Characterizing oxygenated volatile organic compounds and their sources in rural atmospheres in China

J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S 81 (2 0 1 9 ) 1 48–1 5 5

Available online at www.sciencedirect.com

ScienceDirect www.elsevier.com/locate/jes

Characterizing oxygenated volatile organic compounds and their sources in rural atmospheres in China Yu Han, Xiaofeng Huang*, Chuan Wang, Bo Zhu, Lingyan He Key Laboratory for Urban Habitat Environmental Science and Technology, School of Environment and Energy, Peking University Shenzhen Graduate School, Shenzhen 518055, China

AR TIC LE I N FO

ABS TR ACT

Article history:

Oxygenated volatile organic compounds (OVOCs) are important precursors and products of

Received 29 August 2018

atmospheric secondary pollution. The sources of OVOCs, however, are still quite uncertain,

Revised 22 January 2019

especially in the atmosphere with much pollution in China. To study the sources of OVOCs in

Accepted 23 January 2019

rural atmospheres, a proton transfer reaction mass spectrometry (PTR-MS) was deployed at a

Available online 2 February 2019

northern rural site (WD) and a southern rural site (YMK) in China during the summer of 2014 and 2016, respectively. The continuous observation showed that the mean concentration of

Keywords:

TVOCs (totally 17 VOCs) measured at WD (52.4 ppbv) was far higher than that at YMK

OVOCs

(11.1 ppbv), and the OVOCs were the most abundant at both the two sites. The diurnal

Photochemical age

variations showed that local sources of OVOCs were still prominent at WD, while regional

Anthropogenic sources

transport influenced YMK much. The photochemical age-based parameterization method

Primary emission

was then used to quantitatively apportion the sources of ambient OVOCs. The anthropogenic

Secondary formation

primary sources at WD and YMK contributed less (2%–16%) to each OVOC species. At both the sites, the atmospheric background had a dominant contribution (~ 50%) to acetone and formic acid, while the anthropogenic secondary formation was the main source (~ 40%) of methanol and MEK. For acetaldehyde and acetic acid, the biogenic sources were their largest source (~ 40%) at WD, while the background (39%) and anthropogenic secondary formation (42%) were their largest sources at YMK, respectively. This study reveals the complexity of sources of OVOCs in China, which urgently needs explored further. © 2019 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.

Introduction Volatile organic compounds (VOCs) play an important role in tropospheric photochemical reactions, contributing to the formation of secondary pollutants such as ozone (O3), secondary organic aerosol (SOA) and peroxyacetyl nitrate (PANs) with nitrogen oxides (NOx), under the action of ultraviolet radiation (Atkinson, 2000; Derwent et al., 2003; Seinfeld and Pandis, 2006). These compounds are known to

seriously affect the quality of the regional atmospheric environment and human health (Finlayson-Pitts and Pitts, 1997; Bashkin, 2009). Oxygenated volatile organic compounds (OVOCs) are an important subset of VOCs and have strong activities and complex sources that enter the atmosphere, not only through anthropogenic emissions sources, such as vehicle emissions, biomass combustion and industrial activities, but also through biogenic sources. Furthermore, several OVOCs are produced by the secondary oxidation reaction of

⁎ Corresponding author. E-mail: [email protected]. (Xiaofeng Huang).

https://doi.org/10.1016/j.jes.2019.01.017 1001-0742 © 2019 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.

149

J O U RN A L OF E N V I RO N M EN TA L S CI EN CE S 81 (2 0 1 9 ) 1 4 8–1 5 5

Table 1 – Summary of the meteorological parameters and basic air pollutants concentrations at two rural sites (average ± 1σ standard deviation).

Meteorology

Species (μg/m3)

Sampling site

Wangdu

Sampling period

2014.6.7–7.6

2016.6.30–7.23

Temperature (°C) Relative humidity (%) Wind speed (m/sec)

