Impact of industrial activities on atmospheric volatile organic compounds in Sihwa-Banwol, the largest industrial area in South Korea

The impact of industrial activities on atmospheric volatile organic compounds (VOCs) in the Sihwa-Banwol complexes, i.e., the largest industrial area in Korea, was investigated. More than 60 VOCs were determined from 850 samples collected from four sites in and around the complexes through a 2-year monitoring campaign from 2005 to 2007. The VOCs of particular concern found in the area were benzene, toluene, ethylbenzene, xylenes, trichloroethylene, and formaldehyde, given their toxicity, concentration, and detection frequency. Toluene was the most abundant one. The VOC concentration rankings were consistent with their emission rankings. Most VOCs had higher concentrations at the industrial sites than at residential sites, indicating a significant impact of industrial emissions. The ambient levels of benzene and formaldehyde were additionally affected by vehicular emissions and secondary formation, respectively. Overall, the VOC levels increased in winter and at night, because of the local weather conditions. In contrast, the formaldehyde concentration increased in summer, owing to its secondary formation in the atmosphere. The ambient VOC levels in Sihwa-Banwol were higher than those in other parts of Korea. Additionally, the cumulative cancer risks posed by the toxic VOCs exceeded a tolerable risk level of 1 × 10⁻⁴ in not only the industrial areas but also the residential areas. The sum of the non-cancer risks in both areas significantly exceeded the threshold criterion of 1. The large amounts of aromatic compounds emitted from the industrial complexes are believed to play a crucial role in the elevated levels of surface ozone in the Seoul metropolitan area during the summer season. Therefore, comprehensive measures for controlling the VOC emissions in the Sihwa-Banwol area need to be prioritized to reduce the health risks for residents of not only this area but also the capital Seoul and its surrounding areas.

Keywords: VOCs; BTEX; HAPs; Risk assessment; POCP; Industrial complex

Electronic supplementary material The online version of this article (10.1007/s11356-020-09217-x) contains supplementary material, which is available to authorized users.

Sihwa-Banwol is a commonly used name in Korea that refers to the two adjacent Sihwa and Banwol National Industrial Complexes in the country. The administrative district of this industrial area spans across the two cities of Siheung and Ansan, in the west end of the Gyeonggi Province, approximately 30 km southwest of the capital—Seoul. The combined population of the two cities as of 2008 was 1.1 million (KOSIS [33]), while the total population of Seoul and Gyeonggi Province was ~ 20 million.

Developed in 1975, the Banwol complex was established for the purpose of relocating pollution-causing industries scattered across the metropolitan area. However, owing to the rapid economic growth in the 1970s, it soon became saturated with ≥ 1000 firms, following which the Sihwa Lake reclamation project was started in 1977 to solve the land-shortage problem. The Sihwa complex was finally opened in 1987 and was twice the size of Banwol. These two complexes significantly contributed to solving the problem of land shortage for industrial factories in the metropolitan area, as well as to the development of the "West Coast Industrial Belt" in Korea. As of 2008, the combined complex accommodated ≥ 10,000 small and medium-sized businesses, and it is the largest industrial area in Korea (KICC [23]).

The Sihwa-Banwol complexes house a variety of densely located air-polluting and odor-causing industries, including machinery, petroleum and chemical, electronics, leather, metal processing, medicine, and rubber product manufacturing (KMOE [30]). Additionally, the geographical position of this industrial area, which is located in the windward direction of the nearby residential areas, causes it to spread air pollutants and odor to the residential areas owing to the sea–land breeze. Moreover, the industrial and residential areas are close to each other; hence, many civil complaints have been filed against the industries for causing odor in Siheung and Ansan since the 1990s (Jeon [20]; Kim et al. [26]; KNIER [31]). Consequently, the Korean Ministry of Environment (MOE) conducted raids on the odor-causing companies in the Sihwa-Banwol area and established comprehensive anti-odor measures and strict control to curb the problem (KMOE [29]). In addition to the MOE, the local governments in the area have implemented numerous projects to reduce the odor (Jeon [20]; Song [51]).

Given numerous civil complaints related to odor, a vast majority of projects and studies pertaining to the Sihwa-Banwol area have focused mainly on the odor issues (Byeon et al. [7]; Choi et al. [10]; Paeng et al. [45]). However, in addition to large amounts of odor-causing substances, many small businesses in various industrial sectors are likely to emit hazardous air pollutants (HAPs), such as toxic volatile organic compounds (VOCs). Many chemicals belonging to the VOC group can hardly be recognized by their smell yet have high toxicity. Hence, residents staying near industrial sources may have the risk of exposure to toxic VOCs without actually sensing them, which is why resolving the issue of VOC-related air pollution in this area should be a high priority, in addition to solving the odor problem. However, few studies have been performed on VOC pollution in the Sihwa-Banwol area.

Air quality management strategies in Korea have traditionally focused on the reduction of emissions of criteria pollutants, such as SO2 and particulate matter. Nevertheless, reducing the human health risks has become the primary goal for many developed countries in recent years (Bruhl et al. [6]; EEA [14]; USEPA [62]; WHO [66]). Risk assessments of specific air pollutants require a variety of input factors, among which ambient air monitoring is a key step toward providing the exposure levels for the general public (WHO [66]). However, with regard to the Sihwa-Banwol area, the ambient concentration data available for non-criteria pollutants, such as VOCs, are not adequate for policy makers to develop an appropriate control strategy.

The occurrence of VOCs in the ambient air is a complex phenomenon influenced by various factors, including vehicular exhaust, the use of organic solvents, industrial activities, atmospheric chemistry, and weather conditions. The environmental significance of VOCs can be summarized in two respects: firstly, most VOCs play an important role in the formation of tropospheric ozone, carbonyls, and secondary organic aerosols in the atmosphere; secondly, some VOCs, such as aromatics, aldehydes, and chlorinated hydrocarbons, may have harmful effects on human health. The toxic VOCs in ambient air are generally associated with vehicular exhaust and industrial emissions (Baek et al. [4]; McCarthy et al. [39]; Tiwari et al. [53]). Although vehicular exhaust is probably the most common source of VOCs in typical urban areas, industrial emissions may be a more significant contributor in industrial cities such as Siheung and Ansan. Therefore, the adverse impacts of industrial activities on the ambient concentrations of VOCs in nearby residential areas are a matter of great concern in industrial cities (Tiwari et al. [53]).

