Analysis of Trace Elements and Ions in Ambient Fine Particulate Matter at Three Elementary Schools in Ohio
By John, Kuruvilla; Karnae, Saritha; Crist, Kevin; Kim, Myoungwoo; Kulkarni, Amol
ABSTRACT
The results from a chemical characterization study of fine particulate matter (PM^sub 2.5^) measured at three elementary schools in Central and Southeast Ohio is presented here. PM^sub 2.5^ aerosol samples were collected from outdoor monitors and indoor samplers at each monitoring location during the period of February 1, 1999, through August 31, 2000. The locations included a rural elementary school in Athens, OH, and two urban schools within Columbus, OH. The trace metal and ionic concentrations in the collected samples were analyzed using an X-ray fluorescence spectrophotometer and ion chromatography unit, respectively. Sulfate ion was found to be the largest component present in the samples at all three of the sites. Other abundant components included nitrate, chloride, ammonium, and sodium ions, as well as calcium, silicon, and iron. The average PM^sub 2.5^ concentrations showed similar temporal variations among the three sites within the study region. PM^sub 2.5^ and its major component, sulfate ion, showed strong seasonal variations with maximum concentrations observed during the summer at all three of the sites. The indoor environment was found to be more contaminated during the spring months (March through May) at New Albany (a suburb of Columbus, OH) and East Athens (rural Ohio area). Potential source contribution function analysis showed that particulate matter levels at the monitoring sites were affected by transport from adjoining urban areas and industrial complexes located along the Ohio River Valley. A preliminary outdoor source apportionment using the principal component analysis (PCA) technique was performed. The results from the PCA suggest that the study region was primarily impacted by industrial, fossil fuel combustion, and geological sources. The 2002 emissions inventory data for PM^sub 2.5^ compiled by Ohio Environmental Protection Agency also showed impacts of similar source types, and this was used to validate the PCA analysis.
(ProQuest-CSA LLC: … denotes formulae omitted.)
INTRODUCTION
Fine particles in the ambient atmosphere can penetrate to the deepest parts of the lungs and, thus, significantly impact human health. Various epidemiological studies conducted have provided evidence for an association between acute particulate matter (PM) exposures and increases in mortality and morbidity among people suffering with respiratory and cardiovascular diseases.1,2 Other health problems associated with acute exposures to PM included acute asthma exacerbations; bronchitis; acute and chronic respiratory symptoms, such as shortness of breath and painful breathing; increases in the number of hospital admissions for cardiovascular problems, including arrhythmia, myocardial infarction, congestive heart failure, and acute coronary events; and premature deaths.3-11
Measurements conducted in several urban atmospheric environments have shown an increasing trend in the concentration of highly dispersed aerosols, that is, aerosols with particle diameters <2 m, and a decrease in the concentration of coarse particles (diameter <10 m).12 In 1997, U.S. Environmental Protection Agency (EPA) proposed new National Ambient Air Quality Standards (NAAQS) for ozone (O^sub 3^) and PM.13 These standards were based on multiyear scientific assessments that linked health effects to present air pollution levels. The standards, which tighten the requirements for attainment, will have significant economic and societal impact for Ohio. For example, according to the new fine PM (PM^sub 2.5^) standard, estimates indicate that 26 counties in Ohio are projected to be in nonattainment. In comparison, currently only one county in Ohio does not meet the existing NAAQS for coarse PM (PM^sub 10^). There are limited data on PM^sub 2.5^ concentrations and its constituents in Ohio. The few historical studies that have been conducted were health-based studies centered on areas with historically high levels of air pollutants, such as Steubenville.14
Because trace metals and ionic concentrations of PM were found to have adverse health effects, the primary focus of the study was to identify the key trace metals and ionic components in indoor and outdoor PM^sub 2.5^ that may be harmful to the children in Ohio.15,16 The study will also provide data for a better understanding of health effects on the general population. PM^sub 2.5^ aerosol samples were collected from an outdoor monitor and an indoor sampler at three elementary schools in Central and Southeastern Ohio between February 1, 1999 and August 31, 2000. The monitoring locations included a rural elementary school in Athens, OH, and two urban school settings, both within Columbus, OH. The trace metal and ionic concentrations in the collected samples were analyzed using an X-ray fluorescence spectrophotometer and ion chromatography (IC) system, respectively. The major objectives included an extensive review of indoor and outdoor PM^sub 2.5^ characteristics within the study area, especially the spatial and temporal distribution of PM^sub 2.5^ and its components. Potential source contribution function (PSCF) analysis, an aid in complex decisions regarding atmospheric transport pathways and identification of pollutant source region, was used to identify key upwind source regions affecting the observed levels of PM^sub 2.5^. Another key objective was to identify major local source categories of PM^sub 2.5^ and its components impacting ambient outdoor atmosphere in the study region using source apportionment techniques, such as principal component analysis (PCA). The results from this study can potentially benefit policy-makers in Ohio to understand local and regional pollution sources, its impact on PM air quality, and its effective management.
EXPERIMENTAL WORK
PM^sub 2.5^ Measurements
Site Locations. The field experiment, which began in February 1999, consists of three longitudinal studies involving a rural elementary school location in Athens and two urban school settings, both within Columbus. Two contrasting sites in Columbus were established: one urban school site located in the south-central side of Columbus (Koebel Elementary School, 39.9440 latitude and – 82.9575 longitude) and one suburban school site located to the northeast of Columbus (New Albany Elementary School, 40.0850 latitude and -82.8160 longitude). Columbus, like numerous other metropolitan areas in the state and across the nation, has historically met the NAAQS for O^sub 3^ and PM^sub 10^ but has the potential of being in nonattainment of the new PM^sub 2.5^ and the new 8-hr averaged O^sub 3^ standards. Koebel is located in the industrial center of the city that includes foundries, plastic facilities, gravel/quarrying operations, and other manufacturing industries. This site is located within 0.3 mi of a major transportation artery. The New Albany site is located ~5 mi northeast of Columbus and is ~20 mi from the Koebel site. The rural site, Athens (East Elementary School, 39.3194 latitude and -82.8111 longitude), was chosen because of its remoteness and proximity to the Ohio River Valley region. Athens is located ~20 mi from the Ohio River Valley, which has numerous coal-fired power generation facilities and industrial operations. It has been used as an upwind remote rural site for the U.S. Department of Energy’s Upper Ohio River Valley Project, a comprehensive PM^sub 2.5^ and precursor gas- monitoring program initiated in early 1999. The monitoring sites selected typify urban, suburban, and rural locations as shown in Figure 1.