26.2 ± 4.7 60.4 ± 20.2 4.2 ± 2.7

29.7 ± 1.8 73.9 ± 8.1 1.5 ± 0.8

Dominant wind direction

North

South

Ozone NO2 PM2.5

107.8 ± 64.9 23.8 ± 14.8 71.0 ± 46.8

137.0 ± 80.8 12.2 ± 7.4 13.3 ± 11.0

nonmethane hydrocarbons (NMHCs) (Kansal, 2009; Chang et al., 2005). In recent years, as the concentration level of PM2.5 continues to decrease, O3 pollution has gradually appeared (Ministry of Ecology and Environment of China, 2017); therefore, quantitatively distinguishing different sources of premise VOCs and developing differentiated control strategies are conducive to maximizing the improvement of O3 pollution. Ambient background observation is an important part of atmospheric chemical observation, reflecting the initial state of environmental quality in a certain period and specific area, which is of great significance to regional environmental quality assessment and environmental pollution trend prediction (Yang, 1990). Legreida et al. (2007) used a newly developed double adsorbent sampling unit coupled to a gas chromatograph–mass spectrometer (GC–MS) to analyze the pollution characteristics of 21 OVOCs at an urban background site in Zürich. Jordan et al. (2009) used proton transfer reaction mass spectrometry (PTR-MS) to continuously observe the rural atmosphere in the northeastern United States for three years and established a high-resolution VOC database to analyze seasonal and diurnal variations. Zhang et al. (2012) used offline gas chromatography and mass spectrometry (offline-GC/MS) to analyze the regional background concentrations and seasonal variation characteristics of atmospheric VOCs in southwestern China and combined these measurements with principal component and factor analysis for source apportionment. Bai et al. (2016) used the same method to analyze the pollution characteristics of VOCs at the Qinghai-Tibet Plateau background site, and adopted the backward trajectory analysis to track the origin of the longtransport of air pollution. Xue et al. (2013) used a photochemical box model based on the Master Chemical Mechanism to probe the O3 yield of the background atmosphere in Waliguan. However, in the eastern region, where air pollution is most prominent in China, there are few VOC observations at the regional background sites. Beijing–Tianjin–Hebei and the Pearl River Delta are the two fastest-growing urban agglomerations in China. High urbanization and industrialization have brought about regional high-concentration O3 pollution (Sin et al., 2001). It is crucial to scientifically formulate air pollution control strategies for urban agglomerations through the study of regional background ambient quality. In this study, with the aim of revealing the pollution characteristics and sources of OVOCs at the northern and southern rural sites, the ambient levels of

Yangmeikeng

six kinds of OVOCs and other NMHCs were simultaneously measured for 30 days at Wangdu (WD, northern rural site) in 2014 and at Yangmeikeng (YMK, southern rural site) in the summer of 2016, using a high-sensitivity PTR-MS at these two rural sites. The characteristics of concentration and diurnal variation were described, the ozone generation potential (OFP) of different OVOCs were calculated, and the contribution of different sources of OVOCs at the northern and southern rural sites were calculated using the photochemical age-based parameterization method.

1. Materials and methods 1.1. Sampling sites and meteorological conditions Wangdu County (38.72°N, 115.5°E) is located in the suburb of Baoding City, Hebei Province, in the North China Plain and is the center of the Beijing–Tianjin–Hebei urban agglomeration. The sampling site is located in the greening base of the Transportation Bureau, where it is surrounded by wheat fields and orchards. Although it serves as a typical rural site, there were indeed some industry across the whole Baoding city. WD was expected to reflect the regional characteristics of ambient VOC pollution in the Beijing–Tianjin–Hebei urban agglomeration to some extent. The sampling head was set on the top of a second-floor container, approximately 10 m from the ground. Shenzhen Environmental Monitoring Yangmeikeng Station (22.55°N, 114.60°E, altitude 44 m) is located in Dapeng Peninsula in the East of Shenzhen City, Guangdong Province. It is surrounded by the sea on three sides, and there are no industrial or mining enterprises around. YMK essentially reflects the pollution status of VOCs in the rural area of South China coastal line. The sampling head was set on the top of the third floor of the experimental building, approximately 15 m from the ground. In this study, PTR-MS was used to continuously monitor VOCs and other gaseous pollutants (O3, CO, etc.) in Wangdu and Yangmeikeng in 2014 and 2016 (June–July), respectively. Table 1 lists the meteorological parameters of the two observation sites.

