Perfluoroalkyl acids in urban stormwater runoff: Influence of land use

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 6 0 1 e6 6 0 8

Available online at www.sciencedirect.com

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

Perfluoroalkyl acids in urban stormwater runoff: Influence of land use Feng Xiao a,b,*, Matt F. Simcik b, John S. Gulliver a,** a b

Department of Civil Engineering, University of Minnesota, Minneapolis, MN 55414, USA Division of Environmental Health Sciences, School of Public Health, University of Minnesota, Minneapolis, MN 55455, USA

article info

abstract

Article history:

Perfluoroalkyl acids (PFAAs) are persistent organic pollutants in the environment and have

Received 29 June 2011

been reported to have nonpoint sources. In this study, six PFAAs with different chain

Received in revised form

lengths were monitored in stormwater runoff from seven storm events (2009e2011) at

9 November 2011

various outfall locations corresponding to different watershed land uses. We found PFAA(s)

Accepted 9 November 2011

in 100% of stormwater runoff samples. Monitoring results and statistical analysis show

Available online 22 November 2011

that PFAAs in stormwater runoff from residential areas mainly came from rainfall. On the other hand, non-atmospheric sources at both industrial and commercial areas contributed

Keywords:

PFAAs in stormwater runoff. The mass flux of PFAAs from stormwater runoff in the Twin

Urban runoff

Cities (Minneapolis and St. Paul, MN) metropolitan area is estimated to be about 7.86 kg/

Stormwater pollution

year. In addition, for the first time, we monitored PFAAs on the particles/debris in

Nonpoint source pollution

stormwater runoff and found high-level PFOS on the particulate matter in runoff collected

PFOS

from both industrial and commercial areas; the levels were so high that the finding could

PFOA

not be explained by the solidewater partitioning or adsorption. PFOS on the particulate

Water quality

matter is suspected to have originated from industrial/commercial products, entering the waste stream as PFOS containing particles. ª 2011 Elsevier Ltd. All rights reserved.

1.

Introduction

Urban stormwater runoff carries various pollutants from the urban watershed, such as nutrients from lawns and landscaped areas, heavy metals from roofs, brake wear and tire wear, and organic pollutants (Baun et al., 2006; Chocat et al., 2007; Davis et al., 2001; Deletic et al., 1997; Eriksson et al., 2007; Genc¸-Fuhrman et al., 2007; Gulliver et al., 2009). While stormwater management practices (SMPs) can remove many contaminants (Weiss et al., 2007), conventional SMPs may not efficiently remove perfluoroalkyl acids (PFAAs) from stormwater. PFAAs are industrially produced compounds that contain a perfluorinated alkyl moiety of varying chain length

and varying functional groups attached to that moiety. They have been produced for both industrial and commercial applications including aqueous fire fighting foam (AFFF), insecticides and as polymers to repel water and stains on paper and textiles including fabric and carpeting (Emmett et al., 2006; Giesy and Kannan, 2001; Prevedouros et al., 2006). They have recently become the target of investigation by environmental chemists due to their persistence and toxicity and their ubiquitous presence in the environment (Emmett et al., 2006; Giesy and Kannan, 2001; Kim and Kannan, 2007; Jin et al., 2009; Martin et al., 2003; Moody et al., 2003; Murakami et al., 2008; Nakayama et al., 2010; Nguyen et al., 2011; Prevedouros et al., 2006; Schultz et al.,

* Corresponding author. Department of Civil Engineering, University of Minnesota, Minneapolis, MN 55414, USA. Tel.: þ1 651 964 5138; fax: þ1 612 626 7750. ** Corresponding author. Tel.: þ1 612 625 4080; fax: þ1 612 626 7750. E-mail addresses: [email protected] (F. Xiao), [email protected] (J.S. Gulliver). 0043-1354/$ e see front matter ª 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2011.11.029

