Science of the Total Environment 626 (2018) 519–527
Contents lists available at ScienceDirect
Science of the Total Environment
j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv
Characteristics and sources of aerosol pollution at a polluted rural sitesouthwest in Beijing, China
Yang Hua a,b, ShuxiaoWang a,b,⁎, Jingkun Jiang a,b, Wei Zhou a,b, Qingcheng Xu a,b, Xiaoxiao Li a,b, Baoxian Liu c,d,Dawei Zhang c,d, Mei Zheng e
a State Key Joint Laboratory of Environment Simulation and Pollution Control, School of Environment, Tsinghua University, Beijing 100084, Chinab State Environmental Protection Key Laboratory of Sources and Control of Air Pollution Complex, Beijing 100084, Chinac Beijing Municipal Environmental Monitoring Center, Beijing 100044, Chinad Beijing Key Laboratory of Airborne Particulate Matter Monitoring Technology, Beijing 100044, Chinae SKL-ESPC and BIC-ESAT, College of Environmental Sciences and Engineering, Peking University, Beijing 100871, China
H I G H L I G H T S G R A P H I C A L A B S T R A C T
• Four season organic aerosol mass spec-tra is obtained and source apportion-ment is conducted in the rural area inBeijing.
• Residential solid fuel burning is themost important source of aerosol pollu-tion in the rural area of Beijing.
• Results focusing on urban Beijing mighthave underestimate the contributionfrom residential emissions.
⁎ Corresponding author at: State Key Joint Laboratory oE-mail address: [email protected] (S. Wang).
Article history:Received 17 November 2017Received in revised form 5 January 2018Accepted 6 January 2018Available online 19 February 2018
Editor: Jianmin Chen
Annual average PM2.5 concentration in south Beijing was 30% and 40% higher than thewhole Beijing city in 2015and 2016, respectively. Few studies have been conducted to investigate what leads to the characteristics andsources of heavy pollution in the south rural area of Beijing. This study conducted an observation with AerosolChemical Speciation Monitor (ACSM) at a southwest rural site (Liulihe) in Beijing during 2014–2016, to investi-gate the seasonal aerosol characteristics and their sources. Positive matrix factorization (PMF) algorithm wasused to distinguish different components of organic aerosolmeasured by ACSM. Biomass burning is an importantpollution source,mainly due to the open burning after harvest season in autumn, regional transport in spring, andlocal residential biofuel use in winter. Coal consumption is the largest primary organic aerosol source in winter.Heavy duty diesel trucks contributed significantly to organic aerosol at night-time in the rural area. Results of thisstudy show residential solid fuel burning is themost important source of aerosol pollution in the rural area of Bei-jing and the results focusing on urban Beijing might have underestimate the contribution from residential emis-sions in the Beijing-Tianjin-Hebei region.
520 Y. Hua et al. / Science of the Total Environment 626 (2018) 519–527
1. Introduction
As the capital city of China, Beijing burdens 21.5 million residents,5.3 million vehicles (Beijing Municipal Bureau of Statistics, 2015) andthe corresponding high anthropogenic emissions of air pollutants. As aresult, Beijing has experienced serious air pollution and associatedhealth impact in recent years (Zheng et al., 2016; Zheng et al., 2015).In 2013, Chinese government released the Air Pollution Preventionand Control Action Plan (2013–2017) to improve the air quality, espe-cially in key regions including Beijing-Tianjin-Hebei area. Annual aver-age PM2.5 concentration in Beijing is targeted from 88 μg/m3 in 2013to 60 μg/m3 in 2017 (H. Zhang et al., 2016; J.K. Zhang et al., 2016). How-ever, the annual average PM2.5 concentration in Beijing was 81 μg/m3
with 179 polluted days in 2015. The average concentration decreasedto 73 μg/m3 with 169 polluted days in 2016, still facing challenges toreach the goal (Beijing Municipal Environmental Protection Bureau,2016; Beijing Municipal Environmental Protection Bureau, 2017). Airpollution in Beijing exhibits a remarkable spatial distribution character-istic of much higher concentration in south and lower concentration innorth. Annual average PM2.5 concentration in south Beijing was 30%higher than the whole city in 2015 (Beijing Municipal EnvironmentalProtection Bureau, 2016). This spatial distribution continued in 2016,with annual average PM2.5 concentration of the southwest site 40%higher than the whole city (Beijing Municipal EnvironmentalProtection Bureau, 2017). To reduce the pollution concentration of thewhole city, much more efforts need to be made on investigation of pol-lution characteristics and sources of the pollution in the south area ofBeijing. Considering that eight out of the most polluted ten cities inChina in 2015 located in south of Beijing (Ministry of EnvironmentProtection, 2016), regional transport contributed significantly to pollu-tion in Beijing (Beijing Municipal Environmental Protection Bureau,2014; Ji et al., 2014; Sun et al., 2015a, 2015b;Wang et al., 2015). As a re-sult, south area of Beijing might be impacted by the regional transportsignificantly. In addition to the regional transport impact, local emissionalso needs more investigation. Since the most polluted area in south ofBeijing are rural area, with different anthropogenic activities fromurban area, the local emission sources might be different. For example,household solid fuel use which was underestimated is proved to be amajor ambient pollution source recently (Liu et al., 2016).
