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 Apportionment of Ambient Total Suspended Particulates and Coarse Particulate Matter in Urban Areas of Jiaozuo,China

Abstract

Approximately 750 total suspended particulates (TSPs) and coarse particulate matter (PM₁₀) filter samples from six urban sites and a background site and >210 source samples were collected in Jiaozuo City during January 2002 to April 2003. They were analyzed for mass and abundances of 25 chemical components. Seven contributive sources were identified, and their contributions to ambient TSP/PM₁₀ levels at the seven sites in three seasons (spring, summer, and winter days) and a “whole” year were estimated by a chemical mass balance (CMB) receptor model. The spatial TSP average was high in spring and winter days at a level of approximately 530 ～g/m³ and low in summer days at 456 ～g/m³; however, the spatial PM₁₀ average exhibited little variation at a level of approximately 325 ～g/m³, and PM_10-to-TSP ratios ranged from 0.58 to 0.81, which suggested heavy particulate matter pollution existing in the urban areas. Apportionment results indicated that geological material was the largest contributor to ambient TSP/PM₁₀ concentrations, followed by dust emissions from construction activities, coal combustion, secondary aerosols, vehicle movement, and other industrial sources. In addition, paved road dust and re-entrained dust were also apportioned to the seven source types and found soil, coal combustion, and construction dust to be the major contributors.     

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.     

The sensitivity of modeled ozone to the temporal distribution of point, area, and mobile source emissions in the eastern United States

Ozone remains one of the most recalcitrant air pollution problems in the US. Hourly emissions fields used in air quality models (AQMs) generally show less temporal variability than corresponding measurements from continuous emissions monitors (CEM) and field campaigns would imply. If emissions control scenarios to reduce emissions at peak ozone forming hours are to be assessed with AQMs, the effect of emissions' daily variability on modeled ozone must be understood. We analyzed the effects of altering all anthropogenic emissions' temporal distributions by source group on 2002 summer-long simulations of ozone using the Community Multiscale Air Quality Model (CMAQ) v4.5 and the Carbon Bond IV (CBIV) chemical mechanism with 12 km resolution. We find that when mobile source emissions were made constant over the course of a day, 8-h maximum ozone predictions changed by ±7 parts per billion by volume (ppbv) in many urban areas on days when ozone concentrations greater than 80 ppbv were simulated in the base case. Increasing the temporal variation of point sources resulted in ozone changes of +6 and −6 ppbv, but only for small areas near sources. Changing the daily cycle of mobile source emissions produces substantial changes in simulated ozone, especially in urban areas at night; results suggest that shifting the emissions of NO_x from day to night, for example in electric powered vehicles recharged at night, could have beneficial impacts on air quality.     

Lagrangian sampling of 3-D air quality model results for regional transport contributions to sulfate aerosol concentrations at Baltimore,MD, in summer 2004

We use ensemble-mean Lagrangian sampling of a 3-D Eulerian air quality model, CMAQ, together with ground-based ambient monitors data from several air monitoring networks and satellite (MODIS) observations to provide source apportionment and regional transport vs. local contributions to sulfate aerosol and PM_2.5 concentrations at Baltimore, MD, for summer 2004. The Lagrangian method provides estimates of the chemical and physical evolution of air arriving in the daytime boundary layer at Baltimore. Study results indicate a dominant role for regional transport contributions on those days when sulfate air pollution is highest in Baltimore, with a principal transport pathway from the Ohio River Valley (ORV) through southern Pennsylvania and Maryland, consistent with earlier studies. Thus, reductions in sulfur emissions from the ORV under the EPA's Clean Air Interstate Rule may be expected to improve particulate air quality in Baltimore during summer. The Lagrangian sampling of CMAQ offers an inexpensive and complimentary approach to traditional methods of source apportionment based on multivariate observational data analysis, and air quality model emissions separation. This study serves as a prototype for the method applied to Baltimore. EPA is establishing a system to allow air quality planners to readily produce and access equivalent results for locations of their choice.     

