at SciVerse ScienceDirect

Atmospheric Environment 59 (2012) 117e124

Contents lists available

Atmospheric Environment

journal homepage: www.elsevier .com/locate/atmosenv

Dissolved organic carbon in the precipitation of Seoul, Korea: Implicationsfor global wet depositional flux of fossil-fuel derived organic carbon

Ge Yan, Guebuem Kim*

School of Earth & Environmental Sciences/RIO, Seoul National University, Seoul 151-747, South Korea

h i g h l i g h t s

< We investigated the source and flux of DOC in precipitation of Seoul.< Fossil fuel combustion was the dominant source.< DOC originated predominantly from local emissions.< The contribution by long-range transport from China was substantial.< Global flux of wet depositional fossil-fuel DOC was estimated to be 36 � 10 Tg C yr�1.

a r t i c l e i n f o

Article history:Received 29 February 2012Received in revised form22 May 2012Accepted 25 May 2012

Keywords:Rainwater DOCSeoulFossil-fuelWet depositional fluxLong-range transport

* Corresponding author. Tel.: þ82 2 880 7508; fax:E-mail address: gkim@snu.ac.kr (G. Kim).

1352-2310/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.atmosenv.2012.05.044

a b s t r a c t

Precipitation was sampled in Seoul over a one-year period from 2009 to 2010 to investigate the sourcesand fluxes of atmospheric dissolved organic carbon (DOC). The concentrations of DOC varied from 15 mMto 780 mM, with a volume-weighted average of 94 mM. On the basis of correlation analysis using thecommonly acknowledged tracers, such as vanadium, the combustion of fossil-fuels was recognized to bethe dominant source. With the aid of air mass backward trajectory analyses, we concluded that theprimary fraction of DOC in our precipitation samples originated locally in Korea, albeit the frequent long-range transport from eastern and northeastern China might contribute substantially. In light of therelatively invariant organic carbon to sulfur mass ratios in precipitation over Seoul and other urbanregions around the world, the global magnitude of wet depositional DOC originating from fossil-fuelswas calculated to be 36 � 10 Tg C yr�1. Our study further underscores the potentially significant envi-ronmental impacts that might be brought about by this anthropogenically derived component of organiccarbon in the atmosphere.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

The contemporary issue of global warming has been mainlyattributed to the elevated carbon dioxide level in the atmospherearising from various human activities since the industrial revolu-tion. A reliable assessment of human impact on future climatechange therefore rests with a full understanding of the globalbiogeochemical carbon cycle. However, the global carbon budgetpertaining to anthropogenic perturbation based solely on inorganiccarbon is unable to account for the fate of ca. 20% of global carbondioxide, which is referred to as “the missing carbon sink”. Previousstudies (Coelho et al., 2008; Jurado et al., 2008; Kieber et al., 2002;Orlovi�c-Leko et al., 2009; Pan et al., 2010; Willey et al., 2000)highlight the significance of incorporating atmospheric organic

þ82 2 876 6508.

All rights reserved.

carbon (OC) into the models of the global carbon cycle. OC in theatmosphere can be ultimately transformed into carbon dioxideeither by in-situ oxidation (Goldstein and Galbally, 2007; Willeyet al., 2000) or by biochemical degradation at the Earth’s surfaceafter being scavenged from the atmosphere. Therefore, it is of greatimportance to integrate an unequivocal representation of thispotentially significant carbon reservoir, especially the anthropo-genic fraction, into the global models. Since precipitation isconsidered to be the dominant scavenging mechanism in theatmosphere, wet depositional flux is generally employed to eluci-date the quantitative information for most airborne species. Anestimation of global rainwater flux of dissolved organic carbon(DOC) of 430 � 150 Tg C yr�1 has been reached by Willey et al.(2000) based on a comprehensive review of worldwide publishedrainwater DOC concentrations during the late 1990s. Nevertheless,large discrepancy (144 Tg C yr�1 for input to the oceans only)between this global approximation and that derived from another

G. Yan, G. Kim / Atmospheric Environment 59 (2012) 117e124118

approach (Jurado et al., 2008, wet depositional flux using models)implies that the rainwater measurement based calculation mightbe subject to considerable uncertainties, which are deemed to be atleast partially attributable to the paucity of quantitative knowledge(Jurado et al., 2008; Kieber et al., 2002), as well as the remarkabletemporal variations in organic carbon levels in rainwater over thelast decade (Willey et al., 2006, 2011; Xu et al., 2009). Consequently,it is imperative to conduct more extensive surveys on rainwaterDOC, especially in Asia, where little information is available.

It has been recognized that OC in the atmosphere is deriveddirectly or indirectly from both living and dead organisms. Thedirect pathway encompasses any kinds of biogenic emissions byplants and animals in terrestrial and marine systems; whereas theindirect mechanism mainly refers to incomplete combustionprocesses involved with relic (fossil-fuel) and modern plant as wellas animal tissues. A large fraction of atmospheric OC resides in theaerosol phase (the remaining are gaseous), which can be eitheremitted directly from any potential sources (primary organiccarbon, POC) or formed as secondary organic carbon (SOC) fromgaseous volatile organic compounds (VOC) via gas-to-particleconversions. The anthropogenic processes (mainly biomass andfossil-fuel burning) have been found to be the dominant sources ofPOC (Chung and Seinfeld, 2002), as well as significant contributorsto SOC, of which the anthropogenically derived fraction wasrecently suggested to be underestimated (Henze et al., 2008).Previous carbon isotopic analyses (13C and 14C) suggest thatapproximately 20e30% of DOC in rainwater is of fossil-fuel origin(Avery et al., 2006; Raymond, 2005).

