Optimized variable source-profile approach for source apportionment

An expanded chemical mass balance (CMB) approach for PM_2.5 source apportionment is presented in which both the local source compositions and corresponding contributions are determined from ambient measurements and initial estimates of source compositions using a global-optimization mechanism. Such an approach can serve as an alternative to using predetermined (measured) source profiles, as traditionally used in CMB applications, which are not always representative of the region and/or time period of interest. Constraints based on ranges of typical source profiles are used to ensure that the compositions identified are representative of sources and are less ambiguous than the factors/sources identified by typical factor analysis (FA) techniques. Gas-phase data (SO₂, CO and NO_y) are also used, as these data can assist in identifying sources. Impacts of identified sources are then quantified by minimizing the weighted-error between apportioned and measured levels of the fitting species. This technique was applied to a dataset of PM_2.5 measurements at the former Atlanta Supersite (Jefferson Street site), to apportion PM_2.5 mass into nine source categories. Good agreement is found when these source impacts are compared with those derived based on measured source profiles as well as those derived using a current FA technique, Positive Matrix Factorization. The proposed method can be used to assess the representativeness of measured source-profiles and to help identify those profiles that may be in significant error, as well as to quantify uncertainties in source-impact estimates, due in part to uncertainties in source compositions.     

Source apportionment with site specific source profiles

A receptor modeling study was performed to identify and apportion the sources of PM10 mass in Granite City, Illinois, an area of historic TSP nonattainment. Samples of the ambient aerosol were collected using a dichotomous sampler. Each sample was analyzed by x-ray fluorescence and instrumental neutron activation analysis. To begin the study, a factor analysis was performed. Two different chemical mass balance (CMB) analyses were then made. The first CMB analysis used only source profiles available from the literature while the second included twelve source profiles developed from dust samples collected in Granite City. Both CMB analyses used 20 of the 33 analyzed elements since many of the source profiles in the literature did not include the other thirteen elements. The results from both sets of CMB analyses were grouped by the predominate wind direction at the site during the time each sample was taken to identify the direction of each source relative to the sampler. It was found that regional sources were the primary contributors to the fine fraction while the coarse fraction was composed of material from local industries. These sources were generally the ones identified during the Regional Air Pollution Study previously conducted in the area. However, the emission profiles from these sources were observed to have changed between the studies. It was also found that the use of the locally generated profiles greatly improved the results of the CMB analysis.     

Review of receptor modeling methods for source apportionment

Efforts have been made to relate measured concentrations of airborne constituents to their origins for more than 50 years. During this time interval, there have been developments in the measurement technology to gather highly time-resolved, detailed chemical compositional data. Similarly, the improvements in computers have permitted a parallel development of data analysis tools that permit the extraction of information from these data. There is now a substantial capability to provide useful insights into the sources of pollutants and their atmospheric processing that can help inform air quality management options. Efforts have been made to combine receptor and chemical transport models to provide improved apportionments. Tools are available to utilize limited numbers of known profiles with the ambient data to obtain more accurate apportionments for targeted sources. In addition, tools are in place to allow more advanced models to be fitted to the data based on conceptual models of the nature of the sources and the sampling/analytical approach. Each of the approaches has its strengths and weaknesses. However, the field as a whole suffers from a lack of measurements of source emission compositions. There has not been an active effort to develop source profiles for stationary sources for a long time, and with many significant sources built in developing countries, the lack of local profiles is a serious problem in effective source apportionment. The field is now relatively mature in terms of its methods and its ability to adapt to new measurement technologies, so that we can be assured of a high likelihood of extracting the maximal information from the collected data.

Implications: Efforts have been made over the past 50 years to use air quality data to estimate the influence of air pollution sources. These methods are now relatively mature and many are readily accessible through publically available software. This review examines the development of receptor models and the current state of the art in extracting source identification and apportionments from ambient air quality data.     

Polycyclic aromatic hydrocarbons in the ambient air of Greek towns in relation to other atmospheric pollutants

Polycyclic aromatic hydrocarbons (PAHs) were determined in the ambient air of six towns in N. Greece. This paper presents the variability of the particle-bound PAHs concentrations and the particle PAH content during the cold and the warm months. Correlations of total PAHs with other atmospheric pollutants were largely different among towns indicating that the relative contribution of emission sources is different in each location. In the warm months PAHs were significantly correlated with vehicular pollutants thus suggesting traffic as the major PAH emmitting source. The same was also deduced from the comparison of the ambient PAH profiles to the profiles of particular sources. The contribution of residential heating was significant in most towns during winter. Principal component analysis of the data did not result in a clear distinction between towns thus suggesting that all are influenced by similar source types. Finally, the risk associated with the inhallation of carcinogenic PAHs in each town was estimated and compared to the risk from more urbanized/industrialized sites in N. Greece.     

