SO-MUM: A Coupled Atmospheric Transport and Multimedia Model

Article pubs.acs.org/est

SO-MUM: A Coupled Atmospheric Transport and Multimedia Model Used to Predict Intraurban-Scale PCB and PBDE Emissions and Fate Susan A. Csiszar,† Sreerama M. Daggupaty,‡ Stephanie Verkoeyen,§ Amanda Giang,∥ and Miriam L. Diamond*,†,⊥ †

Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, Ontario, Canada Air Quality Research Division, Environment Canada, Toronto, Ontario, Canada § Department of Geography, University of Toronto, Toronto, Ontario, Canada ∥ Centre for Global Change Science, University of Toronto, Toronto, Ontario, Canada ⊥ Department of Earth Sciences, University of Toronto, Toronto, Ontario, Canada ‡

S Supporting Information *

ABSTRACT: A spatially resolved, dynamic version of the Multimedia Urban Model (MUM) and the boundary layer forecast and air pollution transport model BLFMAPS were coupled to build Spatially Oriented MUM (SO-MUM), to estimate emissions and fate of POPs in an urban area on a 5 × 5 km2 cell resolution. SO-MUM was used to back-calculate emissions from spatially resolved measured air concentrations of PCBs and PBDEs in Toronto, Canada. Estimated emissions of Σ88PCBs were 230 (40−480) kg y−1, 280 (50−580) g y−1 km−2, or 90 (16−190) mg y−1 capita−1, and Σ26PBDEs were 28 (6−63) kg y−1, 34 (7−77) g y−1 km−2, or 11 (2−25) mg y−1 capita−1. A mass inventory of penta- and octa-BDEs in Toronto was estimated to be 200 tonnes (90−1000 tonnes) or 80 (40−400) g capita−1. Using this estimate and that of 440 (280−800) tonnes of PCBs, estimated emissions of Σ88PCBs and Σ26PBDEs per mass of chemical inventory in Toronto were 0.5 (0.05−1.6) and 0.1 (0.01−0.7) g y−1 kg−1, respectively. The results suggest annual emission rates of 0.04% and 0.01% from the mass inventories with downtown accounting for 30% and 16% of Toronto’s chemical inventory and emissions of PCBs and PBDEs, respectively. Since total PBDE emissions are a function of mass inventory, which is proportional to building volume, we conclude that building volume can be used as a proxy to predict emissions. Per mass inventory emission rates were negatively related to vapor pressure within a compound class, but not consistently when considering all compound congeners.

1. INTRODUCTION

Several studies have developed aggregate urban emission estimates for PCBs12−15 and PBDES9,16 at specific locations, expressed on an areal and/or per capita basis. This approach has typically relied on inverse mass balance modeling and air measurements to derive their estimates. While this method is arguably less data intensive than using emission factors, it has not been specifically connected to chemical mass inventories or activities and, as such, is less useful for focusing source reduction efforts than using activity-specific emission factors. This approach follows the suggestion by Breivik et al.7 that combining inverse modeling and mass inventories could be used to connect emissions and usage. Elevated urban air concentrations of PCBs persist despite bans on new production passed some 40 years ago in the late 1970s in many western countries, and global agreements to ban production, new uses, and promotion of efforts to eliminate

As laid out by the Stockholm Convention on Persistent Organic Pollutants (POPs) and the Convention on LongRange Transboundary Air Pollution (UNECE CLRTAP), a necessary step toward controlling POPs in the environment is identifying emission sources.1,2 Emission sources have been calculated using two main methods. The first method uses activity-specific emission factors and the second uses an aggregate estimate such as emissions on a per capita basis. The advantage of developing emission estimates based on activity-specific emission factors is the ability to clearly identify the activity or sector contributing most to emissions. Perhaps the best examples of emission estimates that use emission factors are those for polychlorinated dibenzodioxins and furans (PCDD/F), polycyclic aromatic hydrocarbons (PAH), and mercury.3−5 Extensive measurement programs have produced series of well vetted emission factors, e.g., mass of PAH emitted per tonne of municipal solid waste incinerated or wood burned.3,6 Emission factors for other POPs such as PCBs and PBDEs are less well studied and more uncertain.7−11 © 2012 American Chemical Society

Received: Revised: Accepted: Published: 436

August 14, 2012 November 20, 2012 November 28, 2012 November 28, 2012 dx.doi.org/10.1021/es3033023 | Environ. Sci. Technol. 2013, 47, 436−445

Environmental Science & Technology

Article

stocks.12,17 PBDEs have been controlled recently with new use bans of penta and octa mixtures by 2004 in Europe and 2008 in Canada.18 Salamova and Hites19 estimated that PBDE air concentrations are decreasing in Great Lakes air with a halving time of 5−7 years. Several lines of evidence suggest that primary, rather than secondary emissions dominate PCB air concentrations, at least in urban and industrial areas where concentrations are highest.20−22 Primary PCB emissions can originate from open sources such as building sealants, light ballasts, paint and floor coatings, and inadvertent releases from closed sources.7,12 For PBDEs, elevated air concentrations have been related to releases from indoors that have PBDE-containing consumer products such as electronics, furniture, textiles, and plastics.23−27 Furthermore, Melymuk et al.28 found that outdoor PBDE air concentrations can be predicted using building volumes and dwelling counts that serve as surrogates for the mass inventory of PBDE-containing products. Diamond et al.12 reported an in-use and in-storage PCB stock of 440 (280−800) tonnes in the built environment of Toronto, Canada and used the one-box, steady-state (SS) version of the Multimedia Urban Model (MUM), to estimate emissions of 90−880 kg y−1 originating from the city; the emissions were inversely modeled from measured outdoor air concentrations at only a few sites. Other efforts to estimate PCB and PBDE emissions using inverse modeling have similarly used single box models and one or a few urban sites.9,13−15 The weakness of the one-box model is that spatial heterogeneity cannot be captured14,29 nor can emissions be linked to a spatially variant chemical inventory. Multibox multimedia models have been applied on a global scale (e.g., Globo-POP31 and CliMoChem32) and have been coupled to atmospheric transport models to assess long-range transport.33 These strategies have also been used on regional scales, for example BETR in North America and Europe.34,35 However, one box models have been used to study POPs on an urban scale.13,36 Multibox fugacity models on smaller scales, such as 5 km and 0.5 km, have been used to study Lindane in the Great Lakes Region;37 dioxins in Japan without a detailed air transport model;38 and volatile organic compounds in Seoul, South Korea under steady-state conditions.39 Morselli et al.15,40 have coupled a one-box multimedia fugacity model to a meteorological model to study urban PCBs and PAHs. We have extended the one-box version of MUM into a multibox, unsteady-state (USS) version and coupled it to BLFMAPS (Boundary Layer Forecast Model and Air Pollution Prediction System),30 which is a chemical air transport model. We call the coupled model Spatially-Oriented MUM or SOMUM. SO-MUM is unique in that incorporates a detailed meteorological transport model into an USS multibox fugacity model developed explicitly for an urban area by including films that cover impervious surfaces which are important for urban POP fate36,41 and will be further discussed by Csiszar et al.42 We illustrate the use of SO-MUM with PCBs and penta- and octa-BDEs, using measured outdoor air concentrations from Melymuk et al.43 to back-calculate air emissions.12,13,15,16 The goal of this application is to estimate spatially resolved, aggregate (not source-specific) emissions to air of these two compound groups, expressed on a city-wide, areal, mass inventory, and per capita basis for the City of Toronto. The advantage of this method is the ability to estimate chemical emissions per unit of mass inventory at an intraurban scale and to account for spatially variable fate and concentrations. This

paper presents the detailed framework of SO-MUM, its assessment, and resulting emission estimates. In a future paper42 we use SO-MUM to explore the relationship between primary versus secondary emissions, intraurban fate, city−lake effects, and chemical export. These lake effects, as well as a detailed discussion of emission scenarios and fate, are presented in this future publication.42