Sample Collection. Continuous ambient PM^sub 2.5^ measurements were carried out using the tapered element oscillating microbalance (TEOM) manufactured by Rupprecht & Patashnick Company. The TEOMs were set to run 24 hr and 7 days per week. Filter samples were collected on 47-mm-diameter Whatman Teflon filters (2-m pores size) using the automatic cartridge collection unit (ACCU) connected externally to the TEOM. The ACCU system was operated on a 24-hr cycle from Monday through Friday. Leak check using a flow adapter and flow check using a Dry-Cal gas flow meter were performed with the change of sensor filter. Field blanks were collected to account for the background concentrations during gravimetric sampling.
Indoor monitors were operated at 10 L/min using flow-controlled indoor sampling pumps (URG model 3000-02Q). Measurements of indoor PM^sub 2.5^ concentrations were made using 2.5-m cyclones (URG model 2000-30EH). Indoor monitors were timed to run from 8:00 a.m. to 3:00 p.m. on Mondays through Fridays throughout the school year. The indoor filter samples were collected on 37-mm-diameter Whatman Teflon filters (2-m pores size). Filters so collected were placed in Petri dishes, double bagged, and stored at 4 C until analysis. As the schools were closed for summer holidays, indoor filter samples could not be collected during June through August 1999.
Gravimetric Measurements. Filters were weighed in a temperature- and humidity-controlledmicroenvironment (environmentally controlled glove box, PLAS-LABS). The Teflon filters were equilibrated before weighing under controlled conditions (22.5 2.5 C and 35 5% relative humidity). New filters were inspected a month before sampling and conditioned in the glove box, then weighed (preweigh). Each filter was weighed again after sampling (postweigh). Before the postweigh, filters were inspected, and all of the damages were recorded. These filters were weighed within 48 hr after the 24-hr conditioning. Each filter was weighed both before and after sampling using a Sartorius analytical microbalance (MC5 UL), with a readability of 1 g at the Air Quality Research Laboratory in Ohio University. Mass of field blanks measured was subtracted from the filter mass to obtain the mass of PM^sub 2.5^ collected on the filter. The filter samples were then shipped to the Department of Environmental Engineering at Texas A&M University-Kingsville for chemical characterization analysis. The trace metal and ionic concentrations in the collected samples were analyzed using an X- ray fluorescence spectrophotometer and an ion chromatograph, respectively.
Sample Analysis. The analysis of trace elements was performed using the Kevex 771-Energy Dispersive X-ray fluorescence instrument (ED-X-ray fluorescence). X-ray fluorescence is a nondestructive technique that can analyze elements from fluorine to uranium in the periodic table. The instrument consists of a spectrometer, secondary targets, rhodium target X-ray tube, and high-resolution Si (Li) solid-state X-ray detector. Filter samples were analyzed under atmospheric environment using secondary excitation with germanium as the secondary target; tube voltage of 50 kV; tube current of 2.9 mA; and a counting time of 100 sec. The intensities measured were converted into elemental concentrations using a nonlinear iterative least-square analysis. Elements found in the samples included silicon, phosphorus, sulfur, chlorine, potassium, calcium, titanium, vanadium, chromium, manganese, iron, cobalt, nickel, copper, zinc, arsenic, cadmium, and tin. The filter samples were wet with 0.5 mL of methanol and then extracted with 15 mL of deionized water for 10 min using the Fisher Scientific FS9 ultrasonic bath. Watersoluble anions and cations present in the samples were then determined using the Dionex DX-500 IC system. The Dionex system included an electrochemical detector, chromatography oven, gradient pump, and an auto sampler. An anion exchange column, anion guard column, anion eluent, and anion suppressor were used for the anion analysis; a cation exchange column, cation guard column, cation eluent, and cation suppressor were used for the cation analysis. The identification of compounds was based on the retention time in the IC column, and the quantification was made using calibration curve acquired with the external standards. ED-X-ray fluorescence and IC were calibrated for every 15 sets of samples. Laboratory blanks and field blanks were analyzed to account for the background concentrations in the laboratory and field.
Data Analysis
Elevated PM^sub 2.5^ concentrations were observed during September 1999 through August 2000 at all three of the monitoring sites. Concentrations during the summer and spring months were notably higher than the remainder of the study period. Figure 2 shows the monthly average PM^sub 2.5^ concentrations for the New Albany site (bypass flow, filter based; main flow, TEOM). Quantitative analysis of the bypass and main flow PM^sub 2.5^ concentrations of TEOM showed overall a similar trend. Minor variations in the monthly concentrations were observed, which could be because of the fact that the main flow channel samples continuously, whereas samples from bypass flow channel were collected on a 5-days-per-week basis.
According to a study conducted in the Upper Ohio River Valley during October 1999 through September 2000 on PM^sub 2.5^ mass and composition measured at the National Energy Technology Laboratory Pittsburgh, PA, site, and at sites in Ohio, including Steubenville, Columbus, and Athens, long-range transport was a significant source contributing to elevated PM levels in the study region. Ohio River Valley and possible urban regions beyond were found to be the significant sources of PM and its precursors in the Pittsburgh area and at other regional sites included in the study.17 The Steubenville Comprehensive Air Monitoring Program Study noted that PM concentrations >35 g/m3 were observed at a site in Steubenville, PA.18 Based on the results from previous studies, continuous PM^sub 2.5^ mass concentration and speciation data were collected at the three monitoring sites located in Central and Southeast Ohio, and the data were used in a source apportionment assessment of regional and local sources using PSCF analysis and PCA in this study.
PSCF Analysis. PSCF was originally developed by Ashbaugh et al.19 and has been used in many previous applications to locate source areas.20-22 In this study, PSCF was used to help identify probable geographic locations of PM^sub 2.5^ concentrations at the three school sites. Backward trajectories from the three schools sites were computed for every hour of each of the sampling time periods to obtain a sufficient number of end points. All of the trajectories were calculated using the HYSPLIT model23 with National Centers for Environmental Prediction/National Center for Atmospheric Research reanalysis data.24 Twenty-four-hour backward trajectories with a starting height of 500 m above the model ground level were typically used. These backward trajectories were used to compute the PSCF values for PM^sub 2.5^ concentration measured at the three school sites.