Profile analysis of ambient and source emitted particle-bound polycyclic aromatic hydrocarbons from three sites in northern Greece

Polycyclic aromatic hydrocarbons (PAHs) adsorbed to ambient PM(10) were determined at three sites in Thessaloniki, northern Greece, during the period June 1997-July 1998. Ambient PAH profiles exhibited significant seasonal and spatial variations. Source PAH profiles were obtained for a number of urban, industrial and geological sources including cement, fertilizer and asphalt production, quarry operations, metal electroplating, metal welding and tempering, steel manufacture, lead and bronze smelters, metal scrap incineration, oil burning, non-catalyst equipped passenger cars, diesel fueled taxies and buses, paved road dust and soil dust. Principal component analysis (PCA) and diagnostic ratios were employed to compare ambient and source PAH profiles in an attempt to recognize compositional patterns. Similarities between the ambient PAH profiles and the profiles of certain sources, such as vehicular emissions, oil burning and metal industries, were identified.     

Source apportionment of ambient particles in steubenville,oh using specific rotation factor analysis

A statistical approach, Specific Rotation Factor Analysis (SRFA), has been developed to identify and apportion sources of ambient air pollutants. To increase the statistical weight of the source tracer elements, this technique is based on the analysis of the covariance matrix. The obtained eigenvectors are obliquely rotated in order to maximize the loadings of the tracer quantities of the corresponding source (specific rotations). Source contributions and profiles are estimated by regressing elemental and mass concentrations on the factor loadings.The ability of the SRFA technique to resolve major aerosol sources and determine their profiles and contributions at the receptor site is demonstrated by applying it to a subset of fine particle elemental and mass concentration data from Steubenville, OH.     

Sensitivity of source apportionment of urban particulate matter to uncertainty in motor vehicle emissions profiles

A sensitivity analysis was conducted to characterize sources of uncertainty in results of a molecular marker source apportionment model of ambient particulate matter using mobile source emissions profiles obtained as part of the Gasoline/Diesel PM Split Study. A chemical mass balance (CMB) model was used to determine source contributions to samples of fine particulate matter (PM2.5) collected over 3 weeks at two sites in the Los Angeles area in July 2001. The ambient samples were composited for organic compound analysis by the day of the week to investigate weekly trends in source contributions. The sensitivity analysis specifically examined the impact of the uncertainty in mobile source emissions profiles on the CMB model results. The key parameter impacting model sensitivity was the source profile for gasoline smoker vehicles. High-emitting gasoline smoker vehicles with visible plumes were seen to be a significant source of PM in the area, but use of different measured profiles for smoker vehicles in the model gave very different results for apportionment of gasoline, diesel, and smoker vehicle tailpipe emissions. In addition, the contributions of gasoline and diesel emissions to total ambient PM varied as a function of the site and the day of the week.     

Sources of fine particle composition in the northeastern US

Fine particle composition data obtained at three sampling sites in the northeastern US were studied using a relatively new type of factor analysis, positive matrix factorization (PMF). The three sites are Washington, DC, Brigantine, NJ and Underhill, VT. The PMF method uses the estimates of the error in the data to provide optimal point-by-point weighting and permits efficient treatment of missing and below detection limit values. It also imposes the non-negativity constraint on the factors. Eight, nine and 11 sources were resolved from the Washington, Brigantine and Underhill data, respectively. The factors were normalized by using aerosol fine mass concentration data through multiple linear regression so that the quantitative source contributions for each resolved factor were obtained. Among the sources resolved at the three sites, six are common. These six sources exhibit not only similar chemical compositions, but also similar seasonal variations at all three sites. They are secondary sulfate with a high concentration of S and strong seasonal variation trend peaking in summer time; coal combustion with the presence of S and Se and its seasonal variation peaking in winter time; oil combustion characterized by Ni and V; soil represented by Al, Ca, Fe, K, Si and Ti; incinerator with the presence of Pb and Zn; sea salt with the high concentrations of Na and S. Among the other sources, nitrate (dominated by NO₃⁻) and motor vehicle (with high concentrations of organic carbon (OC) and elemental carbon (EC), and with the presence of some soil dust components) were obtained for the Washington data, while the three additional sources for the Brigantine data were nitrate, motor vehicle and wood smoke (OC, EC, K). At the Underhill site, five other sources were resolved. They are wood smoke, Canadian Mn, Canadian Cu smelter, Canadian Ni smelter, and another salt source with high concentrations of Cl and Na. A nitrate source similar to that found at the other sites could not be obtained at Underhill since NO₃⁻ was not measured at this site. Generally, most of the sources at the three sites showed similar chemical composition profiles and seasonal variation patterns. The study indicated that PMF was a powerful factor analysis method to extract sources from the ambient aerosol concentration data.     