2. MATERIALS AND METHODS 2.1. Model Descriptions. MUM is a single-box 7compartment (air, upper air, water, soil, sediment, vegetation, and film) fugacity mass balance model.12,44 Csiszar et al.41 extended MUM into an unsteady-state version which takes into account the dynamic nature of films that develop on impervious surfaces. The general unsteady-state mass balance equation for each compartment is d(ZiVfi i ) dmi = dt dt = Ei +

∑ (Djif j ) − DTifi , i = a , ua , w , s , se , v , f i≠j

(1) 3

−3

−1

where mi (mol), Vi (m ), Zi (mol m Pa ), f i (Pa), Ei (mol h−1), and DTi (mol h−1 Pa−1) are the mass, volume, bulk Zvalue, fugacity, specified emission source or sink, and total loss D-value (transport and transformation), respectively, for compartment i; and Dji (mol h−1 Pa−1) is the intercompartmental transfer D-value of chemicals from compartment j to i (see Csiszar et al.41 and the Supporting Information (SI) for further details regarding model equations). Equation 1 was solved simultaneously for each compartment using the Implicit Euler approximation as follows: Zin + 1Vin + 1f in + 1 − ZinVinf in Δt =

Ein + 1

+

∑ (Djin+ 1f jn+ 1 ) − DTin+ 1f in+ 1 i≠j

(2)

with tn = nΔt, where Δt is the time step, which was set to 5 min to match the time step of BLFMAPS. We used the convention whereby tn represents the current time and tn+1 represents one time step into the future. To couple the model to BLFMAPS, the upper air compartment of MUM was removed and, accordingly, the number of mass balance equations was reduced from seven to six. BLFMAPS30 is a modified version of the Boundary Layer Forecast Model (BLFM)45 which is a meteorological prediction model. BLFM was extended from its first version45 to form the air pollution prediction system, BLFMAPS, by including atmospheric transport, diffusion, and deposition for atmospheric pollutants such as particulate metals and trace substances.30 Ma et al.37 used BLFMAPS to study Lindane transport from the Canadian Prairies in order to assess the effects of Lindane on the Great Lakes. As a mesoscale atmospheric model used to simulate boundary layer flow over heterogeneous terrain, BLFM predicts three-dimensional land and water induced air circulation, temperature, and turbulence parameters in the model domain that includes a 5 × 5 km2 horizontal grid resolution over Southern Ontario. In the vertical, it has nonuniformly distributed levels with top boundaries at 1.5, 3.9, 10, 100, 350, 700, 1200, and 2000 m above ground with the top level at 3000 m above ground. The model considers 37 land use 437

dx.doi.org/10.1021/es3033023 | Environ. Sci. Technol. 2013, 47, 436−445


Article

Figure 1. (a) SO-MUM model domain and (b) zoomed in view of City of Toronto with labeled sample sites. The Toronto cells are identified based on coordinates as indicated in (b).

categories from the United States Geological Survey.46 BLFM uses twice daily objectively analyzed weather data from the Global Environmental Multiscale Model (GEM) of Environment Canada.47−50 The meteorological data are updated and initialized on every 12th hour from which BLFM predicts wind velocities, temperatures, and other meteorological variables as a function of the three-dimensional space of the model domain for the next 12 h period, with a time integration step of 5 min. In BLFMAPS chemical transport is modeled by the following equation: ∂Ca ∂C d ⎛ ∂Ca ⎞ = −v ⃗∇Ca − w a + kh∇2 Ca + ⎟ ⎜k z * ∂z dz ⎝ ∂z ⎠ ∂t * * * + (Fa − Sa) − ηCa

of the cells in BLFMAPS. The upper air compartment in MUM was replaced by BLFMAPS and intercell air transport was modeled by BLFMAPS. Equation 3 was modified so that BLFMAPS modeled bulk air advection and diffusion in the horizontal and vertical directions, as well as chemical input via emission, and we removed deposition and reaction since these processes are treated by MUM, as follows: ∂Ca ∂C d ⎛ ∂Ca ⎞ = −v ⃗∇Ca − w a + kh∇2 Ca + ⎟ + Fa ⎜k z * dz ⎝ ∂z ⎠ ∂t ∂z * * *

(4)

All other air processes, such as chemical deposition (gas, particle-associated, and wet) and reactive loss, were modeled in MUM, in addition to processes concerning the other urban compartments (water, soil, sediment, vegetation, and film) as described at the beginning of Section 2.1. MUM was also extended to include tributary transport between cells in Toronto. The chemical input, Ir,i and output, Or,i terms via rivers for a given cell, i, follow the standard fugacity mass balance formulation of

(3)

where Ca is the air concentration; t is time; w* is the vertical velocity in terrain following coordinates; z* = z − h(x,y) (where z is height and h(x,y) is surface elevation); kh and kz are the horizontal and vertical eddy diffusivities, respectively; Fa (pg m−3 s−1) and Sa (pg m−3 s−1) are concentration fluxes due to emission and deposition, respectively; and η is the reaction rate of the chemical in air. Equation 3 is solved with a 5-min timestep using a finite-difference approximation.30,37 BLFM and BLFMAPS have both been previously described and assessed in the literature.30,37,45,46 2.2. Coupled Model: SO-MUM. To couple the models, the one-box version of MUM was extended into a multibox model with cells corresponding to the size (5 × 5 km2) and locations

Ir , i = DA , jfr , j and Or , i = DA , ifr , i

(5)

where DA,i is the water advection D-value (mol h−1 Pa) for cell i, f r,i (Pa) is the water fugacity in cell i, and the subscript j refers to the cell directly upstream of cell i. Other minor tributaries in the city were not connected cell-to-cell and were given constant water outflow rates as a loss pathway in order to prevent 438



Article

downtown cell (5,2), referred to as DT, we took the average of the 0KM and E1 sites which fell in cells adjacent to the DT cell as air concentrations at site S5 in cell DT were representative of those over Lake Ontario rather than the downtown core (Figure 1b).43 For cell DT we assumed an inflow air concentration equal to one-half in the lower air compartment of that measured concentration because of the higher concentrations measured in downtown (see Melymuk et al.43) relative to the rest of the city. Step 2: Normalize back-calculated emissions per unit mass inventory. The mass of chemical stock in each cell was estimated for PCBs using the 2006 mass inventories presented by Diamond et al.12 and Robson et al.;53 the PBDE mass inventory is described in Section 2.4. For the 9 cells with backcalculated emissions we calculated an emission per unit mass inventory, EI (g h−1 kg−1) and then averaged the 9 values

