Tracking Toxaphene in the North American Great ... - ACS Publications

In North America it has been detected as a major contaminant in Great Lakes fish and in Arctic aquatic life (5−11). Although the use of toxaphene ha...
0 downloads 0 Views 776KB Size
Download PDF
Environ. Sci. Technol. 2005, 39, 8123-8131

Tracking Toxaphene in the North American Great Lakes Basin. 1. Impact of Toxaphene Residues in United States Soils JIANMIN MA,* SRINIVASAN VENKATESH, YI-FAN LI, AND SREERAMA DAGGUPATY Air Quality Research Branch, Meteorological Service of Canada, 4905 Dufferin Street, Toronto, Ontario M3H 5T4, Canada

A coupled atmospheric transport model was employed to study six scenarios to assess the contribution of reemission and long-range transport of toxaphene from different sources in the United States to its environmental fate in the Great Lakes ecosystem in the year 2000. Modeled air concentrations at the first model level (1.5 m) range from less than 5 pg m-3 over the upper Great lakes (Lakes Superior and Huron) to several tens of picograms per cubic meter over the lower Great Lakes (Lakes Erie and Ontario) in the summer but drop off to the range from 0.05 to 2 pg m-3 in the wintertime. The modeled toxaphene depositions to the lakes suggest a decreasing trend from the mid-1990s to 2000. Modeling results showed that, on an annual basis, for the Great Lakes basin as a whole, the southeast U.S. sources made the largest contribution to the toxaphene air concentrations and dry and wet depositions at 72%, 78%, and 88% respectively. The model results also showed that a significant proportion of these contributions occur during relatively short episodic events due primarily to the interseasonal changes in atmospheric circulation patterns.

Introduction Toxaphene is a complex mixture of chlorinated hydrocarbons produced by the chlorination of camphene. Globally, it was one of the most heavily used organochlorine pesticides (as an insecticide) (1). In the United States alone, the total usage of toxaphene from 1947 to 1986 (although usage was banned in 1982, there was still some residual usage up to 1986) was about 490 kt. This renders the United States the largest user of toxaphene in the world (2). For the year 2000, estimated emissions of toxaphene from U.S. soils to the air have been as high as 360 t (2, 3). Given its extreme toxicity (4), the environmental fate of toxaphene has been a major concern globally and it has been listed as one of the “Dirty Dozen” toxics to be eliminated under the Stockholm Convention on persistent organic pollutants. In North America it has been detected as a major contaminant in Great Lakes fish and in Arctic aquatic life (5-11). Although the use of toxaphene has been banned in North America since the earlier 1980s, its concentration in fish from all five Great Lakes still remains above health advisory limits, resulting in restrictions on consumption of fish from the Great Lakes (10-12). * Corresponding author phone: (416) 739-4857; fax: (416) 7394288; e-mail: [email protected]. 10.1021/es050945m CCC: $30.25 Published on Web 09/30/2005

Published 2005 by the Am. Chem. Soc.

Considerably high toxaphene air concentrations were detected over the Great Lakes basin, even around Lake Superior, where toxaphene was not used extensively (13, 14). Toxaphene air concentrations measured in 1998 in the Great Lakes basin (14) show values ranging from 1 to 100 pg m-3. Air concentrations measured over Lake Michigan (15) in 2000 ranged from 1 to 30 pg m-3. Over Lake Ontario, also in 2000, measured air concentrations ranged from 10 to 54 pg m-3 (personal communication with Liisa Jantunen). The toxaphene air sampling at Point Petre (PPT) on the north shore of Lake Ontario by the Organics Analysis Laboratory of the Meteorological Service of Canada (MSC) (16) revealed relatively lower concentrations ranging from 0.5 to 13 pg m-3 in 1999 and 2000. Toxaphene usage in the Great Lakes basin is estimated to be only about 1% of the total for North America (14). In Canada, toxaphene is neither manufactured nor registered for use, so that its usage in Canada was unspecified (1). This suggests that contamination in the Great Lakes and Arctic by toxaphene may be not a local issue but attributable to its volatilization from reservoirs accumulated from past applications, followed by long-range transport on continental to global scales (17-19). Given that the United States, especially the southern United States, was the largest user of toxaphene in the world before the mid-1980s, and that a large amount of residue still persists in agricultural and nonagricultural soils (2, 3) in these areas, this region is likely a major source for toxaphene in the Great Lakes basin and the Arctic (3). Efforts have been made to estimate the impacts/fate of toxaphene emissions from North American sources and their subsequent transport in the Great Lakes basin by both field measurements (15, 18, 20-24) and mathematical simulation techniques (17, 25). These studies suggested that the atmospheric transport from the southern United States was a primary pathway of toxaphene to the Great Lakes. However, in a recent investigation of toxaphene budget in North America, using a continental-scale mass balance/contamination/fate model, MacLeod et al. (14) came to the conclusion that the contribution of toxaphene to the Great Lakes basin from local sources, though they were only 1% of the total for North America, was almost as much as that due to transport from other sources across North America. The previous numerical investigations for the transport of toxaphene to the Great Lakes (14, 17, 25) used trajectory and mass balance models to estimate the fate of toxaphene and its mass balance between soil, water, and air. These models provide a framework of temporal trend and fate of toxaphene but they treat the heterogeneous and complex atmospheric environment as a homogeneous compartment with uniform properties. Since the atmosphere is a major pathway for long-range transport of a pollutant, impacts of realistic atmospheric conditions on the budget of toxaphene in the atmosphere should be taken into account in any numerical model. These may include the vertical structure of the atmosphere, such as wind shear and turbulent exchange, and the changes in atmospheric waves and wind system from diurnal to seasonal scales. These changes in atmospheric circulation will inevitably affect atmospheric transport and re-emission from contaminated terrestrial surfaces, and consequently the temporal and spatial distributions of a pesticide. In general, and especially during wintertime when atmospheric long waves [a wave in the major belt of westerlies, (26); for details readers are referred to the companion paper (27)] are active, westerly winds prevail in the free atmosphere. Hence, pollutants from the southeast VOL. 39, NO. 21, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8123

