Measurements of Mercury in Dew: Atmospheric Removal of Mercury

The major source of exposure to humans and many other animal species is the ... During each of these campaigns, the dew collectors were placed on the ...
0 downloads 0 Views 98KB Size
Environ. Sci. Technol. 2002, 36, 2815-2821

Measurements of Mercury in Dew: Atmospheric Removal of Mercury Species to a Wetted Surface ELIZABETH G. MALCOLM AND GERALD J. KEELER* University of Michigan Air Quality Laboratory, Ann Arbor, Michigan 48109

The importance of dew in the mercury cycle was investigated during three sampling periods in the Great Lakes region and one in the Florida Everglades. Mercury concentrations ranged from 1.0 to 22.6 ng/L in dew. Deposition per dew event was, on average, lowest at a remote site on Lake Superior (0.31 ng/m2) and highest in the Florida Everglades (1.4 ng/m2). The estimated mercury deposition to the canopy associated with dew approximately equaled that of precipitation during the wintertime Everglades study. Relative to other trace elements (Mg, Ti, V, Mn, Ni, Cu, As, Sr, Cd, Sb, La, Ce, Pb), mercury was found to be more enriched in rain than dew, suggesting the importance of gas scavenging for precipitation. The fraction of mercury in dew from particulate deposition was estimated to average 40%, with the remaining contribution from reactive gaseous Hg (RGM). RGM, for which little reliable data exists, was measured in the Everglades and was significantly reduced at the start of a dew event, indicating pronounced removal of this soluble mercury species to wetted surfaces. The first estimates of RGM deposition velocities based on mercury flux measurements are reported here and range up to 1.6 cm/s.

Introduction Mercury (Hg) is a known toxin that has been implicated in the decline of the endangered Florida Panther (1), impairment of avian and plant reproductive success (2), and permanent neurological damage in humans (3). The major source of exposure to humans and many other animal species is the consumption of contaminated fish. Atmospheric deposition of naturally and anthropogenically emitted mercury has been found to be a major source of fish contamination in both polluted and “pristine” watersheds throughout the world (4). A detailed understanding and quantification of atmospheric deposition is therefore necessary to devising a solution to the problem of mercury pollution. Deposition of several atmospheric chemical species has been found to be enhanced when surfaces are wetted by dew (5). Dew forms at night when the emission of infrared radiation from surfaces causes them to cool below the dew point. Atmospheric water vapor in contact with this cooled surface condenses and forms a water film. This condensation generally occurs on clear, atmospherically stable nights when wind speeds are minimal. Under these conditions, atmospheric gases and particles may be transferred to the dew where they are more readily absorbed or retained than under dry conditions, resulting in enhanced deposition. * Corresponding author phone: (734) 936-1836; fax: (734) 9367283; e-mail: [email protected]. 10.1021/es011174z CCC: $22.00 Published on Web 06/01/2002

