Hazard/Risk Assessment

COMPARISON OF HUMAN EXPOSURE PATHWAYS IN AN URBAN BROWNFIELD:REDUCED RISK FROM PAVING ROADS

KYLE JAMES,yz RICHARD E. FARRELL,z and STEVEN D. SICILIANO*z§yInterdisciplinary Graduate Program of Toxicology, University of Saskatchewan, Saskatoon, Saskatchewan, Canada

zDepartment of Soil Science, University of Saskatchewan, Saskatoon, Saskatchewan, Canada

§Toxicology Group, University of Saskatchewan, Saskatoon, Saskatchewan, Canada

(Submitted 19 January 2012; Returned for Revision 23 February 2012; Accepted 9 June 2012)

Abstract—Risk assessments often do not quantify the risk associated with soil inhalation. This pathway generally makes a negligiblecontribution to the cumulative risk, because soil ingestion is typically the dominant exposure pathway. Conditions in northern or ruralcenters in Canada characterized by large areas of exposed soil, including unpaved roads, favor the resuspension of soil particles, makingsoil inhalation a relevant risk pathway. The authors determined and compared human exposure to metals and polycyclic aromatichydrocarbons (PAHs) from soil ingestion and inhalation and analyzed the carcinogenic and noncarcinogenic risks before and afterroads were paved in a northern community. To determine the inhalation exposure, three size fractions of airborne particulate matterwere collected (total suspended particulates [TSP], particulate matter with an aerodynamic diameter less than 10mm [PM10], andparticulate matter with an aerodynamic diameter less than 2.5mm [PM2.5]) before and after roads were paved. Road paving reduced theconcentration of many airborne contaminants by 25 to 75%, thus reducing risk. For example, before paving, the carcinogenic riskassociated with inhalation of Cr was 3.4 excess cancers per 100,000 people exposed, whereas after paving, this risk was reduced to 1.6 in100,000. Paving roads reduced the concentrations of total suspended particulates (TSP; p< 0.1) and PM10 (p< 0.05) but not PM25.Consequently, the ingestion of inhaled soil particles was substantially reduced. The authors conclude that resuspended soil is likelyan important source of risk for many northern communities and that paving roads is an effective method of reducing risk from theinhalation of soil particles. Environ. Toxicol. Chem. # 2012 SETAC

Keywords—Risk assessment Soil Particulate matter Metals Polycyclic aromatic hydrocarbons

INTRODUCTON

Declining soil quality is a growing concern worldwidebecause anthropogenic activities result in increasing concen-trations of contaminants in soil, most notably trace metals [1],heavy metals [1–3], and polycyclic aromatic hydrocarbons(PAHs) [4,5]. In recent years, it has been shown that contami-nated soil has the potential to affect humans adversely across theglobe. Studies from the United States [6], Australia [7], andTunisia [8] have demonstrated that lead in soil poses a risk tohuman health, particularly children. Studies from Canada [9]and China [10] showed that arsenic in soil carries elevatedcarcinogenic risk. In Taiwan, heavymetal contamination of paddysoils threatens rice production and food safety, prompting thedevelopment of risk-based site assessments aimed at protectingagricultural production and minimizing human exposure [11].

Humans are exposed to contaminants in soil through variouspathways, and ingestion of soil particles is commonly consid-ered the major route of exposure [12]. Inhaling soil particles inthe form of airborne particulate matter is often not quantified assoil ingestion, which generally results in exposures one to twoorders of magnitude greater than inhalation [13,14]. As a result,some risk assessments focus entirely on ingestion exposure andignore the inhalation pathway [15,16]. However, in studies thatconsider inhalation pathways, work shows that inhalationin certain situations can be an important exposure pathway.For example, Lai et al. [11] found that inhaling soil particles

contributed 17% to the total heavy metal exposure at a site insouthern Taiwan. Also, review and opinion articles have sug-gested that resuspended soil is likely a human health concern[17,18].

In the city of Iqaluit (northern Canada), the Lower BaseRegion (LBR) is a former military base that has been convertedto residential use and has a history of soil contamination [19].This particular site features three factors that contribute toincreased inhalation exposure: unpaved roads, a lack of vege-tative cover, and contaminant enrichment in the <45-mmfraction of soil. As of 2007, none of the roads in Iqaluit werepaved, which resulted in soils being resuspended into the airvia vehicular traffic [20]. Geological material, such as fugitivedust from unpaved roads, paved roads, and agricultural oper-ations in southern California, was found to account for 60%of particulate matter less than 10mm (PM10) [21]. In northernEuropean countries, suspension of road dust is the primarycontributor to PM10, and approximately 90% comes fromnonexhaust sources [22,23]. In addition, during residentialdevelopment, a large majority of the vegetative cover withinthe city was removed and has not recovered. A vegetative covercan reduce wind velocity and increase the surface moisturecontent, effectively binding soil particles into larger aggregatesand reducing the resuspension of soil [24]; this lack of avegetative cover has also contributed to a greater load ofsuspended particulate matter in the local atmosphere. Finally,metal concentrations in the <45-mm fraction in Iqaluit soilare higher than those in the bulk soil [19], and smaller sizefractions of soil are more likely to be resuspended as opposedto the bulk soil [25,26]. Taken together, these conditionssuggest that soil inhalation exposure may be an importantpathway.

Environmental Toxicology and Chemistry# 2012 SETAC

Printed in the USADOI: 10.1002/etc.1952

* To whom correspondence may be addressed(steven.siciliano@usask.ca).

Published online 21 July 2012 in Wiley Online Library(wileyonlinelibrary.com).

1

The Iqaluit LBR has been the site of numerous studies[19,27,28] that have characterized exposure, contaminant con-centrations, and contaminant bioaccessibility. The area is thusan ideal site for testing the importance of the inhalationexposure pathway in a northern urban setting and evaluatingwhether simple municipal activities can reduce exposure path-ways and thereby mitigate exposure. It should be noted thatmany northern municipalities have competing public healthconcerns and risks, so solutions that increase the functionality oraesthetic value of a city and also reduce exposure pathways areparticularly prized. During the summer of 2008 and 2009, theCity of Iqaluit paved the roads within the city. The purpose ofthe present study was to evaluate the relative magnitude of theinhalation versus ingestion pathways and whether paving roadssignificantly decreased the risk of adverse effects from chem-icals of potential concern in the soils of this municipality.