The geographical region covered by the backward trajectories was divided into 0.2 0.2 grid cells. For PM^sub 2.5^ concentrations, the PSCF for the ijth grid cell, PSCF^sub ij^, is computed as follows:
… (1)
where n^sub ij^ is the total number of trajectory segment end points that falls into the ijth cell, and mij is the number of segment end points in the same cell corresponding with trajectories associated with concentration values at each receptor site exceeding a prespecified criterion value. Therefore, cells with a high PSCF value are likely to produce high concentration values at the receptor sites, so they are reasonably assumed to be probable source areas. In this study, the criteria values were the overall geometric mean of PM^sub 2.5^ concentration at each receptor site. To minimize the high uncertainties that result from small values of n^sub ij^, all of the PSCF values were multiplied by an arbitrary weight function factor, W(n^sub ij^), to downweigh those values for which n^sub ij^ was less than three times the average n^sub ij^:
… (2)
PCA. PCA, a multivariate statistical analysis technique, has been widely applied to source apportionment of various air pollutants.25,26 The principal components (PCs) induced by PCA are commonly considered as source categories contributing to pollutant concentrations at a receptor site.27 It has been applied for source apportionment of elemental composition and heavy metals present in PM.28,29 PCA was also used for source apportionment studies conducted by Park and Kim30 in an urban site in Seoul, Korea, for a dataset of 62 samples and 15 variables.
To evaluate the source types contributing to the measured trace elements and ions at the three monitoring sites in Ohio, PCA was used as the source apportionment tool. The components with concentrations less than the detection limit were excluded in the PCA. The effects of nondetects have not been explored in this preliminary source apportionment study. On an average, the dataset consisted of 70 samples and 26 variables for each site. The results of the PCA provided the number of PCs (factors) that basically are eigenvectors obtained from a correlation matrix of variables or species. The entries in each row (factor loadings) indicate the correlation between the given factor and the given variable. Variance extracted by each eigenvector is the eigenvalue of the matrix. The data was VARIMAX rotated, and eigenvectors with variance (eigenvalue) <1 were excluded from the results applying the Kaiser criterion. Each PC with a simple structure clearly assesses the attribute that is reflected by the variables for which the coefficients are high. Those variables that have low coefficients (<0.4) were not used in interpreting the PC. Previous PM studies in Central and Southeast Ohio were limited in scope, and no archive was available for local PM^sub 2.5^ pollution source profiles. Therefore, in this study, the PCs were interpreted using relevant source profiles from EPA's SPECIATE database and from other similar studies.31-36 EPA's SPECIATE database includes source profiles from emission characterization and emission inventory studies in the Ohio River Valley and neighboring regions. It also includes composite source profiles. The project scope did not include a comprehensive carbon analysis because of resource constraints, and, thus, elemental and organic carbon (OC) was not used in this study. Thus, it is likely that presence of some of the key elemental carbon and OC sources, such as mobile emissions, may not be revealed by this analysis.
RESULTS AND DISCUSSION
Spatial and Seasonal Variations of Indoor and Outdoor PM^sub 2.5^
The chemical characterizations of the outdoor and indoor PM^sub 2.5^ concentrations, as well as the spatial and temporal variations, were studied. The average PM^sub 2.5^ concentrations considered here are the arithmetic averages of the filter mass collected at each site during the entire period of study. Arithmetic averages have been used to be consistent with the PM NAAQS. Outdoor PM^sub 2.5^ data were averaged from 8:00 a.m. to 3:00 p.m. to be consistent with the indoor PM^sub 2.5^ data. Average outdoor PM^sub2.5^ concentrations showed spatial variations, indicating homogeneity in the spatial distribution of PM^sub 2.5^ in the study region. Table 1 shows spatial correlation coefficients of outdoor PM^sub 2.5^ concentration and distance between the sites. The R^sup 2^ values were found to range from 0.7192 (see Figure 3) to 0.4562. The R^sup 2^ values showed an inverse relationship with the distance between two monitoring sites.
Figure 4 shows indoor and outdoor PM^sub 2.5^ variation during the study period at all of the sites. The mean indoor PM^sub 2.5^ concentrations at the suburban and rural sites were higher than those observed outdoors at these sites, whereas the outdoor PM^sub 2.5^ concentration was higher than the indoor PM^sub 2.5^ level at Koebel (an urban site). However, these patterns were not consistently observed throughout various seasons of the years. Seasonal variation of indoor and outdoor PM^sub 2.5^ concentrations is shown in Figure 5. Indoor PM^sub 2.5^ concentrations were higher than outdoor only during the spring months (March to May) at New Albany (suburban) and Athens (rural) sites. During the summer (June to August), fall (September to November), and winter (December to February) months, PM^sub 2.5^ levels at all of the outdoor sites were consistently higher than those observed indoors at these sites.
Chemical Component Concentrations
In Table 2, overall averages (February 1999 to August 2000), standard deviation, maximum PM^sub 2.5^ mass, and chemical component concentrations are shown. The average PM^sub 2.5^ mass at the outdoor sites varied from 12.88 g/m^sup 3^ at East Athens (rural site) to 13.55 g/m^sup 3^ at Koebel (urban site). The maximum 24-hr averaged outdoor PM^sub 2.5^ concentrations of 61.34 g/m^sup 3^ was observed at New Albany, the central suburban location. The maximum 24-hr average outdoor PM^sub 2.5^ concentrations at Koebel and East sites were 59.24 g/m^sup 3^ and 61.12 g/m^sup 3^, respectively. Anions (F^sup -^, Cl^sup -^, NO^sub 3^^sup -^, SO^sub 4^^sup -2^, and PO^sub 4^^sup -3^) and cations (Li^sup +^, Na^sup +^, NH^sub 4^^sup +^, K^sup +^, Mg^sup +2^, and Ca^sup +2^) present in the samples were then determined using the IC system. Sulfate ion was found to be the largest component present in the samples at all three of the sites. Other abundant components included nitrate, chloride, ammonium, and sodium ion, as well as calcium, silicon, and iron. The average outdoor sulfate concentrations were slightly higher than the average indoor concentrations at each site and varied between 2.12 g/m^sup 3^ at New Albany to 2.71 g/m^sup 3^ at Koebel. The highest 24-hr sulfate concentration among all of the sites was observed at the Koebel indoor site (24.51 g/m^sup 3^). Average nitrate levels for the outdoor sites varied between 0.20 g/ m^sup 3^ and 0.56 g/m^sup 3^ at New Albany and East, respectively, and from 0.46 g/m^sup 3^ to 0.72 g/m^sup 3^ at East and Koebel indoor sites, respectively. A maximum 24-hr average concentration of 4.24 g/m^sup 3^ was observed at the New Albany indoor site. The average anion and cation concentrations observed in the samples in a decreasing order are shown below:
SO^sub 4^^sup -2^ > NO^sub 3^^sup -^ > Cl^sup -^ (3)
NH^sub 4^^sup +^ > Ca^sup +2^ > Na^sup +^ > K^sup +^ > Mg^sup +2^ (4)
Relatively high levels of sodium, chloride, and potassium were found in the rural samples when compared with the urban samples. Heavy metals, such as titanium, vanadium, manganese, iron, copper, and zinc, were found in all of the samples, and iron was the most abundant element in the filter samples. Because of artifacts of the analytical procedures, concentrations of ions like K^sup +^ were found to be more than the concentration of its elemental form.