United States, the largest toxaphene usage area (2), cannot be transported efficiently northward to the Great Lakes basin. However, during spring and autumn, the transition seasons between winter and summer, the atmospheric long waves are perturbed frequently by strong meridional exchanges of cold and warm air masses between low and high latitudes, thus providing a vehicle for meridional transport of pollutants. During summertime, because the changes in atmospheric motion are subject primarily to local surface heating and cooling, strong continent-scale transports are unlikely to occur frequently. These uncertainties may pose further questions: if the atmospheric level of toxaphene over the Great Lakes depends largely on long-range transport, to what extent do the southern U.S. soils, as the major reservoir of this compound in North America, influence the magnitude of toxaphene air concentration and loadings to the lakes? Does such long-range transport exhibit temporal variations on a seasonal basis? How significant is the influence of the sources proximate to the Great Lakes on the changes in toxaphene air concentration in this region? And finally, can we quantify the toxaphene inputs from the U.S. sources to the Great Lakes ecosystem so as to assess numerically the contribution of the toxaphene residues in the U.S. soils to its distribution in the basin? The objectives of the present study are to investigate the contribution of the major toxaphene reservoirs in the United States to its budget in the Great Lakes basin, based on the previous results (e.g., 14, 21, 25, 28), and to address the questions raised above. In the present study a coupled regional-scale atmospheric transport, soil-air, and waterair exchange model (29, 30) was used to investigate toxaphene pathways in multimedia environments in the North American continent. Though the atmospheric long-range transport of toxaphene to the Great Lakes likely ranges from continental to global scales, identification of the contribution of toxaphene from North American sources to the air concentrations over the Great Lakes is essential for understanding and differentiating the relative importance of multiscale atmospheric transport to the Great Lakes. In this numerical investigation, toxaphene is treated, following MacLeod et al. (14), as a single chemical compound with averaged physicalchemical properties for the mixture. [It should be noted that although in theory there are nearly 16 000 different chlorinated compounds, only a few hundred are detectable in the environment (31, 32)]. The North American toxaphene emission inventory used in this study is that compiled by Environment Canada (2, 3).

Model and Numerical Experiment Setup The coupled model employed in this investigation is a threedimensional atmospheric transport model able to simulate transport and loadings of organochlorine pesticides at any model grid. The meteorological data required to drive the model may be those measured or predicted. The model has been used in previous numerical studies of lindane budget in the Great Lakes (29, 30). Briefly, it is a three-dimensional regional scale dispersion model coupled with a dynamic, three soil layer, fugacity-based soil-air exchange model, and a two-film model to estimate water-air gas exchange. In the present study the horizontal resolution of the model is 24 km and the model domain covers a large portion of Canada and the United States. The coupled model has 12 vertical levels from the surface to 7 km height. The model integration time step is 12 min. The model has been evaluated by comparison with measured lindane air concentrations from the Integrated Atmospheric Deposition Network around the Great Lakes (29) and through a comprehensive model evaluation and sensitivity study (33). The meteorological data (wind and air temperature) at each time step are obtained by interpolating the 6-hourly 8124

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 21, 2005

TABLE 1. Physical and Chemical Properties of Toxaphene at 25 °C physical and chemical properties mol-1)

molecular mass (g molar volume (cm3 mol-1) solid solubility (g m-3) log KOA degradation half-life in soil (days) degradation half-life in air (days) washout ratio

ref 414 251 0.63 9.3 3650 5 1.23 × 104

34 34 35 23 34 14 36

objectively analyzed data from the regional version (24 km horizontal resolution) of the Global Environmental Multiscale Model (GEM), Canada’s numerical weather forecasting model. When air temperature at the first model level above ground (1.5 m) is e-10 °C, water and soil are assumed to be frozen and hence there is no soil/air exchange or volatilization from these surfaces. The model is solved numerically by a finite-difference approximation and operator-splitting scheme. Relevant physical and chemical properties of toxaphene, adopted by the coupled model, are given in Table 1. The Henry’s law constant (in pascals) for toxaphene is given by (32)

log H )