 2002 American Chemical Society

There is limited research devoted to the role of dew in the cycling of trace elements, such as Hg. Studies by Pierson et al. (6), which measured deposition of trace elements to dew, found a dependence of deposition velocity on aerosol size fraction. Elements predominantly associated with fine particles, Pb, Se, Br, and V had the lowest deposition velocities ( 0.1). This result was unexpected, because Hg deposition to dew and frost will likely differsliquid water in dew will potentially increase the absorption of soluble gases (such as RGM) or the larger surface area of frost will increase deposition. While the deposition process is likely different for dew and frost, this difference may be minimal as compared to other sources of variability. Changes in ambient atmospheric Hg concentrations and atmospheric turbulence will probably have a greater influence on Hg deposition to dew or frost. The lack of significant difference for trace elements, which are associated with aerosols and emitted from a variety of sources, supports this hypothesis. Two previous studies (20, 21) also found no significant difference between dew and frost with the exception that frost had elevated Na+ and Cl- concentrations that were attributable to wintertime use of road salt and not general differences in deposition processes (21). More extensive field measurements and controlled laboratory experiments will likely be necessary to determine differences in gaseous and aerosol deposition to dew versus frost. Consequently, all frost and dew events have been combined in subsequent analysis and all referred to as “dew”. A total of 31 dew events were sampled during the four dew investigations. There was a wide variation in the concentrations of mercury in dew water from event to event at each site (Figure 1). In comparison to Dexter 1997 and 1998 (9.8 ( 5.4 (standard deviation), n ) 13, and 4.7 ( 1.7 ng/L, n ) 6) and the Everglades (6.0 ( 4.9 ng/L, n ) 5), the volume-weighted mean concentration was lowest at the most remote sitesEagle Harbor (2.8 ng/L ( 2.4, n ) 7). In Table 1, measured mean dew concentrations are compared with concentrations of precipitation from both the dew intensive and a 3-month period encompassing the dew sampling. Dew concentrations were generally lower than precipitation concentrations (by less than 1 order of magnitude), except at Dexter in 1997, when they were comparable (8.9 and 9.8 ng/L). During the Everglades study, precipitation only occurred once, and no measurements were made before or after the dew study. The volume-weighted mean Hg dew concentrations from the Everglades of 6.0 ng/L is similar to the concentration of 7.0 ng/L for the rain event. In January, February, and March of 1995 and 1996, monthly precipitation concentrations were similar to our study, varying between 5 and 15 ng/L with a volume-weighted mean of 8.5 ng/L (22). One important factor that may cause differences in con-

TABLE 1. Comparison of Volume-Weighted Mean Hg Concentrations (ng/L) Measured in Dew and Precipitation

site

date

Eagle Harbor

8/97

Dexter

10/97

Dexter

9/98

Everglades

3/99

dew mean 2.8 ( 2.4 n)7 9.8 ( 5.4 n ) 13 4.7 ( 1.7 n)6 6.0 ( 4.9 n)5

precipitation events 11.1 24.2, 4.1, 6.5 9.0, 19.1 7.0

3-month precipitation mean 15.2 ( 10.0 n ) 12 8.9 ( 6.7 n ) 12 13.0 ( 14.8 n ) 19 NA

centrations between dew and precipitation is that Hg in dew originates from the nocturnal surface layer, while Hg in precipitation may originate from in-cloud scavenging as well as the from the boundary layer below cloud. The two processes that determine Hg concentrations in dew are hydrological deposition of water and Hg dry deposition onto the dew surface. A close investigation of these two processes is essential to a thorough understanding of factors influencing the measured concentrations. The mean hydrological input per dew event was highest in the Everglades (270 ( 130 mL/m2). The mean water deposition per event in Michigan was significantly higher (p < 0.01) during August (Eagle Harbor: 170 ( 50 mL/m2) and September (Dexter: 150 ( 20 mL/m2) than October/ November (Dexter: 100 ( 40 mL/m2). This seasonal difference is likely due to higher absolute humidity in summer. Frequency of events and hydrological input are expected to vary seasonally, as well as geographically, because length of night, overnight cloud cover, humidity, and wind speeds in the surface layer will all affect the formation, duration, and volume of dew. As with mercury concentrations in dew, nocturnal Hg deposition was quite variable. Eagle Harbor had the lowest mean deposition per event (total Hg deposited during one night of dew; 0.31 ( 0.2 ng/m2) and the Everglades the highest (1.4 ( 0.8 ng/m2). Hg deposition was higher in the Everglades than in Dexter, despite higher Hg concentrations in dew at Dexter. This is the result of larger hydrological deposition in Florida, which led to increased dilution. At Dexter, deposition per event was higher on average, during October 1997 (0.91 ( 0.6 ng/m2) than September 1998 (0.68 ( 0.3 ng/m2), although not statistically significant (p > 0.05). VOL. 36, NO. 13, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2817

