Characteristics of Haze Pollution Episodes and Analysis of a Typical Winter Haze Process in Shanghai

One-year field campaign was conducted from July 2013 to August 2014 at the site of East China University of Science and Technology (ECUST) in urban Shanghai, and mass concentrations and chemical compositions of PM2.5 were measured. Gaseous pollutants (SO2, NO2) and meteorological parameters (wind speed, wind direction, pressure, temperature and relative humidity) were simultaneously obtained. In this study, PM2.5 mass balances on haze and non-haze days were reconstructed and the sum of secondary inorganic aerosols (SIA) and organic matter (OM) accounted for over 80%. The fraction of nitrate in SIA was much higher on haze days than that on non-haze days, while the corresponding fraction of ammonium was lower, implying that the variations of the sources and formation processes of SIA on haze days. In theory, sulfate and nitrate might be almost fully neutralized by ammonium. Moreover, the sulfur oxidation ratio (SOR) values were much higher than the nitrogen oxidation ratio (NOR) values, indicating the greater oxidation capacity of SO2 would occur. On haze days, the high NOR values could be explained by the relatively low temperature, the high NO2 concentration and the potential dominant gas-phase reaction. As for secondary organic aerosol (SOA), the formation processes were usually associated with nitrate formation. In winter, haze pollution episodes occurred more frequently than those in other seasons, associated with the different features of wind speed, wind direction and 72-h backward trajectory. In addition, one case from 17 November 2013 to 4 January 2014 was selected to investigate the formation mechanism of haze pollution episodes. The key factors that affected the haze formation might be the local stable synoptic conditions including weak surface wind, surface temperature inversion and high relative humidity, the long-range transportations from the Northwest and the large amounts of emissions from local sources.


INTRODUCTION
With the rapid industrialization and urbanization, China is facing serious air pollution problems, especially aerosol pollution.Airborne aerosols, such as sulfate, nitrate, ammonium, organic matter (OM), black carbon (BC) and other chemical compositions, can scatter or absorb the incident light, leading to degradation of atmosphere opacity and horizontal visibility and haze formation.In the urban area, haze episodes usually accompany pollution of fine particles (particulate matter with aerodynamic diameter below 2.5 µm, PM 2.5 ), and occurred frequently in China in recent years, in particular concentrated in the megacity clusters such as the Yangtze River Delta Region (YRDR), the Pearl River Delta Region (PRDR) and the Circum-Bohai-Sea Region (CBSR).Moreover, haze pollution has been studied intensively on its impacts on air quality, climate and public health (Kim et al., 2008;Bell et al., 2011;Nguyen and Leelasakultum, 2011).The major factors of haze formation were stable anti-cyclone conditions, decreasing of the height of planetary boundary layer (PBL), the hygroscopic growth of aerosols and heavy anthropogenic emissions in North China (Liu et al., 2013;Zhao et al., 2013).By comparing different chemical characteristics of aerosols on haze and nonhaze days, Wang et al. (2006a) pointed out that (NH 4 ) 2 SO 4 , NH 4 NO 3 , and Ca(NO 3 ) 2 were the main species on spring haze days in Beijing.Meanwhile, (NH 4 ) 2 SO 4 and NH 4 HSO 4 were also most responsible for summertime haze in the mid-Atlantic region (Chen et al., 2003), whereas secondary organic aerosols (SOA) and secondary inorganic aerosols (SIA) were found to be of similar importance during a winter severe haze episode in China (Huang et al., 2014).
Shanghai is the core city of the YRDR and adjacent to the East China Sea, which concentrates many kinds of heavy industry.Although Shanghai local government had invested a lot of money into environmental pollution control in the past decades (such as nearly 684 million dollars in 2014), haze pollution is still one of the biggest air pollution problems.Nearly 34% days were polluted days in 2013 based on announcement by Shanghai Environmental Protection Bureau.In this case, severe haze pollution in Shanghai has attracted great attentions of scientists to investigate its formation mechanism and effects in the past decades.Recently more publications began to concern the formation process of haze episodes.Wang et al. (2014) used Particulate Matter Source Apportionment Technology to quantify the impacts of sources on fine particulate matter during heavy haze episodes in late autumn, and found both local contribution and regional transport contribution could lead to a haze event in Shanghai, which was dominant depending on the meteorological conditions.Based on aerosol monitoring, satellite and lidar observations, Huang et al. (2012) identified three typical haze types in Shanghai, secondary inorganic pollution, dust and biomass burning.The contribution of biomass burning to haze formation in YRDR was also demonstrated by Cheng et al. (2014).Wang et al. (2015) emphasized a more contribution from traffic in Shanghai to haze formation than that in Beijing.In addition, there are also some studies about chemical characterization and sources of particles on haze days in Shanghai (Yang et al., 2005;Lin et al., 2014;Behera et al., 2015;Zhang et al., 2015).
Haze episode is a complicated process, especially in Shanghai with large amounts of anthropogenic sources and complicated meteorological conditions.For some special episodes, the haze might be dominated by the specific meteorological conditions which could accelerate the occurrence of secondary aerosol reactions.We conducted a one-year filed campaign in urban Shanghai and analyzed PM 2.5 with its major components.In particular, the relative contributions and sources of secondary aerosols on haze days were explored.Furthermore, one severe haze event associated with the meteorological conditions and PM 2.5 compositions was investigated to reveal its formation mechanism.