For the present study, three size fractions of airborne partic-ulate matter were collected to characterize human exposure toairborne contaminants: total suspended particulates (TSP),particulate matter with an aerodynamic diameter less than10mm (PM10), and particulate matter with an aerodynamicdiameter less than 2.5mm (PM2.5). For the present risk assess-ment, particulate matter less than PM10 is considered repre-sentative of inhalation exposure, because it reaches the lowerrespiratory tract, and particulate matter greater than PM2.5 isconsidered representative of ingestion exposure, as it is even-tually removed from the respiratory tract by mucociliary actionand ingested [12]. Soil ingestion exposure was determinedusing the reported values from previous studies [19,27].

MATERIALS AND METHODS

Particulate matter

Three Ecotech Mirco-Vol 1100 (American Ecotech) low-flow (3 L min�1) air samplers were set up outside a residentialhome within the Iqaluit LBR to collect TSP, PM10, and PM2.5across 47-mm Whatman glass fiber filters. The samplers wereset to collect at a height of approximately 0.8m to reflect theexposure for a toddler, because they have the largest intake rateto body mass ratio [29]. The filters were replaced once per weekfor a total of seven weeks. Laboratory and field blanks werecollected with each set of filters as a blank correction for bothairborne particulate concentration and chemical analysis. Filterswere cut into sections for analysis using an acetone-washedscalpel. Filters were preconditioned for 24 h under 50% relativehumidity both before and after sampling to determine mass ofparticulate matter. Samples were reweighed 1 h after the initialweighing, and the mean was taken to represent the mass.

Metal analysis

Approximately 30-mg filter samples were microwavedigested with an acid mix of HNO3/H2O2/HF following theprocedures of Laird et al. [28] and analyzed for total metalcontent using inductively coupled plasma–mass spectrometry.Metals analyzed were Be, Al, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn,As, Se, Sr, Mo, Ag, Cd, Sn, Sb, Ba, Hg, Tl, Pb, and U. Both Seand Tl had values below their respective detection limits of 3and 1mg kg�1 and thus were excluded from the risk assessment.The National Institute of Standards and Technology certifiedreference material 1640a (trace elements in water) was includedin each set of analyses for quality assurance purposes; averagerecoveries ranged from a low of 93% for As to a high of 105%for Mo.

Polycyclic aromatic hydrocarbons analysis

Accelerated solvent extraction was used to extract PAHsfrom approximately 30-mg filter samples following the proce-dures of James et al. [30] and analyzed using high-pressureliquid chromatography coupled with fluorescence detection[31]. The PAHs included in the analyses were naphthalene(Nap), fluorene, pyrene, benzo[a]anthrancene, chrysene,benzo[b]fluoranthene (BbF), benzo[k]fluoranthene (BkF), benzo-[a]pyrene (BaP), dibenzo[a,h]anthrancene, benzo[ghi]pyrene,and indeno[1,2,3–cd]pyrene (IcdP). Recovery of individualPAHs from blank filters spiked with a known amount of PAHsranged from 67% for BbF to 117% for chrysene and averaged83%. Fluorene and pyrene had values below the detection limitof 0.64 pg and were excluded from the risk assessment.

Enrichment factors

Enrichment factors (EF) were calculated as previouslydescribed [32]

EF ¼ ½X�atm½Al�atm

� ½X�soil½Al�soil

(1)

where [X]atm and [Al]atm are the concentrations of thecontaminant and Al in the atmosphere, and [X]soil and [Al]soilare themean concentrations of the contaminant andAl in the soil.Aluminum is used as a reference element, assuming that itsanthropogenic contribution is minimal [33]. An EF approaching1 indicates that the soil is the predominant source for thecontaminant. An EF>5 suggests that a significant fraction of thecontaminant can be attributed to a nonsoil source. An EF >100suggests that the element comes from an anthropogenic source.Given that the contaminant concentration in the <45-mm sizefraction of the soil wasmore likely to be resuspended [25,26] andhad higher contaminant concentrations [19,27], this fractionwasused to calculate enrichment factors. Using bulk soil would haveresulted in larger enrichment factors, when in fact the soil isnaturally enriched.

Exposure assessment

The hypothetical receptor of concern in this scenario was anindividual who lives (and works) within the LBR. The primaryfocus was on toddlers, because they have the highest intake rateto body mass ratio [29]; therefore, if the risk assessment protectstoddlers, it will protect the other receptors.

Three pathways were examined for exposure comparison:incidental soil ingestion, particulate matter inhalation, andparticulate matter ingestion following tracheal deposition. Soilingestion represents the fraction of soil that adheres to handsand is subsequently ingested. Dermal exposure was excludedbecause the additional pathway did not influence the primaryresults of the study. Particulate matter inhalation representsinhaled airborne contaminants that reach the respiratory tract,and particulate matter ingestion represents larger airbornecontaminants that are inhaled but are swept into the esophagusby mucociliary action and ingested. Exposure via soil ingestion,particulate matter inhalation, and particulate matter ingestion iscalculated as

EDI ðmgkg�1 d�1Þingestion-soil

¼ ½X�soil � IRsoil � AFgit � YE

BW� LE(2)