Dew samples were also analyzed for several trace elements (Mg, Ti, V, Mn, Ni, Cu, As, Sr, Cd, Sb, La, Ce, Pb). Unlike Hg, these elements in dew result only from particulate and not gaseous deposition because they are only associated with aerosols in the ambient atmosphere. Mean event deposition of 13 trace elements is presented in Table 1 in the Supporting Information. Trace element deposition was similar to Hg deposition in that the lowest deposition was at Eagle Harbor and the highest at Dexter in 1997 and the Everglades. These variations have likely resulted from the influence of local and regional sources of particular elements at the different sampling sites or temporal differences in atmospheric turbulence. Deposition of elements with anthropogenic sources (Cu, As, Pb) (23) varied by at least an order of magnitude between sites, with elevated levels at Dexter during Fall 1997 and in the Everglades. An element of predominantly crustal origin, Ti, however, varied only by a factor of 2. Trace element concentrations in both dew and precipitation ranged over 1 order of magnitude. The ratio of the mean dew concentration to the mean 3-month rain concentration varied for trace elements from 0.2 (Cd at Eagle Harbor) to 38 (Mg at Dexter). During the Dexter studies, this ratio was nearly always greater than 1, with a mean of 17 and range of 3-38 in 1997 and a mean of 7 and range of 0.8-38 in 1998. At Eagle Harbor, the ratio was generally smaller, with a mean of 1.8, ranging from 0.2 to 8. Many factors will affect this ratio, including deposition rates to dew and in-cloud scavenging rates. Concentrations of the elements available for removal by dew or clouds will also vary spatially and temporally. The ratio of Hg concentrations in dew to rain were lower than for other elements at 0.2, 1.1, and 0.4 for the Eagle Harbor and Dexter 1997 and 1998, respectively. The lower ratio for Hg suggests that, relative to the other elements, Hg is more concentrated in rain than dew. One possible explanation is that in-cloud scavenging of gaseous Hg, particularly RGM, may be very efficient. More research on processes controlling incorporation of Hg into rain and dew are needed to investigate these results. Although concentrations of trace elements were higher in dew than rain, average deposition from a rain event was 7-550 times greater than deposition to the collector during a dew event. The relative importance of deposition of dew versus rain will depend on the relative frequency of dew and rain events and the area of the surface to which dew deposits. During Dexter 1997, the elements with the smallest ratio of rain to dew deposition were predominantly crustal elements (Mg ) 7, Ti ) 10, Mn ) 12, Sr ) 8, La ) 12). This can be explained by higher dry deposition velocities caused by the larger size of crustal versus anthropogenic aerosols. Consideration of the Canopy. Given that dew formation is an atmospheric deposition process, total deposition to an ecosystem is dependent on the canopy surface characteristics. Northern hardwood and conifer forests often have a leaf area index (LAI, a measure of canopy surface area per unit area of ground) of 4-6 (24, 25). Consequently, dew deposition to vegetation will be greater than that measured using the flat surrogate surfaces in this study. To estimate the increased deposition due to the larger surface area of the canopy as compared to our collector, measured deposition was multiplied by LAI (Table 2). LAI measured in the sawgrass-cattail canopy of our Everglades site ranged from 2.1 to 6.7 (26). Unlike deciduous trees, sawgrass and cattails have vertical leaves, so the LAI likely underestimates leaf surface area for grass and wetland habitats. For the Michigan studies, Hg deposition to the canopy wetted by dew (adjusted for LAI) was estimated to be about 3-30% of the value of wet deposition. When LAI is considered for the Everglades, the mercury deposition associated with dew (15-48 ng/m2) was similar to that of wet deposition (52 ng/m2) during the 6-day measurement period. By comparison, deposition measure2818

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 13, 2002

TABLE 2. Mean Hg Deposition (ng/m2) per Dew Event Adjusted for Canopy Surface Areaa site Harborb

Eagle Dexterb Dexterb Evergladesc

date

deposition

8/97 10/97 9/98 3/99

1.2-1.8 3.6-5.4 2.7-4.1 3.0-9.6

a Adjustment made by multiplication of mean deposition to collector by leaf area index (LAI). b LAI ) 4-6. c LAI ) 2.1-6.7.

