Deposition of Polybrominated Diphenyl Ethers, Polychlorinated

Specifically, McLachlan et al. proposed an interpretative framework to quantitatively describe deposition from air to plants based on an SOC's KOA (2,...
0 downloads 3 Views 135KB Size
Download PDF
Environ. Sci. Technol. 2007, 41, 534-540

Deposition of Polybrominated Diphenyl Ethers, Polychlorinated Biphenyls, and Polycyclic Aromatic Hydrocarbons to a Boreal Deciduous Forest

velocities are common throughout boreal and temperate deciduous forests. These extraordinarily high deposition velocities of semi-volatile organic compounds to deciduous forest canopies are at the core of the hypothesis of a significant filter effect of forests on a regional and global scale.

Y U S H A N S U , †,§ F R A N K W A N I A , * ,† TOM HARNER,‡ AND YING D. LEI† Department of Chemical Engineering and Applied Chemistry and Department of Physical and Environmental Sciences, University of Toronto at Scarborough, 1265 Military Trail, Toronto, Ontario, Canada M1C 1A4, and Science and Technology Branch, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada M3H 5T4

Forests play an important role in trapping airborne semivolatile organic compounds (SOCs) and transferring them to terrestrial ecosystems with falling leaves (1-5). Higher SOC deposition fluxes were observed under forest canopies than in nearby open areas as a result of this forest filter effect (2, 5, 6). The active “pumping” process driven by forests reduces SOC air concentrations (2, 7) and eventually leads to elevated contaminant levels in forest soils (8-10). Air concentrations of polychlorinated biphenyls (PCBs) showed stronger temperature dependence in a forest than a nearby clearing, presumably because of historically elevated deposition to the forest floor (11). Model calculations suggest that the forest filter effect is most pronounced for SOCs with an octanol/air partition coefficient log KOA between 7 and 11 and an air/water partition coefficient log KAW higher than -6 (2, 12). For such chemicals, uptake in forests may notably decrease air concentrations (12). Modeling simulations on a global scale suggest that the world’s forests markedly decrease the long-range transport of some SOCs (13, 14). These calculations showed that the global forest filter effect is very sensitive to the input parameters for the deposition velocities, especially the dry gaseous deposition velocity to the boreal deciduous forest (14). So far, only one set of deposition velocities of SOCs to temperate forests has been reported (2). In particular, no such empirical data exist for the boreal forests, even though deposition velocities may vary with tree species, canopy structure, and climatic conditions (1, 2, 4). Therefore, additional field measurements are urgently required. In this study, the measurement approach developed by Horstmann and McLachlan (2) was employed to quantify the filter effect of a Canadian deciduous forest canopy for several classes of SOCs, including polybrominated diphenyl ethers (PBDEs), PCBs, and polycyclic aromatic hydrocarbons (PAHs). This study thus not only reports the first deposition velocity to a forest in North America, it also presents the first such data for PBDEs to any surface. A comparison of measurements between different forests may afford further insight into SOC deposition to forests. Considering the large area covered by boreal forests, the new kinetic data should be useful in modeling studies quantifying the role of these forests in SOC cycling on a continental and global scale.

The atmospheric deposition of several groups of semivolatile organic compounds to a deciduous forest in Canada was determined using an indirect technique based on ratios of measured canopy interception and air concentrations. Air (gas and particle phase) and bulk deposition were sampled for 14 months from October 2001 to December 2002 at both a forest and a nearby clearing, and extracts were quantified for polybrominated diphenyl ethers (PBDEs), polychlorinated biphenyls (PCBs), and polycyclic aromatic hydrocarbons (PAHs). Long-term average dry deposition velocities for vapors and particle-bound species were then derived for the canopy growing period. The mean dry gaseous deposition velocity for PBDEs and PCBs to the Canadian deciduous forest was 2.7 ( 0.52 cm‚s-1, which is similar to the only other measured value for a deciduous canopy. Particle-bound deposition velocities to the canopy due to diffusion and impaction were 0.8 cm‚s-1 for the PBDEs and 0.11 cm‚s-1 for the PAHs. Differences in the particle-bound deposition velocities between PBDEs and PAHs and between deciduous canopies in Canada and Germany are explainable by differences in particle size distribution. The interception/concentration ratios for several PAHs were too low to be interpretable as dry gaseous deposition velocities. This is likely because the measured deposition flux under the canopy was less than the deposition flux to the canopy, possibly as a result of photodegradation in the canopy. From the ratio of canopy interception and average gas-phase concentration of less chlorinated PCBs, a predictive relationship between the canopy/air partition coefficient KPA and the octanol/air partition coefficient KOA was derived (KPA ) 110 KOA0.67). Despite differences in local climate and canopy composition and structure, the deposition velocities and the canopy uptake capacity measured in Canada were remarkably similar to those reported in Germany, lending credibility to the suggestion that high gaseous deposition * Corresponding author phone: (416)287-7225; fax: (416)287-7279; e-mail: [email protected]. † University of Toronto at Scarborough. ‡ Environment Canada. § Current address: Science and Technology Branch, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada M3H 5T4. 534