2 Environ. Toxicol. Chem. 31, 2012 K. James et al.

where EDI¼ estimated daily intake (mg kg�1 d�1), [X]soil¼concentration of contaminant in soil (mg kg�1); [Y]air¼ concen-concentration of contaminant in particulate matter less than10mm in diameter (mgm�3); [Z]air¼ concentration of contami-nant in particulate matter greater than 2.5mm in diameter(mgm�3); IRsoil¼ soil ingestion rate (kg d�1); IRair¼ airinhalation rate (m3 h�1); AFgit¼ relative bioavailability (unit-less); AFinh¼ inhalation absorption factor (unitless); toutdoor¼time spent outdoors (h d�1); tindoor¼ time spent indoors (h d�1);BW¼ body weight (kg); YE¼ years exposed (y); and LE¼ lifeexpectancy (y). Both YE and LE variables apply only forcarcinogenic risk. Soil PAH concentration and relativebioaccessibility were obtained from Siciliano et al. [24], andsoil metal concentration and relative bioaccessibility wereobtained from Siciliano et al. [19] and Laird et al. [28],respectively. Contaminant concentrations used are displayed inTable 1. The most conservative values were always used;therefore, in situations inwhichPM2.5concentrationwasgreaterthan PM10 concentration, the PM2.5 value was used. Theinhalation absorption factorwas assumed to be 1. Environmentalcontaminant concentrations and particulate matter concentra-tions were determined experimentally. From the CanadianCompendium, it was assumed that a toddler would weigh16.5 kg, have an inhalation rate of 0.388m3 h�1, and spend1.5 h d�1 outdoors and 22.5 h d�1 indoors [29]. Similarly, it wasassumed that an adult would weigh 70.7 kg, have an inhalationrate of 0.658m3 h�1, and spend 1.5 h d�1 outdoors and 22.5 h d�1

indoors. It is assumed that toddlers and adults have a soilingestion rate of 8.0� 10�5 kg d�1 and 2.0� 10�5 kg d�1,respectively [34]. For this risk assessment, it is assumed thatthe indoor concentrationof airborne toxicants, inmgm�3, is 70%of the outdoor concentration [35,36].

Risk assessment

Carcinogenic risk was determined using Equation 5, andnoncarcinogenic risk was calculated using Equation 6, utilizinghazard quotients (HQ)

Risk ¼ EDI� CSF (5)

HQ ¼ EDI� TDI (6)

where Risk¼ a unitless probability of an individual developingcancer over a lifetime, EDI¼ estimated daily intake(mg kg�1 d�1); CSF¼ cancer slope factor (mg kg�1 d�1)�1;and TDI¼ tolerable daily intake (mg kg�1 d�1). Values forCSF and TDI were obtained from Health Canada and the U.S.Environmental Protection Agency Integrated Risk InformationSystem ([34]; http://www.epa.gov/iris/). Age-dependent adjust-ment factors (ADAFs) were used to account for the susceptibleearly life stages. An adjustment factor of 10 was used for the firsttwo years of life, an adjustment factor of three was used for agesbetween two and 16 years, and no adjustment was made for ages

Table 1. Average concentration of metals and polycyclic aromatic hydrocarbons in soil and airborne particulate matter

ContaminantSoil concentration

(mg/kg)

TSP concentration (mg/m3) PM10 concentration (mg/m3) PM2.5 concentration (mg/m3)

Before paving After paving Before paving After paving Before paving After paving

Ag 0.53 1.1� 10�7 5.5� 10�8 1.7� 10�8 1.0� 10�8 2.5� 10�9 1.2� 10�9

As 2.4 5.1� 10�7 2.2� 10�7 1.0� 10�7 2.5� 10�8 2.3� 10�9 4.3� 10�9

Ba 1142 1.6� 10�3 6.9� 10�5 4.0� 10�4 1.8� 10�5 1.6� 10�5 1.4� 10�5

Be 1.5 4.7� 10�7 3.5� 10�7 5.5� 10�7 1.1� 10�8 1.5� 10�8 7.4� 10�8

Cd 4.6 5.0� 10�7 3.1� 10�7 9.8� 10�8 4.8� 10�8 1.3� 10�8 9.6� 10�9

Cr 67 2.7� 106 1.3� 10�6 3.6� 10�7 1.7� 10�7 3.5� 10�8 7.8� 10�9

Cu 15 1.4� 10�7 8.0� 10�8 2.4� 10�8 1.0� 10�8 2.4� 10�9 3.0� 10�9

Hg 0.97 9.5� 10�10 8.3� 10�10 3.8� 10�10 ND ND 1.4� 10�10

Mn 655 2.2� 10�6 5.7� 10�7 2.6� 10�7 6.5� 10�8 6.9� 10�9 1.3� 10�7

Mo 1.9 1.3� 10�7 6.1� 10�8 2.6� 10�8 5.0� 10�9 2.0� 10�9 1.8� 10�8

Ni 16 4.2� 10�8 2.3� 10�8 1.0� 10�8 4.7� 10�9 1.3� 10�9 4.4� 10�9

Pb 29 3.9� 10�7 2.4� 10�7 7.3� 10�8 7.9� 10�9 3.3� 10�9 2.1� 10�8

Sb 0.21 1.5� 10�7 8.5� 10�8 2.5� 10�8 5.8� 10�9 3.2� 10�10 1.0� 10�8

Sr 216 2.7� 10�5 1.6� 10�5 5.0� 10�6 6.9� 10�7 1.1� 10�7 1.4� 10�7

U 1.0 9.4� 10�8 4.8� 10�8 1.3� 10�8 4.0� 10�9 6.3� 10�10 4.1� 10�10

V 73 7.5� 10�8 2.9� 10�8 5.4� 10�8 2.7� 10�8 ND 3.1� 10�8

Zn 81 2.8� 10�4 2.0� 10�4 2.8� 10�4 4.6� 10�5 2.3� 10�5 6.8� 10�6

Benzo[a]anthrancene 7.0 2.8� 10�7 1.5� 10�7 5.9� 10�7 1.0� 10�7 1.5� 10�7 3.3� 10�7

Benzo[a]pyrene 6.7 1.4� 10�7 7.2� 10�8 5.0� 10�7 1.0� 10�7 1.1� 10�7 2.0� 10�7

Benzo[b]fluoranthrene 11 2.6� 10�7 2.0� 10�7 3.4� 10�7 6.2� 10�8 3.9� 10�7 2.9� 10�7