Non-linear response of wet deposition to emissions reduction: A model study

An Eulerian atmospheric model with complex chemistry (Acidic Deposition and Oxidant Model) and a Lagrangian model with linear chemistry (Ontario Ministry of the Environment Trajectory Model) were used to simulate the wet SO₄²⁻ deposition pattern over eastern North America for 16 days during April 1981.The two model results agree reasonably well with each other when the 16 day average values are compared. They also show reasonable agreement with observed data. Having established the ability of the models to predict deposition patterns for 1981 emissions, reduction scenarios with 50% SO_x and 50% SO_x and NO_x of the 1981 emissions were studied through the Eulerian model. Near the heavy emissions area, the reduction in SO₄²⁻ wet deposition is only about 30–40%. In this respect the linear Lagrangian model departs significantly from the Eulerian model. This non-linearity in response is attributed to the role of oxidants in controlling the conversion of SO₂ to SO₄²⁻.     

The MICS-Asia study: model intercomparison of long-range transport and sulfur deposition in East Asia

An intercomparison study involving eight long-range transport models for sulfur deposition in East Asia has been initiated. The participating models included Eulerian and Lagrangian frameworks, with a wide variety of vertical resolutions and numerical approaches. Results from this study, in which models used common data sets for emissions, meteorology, and dry, wet and chemical conversion rates, are reported and discussed. Model results for sulfur dioxide and sulfate concentrations, wet deposition amounts, for the period January and May 1993, are compared with observed quantities at 18 surface sites in East Asia. At many sites the ensemble of models is found to have high skill in predicting observed quantities. At other sites all models show poor predictive capabilities. Source–receptor relationships estimated by the models are also compared. The models show a high degree of consistency in identifying the main source–receptor relationships, as well as in the relative contributions of wet/dry pathways for removal. But at some locations estimated deposition amounts can vary by a factor or 5. The influence of model structure and parameters on model performance is discussed. The main factors determining the deposition fields are the emissions and underlying meteorological fields. Model structure in terms of vertical resolution is found to be more important than the parameterizations used for chemical conversion and removal, as these processes are highly coupled and often work in compensating directions.     

Source apportionment of ambient total suspended particulates and coarse particulate matter in urban areas of Jiaozuo, China

Approximately 750 total suspended particulates (TSPs) and coarse particulate matter (PM10) filter samples from six urban sites and a background site and >210 source samples were collected in Jiaozuo City during January 2002 to April 2003. They were analyzed for mass and abundances of 25 chemical components. Seven contributive sources were identified, and their contributions to ambient TSP/PM10 levels at the seven sites in three seasons (spring, summer, and winter days) and a "whole" year were estimated by a chemical mass balance (CMB) receptor model. The spatial TSP average was high in spring and winter days at a level of approximately 530 microg/m(3) and low in summer days at 456 microg/m(3); however, the spatial PMo0 average exhibited little variation at a level of approximately 325 microg/m(3), and PM10-to-TSP ratios ranged from 0.58 to 0.81, which suggested heavy particulate matter pollution existing in the urban areas. Apportionment results indicated that geological material was the largest contributor to ambient TSP/PM10 concentrations, followed by dust emissions from construction activities, coal combustion, secondary aerosols, vehicle movement, and other industrial sources. In addition, paved road dust and re-entrained dust were also apportioned to the seven source types and found soil, coal combustion, and construction dust to be the major contributors.     

An urban emissions inventory for South America and its application in numerical modeling of atmospheric chemical composition at local and regional scales

This work describes the development of an urban vehicle emissions inventory for South America, based on the analysis and aggregation of available inventories for major cities, with emphasis on its application in regional atmospheric chemistry modeling. Due to the limited number of available local inventories, urban emissions were extrapolated based on the correlation between city vehicle density and mobile source emissions of carbon monoxide (CO) and nitrogen oxides (NOx). Emissions were geographically distributed using a methodology that delimits urban areas using high spatial resolution remote sensing products. This numerical algorithm enabled a more precise representation of urban centers. The derived regional inventory was evaluated by analyzing the performance of a chemical weather forecast model in relation to observations of CO, NOx and O₃ in two different urban areas, São Paulo and Belo Horizonte. The gas mixing ratios simulated using the proposed regional inventory show good agreement with observations, consistently representing their hourly and daily variability. These results show that the integration of municipal inventories in a regional emissions map and their precise distribution in fine scale resolutions are important tools in regional atmospheric chemistry modeling.     