6602

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 6 0 1 e6 6 0 8

2006; Zushi et al., 2008). Because of the strong carbonefluorine bonds, PFAAs are not easily degraded by physical or chemical mechanisms once in the environment. With the relatively high solubility, low adsorption potentials (Ahrens et al., 2011; Higgins and Luthy, 2006; Pan and You, 2010; Tang et al., 2010; Xiao et al., 2011), and very low or possibly negligible volatility (Emmett et al., 2006; Giesy and Kannan, 2001; Prevedouros et al., 2006), they are also not readily removed by conventional drinking/waste water treatment processes (Loganathan et al., 2007; Takagi et al., 2008; Yu et al., 2009). PFAAs such as perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) have been found in fish, birds and mammals from mid-latitudes to the poles (Fei et al., 2007; Giesy and Kannan, 2001; Houde et al., 2011; Loi et al., 2011) and in human blood at concentrations of several to tens of mg/L (Calafat et al., 2007; Olsen et al., 2007). PFAAs in blood serum can disrupt human hormone activity (Weiss et al., 2009) and their developmental toxicity to mammals has been reported (Lau et al., 2004). Simcik and Dorweiler (2005) documented that PFAAs in urban water bodies mainly come from non-atmospheric sources, including wastewater treatment plants (WWTPs) (point source pollution) and surface runoff (nonpoint source pollution). In 2007, several metropolitan lakes in Twin Cities of Minneapolis and St. Paul, Minnesota (U.S.) were labeled impaired for contamination with a suite of PFAAs, primarily PFOS in fish (MPCA, 2008). While Minnesota is home to the 3 M corporation that produced and disposed of many of these chemicals, many of the lakes listed as impaired have no connection to 3 M’s production or disposal (MPCA, 2008). In addition, none of these lakes receives direct wastewater discharge. We hypothesize that a significant source of PFAAs to surface waters is urban runoff, which receives PFAAs from transportation, industrial and residential sources within its watershed. Compared to point source pollution, fewer references are available for the environment contamination of PFAAs from nonpoint sources. In Japan, Zushi et al. (2008) found that a river received a higher load of PFAAs from stormwater runoff than from WWTPs. Furthermore, different PFAA profiles between WWTPs effluents and street runoff were found in Japan (Murakami et al., 2009b). Street runoff was found to contain more long-chain (>8 perfluorinated carbon) and even-chain perfluorocarboxylates than WWTP effluents (Murakami et al., 2008, 2009a, 2009b). Aside from these studies in Japan, however, only one published report has been available about PFAAs in surface runoff (Kim and Kannan, 2007). Therefore, a major question regarding the global distribution is whether urban runoff should be viewed as a potential source of PFAAs to urban waters. Once significant sources of contaminants are identified and confirmed, reducing these inputs to receiving waters can be addressed (Davis et al., 2001). Ha and Stenstrom (2003) pointed out that the concentrations and relative abundances of contaminants in stormwater runoff are closely related to various types of land use. However, little is known about the influences of land uses on the levels of PFAAs in urban runoff. The objective of this paper is to monitor stormwater runoff and evaluate the concentrations of PFAAs present in urban runoff. The paper will present results from an urban runoff study, in which runoff was collected and analyzed from different storm events at various outfall locations

corresponding to different land uses (industrial, residential, and commercial areas).

2.

Materials and methods

2.1.

Chemicals

PFAAs (perfluoroheptanoic acid (PFHpA, 99 percent), PFOA (w95 percent), perfluorononanoic acid (PFNA, 97 percent), PFOS (98 percent), perfluorodecanoic acid (PFDA, 98 percent), and perfluoroundecanoic acid (PFUnDA, 95 percent)) were purchased from SigmaeAldrich (Milwaukee, WI, USA & Steinheim, Switzerland) (Table S1 in the Supplementary data). Isotopically labeled surrogate standards, 13C4-PFOA, 13C5PFNA, 13C4-PFOS, and 13C2-PFDA, were acquired from Cambridge Isotope Laboratories (MA, USA). The labeled PFAAs were dissolved into methanol (Optima grade, Fisher Scientific, Hanover Park, IL) to 600 ng/mL of each PFAA as internal standards. Methanol and reverse osmosis (R.O.) or HPLC water (HPLC grade, Fisher Scientific, Hanover Park, IL) were used to clean the containers and equipment. PFAAs were not found in the methanol or the water when this was checked using a Hewlett Packard 1100 MSD mass spectrometer (MS) equipped with an electrospray ionization (ESI) source and interfaced to a high-pressure liquid chromatograph (HPLC).

2.2.

Site selection and sample collection

The samples of stormwater runoff were taken from three residential sites, labeled CSCC (44 590 1700 N, 93 120 5400 W), CTC (44 580 4600 N, 93 110 0000 W) and Mayo (44 580 2000 N, 93 130 5600 W) (Fig. S1 in the Supplementary data), one commercial and heavily trafficked area (Dinkytown, 44 580 5200 N, 93 140 1300 W) (Fig. S1), and multiple sites near an industrial area of suspected PFAA contamination (Fig. S2 in the Supplementary data). Land use in these sites is characterized with percent urban land use ranging from 45.8 to 62.9% and with a 1.3e3.1% increase in households from 2000 to 2010 (Metropolitan Council, 2011). Dinkytown is within the same country as CSCC and Mayo; however, it a busy commercial center with a mix of bars, retail and entertainment uses. Fig. S3 in the Supplementary data shows the land use information of Dinkytown. Runoff was collected for seven separate rain events with a precipitation rate larger than 8 mm/h from 2009 to 2011 (event 1: Sep 25 2009; event 2: Aug 20 2010; event 3: Sep 01 2010; event 4: Sep 15 2010; event 5: Sep 23 2010; event 6: Oct 26 2010; event 7: May 12 2011). At each site, about 4 L of runoff samples were collected by lowering a high density polyethylene container into the stormwater flow from street level down drain holes. Samples were taken after the initial surface runoff (first flush), in spite of the fact that the concentrations of short- to mediumchain-length PFAAs (PFOA, PFNA, and PFOS) were not changed significantly in runoff in a fixed-point hourly monitoring experiment (Zushi and Masunaga, 2009a). All the containers and equipment were pre-cleaned by rinsing with Optima grade methanol (five times) followed by R.O. water (ten times) and then rinsed with runoff water (three times) at each site before collecting samples. The collected samples were transported to the lab and refrigerated at 4  C for further treatment.