anomalous chemical build-up in the tributary compartment. Thus, water outflow from these cells represented a loss route rather than a transport route to another cell or compartment. MUM and BLFMAPS were loosely coupled into SO-MUM by moving them forward separately for a given time step as follows: BLFMAPS receives a chemical emission into air and MUM is then moved forward by Δt, the resulting air concentration is inputted into BLFMAPS which is then moved forward by Δt. This process is repeated for the entire simulation period. To pass concentrations between the two models a volume-weighted average of the air concentrations from the first 4 BLFMAPS air layers (i.e., up to 100 m) were used as input into the single, vertical air compartment of MUM which was set to a 100 m height. We chose a vertical air height of 100 m in MUM to align with that of BLFM and since it was within the average boundary layer height as predicted by BLFM (∼250 m). Thus, we assumed a mixed 100 m air compartment rather than varying air concentration with height within the bottom air layer; this errs on the side of larger advective losses as horizontal wind speeds increase with height and air concentrations generally decrease with height.52 The resulting MUM-calculated air concentrations were then used as input to the first 4 layers of BLFMAPS such that these levels had the same air concentrations at the beginning of each time step; chemicals in these layers were subsequently mixed vertically during BLFMAPS calculations. BLFMAPS was written in Fortran, and for ease of compatibility, MUM was also written in Fortran and added as a subroutine within the framework of BLFMAPS. The model domain of SO-MUM contained 494 cells making up an area of 12350 km2 (19 × 26 cells or 95 km ×130 km) and was centered around Toronto which was made up of 39 cells with a land area of 820 km2 (Figure 1). Details of the model parametrization including physical−chemical properties as well as environmental data can be found in the SI. 2.3. Estimating Emissions. Chemical emissions for cells within Toronto were estimated using a two-step process where, first, emissions to air were back-calculated using the original steady-state 7-compartment version of MUM (described in Section 2.1) and measured air concentrations (see Diamond et al.12 for this method). Second, those rates were scaled to estimated chemical mass inventories. We assumed one bulk emission into the lower air compartment of each cell (100 m height). Air concentrations were those measured by Melymuk et al.43 using PUF (polyurethane foam) passive air samplers (PAS) deployed in 2008 for 3-month periods throughout the city. These concentrations were treated as bulk-phase (rather than gas-phase) consistent with the finding of Melymuk et al.51 See Melymuk et al.43 for details of the measurement campaign. Step 1: Back-calculate emissions. Air measurements were available for 9 cells (corresponding to sites W10, W5, W1/ 0KM/N1, E1/E5, E10, E20, N5, and N10, Figure 1b) within the city. Chemical emission rates from within these 9 cells were back-calculated using steady-state MUM assuming chemical inputs from air advected from adjacent cells at a concentration equal to that in the said cell in the lower air compartment, consistent with the findings of Melymuk et al.,43 and onequarter that concentration in the upper air compartment, as air concentrations have been found to be lower at higher air levels52 and BLFMAPS predicted a reduction in concentration of approximately one-quarter between the 100 and 350 m air levels. For cells with more than one air sampling site (e.g., sites E1 and E5 in cell (5,3)), we averaged measured values. For the

EI =

1 n

n

∑ j=1

Ej mj

(6)

where Ej is the back-calculated emission and mj is the chemical mass in cell j, where j refers to one of the n cells. Next, the process was extended to the 30 cells lacking measured air concentrations. For all cells, we multiplied EI by the mass inventory in each cell to estimate the cell-specific chemical emission rate. Thus, emissions for all cells were set as the product of EI, as determined in eq 6, and the chemical inventory in that cell. As SO-MUM is time-dependent, we initialized the compartmental (water, soil, sediment, vegetation, and film) concentrations in each cell by setting the values to those calculated by the steady-state (SS) version of MUM (with further details in SI). Initial SO-MUM air concentrations were set to zero in all cells. SO-MUM had a run-up time of 2 weeks before recording data. 2.4. Mass Inventories. Diamond et al.12 and Robson et 53 al. reported a spatially resolved PCB stock of 440 (280−800) tonnes in Toronto which included PCBs in-use, in-storage (both recorded as point locations), and contained within building sealants (on a per city ward basis, with an areal range of 47−390 km2). This inventory was used in the emission calculations by assigning the data to each SO-MUM cell. The spatial mass inventory for penta- and octa-BDE mixtures was developed for Toronto by first estimating the amount of penta and octa containing products including computers, printers, televisions, furniture (chairs, sofas, beds, and pillows), and cars, and second, converting to the mass of penta and octa based on concentrations in these products (we excluded deca as we did not have measured air concentrations for BDE-209). We used PBDE product concentrations from Morf et al.54,55 for computers, printers, and televisions, Allen et al.24 for furniture, and Morf et al.54 and Imm et al.56 for cars. Further details of the inventory calculations can be found in the SI and Tables S9− S10. 2.5. Model Runs. SO-MUM was run over spring (April− June) 2008 for 5 PCB congeners (28, 52, 101, 153, 180) and 5 PBDE congeners (28, 47, 100, 154, 183) which range in physical−chemical properties and generally had the highest measured concentrations within their respective homologue groups.43 We ran various sensitivity analyses to assess the effects of wind speeds and dry deposition velocity on model outcomes. Details and results are presented in the SI. 439



Article

3. RESULTS AND DISCUSSION 3.1. Model Assessment. As with most environmental models, SO-MUM cannot be independently evaluated as we used measured PCB and PBDE air concentrations43 to initialize the model and to calculate emissions. However, we can assess the ability of the model to reproduce those measured air concentrations (at a height of 10 m which is close to the deployment height of the PAS) by comparing measured and modeled concentrations in 16 cells (9 within and 7 outside the city) averaged over the 3 month simulation period. It should be noted that the model cell air volumes were larger than those captured by the PAS, however, concentrations within a cell could be expected to be within 5 times of measured concentrations, the factor by which concentrations varied across the city for both chemical classes. We also compared the modeled concentrations in the DT cell to urban concentrations reported for Toronto in previous years.57−60 Means and ranges in modeled concentrations were obtained (Figure S1) using ranges in estimated emissions; we used the mean concentrations in our assessment of the model. The means and ranges of emissions were derived from the mean emission rates per mass chemical inventory, EI (eq 6), plus/minus the standard deviation of the EIs (or set to the minimum EI when the standard deviation was larger than the mean, Table S16) calculated for the 9 cells. Finally, we compared measured and modeled concentrations of PCBs and PBDEs in surfaces films, precipitation, and tributaries. Modeled Σ5PCBs air concentrations were within an order of magnitude of measured values with a best fit slope (BFS) of modeled vs measured values of 0.6 (r2 = 0.7) (Figure S1a,c). The regression excluded the outlier of site W20 which had unexplained, high measured concentrations of only tri- and tetra-PCBs but not higher chlorinated congeners.43 The measured (average of sites 0KM and E1, Figure 1b) and modeled Σ5PCBs concentrations in the DT cell were 150 and 86 pg m−3, respectively, and were within an order of magnitude but lower than previously measured values in downtown Toronto of 350 pg m−3 (Σ5PCBs) in spring 2001,60 210 pg m−3 (Σ5PCBs) annual 2000−2001,57 and 360 pg m−3 (Σ10PCBs) in spring 200258 (Table S15) which, hopefully, reflects a diminishing PCB stock over time.12 In Canada, the federally estimated in-use PCB askarel and mineral oil mass inventory consistently decreased annually between 1998 and 2005 (with a decrease of ∼30% between 1998 and 2005).61 Modeled Σ5PBDE air concentrations were within an order of magnitude of measured values (except for site W60 outside the city, Figure S1b,d) with a BFS of measured vs modeled concentrations of 0.5 (r2 = 0.1) where the low r2 indicates that concentrations were not consistently over- or underestimated. The BFS rose to 0.7 (r2 = 0.4) when DT was excluded since the model underestimated the air concentration here by 3 times. The measured (average of sites 0KM and E1, Figure 1b) and modeled concentrations of BDE-47 in the DT cell were 50 and 14 pg m−3 which were within an order of magnitude of previously measured values in downtown Toronto of 15 pg m−3 in spring/summer 2001,59 22 pg m−3 annual 2000−2001,57 and 55 pg m−3 in spring 200258 (Table S15). The model reproduced the urban−suburban gradient in PCB concentrations in Toronto of 12-fold (7−86 pg m−3) between DT and a suburban cell, (1,3). Similar urban−rural gradients were reported for Toronto, Chicago, Philadelphia, and Cleveland.43,62−64 Modeled PBDE concentrations reproduced