1.2. Measurements of OVOCs and other air pollutants In this study, a commercial high-sensitivity PTR-MS (Ionicon Analytik GmbH, Austria) was employed to conduct

150

J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S 81 (2 0 1 9 ) 1 48–1 5 5

measurements. VOC species with proton affinities greater than that of water can be protonated by hydronium ions (H3O+), generated inside the ion source and then detected via a QMS (Lindinger et al., 1998). The instrument and its detailed operational parameters can be found in a recent review paper (de Gouw and Warneke, 2007) and are not described further here. The ambient air sampling was collected by a sampling head and continuously drawn through a 1/4” Teflon tube (4.0 mm ID, 8-m length) into the gas-handling system by a drainage-pump at a flow rate of 1–1.5 L/min. The inlet flow rate adjusted by a needle valve and kept at 200–250 mL/min. During the sampling period, the PTR-MS measured a total of 24 masses in selected ion mode (SIM) at a time resolution of 25 sec (most of the masses were recorded for 1 sec in each cycle with the exception of m/z 137, which was recorded for 2 sec). Background checks were conducted for 30 out of every 300 scan cycles. Over 99% VOCs were removed from the ambient air by heating to 360°C with a custom-built catalytic converter. The data obtained by PTR-MS were processed into a time resolution of 5 min. In addition, a 49i O3 analyzer and a 48i CO analyzer (Thermo Scientific, USA) were carried out for observations of other gases.

1.3. OVOCs calibration The OVOCs measured by PTR-MS include methanol, acetaldehyde, acetone, methyl ethyl ketone (MEK), formic acid and acetic acid, all of which were calibrated by two types of standard gases. One type was a cylinder gas containing 63 target compounds, according to the TO15 method proposed by the US EPA, which was used to calibrate methanol, acetaldehyde, acetone and MEK. Another type of standard gas was used for formic acid and acetic acid, which were calibrated using a permeate tube (Valco Instruments Co. Inc., USA) and a permeation oven (Dynacalibrator Model 230, Valco Instruments Co. Inc., USA), due to the high activity and viscosity. In the two campaigns, a total of 10 calibrations were carried out. The correlations in the calibration results for all species exceeded 0.99, which satisfied the research requirements of this study.

1.4. Calculation of O3 formation potential As an important precursor of atmospheric photochemical pollution, changes in composition and concentration of VOCs have important influences on the relative contribution of O3 formation due to the large differences in chemical reactivity between different species. This study employed the maximum incremental reactivity (MIR) to quantitatively assess the O3 formation potentials (OFPs) of different types of VOCs, and screened out the dominant components of O3 formation in ambient VOCs (Carter, 1994; Barletta et al., 2008). The OFP can be calculated by Eq. (1). OFPi ¼ ½VOCi   MIRi

ð1Þ

where [VOCi] (ppbv) is the mixing ratio of the respective VOCi measured by PTR-MS, and MIRi is the maximum incremental reactivity (gO3/gVOC) derived from Carter (1994).

1.5. Source apportionment method of OVOCs To evaluate the relative contributions of different sources, this study adopted a photochemical age-based parameterization method to quantitatively distinguish the primary and secondary sources of typical OVOCs. The concentration of OVOCs was attributed to anthropogenic primary sources, anthropogenic secondary emissions, biogenic sources and background sources, as shown in Eq. (2) (de Gouw and Warneke, 2007).  