-3209 + 10.42 T

where T is air temperature. To identify quantitatively the contribution of different sources in the United States to the toxaphene budget over the Great Lakes, the numerical simulations were performed for six model scenarios. These scenarios consist of (1) all sources in the United States, (2) southeast U.S. sources only, (3) northeast U.S. sources only, (4) northwest U.S. sources only, (5) southwest U.S. sources only, and (6) U.S. west-coast sources only. These regions in the model domain are displayed in Figure 1a. The six model scenarios are set up on the basis of considerations of the source strength and locations, and the potential effects of meteorological conditions on the atmospheric transport of toxaphene to the Great Lakes basin. For example, the southeast United States includes nine states (Figure 1a) where there were the highest toxaphene soil residues. However, prevailing large-scale wind systems may not favor the transport of pollutants from this region to the Great Lakes. Compared with the southeast U.S. region, toxaphene residues in the northwest and southwest U.S. source regions are much lower, but once volatilized into the air, they can be readily transported to the Great Lakes region by the prevailing continental wind systems. The presence of the Rockies and other mountain ranges in western North America introduce additional constraints on the long-range transport of pollutants from the west coast to the Great Lakes basin. The northeast United States includes the midwestern, New England, and mid-Atlantic states where the toxaphene residues were substantially lower than those in the southeast United States but more proximate to the Great Lakes basin, and thus can be regarded as the local source region. The model is run from January 1 to December 31, 2000, and driven by objectively analyzed meteorological data from the GEM.

Toxaphene Residues in U.S. Soils A toxaphene soil residues inventory in the United States in 2000 on a 1/6° × 1/4° latitude and longitude grid system was used (2, 3). These gridded residues were then interpolated to the model grids (Figure 1a). Figure 1b shows interpolated toxaphene soil residues in the model domain. At the beginning of 2000 it was estimated that a total of 29 100 t of toxaphene residues were still left in U.S. agricultural soils

FIGURE 1. (a) Model domain, grids, and the five regions in the United States specified for the five model scenarios. (b) Toxaphene soil residues on January 1, 2000 (t cell-1, 1 cell ) 24 km × 24 km). due to accumulation from past use of this pesticide (3). Of that, about 22 700 t was in the southeast U.S. source region (Figure 1a), 3500 t in the southwest, 1800 t in the northeast, 700 t in the west-coast region, and 400 t in the northwest, respectively. Relatively large values of gridded toxaphene soil residues are found from Arkansas to South Carolina, with the largest residue of 142.6 t cell-1 (1 cell ) 24 km × 24 km) at model grid (142, 40) in Alabama. In this study, toxaphene is introduced into the air only by reemission (volatilization) of its historical accumulated residues in the soil (no background air concentrations). The residues are assumed to be initially uniformly mixed in the top 10 cm of the soil (3). As described in detail in refs 23 and 29, the threelayer soil model consists of an exchange layer (0-0.1 cm from the surface), a buffer layer (0.1-1.0 cm), and a reservoir layer (1-10 cm). The toxaphene concentrations in the

reservoir layer are assumed to be influenced mainly by degradation. The buffer layer is the one where there is active exchange between the reservoir layer below and the soil/air exchange layer above. The soil/air exchange layer is the layer that responds most rapidly to changes in conditions (both meteorological and toxaphene air concentrations) in the atmospheric layer immediately above it. The fate of toxaphene in the atmosphere is therefore determined predominantly by degradation, emission from the soil, transport, and deposition in the atmosphere.

Results Modeled Mean Air Concentration. For scenario 1 (all sources included) the modeled monthly averaged daily air concentrations at 1.5 m height for selected months in the different seasons of 2000 are shown in Figure 2. High concentrations VOL. 39, NO. 21, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8125

TABLE 2. Modeled Annual Dry, Wet, and Total Depositions to Each Lake in 2000a depositions (kg year-1) dry wet total (dry + wet)

Lake basinOntario Erie Huron Michigan Superior wideb 0.66 2.68 3.34

2.24 4.37 6.61

1.57 2.39 3.96

5.05 0.73 5.78

1.23 0.24 1.47

10.75 10.41 21.16

a Estimated by summation of daily depositions at model grids within each lake from January 1 to December 31, 2000. b Basin-wide depositions are the sum of annual dry, wet, and total depositions over all five lakes.