In this context, the objective of the present study was to evaluate the impact of industrial activities on the concentrations of atmospheric VOCs in the Sihwa-Banwol area. We investigated the occurrence of a variety of VOCs and their distribution characteristics in the ambient air, along with the spatial and temporal variations in their concentrations. The obtained VOC measurements were interpreted in relation to the emission data from various sources in the industrial complexes. Additionally, we attempted to identify and prioritize important VOCs in this area with respect to their contributions to health risks and surface ozone formation. The obtained results indicate the major VOCs that need to be controlled in this area.

Four sampling sites (two industrial and two residential) were selected to measure the atmospheric VOC concentrations in the Sihwa and Banwol areas (Fig. 1). All the sites were located at the same place as the national ambient air-quality monitoring stations in the area, i.e., Jeongwang2-dong (site A, Sihwa industrial), Jeongwangbon-dong (site B, Sihwa residential), Wonsi-dong (site C, Banwol industrial), and Ansan Station (site D, Banwol residential). The two industrial sites (A and C), were located on rooftops at elevations of ~ 10 m from the ground, with relatively low-traffic 6-lane and 4-lane roads running adjacent to the respective sampling sites. The two residential sites (B and D) were also located on rooftops, ~ 14 to 20 m above ground level. Site B was surrounded by apartment complexes and located near a shopping center with relatively heavy traffic, and site D was located at a subway station with a nearby parking lot and an 8-lane roundabout road.

Graph: Fig. 1 Locations of VOC sampling sites in the Sihwa and Banwol area, Site A: Sihwa industrial site; Site B: Sihwa residential site; Site C: Banwol industrial site; Site D: Banwol residential site. AWS Automatic Weather Station

The VOC sampling was conducted over a period of 2 years from 2005 to 2007, with August 2005 to July 2006 comprising the first year and October 2006 to September 2007 comprising the second year. Depending upon the objective for each year, the monitoring protocol between the first and second years differed slightly with regard to the duration of the sampling period and the frequency of sample collection. This is because the main objective during the first year was to determine the optimal method for measuring the VOCs by conducting trials involving two sampling methods: adsorbent tube and canister. The adsorbent-tube method was evaluated as the more suitable one for this study and was thereafter primarily utilized to collect the VOC samples in the second year.

In the first year, samples were collected for seven consecutive days at each sampling site for two seasons: summer (June to August) of 2005 and spring (March to May) of 2006. In the summer season, six samples were collected per day (from 09:00 to 09:00, at intervals of 4 h), and in the spring, two samples were collected per day (the first from 09:00 to 13:00, and the second from 13:00 to 17:00). In the second year, samples were collected for 10 consecutive days at each site for four seasons, i.e., including the autumn (September to November) and winter (December to February) seasons. In this case, sampling at the industrial sites was conducted six times a day (similar to summer sampling in the first year), and sampling at the residential sites was conducted twice a day (similar to spring sampling in the first year). The sampling duration for all samples was 4 h, but some samples were lost because of equipment malfunction. Thus, a total of 850 samples were obtained in this study: 580 from the industrial sites and 270 from the residential sites. The total number of samples obtained from each site, according to the season, is summarized in Table S1 in the Supplementary Materials (SM).

The weather conditions corresponding to the sampling periods are presented in Table S2 (SM). Hourly weather data for the measurement period were obtained from the Automatic Weather Station in Siheung City (KMA [28]). The rainfall in the study area during each measurement period followed the typical pattern observed in South Korea, with the largest amount in the summer and the smallest in the winter. The seasonal wind directions in the study area were mostly in the range of west to north, although the wind speed varied substantially according to the season (far lower in winter and autumn than in spring and summer).

The VOC measurement method adopted in this study was principally based on the TO-17 methodology established by the US Environmental Protection Agency (USEPA) (USEPA [55]). The VOCs were collected by drawing air through a stainless steel tube (0.6 cm × 9 cm, Perkin Elmer, UK) containing 100 mg of Carbotrap C or Tenax-TA (at the front) and 300 mg of Carbotrap (at the end of the tube). This combination is known to have an excellent adsorption and desorption capacity for variety of VOCs having different physical and chemical properties (Ma et al. [38]; Woolfenden [67]). For continuous sampling over a single day at each site, sequential automatic tube samplers (STS25, Perkin Elmer, UK) were used, together with low flow rate pumps (FLEC Air Pump 1001, Chematec, Denmark), equipped with built-in mass flow controllers. The flow rate was set at 50 mL/min per sample. Swagelok stainless-steel caps, nuts, and polytetrafluoroethylene (PTFE) ferrules were used to seal the tubes, and each tube was placed in a 50-mL glass vial with a PTFE septum cap during the transport and storage of the samples. Prior to sampling, the adsorbent tubes were pre-conditioned using a thermal conditioner (TC-20, Markers Inc., UK) at 250 °C for 2 h, with a helium gas flow of 80 mL/min. The preconditioning was repeated at 300 °C for another 2 h. To ensure that the tubes were absolutely clean for further sampling, one tube was randomly selected from each batch of 10 tubes and analyzed via thermal desorption and gas chromatography–mass spectrometry (GC–MS).

The analysis of VOCs was performed using an automatic thermal desorption apparatus (Unity/Ultra, Markes Inc., UK) and a GC–MS system (HP6890/5973, Hewlett-Packard, USA). The thermal desorption of VOCs adsorbed onto the tube was conducted at 300 °C with a flow rate of 50 mL/min for 10 min (primary desorption). The eluted VOCs were transferred to a cold trap (packed with 12 mg of Tenax-TA and 47 mg of Carbotrap) maintained at − 15 °C, following which the cold trap was rapidly heated to 320 °C (secondary desorption). During the secondary desorption, the VOCs were effectively injected as a narrow band into a capillary column (Rtx-1, 0.32 mm × 105 m × 1.5 μm, Restek Co., USA). The temperature of the GC oven was initially set at 50 °C for 10 min, and it was increased to 250 °C at a rate of 5 °C/min. The transfer line and the valves of the thermal desorber were both maintained at 200 °C. The flow rate of the carrier gas (helium) in the column was maintained at 1.13 mL/min, and the outlet split flow of the thermal desorber was set at 14 mL/min.

The target VOCs differed slightly between the first and second years owing to a variation in the standard mixtures used for the calibration. For the first-year samples, a standard gas mixture of 41 VOCs for USEPA TO-14A (1 ppm, Restek, USA) was used, while a mixture of 61 VOCs for USEPA TO-15/17 (1 ppm, Supelco, USA) was employed in the second year. Acrylonitrile was not included in the 61-VOC mixture but was included in the 41-VOC mixture. Among the 62 VOCs, four highly volatile compounds (propylene, ethanol, Freon 12, and chloromethane) were not determined, because of the inefficient collection of these compounds by the adsorbent tubes used in this study (Seo et al. [50]).