½OVOC ¼ EROVOC  Tracerap  exp − kOVOC −kTracerap ½OHΔt   kprecursor þERprecursor  Tracerap  kOVOC −kprecursor
exp −kprecursor ½OHΔt − expð−kOVOC ½OHΔtÞ
 exp −kTracerap ½OHΔt
þERbiogenic  isoprenesource þ ½background ð2Þ Among them, [OVOC] (ppbv), [Tracerap] (ppbv) and [background] (ppbv) represent the concentration of ambient OVOC, the tracer concentration of the anthropogenic primary sources and the background sources of the OVOC, respectively. kOVOC, kTracerap and kprecursor represent the OH rate constants of the OVOC, tracer and precursor, respectively; kOVOC and kTracerap can be obtained from Atkinson and Arey (2003). EROVOC and ERprecursor are emission factors of the OVOC and precursors relative to the tracer. [OH]Δt (molecule/cm3 sec) represents the exposure level of OH radicals. ERbiogenic is the emission factor of the OVOC, relative to the isoprene concentration emitted from biogenic sources. The values of [OH]Δt and isoprenesource were calculated by isoprene and its photochemical products methyl vinyl ketone (MVK) and methacrolein (MACR). The parameters of EROVOC, ERprecursor, ERbiogenic, kprecursor and [background] are calculated from a linear least-squares fit in Eq. (2). This approach was successfully used in New England (de Gouw and Warneke, 2007), Beijing (Liu et al., 2015; Yuan et al., 2012) and Shenzhen (Zhu et al., 2018). More comprehensive information on this method is described in de Gouw and Warneke (2007). The photochemical age-based parameterization method applied in OVOCs sources appointment has some assumptions: (1) anthropogenic emissions of OVOCs and their precursors are proportional to a primary marker; (2) the removal of OVOCs is dominated by the reactions with OH radicals; (3) biogenic sources of OVOCs (biogenic primary sources and biogenic secondary sources) are proportional to the emission of isoprene; (4) the photochemical ages for sampled air masses can be determined (Chen et al., 2014; de Gouw and Warneke, 2007). It is necessary to select a suitable tracer for anthropogenic sources to fit the source of OVOCs using photochemical age-based parameterization method. In this study, we adopted benzene, which is mainly produced by vehicle emissions and industrial activities at the rural sites, as the tracer of anthropogenic primary sources (Zhang et al., 2013). The tracer not only has a reaction rate constant similar to that of acetylene (Atkinson and Arey, 2003), which was a tracer used in research by de Gouw and Warneke (2007). At the same time, other studies have chosen benzene as the tracer for anthropogenic primary sources (Ma et al., 2016).

151

J O U RN A L OF E N V I RO N M EN TA L S CI EN CE S 81 (2 0 1 9 ) 1 4 8–1 5 5

2. Results and discussion 2.1. Comparison of OVOCs concentration levels in rural atmospheres The continuous online observation at the northern and southern rural sites indicated that the concentration of 17 kinds of VOCs measured by PTR-MS in the North (52.4 ppbv) was higher than that in the South. Among the total measured VOC species, 6 kinds of OVOCs concentrations were the most abundant (89.5%), and 9 kinds of aromatics and biogenic VOCs (BVOCs) accounted for less than 10%; the top 5 dominant species at WD were methanol, formic acid, acetic acid, acetone and acetaldehyde. The mean concentration of TVOCs at the southern rural site was only 11.1 ppbv, which was much lower than that of the northern site. The concentrations of OVOCs were still the largest (74.1%), but the concentrations of BVOCs (13.6%) were significantly higher than those of the North (3.3%). Methanol, acetone, monoterpene, acetaldehyde and formic acid were the top five dominant species at YMK. A comparative study on six kinds of OVOCs conducted among eight areas is listed in Table 2. Shangdianzi, Gucheng, Zürich and New Hampshire were selected as regional background sites (Legreida et al., 2007; Jordan et al., 2009; Liu, 2010), while Beijing and Shenzhen were selected as typical urban sites (Liu et al., 2015; Zhu et al., 2018). It can be seen from Table 2 that there were large differences in the concentration levels of OVOCs at different sites. The concentration level of acetone and MEK at the regional background sites was significantly lower than that of the typical urban sites. The concentration of domestic rural sites was significantly higher than that of foreign rural sites. The methanol concentration at WD was approximately 10 times that of New Hampshire, suggesting that domestic methanol had strong anthropogenic emissions. Compared with the northern rural site, the concentrations of the OVOCs in the South were at a lower level. However, the

concentration of MEK was slightly higher than, and the concentrations of methanol, acetaldehyde and acetone at YMK were similar to those of the foreign background sites.

2.2. O3 formation potentials in ambient OVOCs The OFP of 11 kinds of NMHCs measured by PTR-MS during the initial stage of photochemical reactions (10:00–11:00) and the whole day at two rural sites are shown in Fig. 1. The analysis results revealed that the contribution rate of different species in the northern rural areas during the two monitoring periods were essentially consistent. Methanol, acetaldehyde and isoprene were the top three species in OFP, respectively. The cumulative contribution from these three species reached 74%–77%. The OFP orders for different VOCs at the southern rural site were significantly different between the initial photochemical reaction and the whole day. Isoprene, monoterpene and acetaldehyde were the species with the largest variation, their OFP in the initial stage of the photochemical reaction were 94%, 74% and 38% higher than that of the whole day, respectively. The three species were also the most dominant species at the southern rural site, with a cumulative contribution of 64%–72%. Although the mean concentration of methanol accounted for the highest content of rural components, the OFP ranked only fourth. This is mainly due to the higher MIR of isoprene, acetaldehyde and monoterpene and the higher temperature and stronger radiation in the initial stage of the photochemical reaction in the southern region, which exacerbated biogenic emissions (Alex et al., 1993). In contrast to the typical urban areas, the OFP contribution of toluene at the rural sites was not prominent; this was closely related to the reduction of emission sources.

