AIR MONITORING OF PERSISTENT ORGANIC POLLUTANTS IN THE ...

Environmental Monitoring and Assessment (2005) 100: 201–216

© Springer 2005

AIR MONITORING OF PERSISTENT ORGANIC POLLUTANTS IN THE GREAT LAKES: IADN VS. AEOLOS MATT F. SIMCIK Division of Environmental & Occupational Health, School of Public Health, University of Minnesota, 420 Delaware Street SE, Minneapolis, Minnesota, U.S.A. (∗ e-mail: [email protected])

(Received 26 May 2003; accepted 2 December 2003)

Abstract. When designing a monitoring campaign, one has to consider many factors in the decision to perform a long-term synoptic monitoring program or a short-term intensive study. Each has its own advantages and disadvantages. This paper compares and contrasts the information obtained from two studies conducted on the Laurentian Great Lakes. One, the Integrated Atmospheric Deposition Network (IADN), is a long-term synoptic monitoring study and the other, the Atmospheric Exchange Over Lakes and Oceans (AEOLOS), was a short-term intensive study. The advantages of long-term synoptic monitoring programs are providing greater spatial information, the relative influence of long and short-range transport on the regional background, gross loadings representative of the majority of each lake and long-term temporal trends. Short-term intensive studies provide more information on the processes governing sources, transport and deposition, such as the urban/industrial influence on adjacent large water bodies, specific sources to an urban/industrial area and short-term fluctuations in concentrations due to meteorology, source strength and photochemical reactions. Using information provided by both the IADN and AEOLOS studies, areas of urban influence are predicted for each of the five Great Lakes. Keywords: atmospheric loadings, POPs, source apportionment, spatial coverage, temporal trends

1. Introduction When establishing an air monitoring program for persistent organic pollutants (POPs) it is important to consider both long-term synoptic efforts and short-term intensive studies. Factors necessary in the decision to choose one over the other include logistical capabilities, monetary constraints and desired information such as long-term trends versus short-term variability, background loadings versus urban inputs and resolution of source information. In an ideal world one could operate both in tandem, but rarely is this type of effort possible. Synoptic and intensive monitoring studies provide different information regarding spatial coverage, temporal trends, source information and atmospheric deposition loadings. They both have distinct advantages and disadvantages, but combining information from both results in a much greater impact. Because of periodic sampling, long-term synoptic studies are best suited for large regional background areas so as to minimize shortterm fluctuations in meteorology and source strength. Short-term intensive studies, on the other hand, are best suited to areas with maximal short-term fluctuations

202

M. F. SIMCIK

Figure 1. IADN and AEOLOS sites on the Laurentian Great Lakes. Notes: stars indicate IADN master stations and circles indicate IADN satellite stations; IIT site from AEOLOS and Chicago IADN satellite station are collocated.

near sources where minute changes in meteorology and source strength can have a larger impact on ambient concentrations and loadings. This paper will consider two studies conducted on the Laurentian Great Lakes to compare and contrast the information provided. The Integrated Atmospheric Deposition Network (IADN) is what can be characterized as a long-term synoptic study. IADN is a combined effort under the direction of the U.S. and Canadian governments that began in November, 1990 and continues in operation today. The Atmospheric Exchange Over Lakes and Oceans (AEOLOS) was a multi-institutional intensive study, most of which was performed on southern Lake Michigan near Chicago, IL during May and July, 1994 and January, 1995. 1.1. IADN The IADN program samples air and precipitation from sites distributed among all five Laurentian Great Lakes. It was established under Annex 15 of the Great Lakes Water Quality Agreement between the United States and Canada. On the Canadian side the monitoring is conducted by Environment Canada and on the American side by Indiana University under the supervision of the United States Environmental Protection Agency’s Great Lakes National Program Office. The objectives of IADN are 1) to determine the atmospheric loadings of toxic contaminants to each of the Great Lakes and their temporal and spatial trends, 2)

AIR MONITORING OF PERSISTENT ORGANIC POLLUTANTS IN THE GREAT LAKES

203

TABLE I Summary of sampling information for IADN and AEOLOS

Duration Spatial Coverage (water) Air samples Rain samples Analytes Sampling and analytical methods Additional sampling

IADN

AEOLOS

1990 – present ∼ 250,000 km2 24 hr/12 days Monthly composites PAHs, PCBs, pesticides, some metals Comparable

May 1994 – January 1995 ∼ 1000 km2 2 × 12 hr/day By event PAHs, PCBs, pesticides, metals (inc. Hg)