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 2, 2007

Introduction

Theory The octanol/air partition coefficient KOA has been employed to categorize the atmospheric deposition behavior of different SOCs to vegetation, including grassland (15) and forests (2). Specifically, McLachlan et al. proposed an interpretative framework to quantitatively describe deposition from air to plants based on an SOC’s KOA (2, 3, 15) and identified three major categories of deposition to a forest canopy: Category A, gaseous equilibrium partitioning between air and forest canopy for volatile SOCs; Category B, kinetically controlled dry gaseous deposition from air to forest canopy for SOCs with intermediate volatility; Category C, dry particle-bound deposition, primarily by particle diffusion and impaction, 10.1021/es0622047 CCC: $37.00

 2007 American Chemical Society Published on Web 12/08/2006

(diffusion/impaction) (i.e., I ) IP+IG). If a class of chemicals (e.g., the PAHs) is assumed to associate with the same types of particles, the particle-bound interception IP is the product of the particle-bound concentration CP and the vP derived for category C chemicals (see extrapolation of IP/CP line to lower KOA in Figure 1):

IP ) CP‚vP

(3)

The gaseous interception IG is then simply the difference between I and IP:

IG ) I - IP

(4)

The ratio of IG and the gaseous concentration CG can be interpreted as a dry gaseous deposition velocity vG to the forest canopy (eq 5).

vG ) IG/CG

FIGURE 1. Diagram illustrating the indirect measurement of dry gaseous (vG) and particle-bound (vP) deposition velocities to the forest canopy, and the pseudo-canopy/air partition coefficient KPA for different categories of SOCs. from air to forest canopy for very involatile SOCs (Figure 1). Deposition of persistent SOCs to the canopy can be quantified indirectly by measuring concurrently air concentrations and deposition fluxes under the canopy. For instance, long-term average dry gaseous and particle-bound deposition velocities for SOCs in Categories B and C were previously measured in a deciduous and coniferous forest in Southern Germany (2). When determining deposition velocities to a forest with the indirect technique by Horstman and McLachlan (2), air concentrations, C, and deposition fluxes, F, of SOCs are quantified at both the forest and a nearby clearing using air and bulk deposition samplers, respectively (Figure 1). The deposition flux of a SOC to the clearing, FC, is mainly composed of wet deposition and particle sedimentation, whereas the deposition flux in the forest, FF, additionally includes dry gaseous and particle-bound deposition (diffusion/impaction) to the canopy (2). Canopy interception, I, is then calculated by subtracting FC from FF:

I ) FF - F C

(1)

As a quantitative measure of the ability of the forest canopy to take up SOCs, canopy interception I represents dry gaseous and particle-bound deposition (diffusion/impaction) from air to forest canopy. If a SOC is persistent in the canopy, its deposition velocity to the canopy can be quantified indirectly from the measured canopy interception I and air concentration C (2). (i) For very involatile SOCs (Category C), I is entirely due to particle-bound deposition (diffusion/impaction) (i.e., I ) IP). A dry particle-bound deposition velocity vP can be directly calculated from IP and a particle-bound concentration CP:

vP ) IP/CP

(2)

The uptake capacity of the canopy for chemicals in Category C is quite large and no chemical equilibrium is reached during the leaf’s life span on the trees; therefore, vP is independent of KOA (Figure 1) and only depends on the deposition of the particles. (ii) For SOCs with intermediate volatility (Category B), I is composed of both gaseous and particle-bound deposition

(5)

For SOCs of category B, vG is independent of KOA, reflecting a kinetically controlled deposition process. vG would also be applicable to other SOCs in Categories A and C (indicated by extrapolation of the IG/CG line to higher and lower KOA in Figure 1), but CG of SOCs in category C is often not quantifiable because they are mostly particle-bound. (iii) For very volatile SOCs (Category A), the particle-bound percentage is negligible and chemicals reach pseudoequilibrium between air and canopy during the leaf’s lifetime because of low uptake capacity of the canopy. The calculated IG/CG ratio is thus correlated to KOA and can no longer be interpreted as a deposition velocity. Instead, it is an indicator of canopy uptake capacity:

log [(IG/CG)/(cm‚s-1)] ) m‚log KOA - b

(6)

A dimensionless pseudo-plant/air partition coefficient (KPA) can be derived using the following equation:

KPA ) (IG/CG)‚t‚(1/V)

(7)

where V is the volume density of vegetation in the canopy (m3‚m-2), and t is the canopy’s exposure time to the air (2).