toxaphene in Alabama drops off to 313 pg m-3, and air concentrations of toxaphene in the Great Lakes basin drop down to a range from 0.05 to 2 pg m-3. The significant reductions in air concentration from summer to winter imply temperature dependence of soil/air exchange (volatilization). Figure 3 displays monthly averaged toxaphene air concentrations at selected model grid points near the center of the each lake from model scenario 1 and the monthly mean air temperature (in kelvins) at 1.5 m height averaged over all the Great Lakes. As shown, air concentrations over all lakes tend to increase dramatically in September, with values of 243 and 174 pg m-3 near the center of Lakes Erie and Michigan, respectively. As seen in Figure 3, this rise in the air concentrations is not directly correlated with changes in the mean air temperature in this region. Strong long-range transport from other U.S. source regions is a possibility. Analysis of wind and sea-level pressure in September revealed the occurrence of a typical deformation flow (the axis of deformation flow is an area where strong convergence occurs) in the southern United States during the early to midSeptember period. This flow system is a strong wind convergence zone associated with a frontogenesis process, with the front extending from the southern United States to the Great Lakes and enabling the efficient transport of large amounts of toxaphene from the southern U.S. source regions to the Great Lakes basin. A detailed investigation for this event is presented in the companion paper (27).

FIGURE 2. Monthly average modeled daily toxaphene air concentration (pg m-3) for (a) April, (b) July, and (c) December 2000. ranging from 1900 to 3400 pg m-3 during the summer were found in the southeast United States, with the largest value at model grids in Alabama. These are consistent with the strong emissions occurring in this region. In the Great Lakes basin, in the summer, air concentrations range from less than 5 pg m-3 over the northern portion of the Great Lakes (Lakes Superior and Huron) to several tens of picograms per cubic meter over the lower Great Lakes (Lakes Erie and Ontario). In December, the maximum air concentration of 8126

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 21, 2005

Modeled Depositions to the Lakes. Modeled annual dry, wet, and total (dry + wet) deposition fluxes to each lake are displayed in Figure 4a and also listed in Table 2. These fluxes were obtained by summing the modeled daily fluxes at all the over-water grid points on each lake (30). Lakes Erie and Michigan receive more toxaphene than the other three lakes. In the eastern Great Lakes (Lakes Erie, Ontario, and Huron), as a result of higher precipitation rates in this region, wet deposition contributes more to the total deposition, while dry deposition is higher in the two upper lakes (Michigan and Superior). These results are also similar to numerical modeling results of Ma et al. (30) for lindane depositions to the lakes (1998-1999), which showed lower wet deposition to Lakes Michigan and Superior. Because of the constant washout ratio used in this modeling study, the lower wet depositions to the two upper lakes (Lake Superior in particular) are determined largely by lower precipitation rates predicted over the two lakes by the GEM model. As noted in ref 30, the uncertainties in predicted precipitation rates from the GEM model may, in turn, result in uncertainties in the wet deposition flux estimates. Comparing the total deposition values of 18.8, 13.6, and 5 kg year-1, respectively, for Lakes Superior, Michigan, and Ontario, estimated by Swackhamer et al. (13) for the mid-1990s, to the respective modeled values of 1.5, 5.8, and 3.3 kg year-1 from this study, we find a clear decreasing trend for toxaphene loading in each of the three lakes from the mid-1990s to 2000.

FIGURE 3. Modeled daily toxaphene air concentration (pg m-3) averaged over each month at a model grid near the center of Lake Ontario (172, 103), Lake Erie (168, 95), Lake Huron (116, 112), Lake Michigan (148, 103), and Lake Superior (149, 136).

FIGURE 4. (a) Modeled annual dry, wet, and total (dry + wet) depositions (kg year-1) to each lake in 2000, estimated by summation of daily depositions at all model grids within each lake from January 1 to December 31, 2000. (b) Monthly dry, wet, and total depositions (kg month-1) over all five lakes. The temporal trend of both dry and wet deposition fluxes (Figure 4b) is consistent with changes in the air concentrations as seen from the model output (Figure 3), showing larger values of the fluxes during May and June (note that the scale is logarithmic) compared to those during colder periods of the year, as well as a dramatic increase in September. The dry, wet, and total deposition fluxes to the lakes are about 7, 9, and 16 kg month-1 in September, respectively, but are each less than 1 kg month-1 in other months (Figure 4b), suggesting again that the strong long-

range transport occurring in September exerts a substantial impact on the toxaphene budget over the Great Lakes. This anomalous behavior in September is examined in more detail in the companion paper (27). Impacts of U.S. Sources. To assess the contribution of different toxaphene sources in the United States to its budget over each lake, we have, for each of the model scenarios 2-6 (considering toxaphene sources in the different regions of the United States), computed the ratios of annually averaged daily air concentrations and depositions averaged over all VOL. 39, NO. 21, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8127

by Lake Erie at 79%, Lake Huron at 73%, Lake Ontario at 69%, and Lake Superior at 59%. The overall contribution to the Great Lakes from the northeast United States, the second major source region, is less than 15%. The relative contributions of the various source regions to total deposition follows the same pattern as those for toxaphene air concentrations, with the northeast and southwest U.S. sources making almost similar contributions at ∼7% to the total deposition, due mainly to the wet deposition. The modeling results are consistent with other recent studies (15, 21) which have suggested that the cotton-growing region in the southern United States is the current source of toxaphene in the northern United States and the Great Lakes basin.

