Source Apportionment of VOCs and O3 Production Sensitivity at Coastal and Inland Sites of Southeast China

Received: August 3, 2022
Revised: October 11, 2022
Accepted: October 17, 2022

 Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Cite this article:

Chen, G., Liu, T., Ji, X., Xu, K., Hong, Y., Xu, L., Li, M., Fan, X., Chen, Y., Yang, C., Lin, Z., Huang, W., Chen, J. (2022). Source Apportionment of VOCs and O₃ Production Sensitivity at Coastal and Inland Sites of Southeast China. Aerosol Air Qual. Res. 22, 220289. https://doi.org/10.4209/aaqr.220289

HIGHLIGHTS

• Online and offline VOCs were measured at coastal and inland sites of southeast China.
• C2-C4 alkene species contributed the largest to the OH radical loss and ozone production.
• The contributions of biogenic sources and fuel evaporation were larger at inland sites.
• O₃ production at coastal and inland sites were mainly VOC control.

Keywords: Volatile organic compounds, Source apportionment, Ozone formation, Sensitivity analysis, Emission reduction

1 INTRODUCTION 

Volatile organic compounds (VOCs) play a significant role in the formation of ozone (O₃) and secondary organic aerosol (SOA) (Zhao et al., 2020; Song et al., 2021), which can enhance atmospheric oxidation capacity, and affect air quality and human health (Bari and Kindzierski, 2018; Agathokleous et al., 2020). In China, although the concentrations of PM_2.5 show a decreasing trend, the photochemical pollution of O₃ has attracted more and more attention. As the precursor of O₃, VOCs have a wide variety of sources and reactive activities, which brings great challenges to the control of O₃ formation. Therefore, it is vital to identify particular VOC sources and reduce VOC emissions to decrease O₃ formation.

Population densities, industrial structure, and local regulations for VOC emissions directly affect the concentrations and sources of VOCs (Li et al., 2019a). In developed urban areas, the average concentrations of TVOCs were relatively higher than those in suburban or remote areas, and vehicle exhausts, solvent usage, and industrial emissions were generally primary contributors of VOCs (Liu et al., 2017; Song et al., 2018; Liu et al., 2021). In addition, combustion sources might be a major source of VOCs in medium-sized urban regions due to frequent agricultural activities such as biomass burning (Qin et al., 2021). Additionally, different VOC groups have significant effects on O₃ formation, attributed to the difference of their chemical reactivity. Previous studies suggested that aromatics and alkenes contribute the largest to the ozone formation potential (OFP) (Song et al., 2018; Li et al., 2019b). Thus, aromatics and alkenes were considered as key VOC groups to control their emissions. However, controlling atmospheric O₃ formation is much more complex than what had been expected, due to the fact that O₃ formation is non-linearly related to VOCs and NO_x. Field observation and model results indicated that O₃ formation is generally more sensitive to VOCs in most developed urban regions, but more sensitive to NO_x in rural regions (Li et al., 2018; Song et al., 2021; Meng et al., 2022).

Current studies of atmospheric VOCs in China were mainly concentrated in developed regions, such as Beijing-Tianjin-Hebei (BTH), Yangtze River Delta (YRD), Pearl River Delta (PRD), and Chengdu-Chongqing regions (Yuan et al., 2013; Gong et al., 2017; Liu et al., 2017; Li et al., 2018; Zhao et al., 2020; Liu et al., 2021; Song et al., 2021). However, few studies were carried out in developing urban or county areas, with different levels of urbanization and industrialization. In these areas, biogenic VOCs and their interaction with anthropogenic VOCs and NO_x emissions were undoubtedly essential in affecting O₃ formation. Ningde is a developing coastal city in the southeast of China (Fig. 1). With the relatively low PM_2.5 concentrations, O₃ pollution in Ningde has frequently occurred in recent years, and O₃ became primary pollutants in the evaluation system of ambient air quality from 2015 to 2020 (http://fjaqi.fjemc.org.cn/). Due to convenient traffic conditions, the coastal areas in Ningde have more developed industrial and economic levels than those in inland areas (Table S1). At the same time, population in coastal areas has been increasing in the past twenty years, due to the process of rapid urbanization and industrialization (Fig. S1). Given that different levels of urbanization and industrialization may bring different influence on O₃ and its precursors, it is necessary to study sources of VOCs and the sensitivity of O₃ formation in two types of areas with different levels of urbanization and industrialization to control O₃ concentrations in the ambient air.