Sensitivity of Source Apportionment of Urban Particulate Matter to Uncertainty in Motor Vehicle Emissions Profiles

Abstract

A sensitivity analysis was conducted to characterize sources of uncertainty in results of a molecular marker source apportionment model of ambient particulate matter using mobile source emissions profiles obtained as part of the Gasoline/Diesel PM Split Study. A chemical mass balance (CMB) model was used to determine source contributions to samples of fine particulate matter (PM_2.5) collected over 3 weeks at two sites in the Los Angeles area in July 2001. The ambient samples were composited for organic compound analysis by the day of the week to investigate weekly trends in source contributions. The sensitivity analysis specifically examined the impact of the uncertainty in mobile source emissions profiles on the CMB model results. The key parameter impacting model sensitivity was the source profile for gasoline smoker vehicles. High-emitting gasoline smoker vehicles with visible plumes were seen to be a significant source of PM in the area, but use of different measured profiles for smoker vehicles in the model gave very different results for apportionment of gasoline, diesel, and smoker vehicle tailpipe emissions. In addition, the contributions of gasoline and diesel emissions to total ambient PM varied as a function of the site and the day of the week.     

Source characterization of ozone precursors by complementary approaches of vehicular indicator and principal component analysis

Measurements of speciated non-methane hydrocarbons (NMHCs) were conducted in an ozone non-attainment metropolis with pronounced industrial emissions in addition to traffic ones. Highly variable and complex natures of industrial sources make their composition profiles difficult to determine. In the circumstances of no reliable source profiles, two simple complementary approaches were attempted to characterize sources of NMHCs. First, a robust vehicular indicator, 3-methylpentane (3MC5A), which is an intrinsic component of gasoline, was used to estimate contributions of traffic versus non-traffic sources for major NMHCs with high ozone-forming potentials (OFPs), such as ethene, toluene, xylene, isoprene, etc. Second, the method of principal component analysis (PCA) was employed to further discern non-traffic emissions into various source groups. A total of 454 ambient samples were sampled in the urban-industrial complex metropolis (Kaohsiung, Taiwan) to build up a large dataset to be tested by the two complementary approaches. It was found that four types of emissions, i.e., traffic, household fuel leakage, industrial, and biogenic, were responsible for the observed ambient NMHCs. The industrial contribution was significant for ethene and toluene (with 48–67% and 33–62%, respectively), whereas xylene was found to be mainly vehicular. In addition, isoprene revealed its biogenic nature. OFPs arising from vehicular, industrial and biogenic contributions could be further assessed for the purpose of emission control of NMHCs in the ozone non-attainment area.     

Source apportionment of exposures to volatile organic compounds. I. Evaluation of receptor models using simulated exposure data

Four receptor-oriented source apportionment models were evaluated by applying them to simulated personal exposure data for select volatile organic compounds (VOCs) that were generated by Monte Carlo sampling from known source contributions and profiles. The exposure sources modeled are environmental tobacco smoke, paint emissions, cleaning and/or pesticide products, gasoline vapors, automobile exhaust, and wastewater treatment plant emissions. The receptor models analyzed are chemical mass balance, principal component analysis/absolute principal component scores, positive matrix factorization (PMF), and graphical ratio analysis for composition estimates/source apportionment by factors with explicit restriction, incorporated in the UNMIX model. All models identified only the major contributors to total exposure concentrations. PMF extracted factor profiles that most closely represented the major sources used to generate the simulated data. None of the models were able to distinguish between sources with similar chemical profiles. Sources that contributed <5% to the average total VOC exposure were not identified.     

