provinces, the adjacent downwind areas of the NCP, with an average of 24.2 and 13.9 µg m⁻³ during the episode, respectively. The PM_2.5 contributions in Jilin and Shaanxi provinces are also appreciable, with an average of 9.6 and 6.5 µg m⁻³, respectively. The average percentage contributions of NCP emissions to the PM_2.5 level in Liaoning, Jilin, Shanxi, and Shaanxi provinces are 40.6, 27.5, 32.2, and 20.9 %, respectively. The NCP emissions contribute remarkably to the O₃ level in Liaoning province, with an average of 46.5 µg m⁻³, varying from 23.9 to 69.5 µg m⁻³. The O₃ level in Shanxi province is also influenced considerably by NCP emissions, with an average contribution of 35.1 µg m⁻³. The O₃ level in Shanxi province is also influenced considerably by NCP emissions, with an average contribution of 35.1 µg m⁻³. The average O₃ contributions of NCP emissions to Jilin and Shaanxi provinces are 28.7 and 20.7 µg m⁻³, respectively. The average percentage contributions of NCP emissions to the afternoon O₃ level in Liaoning, Jilin, Shanxi, and Shaanxi provinces are 27.4, 19.5, 21.2, and 15.8 %, respectively. However, the effect of NCP emissions on the air quality in Inner Mongolia is generally insignificant. Therefore, effective mitigation of NCP emissions not only improves the local air quality, but is also beneficial to the air quality in northeast and northwest China during the Asian summer monsoon season.

1 Introduction

With the rapid growth of industrialization, urbanization, and transportation, China has recently experienced severe air pollution with high levels of fine particulate matter (PM_2.5) and ozone (O₃) (e.g., Chan and Yao, 2008; Zhang et al., 2013; Kurokawa et al., 2013; G. Li et al., 2017b). Although the Chinese State Council issued the “Atmospheric Pollution Prevention and Control Action Plan” in September 2013, with the aim of improving China's air quality, heavy haze or photochemical smog still frequently plagues China, especially in the North China Plain (NCP), Yangtze River Delta (YRD), and Pearl River Delta (PRD). Elevated O₃ and PM_2.5 concentrations in the atmosphere not only perturb regional and global climates, but also exert adverse effects on air quality, ecosystems, and human health (Weinhold, 2008; Parrish and Zhu, 2009).

The NCP has become one of the most polluted areas in the world due to a large amount of emissions of pollutants and frequent occurrence of unfavorable meteorological situations, as well as the topography (e.g., Tang et al., 2012; Zhang et al., 2013; Zhuang et al., 2014; Pu et al., 2015; Long et al., 2016). Heavy haze with extremely high PM_2.5 concentrations often covers the NCP during wintertime, partially attributable to coal combustion for domestic heating (e.g., He et al., 2001; Cao et al., 2007; H. Li et al., 2017). However, even in summer, with improvement of the evacuation conditions and increase of precipitation, photochemical smog with high levels of PM_2.5 and O₃ still engulfs the NCP (e.g., Gao et al., 2011; Hu et al., 2014; Cao et al., 2015; Wu et al., 2017). The PM_2.5 concentrations during summertime in the NCP are generally lower than those in winter, but are still much higher than 35 µg m⁻³, the first grade of National Ambient Air Quality Standards (NAAQS) in China (Feng et al., 2016; Z. S. Wang et al., 2016; Sun et al., 2016). The average summertime PM_2.5 concentrations in the NCP are 77.0 ± 41.9 µg m⁻³ in 2013, much more than those in other regions of China and also exceeding the second grade of NAAQS (Hu et al., 2014). In addition, increasing O₃ precursor emissions have caused serious O₃ pollution during summertime in the NCP (e.g., Zhang et al., 2009; Xu et al., 2011; Kurokawa et al., 2013). G. Li et al. (2017b) reported that the maximum 1h O₃ concentration exceeded 200 µg m⁻³ in almost all the cities in eastern China from April to September 2015, mainly concentrated in the NCP and YRD, showing widespread and persistent O₃ pollution. Ma et al. (2016) found a growth trend of surface O₃ at a rural site in the NCP from 2003 to 2015, with an average rate of 1.13 ppb per year. Wu et al. (2017) showed that the average afternoon O₃ concentration in the summer of 2015 in Beijing was about 163 µg m⁻³.