Field Observation Sampling Site and Collection
The sampling site is on the roof top of a building in East China University of Science and Technology (ECUST) (N31°8'43", E121°25'31") (Fig. 1), around 15 m above ground, which is located in the south western part of Shanghai and surrounded by a residential area with few major anthropogenic industrial sources.
From July 2013 to August 2014, 24-h PM 2.5 samples were collected every third day using a five-channel Spiral Ambient Speciation Sampler (SASS, Metone Instrument Inc., OR, USA) at a flow of 6.7 L min -1 .The first channel was used to collect PM 2.5 with a 47 mm Telfon filter (PALL Life Sciences, MI, USA) for PM 2.5 masses and elements analysis.The second channel collected the particle for water-soluble inorganic ions analysis with a 47 mm pre-cleaned Nylon filter (PALL Life Sciences, MI, USA) preceded by a MgOcoated denuder.The last three channels were used to collect PM 2.5 for carbonaceous components on 47 mm quartz fiber filters (PALL Life Sciences, MI, USA).Before sampling, the fresh quartz filters were treated on pre-fired backed at 600°C for 4 h to remove any volatile components.Once sampled, filters were stored at -18°C in a freezer prior to chemical analyses.Field blanks were collected at the end of each month and were analyzed in parallel to the exposed filter samples as a part of QA/QC procedure.
PM 2.5 mass concentrations were determined gravimetrically by weighing the Telfon filters before and after sampling using an electronic microbalance with 0.01 mg sensitivity (Sartorius, Germany).The difference among three weighs of each filter was below 0.02 mg.PM 2.5 masses were deduced from the gravimetric measurements done before and after sampling and the corresponding concentrations were equal to masses divided by the sampled air volumes.Prior to weighting, the Telfon filters were equilibrated for at least 24 h at temperature of (25 ± 1)°C and relative humidity (RH) of (40 ± 5)%.

Automatic Particulate Matter and Gases Monitor
The daily concentrations of ambient air pollutants including PM 2.5 , PM 10 , SO 2 and NO 2 were obtained from the nearest automatic air quality monitoring station approved by Chinese Environmental Protection Agency, Xuhui Station (Fig. 1) (data from http://www.semc.com.cn/home/index.aspx).

The Meteorological Data
The meteorological data of Hongqiao Airport including wind speed, temperature, relative humidity (RH), pressure and visibility were obtained from Weather Underground (data from http://www.wunderground.com/).Wind speed data were obtained at 00:00, 06:00: 12:00 and 18:00 LTC, and other parameters were daily data.The Hongqiao airport is nearly 8 km far away from sampling site.

Carbonaceous Components Analysis
Organic carbon (OC) and elemental carbon (EC) were analyzed by a Desert Research Institute (DRI) Model 2001 carbon analyzer (Atmoslytic Inc., Calabasas, CA, USA), following the IMPROVE TOR protocol (Chow et al., 2007), and more details were reported in our previous papers (Zhao et al., 2015a, b) Half of each quartz filter was analyzed by a Total Organic Carbon (TOC) analyzer (VCPH-TOC, SHIMADZU, Japan) for water-soluble organic carbon (WSOC).The pretreatment process of each sample for WSOC was the same as that for water-soluble inorganic ions.
The triplicate analyses of laboratory standards showed that both analytical uncertainties for TC (TC = OC + EC) and WSOC were within 5%.