none

Dry deposition, particle size distributions

Comparable

report self-consistent air and precipitation concentrations and loadings over time to determine long-term trends and 3) to determine sources of these toxic contaminants to the Great Lakes. There is one Master Station per lake placed to be representative of the regional background air removed from local sources. Additionally, there are several satellite sites placed to investigate urban influences. Specific site locations are summarized in Figure 1. Air is sampled using highvolume air samplers equipped with glass fiber filters (GFF) and either XAD-2 polymer resin or polyurethane foam (PUF) to collect the gas and particle phase contaminants, respectively. The US sites employ the XAD-2 resin and the Canadian sites employ PUF. Sampling frequency is every 12 days for 24 hr resulting in approximately 800 m3 of air per sample collected at the US sites and 350 m3 of air per sample collected at the Canadian sites. Extensive comparisons between the two methods have been performed (Hoff et al., 1996), and the methods were determined to be comparable. Precipitation is sampled with a standard wet-only sampler equipped with a heated rain sensor to prevent condensation and heated collection funnel to melt frozen precipitation. Extraction of precipitation is done continuously in the field using a flow through system of XAD-2 resin. The resin is changed every month resulting in monthly composite precipitation samples. All samples are analyzed for several semivolatile organic compounds (SOCs) including polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs) and several pesticides. Table I summarizes the IADN sampling program. 1.2. AEOLOS The AEOLOS study was a multi-institutional collaboration whose working hypothesis was that emissions of hazardous air pollutants into the coastal urban atmo-

204

M. F. SIMCIK

sphere enhance atmospheric deposition to adjacent waters. The first three sampling campaigns associated with AEOLOS were performed in Chicago, IL and over southern Lake Michigan. Air and precipitation was sampled at three sites around southern Lake Michigan, in the city of Chicago, over Lake Michigan aboard the USEPA ship, RV Lake Guardian, and across the lake in South Haven, MI (inset of Figure 1). Air samples were collected using high volume air samplers equipped with glass fiber filters and polyurethane foam plugs to collect the particle and gas phase compounds, respectively. Samples were collected for 12 hr both day and night for two days in May, 1994, eleven days in July, 1994 and four days in January, 1995. Precipitation events were collected using wet-only samplers on an event basis and kept separate with dissolved- and particle-associated compounds segregated. Additionally, size-segregated particulate samples were collected using both a Berner-type impacter and a multiple orifice uniform deposit impactor (MOUDI), and dry deposition was estimated using surrogate surfaces. All samples were analyzed for a similar suite of SOCs to IADN. Additionally, some research groups sampled air, depositing particles and precipitation for SOCs, metals and mercury. Table I summarizes the AEOLOS sampling program.

2. Spatial Coverage Long-term synoptic monitoring programs often spread their resources over time and space. As such they are best suited for studying large-scale spatial heterogeneity including entire lake systems such as the Great Lakes. Short-term synoptic studies, on the other hand, are better suited to small-scale spatial heterogeneity that describes a portion of a system. When the system of interest involves large water bodies, over-water sampling presents a challenge. Long-term synoptic programs are often unable to sample over-water, and therefore forced to measure air concentrations and atmospheric deposition over land. This over-land sampling is blind to the spatiotemporally varying relationship between over-water and overland concentrations and deposition. Short-term intensive studies can often take advantage of their short duration to sample air and deposition over water such that a true measure of over-water trends and loadings are possible. The IADN long-term synoptic monitoring program was designed to characterize the entire airshed of the Great Lakes basin. By doing so, IADN has been able to characterize spatial heterogeneity among the five lakes. Results of both air and precipitation concentrations indicate that Lake Superior’s air is the least contaminated of the five Great Lakes and that in general air concentrations near Lake Michigan, Lake Erie, and Lake Ontario are similar and among the highest (Cortes et al., 1998; Simcik et al., 1999; Simcik et al., 2000). IADN relies on one Master Site per lake to be representative of the entire area of its respective lake. These sites were selected to be indicative of regional background signals removed from urban influence. As such they do represent the majority of the lake, but neglect urban influences by


205

design. With the addition of satellite sites, IADN has attempted to gain insight into spatial heterogeneity within each lake, but this coverage is still limited. Another limitation is that land-based air sampling may not adequately represent over-water air concentrations and loadings. Investigations by IADN into the efficacy of landbased sampling have been limited to segregating air samples with wind directions indicating over-water and over-land air masses (Simcik et al., 1999). Because of the absence of over-water samples, no further conclusions about this limitation is available from IADN. The AEOLOS intensive study sampled a relatively small area to investigate spatial heterogeneity between over-land and over-water air concentrations, urban influences on the adjacent coastal atmosphere and the extent of that influence. Results indicated that the urban influence is great, but does not extend great distances over the lake. Gas-phase concentrations of PCBs and PAHs indicate that air over southern Lake Michigan just 10 km off-shore were approximately an order of magnitude lower than those measured over-land in the city of Chicago, IL (Simcik et al., 1997). However, the over-water concentrations were still higher than the over-land site across Lake Michigan at a rural site in South Haven, MI (Simcik et al., 1997). Furthermore, the over-water concentration varied dramatically when the wind was coming to the sampler aboard the ship from Chicago versus coming down the axis of the lake. Simcik et al. (1997) concluded that the urban/industrial area around Chicago, IL increased the over-water gas-phase concentrations of PAHs and PCBs by a factor of 12 and 4 times that of background concentrations, respectively. Similar results were reported for SOCs associated with precipitation events measured by AEOLOS (Offenberg and Baker, 1997). However, because this study was concentrated on the urban influence on adjacent coastal air, no comparison to the larger Great Lakes system was possible. In general, long-term synoptic monitoring programs are best suited to largescale spatial heterogeneity on the order of the entire Great Lakes system (i.e. hundreds of thousands of square kilometers) while intensive studies are better suited to investigate spatial heterogeneity arising from mesoscale effects and transport processes in smaller scale situations. Long-term synoptic programs are often forced to employ only land-based sites in which conventional methods do not reveal microscale effects between over-land and over-water concentrations. Shortterm intensive studies can take advantage of over-water sampling to investigate this relationship, but are usually limited to a small portion of the entire Great Lakes system.