China is located in a large monsoon domain, and the Asia summer monsoon (ASM) tends to substantially influence the distribution and trans-boundary transport of air pollutants in China. Zhu et al. (2004) have proposed that the summertime high O₃ concentration over western China is due to the monsoonal transport from eastern China and long-range transport from south/central Asia and even Europe. Zhao et al. (2010) have also indicated that O₃ transported from south/central Asia to western China increases from May to August because of the northward movement of the India summer monsoon. Huang et al. (2015) have suggested that an earlier onset of the ASM would lead to more O₃ enhancement in the lower troposphere over the NCP in later spring and early summer. Numerous studies have also reported that the strength and temporospatial extension of the ASM influences the spatial and temporal distribution of aerosol mass concentrations over eastern China (Cao et al., 2015; Li et al., 2016; Cheng et al., 2016). For example, Zhang et al. (2010) have emphasized that the east ASM plays a major role in determining the seasonal and interannual variations of the PM_2.5 concentration over eastern China. Using the GEOS-CHEM model, Zhu et al. (2012) have shown that the weakening of the ASM increases the aerosol concentration in eastern China. Wu et al. (2016) have pointed out that the regional transport and tempo-spatial distribution of air pollutants are directly influenced by the East Asian monsoon at seasonal, interannual, and decadal scales.

During the ASM season, meteorological conditions are characterized by prevailing southwesterly–southeasterly winds over eastern China. Air pollutants originated from the NCP are likely to be transported northwards and affect the air quality in its downwind areas, so it is imperative to quantitatively evaluate the effect of NCP emissions on the air quality in its neighboring regions. Previous studies have concentrated on the composition, characteristics, and sources of the air pollutants over the NCP (e.g., Han et al., 2006; Liu et al., 2012; Zhao et al., 2013; Li et al., 2015). However, few studies have been performed to investigate the effect of trans-boundary transport of air pollutants originated from the NCP on the air quality in northeast and northwest China under the prevailing southerly wind associated with the ASM.

In this study, we first analyze the role of synoptic situations during the ASM (from May to September) in the trans-boundary transport over northern China and further evaluate the contribution of trans-boundary transport of pollutants originated from the NCP to the air quality in northeast and northwest China using the WRF-CHEM model. The model configuration and methodology are described in Sect. 2. Analysis results and discussions are presented in Sect. 3, and conclusions are given in Sect. 4.

Table 1 WRF-CHEM model configurations.

2 Model and methodology

2.1 WRF-CHEM model and configuration

A persistent air pollution episode with high levels of PM_2.5 and O₃ from 22 to 28 May 2015 in northern China is simulated using the WRF-CHEM model, which was developed by Li et al. (2010, 2011a, b, 2012) at the Molina Center for Energy and the Environment. Table 1 provides detailed model configurations and Fig. 1 shows the WRF-CHEM model simulation domain. It is worth noting that the horizontal resolution of 10 km adopted in this study is the lower bound for the WRF model to turn on the cumulus scheme, so the new Kain–Fritsch scheme is used in the present study (Table 1). Further description of the model is presented in the Supplement.

Figure 1 WRF-CHEM simulation domain with topography. The blue circles represent centers of cities with ambient monitoring sites and the red circle denotes the NCNST site. The size of the blue circle represents the number of ambient monitoring sites of cities.

The key characteristics of the aerosol pollution in China are frequently associated with rather efficient secondary formation, including aerosol nucleation and rapid growth under favorable conditions (Zhang et al., 2012, 2015; Qiu and Zhang, 2013; Guo et al., 2014). The new particle production rate in the WRF-CHEM model is calculated due to the binary nucleation of H₂ SO₄ and H₂O vapor. The nucleation rate is a parameterized function of temperature, relative humidity, and the vapor-phase H₂SO₄ concentration, following the work of Kulmala et al. (1998), and the new particles are assumed to have a diameter of 2.0 nm. Recent studies have shown that organic vapors are involved in the nucleation process (Zhang et al., 2012) and further studies need to be conducted to consider the contributions of organic vapors to the nucleation process. The secondary organic aerosol (SOA) formation is simulated using a nontraditional SOA model including the volatility basis-set modeling method in which primary organic components are assumed to be semi-volatile and photochemically reactive and are distributed in logarithmically spaced volatility bins (Li et al., 2011a). The contributions of glyoxal and methylglyoxal to the SOA formation are also included in the SOA module. The SOA formation from glyoxal and methylglyoxal is parameterized as a first-order irreversible uptake by aerosol particles, with a reactive uptake coefficient of 3.7 × 10⁻³ for glyoxal and methylglyoxal (Zhao et al., 2006). The simulation of inorganic aerosols in the WRF-CHEM model adopts the ISORROPIA version 1.7 (Nenes et al., 1998).