Field observations have been carried out at urban sites in Beijing toexplore the pollution characteristics and sources. The design of AerosolChemical Speciation Monitor (ACSM) enables it easier to conduct long-term continues monitoring of non-refractory particulate matter withaerodynamic diameters smaller than 1 μm (NR-PM1), providing an ad-vanced technique to look into pollution sources and process (Ng et al.,2011). For example, pollution was characterized by high contributionof secondary species and oxygenated organic aerosol (OOA) from re-gional scale in summer in Beijing (Sun et al., 2012). In winter, coal com-bustion organic aerosol (CCOA)was resolved in several studies at urbansites (Sun et al., 2013; H. Zhang et al., 2016; J.K. Zhang et al., 2016; Elseret al., 2016). With the implementation of Air Pollution Prevention andControl Action Plan, coal has been replaced by gas energy for heatingseason in downtown of Beijing. The contribution from CCOA to NR-PM1 reduced from 17% (Sun et al., 2013) in 2011–2012 to 12% in 2014(H. Zhang et al., 2016; J.K. Zhang et al., 2016).
However, most of the previous studies were carried out at urbansites. These results are disable to explain the reason why pollution ismuch more severe in the rural area than the urban area in Beijing.Few investigations have been conducted in the rural area. To have an in-sight of the pollution characteristics and sources in the rural area, weconducted continuous sampling in four seasons at a rural site southwestin Beijing, which is the most polluted site in 2016 among all the sites inBeijing (BeijingMunicipal Environmental Protection Bureau, 2017). Theorganic aerosol mass spectra were obtained by ACSM, providing resultsof pollution characteristics and sources in the most polluted area inBeijing.
2. Field observation and analysis methods
2.1. Field observation site and sampling methods
The observation was conducted in four seasons (October 22nd toNovember 11th, 2014; March 30th to April 30th, 2015; August 11th toSeptember 7th, 2015; December 5th, 2015 to January 7th, 2016) from2014 to 2016.
The sampling site was located at a rural site (Liulihe site, 116°2′E,39°36′N, Fig. S1) in the southwest in Beijing. The site was located onthe border of Beijing and Hebei province. This site is in Fangshan Dis-trict, which is a heavy polluted region in Beijing. The sampling heightwas 3 m, on the roof of a one-storey sampling station which was 500m away from the traffic road. All the time discussed in this article islocal time.
The sampling instruments (Table S1) included weather station(WXT520, VAISALA, Finland), gaseous pollutants monitors (API100/200/400E, Teledyne, USA) and PM2.5/PM10 monitors (TEOM1405/1400a, Thermo Scientific, USA). Time resolution of these instrumentsis 5 min and averaged into 1 h.
ACSM was used to measure species including organic matter (OM),nitrate, sulfate, ammonium and chloride of NR-PM1 (Ng et al., 2011).The ACSMwas calibrated with Ionization Efficiency (IE), which was de-termined with SMPS. DMA (Differential Mobility Analyzer) was gener-ated to select NH4NO3 particles with a size of 300 nm mobilitydiameter and counted by CPC (Condensation Particle Counter). The re-sults were compared with ACSM data.
2.2. Back trajectory analysis and satellite data
HYSPLIT (Hybrid Single Particle Lagrangian Integrated Trajectory)model was used to analyze the regional transport. Trajstat, a Geograph-ical Information System (GIS)-based software into which the HYSPLITmodel was loaded (Wang et al., 2009) and used to calculate the backtrajectory. The model was run every 24 h and four staring height wereset to be 300m, 500m, 1000 m, and 1500 m above sea level. To investi-gate the air mass transport impact, the back trajectories were clustered.Euclidean distance mode was selected. The method is described inWang et al. (2009). The transport differences at different heights areprovided in supplement to explain the reason for height selection. Firepoint maps were obtained from https://firms.modaps.eosdis.nasa.gov/firemap/. Planetary boundary layer (PBL) height data were obtainedfrom the Global Data Assimilation System (GDAS) model (http://www.ready.noaa.gov/READYamet.php).
2.3. Data calibration and PMF analysis
Collection efficiency (CE) was used to calibrate the ACSM data tocompensate the particle loss. Based on the monitoring site condition,the following formula was used for calibration (Middlebrook et al.,2012).
CE ¼ max 0:45;0:0833þ 0:9167� ANMFð Þ
ANMF is characterized by the ammonium nitrated mass fraction(ANMF). Ionization Efficiency (IE) was determined with SMPS. DMA(Differential Mobility Analyzer) was generated to select NH4NO3 parti-cles with a size of 300 nmmobility diameter and counted by CPC (Con-densation Particle Counter). The results were compared with ACSMdata.