In this study, VOCs, criteria air pollutants and metrological parameters were simultaneously measured in coastal and inland areas in Ningde, using the positive matrix factorization (PMF) and the observation-based model (OBM), to determine major sources of VOCs and their effects on O₃ formation. The study aims to (1) explore seasonal patterns of VOCs at coastal sites and inland sites, and (2) quantify the major sources of VOCs, and (3) identify key VOC species of O₃ formation and analyze the sensitivity of O₃ formation. This study will provide a helpful reference for the control measures to reduce O₃ concentrations in the ambient air in developing urban areas with different levels of urbanization and industrialization.

 
2 METHODS

 
2.1 Study Area

A total of eight offline sampling sites were selected, including Yizhong (YZ), Jiancezhan (JCZ), Fu’an (FA), Fuding (FD), Xiapu (XP) as coastal sites (CS), and Gutian (GT), Pingnan (PN), Zhouning (ZHN) as inland sites (IS) (Fig. 1). They are all typical urban sites without direct pollutant emissions, and details about the study sites were summarized in supplementary material. Additionally, the online VOC observation was conducted in the Ningde Meteorological Bureau (119.54°E, 26.64°N) in coastal regions in Ningde, and the detailed description of the online analytical instrument of VOCs can be found in supplementary material. In this study, the online monitoring of VOCs was mainly used to analyze the relation of O₃ formation sensitivity and propose control strategies for both VOC and NO_x emission in the area of Ningde.

Fig. 1. Observations at coastal sites (CS, black dot) and inland sites (IS, red triangle) in Ningde in the southeast of China.

 
2.2 Sampling and Observation

Offline VOC sampling was conducted simultaneously at eight sampling sites during May, August, and November in 2019, representative for spring, summer and autumn, respectively, and every month had 3–5 days for VOC sampling. Manual VOC samples were taken each day at, 6–7, 11–12, 13–14, 17–18, 20–21, 23–24, respectively. A total of 354 samples at CS and 103 samples at IS were collected with 3.2 L Summa canisters (model 3100A, Entech Instruments Inc., USA). To avoid interference from targeted VOCs, all canisters were flushed by ultra-pure N₂ at least three times before VOC sampling.

Offline VOC samples were analyzed based on a Model 7100 Pre-concentrator (Entech Instruments Inc., USA) combined with a gas chromatograph equipped with a mass spectrometer (GC/MS, Shimadzu, Japan), according to our previous studies (Hu et al., 2018; Hong et al., 2019). Briefly, we used liquid-nitrogen glass beads cryogenic trap to condense VOC samples at –150°C. And the trapped VOCs were transferred through pure He into a further trap at –30°C with Tenax as absorbent. Then, the target VOCs were transferred to a third capillary cryo-focus trap by heating the secondary trap at –160°C. Finally, the preconcentrated VOCs were transferred into the GC/MS system. The GC oven temperature was programmed to be initially at –50°C holding for 7 min, then raised to 180°C at 4°C m^–1, finally, raised to 220°C at 15°C m^–1 holding for 3 min. The ionization technique was electron impacting (EI, 70 eV) with the ion source temperature of 230°C. The calibration curve should be recalibrated when its responses were in excess of ± 15% of the original calibration curve. The correlation coefficients (R²) of the standard curve varied from 0.990 to 1.000. Totally, 102 VOCs were detected, including 28 alkanes, 10 alkenes, 18 aromatics, 11 oxygenated VOCs (OVOCs), 34 halocarbons, and carbon disulfide.