For the discussion convenience, northern China is divided into three regions (Fig. S1 in the Supplement): (1) the North China Plain (including Beijing, Tianjin, Hebei, Shandong, Henan, and the south of Jiangsu and Anhui, hereafter referred to as the NCP), (2) northeast China (including Heilongjiang, Jilin, Liaoning, and the east part of Inner Mongolia, hereafter referred to as NEC), and (3) northwest China (including Shanxi, Shaanxi, and the west part of Inner Mongolia, hereafter referred to as NWC). During the episode, the observed average daily PM_2.5 concentration was 75.5 µg m⁻³ and the mean O₃ concentration in the afternoon was 151.2 µg m⁻³ in the NCP. Figure S2 presents the distributions of the anthropogenic emission rates of volatile organic compounds (VOCs), nitrogen oxide (NO_x), organic carbon (OC), and SO₂ in mainland China, showing that the high emission rates of VOCs, NO_x, OC, and SO₂ are generally concentrated in the NCP. It is worth noting that uncertainties in the emission inventory used in this study are rather large, considering the rapid changes in anthropogenic emissions that are not fully reflected in the current emission inventory and the complexity of pollutants' precursors.

2.2 Data and methodology

In the present study, the model performance is validated using the hourly measurements of O₃, NO₂, and PM_2.5 concentrations released by China's Ministry of Environment Protection (China MEP), which can be accessed at http://www.aqistudy.cn/ (last access: 8 June 2018). In addition, the simulated submicron sulfate, nitrate, ammonium, and organic aerosols are compared to the measurements by the Aerodyne Aerosol Chemical Speciation Monitor (ACSM), which was deployed at the National Center for Nanoscience and Technology (NCNST), Chinese Academy of Sciences in Beijing (Fig. 1). The observed mass spectra of organic aerosols are analyzed using the positive matrix factorization (PMF) technique and four components are identified: hydrocarbon-like organic aerosol (HOA), cooking organic aerosol (COA), coal combustion organic aerosol (CCOA), and oxygenated organic aerosol (OOA). HOA, COA, and CCOA are interpreted together as a surrogate of primary organic aerosols (POA), and OOA is a surrogate of secondary organic aerosols (SOA). Furthermore, the reanalysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF) are used to analyze the synoptic patterns during the ASM season from May to September 2015.

The mean bias (MB), root mean square error (RMSE), and the index of agreement (IOA) are utilized to evaluate the performance of the WRF-CHEM model simulations against measurements. To assess the contributions of NCP emissions to the near-surface concentrations of O₃ and PM_2.5 in NEC and NWC, the factor separation approach (FSA) is used in this study (Stein and Alpert, 1993; Gabusi et al., 2008; Li et al., 2014). The detailed description of methodology can be found in Sect. S2.

Figure 2 (a) Geopotential height at 500 hPa and (b) the mean sea level pressure with wind vectors during the summer season in 2015.

3 Results and discussion

3.1 Synoptic patterns during the ASM season

The ASM prevails from May to September each year in China, with strong winds blowing from oceans to eastern China and bringing warm and moist airflow to the continent. Furthermore, the western Pacific subtropical high gradually intensifies and moves from south to north, influencing the weather and climate over China and also transporting water vapor from the sea to eastern China. During the ASM season, due to the influence of the western Pacific subtropical high, rain belts and associated deep convection move from southeastern China to northern China (Ding, 1992; Ding and Chan, 2005; Kang et al., 2002; Lau, 1992; Lau et al., 1988). Figure 2 shows the geopotential heights at 500 hPa and mean sea level pressure with wind vectors during the ASM season in 2015. At 500 hPa, the main part of the subtropical high, which is represented by the scope of the contour of 5880 geopotential meter, is located in the northwestern Pacific Ocean. The mean ridgeline of the western Pacific subtropical high is located at 25^∘ N, moving from south to north from May to September, which substantially affects the synoptic conditions in China. Flat westerly wind at 500 hPa prevails over the NCP and its surrounding regions, indicating a stable weather condition. The mean sea level pressure shows that most of areas in the NCP are continually influenced by the ASM and the high-pressure system centering in the western Pacific, causing the prevailing southeasterly–southwesterly wind over the NCP and its surrounding areas. The detailed description of the synoptic conditions during the study episode can be found in the Supplement.