2.3. Diurnal variations in ambient OVOCs The average diurnal variations of typical VOCs and major gaseous pollutants in the summer of the northern and southern rural sites are shown in Fig. 2. The diurnal variations

Table 2 – Comparison of the oxygenated volatile organic compounds (OVOC) concentrations among rural sites and urban sites (unit: ppbv). Location

Characteristic Period

Methanol Acetaldehyde Acetone MEK Formic acid

Zürich, Switzerland Shangdianzi, China WD, China YMK, China New Hampshire, the U.S.

Background

Summer 2005 2011.6

3.05

2014.6–7 2016.6–7 Summer 2007 Summer 2006 Summer 2005 2011.6 Summer 2008 Summer 2016

Background Rural Rural Rural

Gucheng, China Beijing, China

Rural

Shenzhen, China

Urban

Urban

Acetic acid

Method

Reference

GC–MS HP 6890/HP 5973N Offline-GCMS

Legreida et al. (2007) Bai et al. (2016) This work This work Jordan et al. (2009)

0.80

1.81

0.16

4.05

3.45

1.03

26.04 3.90 2.61

2.39 0.73 0.47

4.78 1.77 2.01

1.27 0.54 0.21

2.76

0.49

2.13

0.20

0.54

2.81

0.64

2.19

0.25

0.70

7.55

4.34

1.47

Offline-GCMS

14.96

2.72

3.96

1.16

PTR-MS

14.47

1.98

3.78

1.33

7.60 0.67

1.21

5.23 0.57 0.62

1.68

PTR-MS PTR-MS PTR-MS

PTR-MS

Bai et al. (2016) Liu et al. (2015) Zhu et al. (2018)

152

J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S 81 (2 0 1 9 ) 1 48–1 5 5

Fig. 1 – Mixing ratios of the measured volatile organic compounds (VOCs), the calculated initial stage of photochemical reactions (10:00–11:00) and the whole day of the ozone formation potential (OFP) at the northern and southern rural sites.

of the same species at WD and YMK showed great differences. The diurnal variations of OVOCs at WD were similar to that of the typical primary pollutants, i.e., benzene and CO, whose concentrations at night were significantly higher than those in the daytime, indicating that the daytime vertical mixing in the atmospheric boundary layer effectively diluted the local pollutants at the ground level. Considering the high TVOC concentrations at WD, the rural emissions (e.g., rural

enterprises) of OVOCs and/or their precursors surrounding WD were inferred to be prominent. On the other hand, the concentration decrease of OVOCs during the daytime could be, also related to their oxidation consumption by OH radicals (Wu et al., 2011; Zhang et al., 2014). In contrast, the diurnal variation characteristics of OVOCs at YMK generally showed higher concentrations during the daytime than during the nighttime. CO had a consistent concentration during the day

Fig. 2 – Diurnal variations in the concentrations of typical VOCs and gaseous pollutants in the summer of the northern and southern rural sites.

153

J O U RN A L OF E N V I RO N M EN TA L S CI EN CE S 81 (2 0 1 9 ) 1 4 8–1 5 5

Table 3 – Relative contributions from different sources to OVOCs at the northern and southern rural sites. Compounds

R

Anthropogenic primary (%)

Anthropogenic secondary (%)

Biogenic (%)

Background (%)

WD Methanol Acetaldehyde Acetone MEK Formic Acid Acetic Acid

0.81 0.76 0.74 0.74 0.68 0.71

2 8 5 3 7 13

39 29 20 39 28 18

18 38 20 30 15 39

41 25 55 28 50 30

YMK Methanol Acetaldehyde Acetone MEK Formic acid Acetic acid

0.93 0.85 0.87 0.88 0.91 0.90

16 10 5 9 10 15

36 25 19 38 14 42

22 26 20 16 30 28

26 39 56 37 46 15

and night at YMK, reflecting that local emissions were minimal, while the higher concentrations of benzene in the daytime suggested that the daytime vertical mixing in the atmospheric boundary layer would increase the ground level VOC pollution through high altitude regional transport. In addition, the stronger daytime increasing of OVOCs was similar to that of O3, implying that secondary formation had a significant contribution.