3. Temporal Trends By definition, long-term synoptic studies are designed to discern long-term temporal trends. A resulting limitation is that they are required to minimize short-term fluctuations in concentrations due to meteorology and other factors in order to re-

206

M. F. SIMCIK

veal long-term trends. To accomplish this, most long-term synoptic studies sample for 24 hr at a time separated by several days over years. The 24 hr samples integrate any diurnal variations that may occur. Samples can be spread out over each year to study seasonal differences within any long-term trends that may be present. Conversely, short-term intensive studies are designed to study and take advantage of short-term variability in concentrations due to meteorological, photochemical and source strength variability. Therefore, short duration samples are taken continuously over a short period of time. As mentioned above, there are often seasonal variations in atmospheric concentrations of SOCs. This seasonal variability is partially due to temperature differences and its effect on volatilization from surfaces. The IADN program accounts for this temperature dependence with the aid of the Clausius-Clapeyron equation: ln P =

−HSA + const RT

(1)

where P is the partial pressure of the gas-phase contaminant, HSA is the enthalpy of phase change, R is the molar gas constant and T is the ambient temperature in Kelvin. By plotting the natural log of the partial pressure versus reciprocal temperature one can often obtain a straight line with a slope equal to the negative of the ratio of the enthalpy of phase change to the molar gas constant. The resulting slopes can then be used to adjust the partial pressures (and hence concentrations) of individual SOCs to a reference temperature of 288 K, the global tropospheric mean. These adjusted partial pressures are regressed versus time and the result for many banned chemicals at many sites is a first-order decrease in concentration over the time span of IADN. These first-order decreases have been presented by the IADN program as halflives. For gas-phase PCBs, each individual congener concentration adjusted to 288K is summed to give the temperature corrected total PCB concentration. The half-life of gas-phase total PCBs for Lakes Superior, Michigan and Erie ranged from 2.7 to 3.5 yr (Simcik et al., 1999). Cortes et al. (1998) found that many pesticides exhibited half-lives in atmospheric concentration ranging from 2.5 to 12 yr. While many others did not show significant half-lives, they were still decreasing and Cortes and Hites (2000) concluded that an additional two years of sampling would produce significance. Precipitation collected by IADN was also shown to exhibit decreasing volume weighted mean concentrations for many SOCs including both PCBs and pesticides. The first-order half-lives corresponding to these decreases ranged from 0.8 ± 0.2 yr for DDT near Lake Huron to 6.9 ± 3.5 yr for total PCBs near Lake Michigan (Simcik et al., 2000). AEOLOS was designed to measure SOCs from 12 hr air samples day and night for three different intensive periods, thus observing and explaining short-term variability. Simcik et al. (1997) observed great variation in gas-phase PCB and PAH concentrations both in Chicago and over Lake Michigan. These variations were attributed to meteorology and photochemistry. Gas phase PCB concentrations were


207

much higher in July than either May or January as a result of temperature effects explained by the IADN study. However, an anomaly unique to AEOLOS occurred during the second week of July sampling when wind shifted from the Southwest to the North bringing clean air down the axis of the lake. This caused a drastic reduction in concentration for both PCBs and PAHs. Gas-phase PCB concentrations during this North wind episode measured over-water (100–600 pg m−3 ) were comparable to concentrations measured at the IADN Master Site for Lake Michigan (160 ± 10 pg m−3 ) chosen to represent the regional background (Hillery et al., 1997; Simcik et al., 1997). From the difference between air masses passing over Chicago and those passing over Lake Michigan, Simcik et al. (1997) concluded that the urban/industrial area resulted in gas-phase PCB and PAH concentrations 4 and 12 times background. Another anomaly observed during AEOLOS was diurnal variation in gas-phase PAHs that was attributed to a daytime loss due to photochemical reactions (Simcik et al., 1997). As a result of this daytime loss, the urban influence could be separated into a daytime increase of 5 times and a nighttime increase of 18 times regional background concentrations. Because AEOLOS sampled precipitation on an event basis, Offenberg and Baker (2002) observed differences between the SOC scavenging of rain originating from convective and frontal storms. This resolution is lost in long-term synoptic studies like IADN that integrate precipitation over many events and fail to sample air and rain at the same time. Long-term synoptic monitoring programs such as IADN are well suited to describing temporal trends in atmospheric contaminants that have been banned from use because they offer a decreasing trend that behaves in a first order manner. It is only by sampling for long periods of time that one can make discoveries of changes that have half-lives on the order of years. However, in areas near urban/industrial sources where there is much higher variability in concentration due to factors other than overall decreases in regional atmospheric signal, it becomes difficult to observe overall trends and requires a much longer time period to observe significant changes (Cortes and Hites, 2000). Furthermore, some fresh urban sources such as places where PCBs are stored or disposed of would not necessarily decrease at a rate similar to that observed for the regional background, if at all. Therefore, lack of a decreasing trend and higher concentrations observed by IADN have lead to the investigation of continuous sources of these banned chemicals. Intensive studies such as AEOLOS are incapable of describing long-term trends, but can lend information on diurnal variations and short-term meteorological differences. These types of observations would be impossible in a long-term monitoring program where sampling is separated by a number of days. It would be difficult to know if changes in concentration were due to changes in wind direction, changes in source strengths, or changes in temperature. Therefore, long-term synoptic monitoring is best suited for assessing long-term trends and short-term intensive studies are best suited for studying short-term changes in concentrations brought about by varying meteorology, source strengths and photochemistry.