Interlaboratory comparison of Receptor Model results for Houston aerosol

In exercises for the Mathematical and Empirical Receptor Models Workshop (Quail Roost II), nine investigators independently applied receptor models to apportion ambient aerosol mass concentrations to components from emission sources in Houston, TX. The receptor model results were based upon ambient aerosol compositional data supplied to investigators that included mass, elemental and ionic species concentrations for consecutive 12-h sampling periods at the University of Houston between 10 and 19 September 1980. Some investigators performed additional analyses by X-ray powder diffraction, scanning electron microscopy with automated image analysis and X-ray energy spectroscopy and light microscopy. In most cases, the fine (0.2–2.5 μm) and coarse (2.5–15 μm) fractions were analyzed separately by the receptor models. The models tested include effective variance type chemical mass balance, weighted ridge regression, multiple linear regression, factor analysis and target transformation factor analysis. The number of source classes determined ranged from 4 (by multiple linear regression) to 23 (by light microscopy). Results are presented and compared in eight general emissions categories.     

Source contribution of atmospheric aerosols in Bangkok metropolitan area by CMB8 receptor model

The CMB₈ model was applied for source apportionment of particulate matter in Bangkok area. The 24 h of ambient data were collected and analysed for elemental composition during December 1996 to January 1997 by high volume air samplers at a station in Bangkok, Thailand. Seven source profiles and the average mass concentrations of 42 ambient data were used to run the chemical mass balance (CMB₈) model. The source apportionment by CMB₈ gave similar results comparing with factor analysis – multiple regression (FA–MR) model of the same data. The results revealed that major sources of particulate matter in Bangkok were: soil (33%), road dust (33%) and automobile (15%). The minor source contributions were: sea salt (4.34%), refuse incineration and biomass burning (0.84%), steel mill (0.62%) and fuel oil combustion (0.35%). The lack of source profile for biomass or open burning in Bangkok resulted in much lower predicted contribution of this source when compared to that from FA–MR. When apply this CMB₈ model with daily ambient data, the result revealed that one fourth of daily CMB₈ source apportionment had high value of ² (>4). These exceedance values of Chi² also point out that one of the selected sources (biomass burning) may not be the true contributing sources. Presumably, accurate biomass burning source profile is needed to improve the CMB calculation of source contributions for particulate matter in Bangkok metropolitan area.     

Locations and preferred pathways of possible sources of Arctic aerosol

The chemical composition of particles collected at Alert, Northwest Territories, Canada, show strong, persistent seasonal variations. In a previous study, a 2-way/3-way mixed factor model was performed on the weekly average concentrations of 24 aerosol components measured over the period from 1980 to 1991. The Multilinear Engine (ME), a new mathematical technique, was used to obtain the solution. The two modes of the 2-way model consist of the source composition profiles and mass contributions over the 11 yr, while for the three modes of the 3-way model, source profiles, mass contributions variations over the weeks within a year, and the year-to-year variation over the 11 yr within the measurement period. Five 2-way and two 3-way factors were found to provide a good fit to the data and were easily interpreted. In this investigation, potential source contribution function (PSCF) analysis was applied to the source contributions derived from the ME analysis by incorporating meteorological information in the form of 5-d air parcel back trajectories. The potential locations and/or the preferred pathways of these possible sources were then determined by the PSCF analysis.     

Combined source apportionment,using positive matrix factorization–chemical mass balance and principal component analysis/multiple linear regression–chemical mass balance models

The methods of positive matrix factorization–chemical mass balance and principal component analysis/multiple linear regression–chemical mass balance were studied in this paper, for combined source apportionment. Due to the high similarity among the source profiles, several problems would raised when only one receptor model was applied. For example, the collinearity problem would result in the negative contributions when applying CMB model; certain sources would not to be separated out when applying PCA or PMF model. In this study, PCA/MLR–CMB model and PMF–CMB were attempted to resolve the problem, where the combined models were applied to study the synthetic and ambient datasets. In synthetic dataset, there were seven sources (six actual sources from real world, and one unknown source). The results obtained by the combined models show that the combined source apportionment technique is feasible. In addition, an ambient dataset from a northern city in China was analyzed by PCA/MLR–CMB model and PMF–CMB model, and these two models got the similar results. The results show that coal combustion contributed the largest fraction to the total mass.     