The carbonyl compounds were collected on 2,4-dinitrophenylhydrazine (DNPH)-coated silica cartridges (LpDNPH S10L, Supelco Inc., USA) at a flow rate of 1 L/min for 3 h. To avoid sampling defects, ozone scrubbers (Supelco Inc., USA) were placed in front of the cartridges (USEPA [56]). The carbonyls were extracted with 3 mL of acetonitrile and then analyzed via high-performance liquid chromatography (HPLC) with ultra-violet detection at 360 nm (Shimadzu HPLC system, Japan). A total of 15 carbonyls were determined using a carbonyl-DNPH mixture (Supelco, USA). Details regarding the carbonyl sampling and analytical method used in this study can be found elsewhere (Seo and Baek [47]; Seo et al. [49]; USEPA [56]). The carbonyl samples were taken only for two seasons of the first year (i.e., summer and spring). Therefore, the number of data points for each carbonyl compound was 98.

The VOC concentrations determined in this study were evaluated for uncertainty with regard to the analytical precision, method detection limit (MDL), and duplicate precision for field samples. The analytical precision was evaluated as the relative standard deviation (RSD), and the RSDs of all the target VOCs were found to be ≤ 10% (n = 10) for the within-run precision (repeatability), while the between-run precision (reproducibility) ranged from 10 to 25% (n = 10), depending on the individual compounds. The MDL for each target VOC was evaluated according to the USEPA guidelines (USEPA [58]) and was estimated to range between 0.01 and 0.10 ppb for a typical air sampling volume of 15 L. Owing to the difference in molecular weight, the MDLs of the chlorinated hydrocarbons were higher than those of the aliphatic and aromatic compounds. The evaluation of duplicate precision for VOC samples collected in the field under similar conditions is an important task, because a replicate analysis for the adsorbent tubes is practically impossible (USEPA [55]). A study by Baek ([2]) revealed the mean duplicate precision (MDP) to be within 36% (n = 179) for all target VOCs, except dichloromethane (DCM, 66.8%), which is marginally higher than the USEPA recommended range of 30% (USEPA [55]). Moreover, the uncertainties of chlorinated compounds generally tend to be greater than those of non-chlorinated ones, mainly owing to their extremely low concentrations in ambient air.

At the time this study was undertaken, no standard reference materials were available for the atmospheric VOC samples. Hence, it was not possible to estimate the audit accuracy. Thus, an alternative approach was adopted, wherein the audit accuracy was evaluated through inter-laboratory comparisons, with a "third party" laboratory at the Korea Research Institute of Standards and Science analyzing 26 pairs of duplicate samples (Baek [2]). The results of the inter-laboratory study indicated that the MDPs for toluene and trichloroethylene (TCE) were 27.5% and 19.2%, respectively, and that for benzene was 32.1%. Most VOCs exhibited MDPs ≤ 30%, although the more volatile compounds, such as DCM, appeared to be less accurate (50.7%). The analytical precision of the carbonyls was also evaluated according to the RSDs of the peak areas for a standard mixture. The within-run precision appeared to be ≤ 2.8%, and the between-run precision was ≤ 5.0%. The MDLs for carbonyls were estimated to be between 1.1 and 2.3 ng/mL, which are equivalent to concentrations of 0.01 and 0.02 ppb in the air, according to a typical sampling volume of 180 L at 20 °C and 1 atm.

In this study, the potential health risk was evaluated only for long-term (chronic) inhalation exposure to VOCs, including cancer and non-cancer effects. For the risk assessment, we applied a point estimation approach that has been used in several studies (Bari and Kindzierski [5]; Fox et al. [15]; Jia and Foran [21]; Tam and Neumann [52]; Wu et al. [69], [68]). The cancer risk for a specific VOC was calculated by multiplying an ambient concentration by an inhalation unit risk (UR) value for the VOC (Eq. (1)), and then the cumulative cancer risk was estimated as the sum of the cancer risks for all available risk-posing VOCs (Eq. (2)). The UR is defined as the individual lifetime excess risk due to a chronic lifetime exposure via inhalation of one unit (1 μg/m3) of pollutant concentration (WHO [65]). In this study, the UR values were taken from the USEPA ([59], [60]) and the California EPA (Cal [8]).

1 $$\text{Cancer risk}=\mathrm{HAP}\text{concentration}\left(\mathrm{\mu g}/\mathrm{m}3\right)\times \mathrm{UR}{\left(\mathrm{\mu g}/\mathrm{m}3\right)}^{-1}$$

Graph

2 $$\text{Cumulative cancer risk}=\sum \text{individual cancer risks}$$

Graph

The chronic non-cancer risk for a specific VOC was evaluated as the hazard quotient (HQ), which was calculated using a measured concentration and a benchmark value for the VOC (Eq. (3)). In this study, the non-carcinogenic benchmark values were taken from reference concentrations (RfC) provided by the USEPA (USEPA [60], [61], [54]), reference exposure levels (REL) provided by the California EPA (Cal [8]), and minimal risk levels (MRL) provided by the US Agency for Toxic Substances and Disease Registry (ATSDR) (ATSDR [1]). The benchmark value is the concentration (in mg/m3) of a pollutant below which long-term exposure to the general population is not expected to result in adverse effects (USEPA [64]). The sum of the HQs for each HAP that affects the same target organ or organ system is defined as the hazard index (HI), as follows:

3 $$\text{Hazard quotient}=\mathrm{HAP}\text{concentration}\left(\mathrm{\mu g}/\mathrm{m}3\right)/\text{Benchmark value}$$

Graph

4 $$\text{Hazard index}=\sum \text{individual hazard quotients}$$

Graph

In this study, the toxicity data were first obtained from the USEPA Integrated Risk Information System (IRIS). When the information for a specific chemical was not available in the IRIS, the estimate provided by (in order of preference) the California EPA, the ATSDR, the Health Effects Assessment Summary Tables (HEAST) (USEPA [54]), or the Provisional Peer-reviewed Toxicity Values (PPRTVs) (USEPA [61]) was used. Chemicals without any UR, RfC, REL, and MRL estimates were not included in the risk assessment process. The toxicological data applied to the VOCs measured in this study are presented in Table S3 (SM).