Benzo[ghi]pyrene 4.3 7.0� 10�8 ND 4.0� 10�7 1.0� 10�7 1.6� 10�7 3.7� 10�7

Benzo[k]fluoranthrene 5.3 2.8� 10�7 1.4� 10�7 2.4� 10�7 2.3� 10�8 5.7� 10�8 1.9� 10�7

Chyrsene 14 5.4� 10�7 6.7� 10�7 1.2� 10�6 8.9� 10�8 7.0� 10�7 8.6� 10�7

Dibenzo[a,h]anthrancene 2.0 ND 2.1� 10�8 ND 3.3� 10�8 ND 3.3� 10�8

Indeno[1,2,3-cd]pyrene 8.5 ND ND 3.1� 10�7 ND ND 2.7� 10�7

Naphthalene 0.31 1.2� 10�8 6.1� 10�9 6.6� 10�8 8.5� 10�9 8.6� 10�7 5.0� 10�9

ND¼ not detected; TSP¼ total suspended particulates; PM10¼ particulate matter less than 10mm; PM2.5¼ particulate matter less than 2.5mm.

EDI ðmgkg�1 d�1Þinhalation-PM ¼ ð½Y �air � IRair � AFinh � toutdoor þ ½Y �air � 70%� IRair � AFinh � tindoorÞYEBW� LE

(3)

EDI ðmgkg�1 d�1Þingestion-PM ¼ ð½Z�air � IRair � AFgit � toutdoor þ ½Z�air � 70%� IRair � AFgit � tindoorÞYEBW� LE

(4)

Paving roads reduces risk from soil inhalation Environ. Toxicol. Chem. 31, 2012 3

16 years and older [37]. Potency equivalency factors were usedto calculate carcinogenic risk from PAHs and were obtainedfrom the Canadian Council of Ministers of the Environment[38]. Briefly, carcinogenic risk is expressed as the sum total BaPequivalents for PAHs with the same mode of toxic action usingtheir relative potency to BaP.

Statistical analysis

The Anderson–Darling test was applied to all data to test fornormality. Significant difference was determined by Student’st test.

RESULTS

Whereas soils in the Iqaluit LBR were the primary sourceof V, Cr, Mn, Ni, Cu, Hg, and Pb in the airborne particulatematter (Fig. 1), PAHs and the metals Be, Zn, Sb, and Ba werederived primarily from anthropogenic (nonsoil) sources.Nonsoil sources also contributed to the As, Sr, Mo, Ag, Cd,and U found in the airborne particulates (Fig. 1). The EF forPAHs in the TSP was only approximately 1; however, because77% of PAHs present in the TSP originated in the PM2.5fraction, which has an EF for PAHs of approximately 100, itcan be concluded that PAHs in airborne particulate matter weremost likely of anthropogenic origin.

When comparing the EDI from each exposure pathway(Fig. 2), intake from soil ingestion was generally one to twoorders of magnitude greater than the combined intake ofairborne particulates via both inhalation and ingestion. Indeed,Ag and Zn were the only elements for which intake through soilingestion was less than the combined intake of airborne partic-ulate matter. Exposure to Ag is not a concern, because total Agexposure is 10,000 times lower than the TDI, and the enrich-ment factor for Zn (�500; Fig. 1) indicates that soil was not theprimary source of this metal. Likewise, inhalation accounts fora large proportion of the EDI of Ba (33%) and Be (48%), andagain the EFs for these metals (�37 and 36, respectively)indicate that they derive primarily from anthropogenic sources.With the exception of these metals, the ingestion of airborne

particulates contributes �1% to the total EDI for most metalsand the PAHs, with particulate inhalation contributing 1 to 10%and soil ingestion contributing 90 to 99% of the total EDI.For PAHs, in particular, the ingestion of airborne particulatesrepresents only a minor exposure pathway (contributing�0.2%to the total EDI). Indeed, the majority of the airborne PAHswere found in the PM2.5 fraction; therefore, in terms ofthe airborne particulates, inhalation is the primary exposurepathway.

The amounts of TSP (p< 0.1) and PM10 (p< 0.05) materi-als decreased significantly following road paving (Fig. 3). Therewas a high degree of temporal variability in the amounts ofparticulate matter collected, which could be attributed partiallyto the changes in precipitation, vehicular traffic, and windvelocity [24]. Furthermore, the City of Iqaluit sprays the roadswith CaCl2 as a dust suppressant when deemed necessary,which in turn complicates comparing airborne particulatematter concentrations with temporal precipitation and windvelocity. Before the roads were paved, the average concen-trations (mean� standard deviation) of TSP, PM10, and PM2.5were 85� 56mgm�3, 35� 17mgm�3, and 3.8� 2.2mgm�3,respectively. After paving, the average concentrations of TSP,PM10, and PM2.5 were 15� 14.4mgm�3, 6.5� 6.0mgm�3,and 2.3� 1.9mgm�3, respectively. In general, paving the roadsreduced airborne concentrations of the various contaminants by25 to 75% and reduced the amount of particulate matter ingestedand inhaled to 16 and 18% of the prepaving values. Whereas Baexhibited the greatest reduction (96%) in airborne concentra-tion, Hg exhibited the smallest reduction (13%; Fig. 4).