2.4. Source apportionment of OVOCs utilizing the photochemical age-based parameterization method Commonly used OVOCs source apportionment methods include the tracer method (Kanaya et al., 2007; Lin et al., 2012), the multiple linear regression method (Possanzini et al., 2007), the photochemical age method (de Gouw and Warneke, 2007; Yuan et al., 2012), the receptor model method (Duan et al., 2012), and the air quality model, based on precursors for estimating the contribution of different sources to OVOCs (Luecken et al., 2012; Parrish et al., 2012). However, the results obtained by different source analysis methods are very different (Yuan et al., 2012), which indicates that there is great uncertainty in the relative contribution of primary and secondary emissions to the concentration of OVOCs in the environment. Utilizing the photochemical age-based parameterization method described in Section 1.5, the correlation between the modeled and measured concentration of OVOCs at the northern and southern rural sites and the relative contribution of various sources to OVOCs are listed in Table 3. The fitting results of the six kinds of OVOCs showed good correlation, such that the correlation coefficients R of the northern and southern rural sites was greater than 0.68 and 0.85, respectively. It can be seen that anthropogenic primary sources at WD and YMK contributed less to each OVOC species, which was consistent with their rural characteristics. On the other hand, the atmospheric background had a dominant contribution (~50%) to acetone and formic acid at the southern and northern sites, indicating that these two species had significant regional pollution characteristics that were closely related to the more stable chemical properties of these two OVOCs. In addition to the atmospheric background,

the anthropogenic secondary source was the main source of methanol and MEK at the southern and northern sites. For acetaldehyde and acetic acid with two carbon atoms, the southern and northern sites showed significant differences. The biogenic source was the most significant source at WD, while the background (39%) and anthropogenic secondary formation (42%) were their largest sources at YMK, respectively. Yuan et al. (2012) and Zhu et al. (2018) used the same method in the urban areas of Beijing and Shenzhen in the same season to determine that the anthropogenic primary source was the most important source of alcohol ketones, which is quite different from the results of this study. This large difference arose because there were substantially fewer human activities in rural areas than in urban areas. Based on the above analysis, it was concluded that the source of OVOCs at rural sites was complex and was due to comprehensive products of anthropogenic and biogenic emissions through photochemical reactions. To control the OVOCs and the resulting O3 pollution, the identification of the dominant sources of OVOCs is a prerequisite.

3. Conclusions The ambient OVOCs and other NMHCs at the northern and southern rural sites were measured using a PTR-MS during the summer to describe the mean concentration and diurnal variation and to calculate the OFP of different VOCs. The photochemical age-based parameterization method was then applied to quantitatively distinguish the primary and secondary sources of OVOCs. The continuous online observation of PTR-MS showed that the mean concentration of TVOCs at the northern rural site (52.4 ppbv) was much higher than that at the southern site (11.1 ppbv), and the OVOC concentration was the most abundant at the two sites. Acetaldehyde was estimated to be a popular high OFP species at both the northern and southern sites. The diurnal variation characteristics indicated prominent local sources (e.g., rural enterprises) of OVOCs at WD, while regional transport of pollution played a key role for OVOCs at YMK. Based on the photochemical age-based parameterization method for assessing the relative contribution of different sources of OVOCs, the

154

J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S 81 (2 0 1 9 ) 1 48–1 5 5

anthropogenic primary sources at WD and YMK contributed less to each OVOC species. The background sources had a dominant contribution (~ 50%) to acetone and formic acid at both the sites. The anthropogenic secondary sources were the main source (~40%) of methanol and MEK at the two sites. For acetaldehyde and acetic acid, the biogenic sources were their major source (~ 40%) at WD, while the background (39%) and anthropogenic secondary formation (42%) were their largest sources at YMK, respectively.

Acknowledgements This work was supported by the Ministry of Science and Technology of China, China (No. 2017YFC0210004), the National Natural Science Foundation of China, China (No. 91544215) and the Science and Technology Plan of Shenzhen Municipality (No. JCYJ20170412150626172).