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 6 0 1 e6 6 0 8

Details of the sample pretreatment, PFAA quantification by HPLCeESI-MS, recovery and procedural blanks, and the limit of quantification (LOQ) and the limit of detection (LOD) can be found in the Supplementary data.

2.3.

Statistical analysis

To compare means between three or more groups, ANOVA and the following post-hoc test (Tamhane’s T2 test for unequal variance, Pagano and Gauvreau, 2000) were applied to data that were normally distributed. The normality was checked by the ShapiroeWilk test. For data that were not normally distributed, a non-parametric test, ManneWhitneyeWilcoxon (MWW) test, was used to compare the medians between groups. The widely adopted significance level, 0.05, was adjusted during multiple comparisons to avoid increasing the risk of rejecting a true null hypothesis (Pagano and Gauvreau, 2000). For four groups, six comparisons were needed ð¼ 4!=ð4  2Þ!2!Þ and the significance level was adjusted to 0.0083. For the comparison between six groups, the significance level was adjusted to 0.0033. For comparing the proportions between groups, a non-parametric test (Chi-square test) was used and the significance level was set as 0.05. All of the tests were performed using SPSS 17.0 (now known as PASW, for Window, SPSS Inc. IL, USA). The statistical analysis results are provided in the Supplementary data.

Mayo and Dinkytown were found with respect to the mean concentrations of PFHpA, PFNA, and PFDA in stormwater runoff. However, the results of statistical tests show that the mean runoff concentrations of PFOS and PFOA at Dinkytown were significantly higher than the mean concentrations at CTC (two-sided p value ¼ 0.008 for PFOA and p ¼ 0.001 for PFOS) and at Mayo (two-sided p value ¼ 0.007 for PFOA and p ¼ 0.004 for PFOS). Compared to CTC and Mayo, Dinkytown is a busy district containing several city blocks occupied by various small businesses, restaurants, food courts, and bars. Murakami and Takada (2008) found that the PFAA content was significantly higher in heavily trafficked street dust than in residential street dust. Kim and Kannan (2007) found high concentrations of PFOA in runoff collected at a site influenced by heavy traffic. The traffic flow (the number of vehicles passing a reference point per unit of time) around the sampling site at Dinkytown is about 1416 vehicles per hour (15 min counts started at 4:30 p.m.), which is much higher than the residential areas (CSCC, 456; CTC, 216; Mayo, 84 vehicles per hour). Therefore, the vehicular traffic around Dinkytown possibly acts as an important source of PFAAs in runoff. Several studies have found PFAAs in dust from cars (Bjo¨rklund et al., 2009; Goosey and Harrad, 2011). In addition, given that PFAAs have been used in various materials (including commercial food packaging), these materials used in the (fast food) restaurants of Dinkytown could be another source of PFAAs in runoff collected from this area.

3.2.

3.

6603

Profile of PFAAs in urban runoff

Results and discussion

3.1. Concentrations of PFAAs in stormwater runoff in urban residential and commercial areas As shown in Fig. 1, the concentrations of PFAAs in stormwater runoff differ from event to event. The total concentration of PFAAs ranged from 14.3 ng/L for event 4 at CTC to 96.0 ng/L for event 5 at Dinkytown. The most abundant PFAAs in stormwater runoff were PFOS and PFOA. The concentrations ranged up to 42.5 ng/L for PFOS and 30.6 ng/L for PFOA. For the three residential areas (CSCC, CTC and Mayo), the total concentrations of PFAAs in runoff were smaller than 30 ng/L and no PFAA concentration was larger than 20 ng/L. On the other hand, relatively high concentrations of PFOS and PFOA were found in runoff collected from Dinkytown, the commercial area. One-way ANOVA and the following post-hoc test were used to determine the significance of the differences among the four sites (CSCC, CTC, Mayo and Dinkytown) in terms of PFAA concentrations in the stormwater runoff. The values of skewness of the PFAA concentrations in the four areas lie mostly between 0 and 1 (except PFUnDA), indicating that the data of PFAA concentrations can be treated as normally distributed and can be tested by the robust one-way ANOVA and post-hoc tests. The Tamhane’s T2 post-hoc test was chosen for unequal variance (the ratios of the largest standard deviation to the smallest standard deviation are larger than two for all PFAAs cases) (Pagano and Gauvreau, 2000). The one-way ANOVA and Tamhane’s T2 post-hoc test results are shown in the Supplementary data. At a significance level of 0.0083, no significant differences among CSCC, CTC,