the urban−suburban gradient of 4-fold (cells (1,3) and (5,3)), respectively, or 7−28 pg m−3 which is similar to that reported by Harner et al.59 and Melymuk et al.43 for Toronto. In general, DT air concentrations of PCBs and PBDEs were underestimated compared to measured values. Reasons for this discrepancy include an error in the model, that measured values were not representative of the 5 × 5 km2 area of downtown which could be due to fine scale spatial heterogeneity (e.g., 1 km),43 an underestimate of the mass inventories in DT, or that the PCB and PBDE emission rates per mass inventory in DT were higher than in other cells. As downtown has the highest geographic density of PCBs and PBDEs, an air sampling location could have a higher probability of being nearby a high emission source compared to a suburban or rural sampling location. Higher emission rates in DT could be related to higher ventilation rates of office buildings than residential structures.27 Air concentrations of both compound classes were also underestimated at sites outside of the city, especially to the west of the city which is upwind of Toronto since the prevailing winds are from the southwest during spring (maximum underestimate was in cell W60 by 12 times and 30 times for PCBs and PBDEs, respectively). The upwind underestimation is consistent with our neglect of upwind emission sources from the neighboring cities of Hamilton and Burlington at the western end of Lake Ontario. Alternatively, the model is not adequately accounting for chemical transport. To explore these possibilities, we back-calculated PCB emissions based on the average measured concentrations at sites outside of the city (W40, W60, N40, and N80) and ran the model with these emissions originating from each cell outside the city. Adding these emissions beyond the border of Toronto did not significantly change modeled concentrations within Toronto (background vs without background had a BFS of 1.0, r2 = 1.0), however concentrations in cells outside of Toronto increased by 1.6 (W40) to 6 (N80) times. As we are focusing on urban emissions, we present results assuming no emissions from cells outside Toronto. Modeled concentrations were also compared to air measurements from a high-volume air sampler located in the DT cell in order to assess dynamic model estimates. Melymuk et al.51 took these measurements at the 0KM site at 12 day intervals for a 24 h sampling period (9 a.m. to 9 a.m.) during the same time period that passive air samplers were deployed. Hourly modeled concentrations were averaged over the same 24 h period as the high volume sampler data for ease of comparison. Measured and modeled concentrations were within an order of magnitude agreement where the model underestimated concentrations on average by 3 times for PCBs, and overestimated by 2 times and underestimated by 4 times for PBDEs, before and after approximately June first, respectively (Figure S2). The model reproduced measured temporal trends, with concentrations generally increasing starting in June. We compared modeled and measured concentrations in media other than air including surface films (Table S11), precipitation (Table S13), and tributaries (Table S12).65−69 For both ∑5PCBs and ∑5PBDEs, modeled concentrations in surface films, tributaries, and precipitation were all within an order of magnitude of measured concentration ranges (Tables S11−13). Further details on the comparison in these media can be found in SI (Section S2.1). 3.2. PBDE Mass Inventory. The mass of PBDEs (penta and octa mixtures) including vehicles in Toronto was estimated 440



Article

to be 200 tonnes (90−1000 tonnes) or 80 (40−400) g capita−1 with penta and octa making up 60% and 40% of the mass, respectively (Figure S3, excluding cars). The PCB mass inventory in 2006 was estimated to be 180 (110−320) g capita−1 12 which reflects the higher measured outdoor air concentrations of PCBs in comparison to PBDEs in Toronto.43 The mass inventories per capita, however, do not reflect the higher estimated body burdens of PBDEs in comparison to PCBs in North America70 which may be due to differences in relative exposure contributions of dust versus diet.71 The large range in the PBDE stock estimate was due to the range in concentrations of penta in furniture and vehicle textiles24,56 (Table S10). The higher mass of penta agrees with higher market demand for penta than octa in North America in 2005,18 however, our mass inventory likely underestimates the true values for both mixtures as we did not account for all sources (e.g., carpets and other furniture types; and electronics such as entertainment systems and appliances24,55). 3.3. Emission Rate Estimates. Spatially Resolved Estimates. Total means and ranges of Σ5PCB and Σ5PBDE emissions from Toronto were 19 (2−41) kg y−1 and 18 (3−42) kg y−1, respectively. A clear spatial trend in values of EI was not apparent. The low ends of the emission ranges seemed to be unreasonable as these values resulted in underestimated modeled air concentrations by 20 and 8 times on average for PCBs and PBDEs, respectively, assuming a linear relationship between emissions and concentrations. However, the annual means were likely biased high since they were based on measured spring air concentrations which were similar to summer values and 3−4 and 5 times higher on average for PCBs and PBDEs, respectively, than those measured in fall and winter.43 Scaling up from Σ5PCBs to Σ88PCBs using a Σ88PCB profile of air concentrations43 yielded a city-wide emission rate of 230 (40−480) kg y−1, or 280 (50−580) g y−1 km−2, or 90 (16− 190) mg y−1 capita−1. The city-wide estimate was on the same order of magnitude as modeled estimates for New York City (300 kg y−1 or 400 g y−1 km−2, Σ93PCBs)72 and Canada (8− 300 mg y−1 capita−1, based on the default and maximum values of Breivik et al.7). The multibox emission estimate was within the range of the one-box MUM estimate for Toronto of 90− 880 kg y−1 (Σ70PCBs).12 Scaling Σ5PBDEs emissions to Σ26PBDEs (using the same method as for PCBs) yielded a city-wide emission rate of 28 (6−63) kg y−1, 34 (7−77) g y−1 km−2, and 11 (2−25) mg y−1 capita−1. Jones-Otazo et al.16 estimated Σ12PBDE emissions in Toronto of 37−150 kg y−1 using the one-box version of MUM, compared to 28 (5−60) kg y−1 using SO-MUM for the same 12 congeners. The lower values estimated here could have been due to the 7−8 year span between the years modeled (e.g., air concentrations used by Jones-Otazo et al.16 were 45−110 pg m−3 compared to average measured urban concentration of 37 pg m−3 used here) or the difference in spatial resolution. We explore this below. Moeckel et al.,9 also using a multimedia model (without spatial resolution), estimated Σ3PBDE (28, 47, 100) emissions from Zurich, Switzerland of 19 mg y−1 capita−1 which was triple that of our mean estimate for these congeners (6 mg y−1 capita−1). Our estimated range captured the estimate of Batterman et al.23 of 5.5 mg y−1 capita−1 for Σ3PBDE (28, 47, 100) for U.S. household emissions. The estimated Σ88PCB and Σ26PBDE emission rates per mass inventory were 0.5 g y−1 kg−1 (0.1−1 g y−1 kg−1) and 0.1 g y−1 kg−1 (0.03−0.3 g y−1 kg−1), respectively, which equals an