208

M. F. SIMCIK

Figure 2. Non-linearity in Clausius-Clapeyron plots observed by Hoff et al. (1998).

4. Source Information When designing a monitoring program from the point of view of determining sources of atmospheric contaminants, one has to consider the level of knowledge desired. Receptor models, for the purposes of the comparison in this paper, can be placed into two categories: those that require a priori knowledge of which sources contribute to concentrations at the receptor, and characterization of the sources; and those that rely on meteorology and/or short-term variation in source strengths with moderate to no understanding of source characterization. An example of the first type of receptor model is chemical mass balance (CMB). While a CMB can be used in both long-term synoptic studies and short-term intensive studies, they are difficult to employ due to the information required about sources and their characterization. CMBs have been used quite successfully to apportion specific sources of atmospheric particulate matter (Dzubay et al., 1988; Kao and Friedlander, 1995; Kleinman et al., 1980; Kowalczyk et al., 1982; Lewis et al., 1986; Morandi et al., 1987; Pratsinis et al., 1988; Tuncel et al., 1985). However, they are not as applicable to sources of organic contaminants due to the atmospheric lability of many SOCs. Therefore receptor-based models used in both long-term synoptic and shortterm intensive studies must rely on meteorology to alter source strengths. Because long-term synoptic studies are spread out over several years, the relative strength of major local sources can change, and therefore be indiscernible. However there are techniques that can be employed to determine the relative influence between longrange transport and short-range transport. Intensive studies, on the other hand, can be designed to elucidate specific source categories, source areas or even specific point sources because short-term variation in meteorology and source strength can be observed.


209

Information from the IADN program provides gross generalizations about sources of atmospheric contaminants. That is, one can determine whether the atmospheric concentrations result from local sources or long-range transport. This has been done in the IADN study by investigating the enthalpies of phase change that result from the Clausius-Clapeyron plots of Equation 1. In a critical review of gas-phase concentrations of SOCs and their temperature dependence Wania et al. (1998) concluded that the similarity between the observed enthalpy of phase change and the enthalpy of vaporization indicates the degree to which local sources contribute to the atmospheric concentration. Two papers from the IADN program have used a different interpretation of these Clausius-Clapeyron plots to infer longrange versus short-range transport. Hoff et al. (1998) plotted several semivolatile compounds in Clausius-Clapeyron plots and observed a non-linearity from which they concluded that long-range transport is the major contributor to atmospheric contaminants in the Great Lakes region. Their conclusion is depicted in Figure 2, which has been adapted from their publication showing that long-range transport will provide a slope very near zero, while equilibrium partitioning will provide various negative slopes. Hoff et al. (1998) first recognized that other enthalpies besides that of vaporization are more applicable to air-terrestrial or air-water exchange by including Henry’s Law enthalpies. However, they still insisted on vapor pressure (hence the enthalpy of vaporization) to be indicative of local transport. Simcik et al. (1999) concluded that equality between enthalpies of phase change and enthalpies of vaporization is a sufficient, but not a necessary condition to prove local transport. Since what is being observed is not partitioning from the pure liquid phase to the gas phase, but from a sorbed phase to the gas phase, and there is no a priori reason to believe that the enthalpies of these two processes should be equal. Simcik et al. (1999) did not observe the same non-linearity in ClausiusClapeyron plots from other IADN sites, so the same argument could not be applied. However, they used the enthalpies of phase change observed for 50 individual PCB congeners at several sites and compared them to literature values of enthalpies of vaporization. These values were not compared for magnitude, but for correlation of field versus laboratory enthalpies over the range of values represented by the many PCB congeners. The authors concluded that equilibrium partitioning would produce a linear dependency of enthalpies of phase change on enthalpies of vaporization. Using this argument they further concluded that the three U.S. master stations largely are subject to long-range transport, and Chicago is subject to local sources. While the Simcik et al. (1999) study pointed to local sources of PCBs to the Chicago atmosphere, it was not possible to pinpoint them. The AEOLOS study used several techniques to infer specific source information for PAHs, PCBs and metals to the Chicago and adjacent coastal atmosphere (Table II). These include apportioning source types as well as specific point sources within the urban/industrial area. Paode et al. (1999) used EPA’s CMB7 chemical mass balance model to apportion the dry deposited atmospheric particulate matter to several sources using metals. They determined that the major sources included