The PFAA concentrations are presented in a box plot (Fig. 2). Although Dinkytown is a commercial area, the concentrations of PFAAs in runoff collected from Dinkytown were included in Fig. 2 to better represent the multiple land uses of urban areas. The data shown in Fig. 2 are positively skewed (except PFNA); the skewness is caused by the outliers that are large values relative to the mean. For skewed data, the median is a better measure of the center than the mean (Pagano and Gauvreau, 2000). To determine the relative abundance of PFAA in urban stormwater runoff, the MWW test of population medians was applied to the data. For multiple comparisons of the sampling distribution of the sample median among six groups, the adjusted significance level is 0.0033. The MWW test results reveal that the median concentration of PFOA (9.3 ng/L) is not significantly different from the median concentration of PFOS (10.6 ng/L). However, their median concentrations are significantly higher than other PFAAs (PFHpA, PFNA, PFDA, and PFUnDA). The results also show that the median concentrations of PFHpA (0.63 ng/ L), PFNA (3.65 ng/L), and PFDA (1.07 ng/L) are not significantly different from each other. The overall median PFAA concentration is 25.8 ng/L. The frequencies of detection for PFAAs in urban stormwater runoff are presented in Fig. S4 in the Supplementary data. As shown, PFOS and PFOA were detected in all samples from residential areas. A Chi-square test was used to examine whether the differences in the proportion of detection were significant. The test results show that the frequencies of detection of PFOS and PFOA are much higher than the other four PFAAs at a significance level of 0.05. Overall, PFOS, PFOA,

6604

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 6 0 1 e6 6 0 8

PFHpA PFOS

PFOA PFDA

PFNA PFUnA

40

Mayo

15

Dinkytown

ng/L

ng/L

30 10

5

0

20 10

2

3

4

5

6

0

7

2

3

Storm Events

CTC

15

5

6

7

6

7

CSCC

15

ng/L

ng/L

10

10

5

5

0

4

Storm Events

0

2

3

4

5

6

7

2

3

Storm Events

4

5

Storm Events

Fig. 1 e PFAA concentrations in street runoff in different locations during six storm events. Non-detect values were measured at CTC and CSCC for events 5, 6, and 7.

3.3. Comparison between stormwater runoff and rainfall PFAAs were analyzed in rainfall from six rain events, three of which were at the same dates when surface runoff samples were taken (Fig. S5 in the Supplementary data). The median concentrations in rainfall were 1.3, 8.4, 1.3, 6.7, 2.6, and
rainfall were similar to the profiles in runoff from three residential areas (CSCC, CTC, and Mayo) but distinctively different from the profiles in runoff from the commercial area (Dinkytown) (Fig. 3). Both the statistical analysis and Fig. 3 indicate that non-atmospheric sources existed in the Dinkytown that contribute PFOS and PFOA to stormwater runoff.

40 30 ng/L

and PFNA are the major species, accounting for 38.3  3.7% (mean  SEM), 37.0  2.3%, and 11.2  2.6%, respectively, of the total PFAAs in the monitored residential stormwater runoff.

max Q3 median Q1 min mean

20 10 0

PFNA PFDA PFHpA PFUnDA PFOS PFOA Fig. 2 e PFAA concentrations in stormwater runoff collected from Dinkytown, CTC, CSSS, and Mayo. The concentrations of PFAAs less than LOQ were assigned as half of the LOQ value, and the concentrations of PFAAs that were below LOD were assigned as zero.

6605

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 6 0 1 e6 6 0 8

PFOA

PFNA

PFOS

PFDA

the median concentration either in the stormwater runoff at the residential/commercial areas or in rainfall (see Supplementary data for the statistical analysis). The industrial area and the subsequent loadings to stormwater were expected to be responsible for PFOS contamination of nearby Lake Calhoun (MPCA, 2008). Because of their relatively high solubility, PFAAs could be washed out from the products within the industrial source by rain and enter into surface runoff. Many years after the phase out of the production of PFOS by a major manufacturer, the relatively high-level PFOS in runoff around the industrial source was traced to an industry nearby that used PFOS coated materials in their process (MPCA, 2008).

PFUnDA

Rain Residential Areas Commercial Area

Event #4

Rain Residential Areas

Event #6

Event #2

PFHpA

Residential Areas

Commercial Area

Rain

Commercial Area

0

10

20

30 40 50 60 Concentration (ng/L)

70

80

3.5.

90

Fig. 3 e PFAA profiles in stormwater runoff at the residential and commercial areas and in rainfall at three heavy rain events (Event 2: August 20 2010; Event 4: Sep 15 2010; Event 6: Oct 26 2010).

Table 1 summarizes previously reported concentrations of PFAAs in runoff. The profiles of PFAAs in U.S. stormwater runoff are different from those in Japan and Singapore (Murakami et al., 2009b; Nguyen et al., 2011; Zushi and Masunaga, 2009b), where PFNA is a primary PFAA in the environment (Table 1). One possible reason for the difference is that the ammonium salt of PFNA is primarily manufactured in Japan while the ammonium salt of PFOA is largely manufactured in the U.S. (Murakami et al., 2009b).

3.4.