TABLE 3. Ambient Atmospheric Mercury Concentrations during Dew Sampling Campaigns

site

date

n

total gaseous mercury (ng/m3)

Eagle Harbor Dexter Dexter Everglades

8/97 10/97 9/98 3/99

7 13a 5a 5a

1.2 ( 0.2 1.5 ( 0.2 1.5 ( 0.1 1.9 ( 0.2

particulate mercury (pg/m3) 8(4 14 ( 8 13 ( 7 25 ( 11

a n ) number of particulate samples and number of nights that total gaseous mercury was continuously measured in 5 min samples.

ments from 1995 and 1996 of Guentzel et al. (22) averaged 60 and ranged from 13 to 109 ng/m2 for the months of January, February, and March and were significantly higher during the summertime wet season. This suggests that dew is an important deposition pathway to the Everglades wetland ecosystem during the dry season. Total deposition to forest canopies in North America has been estimated from measured throughfall flux, using the assumption that the source of Hg in throughfall is precipitation plus dry deposition washed off of leaves (27). Using this method, mean dry deposition to mixed hardwood forests was estimated as 0.9 (0.02-2.86), 2.5 (0.7-8.6), and 1 (0.16.6) ng/m2/h for the Walker Branch Tennessee, Lake Champlain, and Lake Huron Watersheds, respectively (27-29). Hg deposition to dew from Michigan was multiplied by the LAI of each watershed and compared to the total dry deposition estimates from throughfall. Results suggest that deposition to dew contributes on the order of 10-30%, 2-7%, and 5-23% of total dry deposition to each of these North American forests, respectively. This is reasonable because larger deposition of particulate Hg is expected to occur during daytime as a result of higher turbulence and wind speeds during the day. The uncertainties in this estimate are quite large because the dew and throughfall studies did not perfectly coincide geographically or temporally. Ambient Atmospheric Mercury Concentrations. Ambient atmospheric concentrations, in part, determine dry deposition rates and, therefore, will influence mercury deposition rates to dew. The influence of geographic location on the differences observed in dew concentration and deposition corresponded with the regional differences in measured ambient atmospheric mercury concentrations. Total gaseous elemental mercury (TGM) and particulate mercury (Hg(p)) were lowest at Eagle Harbor (1.2 ng/m3 and 7.6 pg/m3) and highest in the Everglades (1.9 ng/m3 and 25 pg/m3), which was reflected in Hg deposition to dew (Table 3). Past research indicates that, of the three sites, Eagle Harbor, the most remote site, was affected least by local anthropogenic mercury sources (30). Reactive gaseous mercury (RGM) was only measured during the Everglades investigation and for 2 days during the Dexter 1998 investigation. These measurements, along with past studies of other mercury species at the sites and documented emission of known mercury sources, suggest that concentrations of RGM at Eagle Harbor are typically very low.

FIGURE 2. Modeled contribution of particulate mercury (Hg(p)) deposition to measured total mercury (Hg(t)) deposition in dew. Error bars indicate estimated uncertainty for particulate mercury contributions. Particulate Mercury Deposition to Dew. The source of mercury in dew is a combination of deposited particulate, elemental, and reactive gaseous forms. The relative contribution of each of these types of Hg will vary by event dependent on meteorological conditions, ambient atmospheric concentrations, particulate characteristics, and RGM speciation. Quantifying the relative contribution that the three atmospheric species make to Hg levels in dew requires knowledge of particulate or gaseous deposition rates. Particulate deposition velocities for the Dexter 1997 and Everglades dew events were calculated using a model by Sehmel and Hogdson (31). Input parameters required for this calculation are mean wind speed and temperature, which were measured, and particle size (0.68 and 3.8 µm) and particle density (2 g/cm3), which were estimated. The modeled deposition velocities ranged from 0.003 to 0.008 cm/s for fine Hg aerosol and 0.09 to 0.12 cm/s for coarse Hg aerosol. These velocities agree well with those predicted by other researchers. Shannon and Voldner (32), for example, estimate seasonally averaged deposition velocities of 0.24 cm/s for summer and 0.14 cm/s for autumn, both of which should be higher than our nighttime values because turbulence is higher during the day. Particulate mercury deposition was then calculated according to these deposition velocities and measured particulate mercury concentrations. Deposition of coarse Hg particles dominated total Hg(p) deposition to dew (>85%) because of higher deposition velocities for coarse aerosols. Figure 2 shows a comparison of the estimated deposition of Hg to dew from particulate and the total deposition of Hg to dew measured with the dew collectors. When this methodology is applied to the Dexter and Florida studies, it was estimated that particulate deposition accounts for, on average, 40% of the mercury deposition to dew (σ ) 30%), with contributions ranging from 14% to 116% per event. Reactive Gaseous Mercury Deposition to Dew. Although RGM is less abundant than Hg0(g), it is much more reactive and soluble (8) and, thus, is expected to contribute significantly to total mercury deposition, especially to water and wetted surfaces. Some of the first continuous measurements of RGM were made in the Florida Everglades during our dew sampling research. RGM exhibited a diurnal behavior, which