Figure 3 Relationships of observed PM_2.5 and O₃ concentrations in the NCP with those in NEC during May to September from 2013 to 2016.

Figure 4 Same as Fig. 3, but for NWC.

In the region controlled by the western Pacific subtropical high, a subsidence airflow is dominant, with calm or weak winds, and the temperature is extremely high due to the strong sunlight, which is favorable for the accumulation and formation of air pollutants. The air pollutants are likely to be transported from south to north under the persistent effect of southerly winds.

Figures 3 and 4 present the relationship of PM_2.5 and O₃ concentrations in the NCP with those in NEC and NWC during the ASM season from 2013 to 2016, respectively. The observed PM_2.5 and O₃ concentrations in the NCP exhibit a positive correlation with those in NEC and NWC, with the correlation coefficients generally exceeding 0.55. There are two possible reasons for the positive correlation of PM_2.5 and O₃ concentrations between the NCP and its surrounding regions. One is that when the NCP and its neighboring areas are controlled by the same large-scale synoptic pattern, the concentrations of air pollutants generally vary synchronously. The other is the trans-boundary transport of air pollutants originated from the NCP to its surrounding areas due to the southerly wind associated with the ASM. The correlation coefficients of PM_2.5 and O₃ concentrations in the provinces of NEC with those in the NCP generally decrease from south to north, with coefficients of 0.69, 0.56, and 0.52 for PM_2.5, and of 0.86, 0.76, and 0.76 for O₃ in Liaoning, Jilin, and Heilongjiang, respectively. The decreasing trend of the correlation coefficients also exists from east to west in NWC, with coefficients of 0.69 and 0.62 for PM_2.5, and 0.87 and 0.84 for O₃ in Shanxi and Shaanxi, respectively. Hence, when severe air pollution occurs in the NCP in summer, the air quality in its adjacent provinces is likely to be deteriorated, possibly caused by the trans-boundary transport of air pollutants originated from the NCP.

It is worth noting that the intensity of ASM substantially influences the temporal variation and spatial distribution of air pollutants (Wu et al., 2016). The East Asia summer monsoon index proposed by Zhang et al. (2003) is defined as a difference of anomalous zonal wind between the 10^∘–20^∘ N, 100^∘–150^∘ E and 25^∘–35^∘ N, 100^∘–150^∘ E regions at 850 hPa during summer (June–August). A year of monsoon index greater than or equal to 2 is defined as a strong summer monsoon year, and a year of monsoon index less than or equal to −2 is defined as a weak summer monsoon year. The monsoon index calculated by the China Meteorological Administration shows that the intensity of the summer monsoon in 2015 was close to the normal (Fig. S5).

Figure 5 Comparison of measured (black dots) and predicted (blue line) diurnal profiles of near-surface hourly (a) PM_2.5, (b) O₃, and (c) NO₂ averaged over all ambient monitoring stations in northern China from 22 to 28 May 2015.

3.2 Model performance

3.2.1 PM_2.5, O₃ and NO₂ simulations in northern China

Figure 5 shows the temporal variations of observed and simulated near-surface PM_2.5, O₃, and NO₂ concentrations averaged over monitoring sites in northern China. The WRF-CHEM model generally simulates the diurnal variation of PM_2.5 concentrations in northern China well, with an IOA of 0.91. The model successfully reproduces the temporal variations of surface O₃ concentrations compared with observations in northern China, e.g., peak O₃ concentrations in the afternoon due to active photochemistry and low O₃ concentrations during nighttime caused by NO_x titration, with an IOA of 0.98. However, the model underestimation still exists in simulating the O₃ concentration, with an MB of −4.0 µg m⁻³. The model also reasonably produces the NO₂ diurnal profiles, but frequently overestimates the NO₂ concentrations in the late evening due to the simulated low planetary boundary layer (PBL) height, and underestimates the concentration in the early morning because of the uncertainties in the NO_x emissions. The further analysis of the model performance of PM_2.5, O₃ and NO₂ concentrations in northern China can be found in the Supplement.