Industrial area

Monitoring similar to the residential and commercial areas was conducted near a suspected industrial source (Table 2). Relatively high concentrations of PFOS (median concentration ¼ 55.4 ng/L) were measured in the runoff, without any other PFAAs detected. The MWW test results indicate that the median PFOS concentration in stormwater runoff around the industrial area is significantly higher than

PFAAs on the solid particles in stormwater runoff

Urban runoff transports a wide gradation of anthropogenic aqueous complexes and particulate matter (PM) (Kim and Sansalone, 2008). No previous studies have yet been performed to determine the level of PFAAs on PM in stormwater runoff. When one considers the aqueous concentrations of PFAAs in the runoff (a few tens of ng/L), the total PM (TPM) of stormwater runoff (e.g., 100 mg/L), and the low solidewater partitioning coefficient (Kd) (several L/kg, Ahrens et al., 2011; Higgins and Luthy, 2006; Pan and You, 2010), no PFAAs are expected to be detected on the solid particles in the runoff. However, high-levels of PFOS were detected on the solid particles in stormwater runoff collected from the industrial area and the commercial area (Table 2). The TPM of the stormwater runoff at the two areas ranged from 22 to 137 mg/ L. Therefore, solidewater partitioning cannot be the mechanism responsible for the high PFOS levels on PM. Otherwise the Kd value of PFOS will reach 104.0 L/kg, which would be orders of magnitude higher than ever before reported in the literature (Higgins and Luthy, 2006; Pan and You, 2010) and our adsorption experimental results (Xiao et al., 2011). Therefore, the PFOS should be inherent in the particles/debris, possibly including the debris of textiles and carpet, and several industrial polymers containing PFOS. The debris from these products and solids could have been carried by stormwater into surface runoff. For the Dinkytown area, which also had substantial particulate PFOS concentrations, fine PM from vehicular traffic (Murakami and Takada, 2008) and the debris

Table 1 e Concentrations of PFAAs in stormwater runoff (ng/L) in previous studies and in the present study.

a

Japan Japanb Singaporec U.S.d U.S.e

PFHpA

PFOA

PFNA

PFOS

PFDA

PFUnDA

e 5.7e6.4 <0.7e10.5
17e174 22.5e28.4 7.2e38.2 0.51e29.3 3.5e30.6

4.7e70 84.9e92.8 0.9e78.3
2.8e50 38.7e44.1 4.5e139.0
1.8e77 1.5e5.0 0.7e28.2 n.d.e8.4 n.d.e10.6

n.a.e45 3.3e6.7 0.2e3.1 n.d.e2.0 n.d.e2.9

n.a.: not available; n.d.: not detected. a Murakami et al., 2009b b Zushi and Masunaga, 2009a c Nguyen et al., 2011 d Kim and Kannan, 2007 e The current study (including PFOS concentrations in runoff collected from the industrial area).

6606

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 6 0 1 e6 6 0 8

Table 2 e PFAAs in stormwater runoff around an industrial source and on the solid particles in stormwater runoff. PFHpA

PFOA

PFAAs in stormwater runoff (ng/L) collected at an industrial 36th and Brunswick n.d. 36th and Alabama n.d. 36th and Yosemite n.d. 36th and Kenwood n.d. 351/2 n.d. PFAAs on particles in stormwater runoff, ng/gTPM n.d. 36th and Brunswicka n.d. 36th and Kenwooda 351/2a n.d. n.d. Dinkytownb n.d. Dinkytownc

PFNA

PFOS

PFDA

PFUnDA

area (Sep 25 2009) n.d.
50.0 8.7 55.4 156.0 59.6

n.d. n.d.
n.d. n.d. n.d. n.d. n.d.

n.d. n.d. n.d. n.d. n.d.

280.4 120.5 590.0 19.8 45.9

n.d. n.d. n.d. n.d. n.d.

n.d. n.d. n.d. n.d. n.d.

n.d. n.d. n.d. n.d. n.d.

n.d.: not detected; LOD: detection limit; TPM: total particulate matter. a Sep 25 2009. b Sep 01 2010. c Sep 15 2010.

of commercial food packaging (such as microwave popcorn bags, fast food and candy wrappers, and pizza box liners, Weise, 2005) are suspected to contribute PFOS on the particles in stormwater runoff.

3.6.

Mass flow of PFAAs from stormwater runoff

The equation used by Murakami et al. (2009b) was modified for estimating the fluxes of background PFAAs from surface runoff in the MinneapoliseSt. Paul metropolitan area where the measurements were taken: M ¼ SRPFAA  SV ¼ SRPFAA  P  k  AI