emission of 0.05% (0.01−0.1%) and 0.01% (0.006−0.07%) of the inventory mass per year, respectively (Table S16). The ranges in emission rates per mass inventory represent the ranges in estimated emissions divided by the “best estimate” mass inventories. Using the ranges in mass inventories and emission rates results in Σ88PCB and Σ26PBDE emission rates per mass inventory of 0.05−1.6 and 0.01−0.7 g y−1 kg−1, respectively. Since PBDE emissions are proportional to the mass inventory which, in turn, is a function of building volume, we conclude that building volume is a reasonable proxy for emissions. The emission rates per inventory were surprisingly similar for the two chemical classes despite the dissimilarity in their usage profiles and time spans of use. These rates imply that if the mass of each compound class is not removed from the city it will take thousands of years for the stock to be released, assuming a constant emission rate over time. We calculated congener-specific values of EI or EIi using the congener-specific emissions per total mass inventory normalized to the composition of Aroclor (weighted average of 1242, 1254, 1260) and DE-7173−75 (Table S17). When expressed on a percentage basis, the annual value of EIi for BDE-47 of 0.03 (0.003−0.2)% was within the range predicted by Alcock et al.11 of 0.01−0.7%. Within a chemical class, EIi generally increased with volatility or decreased with increasing Koa (Figure 2),

Figure 2. Congener specific emission rates per mass inventory, EIi (g y−1 kg−1) for (a) PCBs and (b) PBDEs.

supporting the assumption that emissions can be related to Koa made by Breivik et al.7 and Prevedouros et al.10 Although values of EIi for PCBs and PBDEs were very similar, they did not follow a consistent trend according to Koa when considered together. This dissimilar pattern may be due to uncertainties in model estimates and/or the use of PBDEs indoors in comparison to PCBs which have more outdoor uses in capacitors and transformers, building sealants, and paints.23,27,43 We anticipate that as the mass of the more volatile chemicals is depleted at a faster rate (i.e., they have a higher emission rate per mass chemical), the congener-specific EIs will gradually lose their relationship to Koa over time.11,53 This trend can be seen in Figure 2 where the PCB EIs differ by 2 times (between CB28 and 180) in comparison to PBDEs which differ 20 times 441



Article

Figure 3. Estimated emission rates (kg y−1) in Toronto cells of (a) Σ5PCBs and (b) Σ5PBDEs.

spatial resolution in SO-MUM provided an improved estimate of emissions that would merit the additional complexity of this model, we estimated emissions using the one-box steady-state version of MUM (with geographical properties averaged across the city) using the 2008 data set of Melymuk et al.43 which we used to estimate the spatially resolved emissions. We calculated emissions in two ways, one in which we averaged the air concentrations in the 9 cells used to calculate the emission rates per mass inventory, and the second in which we averaged the 0KM and N10 sampling sites to represent high urban concentrations, following the method of Diamond et al.12 A range in emission rates was obtained by setting congenerspecific inflow air concentrations to the average of the sampling sites outside of the city for a “high” emission scenario and to the same as the urban concentration for a “low” emission scenario.12 Using both methods, one-box MUM provided emission estimates within the range estimated using the multibox MUM (Table S18). Using the average of all measured air concentrations (22 and 19 kg y−1 for PCBs and PBDEs, respectively) gave results closest to those of multibox MUM (19 (2−41) and 18 (3−42) kg y−1 for PCBs and PBDEs, respectively) rather than using the highest measured concentrations from 0KM and N10. These results suggest that the one-box model can be used with a spatially representative average urban air concentration to estimate aggregate emissions similar to those made using the multibox model. The advantage of the multibox model is presumably more accurate emission estimates and use of the model to relate emissions to local air concentrations and to emissions from the city as a whole. Spatially resolved emission estimates could also be used to identify source locations which could inform chemical removal strategies. Differing emission patterns can also lead to differing fates of the urban air plume which is further investigated in Csiszar et al.42 However, these advantages come with considerably greater data requirements, model complexity, and computational demands relative to the one-box version. 3.4. Sensitivity Analysis. On a daily basis, modeled air concentrations were most strongly correlated with wind speed, were weakly correlated with temperature, and were not correlated with precipitation (a further discussion of this can

(between BDE-28 and -183) relative to the difference in Log Koa of 1.3 times and 1.4 times, respectively. Mean Σ5PCB emission rates ranged from 0.04 to 430 g y−1 km−2 in the (8,7) and DT cells, respectively, a span of 4 orders of magnitude (Figure 3a). Cell (8,7) had no known in-use or in-storage PCB mass inventory versus 36 sites in the DT cell.12 Conversely, Σ5PBDE emission rates ranged from 2 to 60 g y−1 km−2 in (8,7) and (5,3) DT cells (Figure 3b). Thus, considering the city as a whole, PCBs act more as a point source in Toronto where 34% of the emissions originated from the DT cell. In comparison, PBDE emissions are spread more evenly throughout the city with near-downtown cells (4,3) and (5,3) accounting for 16% of the total emissions. The point source nature of PCBs stems from the intense electricity usage in the 1960s and 1970s downtown office towers, hospitals, and public buildings that dominated the 2006 PCB mass inventory; PCBs were not used in newer areas of the city built after the 1977 ban on new uses and have lower usage in residential areas.12 PBDE-containing products also have high usage intensities in the downtown, but are also used in virtually all buildings in Toronto (except perhaps those constructed since 2005) and hence they act more as a diffuse source throughout the city. The mass inventory and emission patterns explain the lower urban−rural gradient of PBDEs versus PCBs seen in air and soil.43,59 As will be discussed in greater detail by Csiszar et al.,42 SOMUM estimated that primary emissions clearly dominate total emissions. SO-MUM calculations estimated that during the simulation period, urban media (soil, vegetation, films, and tributaries) transferred 7 and 1 g d−1 to air via volatilization compared to estimated primary emissions of 52 and 49 g d−1 for Σ5PCBs and Σ5PBDEs, respectively. Thus, our analysis of multimedia exchange confirms previous conclusions that “fresh” emissions from PCB and PBDE stocks are by far the main contributors to urban air concentrations.9,10,20,76,77 One-Box Emission Estimates. As mentioned above, emission estimates for PBDEs using SO-MUM were lower than those using a single-box model for Toronto but similar for PCBs. These previous estimates, however, were made with concentrations measured in 2000−2001 and with fewer measurement locations. To explore whether the increased 442