As a result of road paving, carcinogenic risk from airbornecontaminants was reduced (Fig. 5). Moreover, this reductionwas generally much greater for toddlers than for adults. Notsurprisingly, paving roads was more effective at reducing riskfrom carcinogens that have only inhalation CSFs (i.e., Cr andCd) as opposed to carcinogens that have both inhalation andingestion CSFs (i.e., As and BaP equivalents). Conversely,Be has only an inhalation CSF, and paving roads was noteffective at reducing carcinogenic risk; however, the soil isnot the primary source (Fig. 1). After paving roads, incremental

Fig. 1. Enrichment factors (EFs) for three size fractions of airborne particulatematter. Each vertical bar is themeanof three to fivemeasurements, and the error baris the standard error of the estimate. An EF of approximately 1 (—) indicates that the soil is the dominant source of the contaminant, an EF greater than 5 (- - -)suggests that a significant fraction can be attributed to a nonsoil source, and an EF greater than 100 (– – –) suggests that the element comes from an anthropogenicsource.Metals are displayed on the left of the graph and polycyclic aromatic hydrocarbons (PAHs) on the right. Polycyclic aromatic hydrocarbons are abbreviatedas follows: Nap¼ naphthalene; BbF¼ benzo[b]fluoranthrene; BkF¼ benzo[k]fluoranthrene; BaP¼ benzo[a]pyrene; IcdP¼ indeno[1,2,3–cd]pyrene; BaPEq¼ benzo[a]pyrene equivalents.

4 Environ. Toxicol. Chem. 31, 2012 K. James et al.

lifetime carcinogenic risk to a toddler from Cr is reduced by46%, from Cd by 48%, and from Be by only 13%, whereas thecarcinogenic risk to a toddler from As and BaP equivalents isreduced by only 12 and 5%, respectively. Hazard quotientspresented negligible risk from all exposure pathways and aredisplayed in Table 2. Notably, the hazard quotients from theinhalation pathway for Be, Cr, and Mn are larger than theiringestion counterparts.

DISCUSSION

The results from this risk assessment indicate that pavingroads is an effective method of reducing risk from the inhalationof contaminated soil. Paving roads does not eliminate theresuspension of soil particles from roads; instead, the rate ofsoil resuspension is limited by the rate at which soil particulatesare deposited onto the roads [39]. In Iqaluit, mean PM2.5concentrations (3.1mgm�3) are lower than those reported fora number of other Canadian cities (6.7–11.8mgm�3 [40], 7.2–20.9mgm�3 [41]) but are comparable to the concentrationsreported for a rural location in New Brunswick, Canada(4.5mgm�3 [40]). Additionally, individual airborne PAHconcentrations in Iqaluit (ranging from 1 to 5� 10�4mgm�3)are similar to those found in rural locations across Canadaand lower than the concentrations reported for urban counter-parts [42,43]. Prepaving PM10 (35� 17mgm�3) and TSP(85� 56mgm�3) concentrations in Iqaluit are consistent withthose found across Canada [41] (i.e., 27� 16mg PM10m�3 and55� 38mg TSP m�3). However, this same study reports thatPM2.5 contributes an average of 51% of the PM10 mass and30% of the TSP mass. In Iqaluit, however, PM2.5 contributesonly 11% of the PM10 mass and only 4% of the TSP mass,indicating that the airborne particulate matter consists mainlyof larger fractions. Paving roads reduced the TSP and PM10fractions to 15� 14mgm�3 and 6.5� 6.0mgm�3, respectively.As a result, concentrations of airborne particulate matter inIqaluit are among the lowest reported values for Canadian cities[41]. As expected, paving roads did not significantly reduce theconcentration of PM2.5, because most soil particles are greater

than 2.5mm in diameter [44]. Based on the low concentration ofPM2.5 and the reduction of TSP and PM10 after paving roads, itbecomes apparent that vehicular traffic on unpaved roads is themajor source of airborne particulate matter for Iqaluit, whichsupports the findings of Forsberg et al. [22] and Omsted et al.[23]. This indicates that under certain conditions, the primarysource of particulate matter can be geological material.

In the present case study, soil inhalation is the dominantpathway of carcinogenic risk for Cr and Cd. Despite the fact thatinhalation of particulate matter contributes a maximum of only10% of the EDI, inhalation hazard quotients are comparable totheir ingestion counterparts; for Be, Cr, and Mn, the inhalationhazard quotient is greater than ingestion. The results generatedhere show that whereas soil ingestion dominates in terms oftotal exposure, this is only half of the equation in terms of riskassessment. The other half is TDI or CSF, and for manycontaminants there are TDIs and/or CSFs for both the inhalationand ingestion pathways ([34]; http://www.epa.gov/IRIS/). Thus,

Fig. 2. Estimated daily intake (EDI; in mg kg�1 d�1) for a toddler. Inhalation of particulate matter (PM) and ingestion of PM are stacked for comparison againstingestion of soil. Contaminants are ordered from highest to lowest exposure from the combined PM exposure. Polycyclic aromatic hydrocarbons are abbreviatedas follows: Nap¼ naphthalene; BbF¼ benzo[b]fluoranthrene; BkF¼ benzo[k]fluoranthrene; BaP¼ benzo[a]pyrene; IcdP¼ indeno[1,2,3–cd]pyrene; BaP Eq¼benzo[a]pyrene equivalents.

Fig. 3. Average concentration of airborne particulate matter (n¼ 3 or 4).Error bars represent the standard deviation of the mean. By Student’s t test,paving roads significantly reduces the concentration of total suspendedparticulates (TSP; p< 0.1) and particulate matter less than 10mm (PM10;p< 0.05) but not particulate matter less than 2.5mm (PM2.5).

Paving roads reduces risk from soil inhalation Environ. Toxicol. Chem. 31, 2012 5

whereas exposure from soil ingestion can be one to two ordersof magnitude greater than that from soil inhalation, the inha-lation pathway often has a TDI that is much lower, or a CSF thatis much higher, than the ingestion pathway, making soil inha-lation relevant for risk assessment. The soil inhalation risk isattributed mainly to exposure from PM10, the respirable frac-

tion of airborne particulate matter, resulting from the vehiculartraffic on unpaved roads. Before paving roads, the City ofIqaluit used CaCl2 as a dust suppressant to control soil resus-pension, so it is highly probable that the effect of paving roadshas a greater capacity to reduce airborne particulate concen-trations then what has been displayed here (Fig. 3). The

Fig. 4. Percentage of airborne concentration after paving roads. Percentages were calculated as the mean postpaving concentrations (n¼ 3) divided by themean prepaving concentrations (n¼ 4; in mgm�3) from total suspended particulates (TSP). Polycyclic aromatic hydrocarbons are abbreviated asfollows: Nap¼ naphthalene; BbF¼ benzo[b]fluoranthrene; BkF¼ benzo[k]fluoranthrene; BaP¼ benzo[a]pyrene; IcdP¼ indeno[1,2,3-cd]pyrene; BaP Eq¼benzo[a]pyrene equivalents.