(1)

where SRPFAA is the PFAA concentration in the surface runoff; SV is the surface runoff volume; k is a runoff coefficient that converts rainfall to runoff; P is the average annual precipitation (747 mm); and AI is the impervious area of the region (679.9 km2, see Table S2 in the Supplementary data). The value of the runoff coefficient, k, used for the calculation is 0.6 taking account of the connectivity of impervious area (see Table S3 in the Supplementary data). The results are as follows: PFHpA, 0.19  0.79; PFOS, 3.23  3.74; PFOA, 2.83  2.48; PFNA, 1.11  1.04; PFDA, 0.32  0.95; PFUnDA, 0  0.30 kg/year. The relatively large standard deviation reflects that the data of PFAA concentrations at different sampling sites are skewed (also see Section 3.2). The total flux of PFAAs was estimated as w7.86  9.30 kg/year. These values do not include the flux of PFAAs from industrial areas because they have not been quantified, and should not be seen as urban background concentrations. These values are smaller but close to the PFAA fluxes from street runoff in Japan estimated by Murakami et al. (2009b). The mass flux of PFOS in urban runoff is one fourth of the total PFOS flux at the confluence (Minnesota) to upper Mississippi river (Nakayama et al., 2010), indicating that urban runoff is an important source of PFOS to urban waters. As more evidence accumulates that urban areas can contribute PFAA pollution to surface water through surface runoff and storm sewers, the ability of SMPs to remove PFAAs requires investigation in the future. Many stormwater SMPs are being implemented in

metropolitan areas to remove PM and particle-associated pollutants from stormwater runoff. However, to date, no information is available on the removal of PFAAs by SMPs. The loads of PFAAs from rainfall were computed by multiplying the concentrations determined for rainfall, the average annual rainfall per year and the metropolitan area in the MinneapoliseSt. Paul region. The results are 1.89  0.43 (PFOA), 1.48  0.30 (PFOS), 0.29  0.65 (PFNA), 0.28  0.34 (PFHpA), 0.25  0.10 (PFDA), 0  0.08 (PFUnDA) kg/year. The total load of PFAAs was estimated as 4.17  1.92 kg/year. When multiplied by the runoff coefficient, k ¼ 0.6, rainfall can account for a PFAA load of 2.52  1.15 kg/year, which is about one third of PFAA load computed from runoff concentrations applied to Eq. (1) of 7.86 kg/year.

4.

Conclusions

PFAAs have been identified as a potential environmental problem within water systems. This paper contributes to previously very limited knowledge of the levels and fate of these chemicals in urban runoff. In spite of the announcement by a major US manufacturer to phase out the manufacture of PFOS-based fluorochemicals, we detected relatively high-level PFOS in stormwater runoff from industrial and commercial areas. Substantially higher runoff concentrations of PFOS (upto 156 ng/L) were found near a suspected industrial source, where PFOS was used in past decades. Statistical tests show that there were nonatmospheric sources of PFOS and PFOA in industrial and commercial areas. This conclusion is further supported by the detection of high-level PFOS on the particles/debris (19.8e590 ng PFOS/gTPM) in runoff collected from the two areas. Stormwater washes off PFAAs from unknown sources in these areas and transports them to urban waters. Excluding runoff from the industrial area, the median stormwater runoff concentration in the MinneapoliseSt. Paul, Minnesota, USA metropolitan area was estimated at 25.8 ng/L, which can be seen as a background runoff concentration in this metropolitan area. PFOS and PFOA are the most

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 6 0 1 e6 6 0 8

frequently detected PFAAs in stormwater runoff. The frequencies of detection of PFAAs in urban runoff were found to be PFOS (100%), PFOA (100%), PFNA (67%), PFDA (50%), PFHpA (44%), and PFUnDA (22%). The greater PFAA flux from stormwater runoff than from atmospheric inputs implies that PFAAs in the urban environment are solubilized in rain and transported. For vacant/ remote areas, the background runoff concentration can be mainly attributed to rainfall.

Acknowledgments We acknowledge funding for this research from the Minnesota Water Resources Center (2009 MN 253B). We thank Streets Summer of Minnesota Pollution Control Agency and Michael Perniel and Timothy Brown of Minneapolis Park & Recreation Board for collecting runoff samples for this study.

Appendix. Supplementary material Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.watres.2011.11.029.

references

Ahrens, L., Yeung, L.W.Y., Taniyasu, S., Lam, P.K.S., Yamashita, N., 2011. Partitioning of perfluorooctanoate (PFOA), perfluorooctane sulfonate (PFOS) and perfluorooctane sulfonamide (PFOSA) between water and sediment. Chemosphere 85 (5), 731e737. Baun, A., Eriksson, E., Ledin, A., Mikkelsen, P.S., 2006. A methodology for ranking and hazard identification of xenobiotic organic compounds in urban stormwater. Science of Total Environment 370 (1), 29e38. Bjo¨rklund, J.A., Thuresson, K., de Wit, C.A., 2009. Perfluoroalkyl compounds (PFCs) in indoor dust: concentrations, human exposure estimates, and sources. Environmental Science & Technology 43 (7), 2276e2281. Calafat, A.M., Wong, L., Kuklenyik, Z., Reidy, J.A., Needham, L.L., 2007. Polyfluoroalkyl chemicals in the U.S. population: data from the National Health and Nutrition Examination Survey (NHANES) 20032004 and comparisons with NHANES 19992000. Environmental Health Perspective 115 (11), 1596e1602. Chocat, B., Ashley, R., Marsalek, J., Matos, M.R., Rauch, W., Schilling, W., Urbonas, B., 2007. Toward the sustainable management of urban storm-water. Indoor and Built Environment 16 (3), 273e285. Davis, A.P., Shokouhian, M., Ni, S., 2001. Loading estimates of lead, copper, cadmium, and zinc in urban runoff from specific sources. Chemosphere 44 (5), 997e1009.  Ivetic, M., 1997. Modelling of storm Deletic, A., Maksimovic, C, wash-off of suspended solids from impervious surfaces. Journal of Hydraulic Research 35 (1), 99e118. Emmett, E.A., Shofer, F.S., Zhang, H., Freeman, D., Desal, C., Shaw, L.W., 2006. Community exposure to perfluorooctanoate: relationships between serum concentrations and exposure sources. Journal of Occupational and Environmental Medicine 48 (8), 759e770.