3.2.2 Aerosol species simulations in Beijing

Figure 6 presents the temporal variations of simulated and observed aerosol species at the NCNST site in Beijing from 22 to 28 May 2015. Generally, the WRF-CHEM model predicts the temporal variations of the aerosol species against the measurements reasonably, especially for POA and nitrate aerosol, with IOAs of 0.81 and 0.90, respectively. The model has difficulties in simulating the SOA concentrations well, with an IOA and MB of 0.69 and −3.6 µg m⁻³, respectively. It is worth noting that many factors influence the SOA simulation, including measurements, meteorology, precursors emissions, and SOA formation mechanisms and treatments (Bei et al., 2012, 2013). The model reasonably tracks the temporal variation of the observed sulfate concentration, but the bias is still large, and the MB and IOA are −1.2 µg m⁻³ and 0.68, respectively. The sulfate source in the atmosphere is various, including SO₂ gas-phase oxidation by hydroxyl radicals (OH) and stabilized Criegee intermediates (sCI), aqueous reactions in cloud or fog droplets, and heterogeneous reactions on aerosol surfaces, as well as direct emissions from power plants and industries (G. Li et al., 2017a). G. Wang et al. (2016) have also reported that the aqueous oxidation of SO₂ by NO₂ is important for efficient sulfate formation. Considering that the model fails to resolve convective clouds well due to the 10 km horizontal resolution, the sulfate formation from the cloud process is generally underestimated. Additionally, a large amount of SO₂ is emitted from point sources, such as power plants or agglomerated industrial zones, which is much more sensitive to wind field simulations (Bei et al., 2010). The model performs reasonably well in simulating the ammonium aerosol, with an IOA and MB of 0.77 and −0.4 µg m⁻³, respectively.

Figure 6 Comparison of measured (black dots) and simulated (black line) diurnal profiles of submicron aerosol species of (a) POA, (b) SOA, (c) sulfate, (d) nitrate, and (e) ammonium at the NCNST site in Beijing from 22 to 28 May 2015.

Figure 7 Pattern comparison of simulated vs. observed near-surface PM_2.5 at 08:00 BJT from 23 to 28 May 2015. Colored circles show PM_2.5 observations, color contours show the PM_2.5 simulations, and black arrows represent the simulated surface winds.

3.2.3 Simulations of the spatial distribution of PM_2.5 and O₃ concentrations

The peak PM_2.5 concentration generally occurs from 08:00 to 10:00 Beijing Time (BJT) due to rush hour. Figure 7 provides the distributions of calculated and observed near-surface PM_2.5 concentrations along with the simulated wind fields at 08:00 BJT from 23 to 28 May 2015. The calculated PM_2.5 spatial patterns generally agree well with the observations at the monitoring sites. The NCP experiences more severe PM_2.5 pollution than its surrounding areas, with PM_2.5 concentrations frequently exceeding 150 µg m⁻³ in the Beijing–Tianjin–Hebei region. During the study episodes, the pollutants are likely to be transported to NEC and NWC under the prevailing southwesterly or southeasterly winds in northern China, causing the PM_2.5 concentrations in most areas of NEC and NWC to be frequently higher than 75 µg m⁻³.

Figure 8 Same as Fig. 7, but for near-surface O₃ at 14:00 BJT.

The O₃ concentration during summertime generally reaches its peak from 14:00 to 16:00 BJT in northern China (Fig. 5). Figure 8 shows the spatial distribution of calculated and measured near-surface O₃ concentrations at 14:00 BJT from 23 to 28 May 2015, along with the simulated wind fields. Generally, the simulated O₃ spatial patterns are consistent with the observations, but the model overestimation or underestimation still exists. The simulated high O₃ concentrations at 14:00 BJT, exceeding 200 µg m⁻³, are frequently concentrated in the NCP, which is also consistent with the measurements. The O₃ transport to NEC and NWC from the NCP is obvious when the winds are southeasterly or southwesterly, inducing severe O₃ pollution in NEC and NWC.

In general, the simulated variations of PM_2.5, O₃, NO₂, and aerosol species are in good agreement with observations, indicating that the simulations of meteorological conditions and chemical processes and the emission inventory used in the WRF-CHEM model are reasonable, providing a reliable basis for the further investigation.