As many VOCs are known to play an important role in the formation of ozone in the lower troposphere, OFP for each VOC was calculated to estimate the contribution of the VOCs to the photochemical ozone creation in the atmosphere. In this study, the concept of the photochemical ozone creation potential (POCP) suggested by Derwent et al. ([13]) was utilized to calculate the OFPi for a given VOCi, which is defined as follows:

5 $$\text{OFPi}=\text{POCPi}\times \text{concentration}\left(\mathrm{\mu g}/\mathrm{m}3\right)\text{of VOCi}$$

Graph

For developing an effective management scheme for a specific pollutant, information on three factors is crucial: the abundance, ubiquity, and emission sources of the target pollutant. Accordingly, we investigated the detection frequency (DF) of the target VOCs. We found that 11 VOCs were detected in ≥ 50% of the samples, and 9 other VOCs were found in 20–50% of the samples. The VOCs that were detected in all samples were benzene and toluene, and 14 halogenated compounds were not detected in any samples. Table 1 presents the concentrations of selected VOCs that exhibited a high DF and are of great environmental concern. For example, benzene, formaldehyde, and TCE are of utmost concern because of their carcinogenic nature and ubiquity (IARC [19]; USEPA [63]). Exposure to other VOCs, such as toluene, ethylbenzene, and xylenes, also causes adverse health effects on humans (WHO [65]). Although 1,3-butadiene and vinyl chloride are carcinogenic to humans, they were rarely detected in the samples (DFs of 7.7% and 2.5%, respectively) and appeared at very low concentrations (≤ 0.05 ppb).

Concentrations (in ppb) of selected VOCs in the Sihwa-Banwol area

*The numbers of data for these compounds were 638 (total), 238, 240, 80, and 80 for each site from the left to the right, respectively **The numbers of data for these compounds were 98 (total), 24, 25, 24, and 23 for each site from the left to the right, respectively

As shown in Table 1, the distribution of VOCs across the sampling sites exhibited wide variability, as indicated by their mean, standard deviation, and maximum concentrations. For example, the average concentration of toluene between the four sampling sites ranged from 6.3 ppb (at the Sihwa residential site) to 16.3 ppb (at the Banwol industrial site), with a maximum of 92.3 ppb (at the Sihwa industrial site). Similarly, the TCE ranged from 0.7 ppb (Sihwa residential) to 4.2 ppb (Banwol industrial), with a maximum of 32.3 ppb (Banwol industrial). This indicates that the occurrence of VOCs in the Sihwa-Banwol area might be complexly affected by various factors such as industrial emissions, and their spatial and temporal variations.

Regarding the average VOC concentrations for the four sites, the highest concentration was recorded for toluene (13.1 ppb), followed by ethyl acetate (9.8 ppb), 2-butanone (8.3 ppb), 2-propanol (4.1 ppb), and formaldehyde (3.9 ppb). The high concentration of toluene is attributed to its wide use as a common solvent in the manufacturing industry, mainly for making chemical compounds and products, automobiles, and rubber and plastic products (KNICS [32]). Although high concentrations were also obtained for ethyl acetate, 2-butanone, and 2-propanol, they are known to be less toxic. In contrast, formaldehyde is an identified carcinogen that is equally ubiquitous in indoor and outdoor air, being regarded as both as a primary and secondary pollutant. Another carcinogenic VOC is TCE, which is predominantly associated with industrial activities, being widely used as an organic solvent to remove grease from metal parts in the electronics and machinery industries (Baek et al. [4]; NTP [44]). The average concentration and DF of TCE appeared to be relatively high, i.e., 2.7 ppb and 80.1%, respectively. These results indicate that industrial activities may significantly influence the VOC concentrations in the ambient air of the Sihwa-Banwol area.

Benzene, which is also a well-known carcinogen, exhibited an average concentration of 0.7 ppb. At the time of this study, there was no ambient air-quality standard for benzene in Korea. However, in 2010, a standard of 5 μg/m3 (approximately 1.5 ppb) as an annual average was established (KMOE [29]). The major sources of atmospheric benzene in typical urban areas in Korea are mostly motor vehicle exhaust and fugitive emissions from gas service stations (Kim and Lee [27]; Na et al. [42]). As of 2007, 4.1 million cars were registered in Gyeonggi Province, and 35% of the total number of vehicles were diesel cars (KMOE [30]). Although the benzene concentration measured in this study did not exceed the Korean standard, the benzene presents a carcinogenic risk, even at low concentrations in the ambient air, as it is one of the non-threshold pollutants (IARC [18]). Therefore, routine monitoring and emission reductions of benzene are necessary in urban as well as industrial areas. The average values of other VOCs, such as toluene, TCE, xylenes, and styrene, also did not exceed the guideline values specified by the World Health Organization (WHO [65]).

The general statistics of air pollution emission sources in the Sihwa-Banwol area were available from the air emission source survey program of the Korea National Institute of Environmental Research (KMOE [30]). In Korea, businesses emitting air pollutants are classified into five groups, according to their annual emissions. Conventionally, Class I to III businesses are regarded as large emission sources and Classes IV and V businesses are treated as small and miscellaneous sources (KMOE [29]). As of 2006, the study area housed a total of 375 businesses of Classes I to III, accounting for 13.7% of the national total. Most of them (275 companies) were located in the Banwol industrial complex, and the rest were in the Sihwa complex. Majority of the companies were engaged in machinery (47%), electricity/electronics (16%), petroleum and chemical (9%), iron-steel (7%), apparel and textile (4%), and plating industry.

In addition to the sources, the emission data of toxic chemical substances in the Sihwa-Banwol area were investigated using a pollutant release and transfer register (PRTR) database. The Korean PRTR includes only those companies producing or using at least 1 ton/year of a specific hazardous substance (KNICS [32]). As most of the VOCs belong to the category of hazardous chemicals, a certain amount of their emission data could be obtained from the PRTR database. Table 2 presents the emission data and the number of sources for the top 20 VOCs (in descending order of their emission quantity) in the two industrial complexes, as of 2006. According to the table, the VOC emissions in Banwol were significantly larger than those in Sihwa. The total emission amount for the two complexes decreased in the following order: methanol, toluene, ethyl acetate, 2-propanol, 2-butanone, TCE, xylenes, and formaldehyde. Interestingly, excluding methanol, the emission rankings of the VOCs in Table 2 were consistent with the order of the measured levels of the VOCs described previously in "Occurrence of volatile organic compounds in Sihwa-Banwol." Therefore, the ambient VOC concentrations measured in the Sihwa-Banwol area were correlated with the VOC emissions from the industrial complexes.