FIGURE 5. Model estimated ratios (labeled s2/s1 through s6/s1) of model scenarios 2-6 and scenario 1 over the Great Lakes for (a) annual averaged toxaphene air concentrations, (b) annual accumulated dry depositions, and (c) annual accumulated wet depositions. s1, all sources; s2, SE sources only; s3, NE sources only; s4, NW sources only; s5, SW sources only; s6, U.S. west coast sources only grid points within each lake to those from model scenario 1 (consisting of all sources). The air concentration ratios can be defined by

rc2 )

∑c , r ∑c 2 1

c3

)

∑c , ... ∑c 3 1

where rc2, rc3, ..., rc6 are ratios of annual mean daily air concentration produced from the model scenarios 2-6 (c2, c3, ..., c6) and that from model scenario 1 (c1) averaged over all five lakes. Since the governing equations (e.g, the atmospheric dispersion equation) in our coupled modeling system are linear, the solutions can be linearly combined, and hence the above approach can be used to determine the contributions from individual toxaphene source regions to the receptor regions. Figure 5 shows the ratios of the modeled annual averaged daily air concentrations (a), dry deposition (b), and wet deposition (c) over the five lakes. On an annual basis, the southeast U.S. sources made the largest contributions to the toxaphene levels in the air and the depositions to all lakes (or basin-wide deposition) at 72% for the air concentration, 78% for the dry deposition, and 88% for the wet deposition. The second major source of toxaphene over the Great Lakes is the northeast United States, followed by the southwest, northwest, and west-coast sources. The results also showed that the southeast U.S. sources contributed 82% of toxaphene air concentrations over Lake Michigan, followed 8128

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 21, 2005

Having established the relative importance of the various U.S. source regions to the toxaphene budget over the Great Lakes, we may now further explore the temporal trend of the impacts of U.S. sources on the toxaphene air concentrations over the Great Lakes basin. From the results presented above, if we regard the northeast U.S. (which includes the Great Lakes basin) source as the local source (Figure 1a) for the toxaphene budget over the Great Lakes, the toxaphene soil residues in this source region together with the major source region in the southeast United States account for over 85% of toxaphene air concentration and depositions across the Great Lakes in the year 2000 (Figure 5a), of which over 70% was from the southeast United States However, because motion of a pollutant is directly associated with the dynamics of motion in the atmosphere, the changes in mean atmospheric pressure and flow systems with seasons would exert a strong effect on transport and distribution of the pollutant in the air. As a result, the relative importance of the local source in the northeast United States and the major source in the southeast United States to the toxaphene air concentrations over the Great Lakes may also change with seasons. Figure 6 illustrates monthly ratios of toxaphene air concentrations averaged over each of the Great Lakes from model scenarios 2 and 1 (s2/s1; Figure 6a), as well as from model scenarios 3 and 1 (s3/s1; Figure 6b). Although the temporal trends of the air concentration ratios are not identical for each lake, the results indicate that the toxaphene residues in the southeast U.S. soils tend to make larger fractional contributions to the toxaphene budget over the lakes during the spring and autumn periods than during the summer period of the year. This can be seen more clearly from changes in the mean ratios (thick dark line) in Figure 6a, which shows that the ratios in the summertime (JuneAugust) are below 0.5 and increase dramatically thereafter, with peak values in September. In general, as shown in Figure 6a, the relatively large ratios are observed in the spring and autumn, which are the transition seasons and also periods of severe weather with powerful storms. An almost inverse temporal trend (displayed by the thick dark line in Figure 6b) occurs in the ratios of the air concentrations s3/s1. There is a high but negative correlation coefficient of -0.92 between the mean ratios in Figure 6a,b. These results may carry significant implications due to the influence of the seasonal changes in meteorological conditions on toxaphene’s intercompartmental transport processes. First, it is known that midlatitude atmospheric circulations (pressure and wind systems) undergo a very important shift from winter to summer, due primarily to strong seasonal temperature variations over land. Briefly, the time-averaged circulations in the Earth’s atmosphere characterized by the large continental to global scale perturbations are more prominent in relatively cold seasons than in summer, as featured by large, well-behaved, and relatively regular atmospheric waves (26). These atmospheric long waves favor long-range transport of air masses and pollutants. As late spring and summer approach, atmospheric waves often can be small and subtle, rendering continental-scale atmospheric long-range trans-