In addition, the concentrations of criteria air pollutants (O₃, SO₂, CO, NO, NO₂) and meteorological parameters including temperature (T), relative humidity (RH), and atmospheric pressure (P) were synchronously detected at each sampling site. NO_x (NO/NO₂), CO, SO₂, and O₃ were measured using gas analysis instruments (Thermo Fisher Scientific 42i, 48i, 43i, and 49i, respectively, USA).

 
2.3 PMF Analysis

The PMF model is widely used to determine sources and contributions of VOCs. Its principle and instructions have been introduced in previous studies (Bon et al., 2011; Yuan et al., 2012; Abeleira et al., 2017). Details about the model were summarized in the supplementary material.

 
2.4 OBM Analysis

The OBM was used to explore the atmospheric oxidation processes and evaluate the sensitivity of O₃ production to its precursors by comparing their relative incremental reactivity (RIR) values, calculated by Eq. (1).

where, ΔP(O₃)/P(O₃) represents the change of the O₃ production rate; ∆X/X is the percentage of O₃ precursors reduction, adopted as 20%. Details of the OBM can be found in previous studies (Liu et al., 2022a, 2022b). In brief, the observed data including VOCs, air pollutants (i.e., CO, NO₂, NO, SO₂, O₃), and meteorological parameters (i.e., T, P, and RH) were input to constrain the model. In this study, the Empirical Kinetics Modeling Approach (EKMA) curve takes the average values of measured species other than NO_x and VOCs as the constraints of the model. Given that the control of biogenic VOC emissions (BHC) are more difficult than the control of anthropogenic VOC (AHC) emissions, isoprene is set as a fixed value and no sensitivity experiment is performed on it. A total of 361 scenarios were performed, assuming that 19 emission scenarios were for NO_x and AHC activity, respectively, to simulate O₃ production under the corresponding scenario to plot O₃ isopleth diagrams. In this study, S(VOCs), S(NO_x) and S(O₃) is VOC, NO_x, and O₃ concentrations, respectively.

 
3 RESULTS AND DISCUSSION

 
3.1 Characteristics of VOCs

3.1.1 Overview of characteristics

Average concentrations and percentages of VOCs during the sampling period are shown in Fig. 2. The average TVOC concentrations at CS and IS were 49.1 ± 14.4 and 28.4 ± 9.6 ppb, respectively. Compared to other urban regions, TVOC concentrations at IS were much lower than those in densely populated megacities, such as Beijing (64.9 ppb) (Gao et al., 2018), Shanghai (57.5  ppb) (Liu et al., 2019), Guangzhou (63.6 ppb) (Zou et al., 2015). However, TVOC concentrations at CS were close to those in the megacities, reflecting the rapid urbanization in coastal regions of southeast China resulting in more anthropogenic pollutant emissions. At CS, the concentrations of alkanes were the highest (21.6 ± 9.0 ppb) among all VOC groups, followed by OVOCs (11.3 ± 4.2 ppb), halocarbons (9.4 ± 3.4 ppb), aromatics (3.9 ± 2.3 ppb), and alkenes (2.4 ± 1.8 ppb). Similar pattern of VOC groups at IS was found, with the predominant species of alkanes (14.1 ± 5.9 ppb), followed by OVOCs (5.9 ± 2.3 ppb), halocarbons (4.7 ± 2.1 ppb), alkenes (1.9 ± 1.2 ppb), and aromatics (1.7 ± 1.2 ppb). Overall, the concentrations of different types of VOCs at CS were higher than those at IS, due to stronger anthropogenic emissions at CS. In coastal regions in Ningde, there are a busier traffic network and higher population density than those in inland regions, as well as the rapid development of port and industry (Table S1), causing more frequent anthropogenic activities in coastal regions. According the statistics of 2017 Ningde air pollutants emission inventory, the amounts of VOC emissions from industrial production at CS and IS were 20970.4 and 1374. 4 ton per year, respectively (Fig. S2).

Fig. 2. The average concentrations and percentages of different types of VOCs during the sampling period at CS and IS.