The NR-PM1 concentration (OM + SO42− + NO3
− + NH4+ + Cl−)
measured by ACSM tracks well with the PM2.5 concentration measuredby TEOM during the four seasons (Fig. S2). All the correlation coeffi-cients (R2) are above 0.60. It is noticed R2 are lowest in spring and theslope is 0.32, relatively much lower than other seasons and other stud-ies (Sun et al., 2012; Aurela et al., 2015). It is caused by the dust events in
521Y. Hua et al. / Science of the Total Environment 626 (2018) 519–527
spring during which PM10 concentration reached to 1000 μg/m3. Thedust led to significant increase of coarse particulate matters and refrac-tory chemical components. As a result, the correlation and slope arelowest in spring. If the dust episodes are neglected, the R2 and slopewill increase to 0.82 and 0.41, respectively (Fig. S2).
Positive matrix factorization (PMF) with the PMF2.exe algorithmwas used via PMF Evaluation Tool panel (PET v2.04) to distinguish dif-ferent components of OA measured by ACSM (Paatero and Tapper,1994; Ulbrich et al., 2009). Onlym/z b100was used for PMF consideringthe low signals above 100 and low signal-to-noise caused by low trans-mission efficiency (TE) (Ng et al., 2011). The factors were determinedfollowing the steps for choosing best solutions described in Zhang etal. (2011). The mass spectra of OA factors in different seasons areshown in supplement (Fig. S3).
(a) autu
(b) spr
(c) sum
(d) win
Fig. 1. Time series of OA factors and the corresponding t
Four OA factors are determined in autumn: hydrocarbon-like organ-ic aerosol (HOA), biomass burning oxygenated aerosol (BBOA), semivolatile oxygenated organic aerosol (SVOOA) and low volatile oxygen-ated organic aerosol (LVOOA). HOA online variation followed withNOx well (Fig. 1), with a Pearson correlation R2 of 0.46 (Sun et al.,2011). BBOA online variation followed with m/z 60 with a R2 of 0.52.The m/z 60 is an El fragment ion of levoglucosan, a marker for biomassburning (Alfarra et al., 2007; Schneider et al., 2006). Three factors areapportioned in spring: HOA, BBOA and OOA. HOA also correlated withNOx well with a R2 of 0.55 (Fig. 1). BBOA correlated with m/z 60 well.Meanwhile, themass spectra shows the BBOAwasmuch aged, especial-ly compared with the profile of fresh emitted BBOA (Alfarra et al., 2007;Schneider et al., 2006). In summer season, generally POA (primary or-ganic aerosol) concentration is low, meanwhile there was an emission
mn
ing
mer
ter
racer compounds at Liulihe site in the four seasons.
522 Y. Hua et al. / Science of the Total Environment 626 (2018) 519–527
control plan conducted for the national parade during the observation(Xu et al., 2017). As a result, the POA concentration during the controlplan was too low to be apportioned. PMF was run separately beforeand during the control plan. Two factors are found before the controlplan, one is POA and the other one is OOA. Although POA showed aHOA-like profile, it is uncorrelated with NOx (Fig. 1). Meanwhile, POAwas associated with m/z 60 (R2 = 0.20) (Fig. 1). This characteristicshows POA was a mixture of HOA and BBOA, both of which were toolow to be apportioned. Organic aerosol was highly aged in summer,only one factor is resolved during the control plan. Four factors are de-termined in winter: HOA, CCOA, BBOA and OOA. BBOA correlated withm/z 60 well with a R2 of 0.79 (Fig. 1). Considering coal combustion isthe dominant emission source of chloride, especially in winter, chlorideis used as a tracer of coal combustion here (Sun et al., 2013). CCOAhad astrong correlation with chloride (R2 = 0.78). Except the vehicle emis-sion, coal emission is an important source of NOx. When it comes intowinter, coal combustion for residential heating increases and NOx
can't be the tracer of HOA in winter.
3. Results and discussion
3.1. Pollutants concentration in different seasons
Seasonal average pollutants' concentration at Liulihe site is shown inTable 1. Concentrations of both gaseous and particle pollutants werehighest in winter season, especially for PM2.5, NO2 and CO. The averagePM2.5 concentration was 123, 89, 41 μg/m3 in the autumn, spring andsummer, respectively, and it increased to 278 μg/m3 inwinter. Similarly,the average CO concentration was 0.4–0.8 μg/m3 in the other seasonsand reached to 5.0 μg/m3 in winter. Meanwhile, NO2 concentrationwas 22–54 μg/m3 in spring, summer and autumn and increased to 93μg/m3 in winter. SO2 concentration was similar in all seasons, with arange of 4–14 μg/m3.
It is noticedNO2 and PM2.5 average concentration inwinter at Liulihesite were 93 μg/m3 and 278 μg/m3, respectively, much higher than thecity average concentrations which were only about 75 μg/m3 and 150μg/m3 in December 2015 (Table 1; Beijing Municipal EnvironmentalProtection Bureau, 2016). Meanwhile, the concentrations of gaseouspollutants at Liulihe site were only slightly higher than other urbansites in Beijing in spring, summer and autumn (Table 1, BeijingMunicipal Environmental Protection Bureau, 2015; Beijing MunicipalEnvironmental Protection Bureau, 2016). Sources at Liulihe site in win-ter might be different from those in urban areas. Recent studies havefound that NO2 plays an important role in sulfate formation in winterin Beijing (Xie et al., 2015; Cheng et al., 2016; Wang et al., 2016). SO2
is trapped by alkane aerosol components and oxygenated byNOx to sul-fate. The high neutralizing capability of atmosphere in Northern Chinasustains this reaction rate high. The important role of NO2 requiresmore investigation on the reason why the NO2 in the rural area ismuch higher. Previous emission inventory of NOx shows the largestsources of NOx in Jing-Jin-Ji region are coal combustion (50%), industrialprocessing (14%) and traffic (31%) (Zhao et al., 2012).