REFERENCES Alex, B.G., Patrick, R.Z., Peter, H., 1993. Isoprene and monoterpene emission rate variability: model evaluations and sensitivity analyses. J. Geophys. Res.-Atmos. 98, 12609–12617. Atkinson, R., 2000. Atmospheric chemistry of VOCs and NOx. Atmos. Environ. 34, 2063–2101. Atkinson, R., Arey, J., 2003. Atomospheric degradation of volatile organic compounds. Chew. Rev. 103, 4605–4638. Bai, Y., Bai, Z.P., Li, W., 2016. Characteristics and sources analysis of atmospheric volatile organic compounds in the Tibetan Plateau. Acta. Sci. Circum. 36, 2180–2186. Barletta, B., Meinardi, S., Simpson, I.J., Zou, S.C., Rowland, F.S., Blake, D.R., 2008. Ambient mixing ratios of nonmethane hydrocarbons (NMHCs) in two major urban centers of the Pearl River Delta (PRD) region: Guangzhou and Dongguan. Atmos. Environ. 42, 4393–4408. Bashkin, V.N., 2009. Environmental Chemistry: Asian Lessons. Kluwer Academic Pub., Netherlands, Netherlands. Carter, W., 1994. Development of ozone reactivity scales for Volatile Organic Compounds. J. Air Waste Manage. Assoc. 44, 881–899. Chang, C.C., Chen, T.Y., Lin, C.Y., Yuan, C.S., Liu, S.C., 2005. Effects of reactive hydrocarbons on ozone formation in southern Taiwan. Atmos. Environ. 39, 2867–2878. Chen, W.T., Shao, M., Lu, S.H., Wang, M., Zeng, L.M., Yuan, B., et al., 2014. Understanding primary and secondary sources of ambient carbonyl compounds in Beijing using the PMF model. Atmos. Chem. Phys. 14 (6), 3047–3062. de Gouw, J.A., Warneke, C., 2007. Measurement of volatile organic compounds in the Earth's atmosphere using proton-transferreaction mass spectrometry. Mass Spectrom. Rev. 26, 223–257. Derwent, R.G., Jenkin, M.E., Saunders, S.M., Pilling, M.J., Simmonds, P.G., Passant, N.R., et al., 2003. Photochemical ozone formation in northwest Europe and its control. Atoms. Environ. 37, 1983–1991. Duan, J.C., Guo, S., Tan, J.H., Wang, S.L., Chai, F.H., 2012. Characteristics of atmospheric carbonyls during haze days in Beijing, China. Atmos. Res. 114–115, 17–27. Finlayson-Pitts, B.J., Pitts, J.N., 1997. Tropospheric air pollution: ozone, airborne toxics, polycyclic aromatic hydrocarbons, and particles. Science 276, 1045–1052. Jordan, C., Fitz, E., Hagan, T., Sive, B., Frinak, E., Haase, K., et al., 2009. Long-term study of VOCs measured with PTR-MS at a