Source apportionment of fine particles utilizing partially speciated carbonaceous aerosol data at two rural locations in New York State

A source apportionment study was conducted at two rural locations, Potsdam and Stockton, to assess the in-state/out-of-state sources of PM_2.5 and Hg in New York State. At both locations, samples were collected between November 2002 and August 2005 and analyzed for fine PM mass and its chemical constituents. The measured chemical constituents included elements, cations, anions, organic and elemental carbon (OC and EC), black carbon (BC), and water-soluble short-chain (WSSC) organic acids. Positive matrix factorization (PMF) was applied to the measured concentrations and eight and seven factors were resolved at Potsdam and Stockton, respectively. Four factors were resolved in common between the two locations including secondary sulfate, secondary nitrate, secondary OC, and a crustal factor. The factor profiles of mixed industrial and motor vehicle factors resolved at Potsdam were different compared with the corresponding profiles for these factors at Stockton. A resuspended road salt factor was identified at Potsdam, while an aged sea salt factor was identified at Stockton. At Potsdam, a wood smoke factor was also resolved. Among the resolved factors, secondary sulfate was the highest contributor to the measured mass at both sites. Potential source contribution function (PSCF) analysis indicated the Ohio River Valley region as a common potential source region for this factor at both locations. For the secondary nitrate factor, at Potsdam PSCF analysis indicated the Midwestern US (NO_x emissions), and the US farm belt (ammonia emissions) as potential source regions, while at Stockton, the Midwestern US (power plant NO_x emissions) was indicated as a major potential source region.     

Compositions of Particles from Selected Sources in Philadelphia for Receptor Modeling Applications

As a part of a receptor model study of the Philadelphia, PA atmosphere, particulate samples were collected from seven air pollution sources in the area: two oil-fired power plants, a coal-fired power plant, a fluidized catalytic cracker, a refuse incinerator, a secondary aluminum smelter and an antimony ore roaster. Samples were collected In two size fractions with a dilution source sampler connected to a modified dichotomous sampler. Masses of collected material were determined gravlmetrlcally. Samples were analyzed for elements by x-ray fluorescence followed by Instrumental neutron activation analysis of some samples. Other samples were analyzed by chemical methods for volatile and nonvolatile carbon, SO₄ ^2? and NH₄ ⁺. Data are presented for up to 46 elements and species on fine (<2.5-μm aerodynamic equivalent diameter) and coarse (2.5 μm < diam < 7-10 μm) particles from each source. Although the data were collected for use in Philadelphia, they should be of value for receptor modeling of other areas having similar sources. The most unexpected results were the large amounts of rare earth elements on particles from the catalytic cracker (e.g., 0.31 percent La In fine fraction) and the oil-fired power plants (120 and 420 ppm La in fine fraction). Substantial amounts of primary SO₄ ^2? are released from the oil-fired plants, the SO₄ ^2? concentrations accounting for 40-45 percent of the fine particulate mass.     

Source apportionment of particulate matter (PM2.5) in an urban area using dispersion,receptor and inverse modelling

Air pollution emission inventories are the basis for air quality assessment and management strategies. The quality of the inventories is of great importance since these data are essential for air pollution impact assessments using dispersion models. In this study, the quality of the emission inventory for fine particulates (PM_2.5) is assessed: first, using the calculated source contributions from a receptor model; second, using source apportionment from a dispersion model; and third, by applying a simple inverse modelling technique which utilises multiple linear regression of the dispersion model source contributions together with the observed PM_2.5 concentrations. For the receptor modelling the chemical composition of PM_2.5 filter samples from a measurement campaign performed between January 2004 and April 2005 are analysed. Positive matrix factorisation is applied as the receptor model to detect and quantify the various source contributions. For the same observational period and site, dispersion model calculations using the Air Quality Management system, AirQUIS, are performed. The results identify significant differences between the dispersion and receptor model source apportionment, particularly for wood burning and traffic induced suspension. For wood burning the receptor model calculations are lower, by a factor of 0.54, but for the traffic induced suspension they are higher, by a factor of 7.1. Inverse modelling, based on regression of the dispersion model source contributions and the PM_2.5 concentrations, indicates similar discrepancies in the emissions inventory. In order to assess if the differences found at the one site are generally applicable throughout Oslo, the individual source category emissions are rescaled according to the receptor modelling results. These adjusted PM_2.5 concentrations are compared with measurements at four independent stations to evaluate the updated inventory. Statistical analysis shows improvement in the estimated concentrations for PM_2.5 at all sites. Similarly, inverse modelling is applied at these independent sites and this confirms the validity of the receptor model results.     