Top 20 VOCs with regard to emission amounts from the Sihwa and Banwol industrial complexes

*These compounds are listed as hazardous air pollutants (HAPs) by the Ministry of Environment in Korea.

Although the PRTR data play an important role in identifying the sources and emissions of toxic chemicals from large industries, it is difficult to validate these data, as they are voluntarily reported by the concerned company or industry. However, this study clearly demonstrated that the ambient air monitoring data can provide valid and robust information that can be easily compared to the emission source data. Furthermore, the monitoring data can provide additional information that may be missing from the emission inventories or source investigations, such as the PRTR program. Therefore, the two approaches, i.e., the emission inventory and ambient monitoring, should be used symbiotically to develop adequate control strategies for HAPs, particularly VOCs.

According to the PRTR data, the VOC emissions were significantly larger in Banwol than in Sihwa (Table 2). Thus, we investigated whether there was a significant difference between the atmospheric VOC levels of the two complexes. The cumulative probability distributions for the two groups were compared, and the results are illustrated in Fig. S1 in the Supplementary Materials (SM). The mean and median concentrations of toluene were slightly higher in Banwol than in Sihwa, but the difference was not significant (p > 0.05). The TCE concentrations were significantly higher (by a factor of 1.5) in Banwol than in Sihwa (p < 0.05). As shown in Table 2, the emission amount of TCE was ~ 2 times higher in Banwol. Benzene was another VOC with larger emissions in the Banwol complex, although the mean concentrations were higher in Sihwa (0.83 ppb versus 0.69 ppb in Banwol), with a statistically significant difference (p < 0.05). The traffic volume around the Sihwa sampling site was significantly larger than that around the Banwol site, indicating that the benzene concentrations in the area were likely affected by both industrial and automobile emissions. However, in contrast to TCE and benzene, ethylbenzene, xylenes, and trimethylbenzenes (TMBs) exhibited no significant difference (p > 0.05) between the two sites. Formaldehyde also did not exhibit any difference between the two sites, but acetaldehyde, which is a notorious malodorous chemical, had a higher concentration in Banwol. Hence, acetaldehyde can be presumed to be one of the causative substances of the odor complaint in Ansan city, where the Banwol complex is located.

Overall, the differences in the measured concentrations of aromatic VOCs between the two industrial complexes were not as distinct as the differences in their emission data. Therefore, it can be inferred that significant amounts of VOCs are emitted from a large number of small industries, which were not included in the PRTR survey. For example, aromatic organic solvents such as toluene, ethylbenzene, and xylenes are widely utilized in diverse industries of various types and sizes. In contrast, TCE is strongly related to only a specific type of industry, such as electronics and/or machinery. Therefore, this study suggests that the PRTR survey in Korea be revised to include smaller companies as well, which produce or use specific hazardous chemicals of less than 1 ton/year. Although the VOC emissions from individual small-sized industries are low, the cumulative impact can be significantly high, as the overall number of such industries is large.

To evaluate the impacts of industrial activities on the ambient levels of VOCs in the residential areas, the entire data from the four sites were divided into two groups: industrial and residential. For maintaining consistency, a comparison between the two groups was performed for only the VOC data that were measured simultaneously at each site, as the sampling at the residential sites was conducted only during daytime (i.e., 09:00–13:00 and 13:00–17:00). The cumulative probability distributions of the selected VOCs for the two groups are presented in Fig. 2. In contrast to the fluctuating and small difference in the VOC concentrations between the two industrial sites (Fig. S1 (SM)), a distinct variation was observed between the industrial and residential groups. The concentrations of most VOCs at the industrial sites were significantly higher than those at the residential sites. In particular, the VOCs closely related to industrial activities (such as toluene, ethylbenzene, xylenes, and TCE) exhibited a striking difference. These results indicate that industrial activities are the main cause of VOC pollution in the ambient air of both the industrial and residential areas of Siheung and Ansan, even though the contribution of mobile emissions cannot be completely ruled out. The only exception was formaldehyde, which did not differ significantly between the two groups. The atmospheric behavior of formaldehyde is presumed to be different from that of other VOCs, as it is not only produced from industrial sources and motor vehicle exhaust but can also be formed in the atmosphere via photochemical oxidation of hydrocarbons under strong solar radiation at a high temperature (Ho et al. [17]; Liu et al. [36]; Lü et al. [37]).

Graph: Fig. 2 Comparison between the VOC concentrations in the residential and industrial areas of Sihwa-Banwol

To investigate the seasonality of the VOC concentrations in the Sihwa-Banwol area, data for different seasons were compared, as shown in Fig. 3. The concentrations were generally the highest in winter, followed by autumn, while spring and summer did not exhibit any consistent pattern. Assuming that the industrial activities did not vary significantly according to the season, the emission of volatile substances is usually expected to be larger in summer than in winter. However, the VOC concentrations measured in this study exhibited the opposite trend, except for formaldehyde. The higher ambient levels of benzene, toluene, ethylbenzene, and xylenes (BTEX) in winter can likely be attributed, to some extent, to the combustion of heating fuels during this season (Kim et al. [24]; Kim and Lee [27]). In the Sihwa-Banwol area, the heating fuels used in residential and commercial sectors are mostly town gas. In industrial sector, heavy oils (≤ 0.3% of sulfur content) and light oils (≤ 0.1% of sulfur content) are also used together with the town gas (Kim and Lee [27]). However, some VOCs that are not related to fuel combustion, such as TCE and DCM, also exhibited higher concentrations in winter. Therefore, the increased concentrations in winter can be attributed to the influence of meteorological conditions. Similar results have been reported for other areas of South Korea (Baek et al. [4]; Han et al. [16]; Na and Kim [41]). In temperate regions, the atmosphere during the winter season tends to be stable and often unfavorable for atmospheric dispersion. According to the meteorological data for this area (Table S2 (SM)), the average wind speed in winter and autumn is significantly lower than that in spring and summer. This is important because wind speed plays a vital role in the dispersion of air pollutants, as the persistence of the atmospheric concentrations emitted from ground-level sources is inversely related to the wind speed. Because the seasonal weather conditions in a particular area exhibit a cyclic or repetitive pattern every year, air pollution may always become worse in winter compared with the other seasons in the Sihwa-Banwol area, which is surrounded by a lake and the coast on two sides.