rural site in New Hampshire with urban influences. Atmos. Chem. Phys. 9, 4677–4697. Kanaya, Y., Cao, R., Akimoto, H., Fukuda, M., Komazaki, Y., Yokouchi, Y., et al., 2007. Urban photochemistry in central Tokyo: 1. Observed and modeled OH and HO2 radical concentrations during the winter and summer of 2004. J. Geophys. Res.-Atmos. 112, D21312. Kansal, A., 2009. Sources and reactivity of NMHCs and VOCs in the atmosphere: a review. J. Hazard. Mater. 166, 17–26. Legreida, G., Loovb, J.B., Staehelinb, J., Hueglina, C., Hilla, M., Buchmanna, B., et al., 2007. Oxygenated volatile organic compounds (OVOCs) at an urban background site in Z̈urich (Europe): seasonal variation and source allocation. Atmos. Environ. 41, 8409–8423. Lin, Y.C., Schwab, J.J., Demerjian, K.L., Bae, M.S., Chen, W.N., Sun, Y., et al., 2012. Summertime formaldehyde observations in New York City: ambient levels, sources and its contribution to HOx radicals. J. Geophys. Res.-Atmos 117, D08305. Lindinger, W., Hansel, A., Jordan, A., 1998. On-line monitoring of volatile organic compounds at pptv levels by means of protontransfer-reaction mass spectrometry (PTR-MS) - Medical applications, food control and environmental research. Int. J. Mass Spectrom. 173, 191–241. Liu, X.W. (Ed.), 2010. Research of the Regional Characteristics of Gaseous Pollutants and Atmospheric Transport in Beijing and its Surrounding Areas. Chinese Academy of Meteorological Sciences, Beijing, China Master Thesis. Liu, Y., Yuan, B., Li, X.P., Shao, M.Z., Lu, S., Li, Y., et al., 2015. Impact of pollution controls in Beijing on atmospheric oxygenated volatile organic compounds (OVOCs) during the 2008 Olympic Games: observation and modeling implications. Atmos. Chem. Phys. 15, 3045–3062. Luecken, D.J., Hutzell, W.T., Strum, M.L., Pouliot, G.A., 2012. Regional sources of atmospheric formaldehyde and acetaldehyde, and implications for atmospheric modeling. Atmos. Environ. 47, 477–490. Ma, Y., Diao, Y.W., Zhang, B.J., Wang, W.W., Ren, X.R., Yang, D.S., et al., 2016. Detection of formaldehyde emissions from an industrial zone in the Yangtze River Delta region of China using a proton transfer reaction ion-drift chemical ionization mass spectrometer. Atmos. Meas. Tech. 9, 6101–6116. Ministry of Ecology and Environment of the People's Repubic of China, 2017. Bulletin of Ecology and Environment of the People's Repubic of China. pp. 7–14. Parrish, D.D., Ryerson, T.B., Mellqvist, J., Johansson, J., Fried, A., Richter, D., et al., 2012. Primary and secondary sources of formaldehyde in urban atmospheres: Houston Texas region. Atmos. Chem. Phys. 12, 3273–3288. Possanzini, M., Tagliacozzo, G., Cecinato, A., 2007. Ambient levels and sources of lower carbonyls at Montelibretti, Rome (Italy). Water Air Soil Pollut. 183, 447–454. Seinfeld, J.H., Pandis, S.N., 2006. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change. Wiley, New York. Sin, D.W.M., Wong, Y.C., Louie, P.K.K., 2001. Trend of ambient carbonyl compounds in the urban environment of Hong Kong. Atmos. Environ. 35, 5961–5969. Wu, Q.Z., Wang, Z.F., Gbaguidi, A., Gao, C., Li, L.N., Wang, W., 2011. A numerical study of contributions to air pollution in Beijing during CAREBeijing-2006. Atmos. Chem. Phys. 12, 5997–6011. Xue, L.K., Wang, T., Guo, H., Blake, D.R., Tang, J., Zhang, X.C., et al., 2013. Sources and photochemistry of volatile organic compounds in the remote atmosphere of western China: results from the Mt. Waliguan Observatory. Atmos. Chem. Phys. 13, 8551–8567. Yang, L.Y., 1990. Several problems in the monitoring of atmospheric background pollution. Plateau Meteorol. 9, 337–346.

J O U RN A L OF E N V I RO N M EN TA L S CI EN CE S 81 (2 0 1 9 ) 1 4 8–1 5 5

Yuan, B., Shao, M., de Gouw, J., Parrish, D.D., Lu, S.H., Wang, M., et al., 2012. Volatile organic compounds (VOCs) in urban air: How chemistry affects the interpretation of positive matrix factorization (PMF) analysis. J. Geophys. Res.-Atmos 117 (D24302), 1–17. Zhang, J.K., Wang, Y.S., Wu, F.K., Sun, J., 2012. Study on atmospheric VOCs in Gongga Mountain base station. Environ. Sci. 33, 4159–4166. Zhang, Y., Wang, X., Barletta, B., Simpson, I.J., Blake, D.R., Fu, X., et al., 2013. Source attributions of hazardous aromatic hydrocarbons in urban, suburban and rural areas in the Pearl River Delta (PRD) region. J. Hazard. Mater. 250, 403–411.

155

Zhang, R.H., Li, Q., Zhang, R.N., 2014. Meteorological conditions for the persistent severe fog and haze event over eastern China in January 2013. Sci. China Earth Sci. 57, 26–35. Zhu, B., Han, Y., Wang, C., Huang, X.F., Xia, S.Y., Niu, Y.B., et al., 2018. Understanding primary and secondary sources of ambient oxygenated volatile organic compounds (OVOCs) in Shenzhen utilizing photochemical age-based parameterization method. J. Environ. Sci. Available https://doi.org/10.1016/j.jes.2018.03.008 Accessed March 9, 2018.