210

M. F. SIMCIK

TABLE II Source information from the AEOLOS study Contaminant

Technique used

Major sources

Reference:

Atmospheric Particles

CMB

(Paode et al., 1999)

PAHs PCBs

FA/MR PSCF

Light duty gasoline vehicles, soil dust, heavy duty diesel vehicles Coal, natural gas, vehicles Sludge drying beds, landfill, transformer storage yard

(Simcik et al., 1999) (Hsu et al., 2003)

light duty gasoline vehicles, soil dust and heavy-duty diesel vehicles. Other sources included limekilns, coke dust, aluminum foundries, coal-fired power plants, paint spray booths and municipal incineration. While the authors conducted the CMB on several different size particles collected with various dry deposition samplers, the results for course and fine particles were similar. Simcik et al. (1999) used a factor analysis/multiple linear regression model to apportion PAHs to different combustion sources. In general the factor analysis was used to identify tracers of specific sources and the multiple linear regression was used to apportion those tracers to the total PAH concentration measured over the course of the study. Unlike CMB this type of model cannot apportion specific samples, but the aggregate of many samples. Results from Simcik et al. (1999) indicate that the major sources of PAHs to the Chicago and adjacent coastal atmosphere were coal combustion, natural gas combustion and vehicle emissions. These results were supported by the fuel use information for Illinois for the time period of sampling. Hsu et al. (2003) used probability source contribution function (PSCF) to apportion specific geographic sources of PCBs to the Chicago atmosphere. In general, the PSCF model segregates all air concentration measurements of a given contaminant into high and low values (i.e. above or below the mean), constructs a grid over a map, and using back trajectories identifies those grid cells with a higher percentage of trajectories that are associated with high concentrations. The highlighted cells allow one to indicate potential source areas. Hsu et al. (2003) found three areas highlighted in the Chicago area including sludge drying beds, a landfill and a transformer storage yard. These areas were confirmed as being sources by sampling upwind and downwind of each site. The advantages of long-term synoptic monitoring programs over short-term intensive studies is that sites located over a large spatial area can be compared in relation to the relative influence of short-range and long-range transport to atmospheric concentrations over many years. Short-term intensive studies are confined to a single area and a specific point in time. However, given the short period of time


211

that the source apportionment is applicable, the results of determining sources can be produced much more rapidly than in a long-term synoptic study. Additional advantages to the short-term intensive studies are that they can identify specific source areas and categories and do so for many different contaminants.

5. Loadings Since long-term synoptic studies are best suited for regional background monitoring, sites are often chosen far from potential sources and the resulting loadings calculations reflect the regional background. While long-term studies can provide a baseline for regional background loadings, they usually are designed to specifically miss major urban influences. Short-term intensive studies, on the other hand, can be designed to specifically examine the loading that a particular area may have on adjacent water bodies. However, without information from a long-term intensive study the urban load is without context. IADN takes advantage of the long-term synoptic approach because so much of the Great Lakes basin is relatively sparsely populated. IADN combines measured air concentrations with literature values for dissolved phase concentrations in each lake to calculate air-water exchange fluxes and dry and wet deposition loadings (Hillery et al., 1998; Hoff et al., 1996). Because of its long-term monitoring, IADN is also able to track long-term changes in loadings. This becomes most significant for chemicals that have ceased production such as many of the banned pesticides and PCBs. A recent article by Buehler and Hites (2002) reports the temporal trends in Great Lakes loadings for PCBs and alpha and gamma hexachlorocyclohexane (α-HCH and γ -HCH) over the course of sampling for IADN. Both α-HCH and γ -HCH are supersaturated in the atmosphere with respect to the dissolved phase concentration in the lakes, and therefore the overall loading is into the lakes, but there has been a decreasing trend in that loading from 1992 to 1996. PCBs, on the other hand, are supersaturated in the dissolved phase providing a net efflux from the lakes into the atmosphere. In this case as well, there has been a decreasing trend in magnitude of net flux over the period of 1992 to 1996. While the data presented by Buehler and Hites (2002) combined all five Great Lakes, there is considerable difference among loadings to each of the lakes. On an areal basis, Lake Superior has the lowest atmospheric loading while Lakes Michigan and Ontario have the highest. This is not surprising since the watersheds of the lower lakes are much more heavily populated and contain more industry. Therefore, even though the IADN monitoring provides regional background loadings estimates, population and industry in the region impact those estimates. Furthermore, there are pockets of industry that may affect the local atmospheric deposition. Cortes et al. (2000) recorded urban influence on atmospheric PAHs measured at Sturgeon Point resulting from its proximity to Buffalo, NY. Buehler et al. (2001) even noticed an influence from Duluth, MN/Superior, WI on the sampling