Quantification of self pollution from two diesel school buses using three independent methods

We monitored two Seattle school buses to quantify the buses’ self pollution using the dual tracers (DT), lead vehicle (LV), and chemical mass balance (CMB) methods. Each bus drove along a residential route simulating stops, with windows closed or open. Particulate matter (PM) and its constituents were monitored in the bus and from a LV. We collected source samples from the tailpipe and crankcase emissions using an on-board dilution tunnel. Concentrations of PM₁, ultrafine particle counts, elemental and organic carbon (EC/OC) were higher on the bus than the LV. The DT method estimated that the tailpipe and the crankcase emissions contributed 1.1 and 6.8 μg m^?3 of PM_2.5 inside the bus, respectively, with significantly higher crankcase self pollution (SP) when windows were closed. Approximately two-thirds of in-cabin PM_2.5 originated from background sources. Using the LV approach, SP estimates from the EC and the active personal DataRAM (pDR) measurements correlated well with the DT estimates for tailpipe and crankcase emissions, respectively, although both measurements need further calibration for accurate quantification. CMB results overestimated SP from the DT method but confirmed crankcase emissions as the major SP source. We confirmed buses’ SP using three independent methods and quantified crankcase emissions as the dominant contributor.     

Wood smoke as a source of particle-phase organic compounds in residential areas

The objective of this study was to investigate the organic composition of wood smoke emissions and ambient air samples in order to determine the wood smoke contribution to the ambient air pollution in the residential areas. From November 2005 to March 2006 particle-phase PM₁₀ samples were collected in the residential town Dettenhausen surrounded by forests near Stuttgart in southern Germany. Samples collected on pre-baked glass fibre filters were extracted using toluene with ultrasonic bath and analysed by gas chromatography mass spectrometry (GC-MS). 21 polycyclic aromatic hydrocarbons (PAH) including 16 USEPA priority pollutants, different organic wood smoke tracers, primarily 21 species of syringol and guaiacol derivatives, levoglucosan and its isomers mannosan, galactosan and dehydroabietic acid were detected and quantified in this study. The concentrations of these compounds were compared with the fingerprints of emissions from hardwood and softwood combustion carried out in test facilities at Universitaet Stuttgart and field investigations at a wood stove during real operation in Dettenhausen. It was observed that the combustion derived PAH was detected in higher concentrations than other PAH in the ambient air PM₁₀ samples. Syringol and its derivatives were found in large amounts in hardwood burning but were not detected in softwood burning emissions. On the other hand, guaiacol and its derivatives were found in both softwood and hardwood burning emissions, but the concentrations were higher in the softwood smoke compared to hardwood smoke. So, these compounds can be used as typical tracer compounds for the different types of wood burning emissions. In ambient air samples both syringol and guaiacol derivatives were found which indicates the wood combustion contribution to the PM load in such residential areas. Levoglucosan was detected in high concentrations in all ambient PM₁₀ samples. A source apportionment modelling, Positive Matrix Factorization (PMF) was implemented to quantify the wood smoke contribution to the ambient PM₁₀ bound organic compounds in the residential area.     