Experimental Section Field Sampling. The deciduous forest, located in the Great Lakes region of Southern Ontario, Canada, is dominantly composed of mixed red maple (Acer rubrum) and large-tooth aspen (Populus grandidentata). During summer, it has a leaf area index (LAI) of approximately 5 (16). The clearing is located about 3 km west of the forest site. Simultaneous sampling of air (gas and particle phase) and bulk deposition for SOCs was conducted from October 2001 to December 2002 at both sites. High volume (HiVol) air samplers were deployed under the forest canopy and at the clearing every 12 days to coincide with air measurements of the Integrated Atmospheric Deposition Network (17). 750∼900 m3 volume of air was pumped through two glass fiber filters (GFFs) and two polyurethane foam (PUF) plugs in series over 24 h. Two sets of 16 bulk deposition samplers were deployed at both forest and clearing in a grid fashion at intervals of 5 meters. These samplers, consisting of 4 liter glass jars (I.D. of opening 9.7 cm) mounted at a height of 1.5 m above ground, were collected monthly. Two parallel deposition samples were obtained by combining eight glass jars at each site. Detailed information on field sampling can be found in the Supporting Information (SI) and refs 11 and 18. Sample Extraction, Cleanup, and Analysis. PUFs and GFFs were Soxhlet-extracted overnight using petroleum ether and dichloromethane (DCM), respectively. Rainwater was VOL. 41, NO. 2, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

535

TABLE 1. Deposition Fluxes to Forest (FF) and Clearing (FC) during Half-year Period, Arithmetic Mean of Particle-bound (C h P), and h Ratios, and log KOA Values at 25 °C for PAHs (BDL ) Gaseous (C h G) Atmospheric Concentration in the Clearing, Derived I/C Below Method Detection Limit)

compound fluorene phenanthrene fluoranthene pyrene benz[a]anthracene chrysene benzo[b]fluoranthene benzo[k]fluoranthene benzo[a]pyrene indeno[123,cd]pyrene dibenz[ah]anthracene benzo[ghi]perylene

deposition flux (ng‚m-2‚half-yr-1) FF FC 3310 18100 14700 20300 1600 4030 4610 1620 2070 1780 257 2040

Ch P (pg‚m-3)

Ch G (pg‚m-3)

IG/Ch G a (cm‚s-1)

1.6 12 17 30 15 20 75 36 68 81 13 92

1670 5700 625 871 44 100 18 7.7 BDL BDL BDL BDL

(0.011) (0.018) (0.12) (0.11) (0.16) (0.19) (0.75) (0.49)

254 1560 2400 4420 221 607 1130 366 524 415 35 603

IP/Ch P b (cm‚s-1)

log KOA (25 °C)

0.14 0.11 0.11 0.10

6.79,c 6.97d 7.57,c 7.78d 8.88,c 8.89d 8.80,c 8.93d 10.10d 10.42d 11.75d 11.68d 11.86d 13.03d 13.21d 12.98d

a These numbers are too low to be interpreted as net dry gaseous deposition velocities (v ) to the canopy. b These numbers are interpreted G as net dry particle-bound deposition velocities (vP) to the canopy (chemicals in Category C), see text for details. c Ref 19. d Calculated from subcooled liquid vapor pressure (PL) using an equation in ref 20. PL is from ref 21.

concentrated with a solid-phase extraction cartridge and then eluted with a mixture of acetone and ethyl acetate. The elute was combined with other solid phases of the bulk samples (e.g., leaves and filtered particles) and Soxhlet-extracted overnight using DCM. Air samples were cleaned and fractionated using a silicic acid/alumina (S/A) column as described previously (18). Bulk deposition samples were first run through S/A columns, and then cleaned by gel permeation. Mirex (100 ng) was added as an internal standard to all samples and standards prior to analysis. Instrumental analysis was conducted on an Agilent 6890 gas chromatograph equipped with an auto-sampler (Agilent 7683) and a mass spectrometric detector (Agilent 5973). PBDEs were analyzed in negative chemical ionization mode, whereas PCBs and PAHs were analyzed by electron impact in selected ion monitoring mode. Description of methods can be found in the SI and refs 11, 18. Quality Assurance/Quality Control. Method detection limit (MDL) of air samples was defined as mean field blank plus three time standard deviation. MDLs of air samples for PBDEs, PCBs, and PAHs were reported previously (11, 18). All data were blank corrected provided they exceeded the MDLs, but no data were adjusted for recoveries. Laboratory procedure blanks and thimble blanks were not detectable. For bulk deposition samples, a mixture of 13C12 PCBs (namely, 13 C12 PCB-28, 52, 101, 138, 153, 180, and 209) was added to the filtered water if rainwater was collected in those samples; otherwise, surrogates were directly spiked prior to Soxhlet extraction. Recoveries in bulk deposition samples ranged from 59 ( 18% to 75 ( 21% (n ) 48) (Table S1). The two parallel bulk deposition samples are in good agreement and difference between them was mostly less than 20% (Fig. S3). A larger difference (up to 50%) was observed during October and November when leaves fell to the ground, reflecting the spatial heterogeneity under the forest canopy. Since the bulk of the annual deposition under the forest occured during these 2 months, a relatively high uncertainty in the deposition flux during fall directly impacts on the uncertainty of the derived deposition velocities. A separate experiment with urban atmospheric particulate matter (Standard Reference Material SRM 1649a, NIST) was conducted to assess the potential of SOC loss from the bulk deposition samplers by photolysis and volatilization. This experiment showed that recoveries of PAHs and PCBs from SRM are relatively high; most are higher than 80% and even above 90% for many of the SOCs subject to the forest filter effect (see the SI). Therefore, the bulk deposition sampler is 536

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 2, 2007

judged suitable to collect deposition of SOCs which are subject to effective forest filtering. Quality assurance/quality control is detailed in the SI.