Fig. 5. Carcinogenic risk to toddlers and adults before andafter paving roads.Carcinogenic risk forCr exposure is expressedper 100,000 individuals, and all othersare expressed per 1,000,000 individuals. Polycyclic aromatic hydrocarbon carcinogenicity has been converted to benzo[a]pyrene equivalents using CanadianCouncil of Ministers of the Environment guidelines. Age-dependent adjustment factors were used to account for the susceptible early life stages.

6 Environ. Toxicol. Chem. 31, 2012 K. James et al.

inhalation risk to human health in Iqaluit is a function ofunpaved roads providing an infinite reservoir of contaminantsto be resuspended [20] and the lower TDI or higher CSF for theinhalation pathway.

According to Siliciano et al. [19], Ag, As, and Cd haveincreasing concentration with decreasing particle size in soil,and their enrichment factors may be overestimated because ofthe natural enrichment; however Cr, Ni, Co, Sb, and Hg havedecreasing concentrations with decreasing particle size, andtheir enrichment factors may be underestimated. Notably, Baconcentration was not linked to particle size, and enrichmentfactors for Ba ranged from 29 to 48, indicating that there is asignificant nonsoil source for Ba; however, the reduction inairborne concentration of Ba after paving was greater than thatfor any other contaminant. Reduction in Ba concentration frompaving roads is currently unknown to the authors but is underinvestigation.

Not explored here is the potential of contaminants to leachinto the soil from the pavement. Undoubtedly, chemicalspresent in concrete or asphalt can leach into the surroundingsoil and potentially increase the carcinogenic risk. The con-tribution of hazardous contaminants leaching into the soil iscurrently unknown; however, other researchers have found therisk associated with contaminant leaching from concrete andasphalt pavement to be minimal [45–47]. In contrast, researchby Van Metre et al. [48] has shown that asphalt and coal tarsealant can provide a significant source of PAHs in dust, water,and sediment. Notably, sealant makes the significant contribu-tion, whereas the contribution from asphalt or concrete isnegligible in comparison [49].

The risk assessment described herein utilizes various levelsof conservatism. The inductively coupled plasma–mass spec-trometry method used for metal analysis reports the total metalcontent within the sample and does not determine speciation.Metal toxicity varies with speciation, so, when multiple TDIsand CSFs were available based on metal speciation, the mostconservative value was always used. In the case of Cr, carcino-genic risk could be as low as 2.3 in 1,000,000 as opposed to the1.6 in 100,000 if the primary species present is trivalent Cr andnot hexavalent Cr [33]. Comparing the environmental condi-tions in Iqaluit to published values [50] indicates that hexava-

lent Cr may be the dominant species, and, without furtheranalysis, it is impossible to rule out hexavalent Cr as thedominant species, so the most conservative CSF was used.Furthermore, we used the most conservative exposure factorsfrom the Canadian Compendium, which does not explicitlyaccount for changes in human activity occurring during longwinter months; in Iqaluit such months range from Septemberuntil late May. For the exposure assessment, we overestimateexposure by quantifying the 2.5- to 10-mm size fraction for bothinhalation and ingestion exposures. We justify this overestima-tion because the ingestion exposure makes a negligible con-tribution to the cumulative exposure (�1% to the total EDI),and, although particulate matter is swept into the GI tract, it isunknown whether individual contaminants are completelyabsorbed while in the lower respiratory tract.

In judging from the data presented here, vehicular traffic onunpaved roads can be a major source of airborne particulatematter. Although soil ingestion is still the dominant exposurepathway, soil inhalation can result in excessive risk to humanhealth. Paving roads was found to be an effective measure formitigating risk by reducing the amount of soil particles resus-pended into the atmosphere.

Acknowledgement—The present study was supported by a Natural SciencesandEngineeringResearchCouncil ofCanada strategic grant toS.D. Sicilianoand R.E. Farrell. The scientific research licence from the Nunavut ResearchInstitute is 0101107N-M. Support from the Nunavut Research Institute isgreatly appreciated. The authors especially thank B. Laird for assistance incritical review.

REFERENCES

1. Nriagu JO, Pacyna JM. 1988. Quantitative assessment of worldwidecontamination of air, water and soils by trace-metals. Nature 333:134–139.

2. Grzebisz W, Ciesla L, Komisarek J, Potarzycki J. 2002. Geochemicalassessment of heavy metals pollution of urban soils. Pol J Environ Stud11:493–499.

3. Loska K, Wiechula D, Korus I. 2004. Metal contamination of farmingsoils affected by industry. Environ Int 30:159–165.

4. Menzie CA, Potocki BB, Santodonato J. 1992. Exposure to carcinogenicPAHs in the environment. Environ Sci & Technol 26:1278–1284.

5. Srogi K. 2007. Monitoring of environmental exposure to polycyclicaromatic hydrocarbons: A review. Environ Chem Lett 5:169–195.