Graph: Fig. 3 Seasonal variation of the VOC concentrations in the Sihwa-Banwol area: SP spring, SM summer, AT autumn, WT winter. Number of data-points for VOCs: spring (n = 215), summer (n = 315), autumn (n = 160), winter (n = 160), total (n = 850). Number of data-points for carbonyls: spring (n = 52), summer (n = 44), total (n = 96)

In contrast to the other VOCs, the formaldehyde concentrations increased in summer (even compared with spring), possibly owing to its secondary formation as a byproduct of photochemical reactions during this season, as well as the increased volatile emissions associated with higher temperatures. Similarly, acetaldehyde, 2-butanone, and acetone are known to be formed in the atmosphere as secondary pollutants (Liu et al. [36]). However, in this study, these carbonyls did not exhibit increased levels in summer, indicating that their concentrations were predominantly affected by the local emissions from various industrial activities.

To understand the diurnal variation in the VOC concentrations, data from the two industrial sites (sites A and B), where 24-h monitoring was performed, were combined. The VOC measurements were sorted into groups corresponding to the time interval during which they were collected (01:00–05:00, 05:00–09:00, 09:00–13:00, etc.). Hence, regardless of the season, each group represented data collected during a specific time-interval within the 24 h. Figure 4 shows the diurnal changes in selected VOC concentrations in the Sihwa-Banwol area. Numerous studies (Baek et al. [3]; Choi et al. [11]; Seo et al. [48]) have indicated that VOC concentrations in large urban areas in Korea reach the highest levels between 8 and 10 a.m. in the morning, coinciding with rush hours involving heavy traffic, and the lowest levels occur between 2 to 4 p.m., when convection and photochemical reactions are active. However, the diurnal pattern in smaller industrial cities such as Siheung and Ansan is expected to differ from that in typical urban areas, as the impact of stationary emission sources on the VOC concentrations might be far larger than that of mobile sources.

Graph: Fig. 4 Diurnal variation of the VOC concentrations in the Sihwa-Banwol area (the error bar indicates the standard deviation)

As shown in Fig. 4, the lowest concentrations were commonly observed in the afternoon, between 1 and 5 p.m. However, the time intervals during which the concentrations increased varied among the VOCs. For instance, benzene and 1,2,4-TMB, which are closely related to vehicular exhaust (Na et al. [43]), exhibited elevated levels during the morning rush hours. In contrast, toluene, ethylbenzene, xylenes, styrene, and TCE, which largely originate from industrial emissions, tended to increase in the evening and at night rather than in the morning. Given that most industries in the Sihwa-Banwol complexes operate continuously for 24 h a day, the daily variations in their VOC emissions may not be large. However, the atmospheric mixing height is reduced at night owing to stable atmospheric conditions, which explains the elevated VOC levels at night. Similar phenomena have been observed in other small and medium-sized industrial cities in Korea (Baek et al. [4]; Kim et al. [25]; Lee et al. [35]; Seo et al. [49]). The highly volatile DCM exhibited slightly higher concentration levels during the day than at night, although the difference was insignificant.

To investigate the relationships among the different VOCs, PCA was performed for the datasets of each of the four sites. PCA is a common multivariate statistical technique used to investigate the relationships among multiple interrelated variables, wherein variables with similar behavior are grouped into the same principal component (PC). The PCA analysis in this study was conducted using the SPSS statistical package (SPSS Statistics 25, IBM, USA), and the results are presented in Table 3. For each dataset, three or four PCs with an eigenvalue greater than 1 were selected. The first PC of the Sihwa industrial site had high loadings for TCE and the most aromatic VOCs, such as BTEX, styrene, and TMBs, while the second PC was strongly related to chloroform and 1,2-dichloroethane. The third PC appeared to be associated with DCM and tetrachloroethylene. Similarly, the first PC of the Banwol industrial site was highly correlated with BTEX, styrene, and TCE. TMBs and benzene were correlated with the second PC. The third PC showed a close relation to DCM, and the fourth PC was closely associated with chloroform and 1,2-dichloroethane. Overall, the PCA results for the two residential sites exhibited similar patterns to those for the industrial sites.

Principal component loadings for the VOC data measured in the Sihwa-Banwol area

Note: PC loadings larger than 0.50 were italicized

According to the PCA analysis, the chlorinated VOCs in the Sihwa-Banwol area behaved independently from the aromatic VOCs, with the exception of TCE. This implies that the chlorinated VOCs in the ambient air in the area are emitted predominantly by industrial sources, while the aromatic compounds originate from both industrial and vehicular emissions. Another notable observation was that the PCA results for the residential sites were very similar to those of their neighboring industrial sites (Table 3). The PC loading pattern of the Sihwa residential site was similar to that of the Sihwa industrial site, and the pattern of the Banwol residential site matched that of the Banwol industrial site. These results indicate that the concentrations of VOCs in the residential areas in the cities of Siheung and Ansan were significantly influenced by the nearby industrial complexes.

The sources of atmospheric VOCs are diverse, and their occurrences in the ambient air vary according to the local and regional characteristics. Therefore, the VOC levels measured in this study were compared with those in other areas in Korea. Until recently, VOC studies in Korea were not as abundant or intensive as studies on other pollutants. Therefore, for obtaining reference data, papers in domestic journals were carefully examined with regard to the method of sampling and analysis, QC/QA, and amount of data. The mean levels of the major VOCs measured in the Sihwa-Banwol complexes in this study and those measured across different areas of Korea in other studies are presented in Table S4 (SM). The references from which the VOC data for other areas were obtained are listed in chronological order and were divided into two groups: industrial and non-industrial. The table presents only six VOCs, which were the most commonly measured VOCs in the cited studies.

The VOC levels were generally higher in industrial areas than in residential and/or commercial areas. However, the BTEX concentrations at the roadside in large urban areas were similar to those of industrial areas, which are attributed to the large influence of vehicular emissions. The benzene levels were significantly higher in Ulsan and Yeosu than in any other industrial areas, owing to the presence of large petrochemical industries. Similarly, the TCE levels were significantly elevated in Gumi (#1), Sihwa, and Banwol, reflecting the influence of the electronics and machinery industries. The Gumi industrial complex was developed mainly for the electronics industry and is known as the "mecca of the electronics industry" in Korea (Baek et al. [4]). Likewise, the majority of the companies in the Sihwa-Banwol area are engaged in machinery and electricity/electronics (KICC [23]). Toluene exhibited the highest concentration at all sites except for the petrochemical industrial areas such as Ulsan and Yeosu.