Results and Discussion Bulk Deposition Fluxes. The forest canopy developed in May and all leaves fell to the forest floor between October and early November. Data analysis focused on samples collected in the time period May to November 2002 when leaves were on trees, even though the chemicals of interest were analyzed and quantified in all monthly bulk deposition samples (October 2001 to November 2002) (Table S3). The 6-month cumulative deposition fluxes to forest FF and clearing FC are shown in Table 1 for PAHs and in Table 2 for PBDEs and PCBs. Reflecting the uptake of SOCs by the forest canopy, the depositional flux FF was higher than FC for all compounds. Air Concentrations. The concentrations of PBDEs, PCBs, and PAHs in air were reported in two previous papers (11, 18). A forest canopy may reduce SOC concentrations in the atmosphere (2, 5, 7), if chemical net transfer is from air to surfaces. Therefore, air concentrations measured in the clearing were used to represent ambient background conditions. Consistent with the bulk deposition samples, air concentrations of gaseous (C h G) and particle-bound (C h P) SOCs in the clearing are averaged for the period of 6 months and are shown in Table 1 for PAHs and in Table 2 for PBDEs and PCBs. Arithmetic mean concentrations were employed. Medians or geometric means would yield slightly lower air concentrations. Dry Deposition Velocities for PAHs. Canopy interception I of PAHs is calculated using eq 1. By assuming constant air concentrations during the period of 6 months, long-term averaged dry deposition velocity can be calculated from the ratio of the cumulative forest canopy interception and the average air concentration. Four PAHs (i.e., benzo[a]pyrene, indeno[123-cd]pyrene, dibenz[ah]anthracene, and benzo[ghi]perylene) are virtually 100% particle-bound in the air (18) and thus belong into category C (Figure 1). Dry particlebound deposition velocities vP are calculated for these four PAHs using eq 2 and are listed in Table 1. The values of vP obtained for these four PAHs are in good agreement, indicating independence of physical-chemical properties. Most studies on the particle size-distribution of PAHs and PCDDs/PCDFs suggest that different members of one group of chemicals are typically associated with the same types of

TABLE 2. Deposition Fluxes to Forest (FF) and Clearing (FC), and Arithmetic Mean of Particle-bound (C h P) and Gaseous (C h G) h G Ratio; and log KOA Values at 25 °C for PBDEs Atmospheric Concentrations in the Clearing during Half-year Period; Derived IG/C and PCBs (BDL ) Below Method Detection Limit)

compound BDE17 BDE28 BDE47 BDE66 BDE85 BDE99 BDE100 BDE153 BDE154 PCB 8 PCB 17 PCB 18 PCB 28 PCB 44 PCB 52 PCB 87 PCB 95 PCB 99 PCB 101 PCB 110 PCB 118 PCB 138 PCB 149 PCB 151 PCB 153 PCB 180 PCB 187

deposition flux (ng‚m-2‚half-yr-1) FF FC 89.0 216 5470 160 196 5010 1090 330 308 227 171 608 738 605 1050 815 984 591 1670 1540 890 215 589 128 775 77 92

6.4 15 1140 23 41 1200 242 80 71 2.9 3.7 23 25 13 34 6.8 22 7.4 47 23 5.3 4.8 11 1.1 36 6.9 7.8

Ch P (pg‚m-3) BDL BDL 0.80 BDL 0.10 2.5 0.51 0.49 0.25 BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL BDL

Ch G (pg‚m-3)

IG/Ch G (cm‚s-1)

log KOA (25 °C)

0.20 0.42 16 0.41 0.32 10 2.3 0.49 0.47 5.8 2.3 8.3 6.4 2.4 4.2 1.6 2.2 1.2 3.1 2.7 2.3 0.43 1.4 0.37 1.6 0.12 0.20

2.6a 3.0a 1.7a 2.1a 2.8a 2.1a 2.2a 2.4a 2.8a 0.24b 0.44b 0.47b 0.70b 1.6b 1.5b 3.1a 2.8a 3.1a 3.4a 3.6a 2.4a 3.1a 2.6a 2.2a 2.9a 3.6a 2.7a