Temporal Variation in PM^sub 2.5^ and Sulfate Concentrations
In the Eastern United States, PM tends to be more acidic because of the presence of significant fractions of sulfates and nitrates.37 Local sources play a major role in the emissions of sulfates, particularly in this study region. There are >50 coal-fired plants, heavy and light industries, and transportation emission sources near the monitoring sites within the study region.38
Temporal variations in PM^sub 2.5^ mass and its chemical components were analyzed by using 30-day moving averages for the study period (February 1999 to August 2000). Sulfate ion was found to be the most significant fraction of the PM^sub 2.5^ mass. Its variations with PM^sub 2.5^ mass at the New Albany outdoor site are shown in Figure 6. The peak concentrations of sulfate ions often coincided with the peak PM^sub 2.5^ concentrations. Sulfate constitutes a major fraction of summertime PM^sub 2.5^ in Ohio. As shown in Figures 5 and 6, the PM^sub 2.5^ and sulfate ion concentrations tend to be low during the late fall (October and November) and early winter (December and January) months. PM^sub 2.5^ concentrations at Koebel and East Athens monitoring sites showed similar but more dampened seasonal variations than the sulfate concentrations as shown in Figure 7. Speciated indoor data for June through August 1999 were not collected, because the school was not in session during this period. As illustrated in Figure 7, the sulfate concentrations gradually increased from the winter months (December through February) to the summer months (May through August) at the New Albany site. Sulfate ion concentrations also showed a seasonal variation that was strongly influenced by meteorological factors, such as wind speed, wind direction, and humidity levels within the study region. In general for all of the sites, lower sulfate concentrations from October to March and higher concentrations during the summer months were observed.
PSCF Analysis
The PSCF analysis in this study was applied to the observed PM^sub 2.5^ data at the three monitoring sites for 1999- 2000 to produce probability maps of PM^sub 2.5^ source locations. The results of PSCF analysis for the three monitoring sites during 1999- 2000 are presented in Figures 8-10. The analysis of the data from the New Albany site showed that the highest level (red) of PSCF values was observed over Columbus, Dayton, and Cincinnati urban areas and the Ohio River Valley, where major coal-fired power plants and industrial emission facilities are located. The second highest (orange) level of PSCF values were observed over the major urban areas in adjoining states, including the eastern portion of Indiana, Northern Kentucky, Western Pennsylvania, and the northwestern region of West Virginia. As shown in Figure 9, the potential source areas of the Koebel site were very similar to those of the New Albany site, suggesting that both sites located in Columbus were impacted by sources from major urban areas in Ohio and adjacent states, as well as industrial sources in the Ohio River Valley region. The PSCF plot for the East Athens site, as shown in Figure 10, suggests that the major sources in the Ohio River Valley and key urban areas in Western Ohio and Northern Kentucky had the largest impact on observed PM^sub 2.5^ levels. These results indicate that high- probability source regions for each monitoring location in Central and Southeast Ohio were commonly identified to be major urban areas in Ohio and surrounding states, including Indiana, Kentucky, West Virginia, and Pennsylvania, as well as the industrialized Ohio River Valley region. Also, because high levels of PSCF values were observed over the Ohio River Valley, the analyses indicated that the Ohio River Valley acts as one of the main regional sources of PM and its precursors in Ohio. Coal-fired plants located near the monitoring sites are potentially major contributors of the elevated sulfate levels in PM observed at these sites.
Preliminary Source Apportionment Using PCA
PCA for Koebel. The Koebel site is located in the south side of Columbus, the most populous city in the state of Ohio and one of the nation’s fastest growing metropolitan areas. This site is in the industrial center of the city that includes foundries, plastic facilities, gravel/quarrying operations, and other manufacturing industries.31 The monitoring site is located within 0.3 mi of a major transportation artery. Table 3 shows PCA results for the Koebel outdoor monitoring site. The factor loadings >0.7 are shown in footnote a.
Eight factors (PCs) explain 71% of total variance of the Koebel outdoor data, and the first three factors represent >43% of the total variance. The Koebel dataset consisted of 74 samples and 26 variables. The factors appear to represent mixed sources or a broad source category rather than specific individual point sources, which are difficult to interpret. The inclusion of unrelated sources in a PC is because of the mixing in the atmosphere of multiple source emissions. PCA could not separate the influence of sources located near each other or consistently along the same air mass trajectory to the receptor site. The extent to which this commingling of sources occurs is dependent on the relative locations of the sources and the receptor site.32
The first factor (factor 1) was found to have high loadings of Cd, Cr, As, and Ni and appears to be an industrial source. Industrial smelting operations were the major sources of Cd and As. They are also emitted from refineries, furnaces, and metal production activities. Ni and SO^sub 4^^sup -2^ are generally associated with oil-fired power plants, but they also originate from other sources, such as steel manufacturing, steel foundry, and in minor concentrations from other industrial activities.31 P and K with moderate factor loadings also suggest the presence of an industrial source, because they are typically associated with higher factor loadings of Cd and As.
Factor 2 most probably represents a mixed urban source. Cu is one of the markers for incineration, municipal incineration, or sewage sludge incineration.34 Emissio\ns from municipal incinerators are heavily enriched in chlorides, which is typically associated with this PC under moderate factor loadings. These emissions arise mainly from the combustion of plastics and metals that form volatile chlorides. A composite incineration profile also includes Ti.31 Mn is emitted in small concentrations from various industrial sources, and NO^sub 3^^sup -^ indicates an industrial combustion source.
The third factor (factor 3) was found to have high loadings of Si, S, and Cl and moderate loadings of K and Ca. The elements such as Si, K, and Ca are found in soil or crustal profiles.31 However, elements such as Si, S, and Ca in the factor can also be correlated with fossil fuel combustion. Cl is observed in external coal combustion source profiles as well.31
Factor 4, with high loadings of NH^sub 4^^sup +^ and Zn, indicates a waste disposal source associated with certain industrial activities. Source categories for NH^sub 3^ are generally divided into emissions from undisturbed soil (natural) and emissions that are related to human activities, such as fertilized lands and domestic and farm animal waste. Zn is another incineration marker other than Cu mentioned in factor 2.34 Domestic waste is incinerated in municipal incinerators or other industrial incinerators that contain various metals, including Zn. Zn is also used as an additive in plastics or rubber.