has also been observed by other authors (33, 34). This behavior could be caused by a daytime source of RGM and fast removal at night. As Figure 3 illustrates, ambient RGM concentrations decreased sharply at the start of each dew event, as indicated by surface wetness values exceeding zero (marked by arrows). RGM concentrations were significantly reduced immediately after dew formation (p ) 0.03). This suggests that the presence of dew increased the rate of dry deposition, thereby depleting RGM, and reducing its atmospheric concentrations within a few meters of the surface. Unlike Hg0(g), which is known to be easily re-emitted after deposition, Hg(p) and RGM are assumed to remain relatively immobile after deposition to a water surface (8). This suggests that the total Hg measured in our dew samples included virtually all of the nocturnally deposited Hg(p) and RGM. Assuming that the source of Hg in dew is a combination of Hg(p) and RGM, our estimates of particulate deposition indicate that RGM contributes up to 86% and an average of 60% of the remaining mercury in dew. Contributions in this range require deposition velocities on various nights in the Everglades from 0 to 1.6 cm/s (means of 0, 1.2, 1.4, 0.9, and 0.5 cm/s for each event). On 2/27/99, the deposition velocity is calculated as zero because the predicted Hg(p) deposition can account for all the Hg measured in dew. Previous to this study, there has been little research on RGM deposition velocities. Pai et al. (35) found that total Hg dry deposition levels predicted by deposition and transport models are highly sensitive to the chosen RGM deposition velocity. Previous modelers have assumed that the deposition velocity for RGM equals that of other reactive gaseous species, such as HNO3 (36). These velocities are typically in the same range as our indirectly measured estimate. For example, Bullock et al. (26) assumed 0.3 cm/s at night, and Shannon and Voldner (32) assumed 0.72-1.4 cm/s over land. To understand the important processes controlling dry deposition, a resistance model is often conceptualized (37) which breaks dry deposition into three steps. The first is aerodynamic transport of the gas through the boundary layer to the quasi-laminar sublayer of air just above the surface; second is molecular diffusion through this sublayer to the dew surface; and third is uptake by the surface, in this case, the dew. Henry’s law gives us a way to evaluate this last VOL. 36, NO. 13, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2819