Element Analysis
Half of each Telfon filter was digested with 7 mL concentrated HNO 3 and 1 mL concentrated H 2 O 2 in a highpressure Telfon vessel by a microwave digestion system.The microwave-assisted extract was performed at 180°C for 20 min at 800 W. After extraction, the mixture was heated progressively to 130°C in a heating block until completely dry and diluted to 20 mL with ultrapure water (18 MΩ cm).Element of Al was analyzed by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS, X series, Thermo, USA).For quality control, the recovery of each fraction was in the range of 80-120%.Table 1 presents the detection limit (S/N = 3) of each species analyzed in the study.

Trajectories Calculation and PSCF Model
72-h backward air trajectories arriving at the sampling site were calculated using the HYSPLIT model.Meteorological data came from NCEP/NCAR Reanalysis meteorological database.In the study, the model was run at 00:00, 06:00, 12:00, 18:00 LTC and the arrival height was set at 500 m.
PSCF model is a method to identifying regional sources based on the HYSPLIT model.The zone of concern is divided into i × j small equal grid cells.The PSCF values were calculated by Eq. ( 1).The grid covered area of (10-55)°N and (80-135)°E was included and the accuracy of each grid was 0.5° × 0.5°.
where n ij was the total number of endpoints that pass through the grid cell (i, j) and m ij was the number of endpoints related to the "polluted" trajectories in the same grid cell.In the study, the 75 th percentiles of SO 2 and NO 2 were treated as the "polluted" threshold (Waked et al., 2014).Moreover, the weighting function W ij could be defined to reduce the effects of artifacts usually linked to the small numbers of endpoints.

General Descriptions of PM 2.5 Compositions
Haze pollution episodes were defined according to "Observation and forecasting levels of haze" (QX/T 113-2010).Among all 136 sampling days, 48 haze days were found.Table 2 presents the levels of PM 2.5 , its compositions and other air pollutants on haze and non-haze days.The average PM 2.5 concentration on haze days was (80.56 ±  59.90) µg m -3 , about 2.5 times as high as that on non-haze days ((33.41 ± 18.00) µg m -3 ), and the results of t-test (p < 0.01) also suggested the significant difference.The maximum daily concentration during the sampling period was 289.60 µg m -3 obtained on December 5, 2013 which was the serious haze day.Meanwhile, the higher concentrations of SO 2 and NO 2 on haze days were also obtained.
The chemical compositions of PM 2.5 were divided into six major parts as follows: SIA, SOA, primary organic aerosols (POA), EC, sea salt aerosols and mineral matter.SIA were defined as the sum of three secondary inorganic ions (sulfate, nitrate and ammonium).The concentration of SOC was calculated by using EC as the tracer from the following equation (Castro et al., 1999): (3) The minimum OC/EC ratio in each season was obtained by subtracting the special values for the days where rain and blowing dust or others caused significant changes to the ratios (Lim and Turpin, 2002).SOA could be estimated by applying a multiplicative factor of 1.6 to SOC for urban aerosols (Gilardoni et al., 2007), so was OM from OC. POA concentration was given by OM minus SOA.Sea salt aerosols and mineral matter were calculated based on the following equations (Chan et al., 1997): Fig. 2 shows the relationship between measured PM 2.5 and the sum of above six major parts.A good linear correlation (R 2 = 0.919) could reflect the accuracy of reconstruction.The gap between the sum of reconstructed PM 2.5 compositions and gravimetric mass was marked as "others".
As shown in Fig. 3, SIA were the most abundant components, accounting for about 50% of PM 2.5 mass on both haze and non-haze days, whereas the fractions of sulfate, nitrate and ammonium in SIA on non-haze and haze days were somewhat different.The fraction of nitrate was much higher on haze days than that on non-haze days, while the corresponding fraction of ammonium was lower, implying that the changes in the sources and formation processes of SIA on haze days.Nitrate could be a marker of mobile sources (Xiu et al., 2004;Wang et al., 2015), and the increase of nitrate indicated the enhanced contributions of vehicle emissions on haze days.In addition, OM was also an important component, and the proportions of SOA and POA were similar in OM.Compared with non-haze days, the decreasing contributions of sea salt aerosols and the increasing contributions of mineral matter were found on haze days.Furthermore, the severe haze pollution events were driven to a large extent by secondary aerosol formations and Huang et al. (2014) found that secondary aerosols contributed about 30-77% of PM 2.5 mass in four Chinese cities (Beijing, Shanghai, Guangzhou and Xi' an) in January 2013.