Seoul is the sixth largest city around the globe in terms ofpopulation density and is characterized by intensive industrializa-tion and urbanization. Moreover, Seoul is subject to substantialinfluences from long-range transport of air pollutants (Kim et al.,2009), since Northeast Asia is one of the largest fossil-fuelcombustion regions, with China being recognized as a majorsource for anthropogenic carbonaceous aerosols (around a quarterof the global budget) (Cooke et al., 1999). Therefore, Seoul ispresumably an ideal site to investigate atmospheric OC whichstems from or is controlled by human activities. This study presentsthe abundance level of DOC in precipitations at a location in south-central Seoul over a one-year period, as a contribution to theintegrity of the global wet depositional DOC dataset. Furthermore,on the basis of the identification of DOC sources in our site, we haveattempted to undertake a quantitative evaluation of the globalmagnitude of fossil-fuel derived DOC in rainwater.

2. Experimental section

2.1. Study site and sample collection

Seoul, as the capital of Korea, is a metropolitan city with over 10million residents inhabiting an area of 605.36 km2. It is located inthe midwestern part of the Korean Peninsula, approximately 30 kmto the east of the Yellow Sea. The climate of Seoul is rather complexwith both continental and oceanic features. Our sampling durationis considered to be a wet year with a total precipitation depth of1816 mm, in comparison with a thirty-year average of 1451 mm(1981e2010, Korea Meteorological Administration). The prevailingwind systems for Seoul are westerlies (southwesterly in thesummer and northwesterly in the winter). The primary land-usetypes of Seoul in descending order of areal percentage are resi-dential (35%), forest (26%), and industrial (12%). Agriculturalactivities are generally found in the suburban area of western andeastern parts of Seoul. In addition, the western coastal city ofInchon and the surrounding Kyunggi province (along with Seoulare known as the Seoul Metropolitan Region) which are highly

industrialized and urbanized may also exert significant impacts onthe atmospheric environment of Seoul.

The sampling campaign was carried out on a four-storeybuilding rooftop at the Gwanak campus of Seoul National Univer-sity in a mixed commercial and residential area of Seoul (37.5� N,127� E) from October 2009 to September 2010. There were nopotential pollution sources (e.g. exhaust hoods) in sight around thesampling spot or any significant emission sources (e.g. smokestack)in the nearby regions. Precipitation samples (rain and snow) werecollected on an event basis using a borosilicate glass vessel (pre-combusted at 500 �C) and a polypropylene vessel (pre-washedwithdilute acids) for DOC and other chemical constituents, respectively.The apparatus was manually deployed on a supporting rack 1.2 mabove the rooftop at the onset of the precipitation events, andretrieved immediately after cessation. In the case of the eventswhich began or ended at night, samplers were placed in the lateevening or recovered in the early morning of the next day (usuallywithin 5 h). Therefore, the bulk samples we collected were alsoaffected by dry deposition. However, its contribution is expected tobe minimal, considering the rather limited exposure time to dryconditions and the fact that wet scavenging on the whole is thedominant removal process for most airborne species includingcarbonaceous aerosols (w80%) (Kanakidou et al., 2005). Aftercollection, the samples were transported to a laminar flow cleanroom inside the building for processing and preservation. The snowsamples were allowed to thaw at room temperature before furthertreatment. Subsamples for DOC and major ions were filtered usingan acid-washed plastic syringe joined to a syringe filter (WhatmanGF/F, pore size 0.7 mM). The filtration set was thoroughly rinsedwith ultra-pure water before use and pre-conditioned with analiquot of sample to minimize the risk of contamination. DOCsamples were placed in 20 mL pre-muffled glass ampoules andpreserved with 6M pure HCl, followed by fire sealing using a hand-held butane burner. Aliquots for the measurement of ions werecollected in polypropylene conical tubes and kept frozen at �20 �Cuntil analysis. Following the rigorous ultra-clean protocols, ina class-100 clean bench, the subsamples for trace elements (unfil-tered) were transferred to pre-cleaned HDPE bottles and thenacidified with 6M ultra-pure HNO3 (Yan et al., 2012).

2.2. Analysis of precipitation samples

DOC abundances in precipitation samples were determined byhigh temperature catalytic oxidation (HTCO) using a ShimadzuTOC-VCPH total organic carbon analyzer equipped with an ASI-Vauto-sampler. The standards were prepared from reagent gradepotassium hydrogen phthalate in ultra-pure water (resistivity:18 MU cm). The acidified samples were sparged with carbondioxide free carrier gas (UHP oxygen) at a flow rate of 150 mL/minfor 2 min to remove inorganic carbon. Then the samples wereinjected into a combustion column packed with Pt coated aluminabeads heated to 720 �C. The carbon dioxide evolving from com-busted organic carbon was detected by a non-dispersive infrareddetector (NDIR). The quality of the data was assured by insertingcertified referencematerial (CRM fromUniversity of Miami) in eachrun of samples (Kim and Kim, 2010). All of the measured results forCRM agreed to within 5% of its authentic value. The blank duringcollection and filtration for DOC was determined as the differencebefore and after the sampling and processing procedures usingultra-pure water. The blank levels are generally found to be underthe detection limit (i.e. < 5.0 mM), so blank correction was notperformed.

The trace elements were analyzed using an inductively coupledplasma mass spectrometer (ICP-MS, Model: X-II, Thermo Inc., UK)equipped with a MicroMist nebulizer (Glass expansion, USA). The

G. Yan, G. Kim / Atmospheric Environment 59 (2012) 117e124 119

concentrations for anions (SO42� and NO3

�) and cations (Naþ) weremeasured by suppressed ion chromatography (Model ICS 2000)and ICP-MS, respectively. The detection limits which were calcu-lated as three times of the standard deviation of blanks are0.10 mg L�1 for sulfate, 0.03 mg L�1 for nitrate, 0.01 mg L�1 forsodium, 0.24 mg L�1 for aluminum, 0.11 mg L�1 for vanadium,respectively.