For VOC species, the concentrations of the top 10 VOC species at CS and IS accounted for 56.9% and 65.8% of TVOC concentrations, respectively (Table S4). Meanwhile, propane and acetone were the most abundant VOC species at CS and IS in Ningde, similar to those in other cities of Southeast China, such as Fuzhou city and Xiamen city (Hong et al., 2019), likely related to emissions from the transportation process. In addition, alkane species, including propane, iso-pentane, iso-butane, and n-butane, were the largest contributors of the top 10 VOC species at CS and IS. Also, n-butene, toluene, dichloromethane and chloromethane, were the most abundant species of alkenes, aromatics, and halocarbons at CS and IS, respectively.

 
3.1.2 Spatial and temporal variations

Table S5 shows spatial and temporal variations of VOC groups at JCZ, YZ, XP, FA and FD as CS, at GT, PN, and ZHN as IS. At CS, TVOC concentrations in autumn were the highest, followed by summer and spring; at IS, TVOC levels in autumn were also the highest, followed by spring and summer, due to the difference of the emission intensity of anthropogenic activities, meteorological factors during the sampling period (Li et al., 2019a), similar seasonal patterns in Guangzhou (Zou et al., 2015). In addition, TVOC levels at XP among CS were the largest, followed by JCZ, YZ; among IS, GT had the highest TVOC levels. Among VOC groups, alkanes were the most abundant groups because of its widespread sources and longevity, accounting for 45.7%, 47.5%, 45.9%, 36.9%, 37.2%, 52.3%, 47.2%, 50.1% of TVOC levels at JCZ, YZ, XP, FA and FD as CS, at GT, PN, and ZHN as IS, respectively. At CS and IS, the highest levels of alkanes were observed at XP attributed to stronger contributions of propane, n-butane, and isobutane, mainly emitted from the transportation process. It is reasonable because XP has developed tourism and maritime transportation, causing higher concentrations of related VOC species.

In this study, major anthropogenic VOC species including propane, n-butene, chloromethane, benzene (B), toluene (T), ethylbenzene (E), and m/p-xylene (X) were selected to compare the difference of emission sources at CS and IS (Table 1). Propane is an indicator of vehicle exhausts, liquefied petroleum gas (LPG) evaporation, and combustion emissions (An et al., 2014). The average concentrations of propane presented a significant difference between CS and IS in summer and autumn as a result of large amount of transportation emissions at CS. n-Butene is generally used as raw material for industrial products (Chen et al., 2014), serving as an indicator of industrial sources. The concentrations of n-butene at CS were higher than those at IS during the sampling period, especially in autumn, consisted with more industrial production activities at CS. Chloromethane is mainly associated with combustion emissions (McCulloch et al., 1999; Harper, 2000; Keppler et al., 2005; Liu et al., 2008; Song et al., 2018, 2021). In our study, chloromethane levels at IS were higher than those at CS, especially in autumn, due to the influence of more frequent agricultural biomass burning at IS during the harvest season. BTEX are mainly emitted from vehicle exhaust, industrial production, and solvent usage (An et al., 2014; Dumanoglu et al., 2014). In this study, seasonal variations of BTEX at CS and IS were not obvious, probably related to the combined effects of complex sources and photochemical loss (Yuan et al., 2012; Chen et al., 2014). The concentrations of benzene at CS and IS were similar, but ethylbenzene, m/p-xylene, and toluene at CS were higher than those at IS, implying that the impacts of vehicle exhaust, industrial production, and solvent usage emissions may be stronger at CS.