Secondary Inorganic Aerosols
Ammonia (NH 3 ) was important alkaline gas and could neutralize acidic sulfate and nitrate by gas phase and aqueous phase reactions.(NH 4 ) 2 SO 4 was preferentially formed and the least volatile and NH 4 NO 3 was formed next (Zhang et al., 2008;Li et al., 2013).In PM 2.5 , the molar equivalent ratio of ammonium to the sum of sulfate plus nitrate (e.g., )) which was more than 1 indicate that sulfate and nitrate was almost fully neutralized (Pathak et al., 2009).In the study, the average ratios were 1.31 and 1.70 on haze and non-haze days, respectively, different from some previous studies in Shanghai (Zhao et al., 2015a).It was suggested that the compositions of SIA had changed, especially sulfate and nitrate decreased significantly due to some efforts to control the gaseous precursor emissions such as SO 2 and NO 2 .On the contrary, the increasing trend of ammonium was clearly observed in recent years (Lin et al., 2014).Meanwhile, the larger ratios were mainly obtained in spring and summer.In these seasons, less emission from coal combustion might be responsible for the larger ratios and the concentrations of ammonium which originated from animal farming, fertilizers and organic decomposition were relatively higher.In this case, (NH 4 ) 2 SO 4 and NO 3 NH 4 were dominant forms of SIA in the ammonium-poor environment (Qiao et al., 2015).But the ratios on haze days  were lower than those on non-haze days, which meant that the partial neutralization of acidic aerosols by ammonia was greater during haze episodes.Increased mineral matter might be might be an explanation of the lower ratios on haze days.In addition, the average cations to anions ratio measured in the study, i.e., ( ), was above 1 during the whole period, while the significant lower level found during haze episodes, especially in winter, revealed the acidity of aerosols became strong in this environment.SO 2 and NO 2 , major criteria air pollutants, were also the precursors of sulfate and nitrate.In ambient air, SO 2 and NO 2 emitted from primary sources could react with the oxidants (mainly •OH and O 3 ) and then formed sulfate and nitrate.Fig. 4 shows the potential source location maps for high concentrations of SO 2 and NO 2 in Shanghai.The PSCF results were similar and the source areas both located in the YRDR.The air mass from the Northwest was the dominant long-range transportation, whereas the transportation intensity of SO 2 was lower than that of NO 2 .Some effective efforts to control coal consumption in the northwestern areas could lead to the significant decrease of SO 2 .In addition, sulfur oxidation ratio (SOR) and nitrogen oxidation ratio (NOR) could be indicators of the secondary transformation processes (Squizzato et al., 2013).The SOR and NOR values could be calculated by the following equations: where n was the molar concentration.The average SOR value of 0.24 was much higher than the NOR value of 0.09, suggesting the greater oxidation capacity of SO 2 .In this case, emitted SO 2 might be quickly transformed into sulfate along the pathway, which could explain the northwestern area had less influence on SO 2 than that on NO 2 shown in Fig. 4 Moreover, the SOR values were similar on haze and nonhaze days, both higher than 0.20, showing that SO 2 was universally oxidized in the atmosphere (Lyu et al., 2015).But the higher average NOR value of 0.13 on haze days suggested greater oxidation of NO 2 and more nitrate could exist in the atmosphere, which were explained by the specific meteorological conditions and possibly increased oxidation capacity of haze episodes.Squizzato et al. (2013) reported that low temperature and high relative humidity might be favorable to gas-to-particle conversion processes of NO 2 , in particular for ammonium nitrate formation.Considering the ammonium-rich environment during the sampling period, ammonium neutralized nitrate fully and the gas-phase reaction NH 3(g) + HNO 3(g) ⇌ NH 4 NO 3(s) became dominant.Hence a negative significant correlation between temperature and NOR (r = -0.497,p < 0.01) was observed in the study.