Article

(3) Gullett, B. K.; Touati, A.; Hays, M. D. PCDD/F, PCB, HxCBz, PAH, and PM emission factors for fireplace and woodstove combustion in the San Francisco Bay region. Environ. Sci. Technol. 2003, 37 (9), 1758−1765. (4) Marr, L. C.; Kirchstetter, T. W.; Harley, R. A.; Miguel, A. H.; Hering, S. V.; Hammond, S. K. Characterization of polycyclic aromatic hydrocarbons in motor vehicle fuels and exhaust emissions. Environ. Sci. Technol. 1999, 33 (18), 3091−3099. (5) Pirrone, N.; Keeler, G. J.; Nriagu, J. O. Regional differences in worldwide emissions of mercury to the atmosphere. Atmos. Environ. 1996, 30 (17), 2981−2987. (6) Dyke, P. H.; Foan, C.; Fiedler, H. PCB and PAH releases from power stations and waste incineration processes in the UK. Chemosphere 2003, 50 (4), 469−480. (7) Breivik, K.; Sweetman, A.; Pacyna, J. M.; Jones, K. C. Towards a global historical emission inventory for selected PCB congeners - a mass balance approach. 2. Emissions. Sci. Total Environ. 2002, 290 (1− 3), 199−224. (8) Breivik, K.; Vestreng, V.; Rozovskaya, O.; Pacyna, J. M. Atmospheric emissions of some POPs in Europe: A discussion of existing inventories and data needs. Environ. Sci. Policy 2006, 9 (7−8), 663−674. (9) Moeckel, C.; Gasic, B.; Macleod, M.; Scheringer, M.; Jones, K. C.; Hungerbuhler, K. Estimation of the source strength of polybrominated diphenyl ethers based on their diel variability in air in Zurich, Switzerland. Environ. Sci. Technol. 2010, 44 (11), 4225−4231. (10) Prevedouros, K.; Jones, K. C.; Sweetman, A. J. Estimation of the production, consumption, and atmospheric emissions of pentabrominated diphenyl ether in Europe between 1970 and 2000. Environ. Sci. Technol. 2004, 38 (12), 3224−3231. (11) Alcock, R. E.; Sweetman, A. J.; Prevedouros, K.; Jones, K. C. Understanding levels and trends of BDE-47 in the UK and North America: An assessment of principal reservoirs and source inputs. Environ. Int. 2003, 29 (6), 691−698. (12) Diamond, M. L.; Melymuk, L.; Csiszar, S. A.; Robson, M. Estimation of PCB stocks, emissions, and urban fate: Will our policies reduce concentrations and exposure? Environ. Sci. Technol. 2010, 44 (8), 2777−2783. (13) Gasic, B.; Moeckel, C.; Macleod, M.; Brunner, J.; Scheringer, M.; Jones, K. C.; Hungerbuhler, K. Measuring and modeling shortterm variability of PCBs in air and characterization of urban source strength in Zurich, Switzerland. Environ. Sci. Technol. 2009, 43 (3), 769−776. (14) Gasic, B.; MacLeod, M.; Klanova, J.; Scheringer, M.; Ilic, P.; Lammel, G.; Pajovic, A.; Breivik, K.; Holoubek, I.; Hungerbuhler, K. Quantification of sources of PCBs to the atmosphere in urban areas: A comparison of cities in North America, Western Europe and former Yugoslavia. Environ. Pollut. 2010, 158 (10), 3230−3235. (15) Morselli, M.; Ghirardello, D.; Semplice, M.; Di Guardo, A. Modeling short-term variability of semivolatile organic chemicals in air at a local scale: An integrated modeling approach. Environ. Pollut. 2011, 159 (5), 1406−1412. (16) Jones-Otazo, H. A.; Clarke, J. P.; Diamond, M. L.; Archbold, J. A.; Ferguson, G.; Harner, T.; Richardson, G. M.; Ryan, J. J.; Wilford, B. Is house dust the missing exposure pathway for PBDEs? An analysis of the urban fate and human exposure to PBDEs. Environ. Sci. Technol. 2005, 39 (14), 5121−5130. (17) Jaward, F. M.; Farrar, N. J.; Harner, T.; Sweetman, A. J.; Jones, K. C. Passive air sampling of PCBs, PBDEs, and organochlorine pesticides across Europe. Environ. Sci. Technol. 2004, 38 (1), 34−41. (18) Ward, J.; Mohapatra, S. P.; Mitchell, A. An overview of policies for managing polybrominated diphenyl ethers (PBDEs) in the Great Lakes basin. Environ. Int. 2008, 34 (8), 1148−1156. (19) Salamova, A.; Hites, R. A. Discontinued and alternative brominated flame retardants in the atmosphere and precipitation from the Great Lakes basin. Environ. Sci. Technol. 2011, 45 (20), 8698−8706.

be found in the SI, Section S2.4). Hence, we focused our sensitivity analysis on the effects of wind speeds on emissions. Emission rates were positively linearly correlated with horizontal wind speed vh, vertical wind speed between the upper and lower air compartments vv, and dry deposition velocity vd (r2 = 1.0) for the ranges scanned which were 0.1−10, 0.01−0.1, and 0.1 to 10 cm s−1 for vh, vv, and vd, respectively. The ranges for the wind speeds were based on the high and low hourly values calculated by BLFM for the spring 2008 simulation period and thus have a relatively high degree of certainty. Emissions varied the most over the range of horizontal wind speeds with PCB emissions varying more than those of PBDEs. The emission rates varied 14, 2, and 1.1 times over the vh, vv, and vd ranges, respectively, for Σ5PCBs and 5, 2, and 2 times, respectively, for Σ5PBDEs (Table S19). Varying vd from 0.1 to 10 cm s−1 resulted in the advective losses of PCBs ranging from 100 to 97% of emissions; net air deposition ranging from 1% to 3% of emissions; and air concentrations ranging from 67 to 65 pg m−3 indicating that model results for PCBs are not particularly sensitive to changes in vd (Table S20), since they are mainly in the gas phase. For PBDEs, varying vd over the same range resulted in advective losses ranging from 93 to 49% of emissions; net air deposition ranging from 6 to 50% of emissions; and air concentrations ranging from 60 to 30 pg m−3. PBDE results were clearly more sensitive to changes in vd due to their higher particle association than PCBs (Table S21). The relationships between vd and the various model outputs were all linear or near-linear (r2 = 0.93− 1.0). Full details of the sensitivity analysis are presented in the SI.

■

ASSOCIATED CONTENT

■

AUTHOR INFORMATION

S Supporting Information *

Additional text, tables and figures as mentioned in the text. This information is available free of charge via the Internet at http:// pubs.acs.org. Corresponding Author

*E-mail: [email protected]; tel: +1 416 978 1586; fax: +1 416 946 5992; mail: 22 Russell St., Toronto, Ontario, Canada M5S 3B1. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Funding was provided by Ontario Graduate Scholarship and Department of Chemical Engineering and Applied Chemistry of the University of Toronto to S.C., and NSERC. This project benefited from financial support from the Great Lakes Commission, Ontario Ministry of Environment (MOE), Environment Canada (EC), and the Toronto and Region Conservation Authority. Our appreciation is extended to L. Melymuk and M. Robson (formerly of University of Toronto), P. Helm (MOE), and L. Jantunen, P. Blanchard, and T. Harner (EC).

■

REFERENCES

(1) Stockholm Convention. Stockholm Convention on Persistent Organic Pollutants (POPs), as amended in 2009; Secretariat of the Stockholm Convention: Châtelaine, Switzerland, 2009. (2) UNECE. Convention on Long-range Transboundary Air Pollution. http://www.unece.org/env/lrtap/ (February 25). 443