Fig. 2 shows the diurnal variation of different gaseous pollutants. Thediurnal variation of O3 was similar with seasons, showing a peak in theafternoon. SO2 concentration also shared similarities with all the sea-sons. SO2 exhibited a pronounced peak in the daytime in all the seasons.NO2 diurnal variation was similar in spring, summer and autumn,
Table 1Average pollutants concentration at Liulihe site in the four seasons.
SO2 (μg/m3) CO (mg/m3) O3 (μg/m3) NO2 (μg/m3) PM2.5 (μg/m3)
showing an increase at night and a decrease in the afternoon. However,the variation changed in winter with an increase around noon and a de-crease at night.
3.2. Source apportionment results andNR-PM1 chemical components in dif-ferent seasons
3.2.1. Source apportionment results and NR-PM1 chemical components inautumn
The average concentration of NR-PM1was53 μg/m3 in autumn. Inor-ganic aerosol and organic aerosol accounted for 44% and 56% of NR-PM1,respectively. Nitrate contributed themost among all the inorganic com-ponents, accounting for 21%. Sulfate and ammonium contributed 8% and11%, respectively, much lower than nitrate (Fig. 3). Table 2 comparesthe observation results of this study with those from other studies car-ried out at urban sites in Beijing. All the urban campaigns listed in theTable 2 used ACSM to obtain the organic aerosol mass spectra. Mean-while, all the source apportionment results in the Table 2were obtainedby PMFmethod. Therefore, the results from literatures and from this re-search used the same measurement and analysis method. The nitrateand ammonium contributions were nearly the same with the autumnobservation results reported in urban area in Beijing, while the sulfatecontribution was much lower than the reported results at the urbansite (16%). What's more, the organic matter contribution at Liulihe sitewas much higher than that at the urban site (47%).
Source apportionment in autumn results shows OOA accounted for60% of OMand POA accounted for 40%. Both BBOA andHOA contributed20% to organic aerosol. The results at urban site show BBOA accountedfor 10% of OA, which is much lower than that at Liulihe site. Exceptopen biomass burning after harvest in autumn, crop residues used inhousehold stoves for cooking is also an important emission source(8.44 kt PM2.5 in 2014, accounting for 12.6% among all the emissionsources) in China (Zhao et al., 2013; Cai et al., in preparation). This ex-plains why BBOA contributed much significantly at Liulihe site. HOAcontributed 20% to OM, comparable to the contribution at urban site.It is noticed COA (cooking organic aerosol) is not distinguished atLiulihe site while it contributed 17% at the urban site.
The diurnal variation of inorganic aerosol shows relatively stablecurve in the whole day except a small peak in the morning, from 8:00am to 10:00 am (Fig. 4). This diurnal variation is consistent to theLVOOA variation. HOA reached to a minimum value in the afternoonand kept highwith smallfluctuations fromevening to earlymorning, in-stead of two peaks in the rush hour. It might be caused by both rushhour traffic and diesel trucks for goods transportation. Transportationof goods to Beijing via heavy-duty diesel trucks (HDDT) is permittedat night only. Liulihe site is at the entrance of HDDT fromoutside into in-side of Beijing city. The BC emitted by HDDT is even more than the sumof all the other categories of vehicles at night (Zhang et al., 2017) so it isnot surprising HOA kept high at night at Liulihe site. BBOA had a signif-icant increase from 5:00 pm to 9:00 pm in the evening and a small peakin the morning from 7:00 am. COA isn't distinguished at Liulihe site.Usually COA will be apportioned in the urban area. As to the rural areawith much lower population density, COA concentration should bemuch lower and not easy to be apportioned. What's more, village-based inventory in Beijing shows biomass and coal are the dominatedfuel for residential cooking (Cai et al., in preparation). As a result, COAis emitted together with BBOA and CCOA, which makes it even harderto be identified. LVOOA diurnal variation was relatively flat in autumn,indicating photochemical reaction is not active in autumn.
3.2.2. Source apportionment results and NR-PM1 chemical components inspring
The average concentration of NR-PM1 was 31 μg/m3 in spring. Inor-ganic aerosol and organic aerosol accounted for 48% and 52% of NR-PM1,respectively (Fig. 3). Similarwith autumn, nitrate contributed 23%of theSIA (Secondary Inorganic Aerosol) species. Sulfate contribution kept the
Fig. 2. Diurnal variation of gaseous pollutants and RH at Liulihe site in the four seasons.
523Y. Hua et al. / Science of the Total Environment 626 (2018) 519–527
samewith autumnwhile ammonium contribution increases to 16%. Thelarge difference between rural site and urban site is still the sulfate andOM contribution. Sulfate contribution was much lower than the urbansite while OM contribution at Liulihe site was much higher than theurban site.