3.3 Effects of NCP emissions on the air quality in NEC and NWC

To evaluate the contribution of NCP emissions to the air quality in its neighboring areas, four model simulations are performed, including f_NS with all anthropogenic emissions from the NCP and non-NCP areas, f_N with anthropogenic emissions from the NCP only, f_S with anthropogenic emissions from the non-NCP areas only, and f₀ without all anthropogenic emissions. Consequently, the air pollutant concentrations in NEC and NWC can be separated into four components, including contributions from the local emissions ($f_{\mathrm{S}}^{\prime }$, f_S−f₀), the trans-boundary transport of NCP emissions ($f_{\mathrm{N}}^{\prime }$,f_N−f₀), the interactions of these two emissions ($f_{\mathrm{NS}}^{\prime }$, $f_{\mathrm{NS}}-f_{\mathrm{N}}-f_{\mathrm{S}}+f_{\mathrm{0}})$, and the background (f₀).

In the present study, the effect of NCP emissions on the PM_2.5 and O₃ concentrations in NEC and NWC is evaluated, considering that they have a long lifetime of several days in the troposphere and often constitute the primary air pollutant during summertime (Seinfeld and Pandis, 2006). However, the trans-boundary transport of PM₁₀ is not considered due to its short lifetime of several hours, caused by the dry deposition and gravity, and the fact that PM₁₀ is generally confined to its source region when the wind is not strong enough (Sun et al., 2006).

Figure 9 Contributions of NCP emissions to the daily mean near-surface PM_2.5 concentration in NEC and NWC from 23 to 28 May 2015.

3.3.1 Contributions of NCP emissions to PM_2.5 levels in NEC and NWC

Figure 9 shows the simulated spatial distribution of daily mean PM_2.5 concentrations contributed by NCP emissions in NEC and NWC from 23 to 28 May 2015. The contribution of trans-boundary transport of NCP emissions to the PM_2.5 concentration is remarkable in Liaoning, frequently exceeding 30 µg m⁻³ in most areas of the province during the episode. NCP emissions also considerably influence the PM_2.5 concentration in Jilin, contributing 5–30 µg m⁻³ in most areas and occasionally exceeding 40 µg m⁻³. The effect of NCP emissions on the PM_2.5 level in Shanxi and Shaanxi is increasingly evident from 23 to 28 May 2015, with the contribution of up to 50–60 µg m⁻³ in the southeast of Shanxi and to a lesser extent of 30–40 µg m⁻³ in the middle part of Shaanxi on 27–28 May. The contribution of trans-boundary transport of NCP emissions to the PM_2.5 level in Inner Mongolia is not significant, which may be attributed to the location of the low pressure and terrain characteristics. Obviously, the effect of trans-boundary transport shows a stepwise characteristic: the closer to the NCP emission sources, the more remarkable the impact on the downwind areas. As a result, Liaoning and Shanxi provinces are substantially influenced by NCP emissions, while Jilin and Shaanxi provinces are affected to a lesser extent.

Table 2 Daily average PM_2.5 contributions (µg m⁻³) of NCP emissions in NEC and NWC from 22 to 28 May 2015.

The impact of NCP emissions on the daily average PM_2.5 concentration in NEC and NWC from 22 to 28 May 2015 is summarized in Table 2. On average, NCP emissions increase the PM_2.5 concentrations by 24.2, 9.6, 13.9, 6.5, and 2.6 µg m⁻³ in Liaoning, Jilin, Shanxi, Shaanxi, and Inner Mongolia, with average percentage contributions of 40.6, 27.5, 32.2, 20.9, and 16.7 %, respectively. Figure 10 shows the episode-averaged PM_2.5 percentage contribution from NCP emissions to the surrounding areas. The NCP emissions markedly affect the air quality in Liaoning, accounting for around 20–50 % of the PM_2.5 concentration during the episode and with the most substantial impact on the west part of the province. The NCP emissions contribute about 15–30 % of the PM_2.5 concentration in Jilin. Shanxi province is also remarkably affected by NCP emissions, with more than 25 % of PM_2.5 concentration contributed by NCP emissions in most areas. Although Shaanxi province is a little far from the NCP, NCP emissions still contribute about 10–35 % of the PM_2.5 concentration. NCP emissions also enhance the PM_2.5 concentration by 5–50 % in the southern edge of Inner Mongolia, which is adjacent to the NCP.