Article

(20) Robson, M.; Harrad, S. Chiral PCB signatures in air and soil: Implications for atmospheric source apportionment. Environ. Sci. Technol. 2004, 38 (6), 1662−1666. (21) Hung, H.; Lee, S. C.; Wania, F.; Blanchard, P.; Brice, K. Measuring and simulating atmospheric concentration trends of polychlorinated biphenyls in the Northern Hemisphere. Atmos. Environ. 2005, 39 (35), 6502−6512. (22) Gioia, R.; Steinnes, E.; Thomas, G. O.; Mejier, S. N.; Jones, K. C. Persistent organic pollutants in European background air: Derivation of temporal and latitudinal trends. J. Environ. Monit. 2006, 8 (7), 700−710. (23) Batterman, S. A.; Chernyak, S.; Jia, C.; Godwin, C.; Charles, S. Concentrations and emissions of polybrominated diphenyl ethers from US houses and garages. Environ. Sci. Technol. 2009, 43 (8), 2693− 2700. (24) Allen, J. G.; McClean, M. D.; Stapleton, H. M.; Webstert, T. F. Linking PBDEs in house dust to consumer products using X-ray fluorescence. Environ. Sci. Technol. 2008, 42 (11), 4222−4228. (25) Hazrati, S.; Harrad, S. Causes of variability in concentrations of polychlorinated biphenyls and polybrominated diphenyl ethers in indoor air. Environ. Sci. Technol. 2006, 40 (24), 7584−7589. (26) Zhang, X. M.; Diamond, M. L.; Robson, M.; Harrad, S. Sources, emissions, and fate of polybrominated diphenyl ethers and polychlorinated biphenyls indoors in Toronto, Canada. Environ. Sci. Technol. 2011, 45 (8), 3268−3274. (27) Björklund, J. A.; Thuresson, K.; Palm Cousins, A.; Sellström, U.; Emenius, G.; De Wit, C. A. Indoor air is a significant source of tridecabrominated diphenyl ethers to outdoor air via ventilation systems. Environ. Sci. Technol. 2012, 46 (11), 5876−5884. (28) Melymuk, L.; Robson, M.; Helm, P.; Diamond, M. Application of land use regression to identify sources and assess spatial variation in urban SVOC concentrations. Environ. Sci. Technol. 2012, submitted. (29) Hollander, A.; Hauck, M.; Cousins, I. T.; Huijbregts, M. A. J.; Pistocchi, A.; Ragas, A. M. J.; van de Meent, D. Assessing the relative importance of spatial variability in emissions versus landscape properties in fate models for environmental exposure assessment of chemicals. Environ. Model. Assess. 2012, 17, 577−587. (30) Daggupaty, S. M.; Banic, C. M.; Cheung, P.; Ma, J. M. Numerical simulation of air concentration and deposition of particulate metals around a copper smelter in northern Quebec, Canada. Geochem.: Explor., Environ. Anal. 2006, 6, 139−146. (31) Wania, F.; Mackay, D. A global distribution model for persistent organic-chemicals. Sci. Total Environ. 1995, 160−61, 211−232. (32) Scheringer, M.; Wegmann, F.; Fenner, K.; Hungerbuhler, K. Investigation of the cold condensation of persistent organic pollutants with a global multimedia fate model. Environ. Sci. Technol. 2000, 34 (9), 1842−1850. (33) Scheringer, M. Long-range transport of organic chemicals in the environment. Environ. Toxicol. Chem. 2009, 28 (4), 677−690. (34) MacLeod, M.; Woodfine, D. G.; Mackay, D.; McKone, T.; Bennett, D.; Maddalena, R. BETR North America: A regionally segmented multimedia contaminant fate model for North America. Environ. Sci. Pollut. Res. 2001, 8 (3), 156−163. (35) Prevedouros, K.; MacLeod, M.; Jones, K. C.; Sweetman, A. J. Modelling the fate of persistent organic pollutants in Europe: Parameterisation of a gridded distribution model. Environ. Pollut. 2004, 128 (1−2), 251−261. (36) Priemer, D. A.; Diamond, M. L. Application of the multimedia urban model to compare the fate of SOCs in an urban and forested watershed. Environ. Sci. Technol. 2002, 36 (5), 1004−1013. (37) Ma, J. M.; Daggupaty, S.; Harner, T.; Li, Y. F. Impacts of lindane usage in the Canadian prairies on the Great Lakes ecosystem. 1. Coupled atmospheric transport model and modeled concentrations in air and soil. Environ. Sci. Technol. 2003, 37 (17), 3774−3781. (38) Suzuki, N.; Murasawa, K.; Sakurai, T.; Nansai, K.; Matsuhashi, K.; Moriguchi, Y.; Tanabe, K.; Nakasugi, O.; Morita, M. Georeferenced multimedia environmental fate model (G-CIEMS): Model formulation and comparison to the generic model and monitoring approaches. Environ. Sci. Technol. 2004, 38 (21), 5682−5693.

(39) Kim, J. H.; Kwak, B. K.; Shin, C. B.; Jeon, W. J.; Park, H.-S.; Lee, S. W.; Choi, K.; Lee, W. G.; Lee, J. H.; Baek, S. H.; Yi, J. Development of a local-scale spatially refined multimedia fate model (LSRMFM) for urban-scale risk assessment: Model formulation, GIS-based preprocessing, and case study. Environ. Model. Assess. 2011, 16 (3), 265−281. (40) Morselli, M.; Ghirardello, D.; Semplice, M.; Raspa, G.; Di Guardo, A. Integration of an atmospheric dispersion model with a dynamic multimedia fate model: Development and illustration. Environ. Pollut. 2012, 164, 182−187. (41) Csiszar, S. A.; Diamond, M. L.; Thibodeaux, L. J. Modeling urban films using a dynamic multimedia fugacity model. Chemosphere 2012, 87 (9), 1024−1031. (42) Csiszar, S. A.; Diamond, M. L.; Daggupaty, S. M. Modeling the influence of urban SVOC emissions on neighboring regions using a coupled multimedia and air transport model. In Prep. (43) Melymuk, M.; Robson, R.; Helm, P.; Diamond, M. L. PCBs, PBDEs, and PAHs in the urban atmosphere: Spatial and seasonal trends and implications for contaminant transport. Sci. Total Environ. 2012, 429, 272−280. (44) Diamond, M. L.; Priemer, D. A.; Law, N. L. Developing a multimedia model of chemical dynamics in an urban area. Chemosphere 2001, 44 (7), 1655−1667. (45) Daggupaty, S. M.; Tangirala, R. S.; Sahota, H. BLFMESO - A 3dimensional mesoscale meteorological model for microcomputers. Bound.-Layer Meteor. 1994, 71 (1−2), 81−107. (46) Daggupaty, S. M. A case study of the simultaneous development of multiple lake-breeze fronts with a boundary layer forecast model. J. Appl. Meteorol. 2001, 40 (3), 289−311. (47) Cote, J.; Gravel, S.; Methot, A.; Patoine, A.; Roch, M.; Staniforth, A. The operational CMC-MRB Global Environmental Multiscale (GEM) model. Part I: Design considerations and formulation. Mon. Weather Rev. 1998, 126 (6), 1373−1395. (48) Cote, J.; Desmarais, J. G.; Gravel, S.; Methot, A.; Patoine, A.; Roch, M.; Staniforth, A. The operational CMC-MRB Global Environmental Multiscale (GEM) model. Part II: Results. Mon. Weather Rev. 1998, 126 (6), 1397−1418. (49) Gong, S. L.; Huang, P.; Zhao, T. L.; Sahsuvar, L.; Barrie, L. A.; Kaminski, J. W.; Li, Y. F.; Niu, T. GEM/POPs: A global 3-D dynamic model for semi-volatile persistent organic pollutants - Part 1: Model description and evaluations of air concentrations. Atmos. Chem. Phys. 2007, 7 (15), 4001−4013. (50) Smyth, S. C.; Yin, D. Z.; Roth, H.; Jiang, W. M.; Moran, M. D.; Crevier, L. P. The impact of GEM and MM5 modeled meteorological conditions on CMAQ air quality modeling results in eastern Canada and the Northeastern United States. J. Appl. Meteorol. Clim. 2006, 45 (11), 1525−1541. (51) Melymuk, L.; Robson, M.; Helm, P. A.; Diamond, M. L. Evaluation of passive air sampler calibrations: Selection of sampling rates and implications for the measurement of persistent organic pollutants in air. Atmos. Environ. 2011, 45 (10), 1867−1875. (52) Farrar, N. J.; Harner, T.; Shoeib, M.; Sweetman, A.; Jones, K. C. Field deployment of thin film passive air samplers for persistent organic pollutants: a study in the urban atmospheric boundary layer. Environ. Sci. Technol. 2005, 39 (1), 42−48. (53) Robson, M.; Melymuk, L.; Csiszar, S. A.; Giang, A.; Diamond, M. L.; Helm, P. A. Continuing sources of PCBs: The significance of building sealants. Environ. Int. 2010, 36, 506−513. (54) Morf, L.; Smutny, R.; Taverna, R.; Daxbeck, H. Selected Polybrominated Flame Retardants - PBDE and TBBPA - Substance Flow Analysis; Swiss Agency for the Environment, Forests, and Landscape: Bern, 2002. (55) Morf, L. S.; Tremp, J.; Gloor, R.; Huber, Y.; Stengele, M.; Zennegg, M. Brominated flame retardants in waste electrical and electronic equipment: Substance flows in a recycling plant. Environ. Sci. Technol. 2005, 39 (22), 8691−8699. (56) Imm, P.; Knobeloch, L.; Buelow, C.; Anderson, H. A. Household Exposures to polybrominated diphenyl ethers (PBDEs) in a Wisconsin cohort. Environ. Health Perspect. 2009, 117 (12), 1890− 1895. 444