The average concentrations of DOC were volume weighted toaccount for any influences exerted by precipitation amount (Topolet al., 1985). The sulfate and vanadium in our samples were cor-rected for contributions by seasalt and surface soil, using sodium(Keene et al., 1986) and aluminum (Taylor and McLennan, 1995) asreference elements, respectively.

2.3. Air mass backward trajectories

The air mass transport pathways for all the precipitation eventswere determined based on air mass back-trajectories producedusing version 4 of the Hybrid Single Particle Lagrangian IntegratedTrajectory (HYSPLIT) model developed at the Air Resources Labo-ratory of NOAA (Draxler and Hess, 1998). Trajectory maps weregenerated using the software TrajStat developed by Wang et al.(2009) and the GDAS database for a 72 h hind-cast starting at the500 m above ground level, corresponding to the well-mixed

Fig. 1. Temporal variations of DOC abundance levels in precipitation: (a) volume-weightedamount for each month during the sampling period from 2009 to 2010.

boundary layer in our location (Kim et al., 2009), which is likelyto contribute more heavily to in-cloud processes and wet deposi-tion (Walker et al., 2000).

3. Results and discussion

3.1. DOC levels in precipitation and the controlling factors

The DOC concentration levels in our precipitation samplesranged from 15 mM to 780 mM, with an annual volume-weightedaverage (VWA) of 94 mM. No significant differences in DOC concen-trations were found between rainwater (VWA¼ 94 mM, N¼ 49) andsnow (VWA ¼ 92 mM, N ¼ 8) samples. The monthly VWA DOCconcentrations showed relatively small variations throughout theyear, amongwhich slightly higher valueswere found in cold seasonsthan those inwarm seasons (Fig.1a). Themaximumwas observed inJanuary, which is associated with the lowest monthly precipitationof the year. The VWA for the growing season (AprileSeptember) andnon-growing season (OctobereMarch) were calculated to be 87 mMand 127 mM, respectively. Several extreme events with concentra-tions as lowas 15 mMhave been observed. Because our sampling siteis subject to frequentmarine influences, especially in summer, theseevents canprobably be linked tomarine airmasses that arenormallyassociated with low DOC content (Willey et al., 2000) and high

average concentrations and (b) integrated wet depositional fluxes and precipitation

Fig. 2. 72 h HYSPLIT air mass backward trajectories for precipitation in Seoul during2009e2010. The precipitation events are categorized to three regimes: local (a), Asiancontinental (b), and marine (c). The frequency (expressed as percentage of totalnumber of events) and average concentrations (VWA � SD) for each regime are alsoshown.

G. Yan, G. Kim / Atmospheric Environment 59 (2012) 117e124120

precipitation volume (i.e., dilution effect, see following section forthe detailed Discussion). With the aid of air mass back trajectoryanalyses, we excluded any extreme marine events, whose precipi-tation depths are greater than 90 mm, and obtained an average of133 mM. It is slightly lower than that reported for terrestrial rain(VWA: 161 mM) on a global scale (Willey et al., 2000), and muchhigher than that observed at coastal sites (Kieber et al., 2002).However, significantly higher values for DOC concentrations werefound in precipitation over Northern China (VWA: 250 mM) (Panet al., 2010), the closest continental region at the same latitude.This seems to be associated with its continental feature (i.e.,neglectable marine inputs) and low annual precipitation amount(average of 635 mm in contrast with 1451 mm in our site). Themonthly wet depositional fluxes of DOC were found to be high insummer and low inwinter,with remarkable variations ranging from0.24 to 4.04 mmol cm�2 month�1 (Fig. 1b). The observed temporaltrend can be largely attributed to the precipitation pattern ratherthan the source strength, as suggested by the significant positivecorrelation between monthly fluxes and precipitation amount(r2 ¼ 0.78, p < 0.01). The relatively low flux with respect to precip-itation amount in September 2010 appears to be resulting from theexcessive marine input in that month.

The DOC abundance in precipitation can be considered asa complex function of multiple controlling parameters. Besides thestrength of the major sources (see next section), the storm origineffect on DOC concentrations has been the most frequently dis-cussed in the literature (Kieber et al., 2002; Pan et al., 2010; Willeyet al., 2006). According to air parcel backward trajectory regimes(72 h), all the events were categorized into three groups, as shownin Fig. 2. The first group (Type I) consists of events with air massesoriginating in Korea or adjacent coastal areas and thereafterremaining around the Korean peninsula. Owing to the relatively“stagnant” nature of these air masses, it is postulated that DOC insamples associated with these events was predominantly emittedby local sources within Korea. For Type II events, the air massesoriginated mostly from eastern and northeastern China, and occa-sionally from Mongolia or Russia. They normally passed throughthe densely populated and industrialized areas of China, andsubsequently entered Korea via the Yellow Sea or the East Sea. Thistype accounts for the largest fraction (47%) of events being inves-tigated, implying the potentially significant contributions byemissions from China. The source areas for Type III events spreadedover the remote Pacific Ocean. These air masses spent a consider-able amount of time traveling over the oceanic areas prior tomaking landfall on Korea. The VWA DOC concentration for Type Ievents was calculated to be 162 mM, which is the highest among allthree groups. This high value can be largely linked to local conti-nental sources, in which anthropogenic processes are expected tobe dominant. Relatively lower values were found in other twogroups, which were 111 mM for Type II and 62 mM for Type III,respectively. The precipitation events in these two groupspresumably received considerable regional influences from long-range transport, which has been recognized to be a significantpathway of dispersion for air pollutants and other airborne species.The DOC in these samples is thus thought to be amixture of organicsubstance derived locally and regionally from both continental andmarine areas. Indeed, it has been suggested that Korea is subject tosubstantial influences by polluted air masses originating fromChina as well as clean marine air masses from the surroundingoceans (Kim et al., 2009). The lower concentrations for Type IIevents (in comparisonwith those of Type I) seem to be the result ofthe inclusion of marine air masses and/or the loss of the labileportion of organic matter during the transport. Likewise, thesignificantly higher values for the events associated withmarine airmasses (Type III) than that of “strictly” oceanic rains (23 mM)

G. Yan, G. Kim / Atmospheric Environment 59 (2012) 117e124 121

(Willey et al., 2000) can thus be explained by the considerablecontribution from local sources.