The fifth factor (factor 5) with high loadings of Co and Sn and moderate loadings of P, K, and SO^sub 4^^sup -2^ is difficult to interpret and is possibly another mixed industrial source. Factor six contains high loadings of Mg^sup +2^ and moderate loadings of Cl^sup -^ and Na^sup +^. The apparent source of water-soluble ions in this factor is road salt mixed with sand. NaCl is by far the most popular of the de-icers, because it is inexpensive, reliable, and easy to store and apply.
Factor 7 contains K^sup +^ with high loading and Ca^sup +2^ and V with moderate loadings. These ions are constituents of a geological source. K^sup +^ is normally considered to be responsible for combustion of the agricultural waste or biomass burning, but because it is associated with other soil and crustal constituents, it most likely indicates a geological source.31 The eighth factor (factor 8) contains Fe with high loadings and SO^sub 4^^sup -2^ with moderate loadings. Iron is normally a crustal marker, whereas SO^sub 4^^sup – 2^ is found in average soil profiles.31,35 This factor is also associated with coal-fired combustion sources. The two chemical components are found in varying degrees in the emissions from the coal-fired power plant.
PCA for New Albany. The New Albany site is located ~5mi northeast of Columbus and is ~20 mi from the Koebel site. Table 3 displays eight factors determined by PCA for the New Albany outdoor site. The New Albany dataset consisted of 80 samples and 26 variables. Eight factors represent >71% of the total variance. The first three factors represent >43% of the variance.
The first factor (factor 1) contains high loadings of Na^sup +^, K^sup +^, Ca^sup +2^, Cl^sup -^, and NO^sub 3^^sup -^ and moderate loadings of Cu and Zn. The source is possibly from an incinerator and also is associated with crustal sources. As discussed earlier in PCA for the Koebel site, chloride is abundantly found in average incineration profiles, whereas Zn and Cu are also indicators of incineration (municipal or sewage sludge incineration). Na^sup +^, K^sup +^, and Ca^sup +2^ correlate well with soil or crustal sources.
The second factor (factor 2) with high loadings of Cr and moderate loadings of P, Mn, Ni, As, Cd, and Sn appears to be an industrial source. Factor 3 contains V, Ti, and Mg^sup +2^. This combination of trace elements and ions most likely represents a crustal or soil source. The fourth factor (factor 4) contains high loadings of Si, S, and Cl. The apparent source for these elements is a fossil fuel combustion source, more likely coal-fired power plants. Si is one of the major components of coal combustion source profiles.29 Factor 5 has high loadings of K and Co and moderate loadings of P, Mn, and Ni. Ni and Co may result from smelting activities or heavy fuel oil combustion. Other elements in this factor, mainly K, Co, and Mn, can also be associated with geological sources. The factor, therefore, represents a mixed industrial and geological source. The sixth factor (factor 6) contains NH^sub 4^^sup +^ and Ca and is possibly a waste disposal, recycling source. The chemical components in this factor are also associated with biomass burning. Factor 7 contains only sulfate. Sulfate is a major particulate constituent released by coalfired power plants. Hence, the source is most likely a power plant. Factor 8 contains Fe, Zn, Cd, and Sn. Fe is normally a crustal source marker.31 The other elements in this factor indicate industrial activities such as incineration, smelting, and metal production.
New Albany is a bedroom community of Columbus located northwest of the urban core and has few commercial facilities and no significant industrial operations.39 The presence of industrial and other sources suggested by PCA at New Albany indicates the transport of PM^sub 2.5^ precursors from within the greater Columbus region, as well as sources along the Ohio River Valley region. This is consistent with the direction of prevailing winds (south and southwest) at New Albany as discussed earlier.
PCA for East Athens. The Athens site located ~75 mi southeast of Columbus is a rural location close to Wayne national forest. Athens is a university town with a population of ~20,000. The only significant local source is the university’s coal-fired power plant. However, Athens is located ~20 mi from the Ohio River Valley, which has numerous coal-fired power generation facilities and industrial operations. Athens has been used as an upwind remote rural site for the Department of Energy’s Ohio River Valley PM^sub 2.5^ monitoring projects.39
PCA was performed for the East Athens outdoor site, and Table 3 shows seven significant factors affecting this site. Factor solution consisted of 52 samples and 26 variables. Seven factors (PCs) explain 75% of the total variance of the data. The first three factors represent >50% of the data variance.
The first factor (factor 1) contains P, K, Cr, Ni, As, Cd, and Sn. All of the elements except Sn have high factor loadings. This factor is typical of an industrial source signature. The second factor (factor 2) contains Si, S, Cl, and Ca. Si, S, and Ca are common constituents of fossil fuel combustion, most likely coal combustion. Cl is also found in coal- and wood-fired external combustion source profiles. As discussed earlier in PCA for the Koebel site, elements such as Si and Ca in this factor are also found to be abundantly available in average soil and crustal profiles. 31 The third factor (factor 3) has high loadings with Mg^sup +2^ and moderate loadings with Co, Na^sup +^, K^sup +^, Cl^sup -^, and NO^sub 3^^sup -^. This factor appears to be a mixture of various sources, mainly fugitive dust in the form of natural soil, automobile exhaust, and roadway salt. Fugitive dust or crustal material does not originate from point sources. The source regions for these materials are difficult to identify, and they can easily mix in with the other factors.32
The fourth factor (factor 4) contains NH^sub 4^^sup +^, Ti, Ca^sup +2^, and SO^sub 4^^sup -2^. A strong factor loading for NH^sub 4^^sup +^ in the rural area may indicate emissions from fertilized lands, animal waste from farms, or even agricultural burning. All of the chemical components in this factor are also observed to varying degrees in agricultural burning (fescue) and wood burning.31 The composition of biomass burning emissions is strongly dependent on the combustion stage (i.e., flaming, smoldering, or mixed) and the type of vegetation (e.g. forest, grassland, or scrub). Emissions from biomass burning are also strongly seasonal and can be highly episodic within the peak emissions season. The burning of fuel wood is confined mainly to the winter months. Forest fires occur primarily during the driest months of the year.