As shown in Table S4 (SM), the VOC concentrations exhibited no apparent trend of change over the years but varied significantly according to the location and the type of industry. The Sihwa-Banwol area was established as a national industrial complex: however, it was not designed for any specific type of industry. Thus, a wide variety of small and medium-sized businesses were established in the complexes, in a dense spatial pattern, and the atmospheric VOCs can hardly be characterized by a specific pattern of appearance. Nevertheless, it is clear that the levels of TCE and toluene in Sihwa-Banwol are significantly higher than those in other industrial areas. The distinct presence of these two chemicals in the ambient air is likely a good indicator of the impact of industrial activities, as they are among the most widely used chlorinated and aromatic solvents in Korea (KNICS [32]).

To assess the health impacts of the VOCs in the Sihwa-Banwol atmosphere, the cancer risk for each VOC was calculated using Eq. (1) for two groups (i.e., industrial and residential areas), as in described in "Spatial variation in volatile organic compound concentrations". For public health protection, the USEPA recommended an acceptable risk level of one per million (1 × 10−6) and a tolerable risk level of one hundred per million (1 × 10−4) for the excess cancer risk (USEPA [57]). The point estimates of excessive cancer risks from inhalation exposure to VOCs are presented in Table 4, together with the ambient mean concentrations (in μg/m3) and URs of each VOC. The VOC concentrations in ppb units were converted into μg/m3 for conditions of 20 °C and 1 atm. Among the ~ 60 target analytes measured in this study, the UR values were available for only 31 (see Table S3 (SM)). Among these 31 VOCs, descriptive statistics were given for the 22 species for which a DF of ≥ 20% of all samples was included in the risk assessment. The VOCs in Table 3 are listed in order of the magnitude of the mean cancer risk estimated for the residential areas.

Carcinogenic risk assessment for VOCs in industrial and residential areas of Sihwa-Banwol

[1] Integrated risk information system (USEPA [59]) [2] Health effect assessment summary tables (USEPA [54]) ND not detected

According to Table 4, the cancer risk levels for individual VOCs were generally higher in the industrial areas than in the residential areas. The risk levels ranged between 7.9 × 10−5 (for TCE) and 1.6 × 10−8 (for 1,1-dichloroethane) in the industrial areas, whereas they ranged from 6.1 × 10−5 (for formaldehyde) to 4.0 × 10−8 (for DCM) in the residential areas. The cumulative cancer risks associated with exposure to the 22 VOCs were 2.32 × 10−4 (industrial areas) and 1.76 × 10−4 (residential areas), both of which exceeded the tolerable risk level proposed by the USEPA ([57]). These levels are comparable to the cumulative risks that were estimated at air pollution hot-spots in southwest Memphis, Tennessee (2.3 × 10−4) (Jia and Foran [21]); Portland, Oregon (2.47 × 10−4) (Tam and Neumann [52]); and California (3.0 × 10−4) (Morello-Frosch et al. [40]) in the USA. Additionally, these levels are significantly higher than those estimated in Edmonton, Canada (≤ 1.0 × 10−5) (Bari and Kindzierski [5]). and Seattle, Washington (4.3–6.0 × 10−5) (Wu et al. [68], [69]). However, a direct comparison of the cumulative cancer risks between this study and other studies must be done with caution, because the results can vary greatly depending on the number of VOC species included in the risk assessment. Comparison of the cumulative cancer risks posed by a variety of VOCs in the Sihwa-Banwol area with those for other areas in Korea was not possible, because no comprehensive risk-assessment data for such a large number of VOCs as were measured in this study were available.

In industrial areas, the top five contributors appeared to be TCE (34.1%), formaldehyde (28.0%), chloroform (10.3%), ethylbenzene (9.5%), and benzene (8.2%), which contributed 90.1% to the cumulative cancer risk. In residential areas, the top five contributors were formaldehyde (34.7%), 1,1,2,2-tetrachloroethane (10.2%), TCE (9.7%), benzene (8.5%), and benzyl chloride (6.8%), which contributed 70.0% to the cumulative risk.

The non-carcinogenic health risk for a specific VOC was estimated using Eq. (3) and was expressed as an HQ. The HQ is the ratio of the potential exposure level to a toxicant and the benchmark level at which no adverse effects are expected. An HQ ≤ 1 indicates that adverse effects are not likely to occur; thus, the hazard from exposure is considered to be negligible. The HQ has been used by the USEPA to assess the non-carcinogenic health risks of air toxics (USEPA [64]).

Among the ~ 60 target analytes measured in this study, non-carcinogenic benchmark values (i.e., RfC, REL, or MRL) were available for 55 species (see Table S3 (SM)). Among these 55 species, 13 species were not included in the non-carcinogenic risk assessment, because of their low DFs (< 20%). The results of the HQ calculations for the remaining 42 VOCs are presented in Table 5. The sums of the HQs for the 42 VOCs were 22.0 and 9.46 for the industrial and non-industrial areas, respectively, and both exceeded the threshold criterion of 1 (USEPA [57]). However, it is unclear whether all the 42 VOCs have impacts on the same target organ or organ systems. Synergistic or antagonistic interactions among pollutants may enhance or mitigate risks in a way that could not be identified in this study. Nevertheless, such high HI values clearly indicate that a large proportion of the residents of Sihwa-Banwol have experienced concurrent VOC exposures that may pose non-cancer risks for various health effects.

Non-carcinogenic risk assessment for VOCs in industrial and non-industrial areas of Sihwa-Banwol

[1] Integrated risk information system (USEPA [59]) [2] Toxicological profiles (ATSDR [1]) [3] Provisional peer-reviewed toxicity values (PPRTVs) assessments (USEPA [60]) [4] Health effect assessment summary tables (USEPA [54]) [5] OEHHA air chemical database (Cal [8]) ND not detected aThe benchmark concentration for a specific toxicant was taken among reference concentrations (RfC) of IRIS (USEPA [59]) and PPRTVs (USEPA [61]), reference exposure levels (REL) of OEHHA (Cal [8]), minimal risk levels (MRL) from ATSDR (ATSDR [1]), and a standard value from the NAAQS (USEPA [59])

As an individual species, the highest non-cancer risk was posed by acrolein (10.0) in the industrial areas, followed by TCE (9.6), ethyl acetate (0.63), formaldehyde (0.51), and acetaldehyde (0.43). These five VOCs contributed 96.2% of the total non-cancer risk in the industrial areas. The top five contributors in the residential areas appeared to be acrolein (5.4), TCE (2.1), formaldehyde (0.48), acetaldehyde (0.39), and benzyl chloride (0.25), which contributed 91.1% of the total non-cancer risks.