Furthermore, DOC in the atmosphere might vary with seasons,since biogenic emission from vegetation, as one of the major OCsources, is largely controlled by seasonality. As a result, relativelyhigher levels of DOC in rainwater have been found in warm (orgrowing) seasons in previous studies (Kieber et al., 2002; Willeyet al., 2000). However, in our location, the reverse pattern wasobserved, for which winter precipitations were generally associ-ated with higher DOC abundances (Fig. 1a). This phenomenon islikely due to the overwhelming anthropogenic contributions (inurban regions mostly fossil-fuel burning relevant processes lackingof temporal variations) which have rendered the seasonality of thenatural biogenic fraction undiscernable (i.e., variations in biogenicinput are not significant enough to impact the overall temporalpattern of DOC), as well as the low precipitation depth in the drywintertime of this region. Similar seasonal variations have beendocumented in other areas where the anthropogenic contributionto rainwater DOC prevails (Pan et al., 2010).

In addition, the DOC concentration levels have been found todecrease with the increasing precipitation amount (Pan et al.,2010). However, the correlations between these two parametersare usually weak and can be absent occasionally, which has beenattributed to continuous supplies of DOC during the process ofprecipitation (Kieber et al., 2002). The logarithmic regressionanalysis indicates DOC levels in our samples are largely controlledby dilution (Fig. 3), in such a manner that DOC is effectively“washed out” from the atmosphere by the initial precipitation (i.e.the DOC concentrations decrease with precipitation amount)(Soyol-Erdene et al., 2011). The precipitation amount, therefore, isprobably an important factor controlling DOC concentrations in ourlocation, if taking into account the fact that the effect of any indi-vidual factor can be easily masked by those of others.

The dry period preceding the precipitation events might beanother controlling factor, since the organic substances emittedlocally may build up in the surface air and compose a large fractionof rainwater DOC through below-cloud scavenging. However, it wasshown to be irrelevant to DOC levels in our samples. Other potentialcontrolling factors include storm type, wind speed, and the scav-enging efficiency of precipitation (i.e. removal rate of airborne gasesand particles by hydrometeors), which are beyond the scope of thisstudy and thus are not discussed here. All the above parameters are

Fig. 3. Concentrations of DOC versus precipitation depth for individual eventsobserved in Seoul during 2009e2010.

not independent but interact with each other. The DOC abundanceis therefore determined by the interplay among multiple control-ling factors. Further studies encompassing all the potentialparameters are thereby essential to elucidate the mechanismdetermining DOC concentration levels in wet deposition.

3.2. Sources of dissolved organic carbon in precipitation in Seoul

In general, biomass burning and fossil-fuel combustion werethought to be the most significant anthropogenic sources for rain-water OC, whereas biogenic emissions from both continental andmarine areas may also contribute substantially. In this work, thesource apportionment for DOC in precipitationwasmade by tracingthe inputs from the potential sources using signature chemicalspecies. The DOC concentrations were found to be correlatedstronglywith those of nitrate and non-seasalt sulfate (Fig. 4a and b),the two widely used pollution indicators in atmospheric studies(Matsumoto and Uematsu, 2004). The positive relationships withone another among these species indicate their common origin-sdhuman activities. The gaseous precursors for nitrate and sulfate(i.e., NO2 and SO2) in Seoul and its neighboring areas have beenreported to be emitted by industries, power plants, house heating,and vehicles (Chae et al., 2004). In particular, approximately 86% ofSO2 emissions were attributed to the first two sources in the prov-ince surrounding Seoul,which is thus expected tobe a crucial sourcearea. Since all the aforementioned processes utilize fossil-fuel as theenergy source, it can be inferred that DOC inprecipitation over Seoulis mainly derived from incomplete combustion of fossil-fuel.

We further employed vanadium, a reliable fingerprint foremissions from fossil-fuel combustion (Tsukuda et al., 2005), tojustify our inference. A significant positive relationship was foundbetween concentrations of DOC and non-crustal vanadium (Fig. 4c),indicative of fossil-fuel combustion as the dominant source forDOC. Fossil-fuels are primarily consumed by four major sectors:transportation, household heating, coal-/oil-firing electricitygeneration, and industrial production. The major fuels for groundtransportation in the urban areas are gasoline and diesel; whereashouse heating in winter season mainly consumes natural gas inKorea. Since the refining process removes most vanadium in fossil-fuels (Hope, 1997), it is unlikely that these two processes would besignificant sources of vanadium content in our samples. On theother hand, the fuels consumed by power generation and industrialproduction are primarily crude oil and coal. In these raw materialsand their fly ashes, vanadium was found to be generally enriched(Tsukuda et al., 2005). In fact, these two anthropogenic processesare very likely to have significant impacts in our location. Seoul isa highly industrialized city with around 18,100 industrial factoriesoccupying ca. 12% of the total area (Lee et al., 2005). The thermalplants (where coal and oil are fired) generate the largest portion(66.5%) of electricity annually in Korea. The majority (94%) of globalfossil-fuel emissions takes place in the Northern Hemisphere (NH)(Cooke et al., 1999). Besides, Chung and Seinfeld (2002) suggestedthat primary organic aerosol was dominated by fossil-fuel emis-sions in Eastern Asia. Accordingly, DOC in rainwater derived fromfossil-fuel is likely to surpass that from biomass burning and othersources in someNH urban regions. It is therefore concluded that theDOC observed in the precipitation over Seoul is producedpredominantly by the combustion of crude fossil-fuels, especiallyduring industrial production and power generation.