OOA contributed 43% and POA contributed 57% in spring. HOA con-tributed 11% to OM in spring. However, BBOA increased significantly to47% in spring. BBOA contribution in spring was the highest among allthe seasons. Meanwhile, mass spectra profile of BBOA in spring wasmuch different from the BBOA profile in the others seasons. The m/z44 fraction is 0.17, much higher than other seasons. It means BBOA in
Fig. 3. Average non-refractory PM1 chemical com
spring was much more aged than the other seasons, as well as muchaged than the fresh biomass burning emission profile (Grieshop et al.,2009). BBOA was more like to be transported to the site rather thanfrom local emission (Cubison et al., 2011). Cluster analysis was appliedto the back trajectories in spring. Four clusters are identified and the av-erage PM2.5 concentration corresponding to each cluster was calculated(Fig. 5). The highest PM2.5 concentration was corresponding to two di-rections. One is the northeast and the other one is the northwest. Ac-cording to the fire point map obtained from MODIS (Fig. 5), openbiomass burning was intensive in the northeast area. It is possibleaged BBOA was transported to the Liulihe site from northeast. OOA
a Results from literatures listed in the table are from J.K. Zhang et al. (2016) (autumn), Sun et al., 2015a, 2015b (spring), Sun et al., 2012 (summer) and J.K. Zhang et al. (2016) (winter).
524 Y. Hua et al. / Science of the Total Environment 626 (2018) 519–527
contributionwas 43% in spring. Considering the BBOA profile wasmuchaged, parts of the BBOA are also oxygenated. OOA contribution is rela-tively higher than 43%.
The diurnal variation of inorganic aerosol was evenmore stable thanautumn, especially for sulfate. Nitrate, ammoniumand LVOOA showed asimilar diurnal variation with autumn while sulfate showed littlechange (Fig. 4). This indicates sulfate might be from regional transport
(a) autumn (u
(b) spring (uni
Fig. 4. Diurnal variation of organic aerosol o
rather than local photochemical reaction, similarly with previous stud-ies (Sun et al., 2012). HOA kept high at night, similar with autumn. Amorning peak appeared at 7:00 am. BBOA diurnal variation showedno increase in themorning like the other seasons, indicating local emis-sion had little impact on BBOA in spring. BBOA was relatively stable,only showing a low concentration in the daytime and an accumulationat night, which was consistent to the aged BBOA profile.
525Y. Hua et al. / Science of the Total Environment 626 (2018) 519–527
3.2.3. Source apportionment results and NR-PM1 chemical components insummer
The average concentration of NR-PM1 in summer was 24 μg/m3. In-organic aerosol and organic aerosol contributed 42% and 58%,
(a) backward trajectory in spring during the observation
Fig. 5. Backward trajectory and fire points
respectively (Fig. 3). Different from autumn and spring, nitrate contrib-uted only 12% to NR-PM1, much lower than that reported at the urbansite (21%) in Beijing. Sulfate contribution increased significantly to14% and ammonium contribution changed little with a value of 16% in
(b) fire points map in spring during the observation
526 Y. Hua et al. / Science of the Total Environment 626 (2018) 519–527
summer. The contributions of sulfate and ammoniumwere similar withwhat reported at the urban site. Similar with autumn and spring, OMcontribution at Liulihe site was higher than the urban site (40%).
The diurnal variations of sulfate and ammonium showed littlechange in summer (Fig. 4). Nitrate showed two pronounced peaks,one was in the early morning at 4:00 am and the other one was at latenight at 11:00 pm. There are two major ways for the nitrate formation.One is gas phase homogenous reaction between ambient ammonia andnitric acid. The other is the heterogeneous hydrolysis reaction of N2O5
on the moist surface of pre-existing aerosols at night (Pathak et al.,2009). The night increase of nitrate was not consistent to the NO2 vari-ation. Considering the much higher RH in summer, the night formationmight be resulted from the heterogeneous hydrolysis reaction of N2O5
under high RH in summer (Pathak et al., 2009).Pollution emission control planwas carried out for Parade during the
summer observation. The average concentration of NR-PM1 decreasedfrom 46 μg/m3 before the control plan to 13 μg/m3 during the plan.Therefore, the source apportionment had to be done separately as twoperiods. OOA contribution reached to the highest in summer. Beforethe control plan, OOA contribution was 83% higher than the other sea-sons. POA contributed only 17%. It is noticed that HOA contributed36% at urban site, which is much higher than that at Liulihe site. The di-urnal variation of POA before control plan showed two peaks in the rushhour and a night increase resulting from HDDT. OOA diurnal variationshowed a much more significant increase in the afternoon, resultingfrom the active photochemical reaction in the summer afternoon. OOAdiurnal variation during the control plan was similar with it is beforethe plan.
3.2.4. Source apportionment results and NR-PM1 chemical components inwinter
The average concentration of NR-PM1 in winter was 122 μg/m3 inwinter. The inorganic aerosol and organic aerosol contributed 42% and58%, respectively (Fig. 3). Although nitrate concentration was thehighest in winter among all the seasons, the contribution was only10%, which was much lower than autumn (21%) and spring (23%).Thiscontribution was the same with the contribution reported at theurban site (10%). Sulfate and ammonium contributed 12% and 13%, re-spectively. Sulfate contribution was lower than its contribution at theurban site while ammonium was higher.