Table 2. Hazard quotients for human exposure at a contaminated brownfield site

Contaminant

HQ inhalation PM

Percentagereduction (%)

HQ ingestion PM

Percentagereduction (%) HQ ingestion soilBefore paving After paving Before paving After paving

Ag — — — 6.3� 10�6 3.2� 10�6 50 6.0� 10�7

As — — — 4.7� 10�4 2..0� 10�5 57 1.4� 10�2

Ba — — — 2.8� 10�2 1.2� 10�3 96 1.5� 10�2

Be 2.0� 10�2 1.8� 10�2 13 1.0� 10�5 7.4� 10�6 26 8.1� 10�5

Cd — — — 1.7� 10�4 1.1� 10�4 38 1.3� 10�3

Cr 3.2� 10�2 1.5� 10�2 46 1.0� 10�4 4.8� 10�5 52 1.8� 10�3

Cu — — — 1.3� 10�6 7.4� 10�7 43 4.1� 10�4

Hg 9.1� 10�7 4.5� 10�7 50 1.0� 10�7 8.7� 10�8 13 1.1� 10�4

Mn 3.7� 10�3 9.4� 10�4 75 8.0� 10�7 2.1� 10�7 75 2.0� 10�3

Mo — — — 7.2� 10�6 3.4� 10�6 53 1.5� 10�5

Ni 2.1� 10�3 1.0� 10�3 55 1.3� 10�6 7.1� 10�7 45 2.7� 10�3

Pb — — — 2.9� 10�5 1.8� 10�5 38 4.4� 10�3

Sb 9.0� 10�5 2.1� 10�5 77 1.8� 10�5 1.0� 10�5 43 1.7� 10�4

Sr — — — 1.2� 10�5 7.1� 10�6 40 2.5� 10�5

U — — — 4.3� 10�5 2.2� 10�5 49 5.1� 10�4

V — — — 2.3� 10�5 8.8� 10�6 61 2.3� 10�4

Zn — — — 1.7� 10�3 1.2� 10�3 28 6.2� 10�4

Napthalene 6.0� 10�6 7.7� 10�7 87 1.7� 10�7 8.6� 10�8 49 7.4� 10�5

HQ¼ hazard quotients; PM¼ particulate matter.

Paving roads reduces risk from soil inhalation Environ. Toxicol. Chem. 31, 2012 7

6. Filippelli GM, Laidlaw MAS. 2010. The elephant in the playground:Confronting lead-contaminated soils as an important source of leadburdens to urban populations. Perspect Biol Med 53:31–45.

7. LaidlawMAS, TaylorMP. 2011. Potential for childhood lead poisoningin the inner cities ofAustralia due to exposure to lead in soil dust.EnvironPollut 159:1–9.

8. Ghorbel M, Munoz M, Courjault-Rade P, Destrigneville C, de ParsevalP, Souissi R, Souissi F, Ben Mammou A, Abdeljaquad S. 2010. Healthrisk assessment for human exposure by direct ingestion of Pb, Cd, Znbearing dust in the former miners’ village of Jebel Ressas (NE Tunisia).Eur J Mineral 22:639–649.

9. Zhang H, Huang GH, Zeng GM. 2009. Health risks from arsenic-contaminated soil in Flin Flon–Creighton, Canada: Integrating geo-statistical simulation and dose–response model. Environ Pollut 157:2413–2420.

10. Qiao M, Cai C, Huang YZ, Liu YX, Lin AJ, Zheng YM. 2011.Characterization of soil heavy metal contamination and potential healthrisk in metropolitan region of northern China. Environ Monit Assess172:353–365.

11. Lai HY, Hseu ZY, Chen TC, Chen BC, Guo HY, Chen ZS. 2010. Healthrisk-based assessment and management of heavy metals-contaminatedsoil sites in Taiwan. Int J Environ Res Publ Health 7:3595–3614.

12. Bright DA, Richardson GM, Dodd M. 2006. Do current standards ofpractice in Canada measure what is relevant to human exposure atcontaminated sites? I: A discussion of soil particle size and contaminantpartitioning in soil. Hum Ecol Risk Assess 12:591–605.

13. Granero S, Domingo JL. 2002. Levels of metals in soils of Alcala deHenares, Spain: Human health risks. Environ Int 28:159–164.

14. The Netherlands National Institute for Public Health and the Environ-ment. 2001. Evaluation and Revision of the CSOIL Parameter Set(RIVM report 711701021). Bilthoven, The Netherlands.

15. Poggio L, Vrscaj B, Hepperle E, Schulin R, Marsan FA. 2008.Introducing a method of human health risk evaluation for planning andsoil quality management of heavy metal-polluted soils—An examplefrom Grugliasco (Italy). Landscape Urban Plan 88:64–72.

16. Guney M, Zagury GJ, Dogan N, Onay TT. 2010. Exposure assessmentand risk characterization from trace elements following soil ingestion bychildren exposed to playgrounds, parks and picnic areas. JHazardMater182:656–664.

17. Sing D, Sing CF. 2010. Impact of direct soil exposures from airbornedust and geophagy on human health. Int J Environ Res Public Health7:1205–1223.

18. Skinner HCW. 2007. The earth, source of health and hazards: Anintroduction tomedical geology.AnnuRevEarthPlanetSci35:177–213.

19. SicilianoSD, JamesK,ZhangGY,SchaferAN,Peak JD.2009.Adhesionand enrichment of metals on human hands from contaminated soil at anArctic urban brownfield. Environ Sci Technol 43:6385–6390.

20. Gillies JA, Etyemezian V, Kuhns H, Nikolic D, Gillette DA. 2005.Effect of vehicle characteristics on unpaved road dust emissions.Atmos Environ 39:2341–2347.

21. Magliano KL, Hughes VM, Chinkin LR, Coe DL, Haste TL, Kumar N,LurmannFW. 1999. Spatial and temporal variations in PM10 andPM2.5source contributions and comparison to emissions during the 1995integrated monitoring study. Atmos Environ 33:4757–4773.

22. Forsberg B, Hansson HC, Johansson C, Areskoug H, Persson K,Jarvholm B. 2005. Comparative health impact assessment of local andregional particulate air pollutants in Scandinavia. Ambio 34:11–19.

23. OmstedtG,BringfeltB, JohanssonC. 2005.Amodel for vehicle-inducednon-tailpipe emissions of particles along Swedish roads. Atmos Environ39:6088–6097.

24. Watson JG, Chow JC. 2000. Reconciling Urban Fugitive DustEmisssions Inventory and Ambient Source Contribution Estimates:Summary of Current Knowledge andNeeded Research. Desert ResearchInstitute, Reno, NV, USA.