3.3. Magnitude of wet depositional DOC flux in Seoul and itsimplications

The annual wet depositional flux of DOC in Seoul was calculatedto be 1.9 g C m�2 yr�1 (the sum of depositional amount per event),

Fig. 4. Plots showing correlations between concentrations of DOC and various sourcetracers: (a) non-sea-salt sulfate, (b) nitrate, and (c) non-crustal vanadium in individualprecipitation samples collected in Seoul during 2009e2010.

G. Yan, G. Kim / Atmospheric Environment 59 (2012) 117e124122

which falls within the range of those reported for continental andcoastal rain DOC fluxes around the world (Willey et al., 2000).Moreover, it is identical to the 10-site-average value over Northern

China (Pan et al., 2010). This value amounts to 1.2 Gg yr�1 over theentire Seoul area if assuming uniform deposition rate (calculated asdepositional flux of 1.9 g C m�2 yr�1 times by the area of605.36 km2). Alternatively, the annual flux of DOC in precipitationof Seoul can be calculated using the equation given below:

FC ¼ FS � Roxidation � fw � RC=S (1)

where FC and FS denote thewetdepositionalfluxofDOCandemissioninfluxof SO2 to the atmosphere, respectively; Roxidation represents theconversion rate of sulfur dioxide to sulfate (64 � 18%)(Intergovernmental Panel on Climate Change (IPCC), 2001), whereasfw indicates the fraction of sulfate subject towet scavenging (83� 6%)(IPCC, 2001); Rc/s is the mass ratio between DOC and sulfur in ourprecipitation samples calculated from the linear regression slope inFig. 4a (C/S ¼ 1.3). Using the published official SO2 emission data(2.8 Gg S yr�1) (Environmental Statistics Yearbook 2010, http://eng.me.go.kr) would yield an annual wet depositional DOC flux of1.9� 0.6 Gg yr�1. It is in general agreement with the value calculatedabove using the area, taking into account any potential variations ofdeposition rate of DOC within Seoul from our sampling location.Similar carbonesulfur ratios forwet depositionwere found in severalother locations under significant anthropogenic influences, includingSouth China (C/S ¼ 1.7, 22.6� N, 113.9� E, 2005e2009) (Huang et al.,2010), The Netherlands (C/S ¼ 1.4, 51.5� N, 4.1� E, 1980e1986)(Nguyen et al., 1990), Puerto Rico (C/S ¼ 1.3, 51.5� N, 18.3� E, anthro-pogenic plume, 2004e2007) (Gioda et al., 2011). The organic matterto sulfate mass ratios in tropospheric aerosols at low altitudes(0.5e1.5 km column) in urban regions were found to be around 1 byaircraft measurements (Heald et al., 2005). If an organic-mass-to-organic-carbon ratio of 1.4e2.2 is taken (Aiken et al., 2008), the C/Sratio for these aerosols is thus calculated to be 1.4e2.1, which isconsistent with those found in precipitation samples. Furthermore,fossil-fuel combustion has been recognized to be the predominantsource for global SO2 emission (Lee et al., 2011). Therefore, byassuming that the DOC to sulfur ratio does not vary dramatically inprecipitation over the regions where fossil-fuel combustion hassignificant impacts, one can extrapolate the above approximation tothe global level. Given the global annual land surface anthropogenicSO2 emission of 52 Tg S (Lee et al., 2011), approximately 36 � 10 TgDOC scavenged by wet deposition is generated from man-madeprocesses (mainly fossil-fuel combustion) each year (Eq. (1)).

Using the DOC flux in global continental rain of 340� 120 Tg peryear (Willey et al., 2000) and the fossil-fuel DOC content in rain-water (Avery et al., 2006; Raymond, 2005) yields an annual netremoval of 85 � 35 Tg fossil-fuel derived DOC by precipitation,which is approximately twice of our estimate. However, the esti-mation of global flux by Willey et al. (2000) was made on the basisof measurements conducted over a decade ago. Rainwater DOCmight have undergone considerable variations since that time, asbeing observed in the continental USA (Willey et al., 2006) andSouthwestern China (Xu et al., 2009). In fact, the global atmo-spheric depositional flux of fossil OC has not yet been explicitlyreported. OC scavenged by precipitation is thought to consist ofboth aerosols (POC and SOC) and gases (VOC). However, due to thelimitation of methodology (HTCO), we actually analyzed non-purgeable DOC in our precipitation samples (Avery et al., 2009). Itis assumed that VOC is completely removed during the initial stepof this widely employed method and thus not quantified in rain-water DOC measurements. We tabulated the flux estimates ofproduction of organic aerosols in the most recent works as Table 1.The best estimates for POC and SOC subject to atmospheric scav-enging were taken to be 10� 6 Tg and 7� 3 Tg, respectively. Takinginto account in-situ oxidation, wet scavenging efficiency, andsolubility, global wet depositional fossil-fuel DOC is calculated to be

Table 1The magnitude of global wet depositional flux of fossil-fuel derived DOC and theproduction of fossil-fuel derived primary and secondary organic carbon in theaerosol phase.