9.30c 9.50c 10.53,c 10.34d 10.82,c 10.49d 11.66c 11.31,c 11.20d 11.13c 11.82,c 12.15d 11.92c 7.40,e 7.20g 7.55g 7.60,e 7.55g 7.92,e 7.89g 8.36,e 8.23g 8.22,e 8.23g 8.92g 8.71,e 8.79,f 8.58g 8.92g 8.80,e 9.06,f 8.92g 9.06,e 8.92g 9.82,f 9.26g 9.51,e 9.80,f 9.60g 9.26g 9.26g 9.37,e 9.73,f 9.60g 9.88,e 10.51,f 10.29g 9.87,e 9.95g

a These numbers are interpreted as kinetically controlled dry gaseous deposition velocities (v ) to the canopy (chemicals in Category B). b These G numbers are lower than the actual gaseous deposition velocity because an equilibrium between air and canopy was approached (chemicals in c d e f g Category A). Ref 22. Ref 23. Ref 24. Ref 25. Calculated from QSPR equations given in ref 26.

particles (27-30). In other words, the particle-bound (diffusion/ impaction) deposition velocity can be assumed to have the same value for a group of related chemicals since it is determined by properties of the particles, instead of physical-chemical properties of the SOCs (i.e., KOA). On the basis of this argument, the average particle-bound deposition velocity vj P calculated for these four PAHs (0.11 cm‚s-1) is applied to the whole group of PAHs. The importance of this assumption for the derived kinetic data is fairly limited, because for most of the Category-B PAH, IP constitutes less than 10% of I (Table S2). As mentioned in the Theory section, the canopy interception I of intermediate PAHs (Category B) consists of both gaseous and particle-bound deposition. The particle-bound interception IP of these PAHs is directly calculated from averaged particle-bound concentration C h P and the averaged vj P derived for PAHs with eq 3. The gaseous interception IG is then calculated from eq 4. Derived IG/CG ratios for PAHs are listed in Table 1, together with their KOA values at 25 °C. As will be discussed in detail below, the IG/CG ratios for PAHs were too low to be interpreted as dry gaseous deposition velocities, presumably because of underestimation of IG. Dry Deposition Velocities for PCBs and PBDEs. Concentrations of particle-bound PCBs were negligible at the sampling sites (11). Therefore, nearly all forest canopy interception of PCBs was due to gaseous deposition (i.e., IG ) I) and the IG/CG ratio is simply calculated from IG and CG (Table 2). Equation 2 cannot be used directly to calculate a particle-bound deposition velocity vP for PBDEs since none of them was 100% particle-bound. By combining eqs 2-5, we obtain the following:

vG ) (I - vP‚CP)/CG

(8)

It is assumed that vP and vG are the same for the group of PBDEs and independent of their physical-chemical properties, since both parameters describe kinetically controlled processes. An average vj P for PBDEs is obtained by choosing the value of vP in eq 8 that yields the minimum standard deviation of vG among nine PBDE congeners (Table 2). The optimized vP for PBDEs is 0.8 cm‚s-1, and the standard deviation of vG is 0.42 cm‚s-1. The gaseous deposition velocities for the PBDEs to the forest canopy are then calculated with eqs 3-5 and are listed in Table 2. Comparison Between PAHs and PBDEs/PCBs. The derived dry particle-bound deposition velocities, vP, for the PAHs are considerably smaller than for the PBDEs (Tables 1 and 2). The difference in the particle-bound deposition velocities (0.8 cm‚s-1 for the PBDEs vs 0.11 cm‚s-1 for the PAHs) could be due to different particle size distributions for the two groups of SOCs. For example, PBDEs may be associated with slightly larger particles than the PAHs, which would result in faster deposition velocities. Horstmann and McLachlan (2) had also noted a marked difference in the particle-bound deposition velocities to a deciduous canopy for PAHs and PCDDs/PCDFs and speculated that it is related to differences in the particle size distribution of these two groups of compounds. Studies on the particle size distribution of different chemical classes may provide direct evidence of differences in dry particle-bound deposition velocities. The averaged IG/CG ratio, interpreted as the dry gaseous deposition velocity vG, for nine PBDE and twelve PCB congeners listed in Table 2 is 2.7 ( 0.52 cm‚s-1. The standard deviation of this number is calculated from 21 individual measurements of the same parameter and thus should provide a reasonable estimate of the random error of the determined value. It does not quantify any systematic VOL. 41, NO. 2, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