25. Clairborn C, Mitra A, Adams G, Bamesberger L, Allwine G,Kantamaneni R, et al. 1995. Evaluation of PM10 emission rates frompaved and unpaved roads using tracer techniques. Atmos Environ29:1075–1089.

26. Chow JC, Watson JG, Houck JE, Pritchett LC, Rogers CF, Frazier CA,Egami RT, Ball BM. 1994. A laboratory resuspension chamber tomeasure fugitive dust size distributions and chemical compositions.Atmos Environ 28:3463–3481.

27. Siciliano SD, Laird BD, Lemieux CL. 2010. Polycyclic aromatichydrocarbons are enriched but bioaccessibility reduced in brownfieldsoils adhered to human hands. Chemosphere 80:1101–1108.

28. Laird BD, Peak D, Siciliano SD. 2011. Bioaccessibility of metal cationsin soil is linearly related to its water exchange rate constant. Environ SciTechnol 45:4139–4144.

29. Richardson GM. 1997. Compendium of Canadian Human ExposureFactors for Risk Assessment. O’Connor Associates Environmental,Ottawa, Ontario, Canada.

30. James K, Peters RE, Laird BD,MaWK,WickstromM, Stephenson GL,Siciliano SD. 2011. Human exposure assessment: A case study of eightPAH contaminated soils using in vitro digestors and the juvenile swinemodel. Environ Sci Technol 45:4586–4593.

31. Marriott PJ, Carpenter PD, Brady PH, McCormick MJ, Griffiths AJ,Hatvani TSG, Rasdell SG. 1993. Optimization of fluorescence detectionfor polyaromatic hydrocarbondetermination byusing high-performanceliquid chromatography. J Liquid Chromatogr 16:3229–3247.

32. Gao Y, Nelson ED, Field MP, Ding Q, Li H, Sherrell RM, Gigliotti C,Van Ry DA, Glenn T, Eisenreich SJ. 2002. Characterization ofatmospheric trace elements on PM2.5 particulate matter over the NewYork–New Jersey Harbor estuary. Atmos Environ 36:1077–1086.

33. Gemenetzis P, Moussas P, Arditsoglou A, Samara C. 2006. Massconcentration and elemental composition of indoor PM2.5 and PM10in university rooms in Thessaloniki, northern Greece. Atmos Environ40:3195–3206.

34. Health Canada. 2004. Federal Contaminated Site Risk Assessment inCanada. Part II.HealthCanadaToxicologicalReferenceValues (TRVs).Contaminated Sites Division, Safe Environments Programme, HealthCanada, Ottawa, Ontario, Canada.

35. Hawley JK. 1985. Assessment of health risk from exposure tocontaminated soil. Risk Anal 5:289–302.

36. Abt E, Suh HH, Allen G, Koutrakis P. 2000. Characterization of indoorparticle sources: A study conducted in the metropolitan Boston area.Environ Health Persp 108:35–44.

37. U.S. Environmental Protection Agency. 2005. Supplemental Guidancefor Assessing Cancer Susceptibility from Early-Life Exposure toCarcinogens. EPA/630/R-03/003F. Washington, DC.

38. CanadianCouncil ofMinisters of the Environment. 2010.Canadian SoilQuality Guidelines for the Protection of Environmental and HumanHealth: Carcinogenic and Other Polycyclic Aromatic Hydrocarbons(PAHs). Winnipeg, Manitoba, Canada.

39. Hussein T, Johansson C, Karlsson H, Hansson HC. 2008. Factorsaffecting non-tailpipe aerosol particle emissions from paved roads: On-roadmeasurements in Stockholm, Sweden. Atmos Environ 42:688–702.

40. Celo V, Dabek-Zlotorzynska E., 2010. Concentration and source originof trace metals in PM2.5 collected at selected Canadian Sites within theCanadian National Air Pollution Surveillance Program. In Zereini F,Wiseman CLS, eds. Urban Airborne Particulate Matter. Springer,New York, NY, USA, pp 19–38.

41. Brook JR, Dann TF, Burnett RT. 1997. The relationship among TSPPM(10) PM(25) and inorganic constituents of atmospheric particulatematter at multiple Canadian locations. J Air Waste Manag 47:2–19.

42. Sanderson EG, Raqbi A, Vyskocil A, Farant JP. 2004. Comparison ofparticulate polycyclic aromatic hydrocarbon profiles in different regionsof Canada. Atmos Environ 38:3417–3429.

43. Blanchard P, Brook JR, Brazal P. 2002. Chemical characterization of theorganic fraction of atmospheric aerosol at two sites in Ontario, Canada.J Geophys Res Atmos 107 (D21).

44. Querol X, AlastueyA, Rodriguez S, Plana F,Mantilla E, Ruiz CR. 2001.Monitoring of PM10 and PM2.5 around primary particulate anthro-pogenic emission sources. Atmos Environ 35:845–858.

45. SadlerR,DelamontC,WhiteP,ConnellD. 1999.Contaminants in soil asa result of leaching from asphalt. Toxicol Environ Chem 68:71–81.

46. Marion AM, De Laneve M, De Grauw A. 2005. Study of the leachingbehaviour of paving concretes: Quantification of heavy metal content inleachates issued from tank test using demineralized water. CementConcrete Res 35:951–957.

47. Kayhanian M, Vichare A, Green PG, Harvey J. 2009. Leachabilityof dissolved chromium in asphalt and concrete surfacing materials.J Environ Manag 90:3574–3580.

48. Van Metre PC, Mahler BJ, Wilson JT. 2009. PAHs underfoot:Contaminated dust from coal-tar sealcoated pavement is widespreadin the United States. Environ Sci Technol 43:20–25.

49. Simon JA, Sobieraj JA. 2006. Contributions of common sources ofpolycyclic aromatic hydrocarbons to soil contamination. Remediat J16:25–35.

50. Kotas J, Stasicka Z. 2000. Chromiumoccurrence in the environment andmethods of its speciation. Environ Pollut 107:263–283.

8 Environ. Toxicol. Chem. 31, 2012 K. James et al.