Method Flux (Tg C yr�1) Reference

DOC wet depositionCarbon isotope based 50e120 Avery et al. (2006),

Raymond (2005) andWilley et al. (2000)

Sulfate budget based 26e46 This studyEmission inventory based 6e18a This study

POC production3e18 Hallquist et al. (2009)5e15 Bond et al. (2004)SOC production1e9 Henze et al. (2008)3e17 Hallquist et al. (2009)6 Spracklen et al. (2011)8 de Gouw and Jimenez (2009)

a The global flux of DOC in precipitation is calculated using the following equa-tion: FC¼ (PPOCþ PSOC)� fw� fdis, where FC represents wet depositional flux of DOC;PPOC and PSOC denote the productions of POC and SOC, respectively; fw indicates thefraction of organic aerosol removed bywet deposition (80� 5% of total wet plus dry)(Kanakidou et al., 2005); fdis. signifies the percentage of rainwater organic carbonaccounted for by dissolved phase (85 � 5%, Pan et al., 2010; Willey et al., 2000).Oxidation of organic aerosol (POC and SOC) to CO/CO2 is assumed to be neglectable(Hallquist et al., 2009).

G. Yan, G. Kim / Atmospheric Environment 59 (2012) 117e124 123

12 � 6 Tg (see the footnote of Table 1 for calculation). It isapproximately one third of our estimate, and one seventh of thatdetermined from global DOC rainwater flux and carbon isotopicratios. These remarkable discrepancies signify the imperativenessof conducting further studies to reduce the uncertainties associatedwith the wet depositional flux of fossil-fuel derived DOC.

The fossil-fuel derived DOC flux based on our calculation, on anannual basis, is equivalent inmagnitude to 8%of the global rainwaterDOC flux reached by Willey et al. (2000) or 15% of the wet deposi-tional OC to the global oceans estimated by Jurado et al. (2008). Ifcompared with inorganic carbon, this figure is less than 1% of globalannual CO2 influx to the atmosphere from fossil-fuel burning, or lessthan 5% of the missing carbon sink in current global carbon cyclingmodels (Schimel et al., 1996). This component of atmospheric OC,although not quantitatively significant in the global carbon budget,may take part in the regional carbon cycle and cause a variety ofhealth and environmental problems. For example, organic aerosolsare able to scatter and reflect solar radiation (direct effect) andinfluence cloud albedo (indirect effect), thereby altering the radia-tive balance of the atmosphere. Fossil-fuel derived organic aerosols(17 � 9 Tg) (Table 1) correspond to a global atmospheric burden of0.34�0.18Tg (ChungandSeinfeld, 2002). It could exert a top-of-the-atmosphere direct radiative forcing of�0.02 to�0.05Wm�2, usingthe value of �36 to �76 W g OC�1 (Chung and Seinfeld, 2002); oralternatively, �0.05 to �0.07 W m�2 (depending on the fraction ofSOC), assuming that the direct radiative effect of organic aerosol isproportional to its burden (Maria et al., 2004). In addition, fossil-fuelcombustion represents a significant source for the toxic organicconstituents present ubiquitously in the atmosphere, such as poly-cyclic aromatic hydrocarbons (PAHs) (Qiao et al., 2006), which areconsidered to be highlymutagenic and carcinogenic. As such, fossil-fuel combustion derived atmospheric OC in urban areas may haveprofound implications on the environment and human health. Ourmagnitude estimation could contribute to the quantitative assess-ment on its impacts in various relevant fields.

4. Conclusions

The annual volume-weighted average concentration of DOC inprecipitation over Seoul was determined to be 94 mM, which falls in

between averages for continental and coastal areas on a globalscale. Its concentration levels were influenced by storm origin andprecipitation amount, yet independent of seasonality or the lengthof dry period prior to precipitation. The DOC was shown to bederived predominantly from local emissions by the combustion offossil-fuels, although the contribution by long-range transport fromChina can be substantial. The global magnitude of fossil-fuelderived organic carbon dissolved in precipitation was furtherquantified to be 36 � 10 Tg C yr�1. This empirical estimate couldcontribute to the constraint of magnitude of atmospheric organiccarbon as well as the quantitative evaluation of the impacts offossil-fuel combustion on the Earth’s environment and humanhealth.

Role of the funding source

This work was funded by the Korea Meteorological Adminis-tration Research and Development Program under Grant CATER2012-7170. G.Y. was partially supported by the BK21 scholarshipthrough School of Earth and Environmental Sciences, SeoulNational University, Korea. The sponsor had no involvement instudy design, data analysis or preparation of the manuscript.

Acknowledgments

We gratefully acknowledge all the EMBL members who sup-ported us in laboratory analyses and the NOAA ARL for the provi-sion of HYSPLIT transport model.

References

Aiken, A.C., DeCarlo, P.F., Kroll, J.H., Worsnop, D.R., Huffman, J.A., Docherty, K.S.,Ulbrich, I.M., Mohr, C., Kimmel, J.R., Sueper, D., Sun, Y.L., Zhang, Q., Trimborn, A.,Northway, M., Ziemann, P.J., Canagaratna, M.R., Onasch, T.B., Alfarra, M.R.,Prevot, A.S.H., Dommen, J., Duplissy, J., Metzger, A., Baltensperger, U.,Jimenez, J.L., 2008. O/C and OM/OC ratios of primary, secondary, and ambientorganic aerosols with high-resolution time-of-flight aerosol mass spectrometry.Environmental Science and Technology 42, 4478e4485.

Avery Jr., G.B., Willey, J.D., Kieber, R.J., 2006. Carbon isotopic characterization ofdissolved organic carbon in rainwater: terrestrial and marine influences.Atmospheric Environment 40, 7539e7545.

Avery Jr., G.B., Dickson Brown, J.L., Willey, J.D., Kieber, R.J., 2009. Assessment ofrainwater volatile organic carbon in southeastern North Carolina, USA. Atmo-spheric Environment 43, 2678e2681.

Bond, T.C., Streets, D.G., Yarber, K.F., Nelson, S.M., Woo, J.-H., Klimont, Z., 2004.A technology-based global inventory of black and organic carbon emissionsfrom combustion. Journal of Geophysical Research 109, D14203. http://dx.doi.org/10.1029/2003JD003697.