3.2 Photochemical Reactive Activity of VOCs

The concentrations, L_OH, and OFP of different VOC groups are presented in Fig. 3, and the L_OH, and OFP of the top 10 VOC species are summarized in Table S6. Totally, the L_OH showed an obvious seasonal pattern at CS and IS, with the maximum value in autumn, the minimum value in spring, likely due to higher concentrations of related VOC species in autumn. Besides, the L_OH at CS were higher than that at IS during the spring, summer, autumn sampling period, indicating more active photochemical activities at CS with higher levels of VOCs. Among the L_OH of VOC groups at CS and IS, the L_OH of alkenes was the largest, followed by alkanes, aromatics, and OVOCs. The levels of alkanes were the highest, but the L_OH of alkanes was lower than that of alkenes. Although the levels of alkenes were very low, the L_OH of alkenes was the highest among all VOC groups, attributed to higher photochemical reactivity of alkenes with the OH radical. Because alkenes had the highest L_OH, it is vital to identify key alkene species that contributed the most abundant to the L_OH. Among the top 10 VOC species, n-butene, isoprene, and propene at CS accounted for 32.6% of the L_OH; n-butene, propene, isoprene, cis-2-butene, and trans-2-butene at IS accounted for 49.4%. Meanwhile, isoprene in summer contributed the largest to the L_OH at CS and IS, and isoprene at IS may be more important for the L_OH in spring and autumn. Overall, the role of C2-C4 alkene species in atmospheric chemistry should not be ignored.

Fig. 3. The concentrations, L_OH, and OFP of different VOC groups at CS and IS in Ningde.
 

The L_OH can only evaluate the photochemical reactivity of different VOC species, and the OFP can determine the contributions of different VOC species to O₃ formation. The largest OFP at CS and IS was in autumn, followed by summer and spring. Similar to the L_OH, alkenes, alkanes, and aromatics were the major contributors to the OFP, and the OFP of alkenes was more than twice as high as the OFP of alkanes, aromatics, OVOCs, and halocarbons. The average OFP at CS (30.3 ppb) was higher than that at IS (21.1 ppb), suggesting that the photochemical O₃ pollution at CS may be more serious. The top 5 VOC species of the OFP at CS were n-butene, propene, toluene, propane, and isoprene, accounting for 56.6%; the top 5 VOC species of the OFP at IS were propene, n-butene, isoprene, propane, and i-pentane, accounting for 58.1%. It should be mentioned that the importance of isoprene to the OFP at IS should be valued.

 
3.3 Source Apportionment of VOCs

3.3.1 Diagnostic ratios

Ratios of certain VOC species, such as i-pentane/n-pentane, o-xylene/m/p-xylene, and cis-2-butene/trans-2-butene, are used to preliminarily identify the sources of anthropogenic VOC emissions. In this study, the correlation coefficient values between i-pentane and n-pentane (0.81 and 0.93), o-xylene and m/p-xylene (0.97 and 0.98), cis-2-butene and trans-2-butene (0.98 and 0.98) were observed during the sampling period (Fig. S3). The ratios of i-pentane/n-pentane can be used to identify combustion and vehicle exhaust sources, i.e., 0.56–0.8 for combustion sources and 2.93 for vehicle emissions (Liu et al., 2008; Yan et al., 2017). Besides, the ratios of o-xylene/m/p-xylene (0.35), cis-2-butene/trans-2-butene (1.14) were corresponded to traffic emissions (Liu et al., 2008; Yuan et al., 2010). In this study, the ratios of i-pentane/n-pentane at CS and IS were 2.64 and 3.44, respectively (Fig. S3(a)). The ratios of o-xylene/m/p-xylene, cis-2-butene/trans-2-butene at CS were 0.55 and 1.29, and were 0.84 and 1.34 at IS, respectively (Figs. S3(b) and S3(c)), indicating that vehicle exhausts may play an important role for VOC emissions at CS and IS.