FIGURE 6. Monthly ratios of toxaphene air concentrations averaged over each of the Great Lakes: (a) scenario 2 to scenario 1 and (b) scenario 3 to scenario 1. s2, SE sources only; s3, NE sources only. port of a pollutant rather weak. Further, atmospheric short waves in summertime are quite often featured by heating and cooling caused by local terrestrial surfaces. Consequently, toxaphene air concentrations over the Great Lakes would imprint more local signatures from those sources that are more proximate to the receptor region during summertime than during the relatively cold seasons, as revealed by Figure 6. Nevertheless, we should point out that the above analysis should not be regarded as an indication that the local sources would dominate the toxaphene budget over the Great Lakes in the summer. In fact, Figure 6b shows clearly that the s3/s1 ratios never exceed 0.5 and are smaller than the s2/s1 ratios throughout the year, even during the summer season. To further explore the soil/air exchange in the southeast (major) and northeast U.S. (local) source regions, we have estimated monthly soil/air fugacity ratios fs/fa at two selected grid cells as reference grids of the major and local sources. These are grids (142, 39), sited in Alabama, and (155, 100), sited in southern Michigan. The toxaphene residues in the soil at these two grids at the beginning of 2000 are 135 and 0.12 t cell-1, respectively. This may provide further insight into soil-air exchange dynamics in these two source regions. The soil/air fugacity ratio is an indicator of equilibrium status (23). Values of fs/fa > 1 represent net transfer out of soil (volatilization), and values < 1 indicate the opposite (deposition). Given that the toxaphene was contained in the top 10 cm layer of soil (3) and that the response of toxaphene residues in the reservoir layer to the changes in its air concentration is relatively slow, the soil fugacity is calculated for the reservoir layer (1-10 cm below the surface) and the air fugacity is calculated at 1.5 m height above the surface.

FIGURE 7. Model calculated toxaphene soil/air fugacity ratio (fs/fa) at model source grids (142, 39) in Alabama and (155, 100) in southern Michigan. Soil fugacity was taken at reservoir layer (1-10 cm) and air fugacity was taken at 1.5 m height. The results are illustrated in Figure 7. It is clear that the temporal trend of the soil/air fugacity ratios at the two grid locations differs substantially. fs/fa at the major source grid cell (142, 39) increases gradually over the course of the year, reaching a maximum in November. This trend is likely to follow changes in the soil capacity for toxaphene and is consistent with the fugacity ratios estimated from measured soil residue levels of toxaphene in an agriculture region of northwest Alabama (23). Interestingly, at the (local) source grid (155, 100) with much lower soil residue, the fugacity ratios become less than 1 from September onward, indicating VOL. 39, NO. 21, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8129

FIGURE 8. Modeled (- - -) at grid (183, 106) and measured (s) daily air concentrations of toxaphene at Point Petre on the sampling days. (Inset) Correlation diagram between measured (ordinate) and modeled (abscissa) concentrations. that this source grid turns to a receptor grid! This occurs because large amounts of toxaphene-laden air migrated from the southern United States to the “local” source and the Great Lakes regions, leading to a substantial increase in toxaphene levels in the atmosphere over these regions. Because fa increases more sharply than fs, the increasing air concentration of toxaphene results in net deposition and a reverse thermodynamic gradient at those weak source grids over which the large amount of toxaphene originated and transported from the southeast United States could dominate ambient air concentrations. The temporal changes in fugacity ratios at the two source grids suggest further that strong volatilizations would take place in the southeast U.S. dominant source region throughout the year. In the northeast U.S. source region, where the toxaphene soil residues were much lower, the net transfer of toxaphene between soil and air may change direction from volatilization to deposition. As a result, the local sources become less important to the toxaphene budget over the Great Lakes basin. Comparison with Measurements. To evaluate the model performance, the modeled air concentrations at 1.5 m height above the ground were compared with measurement of airborne (vapor-phase) toxaphene. The first sampled data set was collected at Point Petre by the Organics Analysis Laboratory of the MSC (16, 37). The air concentrations in 2000 were sampled once per month as a 72-h integrated value. To be consistent with the measurement approach, the modeled daily air concentrations were averaged over 3 days at the model grid (183, 106), the grid closest to Point Petre, which were then compared with the measured toxaphene air concentrations on the sampling days. The results are shown in Figure 8. In general, while the modeled air concentrations correlate well with measurements a (correlation coefficient of r ) 0.9), the model tends to overestimate air concentrations at Point Petre. Large discrepancies occurred on the sampling days in May and June. The modeled air concentrations are about a factor of 2-3 higher than the measurements. We further compared modeled air concentrations with atmospheric samples of toxaphene, collected by James and Hites (15), at Sleeping Bear Dunes on the northeast shore of Lake Michigan [44° 48′ 47′′ N, 86° 03′ 02′′ W, modeled grid (151, 109)] and at Indiana University in Bloomington, IN [39° 10′ 00′′ N, 86° 31′ 17′′ W, model grid (154, 86)]. The air samples were taken once every 12 days from March to December 2000. Results show that 8130