OOA contributed 31% to OM in winter, whichwas the lowest amongall the seasons. HOA accounted for 26% in winter at Liulihe site. BBOAcontribution was 13% in winter, a slightly higher than spring and sum-mer. There was no BBOA apportioned in the urban site (Table 1).CCOA concentration was the highest among POA species (29%),resulting from the heating in winter. The contribution was higher thanthe contribution at the urban site. Total contribution of BBOA andCCOA reached to 42%. Both of them were mainly from householdcooking and heating, which lead to the much more severe pollution atrural sites than urban sites in Beijing.
Nitrate increased in themorning, retained a relatively stable concen-tration in the daytime and decreased in the midnight winter (Fig. 4).This diurnal variationwas nearly the samewith that of NO2. The diurnalvariation of ammonium was similar with nitrate. Sulfate showed twopronounced peaks with one at 10:00 am and the other one at 8:00 pm.
HOA showed two pronounced peaks. Onewas from 7:00 am to 9:00am and the other onewas from 5:00 pm to 1:00 am the next day. BBOAand CCOA showed a similar variation which was high at night and lowin the daytime. The diurnal variation of PBL in winter season (Fig. S5)showed the PBL height started to increase significantly after 8:00 am,which was not consistent to the POA variation. This indicates theboundary layer was unable to explain the high concentration of POAat night. The POA variation was consistent to the residential heating be-haviour. Residents consumemuchmore fuels after they comeback fromwork at night than in the daytime for heating.
LVOOA showed a strongly regular variation with high concentrationin the daytime and low concentration at night-time. This is reasonablethat photochemical reaction in the daytime played the leading role.
3.3. Comparison of characteristics and sources in different seasons betweenthe rural and urban area
To investigate the reason why the rural area is muchmore pollut-ed, results in this study and from the literatures are listed in Table 2.As a whole, OM contribution at Liulihe site was higher than the urbansite in autumn, spring and summer. They are comparable in winterseason.
The diurnal variation of HOA at Liulihe site showed no two pro-nounced peaks in the rush hour like urban site. This is resulted fromthe effects of rush hour and HDDT for goods transportation at ruralarea. The contribution of HOA was comparable in autumn and winterbetween the rural area and urban area. There should be less numberof light-duty vehicles in the rural area, which means HDDT dominatedthe vehicle emission,making the HOA contributionwas comparable be-tween the rural area and urban area.
BBOA concentration was comparable between autumn and spring,andwas twice of that in winter season at Liulihe site. The large quantityof biomassmaterials after the harvest in autumn canmake it become animportant residential consumption fuel.When it comes intowinter sea-son, the biomass material can be used for heating at rural areas. What'smore, long distance transport can bring aged BBOA to Beijing. Whencomparing with the results at the urban site in Beijing (Table 2), thecontribution at Liulihe site was significantly higher than that at theurban site.
CCOA was distinguished only in winter season both in the rural andurban area. The contribution of CCOA in the rural areawas resulted fromresidential heating. Dispersed heating is the major residential heatingmode in the rural area. The heating emission can be huge without anypollutants removal facilities. With the addition of the contribution ofBBOA, residential heating in rural areas in Beijing lead to much higherPM2.5 concentration than urban area.
OOA contribution at Liulihe site was lower than the urban site inall the seasons except winter season. This might be caused by themuch higher primary organic aerosol emission in the rural area inwinter.
4. Conclusions
Four seasons observation was carried out at Liulihe site, a most pol-luted rural site south of Beijing, to investigate the pollution characteris-tics and sources. Compared with the urban area in Beijing, biomassburning and coal combustion contributed much more in the southrural area of Beijing. Biomass burning was a very important emissionsource in all the seasons except summer. Coal combustion was the larg-est emission source in winter. Especially in the most polluted winterseason, residential heating consumes both biomass and coal withoutany pollutants removal facilities, making the pollutants concentrationmuch higher in rural areas. HOA showed a significant effect of HDDTfor goods transportation at night-time in the rural area in Beijing. Thecontribution from residential emissions in the rural area was muchhigher than that reported in the urban area. If the residential emissionsare not investigated in the rural area, the results from urban sites alonewill underestimate the contribution of residential emissions to aerosolpollution in the region.
Acknowledgement
This work was supported by the National Natural Science Founda-tion of China (21625701 & 21521064). The authors also appreciate thesupport fromCollaborative Innovation Centre for Regional Environmen-tal Quality.
527Y. Hua et al. / Science of the Total Environment 626 (2018) 519–527
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.org/10.1016/j.scitotenv.2018.01.047.
References
Alfarra, M., Prevot, A., Szidat, S., Sandradewi, J., Weimer, S., Lanz, V., Schreiber, D., Mohr,M., Baltensperger, U., 2007. Identification of the mass spectral signature of organicaerosols from wood burning emissions. Environ. Sci. Technol. 41 (16), 5770–5777.
Aurela, M., Saarikoski, S., Niemi, J.V., Canonaco, F., Prevot, A.S.H., Frey, A., Carbone, S.,Kousa, A., Hillamo, R., 2015. Chemical and source characterization of submicron par-ticles at residential and traffic sites in the Helsinki metropolitan area, Finland. AerosolAir Qual. Res. 15 (4), 1213–1226.