Figure 10 Average percentage contribution of NCP emissions to PM_2.5 concentrations in NEC and NWC from 22 to 28 May 2015.

Figure 11 Same as Fig. 9, but for the afternoon (12:00–18:00 BJT) O₃ concentration.

3.3.2 Contributions of NCP emissions to O₃ concentrations in NEC and NWC

Figure 11 shows the simulated spatial distribution of the average afternoon O₃ concentrations contributed by NCP emissions from 23 to 28 May 2015. Similar to the PM_2.5 case, the contribution of NCP emissions to the O₃ formation in Liaoning and Jilin province is increasingly enhanced during the episode (except on 26 May), and on 25 and 27 May, NCP emissions account for more than 70 µg m⁻³ of the O₃ concentration in most areas of Liaoning. On 25 and 28 May, NCP emissions contribute more than 70 µg m⁻³ of the O₃ concentration in some regions of Jilin. A lesser impact of NCP emissions on Jilin province on 26 May is due to the weakening of the low pressure. The NCP emissions play a progressively important role in O₃ concentrations in Shanxi and Shaanxi provinces during the episode, especially on 27 and 28 May when the contribution can be up to 60 µg m⁻³. The impact of NCP emissions on O₃ concentrations in Inner Mongolia is insignificant overall.

Table 3 Daily afternoon (12:00–18:00 BJT) average O₃ contributions (µg m⁻³) of NCP emissions in NEC and NWC from 22 to 28 May 2015.

Table 3 summarizes the effects of NCP emissions on the average afternoon O₃ concentration in NEC and NWC from 22 to 28 May 2015. During the episode, NCP emissions substantially influence the O₃ level in Liaoning province, and the afternoon O₃ contribution is about 46.5 µg m⁻³ on average, ranging from 23.9 to 69.5 µg m⁻³. The NCP emissions also contribute an average of 28.7 µg m⁻³ to the O₃ concentration in Jilin province, varying from 12.4 to 45.7 µg m⁻³. The contribution of NCP emissions to Shanxi and Shanxi provinces becomes increasingly significant during the episode, with an average of 35.1 µg m⁻³ for Shanxi province and 20.7 µg m⁻³ for Shaanxi province, respectively. The O₃ concentration in Inner Mongolia is less influenced by NCP emissions, with an average of 8.4 µg m⁻³. Figure 12 illustrates the episode-averaged afternoon O₃ percentage contribution of NCP emissions to the surrounding areas. In NEC, NCP emissions account for 15–35 % of the afternoon O₃ concentration in most areas of Liaoning province and 10–30 % in Jilin province. In NWC, NCP emissions contribute 10–35 % of the O₃ concentration in Shanxi province and 10–25 % in Shaanxi. In Inner Mongolia, the impact of NCP emissions on O₃ formation is small, generally less than 15 %, except in the southern area adjacent to the NCP and Liaoning province, where a contribution of more than 10 % is found. On average, NCP emissions distinctly increase the afternoon O₃ concentrations in Liaoning, Jilin, Shanxi, Shaanxi, and Inner Mongolia, with average percentages of 27.4, 19.5, 21.2, 15.8, and 8.0 %, respectively (Table 3).

Additional sensitivity studies have also been performed to examine the potential influences of the cumulus parameterization on evaluation of the contribution of NCP emissions to the PM_2.5 and O₃ concentrations in NEC and NWC, in which the cumulus parameterization is turned off. The difference of the contribution of NCP emissions to the PM_2.5 and O₃ concentrations in NEC and NWC is less than 0.8 % between the simulations with and without the cumulus parameterization. Furthermore, it is worth noting that uncertainties from meteorological field simulations, emission inventories, and the chemical mechanism used in simulations have a large potential to influence the evaluation of the effect of NCP emissions on the PM_2.5 and O₃ concentrations in NEC and NWC (Carter and Atkinson, 1996; Lei et al., 2004; Song et al., 2010; Bei et al., 2017).

4 Summary and conclusions

Analyses of the synoptic pattern during the ASM season show that the southeasterly–southwesterly winds prevail in northern China, facilitating the trans-boundary transport of air pollutants from the NCP to NEC and NWC. The good relationships of PM_2.5 and O₃ concentrations in the NCP with those in NEC and NWC during the ASM season also indicate the possibility that the air quality in NEC and NWC is influenced by the trans-boundary transport of air pollutants originated from the NCP.