212

M. F. SIMCIK

at Brule River in rural Wisconsin just 40 km east of Duluth compared to the Eagle Harbor master site, 400 km to the east. This influence occurred despite the relatively small size of the Duluth/Superior urban/industrial area, because the regional background is so low. From this, one has to wonder what other urban/industrial areas around the coastal Great Lakes are potential sources not detected by the IADN program. AEOLOS looked specifically at the urban influence on the loadings of toxic contaminants to southern Lake Michigan adjacent to Chicago, IL. Elevated overlake concentrations were observed for PAHs and PCBs (Simcik et al., 1997), resulting in elevated dry deposition (Franz et al., 1998), gaseous deposition (Zhang et al., 1999) and wet deposition (Offenberg and Baker, 1997). The relative enhancement of net gas exchange for PCBs was performed by comparing AEOLOS data with wind from the south and southwest and from the north and northeast. Net gas exchange estimates based on AEOLOS data for wind from the north was net volatilization of 30 ± 17 ng m−2 day. (Zhang, 1996). This is a factor of 20 lower than the estimate from IADN data for the same summer (Hillery et al., 1998). The factor of 20 difference could be a result of fewer number of samples taken during AEOLOS with north wind, or may indicate urban influence even during north winds. Unfortunately, due to limited sampling during a short-term intensive study such as AEOLOS, the reason for the difference cannot be determined emphatically. Enhanced dry deposition was also observed for trace metals and elements (Caffrey et al., 1998; Paode et al., 1998; Sofuoglu and Holsen, 1997; Sofuoglu et al., 1998; Zufall et al., 1998). This enhancement was largely attributed to deposition of rather large particles (> 8 µm) (Franz et al., 1998; Paode et al., 1998; Sofuoglu et al., 1998; Zufall et al., 1998). Atmospheric mercury concentrations also showed elevated values over southern Lake Michigan (Landis et al., 2002) leading to higher dry and wet deposition calculations to the extent that atmospheric deposition is the dominant loading to Lake Michigan and that dry and wet deposition is approximately equal to one another (Landis and Keeler, 2002). While the above AEOLOS papers indicate that elevated concentrations and atmospheric deposition occur over southern Lake Michigan, they also concluded that no urban influence was seen at the rural South Haven site across the lake, indicating that the urban influence is confined to the near-shore area. This has been confirmed by results of the Lake Michigan Mass Balance study conducted in 1995 (Green et al., 2000). Since a large urban/industrial area the size of greater Chicago, IL has only a near-shore effect on atmospheric deposition, one might question whether or not other smaller urban/industrial areas around the Great Lakes have any impact at all. The answer lies is in the fact that any increase in the regional background signal would increase near-shore deposition. It has already been mentioned that increases in atmospheric concentrations result from the influence of Duluth/Superior on Lake Superior. Even though this urban/industrial area is much smaller than Chicago, IL, it exerts an influence because the regional background of Lake Superior also is small. This was confirmed by Simcik et al. (2003), who observed increased


213

Figure 3. Urban areas that have the potential to elevate atmospheric deposition over regional background.

loadings to the sediments of the western arm of Lake Superior, presumably from atmospheric deposition. Therefore, in order to identify potential source areas that might affect the atmospheric loadings not recorded at sites representative of regional background levels, it is important to consider the regional background level for context. Assuming that population density, a surrogate for urban pollution, is a good predictor, then one can identify potential sources around the Great Lakes. Here the population density for each county from U.S. and Canadian census data was taken, and areas with densities one standard error above the geometric mean for each lake were highlighted as possible sources of urban influence to the regional backgrounds (Figure 3). The two least populated lakes produced the fewest number of potential source areas. For Lake Superior, two areas are indicated as potential sources, Duluth/Superior and Thunder Bay, and Lake Huron only produced one potential source area of Saginaw/Bay City. Lakes Erie and Ontario had the highest population density along their respective shorelines, and produced the next most urban areas. Lake Michigan had the most urban areas, with the majority centered around the southern end of the lake from Milwaukee, WI to Gary, IN, and one site in the northern part of the lake, Green Bay, WI. These areas stand out because of the relatively sparsely populated northern and eastern shores of Lake Michigan. It appears from Figures 1 and 3 that the IADN study met its goal of placing sites away from the influence of local sources, and that many of the satellite sites are well placed to capture some of the urban influences. However, Figure 3 also points to other potential urban sources that may be considered in the future should IADN

214

M. F. SIMCIK

TABLE III Summary of Information obtained from the IADN and AEOLOS programs

Spatial Coverage

Sources of Contaminants Loadings Temporal Trends

IADN (Long-Term Synoptic)

AEOLOS (Short-Term Intensive)

System-wide description of regional background concentrations Long-range vs. Short-range transport for each lake

Changes over short distances from urban to over-water to rural

Gross loadings from regional background contamination Long-term changes in concentrations and loadings

Specific source categories and specific source areas within one urban area Urban influence on adjacent coastal waters Diurnal variations due to meterologicaland photochemical conditions

wish to expand its coverage. In summary (Table III), long-term synoptic monitoring programs like IADN have advantages over short-term initiatives in providing greater spatial information, the relative influence of long and short-range transport on the regional background, gross loadings representative of the majority of each lake and long-term temporal trends. On the other hand, short-term intensive studies provide more information on the processes governing sources, transport and deposition, like the urban/industrial influence on adjacent large water bodies, specific sources to an urban/industrial area and short-term fluctuations in concentrations due to meteorology, source strength and photochemical reactions. Therefore, these factors should be considered when choosing a monitoring program that best suits a given situation. The Great Lakes community was fortunate to have had both the IADN and AEOLOS projects available to explore these factors, but the greatest impact on Great Lakes air monitoring results from the combining of information. Together IADN and AEOLOS have identified areas within the greater Chicago urban/industrial complex as continuing sources of PCBs to the atmosphere of Lake Michigan. From the results of AEOLOS, IADN loadings estimates can be modified to include other urban/industrial areas. Most importantly future satellite site selection for IADN should take the results of AEOLOS into consideration.