Source identifications of airborne fine particles using positive matrix factorization and U.S. Environmental Protection Agency positive matrix factorization

The widely used source apportionment model, positive matrix factorization (PMF2), has been applied to various air pollution data. Recently, U.S. Environmental Protection Agency (EPA) developed EPA positive matrix factorization (PMF), a version of PMF that will be freely distributed by EPA. The objectives of this study were to conduct source apportionment studies for particulate matter less than 2.5 microm in aerodynamic diameter (PM(2.5)) speciation data using PMF2 and EPA PMF (version 1.1) and to compare identified sources between the two models. In the present study, ambient PM(2.5) compositional datasets of 24-hr integrated samples collected at EPA Speciation Trends Network monitoring sites in Chicago, IL, and Portland, OR, were analyzed. Both PMF2 and EPA PMF extracted eight sources for the Chicago data and 10 sources for the Portland data. The model-resolved source profiles were similar between two models for both datasets. However, in several sources, the average contributions did not agree well and the time series contributions were not highly correlated. The differences between PMF2 and EPA PMF solutions were caused by the different least-square algorithm and the different nonnegativity constraints. Most of the average source contributions resolved by both models were within 5-95% uncertainty provided by EPA PMF, indicating that the sources resolved by both models were reproducible.     

Receptor modeling for elemental source contributions to fine aerosols in New York State

Fine atmospheric particulate material was collected at five sites in upstate New York and analyzed for its trace element composition by instrumental neutron activation analysis. Of the 3700 daily samples collected over a 2-yr period, 1459 were analyzed for 39 elements, providing a large and detailed data set. Factor analysis (FA) was used to identify potential pollution sources or source regions and to construct inorganic source profiles for each. Following FA, the method of absolute factor scores-multiple linear regression was used to estimate the absolute elemental contribution of each of the identified sources. Factor analysis identified nine sources impacting the sampling region. Seven of these were found to be present in varying degrees among of the sampling sites. The other two sources had more localized impacts and were observed at only one of the sites each. Regional sources (such as the midwestern United States and eastern Canada) and crustal/soil material accounted for the greatest amount of the trace elements measured in the collected material.     

Temporal variations of source impacts at the receptor,as derived from air particulate monitoring data in Ho Chi Minh City,Vietnam

Varimax rotation factor analysis was applied to monthly concentrations of elements in total suspended air particulate (TSP) matter in Ho Chi Minh City collected from December 1992 to November 1996, covering four dry/rainy seasons. Six pollution source types were revealed. Resuspended soil/road dust accounts for 74% of the TSP mass loading. Motor vehicles and a source which emits particulates containing arsenic account for 10% and 9%, respectively. There are three minor sources, namely, cement dust from the nearby construction site, road dust of local traffic origin and burning emissions. The contributions from these source were estimated with high uncertainties. The interpretation of sources was corroborated by studying source profiles and temporal variations of source contributions. The monthly variations of source contributions at the receptor were modelled by using source apportionment techniques. From the variation patterns, emission scenarios for burning, construction and motor vehicle sources were reproduced. Source contributions also exhibit seasonal variability induced by changes of meteorological conditions. No seasonal change was found for the As-containing particulates, suggesting a speculation on their origin as coal fly ash emitting from any local coal burning source.     

Levels of PCDD/FS in ambient air and soil in the vicinity of a municipal solid waste incinerator in Hsinchu

Levels of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) were determined in twenty-one ambient air samples, eight soil samples and two stack gas samples, collected near or in a municipal solid waste incinerator (MSWI) in Hsinchu, Taiwan. A systematic decrease of PCDD/Fs in the ambient air from the northeastern area was observed. PCDD/Fs levels measured in the ambient air range from 0.058 to 0.127 pg-TEQ/m3. Higher PCDD/Fs levels in the ambient air were found during winter. In addition, PCDD/Fs levels measured in the soil range from 0.524 to 5.02 pg-TEQ/g d.m. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) did not provide sufficient evidence that the environmental PCDD/Fs contamination was caused by emissions from the Hsinchu MSWI. An unknown PCDD/Fs source was proposed using congener profile analysis and supported by both PCA and HCA.     

The analysis of pollutant profiles by means of compositional analysis. An application to PCDD/F data

Compositional analysis consists of a group of techniques used to manipulate closed (=compositional) data, i.e. multivariate data summing to a fixed quantity (proportions, percentages). It is based on the analysis of the relations among variables and the use of logarithmic transformations. It has been claimed that this group of techniques should be used to analyse profiles of pollutant sources because profiles themselves are proportions. We show in this paper that, for the exploratory analysis of these data, a good strategy is to combine the analysis done with and without transformation because they give different and complementary insights on the structure of the data. We discuss in particular the study of processes such as the mixing of pollutants produced by different sources and the exponential decay of concentrations with distance to the source found in many studies. The clr transformation is also appropriate for the study of the variables having small proportions, which remain concealed by the abundant variables when analysed without transformation. We present simulations to illustrate these ideas and we also apply these techniques to two data sets of PCDD/F content in moss tissues.