Article

(57) Shen, L.; Wania, F.; Lei, Y. D.; Teixeira, C.; Muir, D. C. G.; Xiao, H. Polychlorinated biphenyls and polybrominated diphenyl ethers in the North American atmosphere. Environ. Pollut. 2006, 144 (2), 434−444. (58) Gouin, T.; Harner, T.; Daly, G. L.; Wania, F.; Mackay, D.; Jones, K. C. Variability of concentrations of polybrominated diphenyl ethers and polychlorinated biphenyls in air: Implications for monitoring, modeling and control. Atmos. Environ. 2005, 39 (1), 151−166. (59) Harner, T.; Shoeib, M.; Diamond, M.; Ikonomou, M.; Stern, G. Passive sampler derived air concentrations of PBDEs along an urbanrural transect: Spatial and temporal trends. Chemosphere 2006, 64 (2), 262−267. (60) Motelay-Massei, A.; Harner, T.; Shoeib, M.; Diamond, M.; Stern, G.; Rosenberg, B. Using passive air samplers to assess urbanrural trends for persistent organic pollutants and polycyclic aromatic hydrocarbons. 2. Seasonal trends for PAHs, PCBs, and organochlorine pesticides. Environ. Sci. Technol. 2005, 39 (15), 5763−5773. (61) Environment Canada. National Inventory of PCBs in Use and PCB Wastes in Storage in Canada -- 2005 Annual Report; Environmental Stewardship Branch Government of Canada: Ottawa, Ontario, 2006. (62) Hu, D. F.; Lehmler, H. J.; Martinez, A.; Wang, K.; Hornbuckle, K. C. Atmospheric PCB congeners across Chicago. Atmos. Environ. 2010, 44 (12), 1550−1557. (63) Persoon, C.; Peters, T. M.; Kumar, N.; Hornbuckle, K. C. Spatial distribution of airborne polychlorinated biphenyls in Cleveland, Ohio and Chicago, Illinois. Environ. Sci. Technol. 2010, 44 (8), 2797− 2802. (64) Du, S.; Wall, S. J.; Cacia, D.; Rodenburg, L. A. Passive air sampling for polychlorinated biphenyls in the Philadelphia metropolitan area. Environ. Sci. Technol. 2009, 43 (5), 1287−1292. (65) Gingrich, S. E.; Diamond, M. L.; Stern, G. A.; McCarry, B. E. Atmospherically derived organic surface films along an urban-rural gradient. Environ. Sci. Technol. 2001, 35 (20), 4031−4037. (66) Butt, C. M.; Diamond, M. L.; Truong, J.; Ikonomou, M. G.; Ter Schure, A. F. H. Spatial distribution of polybrominated diphenyl ethers in southern Ontario as measured in indoor and outdoor window organic films. Environ. Sci. Technol. 2004, 38 (3), 724−731. (67) Melymuk, L.; Robson, M.; Diamond, M. L.; Bradley, L. E.; Backus, S. Wet deposition loadings of organic contaminants to Lake Ontario: Assessing the influence of precipitation from urban and rural sites. Atmos. Environ. 2011, 45 (28), 5042−5049. (68) Melymuk, L.; Robson, M.; Csiszar, S. A.; Diamond, M. L.; Helm, P.; Keltenecker, G.; Backus, S.; Bradley, L.; Blanchard, P.; Jantunen, L. From the city to the lake: Loadings of PCBs, PBDEs, PAHs and PCMs from Toronto to Lake Ontario. In Prep. (69) Robson, M.; Melymuk, L.; Gilbert, B.; Helm, P.; Diamond, M.; Blanchard, P.; Backus, S. Comparison of concentrations and loadings of PCBs and PAHs in urban and rural streams during base flow and storm events. Organohalogen Compd. 2008, 70, 685−688. (70) Schecter, A.; Papke, O.; Tung, K. C.; Joseph, J.; Harris, T. R.; Dahlgren, J. Polybrominated diphenyl ether flame retardants in the US population: Current levels, temporal trends, and comparison with dioxins, dibenzofurans, and polychlorinated biphenyls. J. Occup. Environ. Med. 2005, 47 (3), 199−211. (71) Harrad, S.; Diamond, M. New directions: Exposure to polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls (PCBs): Current and future scenarios. Atmos. Environ. 2006, 40 (6), 1187−1188. (72) Totten, L. A.; Stenchikov, G.; Gigliotti, C. L.; Lahoti, N.; Eisenreich, S. J. Measurement and modeling of urban atmospheric PCB concentrations on a small (8 km) spatial scale. Atmos. Environ. 2006, 40 (40), 7940−7952. (73) La Guardia, M. J.; Hale, R. C.; Harvey, E. Detailed polybrominated diphenyl ether (PBDE) congener composition of the widely used penta-, octa-, and deca-PBDE technical flameretardant mixtures. Environ. Sci. Technol. 2006, 40 (20), 6247−6254. (74) Frame, G. M.; Cochran, J. W.; Bowadt, S. S. Complete PCB congener distributions for 17 aroclor mixtures determined by 3 HRGC

systems optimized for comprehensive, quantitative, congener-specific analysis. HRC-J. High Res. Chromatogr. 1996, 19 (12), 657−668. (75) Christensen, E. R.; Lo, C. K. Polychlorinated-biphenyls in dated sediments of Milwaukee-Harbor, Wisconsin, USA. Environ. Pollut. B 1986, 12 (3), 217−232. (76) Basu, I.; Hafner, W. D.; Mills, W. J.; Hites, R. A. Differences in atmospheric persistent organic pollutant concentrations at two locations in Chicago. J. Great Lakes Res. 2004, 30 (2), 310−315. (77) Kohler, M.; Tremp, J.; Zennegg, M.; Seiler, C.; Minder-Kohler, S.; Beck, M.; Lienemann, P.; Wegmann, L.; Schmidt, P. Joint sealants: an overlooked diffuse source of polychlorinated biphenyls in buildings. Environ. Sci. Technol. 2005, 39 (7), 1967−1973.

445


SO-MUM: A Coupled Atmospheric Transport and Multimedia Model

Recommend Documents