Although the magnitude of the cancer and non-cancer risks for each species differed between the industrial and residential areas, the overall rankings were similar for the two areas. These results suggest that toxic VOCs emitted from the industrial complexes were dispersed or transported to the residential areas of Sihwa-Banwol.

Photochemical ozone formation of individual VOC species has often described through the POCP (Derwent et al. [12], [13]; Jenkin et al. [22]) or maximum incremental reactivity (MIR) (Carter et al. [9]). Although the two methods are slightly different, both were developed for the ranking of VOCs according to their ability to produce ozone. In this study, the OFPs for various VOCs were evaluated by multiplying the POCP value by the mean concentration (in μg/m3) of each VOC. The POCP is generally presented as a relative value: the amount of ozone produced by a certain VOC is divided by the amount of ozone produced by an equally large emission (on a mass basis) of ethylene. Ethylene has been chosen as a reference gas because it is one of the most potent ozone precursors among VOCs. By definition, the calculated POCP values are not absolute values, and these values were recently modified and updated by Jenkin et al. ([22]). The POCP values were available only for 32 VOCs among the VOCs measured in this study, and the OFPs calculated for the 32 species are presented in Table 6.

Ozone formation potentials (OFPs) of VOCs measured in the Sihwa-Banwol area

ND not detected aPOCPE: estimated photochemical ozone creation potential value from Jenkin et al. ([22]) bPOCP: photochemical ozone creation potential value from Derwent et al. ([13])

Although we estimated the OFPs for various VOCs in the Sihwa-Banwol area, the results have many limitations, because we did not measure some olefinic hydrocarbons (such as ethylene, propylene, and isoprene) that have high POCP values. This study was initially designed to focus on toxic VOCs instead of ozone precursors; hence, the adsorbent-tube method was adopted as the VOC measurement method. Therefore, it was technically impossible to measure very volatile C2−C3 alkenes in the atmosphere. Despite these limitations, Table 6 presents a number of important findings: (i) the OFPs for the industrial areas were 2.2 times greater than those for the residential areas; (ii) the aromatics exhibited the most significant contributions to the total OFPs in industrial and residential areas (63.0% and 55.1%, respectively), followed by the carbonyls (8.0% and 18.9%, respectively); and (iii) among the individual VOC species, toluene was the largest contributor to the OFP, accounting for ~ 30% of the total OFP, followed by xylenes, which accounted for ~ 20%. The larger contributions of toluene and xylenes compared with the other VOCs are consistent with the results of other studies performed in the Seoul metropolitan area. Lee et al. ([34]) and Park et al. ([46]) investigated POCPs suited for the metropolitan area by utilizing 56 ozone-precursor VOC data from photochemical assessment monitoring stations in Korea. They found that toluene and xylenes were the largest and second-largest contributors, respectively, to the ozone production. Therefore, the large amounts of aromatic compounds emitted from the Sihwa-Banwol industrial complexes are believed to play an important role in increasing the levels of tropospheric ozone in the Seoul metropolitan area during the summer season.

More than 60 VOCs were identified from a total of 850 samples obtained in the Sihwa-Banwol complexes, i.e., the largest industrial area in Korea, via a two-year air monitoring project. The target VOCs included aromatic, aliphatic, organochlorine, and carbonyl compounds. The obtained data were used to evaluate the impact of industrial activities on the occurrence of atmospheric VOCs, the cancer and non-cancer risks to residents living near the complexes, and the OFPs in this area.

The VOCs of particular concern appeared to be BTEX, TCE, and formaldehyde, according to their toxicity, concentration, and DF. Toluene was the most abundant VOC, followed by ethyl acetate, 2-butanone, 2-propanol, formaldehyde, xylenes, and TCE. The measured concentration rankings for each VOC closely matched the rankings of their emission data in the PRTR database. Although the PRTR data play an important role in identifying the major sources of VOCs, the system must be upgraded to include smaller companies whose usage and/or handling of a specific hazardous chemical is ≤ 1 ton/year.

The impact of industrial emissions on the ambient levels of VOCs in the Sihwa-Banwol area appeared to be significant, as the concentrations of most VOCs at the industrial sites were significantly higher than those at the residential sites. This was supported by the PCA results, as the patterns of PC loadings for the residential datasets were similar to those of the neighboring industrial complex. However, the industrial emissions were not the only sources of benzene and formaldehyde, considering that the difference between the two groups was not as distinct as that for other VOCs. Thus, vehicular emissions in the area are also likely to be an important source of some aromatic VOCs, such as benzene and TMBs. The concentrations of most VOCs tended to increase in winter and at night, which is attributed to meteorological influences. In contrast, the formaldehyde levels increased in summer, owing to the secondary formation in the atmosphere. Overall, the ambient levels of the VOCs in the Sihwa-Banwol area were relatively high compared with those in other regions of Korea, although they did not exceed the guideline values prescribed by the World Health Organization. In particular, the concentrations of TCE and toluene were significantly higher in the Sihwa-Banwol complexes than in other industrial areas in Korea.

The health risk assessment revealed that the cumulative cancer risks associated with exposure to the toxic VOCs were 2.32 × 10−4 (industrial areas) and 1.76 × 10−4 (residential areas), both of which exceed the tolerable risk level of 1 × 10−4 proposed by the USEPA. Additionally, the sum of the non-cancer risks caused by various VOCs appeared to be 22.0 and 9.46 for industrial and non-industrial areas, respectively, significantly exceeding the threshold criterion of 1. Large amounts of aromatic compounds emitted from the Sihwa-Banwol industrial complexes may play a crucial role in the increased levels of surface ozone in the Seoul metropolitan area during the summer season.

The Sihwa-Banwol complexes were originally developed to accommodate the pollution-inducing industries scattered across the Seoul metropolitan area, but ironically emit large amounts of toxic VOCs and ozone precursors to regions further inland in the downwind direction, where more than 20 million people live in the metropolitan area. According to the results of this study, to reduce the health risks for the residents of not only Sihwa-Banwol but also Seoul and its surrounding areas, the implementation of VOC control measures for industrial emissions in the Sihwa-Banwol complexes should be given a high priority within the framework of air-quality management strategies of Korea.

This study was conducted as part of a comprehensive project entitled "Monitoring of Hazardous Air Pollutants in Sihwa-Banwol Area, 2005 to 2007," in collaboration with the National Institute of Environmental Research (NIER) in Korea (contract number 2005-98). We acknowledge Mr. Min-do Lee of NIER for his administrative assistance in the early stage of the project.

Graph: (DOCX 767 kb)

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.