Beijing Municipal Bureau of Statistics, 2015. The National Statistical Yearbook of China in2014.
Beijing Municipal Environmental Protection Bureau, 2014. http://www.bjepb.gov.cn/bjepb/413526/331443/331937/333896/396191/index.html.
Beijing Municipal Environmental Protection Bureau, 2015. http://www.bjepb.gov.cn/bjepb/resource/cms/2015/04/2015041609380279715.pdf.
Beijing Municipal Environmental Protection Bureau, 2016. http://www.bjepb.gov.cn/bjepb/resource/cms/2016/04/2016041514503583104.pdf.
Beijing Municipal Environmental Protection Bureau, 2017. http://www.bjepb.gov.cn/bjhrb/xxgk/jgzn/jgsz/jjgjgszjzz/xcjyc/xwfb/803842/index.html.
Cai, S., Wang, S., Zhao, B., Chang, X., Hao, J., 2018. Pollutant Emissions From ResidentialCombustion and Reduction Strategies Estimated via a Village-based Inventory in Bei-jing (unpublished results).
Cheng, Y., Zheng, G., Wei, C., Mu, Q., Zheng, B., Wang, Z., Gao, M., Zhang, Q., He, K.,Carmichael, G., Pöschl, U., Su, H., 2016. Reactive nitrogen chemistry in aerosol wateras a source of sulphate during haze events in China. Sci. Adv. 2 (12).
Cubison, M., Ortega, A., Hayes, P., Farmer, D., Day, D., Lechner, M., Brune, W., Apel, E.,Diskin, G., Fisher, J., 2011. Effects of aging on organic aerosol from open biomass burn-ing smoke in aircraft and laboratory studies. Atmos. Chem. Phys. 11 (23),12049–12064.
Elser, M., Huang, R.-J., Wolf, R., Slowik, J.G., Wang, Q., Canonaco, F., Li, G., Bozzetti, C.,Daellenbach, K.R., Huang, Y., 2016. New insights into PM2.5 chemical compositionand sources in two major cities in China during extreme haze events using aerosolmass spectrometry. Atmos. Chem. Phys. 16 (5), 3207–3225.
Grieshop, A., Donahue, N., Robinson, A., 2009. Laboratory investigation of photochemicaloxidation of organic aerosol fromwood fires 2: analysis of aerosol mass spectrometerdata. Atmos. Chem. Phys. 9 (6), 2227–2240.
Ji, D., Li, L., Wang, Y., Zhang, J., Cheng, M., Sun, Y., Liu, Z., Wang, L., Tang, G., Hu, B., Chao, N.,Wen, T., Miao, H., 2014. The heaviest particulate air-pollution episodes occurred innorthern China in January, 2013: insights gained from observation. Atmos. Environ.14 (92), 546–556.
Liu, J., Mauzerall, D.L., Chen, Q., Zhang, Q., Song, Y., Peng, W., Klimont, Z., Qiu, X., Zhang, S.,Hu, M., 2016. Air pollutant emissions from Chinese households: a major and under-appreciated ambient pollution source. Proc. Natl. Acad. Sci. 113 (28), 7756–7761.
Middlebrook, A.M., Bahreini, R., Jimenez, J.L., Canagaratna, M.R., 2012. Evaluation of com-position-dependent collection efficiencies for the aerodyne aerosol mass spectrome-ter using field data. Aerosol Sci. Technol. 46 (3), 258–271.
Ministry of Environment Protection, 2016. http://www.gov.cn/xinwen/2016-02/05/con-tent_5039488.htm.
Ng, N.L., Herndon, S.C., Trimborn, A., Canagaratna, M.R., Croteau, P.L., Onasch, T.B., Sueper,D., Worsnop, D.R., Zhang, Q., Sun, Y.L., Jayne, J.T., 2011. An aerosol chemical speciationmonitor (ACSM) for routine monitoring of the composition and mass concentrationsof ambient aerosol. Aerosol Sci. Technol. 45 (7), 780–794.
Paatero, P., Tapper, U., 1994. Positive matrix factorization: a non-negative factor modelwith optimal utilization of error estimates of data values. Environmetrics 5 (2),111–126.
Pathak, R.K., Wu, W.S., Wang, T., 2009. Summertime PM2.5 ionic species in four major cit-ies of China: nitrate formation in an ammonia-deficient atmosphere. Atmos. Chem.Phys. 9 (5), 1711–1722.
Schneider, J., Weimer, S., Drewnick, F., Borrmann, S., Helas, G., Gwaze, P., Schmid, O.,Andreae, M., Kirchner, U., 2006. Mass spectrometric analysis and aerodynamic prop-erties of various types of combustion-related aerosol particles. Int. J. Mass Spectrom.258 (1), 37–49.
Sun, Y., Zhang, Q., Schwab, J., Demerjian, K., Chen,W., Bae, M., Hung, H., Hogrefe, O., Frank,B., Rattigan, O., 2011. Characterization of the sources and processes of organic and in-organic aerosols in New York city with a high-resolution time-of-flight aerosol massapectrometer. Atmos. Chem. Phys. 11 (4), 1581–1602.