 
3.3.2 PMF analysis

According to the PMF model analysis, five VOC sources were identified, including solvent usage, industrial sources, vehicle exhaust and combustion sources, fuel evaporation, and biogenic sources (Fig. S4). Fig. 4 shows the percentages of different sources to VOC concentrations (inner pie chart), the OFP (outer pie chart) in Ningde, and their contributions of VOC sources at CS and IS, indicating that vehicle exhaust and combustion sources (40.9%) were the largest VOC sources, followed by industrial sources (21.7%), fuel evaporation (13.1%), biogenic sources (13.0%), and solvent usage (11.3%). The contributions of different VOC sources to the OFP were as follows: vehicle exhaust and combustion sources (31.8%) contributed the most abundant, followed by fuel evaporation (21.3%), industrial sources (20.0%), solvent usage (15.8%), and biogenic sources (11.3%). Therefore, vehicle exhaust and combustion sources were the major VOC sources and had the largest OFP, similar to Zhengzhou (Li et al., 2017). In addition, the contributions of vehicle exhaust and combustion sources, industrial sources, solvent usage, biogenic sources, and fuel evaporation were 42.2%, 22.4%, 12.7%, 11.7%, and 11.0% at CS, and 34.5%, 18.2%, 4.5%, 19.4%, and 23.4% at IS, respectively. Compared to those at IS, the contributions of vehicle exhaust and combustion sources, industrial sources, and solvent usage at CS were obviously larger. However, the contributions of fuel evaporation and biogenic sources were more predominant at IS. It should be noted that the quantity of solvent usage at CS was over two times higher than that at IS, likely attributed to the influence of road asphalt pavement and household appliances coatings. The percentage of fuel evaporation at IS was larger than that at CS, due to the lack of control measures, such as fuel recovery in oil station. Owing to the high forest coverage rate, the percentage of biogenic sources at IS was almost two times higher than that at CS. Overall, vehicle exhaust and combustion sources, industrial sources, and fuel evaporation were the most abundant VOC sources contributing to O₃ production. In coastal regions in Ningde, the government needs to strengthen the management of road mobile sources and ship transportation emissions, whereas in inland regions in Ningde, biomass burning activities should be strictly enforced to be prevented, and the fuel leakage in oil stations should be inhibited.

Fig. 4. The percentages of (a) VOC concentrations (inner pie chart), OFP (outer pie chart), and five VOC sources at (b) CS and (c) IS.

 
3.4 Sensitivity Analysis of O₃ Production and Policy Implications

3.4.1 Ratio of VOCs and NO_x

The VOCs/NO_x value is usually used to determine whether O₃ production in a region is controlled by VOCs or NO_x. When the VOCs/NO_x value is lower than 8, the VOCs is the key factor for the O₃ formation in the region; the region will be controlled by NO_x if the value is higher than 8. In this study, the average VOCs/NO_x value at CS was 6.3 ± 3.8, with a seasonal order of summer (9.5 ± 3.6) > spring (4.8 ± 3.1) > autumn (4.4 ± 2.1). Besides, the average VOCs/NO_x value at IS was 6.5 ± 3.3, with a seasonal order of summer (6.9 ± 4.2) > autumn (6.8 ± 3.2) > spring (5.9 ± 2.1) (Fig. S5). Therefore, O₃ formation at CS was more sensitive to VOCs in spring and autumn, but it was controlled by VOCs and NO_x in summer; for IS, O₃ formation was mainly controlled by VOCs in spring, summer, autumn.

Because the coastal regions faced the problem of O₃ and its precursors pollution, the following contents focused on analyzing controlling factors of O₃ formation based on the OBM in coastal regions. The online observation of VOCs was only conducted in the Meteorological Bureau in coastal regions of Ningde city, thus, we analyzed the O₃ formation sensitivity and emission reduction strategies at the urban site.

 
3.4.2 RIR analysis

The O₃ formation process was simulated, and RIR values of O₃ precursors were calculated during the sampling period. The RIR values of CO, NO_x, biogenic VOCs (BHC), anthropogenic VOCs (AHC) are displayed in Fig. 5. The RIR values of VOCs and NO_x were positive and negative, respectively, in spring and autumn, indicating that O₃ formation is more sensitive to VOCs. However, the positive RIR values of VOCs and NO_x indicates that O₃ formation is controlled by both VOCs and NO_x in summer. AHC had the largest RIR values, followed by NO_x, BHC, CO, especially in spring and autumn, implying that AHC had the largest effects on O₃ formation. Hence, the control of AHC was more important to decrease O₃ pollution. Among AHC, the RIR values of alkenes were the largest, followed by aromatics, alkanes, especially in spring and autumn. Therefore, the reduction of alkene and aromatic emissions was helpful to decrease O₃ concentrations in spring and autumn, while abating NO_x levels means the aggravation of O₃ pollution. In summer, the reduction of both VOC and NO_x emissions would suppress the formation of O₃.