Factor 5 contains Cu, Mn, and V. Smelter emissions are reported to be enriched in Cu.31 Mn is also found in emissions from various industrial sources including primary lead smelting and metal production and other industrial operations.31 Anthropogenic V in aerosols is generally considered to be a result of combustion of heavy oil fuels.38 There is a lack of these sources near this rural site. The presence of these trace metals in this area suggests the potential transport of pollutants from industrial operations in the Ohio River Valley. The sixth factor (factor 6) contains Fe, which, as discussed earlier, is a crustal source marker.33 Fe is a major constituent of both soil and crustal profiles, and its concentration depends on local geology and climate conditions. Factor 7 has a high loading of Zn and SO^sub 4^^sup -2^ and moderate loadings of Ni. A number of trace elements are greatly enriched over crustal abundances in different fuels, such as Zn and Ni in oil. Another interesting fact about fuel combustion is that the oil-fired power plant profiles contain higher sulfate concentrations than most of the coal-fired power plant profiles. Thus, oil-fired power plants appear to be the primary source associated with this factor.
Summary of PCA
PCA was used for preliminary source apportionment of PM^sub 2.5^ observed at the three monitoring sites in Ohio. On average, the dataset consisted of 70 samples and 26 variables for each site. The PCA explained the majority of the variance of the data matrix and qualitatively interpreted possible sources affecting the three sites that represent urban, suburban, and rural areas of Central and Southeast Ohio. The results appear to represent mixed sources or a broad source categor\y rather than specific individual point sources.
The PCA for the Koebel site, which is in the industrial center of Columbus city, showed the influence of PM^sub 2.5^ sources associated with industrial, soil, crustal, and incineration sources, as well as the impact of road salt mixed with sand. The PCA for New Albany, a suburban site, indicated emissions from fossil fuel combustion sources, soil and crustal sources, and incineration sources, as well as waste disposal sources. The major source categories suggested by PCA at the East Athens site, which is a rural location near Wayne National Forest, included industrial, fossil fuel combustion, soil and crustal, and biomass burning sources. Some of the common sources indicated by PCA at all three of the sites included industrial sources (various chemical and manufacturing processes), fossil fuel combustion sources (coal- fired and oil-fired power plant emissions and external combustion), and geological sources (soil and crustal). The Koebel site PCA showed more urban source emissions when compared with the other two sites because of the impact of locally originating sources, as well as sources from the Ohio River Valley region. The PCA for the East site suggested a distinct source of biomass burning typical of the rural Ohio area. The PCA for the New Albany site indicates similar emission sources to the Koebel site except that this site was also heavily impacted by transport from other areas. The presence of industrial and combustion source emissions at the East and New Albany sites indicates transport of PM^sub 2.5^ precursors from areas such as the urban core of Columbus, greater Columbus industrial sources, and the Ohio River Valley region, respectively.
Ohio Environmental Protection Agency compiles inventory of the actual emissions from various sources. For this study, a rudimentary analysis of PM^sub 2.5^ source contributors was performed using the emissions inventory compiled for the base year of 2002. Both the rudimentary analysis of actual emission sources and the PCA analysis showed similar results. Industrial sources, including automotive manufacturing industries, coating and resins manufacturing industries, packing paper manufacturing industries, glass manufacturing industries, and chemical industries, were found to be the primary sources emitting ~11,000 t of PM. The second largest group was found to be coal-fired power plants, electric-generating power plants, and compressors emitting ~2300 t of PM. Land fills and construction work areas were found to be among the third largest group emitting ~1000 t of PM.40 The PCA analysis highlighted similar PM sources, impacting the monitoring sites in Central and Southeast Ohio, as those listed in the emissions inventory for Ohio.
CONCLUSIONS
The results from a study of PM^sub 2.5^ measured at three elementary schools in Central and Southeast Ohio during February 1999 through August 2000 is presented here. Daily averaged PM^sub 2.5^ concentrations did not exceed the 24-hr NAAQS standard of 65 g/ m^sup 3^ at the three monitoring sites. However, data from 1999 and 2000 suggest that Koebel, East Athens, and potentially New Albany could possibly exceed the annual NAAQS standard of 15 g/m^sup 3^. The average PM^sub 2.5^ concentrations showed similar temporal characteristics among the three sites indicating homogeneity in the spatial distribution of PM^sub 2.5^ within the study region. The water-soluble ions and trace element concentrations in the collected samples were analyzed using an ion chromatograph and X-ray fluorescence spectrophotometer, respectively. The components identified included: F^sup -^, Cl^sup -^, NO^sub 3^^sup -^, SO^sub 4^^sup -2^, PO^sub 4^^sup -3^, Li^sup +^, Na^sup +^, NH^sub 4^^sup +^, K^sup +^, Mg^sup +2^, Ca^sup +2^, Si, P, S, Cl, K, Ca, Ti, Co, Ni, V, Cr, Mn, Fe, Cu, Zn, As, and Cd. The single largest component of analyzed PM^sub 2.5^ mass was SO^sub 4^^sup -2^ at ~25% and ~19% of the outdoor and indoor PM^sub 2.5^ mass, respectively. Other abundant components included NO^sub 3^^sup -^, Cl^sup -^, NH^sub 4^^sup +^, and Na ions, as well as Ca, Si, and Fe.
PM^sub 2.5^ and its major component, sulfate ion, showed strong seasonal variations with maximum concentrations observed during the summer at all three of the sites. Sulfate concentrations increased from winter to summer at all three of the sites. During the summer (June through August), fall (September through November), and winter (December through February) months, the outdoor concentrations of PM^sub 2.5^ at all of the sites were consistently higher than the indoor concentrations at the corresponding sites. Elevated PM concentrations were observed in the indoor environment during the spring months (March through May) at New Albany (a suburb of Columbus) and East Athens (rural Ohio area).
The impact of long-range transport of PM^sub 2.5^ from upwind source regions on the three monitoring sites in Ohio was evaluated using the PSCF analysis. Source regions with the highest impact on these air monitoring locations in Central and Southeast Ohio included major urban areas in Ohio and surrounding states of Indiana, Kentucky, West Virginia, and Pennsylvania, a well as the industrialized areas of the Ohio River Valley region. PSCF analysis provides a reasonable estimate of the influence of upwind regions and is a useful tool for policy planners.
PCA was used for preliminary source apportionment of PM^sub 2.5^. On average, the dataset consisted of 70 samples and 26 variables for each site. The PCA explained the majority of the variance of the data matrix and qualitatively interpreted possible sources affecting the three sites that represent urban, suburban, and rural areas of Central and Southeast Ohio. The results appear to represent mixed sources or a broad source category rather than specific individual point sources.