References Buehler, S. S., Basu, I. and Hites, R. A.: 2001, ‘A comparison of PAH, PCB, and pesticide


215

concentrations in air at two rural sites on Lake Superior’, Environ. Sci. Technol. 35, 2417–2422. Buehler, S. S. and Hites, R. A.: 2002, ‘The Great Lakes’ integrated atmospheric deposition network’, Environ. Sci. Technol. 36, 354A–359A. Caffrey, P. F., Ondov, J. M., Zufall, M. J. and Davidson, C. I.: 1998, ‘Determination of size-dependent dry particle deposition velocities with multiple intrinsic elemental tracers’, Environ. Sci. Technol. 32, 1615–1622. Cortes, D. R., Basu, I., Sweet, C. W., Brice, K. A., Hoff, R. M. and Hites, R. A.: 1998, ‘Temporal trends in gas-phase concentrations of chlorinated pesticides measured at the shores of the Great Lakes’, Environ. Sci. Technol. 32, 1920–1927. Cortes, D. R., Basu, I., Sweet, C. W. and Hites, R. A.: 2000, ‘Temporal trends in and influence of wind on PAH concentrations measured near the Great Lakes’, Environ. Sci. Technol. 34, 356–360. Cortes, D. R. and Hites, R. A.: 2000, ‘Detection of statistically significant trends in atmospheric concentrations of semivolatile compounds’, Environ. Sci. Technol. 34, 2826–2829. Dzubay, T. G., Stevens, R. K., Gordon, G. E., Olmez, I., Sheffield, A. E. and Courtney, W. J. A.: 1988, ‘Composite receptor method applied to Philadelphia aerosol’, Environ. Sci. Technol. 22, 46–52. Franz, T. P., Eisenreich, S. J. and Holsen, T. M.: 1998, ‘Dry deposition of particulate polychlorinated biphenyls and polycyclic aromatic hydrocarbons to Lake Michigan’, Environ. Sci. Technol. 32, 3681–3688. Green, M. L., DePinto, J. V., Sweet, C. and Hornbuckle, K. C.: 2000, ‘Regional spatial and temporal interpolation of atmospheric PCBs: Interpretation of Lake Michigan mass balance data’, Environ. Sci. Technol. 34, 1833–1841. Hillery, B. R., Basu, I., Sweet, C. W. and Hites, R. A.: 1997, ‘Temporal and spatial trends in a longterm study of gas-phase PCB concentrations near the Great Lakes’, Environ. Sci. Technol. 31, 1811–1816. Hillery, B. R., Simcik, M. F., Basu, I., Hoff, R. M., Strachan, W. M. J., Burniston, D., Chan, C. H., Brice, K. A., Sweet, C. W. and Hites, R. A.: 1998, ‘Atmospheric deposition of toxic pollutants to the Great Lakes as measured by the integrated atmospheric deposition network’, Environ. Sci. Technol. 32, 2216–2221. Hoff, R. M., Strachan, W. M. J., Sweet, C. W., Chan, C. H., Shackleton, M., Bidleman, T. F., Brice, K. A., Burniston, D. A., Cussion, S., Gatz, D. F., Harlin, K. and Schroeder, W. H.: 1996, ‘Atmospheric deposition of toxic chemicals to the Great Lakes: A review of data through 1994’, Atmos. Environ. 30, 3505–3527. Hoff, R. M., Brice, K. A. and Halsall, C. J.: 1998, ‘Nonlinearity in the slopes of Clausius-Clapeyron plots for SVOCs’, Environ. Sci. Technol. 32, 1793–1798. Hsu, Y.-K., Holsen, T. M. and Hopke, P. K.: 2003, ‘Locating and quantifying PCB sources in Chicago: Receptor modeling and field sampling’, Environ. Sci. Technol. 37, 681–690. Kao, A. S. and Friedlander, S. K.: 1995, ‘Frequency distributions of PM10 chemical components and their sources’, Environ. Sci. Technol. 29, 19–28. Kleinman, M. T., Pasternack, B. S., Eisenbud, M. and Kneip, T. J.: 1980, ‘Identifying and estimating the relative importance of sources of airborne particulates’, Environ. Sci. Technol. 14, 62–65. Kowalczyk, G. S., Gordon, G. E. and Rheingrover, S. W.: 1982, ‘Identification of atmospheric particulate sources in washington, D.C., using chemical element balances’, Environ. Sci. Technol. 16, Landis, M. S. and Keeler, G. J.: 2002, ‘Atmospheric mercury deposition to Lake Michigan during the Lake Michigan mass balance study’, Environ. Sci. Technol. 36, 4518–4524. Landis, M. S., Vette, A. F. and Keeler, G. J.: 2002, ‘Atmospheric mercury in the Lake Michigan basin: Influence of the Chicago/Gary urban area’, Environ. Sci. Technol. 36, 4508–4517. Lewis, C. W., Baumgardner, R. E., Stevens, R. K. and Russwurm, G. M.: 1986, ‘Receptor modeling study of Denver winter haze’, Environ. Sci. Technol. 20, 1126–1136.