A widespread and severe pollution episode from 22 to 28 May 2015 in northern China is further simulated using the WRF-CHEM model to investigate the impact of the trans-boundary transport of NCP emissions on PM_2.5 and O₃ concentrations in NEC and NWC, when the region is affected by prevailing southeasterly–southwesterly winds associated with the ASM.

In general, the WRF-CHEM model reproduces the temporal variations and spatial distributions of PM_2.5, O₃, and NO₂ concentrations well compared to observations in northern China, although the model biases still exist due to the uncertainties in simulated meteorological fields and the emission inventory. The model also performs reasonably well in simulating the variations of aerosol constituents against the ACSM measurement at the NCNST site in Beijing.

Figure 12 Same as Fig. 10, but for the afternoon (12:00–18:00 BJT) O₃ concentration.

The FSA is used to investigate the contribution of trans-boundary transport of NCP emissions to O₃ and PM_2.5 levels in NEC and NWC. Model results show that NCP emissions contribute approximately an average of 24.2 and 13.9 µg m⁻³ to the PM_2.5 concentration in Liaoning and Shanxi during the episode, with average percentage contributions of 40.6 and 32.2 %, respectively. The NCP emissions enhance the PM_2.5 level by 9.6 and 6.5 µg m⁻³ in Jilin and Shaanxi on average, with percentage contributions of 27.5 and 20.9 %, respectively. The NCP emissions also substantially influence the O₃ concentration in NEC and NWC. The NCP emissions increase the afternoon (12:00–18:00 BJT) O₃ concentration in Liaoning by 46.5 µg m⁻³ on average during the episode, followed by 35.1 µg m⁻³ in Shanxi, 28.7 µg m⁻³ in Jilin, and 20.7 µg m⁻³ in Shaanxi, with average percentage contributions of 27.4, 21.2, 19.5, and 15.8 %, respectively. In contrast, the contribution of the trans-boundary transport of NCP emissions to the PM_2.5 and O₃ concentration in Inner Mongolia is lower, with an average of 2.6 and 8.4 µg m⁻³, respectively. Our results demonstrate that when southerly winds are prevailing in northern China, air pollutants originated from the NCP are likely to be transported northwards and profoundly affect the air quality in NEC and NWC. Stringent control of NCP emissions not only mitigates local air pollution, but is also beneficial for the air quality in NEC and NWC during the ASM season.

It is worth noting that interactions between the air pollution in China and ASM are two-way and their relationships are complicated and interrelated, especially with regard to the aerosol–meteorology interaction. Aerosol impacts on meteorology are significant due to its direct and indirect effects, which further influence the air pollution conditions in the lower troposphere. The aerosol semi-direct effect induced by the light-absorbing aerosols in the atmosphere stabilizes the PBL and thus reduces the PBL height to exacerbate the accumulation of air pollutants within the PBL, particularly for the aging process of black carbon which considerably enhances light absorption (Wang et al., 2013; Khalizov et al., 2009; Peng et al., 2016). In addition, aerosol plays an important role in the process of cloud formation and precipitation via acting as cloud condensation nuclei (CCN) and ice nuclei (IC). Therefore, aerosol–cloud interactions modify temperature and moisture profiles and influence precipitation, leading to potential feedback on the atmospheric chemistry (Wang et al., 2011). In addition, the ASM substantially influences the spatial characteristics of the air pollutants' transport and distribution in eastern China on seasonal, interannual, and decadal scales (Wu et al., 2016). Further studies need to be performed to investigate the impacts of the ASM variation on the air pollutants' transport, which is modulated by climate changes.

Although the model performs well in simulating PM_2.5, O₃, and NO₂ during the episode in northern China, the uncertainties from meteorological fields and the emission inventory still exist. Future studies need to be conducted to improve the WRF-CHEM model simulations and to further assess the contributions of the trans-boundary transport of NCP emissions under specific synoptic patterns, considering the rapid changes in anthropogenic emissions, which are not reflected in the present study. Therefore, more episode simulations during the ASM season should be performed to comprehensively evaluate the contribution of the trans-boundary transport of NCP emissions to the air quality in its downwind regions and support the design and implementation of effective emission control strategies.