FIGURE 3. Reactive gaseous mercury (RGM) concentration versus presence of dew measured as surface wetness in the Florida Everglades. Arrows and surface wetness values above zero indicate the beginning of a dew event. RGM values are for 2-h integrated samples. Surface wetness values are proportional to sensor resistance. resistance. HgCl2 is soluble with a large Henry’s Law constant of 1.4 × 106 M/atm at 25 °C (38). When it is assumed that the first two transport resistances were not limiting, HgCl2 concentrations of approximately 170 ng/L are predicted for the Everglades samples. Under low-temperature conditions, such as in Dexter 1997, Henry’s Law predicts that even more HgCl2 will partition into dew (about 400 ng/L at 10 °C). This is much higher than the measured total mercury concentrations in the dew of 1.8-14 ng/L, indicating that the dew is not saturated with HgCl2 and one of the first two resistances limits dry deposition. This is consistent with model predictions (39) for highly soluble gases. The second resistance, transport across the quasi-laminar sublayer, is generally fast and does not often limit deposition (37). The limiting step is likely the first resistance, aerodynamic transport, considering that dew forms on atmospherically stable nights when wind speeds and, hence, turbulent mixing are low. This resistance model suggests that virtually all of the RGM, which is transported to the surface, will partition into the dew, explaining the depletion of RGM measured by our instruments at approximately 1-m above the canopy. Elemental Gaseous Mercury Deposition to Dew. The deposition of Hg0(g) to dew is expected to be minimal as compared to Hg(p) and RGM. Hg0(g) has a low solubility with a Henry’s Law constant of 0.11 M/atm at 25 °C (38), which predicts that, under ambient conditions, only 4 × 10-5 ng/L will partition into the dew. At lower temperatures, this concentration is still quite low (8 × 10-5 ng/L at 5 °C). The Hg0(g) flux was estimated from gradient measurements made during two nights when dew formed at the Everglades sampling site. No downward (negative) fluxes coincide with the onset of surface wetness. The facts that our measured Hg0 fluxes are near zero when dew is present and that Hg(p) and RGM deposition reasonably account for all of the Hg deposition to dew in this study suggest that Hg0(g) deposition’s contribution to total Hg in dew is minimal. By applying the modified Bowen ratio approach to measured Hg gradients, Lindberg et al. (40) calculated Hg0(g) fluxes over forest canopies in Tennessee and Sweden. Fluxes measured when the canopy was wet with dew or rain (n ) 4) resulted in a downward Hg0(g) deposition velocity of 1.3 ( 1.8 cm/s. Lindberg et al. hypothesized that Hg0 was being oxidized in the dew solution to Hg2+(aq), thus enhancing the deposition of the less soluble Hg0 when the canopy was wet. Using the deposition velocity of 1.3 cm/s applied by Lindberg 2820

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 13, 2002

et al. will produce a 3 orders of magnitude overestimation of total Hg concentration in dew from our study, suggesting that the previous Hg0(g) deposition estimate is much too high and not broadly applicable to other locations or studies. Another plausible explanation for this discrepancy is that the Hg gradient reported by Lindberg et al. was dominated by an RGM and not a Hg0(g) gradient, which is supported by research indicating that gold bead traps can capture both RGM and Hg0(g) (41). This is plausible because the measured gradients were about 10-40 pg/m3 and the average RGM concentration previously measured at this site is 60 pg/m3 (33).

Future Research As the first investigation of Hg in dew, this study demonstrated that dew can be an important deposition pathway for atmospheric Hg, thus revealing the great need for additional dew research. Long-term measurements are necessary (1) to determine if mercury concentrations and deposition associated with dew vary seasonally and (2) to quantify yearly mercury inputs from dew. Because dew formation and ambient mercury concentrations vary geographically, this input is expected to be highly variable by location. Processoriented studies are necessary to investigate dew on real surfaces, such as leaves, to determine the fate of deposited mercury after sunrise when the dew water evaporates, and to determine what fraction may be washed off as throughfall or deposited as litterfall. Because chemical reactions have been found to occur in dew for nitrogen species (42), the effect of aqueous Hg chemistry within dew and its effect on fluxes should be investigated. Last, future research is needed to investigate the applicability and consequence of our findings that RGM is significantly depleted during dew events and that most of the Hg in dew is from RGM and Hg(p) deposition.