Fine Particle Receptor Modeling in the Atmosphere of Mexico City

Abstract

Source apportionment analyses were carried out by means of receptor modeling techniques to determine the contribution of major fine particulate matter (PM_2.5) sources found at six sites in Mexico City. Thirty-six source profiles were determined within Mexico City to establish the fingerprints of particulate matter sources. Additionally, the profiles under the same source category were averaged using cluster analysis and the fingerprints of 10 sources were included. Before application of the chemical mass balance (CMB), several tests were carried out to determine the best combination of source profiles and species used for the fitting. CMB results showed significant spatial variations in source contributions among the six sites that are influenced by local soil types and land use. On average, 24-hr PM_2.5 concentrations were dominated by mobile source emissions (45%), followed by secondary inorganic aerosols (16%) and geological material (17%). Industrial emissions representing oil combustion and incineration contributed less than 5%, and their contribution was higher at the industrial areas of Tlalnepantla (11%) and Xalostoc (8%). Other sources such as cooking, biomass burning, and oil fuel combustion were identified at lower levels. A second receptor model (principal component analysis, [PCA]) was subsequently applied to three of the monitoring sites for comparison purposes. Although differences were obtained between source contributions, results evidence the advantages of the combined use of different receptor modeling techniques for source apportionment, given the complementary nature of their results. Further research is needed in this direction to reach a better agreement between the estimated source contributions to the particulate matter mass.     

An Environmental Decision-Making Tool for Evaluating Ground-Level Ozone-Related Health Effects

Abstract

A computer model called the Ozone Risk Assessment Model (ORAM) was developed to evaluate the health effects caused by ground-level ozone (O₃) exposure. ORAM was coupled with the U.S. Environmental Protection Agency’s (EPA) Third-Generation Community Multiscale Air Quality model (Models-3/CMAQ), the state-of-the-art air quality model that predicts O₃ concentration and allows the examination of various scenarios in which emission rates of O₃ precursors (basically, oxides of nitrogen [NO_x] and volatile organic compounds) are varied. The principal analyses in ORAM are exposure model performance evaluation, health-effects calculations (expected number of respiratory hospital admissions), economic valuation, and sensitivity and uncertainty analysis through a Monte Carlo simulation. As a demonstration of the system, ORAM was applied to the eastern Tennessee region, and the entire O₃ season was simulated for a base case (typical emissions) and three different emission scenarios. The results indicated that a synergism occurs when reductions in NO_x emissions from mobile and point sources were applied simultaneously. A 12.9% reduction in asthma hospital admissions is expected when both mobile and point source NO_x emissions are reduced (50 and 70%, respectively) versus a 5.8% reduction caused by mobile source and a 3.5% reduction caused by point sources when these emission sources are reduced individually.     

Analysis of C1, C2, and C10 through C33 particle-phase and semi-volatile organic compound emissions from heavy-duty diesel engines

To meet increasingly stringent regulations for diesel engines, technologies such as combustion strategies, aftertreatment components, and fuel composition have continually evolved. The emissions reduction achieved by individual aftertreatment components using the same engine and fuel has been assessed and published previously (Liu et al., 2008a, Liu et al., 2008b, Liu et al., 2008c). The present study instead adopted a systems approach to evaluate the net effect of the corresponding technologies for model-year 2004 and 2007 engines. The 2004 engine was equipped with an exhaust gas recirculation (EGR) system, while the 2007 engine had an EGR system, a crankcase emissions coalescer, and a diesel particulate filter. The test engines were operated under the transient federal test procedure and samples were collected with a source dilution sampling system designed to stimulate atmospheric cooling and dilution conditions. The samples were analyzed for elemental carbon, organic carbon, and C₁, C₂, and C₁₀ through C₃₃ particle-phase and semi-volatile organic compounds. Of the more than 150 organic species analyzed, the largest portion of the emissions from the 2004 engine consisted of formaldehyde, acetaldehyde, and naphthalene and its derivatives, which were significantly reduced by the 2007 engine and emissions technology. The systems approach in this study simulates the operation of real-world diesel engines, and may provide insight into the future development of integrated engine technology. The results supply updated information for assessing the impact of diesel engine emissions on the chemical processes, radiative properties, and toxic components of the atmosphere.     