Secondary Organic Aerosols
As calculated in Section 3.1, the estimated mean SOA concentrations were (12.99 ± 10.22) µg m -3 and (5.25 ± 4.14) µg m -3 on haze and non-haze days, respectively.The contributions of SOA were lower than those of SIA and the average SOA/SIA ratio were 0.51 ± 0.55.No significant correlation (p < 0.01) between SOA and SOR was observed.Conversely, SOA had a positive correlation with NOR (r haze day = 0.561, r non-haze day = 0.285, p < 0.01).That indicated the formation processes of SOA might be associated with photochemistry of NO x , or possibly SOA precursors and NO x had similar sources on haze days.In theory, SOA were formed through the gas to particle partitioning of volatile organic compounds (VOC s ) while nitrate could be a marker of mobile sources, therefore it could be concluded that vehicle exhaust emission must played more important roles since the higher correlation of SOA with NOR was found during haze pollution episodes.The correlation coefficients between SOA and WSOC on haze and non-haze days were both higher 0.75.Generally, SOA contained polar functional groups such as hydroxyls, carbonyls and carboxyls generated by atmospheric oxidation (Saxena and Hildemann, 1996;Choi et al., 2012) and most SOA compounds were water soluble.Due to the solubility of most SOA compounds, they could participate in some aqueous phase aerosol reactions easily and were associated with hygroscopic growth of particles.

Seasonal Variations of Haze Pollution Episodes Associated with Meteorological Conditions
Haze pollution episodes occurred most frequently in winter and 60% sampling days were haze days including a severe haze event in Shanghai.In addition, the frequency of haze days in summer, autumn and spring was 22%, 37% and 29%, respectively, showing the seasonal feature of haze pollution episodes.
The frequency distributions of wind speed and direction show different seasonal features (Fig. 5).In winter and autumn, the dominant wind directions were both north, whereas the slower wind speeds were obtained in winter.In spring and summer, the southeastward winds occurred more frequently and the wind speeds were relatively higher.The results of 72 h backward trajectories shown in Fig. 6 were similar.The long-range air masses in winter and autumn originated from the North, especially some northwestern areas, which could carry the anthropogenic pollutants to affect the air quality in Shanghai.And the southeastward marine trajectories were the most frequent in summer.In spring, the trajectories came from different areas such as the Northwest and the East.In addition, some stable meteorological conditions (low surface wind speed and even temperature inversion) in winter were favorable to haze formation in Shanghai (Qiao et al., 2015).Overall, the frequency of haze days was closely related to the meteorological conditions in different seasons.

Analysis of a Typical Haze Pollution Episode
It was mentioned that the regional severe haze episodes occurred in the YRDR in winter during the sampling period.One case from 17 November 2013 to 4 January 2014 was analyzed to investigate the characterization of PM 2.5 and the haze formation mechanism in Shanghai.During this process, all 17 sampling days were conducted and 15 haze days were defined, except 29 November and 17 December, 2013.
The visibility was relatively low with the mean value of 3.09 km from 2 to 9 December, which was associated with low wind speed and pressure (red marked in Fig. 7).The surface pressure decreased gradually, similar with the trend of pressure in the North China Plain during a winter regional haze event (Zhao et al., 2013).The surface weather maps at 08:00 LTC from 30 November to 5 December were given in Fig. 8 (images form http://envf.ust.hk/dataview/hko_wc/current/).A surface high-pressure, which could lead to clear sky, subsidence airflow and relatively stagnant conditions suggested by Chen et al. (2003), persisted from 30 November to 2 December in Shanghai surroundings.On 3 December, high-pressure and low-pressure systems both affected Shanghai.However, Shanghai was again controlled by the high-pressure system from 4 December.This weather condition was unfavorable for the dispersion and accumulation of air pollutants and Xu et al. (2011) also concluded the winter haze pollution episodes were formed favorably under the cold high-pressure system.Meanwhile, the surface temperature inversion could be found from 2 to 9 December (data from http://weather.uwyo.edu/upperair/sounding.html),especially occurred commonly before 5 December, limiting the vertical dispersion of air pollutants.The relative humidity increased and the peak was on 6 December, which favored to hygroscopic growth of particles and might result in the abundance of fine mode particles.The higher PM 2.5 /PM 10 ratios (above 0.8) on these days also demonstrated it.Therefore, air pollutants emitted from local sources could accumulate, formed to particles (especially PM 2.5 ) via secondary aerosol reactions and then caused the formation and evolution of haze pollution episodes.And Wu et al. (2005) studied a severe haze episode over the Guangzhou region of China and also clarified that descending air motion and weak horizontal wind produced very high aerosol concentrations.
In addition, the mass ratio of NO 3 -/SO 4 2-was usually used as an indicator to suggest the relative importance of mobile versus stationary sources of sulfur and nitrogen in the atmosphere (Wang et al., 2006b).The higher ratios from 2 to 8 December (above 1.3) indicated that the local pre-pollutants of mobile sources were over the pollutants from the longrange transportations, whereas the ratios on 23 and 26 November were significantly lower, less than 0.80.Based on HYSPLIT model, 72-h backward trajectories from 22 to 26 November are shown in Fig. 9.The trajectory speed increased gradually and the transportation from the Northwest had more and more influence.On 26 November, the 72-h backward trajectory came from Mongolia to Shanghai, via Gobi Desert, Shanxi Province with large amounts of coal consumption and Jiangsu Province.The PM 2.5 /PM 10 ratio was only 0.34 and coarse mode particles were dominant.Furthermore, Al could be used as the reference element for the crustal material and the Al/PM 2.5 ratio was 0.84, much higher than the mean value of 0.37 during this haze process.The lowest NOR value was obtained, reflecting the atmospheric condition limited the formation of NO 2 .Thus, the haze pollution episode on 26 November was mainly affected by the long-range transportation carrying air pollutants.
The formation mechanism of this severe haze pollution episode in winter 2013 was complicated.The local stable synoptic conditions and long-range transportations both had effects.Apart from the meteorological conditions, the emissions from local sources also made important contributions.As presented in Table 2, the concentrations of most primary species such as EC, POA and Potassium were higher than those on other haze days.