537

deviation, as could, e.g., be introduced by the averaging technique for the air concentrations. Such a high value of the dry gaseous deposition velocity and independence of physical-chemical properties (e.g., KOA) confirm that the rate of SOC deposition to a deciduous forest canopy is controlled by the air-side resistance (2). The difference in the IG/CG ratios for the Category B-PBDEs and PCBs on one hand (2.7 cm‚s-1) and the Category B-PAHs on the other hand ( 90% I, except for BDE-153, when IG ) 76% I) (Table S2). However, a high uncertainty of IG does not explain a consistent underestimation of IG. Neither is it likely that the lower IG/CG values for the PAHs are due to equilibration with the atmospheric gas phase, because they even apply to fairly involatile compounds with log KOA above 10. A low IG/CG would however occur if some PAHs were not completely extractable from leaves, or if the PAHs in these samples were subject to relatively rapid degradation. Real-time visualization of low molecular weight PAHs (anthracene, fluoranthene, and phenanthrene) in foliage revealed that photodegradation of PAHs could potentially lead to rapid loss from leaves (31). In summary, we believe that the IG/CG ratios for the PAHs are too low to be interpretable as dry gaseous deposition velocities. The assumption that what is measured in the bulk deposition samples equals the cumulative fluxes into the canopy may not be valid for the PAHs. In this regard, it is noteworthy that the previous study neither reported IG/CG ratios or gaseous deposition velocities for the two benzo-flouranthenes, even though gaseous concentrations for these substances were reported (2). In fact, a vG for only a single Category B-PAH (chrysene/triphenylene) was reported, presumably because for the other PAHs IG did not contribute at least 20% to the total deposition flux below the canopy I and, therefore, would have resulted in IG/CG ratios with high uncertainties (2). Deriving Pseudo-Canopy/Air Partition Coefficient KPA. The IG/CG ratios derived for less chlorinated PCBs (Category A) in Table 2 decrease with increasing volatility, indicating that these congeners are approaching partitioning equilibrium between air and the forest canopy. In this case, IG/CG ratios can again not be interpreted as gaseous deposition velocities. However, a dimensionless pseudo-canopy/air partition coefficient (KPA) can be derived from these data (2). As most of the leaves had fallen to the ground in October and early November, KOA values at the seasonal temperature of 7 °C (i.e., average ambient temperature in October 2002) are employed. References used in the derivation of KOA values are listed in Table 2. Experimentally determined values were preferably employed, but for some PCB congeners KOA had to be calculated from QSPR equations (26). The relationship between log(IG/CG) and log KOA is plotted in Figure 2. Two segments are identified: equilibrium distribution between air and forest canopy for volatile PCBs (Category A) and kinetically controlled dry gaseous deposition for intermediate PCBs and PBDEs (Category B). A critical log KOA value of around 9.7 (calculated by inserting the average vG of 2.7 cm‚s-1 into eq 9 below), separates these two segments, which agrees with both theoretical prediction (3) and previous field measurement (2). 538

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 2, 2007

FIGURE 2. Relationship between log IG/CG and log KOA at ambient temperature (7 °C). The solid regression line in Category B is the average vG (2.7 cm‚s-1), whereas the line in Category A is derived from PCBs reaching an equilibrium between air and canopy. Triangles (4) identify PBDEs, diamonds (]) PCBs, and crosses (×) PAHs. The critical KOA in this study is slightly lower than the previously reported one (9.7 vs 9.9). This means that during the growing season, a PCB with a log KOA of 9.9 was approaching equilibrium with the German deciduous canopy, but not with the Canadian one. There are two possible explanations for this: First, the leaf-growing season in Canada may have been shorter than in Germany, and therefore, the time period available for deposition to the canopy is also shorter. Second, as discussed above, the gaseous deposition velocity to the Canadian deciduous canopy was slightly smaller, which implies that it would take longer to reach equilibrium.Notethatslightlydifferentambienttemperaturess the average temperature in October was 7 °C at the Canadian forest site but 11 °C in Germanyscannot be responsible, because the KOA in Figure 2 had been temperature adjusted. Linear regression was performed between log(IG/CG) and log KOA for those PCBs that approached equilibrium between air and canopy (Category A). The following regression equation is derived:

log [(IG/CG)/(cm‚s-1)] ) 0.67 log KOA - 6.08 (R2 ) 0.96) (9) Equation 9 is fairly similar to the regression equation reported previously (2).

log [(IG/CG)/(cm‚s-1)] ) 0.76 log KOA - 6.97

(10)

The lines described by eqs 9 and 10 were added to Figure 2, as were horizontal lines representing the kinetically controlled deposition velocities to the canopy (2.7 cm‚s-1 in this study and 3.6 cm‚s-1 in ref 2). The similarity between the two uptake profiles is remarkable and suggests that the uptake of SOCs by the two types of deciduous canopies is alike. In particular the log KOA threshold between Category A and B compounds is similar in the two studies. The pseudo-plant/air partition coefficient KPA is further derived from eqs 7 and 9. Time t refers to 6 months. V was not measured for the Canadian forest, but it had a similar leaf area index (approximately 5 m2‚m-2) as the German forest (M.S. McLachlan, pers. comm.). Therefore, the V of 0.0012 m3‚m-2 measured for the German deciduous forest (2) was also employed to derive a correlation between KPA and KOA here. Inserting above values into eqs 7 and 9, we obtain the following:

KPA ) 110 KOA0.67

(11)

Regressions for three forests are plotted in Figure 3. The average temperatures underlying these regressions are slightly different (7 °C in this study vs 11 °C in previous study). Comparison shows that the relationship for the deciduous

Acknowledgments We thank Michael S. McLachlan for numerous helpful discussions, and Bondi Gevao and John Deary for field sampling assistance. Funding from the Canadian Foundation for Climate and Atmospheric Sciences (CFCAS) is greatly appreciated.