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 21, 2005

the modeled air concentrations captured the variations in the observations but underestimated the magnitudes of air concentrations at these two sites. The third data set from a ship-based sampling of toxaphene air concentrations over water at various locations on Lake Ontario in June 2000 by the MSC (unpublished, personal communication with Liisa Jantunen, 2004) yielded values ranging from 10 to 54 pg m-3. These values are comparable to the modeled daily average concentrations, averaged over all points over Lake Ontario, which ranged from 2 to 70 pg m-3 over the month of June. The differences in the model performance when compared with observations at the different locations may be due to inadequate representation of soil-air exchange processes in the coupled modeling system. However, it is also apparent that the different instrumentation and measurement techniques employed in the different field campaigns (15, 16, 37) may themselves have an influence on the measured values (personal communication with Liisa Jantunen, 2004). The numerical investigation reported in the paper is part of a continuing effort to understand the impact of toxaphene re-emissions in the U.S. source regions on its budget over the Great Lakes basin. The investigation identifies that the southeast United States is a dominant source contributing to toxaphene levels in the atmosphere over the Great Lakes basin and to depositions to the lake waters. Another key finding from this investigation is that the contribution of different sources in the United States to toxaphene budget over the Great Lakes undergoes a significant shift from winter to summer seasons, showing that the impact of the largest toxaphene reservoir in the southeast United States on the Great Lakes ecosystem is more significant during spring and autumn than during the summer season. The sources proximate to the Great Lakes make a greater relative contribution to the toxaphene budget over this ecosystem during the summer than during other seasons of the year, due primarily to the interseasonal changes in atmospheric circulation systems. The quantitative understanding of the air concentrations and budget of toxaphene over the Great Lakes obtained through these model simulations can be used to take appropriate measures, at either the source or receptor regions, to mitigate the impacts of toxaphene.

Acknowledgments We are grateful to Pierrette Blanchard, Liisa Jantunen of Environment Canada for providing measured data on

toxaphene air concentrations, and Philip Cheung, also of Environment Canada, for assistance with the meteorological data. Special thanks to Terry Bidleman for his useful comments on the early draft of this paper. This study is partially funded by the Great Lakes Binational Toxics Strategy program.

Literature Cited (1) Voldner, E. C.; Li, Y. F. Global usage of selected persistent organochlorines. Sci. Total Environ. 1995, 160/161, 201-210. (2) Li, Y. F. Toxaphene in the United States. 1. Usage gridding. J. Geophys. Res. 2001, 106, 17919-17927. (3) Li, Y. F.; Bidleman, T. F.; Barrie, L. A. Toxaphene in the United States. 2. Emissions and residues. J. Geophys. Res. 2001, 106, 17929-17938. (4) Saleh, M. A. Toxaphene: chemistry, biochemistry, toxicity and environmental fate. In Reviews of Environmental Contaminant Toxicology; Ware, G. W., Ed.; Springer-Verlag: New York, 1991; Vol. 118, pp 1-85 (5) Delorme, P. D.; Muir, D. C. G.; Lockhart, W. L.; Mills, K. H.; Ward, F. J. Depuration of toxaphene in lake trout and white suckers in a natural ecosystem following a single I. P. dose. Chemosphere 1993, 27, 1965-1973. (6) Paasivirta, J.; Rantio, T.; Koistinen, J.; Vuorinen, G. P. Studies on toxaphene in the environment. II. PCCs in Baltic and Arctic Sea and lake fish. Chemosphere 1993, 27, 2011-2015. (7) Zhu, J.; Norstrom, R. J. Identification of polychlorocamphenes (PCCs) in the polar bear (Ursus maritimus) food chain. Chemosphere 1993, 27, 1923-1936. (8) Rice, C. P. Homologue concentrations of chlorobornanes in selected samples from the Bering/Chukchi Sea. Chemosphere 1993, 27, 1937-1947. (9) Hargrave, B. T.; Muir, D. C. G.; Bidleman, T. F. Toxaphene in amphipods and zooplankton from the Arctic Ocean. Chemosphere 1993, 27, 1949-1963. (10) Glassmeyer, S. T.; De Vault, D. S.; Myers, T. R.; Hites, R. A. Toxaphene in Great Lakes fish: A temporal, spatial, and trophic study. Environ. Sci. Technol. 1997, 31, 84-88. (11) Pearson, R. F.; Swackhamer, D. L.; Eisenreich, S. J.; Long, D. T. Concentrations, accumulations, and inventories of toxaphene in sediments of the Great Lakes. Environ. Sci. Technol. 1997, 31, 3523-3529. (12) Delta Institute Ann Arbor Statement: Long-range transport of toxic substances into the Great Lakes basin, Ann Arbor, MI, 2004; http://www.Delta-institute.Org/pollprev/lrtworkshop/ AnnArborStatement (PDF) (13) Swackhamer, D. L.; Schottler, S.; Pearson, R. F. Air-water exchange and mass balance of toxaphene in the Great Lakes. Environ. Sci. Technol. 1999, 33, 3864-3872. (14) MacLeod, M.; Woodfine, D.; Brimacombe, J.; Toose, L.; Mackay, D. A dynamic mass budget for toxaphene in North America. Environ. Toxicol. Chem. 2002, 21, 1628-1637. (15) James, R.; Hites, R. A. Atmospheric transport of toxaphene from the Southern United States to the Great Lakes region. Environ. Sci. Technol. 2002, 36, 3474-3481. (16) Shoeib, M.; Brice, K. A.; Hoff, R. M. Studies of toxaphene in technical standard and extracts of background air samples (Point Petre, Ontario) using multidimensional gas chromatographyelectron capture detection (MDGC-ECD). Chemosphere 2000, 40, 201-211. (17) Voldner, E. C.; Schroeder, W. H. Modelling of atmospheric transport and deposition of toxaphene into the Great Lakes ecosystem. Atmos. Environ. 1989, 23, 1949-61. (18) McConnell, L. E.; Kuclick, J. R.; Bidleman, T. F.; Walla, M. D. Long-range atmospheric transport of toxaphene to Lake Baikal. Chemosphere 1993, 27, 2027-2036. (19) Barrie, L. A.; Bidleman, T. F.; Dougherty, D.; Fellin, P.; Grift, N.; Muir, D. C. G.; Rosenberg, B.; Stern, G.; Toom, D. Atmospheric toxaphene in the high Arctic. Chemosphere 1993, 27, 20372046.