Sun, Y., Wang, Z., Dong, H., Yang, T., Li, J., Pan, Chen, P., Jayne, J., 2012. Characterization ofsummer organic and inorganic aerosols in Beijing, China with an aerosol chemicalspeciation monitor. Atmos. Environ. 51, 250–259.
Sun, Y., Wang, Z., Fu, P., Jiang, Q., Yang, T., Li, J., Ge, X., 2013. The impact of relative humid-ity on aerosol composition and evolution processes during wintertime in Beijing,China. Atmos. Environ. 77 (0), 927–934.
Sun, Y., Du, W., Wang, Q., Zhang, Q., Chen, C., Chen, Y., Chen, Z., Fu, P., Wang, Z., Gao, Z.,2015a. Real-time characterization of aerosol particle composition above the urbancanopy in Beijing: insights into the interactions between the atmospheric boundarylayer and aerosol chemistry. Environ. Sci. Technol. 49 (19), 11340–11347.
Sun, Y., Wang, Z., Du, W., Zhang, Q., Wang, Q., Fu, P., Pan, X., Li, J., Jayne, J., Worsnop, D.,2015b. Long-term real-time measurements of aerosol particle composition in Beijing,China: seasonal variations, meteorological effects, and source analysis. Atmos. Chem.Phys. 15 (17), 10149–10165.
Ulbrich, I., Canagaratna, M., Zhang, Q., Worsnop, D., Jimenez, J., 2009. Interpretation of or-ganic components from positive matrix factorization of aerosol mass spectrometricdata. Atmos. Chem. Phys. 9 (9), 2891–2918.
Wang, Y., Zhang, X., Draxler, R.R., 2009. TrajStat: GIS-based software that uses various tra-jectory statistical analysis methods to identify potential sources from long-term airpollution measurement data. Environ. Model. Softw. 24 (8), 938–939.
Wang, L., Liu, Z., Sun, Y., Ji, D., Wang, Y., 2015. Long-range transport and regional sourcesof PM2.5 in Beijing based on long-term observations from 2005 to 2010. Atmos. Res.157, 37–48.
Wang, G., Zhang, R., Gomez, M.E., Yang, L., Zamora, M.L., Hu, M., Lin, Y., Peng, J., Guo, S.,Meng, J., 2016. Persistent sulphate formation from London Fog to Chinese haze.Proc. Natl. Acad. Sci. 113 (48), 13630–13635.
Xie, Y., Ding, A., Nie, W., Mao, H., Qi, X., Huang, X., Xu, Z., Kerminen, V.M., Petäjä, T., Chi, X.,2015. Enhanced sulphate formation by nitrogen dioxide: implications from in situ ob-servations at the SORPES station. J. Geophys. Res. Atmos. 120 (24), 12679–12694.
Xu, W., Song, W., Zhang, Y., Liu, X., Zhang, L., Zhao, Y., Liu, D., Tang, A., Yang, D., Wang, D.,2017. Air quality improvement in a megacity: implications from 2015 Beijing ParadeBlue pollution control actions. Atmos. Chem. Phys. 17 (1), 31–46.
Zhang, H., Wang, S., Hao, J., Wang, X., Wang, S., Chai, F., Li, M., 2016. Air pollution and con-trol action in Beijing. J. Clean. Prod. 112, 1519–1527.
Zhang, J.K., Cheng, M.T., Ji, D.S., Liu, Z.R., Hu, B., Sun, Y., Wang, Y.S., 2016. Characterizationof submicron particles during biomass burning and coal combustion periods in Bei-jing, China. Sci. Total Environ. 562, 812–821.
Zhang, S., Wu, Y., Yan, H., Du, X., Max Zhang, K., Zheng, X., Fu, L., Hao, J., 2017. Black carbonpollution for a major road in Beijing: implications for policy interventions of theheavy-duty truck fleet. Transp. Res. Part D: Transp. Environ. https://doi.org/10.1016/j.trd.2017.07.013.
Zhao, B., Wang, P., Ma, J.Z., Zhu, S., Pozzer, A., Li, W., 2012. A high-resolution emission in-ventory of primary pollutants for the Huabei region, China. Atmos. Chem. Phys. 12(1), 481–501.
Zhao, B., Wang, S., Dong, X., Wang, J., Duan, L., Fu, X., Hao, J., Fu, J., 2013. Environmentaleffects of the recent emission changes in China: implications for particulate matterpollution and soil acidification. Environ. Res. Lett. 8 (2), 024031.
Zheng, S., Pozzer, A., Cao, C., Lelieveld, J., 2015. Long-term (2001–2012) concentrations offine particulate matter (PM2.5) and the impact on human health in Beijing, China.Atmos. Chem. Phys. 15 (10), 5715–5725.
Zheng, G., Duan, F., Ma, Y., Zhang, Q., Huang, T., Kimoto, T., Cheng, Y., Su, H., He, K., 2016.Episode-based evolution pattern analysis of haze pollution: method developmentand results from Beijing, China. Environ. Sci. Technol. 50 (9), 4632–4641.