6607

Eriksson, E., Baun, A., Scholes, L., Ledin, A., Ahlman, S., Revitt, M., Noutsopoulos, C., Mikkelsen, P.S., 2007. Selected stormwater priority pollutants e a European perspective. Science of the Total Environment 383 (1e3), 41e51. Fei, C.Y., McLaughlin, J.K., Tarone, R.E., Olsen, J., 2007. Perfluorinated chemicals and fetal growth: a study within the Danish National Birth Cohort. Environmental Health Perspectives 115 (11), 1677e1682. Genc¸-Fuhrman, H., Mikkelsen, P.S., Ledin, A., 2007. Simultaneous removal of As, Cd, Cr, Cu, Ni and Zn from stormwater: experimental comparison of 11 different sorbents. Water Research 41 (3), 591e602. Giesy, J.P., Kannan, K., 2001. Global distribution of perfluorooctane sulfonate in wildlife. Environmental Science & Technology 35 (7), 1339e1342. Goosey, E., Harrad, S., 2011. Perfluoroalkyl compounds in dust from Asian, Australian, European, and North American homes and UK cars, classrooms, and offices. Environmental International 37 (1), 86e92. Gulliver, J.S., Erickson, A.J., Weiss, P.T., 2009. Stormwater Treatment: Assessment and Maintenance. St. Anthony Falls Laboratory, University of Minnesota, Minneapolis, MN. Ha, H., Stenstrom, M.K., 2003. Identification of land use with water quality data in stormwater using a neural network. Water Research 37 (17), 4222e4230. Higgins, C.P., Luthy, R.G., 2006. Sorption of perfluorinated surfactants on sediments. Environmental Science & Technology 40 (23), 7251e7256. Houde, H., de Silva, A.O., Muir, D.C.G., Letcher, R.J., 2011. Monitoring of perfluorinated compounds in aquatic biota: An updated review. Environmental Science & Technology 45 (19), 7962e7973. Jin, Y.H., Liu, W., Sato, I., Nakayama, S.F., Sasaki, K., Saito, N., Tsuda, S., 2009. PFOS and PFOA in environmental and tap water in China. Chemosphere 77 (5), 605e611. Kim, S., Kannan, K., 2007. Perfluorinated acids in air, rain, snow, surface runoff, and lakes: relative importance of pathways to contamination of urban lakes. Environmental Science & Technology 41 (24), 8328e8334. Kim, J., Sansalone, J.J., 2008. Zeta potential of clay-size particles in urban rainfallerunoff during hydrologic transport. Journal of Hydrology 356 (1e2), 163e173. Lau, C., Butenhoff, J.L., Rogers, J.M., 2004. The development toxicity of perfluoroalkyl acids and their derivatives. Toxicology and Applied Pharmacology 198 (2), 231e241. Loganathan, B.G., Sajwan, K.S., Sinclair, E., Senthil Kumar, K., Kannan, K., 2007. Perfluoroalkyl sulfonates and perfluorocarboxylates in two wastewater treatment facilities in Kentucky and Georgia. Water Research 41 (20), 4611e4620. Loi, E.I.H., Yeung, L.W.Y., Taniyasu, S., Lam, P.K.S., Kannan, K., Yamashita, N., 2011. Trophic magnification of poly- and perfluorinated compounds in a subtrophical food web. Environmental Science & Technology 45 (13), 5506e5513. Martin, J.W., Mabury, S.A., Solomon, K.R., Muir, D.C.G., 2003. Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus Mykiss). Environmental Toxicology and Chemistry 22 (1), 196e204. Metropolitan Council, 2011. Annual Population and Household Estimates (Twin Cities Metropolitan Area). Moody, C.A., Hebert, G.N., Strauss, S.H., Field, J.A., 2003. Occurrence and persistence of perfluorooctanesulfonate and other perfluorinated surfactants in groundwater at a firetraining area at Wurtsmith Air Force Base, Michigan, USA. Journal of Environmental Monitoring 5, 341e345. MPCA, 2008. PFCs in Minnesota’s Ambient Environment: 2008 Progress Report. Minnesota Pollution Control Agency. Murakami, M., Imamura, E., Shinohara, H., Kiri, K., Muramatsu, Y., Harada, A., Takada, H., 2008. Occurrence and