Acknowledgments This manuscript is dedicated to Dr. William R. Pierson, friend and mentor, and would not have been possible without his inspiration, encouragement, and pioneering research on dew chemistry. Working with Bill in the laboratory and field always meant a level of excellence that few of us live up to and that all of us strive to attain. James Barres at the University of Michigan Air Quality Lab (UMAQL) provided indispensable

assistance with the design and execution of the sampling methods. Several additional researchers provided expertise and assistance: Dr. Frank Marsik (UMAQL) with micrometeorology, Dr. Alan Vette (UMAQL) with Hg0 gradients, Dr. Nicola Pirrone (UMAQL) with particulate deposition, Scott Deboe (UMAQL) with data processing, Dr. Matthew Landis (EPA) and Dr. Timothy Dvonch (UMAQL) with RGM, Dr. Marion Hoyer (UMAQL) for methods development, and Dr. Steve Lindberg and Kelly Roy (ORNL) for the LAI data. Additional field assistance was provided by several students and staff of the UMAQL, in particular Yolada Yaiparyoon, Maureen Connell, and Michelle Debuck. Robert Stevens (FL DEP), Tom Atkinson (FL DEP), and Larry Fink (SFWMD) also helped with logistics and planning during the Everglades campaign. Financial support was partially provided by the Lake Superior Basin Trust Fund and the Florida Department of Environmental Protection.

Supporting Information Available Map of sampling sites and table of trace element deposition per dew event for individual studies. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Facemire, C.; Augspurger, T.; Bateman, D.; Brim, M.; Conzelmann, P.; Delchamps, S.; Douglas, E.; Inmon, L.; Looney, K.; Lopez, F.; Masson, G.; Morrison, D.; Morse, N.; Robison, A. Water Air Soil Pollut. 1995, 80, 923-932. (2) Meyer, M. W.; Evers, D. C.; Daulton, T.; Braselton, W. E. Water Air Soil Pollut. 1995, 80, 871-880. (3) Clarkson, T. W. Environ. Health Perspect. 1993, 100, 31-38. (4) Expert Panel on Mercury Atmospheric Processes. Mercury Atmospheric Processes: a Synthesis Report; EPRI/TR-104214; Workshop Proceedings, 1994. (5) Mulawa, P. A.; Cadle, S. H.; Lipari, F.; Ang, C. C.; Vandervennet, R. T. Atmos. Environ. 1986, 20, 1389-1396. (6) Pierson, W. R.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Norbeck, J. M. J. Geophys. Res. 1986, 91, 4083-4096. (7) Pierson, W. R.; Brachaczek, W. W.; Japar, S. M.; Cass, G. R.; Solomon, P. A. Atmos. Environ. 1988, 22, 1657-1663. (8) Schroeder, W. H.; Munthe, J. Atmos. Environ. 1998, 32, 809822. (9) Pleijel, K.; Munthe, J. Atmos. Environ. 1995, 29, 1441-1457. (10) Lindberg, S. E.; Meyers, T. P.; Taylor, G. E.; Turner, R. R., Jr.; Schroeder, W. H. J. Geophys. Res. 1992, 97D2, 2519-2528. (11) Rossmann, R.; Barres, J. A. Sci. Total Environ. 1992, 125, 405415. (12) Lindberg, S. E.; Vette, A. F.; Miles, C.; Schaedlick, F. Biogeochemistry 2000, 48, 237-259. (13) University of Michigan Air Quality Laboratory. Determination of Metals in Ambient Aerosols, Dry Deposition, Precipitation, Runoff, and Lake and River Waters by Inductively Coupled Plasma Mass Spectrometry (ICP-MS); 1995. (14) Landis, M. S.; Keeler, G. J. Environ. Sci. Technol. 1997, 31, 26102615. (15) Burke, J.; Hoyer, M.; Keeler, G.; Scherbatskoy, T. Water Air Soil Pollut. 1995, 80, 353-362. (16) Keeler, G. J.; Glinsorn, G.; Pirrone, N. Water Air Soil Pollut. 1995, 80, 159-168.