Sources and seasonal variation of atmospheric polycyclic aromatic hydrocarbons in Dalian,China: Factor analysis with non-negative constraints combined with local source fingerprints

Vapor- and particulate-phase polycyclic aromatic hydrocarbon (PAH) samples were continuously collected at an urban site in Dalian, China, during the heating and non-heating period. There is strong temperature dependence and obvious seasonal trend for atmospheric PAHs, and significant positive correlations of atmospheric PAHs with SO₂ and CO concentrations were observed. Factor analysis model with non-negative constraints (FA–NNC) combined with local and literature PAH source fingerprints was successful in source identification of particulate PAHs in the atmospheric samples. The results suggested that, in heating period, the main pollution sources were identified as coal-fired boiler emission (56%), residential coal combustion (33%) and traffic emissions (11%). As for non-heating period, the main sources were gasoline engine emission, traffic tunnel emission and coal-fired power plant, and the overall source contributions of traffic emission (gasoline engine + traffic tunnel) were 79% and coal-fired power plant 21%. The current results support our previous study and provide new insights. This can be the first attempt to quantitatively apportion air organic pollutants using receptor models combined with local source fingerprints. The source fingerprints can be used as reference data for source apportionment of atmospheric PAHs of China.     

C1–C8 volatile organic compounds in the atmosphere of Hong Kong: Overview of atmospheric processing and source apportionment

We present measurements of C₁–C₈ volatile organic compounds (VOCs) at four sites ranging from urban to rural areas in Hong Kong from September 2002 to August 2003. A total of 248 ambient VOC samples were collected. As expected, the urban and sub-urban sites generally gave relatively high VOC levels. In contrast, the average VOC levels were the lowest in the rural area. In general, higher mixing ratios were observed during winter/spring and lower levels during summer/fall because of seasonal variations of meteorological conditions. A variation of the air mass composition from urban to rural sites was observed. High ratios of ethyne/CO (5.6 pptv/ppbv) and propane/ethane (0.50 pptv/pptv) at the rural site suggested that the air masses over the territory were relatively fresh as compared to other remote regions. The principal component analysis (PCA) with absolute principal component scores (APCS) technique was applied to the VOC data in order to identify and quantify pollution sources at different sites. These results indicated that vehicular emissions made a significant contribution to ambient non-methane VOCs (NMVOCs) levels in urban areas (65±36%) and in sub-urban areas (50±28% and 53±41%). Other sources such as petrol evaporation, industrial emissions and solvent usage also played important roles in the VOC emissions. At the rural site, almost half of the measured total NMVOCs were due to combustion sources (vehicular and/or biomass/biofuel burning). Petrol evaporation, solvent usage, industrial and biogenic emissions also contributed to the atmospheric NMVOCs. The source apportionment results revealed a strong impact of anthropogenic VOCs to the atmosphere of Hong Kong in both urban/sub-urban and rural areas.     

Source apportionment of VOCs in the Los Angeles area using positive matrix factorization

Eight 3-h speciated hydrocarbon measurements were collected daily by the South Coast Air Quality Management District (SCAQMD) as part of the Photochemical Assessment Monitoring Stations (PAMS) program during the summers of 2001–03 at two sites in the Los Angeles air basin, Azusa and Hawthorne. Over 30 hydrocarbons from over 500 samples at Azusa and 600 samples at Hawthorne were subsequently analyzed using the multivariate receptor model positive matrix factorization (PMF). At Azusa and Hawthorne, five and six factors were identified, respectively, with a good comparison between predicted and measured mass. At Azusa, evaporative emissions (a median of 31% of the total mass), motor vehicle exhaust (22%), liquid/unburned gasoline (27%), coatings (17%), and biogenic emissions (3%) factors were identified. Factors identified at Hawthorne were evaporative emissions (a median of 34% of the total mass), motor vehicle exhaust (24%), industrial process losses (15%), natural gas (13%), liquid/unburned gasoline (13%), and biogenic emissions (1%). Together, the median contribution from mobile source-related factors (exhaust, evaporative emissions, and liquid/unburned gasoline) was 80% and 71% at Azusa and Hawthorne, respectively, similar to previous source apportionment results using the chemical mass balance (CMB) model. There is a difference in the distribution among mobile source factors compared to the CMB work, with an increase in the contribution from evaporative emissions, though the cause (changes in emissions or differences between models) is unknown.