Supporting Information Available Detailed description of bulk deposition sampling; quality assurance/quality control; assessing potential losses from the bulk deposition sampler; derived canopy interception I in the last 2 months; calculated IP and IG for PAHs and PBDEs; measured monthly deposition fluxes for individual PAHs, PBDEs, and PCBs at clearing and forest site. This material is available free of charge via the Internet at http:// pubs.acs.org. FIGURE 3. Relationship between log KPA and log KOA for three different forest canopies. KOA values were adjusted to the ambient temperature during fall (7 °C in this study and 11 °C in the case of the data from ref 2). The 1:1 line is plotted as a reference. canopy in Canada has a slightly shallower slope than the one in Germany. However, the difference in the linear regressions between log KPA and log KOA between variable deciduous forest canopies is small, and in particular much smaller than was observed for five different grassland species (32). This may be because potentially larger differences in the partitioning properties of individual tree species are averaged out when comparing partitioning properties of multi-species forest canopies. Comparison between two deciduous canopies. A dry particle-bound deposition velocity (diffusion/impaction) for PAHs of 0.11 cm‚s-1 reported here contrasts with a value of 0.73 cm‚s-1 measured in Germany. The latter is actually very close to the vP for PBDEs derived in this study (0.8 cm‚s-1). The difference in dry particle-bound deposition velocity could be related to different particle size distribution of PAHs in Germany and Canada. The dry gaseous deposition velocities for SOCs are similar between Canadian (2.7 ( 0.52 cm‚s-1) and German (3.6 ( 0.65 cm‚s-1) deciduous forest, although the former is slightly lower than the latter. Horstmann and McLachlan (2) attributed differences in the deposition velocities to a deciduous and coniferous forest canopy to “differences in the canopy turbulence and the mass transfer directly at the leaf/needle surface”. Similar factors can also explain the differences between the Canadian and German forest. Canopy turbulence may differ because of different meteorological conditions or different surface roughness. For example, the two deciduous canopies differ in terms of tree age, height and species composition. However, the deposition velocities of SOCs to a Canadian deciduous forest are of a similar order of magnitude as the ones reported previously for a German deciduous forest (2). In this study, we have confirmed the extraordinarily high gaseous deposition velocities of SOCs to a deciduous forest canopy, first reported by Horstmann and McLachlan (2) and used since to propose a significant filter effect of forests on a regional (12) and global scale (14). Despite differences in local climate and canopy composition and structure, the dry gaseous deposition velocities measured in Canada were remarkably similar to those reported in Germany, lending credibility to the suggestion that such high deposition velocities are common throughout boreal and temperate deciduous forests. This is particularly important in light of the fact that the gaseous deposition velocity to deciduous forests was identified as one of the most influential parameters in controlling the global forest filter effect for the SOCs (12, 14).

Literature Cited (1) Simonich, S. L.; Hites, R. A. Organic pollutant accumulation in vegetation. Environ. Sci. Technol. 1995, 29, 2905-2914. (2) Horstmann, M.; McLachlan, M. S. Atmospheric deposition of semi-volatile organic compounds to two forest canopies. Atmos. Environ. 1998, 32, 1799-1809. (3) McLachlan, M. S. Framework for the interpretation of measurements of SOCs in plants. Environ. Sci. Technol. 1999, 33, 17991804. (4) Collins, C.; Fryer, M.; Grosso, A. Plant uptake of non-ionic organic chemicals. Environ. Sci. Technol. 2006, 40, 45-52. (5) Nizzetto, L.; Cassani, C.; Di, Guardo, A. Deposition of PCBs in mountains: the forest filter effect of different forest ecosystem types. Ecotoxicol. Environ. Saf. 2006, 63, 75-83. (6) Horstmann, M.; McLachlan, M. S. Evidence of a novel mechanism of semi-volatile organic compound deposition in coniferous forests. Environ. Sci. Technol. 1996, 30, 17941796. (7) Jaward, F. M.; Di Guardo, A.; Nizzetto, L.; Cassani, C.; Raffaele, F.; Ferretti, R.; Jones, K. C. PCBs and selected organochlorine compounds in Italian mountain air: the influence of altitude and forest ecosystem type. Environ. Sci. Technol. 2005, 39, 34553463. (8) Simonich, S. L.; Hites, R. A. Importance of vegetation in removing polycyclic aromatic hydrocarbons from the air. Nature 1994, 370, 49-51. (9) Meijer, S. N.; Steinnes, E.; Ockenden, W. A.; Jones, K. C. Influence of environmental variables on the spatial distribution of PCBs in Norwegian and U.K. soils: implications for global cycling. Environ. Sci. Technol. 2002, 36, 2146-2153. (10) Hassanin, A.; Breivik, K.; Meijer, S. N.; Steinnes, E.; Thomas, G. O.; Jones, K. C. PBDEs in European background soils: levels and factors controlling their distribution. Environ. Sci. Technol. 2004, 38, 738-745. (11) Su, Y.; Wania, F.; Harner, T.; Lei, Y. D.; Shoeib, M. Temperature dependence of the air concentrations of polychlorinated biphenyls and polybrominated diphenyl ethers in a forest and a clearing. Environ. Sci. Technol. submitted. (12) Wania, F.; McLachlan, M. S. Estimating the influence of forests on the overall fate of semi-volatile organic compounds using a multimedia fate model. Environ. Sci. Technol. 2001, 35, 582590. (13) Wegmann, F.; Scheringer, M.; Mo¨ller, M.; Hungerbu ¨ hler, K. Influence of vegetation on the environmental partitioning of DDT in two global multimedia model. Environ. Sci. Technol. 2004, 38, 1505-1512. (14) Su, Y.; Wania, F. Does the forest filter effect prevent semi-volatile organic compounds from reaching the Arctic? Environ. Sci. Technol. 2005, 39, 7185-7193. (15) McLachlan, M. S.; Welsch-Pausch, K.; Tolls, J. Field validation of a model of the uptake of gaseous SOC in Lolium multiflorum (rye grass). Environ. Sci. Technol. 1995, 29, 1998-2004. (16) Neumann, H. H.; den Hartog, G.; Shaw, R. H. Leaf area measurements based on hemispheric photographs and leaflitter collection in a deciduous forest during autumn leaf-fall. Agric. For. Meteorol. 1989, 45, 325-345. (17) Buehler, S. S.; Hites, R. A. The Great Lakes’ Integrated Atmospheric Deposition Network. Environ. Sci. Technol. 2002, 36, 354A-359A. VOL. 41, NO. 2, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