(17) Stevens, R. K.; Schaedlich, F. A.; Schneeberger, D. R.; Prestbo, E.; Lindberg, S.; Keeler G. 5th International Conference Mercury As a Global Pollutant Book of Abstracts; CETEM-Center for Mineral Technology: Rio de Janeiro, 1999; p 7. (18) Marsik, F. M.; Keeler, G. J.; Malcolm, E. G.; Dvonch, J. T.; Barres, J. A.; Lindberg, S. E.; Zhang, H.; Stevens, R. K.; Landis, M. L. 11th International Conference on Heavy Metals in the Environment, 2000. (19) Meyers, T. P.; Hall, M. E.; Lindberg, S. E.; Kim, K. Atmos. Environ. 1996, 30, 3321-3329. (20) Foster, J. R.; Pribush, R. A.; Carter, B. H. Atmos. Environ. 1990, 24A, 2229-2236. (21) Wagner, G. H.; Steele, K. F. J. Geophys. Res. 1992, 97, 2059120597. (22) Guentzel, J. L.; Landing, W. M.; Gill, G. A.; Pollman, C. D. Environ. Sci. Technol. 2001, 31, 863-873. (23) Cheng, M. D.; Hopke, P. K. Atmos. Environ. 1989, 23, 13731384. (24) Jose, S.; Gillespie, A. R. For. Sci. 1997, 43, 56-64. (25) Gower, S. T.; Vogel, J. G.; Norman, J. M.; Kucharik, C. J.; Steele, S. J.; Stow, T. K. J. Geophys. Res. 1997, 102, 29029-29041. (26) Bullock, O. R., Jr.; Benjey, W. G.; Keating, M. H. The modeling of regional-scale atmospheric mercury transport and deposition using RELMAP. In Atmospheric Deposition of Contaminants to the Great Lakes and Coastal Waters; Baker, J. E., Ed.; SETAC Press: Pensacola, FL, 1997; pp 323-347. (27) Lindberg, S. E.; Owens, J. G.; Stratton, W. J. Mercury Pollution: Integration and Synthesis; 1994; 261-271. (28) Rea, A. W.; Keeler, G. J.; Scherbatskoy, T. Atmos. Environ. 1996, 30, 3257-3263. (29) Rea, A. W.; Lindberg, S. E.; Keeler, G. J. Atmos. Environ. 2001, 35, 3453-3462. (30) Burke, J. M. Ph.D. Dissertation, University of Michigan, 1998. (31) Sehmel, G. A.; Hodgson, W. H. Model for Predicting Dry Deposition of Particles and Gases to Environmental Surfaces; PNL-SA-6721; Battelle Pacific Northwest Laboratories: Richland, WA, 1978. (32) Shannon, J. D.; Voldner, E. C. Atmos. Environ. 1995, 29, 16491661. (33) Stratton, W. J.; Lindberg, S. E. Water Air Soil Pollut. 1995, 80, 1269-1278. (34) Sheu, G.; Mason, R. P. Environ. Sci. Technol. 2001, 35, 12091216. (35) Pai, P.; Karamchandani, P.; Seigneur, C.; Allan, M. A. J. Geophys. Res., [Atmos.] 1999, 104, 13855-13868. (36) Petersen, G.; Iverfeldt, A.; and Munthe, J. Atmos. Environ. 1995, 29, 47-67. (37) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change; John Wiley and Sons: New York, 1998. (38) Lin, C. J.; Pehkonen, S. O. Atmos. Environ. 1999, 33, 2067-2079. (39) Chameides, W. L. J. Geophys. Res. 1987, 92, 11895-11908. (40) Lindberg, S. E.; Hanson, P. J.; Meyers, T. P.; Kim, K. H. Atmos. Environ. 1998, 32, 895-908. (41) Sommar, J.; Lindqvist, O.; Stro¨mberg, D. J. Air Waste Manage. Assoc. 2000, 50, 1663-1666. (42) Takenaka, N.; Suzue, T.; Ohira, K.; Morikawa, T.; Bandow, H.; Maeda, Y. Environ. Sci. Technol. 1999, 33, 1444-1447.

Received for review July 30, 2001. Revised manuscript received April 1, 2002. Accepted April 16, 2002. ES011174Z

VOL. 36, NO. 13, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2821