(20) McDonald, J. G.; Hites, R. A. Dilution model for the distribution of toxaphene in the United States and Canada on the basis of measured concentrations in tree bark. Environ. Sci. Technol. 2003, 37, 475-481. (21) Bidleman, T.; Leone, A. Soil-air exchange of organochlorine pesticides in the Southern United States. Environ. Pollut. 2004, 128, 49-57. (22) Hoff, R. M.; Muir, D. C. G.; Grift, N. P.; Brice, K. A. Measurement of PCCs in air in southern Ontario. Chemosphere 1993, 27, 20572062. (23) Harner, T.; Bidleman, T. F.; Jantunen, L. M. M.; Mackay, D. Soil-air exchange model of persistent pesticides in the United States cotton belt. Environ. Toxicol. Chem. 2001, 20, 16121621. (24) Jantunen, L. M. M.; Bidleman, T. F.; Harner, T.; Parkhurst, W. J. Toxaphene, chlordane, and other organochlorine pesticides in Alabama air. Environ. Sci. Technol. 2000, 34, 5097-5105. (25) Rice, C. P.; Samson, P. J.; Noguchi, G. E. Atmospheric transport of toxaphene to Lake Michigan. Environ. Sci. Technol. 1986, 20, 1109-1116. (26) Riehl, H. Introduction to the Atmosphere; McGraw-Hill Inc.: New York, 1965. (27) Ma, J.; Venkatesh, S.; Li, Y. F.; Cao, Z.; Daggupaty, S. M. Tracking toxaphene in the North American Great Lakes basin. 2. A strong episodic long-range transport event. Environ. Sci. Technol. 2005 39, 8132-8141. (28) Swackhamer, D. L.; McConnell, L. L.; Gregor, D. J. Workgroup report on environmental transport and fate. Chemosphere 1993, 27, 1835-1840. (29) Ma, J.; Daggupaty, S. M.; 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, 37743781. (30) Ma J.; Daggupaty, S. M.; Harner, T.; Blanchard, P.; Waite, D. Impacts of lindane usage in the Canadian Prairies on the Great Lakes ecosystem. 2. Modeled fluxes and loadings to the Great Lakes. Environ. Sci. Technol. 2004, 38, 984-990. (31) Hainzl, D.; Burhenne, D.; Parlar, H. Theoretical consideration of the structural variety in the toxaphene mixture taking into account recent experimental results. Chemosphere 1994, 28, 245-252. (32) Jantunen, L. M. M.; Bidleman, T. F. Temperature-dependent Henry’s law constant for technical toxaphene. Chemosphere: Global Change Sci. 2000, 2, 225-231. (33) Daggupaty, S. M.; Ma, J. An evaluation of uncertainty and sensitivity of a regional scale air pollution and pesticide modeling system. 13th Joint Conference on the Applications of Air Pollution Meteorology, Vancouver, Canada, 2004; http:// ams.confex.com/ams/. (34) Mackay, D.; Shiu, W.-Y.; Ma, K.-C. Physical-Chemical Properties and Environmental Fate Handbook; Chapman & Hall: New York, 1998 [CRCnetBASE; CRC Press LLC: Boca Raton, FL, 1999]. (35) Murphy, T. J.; Mullin, M. D.; Meyer, J. A. Equilibration of polychlorinated biphenyls and toxaphene with air and water. Environ. Sci. Technol. 1987, 21, 155-162. (36) Atlas, E.; Giam, C. S. Ambient concentration and precipitation scavenging of atmospheric organic pollutants. Water, Air, Soil Pollut. 1988, 38, 19-36. (37) Shoeib, M.; Brice, K. A.; Hoff, R. M. Airborne concentrations of toxaphene congeners at Point Petre (Ontario) using gaschromatography-electron capture negative ion mass spectrometry (GC-ECNIMS). Chemosphere 1999, 39, 849-871.

Received for review May 17, 2005. Revised manuscript received August 10, 2005. Accepted August 19, 2005. ES050945M

VOL. 39, NO. 21, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8131