Chae, G.-T., Yun, S.-T., Kim, K.-H., Lee, P.-K., Choi, B.-Y., 2004. Atmospheric versuslithogenic contribution to the composition of first- and second-order streamwaters in Seoul and its vicinity. Environment International 30, 73e85.

Chung, S.H., Seinfeld, J., 2002. Global distribution and climate forcing of carbona-ceous aerosols. Journal of Geophysical Research 107 (D19), 4407. http://dx.doi.org/10.1029/2001JD001397.

Coelho, C.H., Francisco, J.G., Nogueira, R.F.P., Campos, M.L.A.M., 2008. Dissolvedorganic carbon in rainwater from areas heavily impacted by sugar cane burning.Atmospheric Environment 30, 7115e7121.

Cooke, W.F., Liousse, C., Cachier, H., Feichter, J., 1999. Construction of a 1� �1� fossilfuel emission data set for carbonaceous aerosol and implementation andradiative impact in the ECHAM4 model. Journal of Geophysical Research 104(D18), 22137e22162.

de Gouw, J., Jimenez, J.L., 2009. Organic aerosols in the Earth’s atmosphere. Envi-ronmental Science and Technology 43, 7614e7618.

Draxler, R.R., Hess, G.D., 1998. An overview of the HYSPLIT_4 modeling system oftrajectories, dispersion, and deposition. Australian Meteorological Magazine 47,295e308.

Gioda, A., Reyes-Rodríguez, G.J., Santos-Figueroa, G., Collett Jr., J.L., Decesari, S.,Ramos, M.C.K.V., Netto, H.J.C.B., Neto, F.R.A., Mayol-Bracero, O.L., 2011. Specia-tion of water-soluble inorganic, organic, and total nitrogen in a backgroundmarine environment: cloud water, rainwater, and aerosol particles. Journal ofGeophysical Research 116, D05203. http://dx.doi.org/10.1029/2010JD015010.

Goldstein, A.H., Galbally, I.E., 2007. Known and unexplored organic constituents inthe Earth’s atmosphere. Environmental Science and Technology 41, 1514e1521.

Hallquist, M., Wenger, J.C., Baltensperger, U., Rudich, Y., Simpson, D., Claeys, M.,Dommen, J., Donahue, N.M., George, C., Goldstein, A.H., Hamilton, J.F.,

G. Yan, G. Kim / Atmospheric Environment 59 (2012) 117e124124

Herrmann, H., Hoffmann, T., Iinuma, Y., Jang, M., Jenkin, M.E., Jimenez, J.L.,Kiendler-Scharr, A., Maenhaut, W., McFiggans, G., Mentel, Th.F., Monod, A.,Prévót, A.S.H., Seinfeld, J.H., Surratt, J.D., Szmigielski, R., Wildt, J., 2009. Theformation, properties and impact of secondary organic aerosol: current andemerging issues. Atmospheric Chemistry and Physics 9, 5155e5236.

Heald, C.L., Jacob, D.J., Park, R.J., Russell, L.M., Huebert, B.J., Seinfeld, J.H., Liao, H.,Weber, R.J., 2005. A large organic aerosol source in the free troposphere missingfrom current models. Geophysical Research Letters 32, L18809. http://dx.doi.org/10.1029/2005GL023831.

Henze, D.K., Seinfeld, J.H., Ng, N.L., Kroll, J.H., Fu, T.-M., Jacob, D.J., Heald, C.L., 2008.Global modeling of secondary organic aerosol formation from aromatichydrocarbons: high- vs. low- yield pathways. Atmospheric Chemistry andPhysics 8, 2405e2420. doi:10.5194/acp-8-2405-2008.

Hope, B.K., 1997. An assessment of the global impact of anthropogenic vanadium.Biogeochemistry 37, 1e13.

Huang, X.F., Li, X., He, L.Y., Feng, N., Hu, M., Niu, Y.W., Zeng, L.W., 2010. 5-year studyof rainwater chemistry in a coastal mega-city in South China. AtmosphericResearch 97, 185e193.

Intergovernmental Panel on Climate Change (IPCC), 2001. In: Houghton, J.T., et al.(Eds.), Climate Change 2001: The Scientific Basis. Cambridge Univ. Press, NewYork, p. 880.

Jurado, E., Dachs, J., Duarte, C.M., Simó, R., 2008. Atmospheric deposition of organicand black carbon to the global oceans. Atmospheric Environment 42, 7931e7939.

Kanakidou, M., Seinfeld, J.H., Pandis, S.N., Barnes, I., Dentener, F.J., Facchini, M.C., vanDingenen,R., Ervens, B.,Nenes,A.,Nielsen, C.J., Swietlicki, E.,Putaud, J.P., Balkanski, Y.,Fuzzi, S., Horth, J., Moortgat, G.K., Winterhalter, R., Myhre, C.E.L., Tsigaridis, K.,Vignati, E., Stephanou, E.G., Wilson, J., 2005. Organic aerosol and global climatemodelling: a review. Atmospheric Chemistry and Physics 5, 1053e1123.

Keene, W.C., Pszenny, A.A., Galloway, J.N., Hartley, M.E., 1986. Sea-salt correctionsand interpretation of constituent ratios in marine precipitation. Journal ofGeophysical Research 91, 6647e6658.

Kieber, R.J., Peake, B., Willey, J.D., Avery Jr., G.B., 2002. Dissolved organic carbon andorganic acids in coastal New Zealand rainwater. Atmospheric Environment 36,3557e3563.

Kim, T.H., Kim, G., 2010. Distribution of dissolve organic carbon (DOC) in thesouthwestern East Sea in summer. Ocean and Polar Research 32 (3), 291e297.

Kim, S.H., Han, Y.J., Holsen, T.M., Yi, S.M., 2009. Characteristics of atmosphericspeciated mercury concentrations (TGM, Hg(II) and Hg(p)) in Seoul, Korea.Atmospheric Environment 43, 3267e3274.