The PCA for the Koebel site, which is in the industrial center of Columbus city, showed the influence of PM^sub 2.5^ sources associated with industrial, soil and crustal, and incineration sources, as well as the influence of road salt mixed with sand. The PCA for New Albany, a suburban site, indicated the influence of emissions from fossil fuel combustion sources, soil and crustal sources, and incineration sources, as well as waste disposal and incineration sources. The major source categories suggested by PCA at the East Athens site, which is a rural location near Wayne National Forest, included industrial, fossil fuel combustion, soil and crustal, and biomass burning sources. The presence of industrial and combustion source emissions at the East Athens and New Albany sites indicate transport of PM^sub 2.5^ precursors from areas such as the urban core of Columbus, greater Columbus industrial sources, and the Ohio River Valley region, respectively. Some of the common sources indicated by PCA at all three of the sites included industrial sources (various chemical and manufacturing processes), fossil fuel combustion sources (coalfired and/or oil-fired power plant emissions, as well as outdoor combustion), and geological sources (soil and crustal). The results from the PCA were validated by comparing against the 2002 emissions inventory data that identified major PM source categories in the state of Ohio.
ACKNOWLEDGMENTS
This research was an integral component of the Pediatric Health Impact Assessment Project sponsored by the Ohio Environmental Protection Agency and the Ohio Air Quality Development Authority. The authors thank the sponsoring organizations for their financial support.
IMPLICATIONS
The National Ambient Air Quality Standards for PM^sub 2.5^ was revised recently. A detailed review of the chemical composition of PM^sub 2.5^ is needed to identify the health effects of PM^sub 2.5^ and its constituents on susceptible population such as children. This study is the first of its kind conducted in elementary schools in Central and Southeast Ohio. The use of potential source contribution function analysis and principal component analysis techniques for the identification of major regional and local source contributors provides the policy planner with an important assessment tool.
REFERENCES
1. D’Ippoliti, D.; Forastiere, F.; Ancona, C.; Agabiti, N.; Fusco, D.; Michelozzi, P.; Perucci, C.A. Air Pollution and Myocardial Infarction in Rome: a Case-Crossover Analysis; Epidemiol. 2003, 14, 528-535.
2. Englert, N. Fine Particles and Human Health-a Review of Epidemiological Studies; Toxicol. Lett. 2004, 149, 235-242.
3. Dockery, D.W.; Pope, C.A., III; Xu, X.; Spengler, J.D.; Ware, J.H.; Martha, E.F.; Ferris, B.G., Jr.; Speizer, F.E. An Association between Air Pollution and Mortality in Six U.S. Cities; N. Engl. J. Med. 1993, 329, 1753-1759.
4. Pope, C.A., III; Dockery, D.W.; Schwartz, J. Review of Epidemiological Evidence of Health Effects of Particulate Air Pollution; Inhal. Toxicol. 1995, 7, 1-18.
5. Dockery, D.W.; Cunningham, J.; Damokosh, A.I.; Neas, L.M.; Spengler, J.D.; Koutrakis, P.; Ware, J.H.; Raizenne, M.; Speizer, F.E. Health Effects of Acid Aerosols on North American Children: Respiratory Symptoms. Environ. Health Perspect. 1996, 104, 500-518.
6. Krewski, D.; Burnett, R.T.; Goldberg, M.S.; Hoover, K.; Siemiatycki, J.; Abrahamowicz, M.; White, W.H. Validation of the Harvard Six Cities Study of Particulate Air Pollution and Mortality; N. Engl. J. Med. 2004, 350, 198-199.
7. Burnett, R.T.; Dales, R.E.; Brook, J.R.; Raizenne, M.E.; Krewski, D. Association between Ambient Carbon Monoxide Levels and Hospitalizations for Congestive Heart Failure in the Elderly in 10 Canadian Cities; Epidemiol. 1997, 8, 162-167.
8. Peters, A.; Liu, E.; Verrier, R.L.; Schwartz, J.; Gold, D.R.; Mittleman, M.; Baliff, J.; Oh, J.A.; Allen, G.; Monahan, K.; Dockery, D.W. Air Pollution and Incidence of Cardiac Arrhythmia; Epidemiol. 2000, 11, 11-17.
9. Peters, A.; Dockery, D.W.; Muller, J.E.; Mittleman, M.A. Incr\eased Particulate Air Pollution and the Triggering of Myocardial Infarction; Circulation 2001, 103, 2810-2815.
10. Annette, P. Particulate Matter and Heart Disease: Evidence from Epidemiological Studies; Toxicol. Appl. Pharmacol. 2005, 207, 477-482.
11. Li N.; Hao M.; Phalen R.F.; Hinds W.C.; Nel A.E. Particulate Air Pollutants and Asthma. A Paradigm for the Role of Oxidative Stress in PM-Induced Adverse Health Effects; Clin. Immunol. 2003, 109, 250- 265.
12. Spurny, K. R.; Hochrainer, D., Eds.; Aerosol Chemical Processes in the Environment; Lewis Publishers: New York, NY, 2000; p 406.
13. U.S. Environmental Protection Agency. Code of Federal Regulations, Part 50, Title 62, 1997.
14. Crist, K.; John, K.; Carlson, B. Characterization of Fine Particulate Matter in Ohio. Presented at the 219th American Chemical Society National Meeting, San Francisco, CA, March 2000.
15. Gilli, G.; Traversi, D.; Rovere, R.; Pignata, C.; Schiliro, T. Airborne Particulate Matter: Ionic Species Role in Different Italian Sites; Environ. Res. 2006, 103, 1-8.
16. Lingard, J.J.N., Tomlin, A.S.; Clarke, A.G.; Healey, K.; Hay, A.W.M., Wild, C.P.; Routledge, M.N. A Study of Trace Metal Concentration of Urban Airborne Particulate Matter and Its Role in Free Radical Activity as Measured by Plasmid Strand Break Assay; Atmos. Environ. 2005, 39, 2377-2384.
17. Anderson, R.R.; Martello, V.D.; White, M.C.; Crist, K.; John, K.; Modey, K.W.; Eatough, J.D. The Regional Nature of PM^sub 2.5^ Episodes in the Upper Ohio River Valley; J. Air & Waste Manage. Assoc. 2004, 54, 971-984.
18. Connell, P.D.; Withum, A.J.; Winter, E.S.; Statnick, M.R.; Bilonick, A.R. The Steubenville Comprehensive Air Monitoring Program (SCAMP): Analysis of Short-Term and Episodic Variations in PM^sub 2.5^ Concentrations Using Hourly Air Monitoring Data; J. Air & Waste Manage. Assoc. 2005, 55, 559-573.
19. Ashbaugh, L.L.; Malm, W.C.; Sadeh, W.D. A Residence Time Probability Analysis of Sulfur Concentrations at Grand Canyon National Park; Atmos. Environ. 1985, 19, 1263-1270.