539

(18) Su, Y.; Lei, Y. D.; Wania, F.; Shoeib, M.; Harner, T. Regressing gas/particle partitioning data for polycyclic aromatic hydrocarbons. Environ. Sci. Technol. 2006, 40, 3558-3564. (19) Harner, T.; Bidleman, T. F. Measurements of octanol-air partition coefficients for polycyclic aromatic hydrocarbons and polychlorinated naphthalenes. J. Chem. Eng. Data. 1998, 43, 40-46. (20) Xiao, H.; Wania, F. Is vapor pressure or the octanol-air partition coefficient a better descriptor of the partitioning between gas phase and organic matter? Atmos. Environ. 2003, 37, 28672878. (21) Lei, Y. D.; Chankalal, R.; Chan, A.; Wania, F. Supercooled liquid vapor pressures of the polycyclic aromatic hydrocarbons. J. Chem. Eng. Data. 2002, 47, 801-806. (22) Harner, T.; Shoeib, M. Measurements of octanol-air partition coefficients (KOA) for polybrominated diphenyl ethers (PBDEs): predicting partitioning in the environment. J. Chem. Eng. Data. 2002, 47, 228-232. (23) Wania, F.; Lei, Y. D.; Harner, T. Estimating octanol-air partition coefficients of nonpolar semi-volatile organic compounds from gas chromatographic retention times. Anal. Chem. 2002, 74, 3476-3483. (24) Ko¨mp, P.; McLachlan, M. S. Octanol/air partitioning of polychlorinated biphenyls. Environ. Toxicol. Chem. 1998, 16, 24332437. (25) Harner, T.; Bidleman, T. F. Measurement of octanol-air partition coefficients for polychlorinated biphenyls. J. Chem. Eng. Data. 1996, 41, 895-899. (26) Li, N.; Wania, F.; Lei, Y. D.; Daly, G. L. A comprehensive and critical compilation, evaluation, and selection of physicalchemical property data for selected polychlorinated biphenyls. J. Phys. Chem. Ref. Data 2003, 32, 1545-1590.

540

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 2, 2007

(27) Kaupp, H.; Towara, J.; McLachlan, M. S. Distribution of polychlorinated dibenzo-p-dioxins and dibenzofurans in atmospheric particulate matter with respect to particle size. Atmos. Environ. 1994, 28, 585-593. (28) Venkataraman, C.; Friedlander, S. K. Size distributions of polycyclic aromatic hydrocarbons and elemental carbon. 2. Ambient measurements and effects of atmospheric processes. Environ. Sci. Technol. 1994, 28, 563-572. (29) Kaupp, H.; McLachlan, M. S. Atmospheric particle size distributions of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) and polycyclic aromatic hydrocarbons (PAHs) and their implications for wet and dry deposition. Atmos. Environ. 1999, 33, 85-95. (30) Kaupp, H.; McLachlan, M. S. Distributions of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) and polycyclic aromatic hydrocarbons (PAHs) within the full size range of atmospheric particles. Atmos. Environ. 2000, 34, 73-83. (31) Wild, E.; Dent, J.; Thomas, G. O.; Jones, K. C. Real-time visualization and quantification of PAH photodegradation on and within plant leaves. Environ. Sci. Technol. 2005, 39, 268273. (32) Ko¨mp, P.; McLachlan, M. S. Interspecies variability of the plant/ air partitioning of polychlorinated biphenyls. Environ. Sci. Technol. 1997, 31, 2944-2948.

Received for review September 14, 2006. Revised manuscript received November 4, 2006. Accepted November 7, 2006. ES0622047