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.sub.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.

INTRODUCTION

Motor vehicles are a major source of particulate matter (PM) to the urban atmosphere. However, results of studies investigating the contributions of vehicle tailpipe emissions to urban PM have often differed, especially concerning the relative impacts of gasoline and diesel vehicles. (1-3) The range of results suggests that the chemical mass balance (CMB) models used for source apportionment are sensitive to variation in the methods used to create chemical source profiles for vehicle emissions. One source of variation in source profiles and models results is the use of different methods for measurement of the operationally defined organic carbon (OC) and elemental carbon (EC) fractions, which has been explored previously. (4) The other major source of uncertainty in vehicle emissions source profiles are the relationships between representative vehicle profiles and actual on-road fleets of vehicles.

Vehicle tailpipe emissions profiles are developed through testing vehicle emissions on chassis dynamometers. The uncertainty in source profiles arises from testing of a small number of vehicles to represent a large fleet in an area, from applicability of tested driving cycles to on-road driving patterns, and from the unknown numbers of high-emitting or poorly-functioning vehicles in on-road fleets that are not adequately represented in testing. Compared with other combustion sources, it is difficult to develop representative profiles for emissions from gasoline and diesel vehicles. Other sources, such as coal burning or open wood burning, have a relatively limited set of possible combustion conditions, and therefore a limited range of emissions compositions. Alternatively, motor vehicle emissions are dependent upon such factors as fuel, engine condition, ambient temperature, driving cycle, and lubricating oil age. (5-10) These factors vary not only between vehicles of different ages and weight classes, but also for an individual vehicle in different circumstances. (11) Therefore, developing useable source emissions profiles that are representative of the actual on-road emissions from a fleet of vehicles requires understanding of the variability of emissions profiles.

In this study, a large number of gasoline and diesel vehicles were tested on chassis dynamometers over several driving cycles. The chemical compositions of the PM emissions were averaged to create profiles for gasoline and diesel tailpipe emissions for source apportionment of ambient PM collected in Los Angeles. Profiles were developed for the most realistic vehicle fleets possible, using data on vehicle ages and weight classes for the Los Angeles area. To investigate the impact of a variety of emissions conditions on the model results, additional profiles were developed for different vehicle ages, for different driving cycles, and for gasoline smoker vehicles with visible plumes. The vehicle profiles were also used to conduct CMB source apportionment of day-of-week composite samples of ambient PM to observe changes in the contribution of gasoline and diesel vehicles to ambient concentrations with the day of the week. (12)

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

METHODS

Ambient Samples

Ambient samples were collected at two sites in Los Angeles (LA), CA, over 3 weeks in July 2001. (13) The first site was located in Azusa, a suburb of LA, in an area that is residential and industrial. The second site was in downtown LA, near a main highway in an industrial area. Daily 24-hr [PM.sub.2.5] (PM that is 2.5 [micro]m in size or smaller) samples were collected with a 92-LPM cyclone and 90-mm filter holder (URG Inc.) on prebaked quartz fiber filters. A punch of each filter was taken for analysis of EC and OC using the National Institute for Occupational Safety and Health 5040 method. (14) The remainder of the filters were composited by the day of the week for detailed organic compound analysis, resulting in seven composites of three filters at each site (i.e., the three Monday Azusa filters were composited). These day-of-week composites were extracted with solvents and analyzed by GC-mass spectroscopy (GCMS). Details of sample preparation (15) and GCMS analysis (16) have been described elsewhere. (13) Filters were spiked with a set of internal standard compounds before extraction. Two sequential extractions were performed with a Soxhlet apparatus, one with dichloromethane and one with methanol, and the two extracts were combined. Rotary evaporation and nitrogen blowdown decreased total extract volume to 0.25 mL. Derivatization of an aliquot of the extract with diazomethane allowed quantification of organic acids as their methyl ester analogs. A portion of this methylated aliquot was then silylated to allow quantification of very polar compounds as their trimethylsilyl derivatives. All three aliquots of extract were analyzed with the same GCMS conditions.

[FIGURE 3 OMITTED]

Data for ambient mass, ionic species (sulfate, nitrate, and ammonium), and trace metals were obtained from measurements made by Desert Research Institute (DRI). (17,18)

Field blanks were collected at regular intervals, and all sample data were corrected by subtraction of blank results. Mass of a species on a filter or in a filter composite, after blank correction, was divided by the total volume of air sampled to obtain concentrations in mass per volume of air.

Source Profiles

Gasoline Engine Emissions. Emissions profiles for gasoline and diesel vehicles were developed in this work. Sample collection has been described completely elsewhere, (10) and briefly described here. To develop the gasoline profile, PM emissions from 54 gasoline vehicles (Table 1) operated on a chassis dynamometer were sampled from diluted exhaust. Emissions were sampled from each vehicle over cold-start and warm-start phases of the unified driving cycles. Analysis for EC, OC, and speciated organic compounds was the same as for ambient samples. The cold-start and warm-start results were averaged by model year class, using data on the age distribution of vehicles in the LA area in 2001 provided by the California Air Resources Board (CARB). The cold-start and warm-start profiles were then averaged on a mass-weighted basis to create a single profile for gasoline vehicle emissions. All species were normalized to OC, and the profile is presented in units of mass species per mass OC. This profile is shown in Figure 1. In all vehicle profiles, uncertainties for each species represent the total measurement uncertainties in the averaged vehicle tests, propagated through the mass-weighted averaging as the square root of the sum of the squares.

Diesel Engine Emissions. Development of the profile for diesel emissions was similar to that of gasoline emissions. A total of 33 diesel vehicles (Table 2) were operated on a chassis dynamometer through several driving cycles. The results from all driving cycles and vehicle weight classes were averaged based on vehicle age data for the LA area, as discussed in other work. (10) The resulting source profile, normalized to OC, is shown in Figure 1. The diesel profile contains a large fraction of EC, which the model relies on to apportion to this source. Whereas EC/organic carbon ratios in the emissions of individual diesel vehicles have been observed in this (10) and other studies (19,20) to vary over more than an order of magnitude, the robust approach taken in this study to average measured emissions to represent on-road vehicle fleets decreased the EC/OC variability.

[FIGURE 4 OMITTED]

Smoker and Microenvironment Profiles. Seven smoker vehicles with either visible emission plumes or emissions of greater than 50 mg x [mi.sup.-1] were also tested during testing of gasoline vehicles (Table 1). (10) Three of the vehicles had emissions profiles similar to what is expected of a "normal" oil-burning smoker vehicle, with high mass fractions of OC, high concentrations of hopane and sterane compounds, and low concentrations of polycyclic aromatic hydrocarbons (PAHs). (10) The other four tested smoker vehicles had a wide range of emission compositions and rates, reflecting the fact that a number of mechanisms other than oil burning can cause a vehicle to be a smoker. A profile of smoker vehicle emissions similar to the normal oil-burning smoker profile was used as the base case in the model, and is shown at the top of Figure 2.

[FIGURE 5 OMITTED]