20. Yli-Tuomi, T.; Hopke, P.K.; Paatero, P.; Basunia, M.S.; Landsberger, S.; Viisanen, Y.; Paatero, J. Atmospheric Aerosol over Finnish Arctic: Source Analysis by the Multilinear Engine and the Potential Source Contribution Function; Atmos. Environ. 2003, 37, 4381-4392.
21. Begum, B.A.; Kim, E.; Jeong, C.-H.; Lee, D.-W.; Hopke, P.K. Evaluation of the Potential Source Contribution Function Using the 2002 Quebec Forest Fire Episode; Atmos. Environ. 2005, 39, 3719- 3724.
22. Hoh, E.; Hites, R.A. Sources of Toxaphene and Other Organochlorine Pesticides in North America as Determined by Air Measurements and Potential Source Contribution Function Analyses; Environ. Sci. Technol. 2004, 38, 4187-4194.
23. Draxler, R.R.; Rolph, G.D. HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory) Model. National Oceanic and Atmospheric Administration Air Resources Laboratory: Silver Spring, MD, 2003; available at http://www.arl.noaa.gov/ready/hysplit4.html (accessed 2006).
24. National Oceanic and Atmospheric Administration. NOAA ARL- Gridded Meteorological Data Archives; available at http:// www.arl.noaa.gov/ss/ transport/archives.html (accessed 2006).
25. Hopke, P.K.; Gladney, E.S.; Gordon, G.E.; Zoller, W.H.; Jones, A.G. The Use of Multivariate Analysis to Identify Sources of Selected Elements in the Boston Aerosol; Atmos. Environ. 1976, 10, 1015-1025.
26. Henry, R.C. History and Fundamentals of Multivariate Air Quality Receptor Models; Chemometrics Intell. Lab. Systems 1997, 37, 37-42.
27. Vallius, M.; Lanki, T.; Tiittanen, P.; Koistinen, K.; Ruuskanen, J.; Pekkanen, J. Source Apportionment of Urban Ambient PM^sub 2.5^ in Two Successive Measurement Campaigns in Helsinki, Finland; Atmos. Environ. 2003, 37, 615-623.
28. Almeida, S.M.; Pio, C.A.; Frietas, M.C.; Reis, M.A.; Transoco, M.A. Source Apportionment of Fine and Coarse Particulate Matter in a Sub-Urban Area at the Western European Coast; Atmos. Environ. 2005, 39, 3127-3138.
29. Vallius, M.; Janssen, N.A.; Heinrich, J.; Hoek, G.; Ruuskanen, J.; Cyrys, J.; van Grieken, R.; de Hartog, J.J.; Kreyling, W.G.; Pekkanen, J. Sources and Elemental Composition of Ambient PM(2.5) in Three European Cities; Sci. Total Environ. 2005, 337, 147-162.
30. Park, S.S.; Kim, Y.J. Source Contributions to Fine Particulate Matter in an Urban Atmosphere; Chemosphere 2005, 59, 217- 226.
31. SPECIATE Software Version 3.1, User Guide and Program; U.S. Environmental Protection Agency: Durham, NC, 1999.
32. Ames, A.R.; Gullu, G.; Beal, J.; Olmez, I. Receptor Modeling for Elemental Source Contributions to Fine Aerosols in New York State; J. Air & Waste Manage. Assoc. 2000, 50, 881-887.
33. Biegalski, S.R.; Landsberger, S.; Hoff, R.M. Source-Receptor Modeling Using Trace Metals in Aerosols Collected at Three Rural Canadian Great Lakes Sampling Stations; J. Air & Waste Manage. Assoc. 1998, 48, 227-237.
34. Park, S.S.; Bae, S.M.; Kim, Y.J. Chemical Composition and Source Apportionment of PM^sub 2.5^ Particles in the Sihwa Area, Korea; J. Air & Waste Manage. Assoc. 2001, 51, 393-405.
35. Ramadan, Z.; Song, X.; Hopke, P.K. Identification of Sources of Phoenix Aerosol by Positive Matrix Factorization; J. Air & Waste Manage. Assoc. 2000, 50, 1308-1320.
36. Kim, B.M.; Teffera, S.; Zeldin, M.D. Characterization of PM^sub 2.5^ and PM^sub 10^ in the South Coast Air Basin of Southern California: Part 1-Spatial Variations; J. Air & Waste Manage. Assoc. 2000, 50, 2034-2044.
37. Sisler, J.F.; Malm, W.C. Interpretation of Trends of PM^sub 2.5^ and Reconstructed Visibility from the IMPROVE Network; J. Air & Waste Manage. Assoc. 2000, 50, 775-789.
38. Khosah, R.P.; Shimshock, J.P. Semi-Annual Technical Progress Report- Comparative Evaluation of Ambient Fine Particulate Matter (PM^sub 2.5^) Data Obtained from Urban and Rural Monitoring Sites along the Upper Ohio River Valley; National Energy Technology Laboratory of U. S. Department of Energy: Pittsburgh, PA, 2003.
39. Liu, B., Crist, K. Characterization of Fine Particulate Matter in Ohio; Presented at the 94th A&WMA Annual Conference, Orlando, FL, June 2001.
40. Ohio Environmental Protection Agency; available at http:// www.epa. state.oh.us/dapc/aqmp/eiu/eiu.html (accessed 2006).
Kuruvilla John and Saritha Karnae
Department of Environmental Engineering, Texas A&M University- Kingsville, Kingsville, TX
Kevin Crist and Myoungwoo Kim
Air Quality Center, Ohio University, Athens, OH
Amol Kulkarni
KBR (Halliburton), Houston, TX
About the Authors
Kuruvilla John is an associate dean of the Frank H. Dotterweich College of Engineering and associate professor of Environmental Engineering at Texas A&M University-Kingsville. Kevin Crist is an associate professor of chemical engineering and director of the Air Quality Center, Institute for Sustainable Energy and the Environment at Ohio University. Myoungwoo Kim is a research scientist in the Institute for Sustainable Energy and the Environment at Ohio University. Saritha Karnae is a research associate in the Department of Environmental Engineering at Texas A&M University-Kingsville. Amol Kulkarni is currently a process engineer with KBR (Halliburton). Address correspondence to Kuruvilla John, Associate Dean, Frank H. Dotterweich College of Engineering, Texas A&M University-Kingsville, MSC 188, 700 University Blvd., Kingsville, TX 78363; phone: +1-361-593-2001; fax: +1-361-593-2106; e-mail: k- john@tamuk.edu.
Copyright Air and Waste Management Association Apr 2007
(c) 2007 Journal of the Air & Waste Management Association. Provided by ProQuest Information and Learning. All rights Reserved.