Lee, P.-K., Yu, Y.-H., Yun, S.-T., Mayer, B., 2005. Metal contamination and solid phasepartitioning of metals in urban roadside sediments. Chemosphere 60, 672e689.

Lee, C., Martin, R.V., van Donkelaar, A., Lee, H., Dickerson, R.R., Hains, J.C.,Krotkov, N., Richter, A., Vinnikov, K., Schwab, J.J., 2011. SO2 emissions and life-times: estimates from inverse modeling using in situ and global, space-based(SCIAMACHY and OMI) observations. Journal of Geophysical Research 116,D06304. http://dx.doi.org/10.1029/2010JD014758.

Maria, S.F., Russell, L.M., Gilles, M.K., Myneni, S.C.B., 2004. Organic aerosol growthmechanisms and their climate-forcing implications. Science 306 (5703),1921e1924.

Matsumoto, K., Uematsu, M., 2004. Free amino acids in marine aerosols over thewestern North Pacific Ocean. Atmospheric Environment 39, 2163e2170.

Nguyen, V.D., Merks, A.G.A., Valenta, P., 1990. Atmospheric deposition of acid, heavymetals, dissolved organic carbon and nutrients in the Dutch delta area in1980e1986. Science of the Total Environment 99, 77e91.

Orlovi�c-Leko, P., Plav�si�c, M., Bura-Naki�c, E., Kozarac, Z., �Cosovi�c, B., 2009. Organicmatter in the bulk precipitations in Zagreb and �Sibenik, Croatia. AtmosphericEnvironment 4, 805e811.

Pan, Y.P., Wang, Y.S., Xin, J.Y., Tang, G.Q., Song, T., Wang, Y.H., Li, X.R., Wu, F.K., 2010.Study on dissolved organic carbon in precipitation in Northern China. Atmo-spheric Environment 44, 2350e2357.

Qiao, M., Wang, C.X., Huang, S.B., Wang, D.H., Wang, Z.J., 2006. Composition,sources, and potential toxicological significance of PAHs in the surface sedi-ments of the Meiliang Bay, Taihu Lake, China. Environment International 32,28e33.

Raymond, P.A., 2005. The composition and transport of organic carbon in rainfall:Insights from the natural (13C and 14C) isotopes of carbon. Geophysical ResearchLetters 32, L14402. http://dx.doi.org/10.1029/2005GL022879.

Schimel, D.S., Alves, D., Enting, I., Heimann, M., Joos, F., Raynaud, D., Wigley, T., 1996.CO2 and the carbon cycle. In: Houghton, J.T., et al. (Eds.), Climate Change 1995:The Science of Climate Change. Cambridge Univ. Press, New York, pp. 76e86.

Soyol-Erdene, T.-O., Han, Y.C., Lee, B., Huh, Y.S., 2011. Sources and fluxes of Pt, Ir, andREE in the Seoul metropolitan area through wet scavenging processes. Atmo-spheric Environment 45, 1970e1978.

Spracklen, D.V., Jimenez, J.L., Carslaw, K.S., Worsnop, D.R., Evans, M.J., Mann, G.W.,Zhang, Q., Canagaratna, M.R., Allan, J., Coe, H., McFiggans, G., Rap, A., Forster, P.,2011. Aerosol mass spectrometer constraint on the global secondary organicaerosol budget. Atmospheric Chemistry and Physics Discussion 11, 5699e5755.

Taylor, S.R., McLennan, S.M., 1995. The geochemical evolution of the continentalcrust. Reviews of Geophysics 33, 241e265.

Topol, L., Levon, M., Flanagan, J., Schwall, R., Jackson, A., 1985. Quality AssuranceManagement for Precipitation Systems. EPA/600/4-82-042a. EnvironmentalProtection Agency, Research Triangle Park, North Carolina.

Tsukuda, S., Sugiyama, M., Harita, Y., Nishimura, K., 2005. Atmospheric bulkdeposition of soluble phosphorous in Ashiu Experimental Forest, Central Japan:source apportionment and sample contamination problem. AtmosphericEnvironment 39, 823e836.

Walker, J.T., Aneja, V.P., Dickey, D.A., 2000. Atmospheric transport and wet depo-sition of ammonium in North Carolina. Atmospheric Environment 20,3407e3418.

Wang, Y.Q., Zhang, X.Y., Draxler, R., 2009. TrajStat: GIS-based software that usesvarious trajectory statistical analysis methods to identify potential sources fromlong-term air pollution measurement data. Environmental Modeling andSoftware 24, 938e939.

Willey, J.D., Kieber, R.J., Eyman, M.S., Avery Jr., G.B., 2000. Rainwater dissolvedorganic carbon: concentrations and global flux. Global Biogeochemical Cycles 1,139e148.

Willey, J.D., Kieber, R.J., Avery Jr., G.B., 2006. Changing chemical composition ofprecipitation in Wilmington, North Carolina, U.S.A.: implications for thecontinental U.S.A. Environmental Science and Technology 40, 5675e5680.

Willey, J.D., Glinski, D.A., Southwell, M., Long, M.S., Avery Jr., G.B., Kieber, R.J., 2011.Decadal variations of rainwater formic and acetic acid concentrations in Wil-mington, NC, USA. Atmospheric Environment 45, 1010e1014.

Xu, G., Lee, X.Q., Lv, Y.C., 2009. Urban and rural observations of carboxylic acids inrainwater in Southwest of China: the impact of urbanization. Journal ofAtmospheric Chemistry 62, 249e260.

Yan, G., Cho, H.-M., Lee, I., Kim, G., 2012. Significant emissions of 210Po by coalburning into the urban atmosphere of Seoul, Korea. Atmospheric Environment54, 80e85.