6608

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 6 0 1 e6 6 0 8

sources of perfluorinated surfactants in rivers in Japan. Environmental Science & Technology 42 (17), 6566e6572. Murakami, M., Takada, H., 2008. Perfluorinated surfactants (PFSs) in size-fractionated street dust in Tokyo. Chemosphere 73 (8), 1172e1177. Murakami, M., Kuroda, K., Sato, N., Fukushi, T., Takizawa, S., Takada, H., 2009a. Groundwater pollution by perfluorinated surfactants in Tokyo. Environmental Science & Technology 43 (10), 3480e3486. Murakami, M., Shinohara, H., Takada, H., 2009b. Evaluation of wastewater and street runoff as sources of perfluorinated surfactants (PFSs). Chemosphere 74 (4), 487e493. Nakayama, S.F., Strynar, M.J., Reiner, L.R., Delinsky, A.D., Lindstrom, A.B., 2010. Determination of perfluorinated compounds in the upper Mississippi river basin. Environmental Science & Technology 44 (11), 4103e4109. Nguyen, V.T., Reinhard, M., Karina, G.Y., 2011. Occurrence and source characterization of perfluorochemicals in an urban watershed. Chemosphere 82 (9), 1277e1285. Olsen, G.W., Burris, J.M., Ehresman, D.J., Froehlich, J.W., Seacat, A.M., Butenhoff, J.L., Zobel, L.R., 2007. Half-life of serum elimination of perfluorooctanesulfonate, perfluorohexanesulfonate, and perfluorooctanoate in retired fluorochemical production workers. Environmental Health Perspective 115 (9), 1298e1305. Pagano, M., Gauvreau, K., 2000. Principles of Biostatistics, second ed. Duxbury Press, Pacific Grove, CA. Pan, G., You, C., 2010. Sediment-water distribution of perfluorooctane sulfonate (PFOS) in the Yangtze River estuary. Environmental Pollution 158 (5), 1363e1367. Prevedouros, K., Cousins, I.T., Buck, R.C., Korzeniowski, S.H., 2006. Sources, fate and transport of perfluorocarboxylates. Environmental Science & Technology 40 (1), 32e44. Schultz, M.M., Higgins, C.P., Huset, C.A., Luthy, R.G., Barofsky, D.F., Field, J.A., 2006. Fluorochemical mass flows in a municipal wastewater treatment facility. Environmental Science & Technology 40 (23), 7350e7357. Simcik, M.F., Dorweiler, K.J., 2005. Ratio of perfluorochemical concentrations as a tracer of atmospheric deposition to surface waters. Environmental Science & Technology 39 (22), 8678e8683.

Takagi, S., Adachi, F., Miyano, K., Koizumi, Y., Tanaka, H., Mimura, M., Watanabe, I., Tanabe, S., Kannan, K., 2008. Perfluorooctanesulfonate and perfluorooctanoate in raw and treated water from Osaka, Japan. Chemosphere 72 (10), 1409e1412. Tang, C.Y., Fu, Q.S., Gao, D., Criddle, C.S., Leckie, J.O., 2010. Effect of solution chemistry on the adsorption of perfluorooctane sulfonate onto mineral surfaces. Water Research 44 (8), 2654e2662. Weise, E., 2005. Engineer: DuPont Hid Facts about Paper Coating. http://www.usatoday.com/money/companies/management/ 2005-11-16-dupont-usat_x.htm. Weiss, J.M., Andersson, P.L., Lamoree, M.H., Leonards, P.E.G., van Leeuwen, S.P.J., Hamers, T., 2009. Competitive binding of polyand perfluorinated compounds to the thyroid hormone transport protein transthyretin. Toxicological Sciences 109 (2), 206e216. Weiss, P.T., Erickson, A.J., Gulliver, J.S., 2007. Cost and pollutant removal of storm-water treatment practices. Journal of Water Resources Planning and ManagementeASCE 133 (3), 218e229. Xiao, F., Zhang, X., Lee Penn, R., Gulliver, J.S., Simcik, M.F., 2011. Effects of monovalent cations on the competitive adsorption of perfluoroalkyl acids on kaolinite: experimental studies and modeling. Environmental Science & Technology Doi:10.1021/es202524y. Yu, J., Hu, J., Tanaka, S., Fujii, S., 2009. Perfluorooctanoic acid (PFOA) in sewage treatment plants. Water Research 43 (9), 2399e2408. Zushi, Y., Masunaga, S., 2009a. First-flush loads of perfluorinated compounds in stormwater runoff from Hayabuchi River basin, Japan served by separated sewerage system. Chemosphere 76 (6), 833e840. Zushi, Y., Masunaga, S., 2009b. Identifying the nonpoint source of perfluorinated compounds using a geographic information system based approach. Environmental Toxicology and Chemistry 28 (4), 691e700. Zushi, Y., Takeda, T., Masunaga, S., 2008. Existence of nonpoint source of perfluorinated compounds and their loads in the Tsurumi River basin, Japan. Chemosphere 71 (8), 1566e1573.