Fig. 5. The RIR values of O₃ precursors in different seasons at the urban site in coastal regions of Ningde.

 
3.4.3 Emission reduction policies

The OBM was also used to make the scenario analysis with reduction by 0-100 % at intervals of 10% for the reduction of anthropogenic VOCs (ΔS(VOCs)/S(VOCs)) and NO_x (ΔS(NO_x)/S(NO_x)) for control policies of O₃ pollution. The EKMA results showed that O₃ formation was more sensitive to VOCs in spring and autumn, and controlled by VOCs and NO_x in summer (Fig. S6), consistent with the results of the VOCs/NO_x value and RIR analysis. In addition, the maximum value of MDA8h O₃ during the observation period was 91.5 ppbv, exceeding the national air quality standard of 81.5 ppbv by 12%, hence, 5%, 10%, and 15% reductions targets of O₃ levels were considered to discuss the reduction proportions of VOCs and NO_x (Fig. 6). As VOC control in spring, the appropriate way to achieve the 5% O₃ reduction target was as follow: S(VOCs) is reduced by 20%, and S(NO_x) is reduced by 3%. It is reasonable to reduce 30% of S(VOCs) and 3% of S(NO_x) to achieve a 10% O₃ control target. In order to realize the 15% O₃ control goal, the reduction of S(VOCs) and S(NO_x) is 50% and 10%, respectively. The reduction schemes in autumn were similar to those in spring. In summer, the 5% of O₃ control target can be realized by the 2% reduction of S(NO_x) and 10% reduction of S(VOCs) and achieving a 15% O₃ control target needed to reduce 4% S(NO_x) and 30% S(VOCs). To reach a 15% reduction of O₃ levels control target, more strict emission reduction measures should be taken. Simultaneously, it is vital to adjust the schemes on the VOCs and NO_x emission reduction in time when O₃ formation sensitivity changes.

 Fig. 6. Simulated O₃ concentrations response to the reduction of S(NO_x) and S(VOCs) in different seasons at the urban site in coastal regions of Ningde.
 

4 CONCLUSIONS

In this study, spatial and temporal characteristics of VOCs at coastal and inland sites with different levels of urbanization were investigated, and sources apportionment of VOCs and the sensitivity of O₃ formation were evaluated using the positive matrix factorization (PMF) and the observation-based model (OBM). The average TVOC concentrations at CS (49.1 ± 14.4 ppb) was obviously higher than those at IS (28.4 ± 9.6 ppb), indicating stronger anthropogenic activities emissions at CS. Alkenes played an important role in the L_OH and OFP, and C2–C4 alkene species contributed the most abundant. The major contributors of VOC sources at CS were vehicle exhaust and combustion sources (42.2%), industrial sources (22.4%), solvent usage (12.7%), while the contributions of fuel evaporation (23.4%) and biogenic source (19.4%) were relatively larger at IS. O₃ formation at CS was controlled by VOCs in spring and autumn, and mainly VOC and NO_x control in summer, but was controlled by VOCs at IS in spring, summer, and autumn, and the control of alkenes and aromatics emissions could reduce O₃ concentrations. The OBM simulated results showed that 20% reduction of S(VOCs) and 3% reduction of S(NO_x) could realize the 5% reduction of S(O₃). It is vital to adjust the schemes on the VOCs and NO_x emission reduction if the changing of O₃ formation sensitivity happens.

 
ACKNOWLEDGEMENTS

This study was funded by the Science and Technology Department of Fujian Province (2022L3025), the National Natural Science Foundation of China (42277091), Center for Excellence in Regional Atmospheric Environment Project (E0L1B20201), Xiamen Atmospheric Environment Observation and Research Station of Fujian Province, and Fujian Key Laboratory of Atmospheric Ozone Pollution Prevention (Institute of Urban Environment, Chinese Academy of Sciences).