CONCLUSIONS
In the study, a one-year filed campaign was carried out in urban Shanghai from July 2013 to August 2014.With the analysis of PM 2.5 compositions and other air pollutants, the comparison of characterization of aerosols on haze and non-haze days was discussed.Meanwhile, a typical winter haze pollution episode from the point of meteorological conditions and chemical characterization was investigated to clarify the formation mechanism.
The PM 2.5 concentrations on haze days were much higher than those on non-haze days, associated with air pollutants (SO 2 and NO 2 ) and PM 2.5 compositions.The  fraction of nitrate in SIA was much higher on haze days than that on non-haze days, while the corresponding fraction of ammonium was lower, implying that the changes in the sources and formation processes of SIA on haze days.Sulfate and nitrate were almost fully neutralized by ammonium in the study.SO 2 and NO 2 , the precursors of sulfate and nitrate, showed the similar source areas and transportation pathways.Meanwhile, the SOR and NOR values could indicate the conversion of SO 2 to SO 4 2-and of NO 2 to NO 3 -.The higher average SOR value (0.24) suggested the greater oxidation capacity of SO 2 than that of NO 2 .The NOR values on haze days were higher than those on non-haze days, which could be explained by the relatively low temperature, the high NO 2 concentration and the dominant gas-phase reactions on haze days.As for SOA, the formation processes were usually associated with nitrate formation.Haze pollution episodes occurred frequently in winter and the corresponding frequency was nearly 60%, much higher than those in other seasons.The different seasonal features of wind direction were found that the northward winds were dominant in autumn and winter, whereas the directions changed to southeast in spring and summer.Moreover, the slowest wind speeds were obtained in winter and the results of 72-h backward trajectories were similar with those of wind roses.One case from 17 November 2013 to 4 January 2014 was selected to investigate the formation mechanism.The formation of severe haze episodes was a combination of many factors.The local stable synoptic conditions including weak surface wind, surface temperature inversion and high relative humidity could cause the accumulation of air pollutants and lead to haze pollution episodes.Furthermore, the long-range transportations form the Northwest carrying large amounts of air pollutants also had significant effects.And the emissions from local sources also made important contributions.

Fig. 2 .
Fig. 2. Relationship between measured PM 2.5 and the sum of PM 2.5 major compositions (SIA, SOA, POA, EC, Sea salt aerosols and mineral matter).

Fig. 8 .
Fig. 8. Surface weather patterns at 08:00 (LTC) from 30 November to 5 December.Blue circle represents the low pressure center, red circle represents the high pressure center and blue dot denotes Shanghai.

Table 1 .
Detection limit of each species analyzed in the study.

Table 2 .
Summary of PM 2.5 , SO 2 , NO 2 and PM 2.5 compositions concentration