216

M. F. SIMCIK

Morandi, M. T., Daisey, J. M. and Lioy, P. J.: 1987, ‘Development of a modified factor analysis/multiple regression model to apportion suspended particulate matter in a complex urban airshed’, Atmos. Environ. 21, 1821–1831. Offenberg, J. H. and Baker, J. E.: 1997, ‘Polychlorinated biphenyls in Chicago precipitation: Enhanced wet deposition to near-shore Lake Michigan’, Environ. Sci. Technol. 31, 1534–1538. Offenberg, J. H. and Baker, J. E.: 2002, ‘Precipitation scavenging of polychlorinated biphenyls and polycyclic aromatic hydrocarbons along an urban to over-water transect’, Environ. Sci. Technol. 36, 3763–3771. Paode, R. D., Sofuoglu, S. C., Sivadechathep, J., Noll, K. E., Holsen, T. M. and Keeler, G. J.: 1998, ‘Dry deposition fluxes and mass size distributions of Pb, Cu, and Zn measured in Southern Lake Michigan during AEOLOS’, Environ. Sci. Technol. 32, 1629–1635. Paode, R. D., Shahin, U. M., Sivadechathep, J., Holsen, T. M. and Franek, W. J.: 1999, ‘Source apportionment of dry deposited and airborne coarse particles collected in the Chicago area’, Environ. Sci. Technol. 31, 473–486. Pratsinis, S. E., Zeldin, M. D. and Ellis, E. C.: 1988, ‘Source resolution of the fine carbonaceous aerosol by principal component – Stepwise regression analysis’, Environ. Sci. Technol. 22, 212– 216. Simcik, M. F. Zhang, H., Eisenreich, S. and Franz, T. P. 1997, ‘Urban contamination of the Chicago/Coastal Lake Michigan atmosphere by PCBs and PAHs during AEOLOS’, Environ. Sci. Technol. 31. 2141–2147. Simcik, M. F., Basu, I., Sweet, C. W. and Hites, R. A.: 1999, ‘Temperature dependence and temporal trends of polychlorinated biphenyl congeners in the Great Lakes atmosphere’, Environ. Sci. Technol. 33, 1991–1995. Simcik, M. F., Eisenreich, S. J. and Lioy, P. J.: 1999, ‘Source apportionment and source/sink relationships of PAHs in the coastal atmosphere of Chicago and Lake Michigan’, Atmos. Environ. 33, 5071–5079. Simcik, M. F., Hoff, R. M., Strachan, W. M. J., Sweet, C. W., Basu, I. and Hites, R. A.: 2000, ‘Temporal trends of semivolatile organic contaminants in Great Lakes precipitation’, Environ. Sci. Technol. 34, 361–367. Simcik, M. F., Jeremiason, J. D., Lipiatou, E. and Eisenreich, S. J.: 2003, ‘Enhanced removal of hydrophobic organic contaminants by settling sediments in western Lake Superior’, J. Great Lakes Res. 29, 41-53. Sofuoglu, S. C. and Holsen, T. M.: 1997, ‘Dry deposition fluxes and mass size distributions of trace elements measured near southern Lake Michigan’, Environ. Res. Forum 7–8, 691–696. Sofuoglu, S. C., Paode, R. D., Sivadechathep, J., Noll, K. E., Holsen, T. M. and Keeler, G. J.: 1998, ‘Dry deposition fluxes and atmospheric size distributions of mass, Al, and Mg measured in southern Lake Michigan during AEOLOS’, Environ. Sci. Technol. 29, 281–293. Tuncel, S. G., Olmez, I., Parrington, J. R., Gordon, G. E. and Stevens, R. K.: 1985, ‘Composition of fine particle regional sulfate component in Shenandoah Valley’, Environ. Sci. Technol. 19, 529–537. Wania, F., Haugen, J.-E., Lei, Y. D. and Mackay, D.: 1998, ‘Temperature dependence of atmospheric concentrations of semivolatile organic compounds’, Environ. Sci. Technol. 32, 1013–1021. Zhang, H.: 1996. Enhanced Air-Water Exchange of Polychlorinated Biphenyls in Southern Lake Michigan in the Chicago Plume. Master of Science. Civil Engineering, Universtiy of Minnesota. Zhang, H., Eisenreich, S. J., Franz, T. R., Baker, J. E. and Offenberg, J. H.: 1999, ‘Evidence for increased gaseous PCB fluxes to Lake Michigan from Chicago’, Environ. Sci. Technol. 33, 2129– 2137. Zufall, M. J., Davidson, C. I., Caffrey, P. F. and Ondov, J. M.: 1998, ‘Airborne concentrations and dry deposition fluxes of particulate species to surrogate surfaces deployed in southern Lake Michigan’, Environ. Sci. Technol. 32, 1623–1628.