The purpose of this research project is to construct and test both mathematical models that relate gaseous pollutants to the formation of new particulate matter in the atmosphere as well as models that relate the emissions of gases and primary particles to atmospheric particulate matter concentrations. In both 1996 and again in 1997, field experiments were conducted in Southern California in which the evolution of the size distribution and chemical composition of the urban aerosol complex was observed using methods that focus on the evolution of the individual aerosol particles. The data from the 1996 and 1997 experiments will be used as a model evaluation data set.
During September and October of 1996 and again in August through November 1997, field experimental programs were conducted in the Los Angeles area in which the evolution of the size distribution and chemical composition of the urban aerosol complex was observed using methods that focus on the evolution of the individual aerosol particles. In 1996, experiments were conducted in which the background marine aerosol was first characterized as it flows across the Pacific coastline in Southern California. Lagrangian air parcels were sampled as they were transported across the urban Los Angeles area from Long Beach to Fullerton to Riverside, CA, in the presence of direct emissions from urban pollution sources and as the aerosol is modified by gas-to-particle conversion processes. In 1997, air parcels were studied as they were transported from the high traffic area of central Los Angeles to a downwind site at Azusa, and nitrate aerosol formation was studied as air parcels were transported across the large ammonia source in the Chino dairy area. Both organic and inorganic aerosol species were sampled simultaneously, (1) by aerosol time-of-flight mass spectrometers that view single particle size and composition, (2) by MOUDI cascade impactors from which particle chemical composition was sampled as a function of particle size, (3) by filter-based fine particle samplers and (4) by electrical aerosol analyzers and optical particle counters that measure particle size distributions directly and continuously.
The purpose of the present project is to apply atmospheric models and other advanced data analysis methods to the study of atmospheric aerosol processes using the data base from the 1996 and 1997 Southern California field studies.
Aerosol processes air quality models have been developed that seek to compute how the size distribution of the chemical composition of the aerosol evolves as the source emissions undergo atmospheric transport and transformation. Nearly all of these models developed to date have assumed that the aerosol is "internally mixed", that is that all particles of the same size have the same chemical composition. We know both from examination of atmospheric particles via electron microscopy with X-ray detection and from Professor Prather's aerosol time of flight mass spectrometry data that this really is not true, but the internal mixture assumption persists because it renders an extremely difficult modeling calculation more tractable and because in the past there were no atmospheric data bases on single particle composition that could be used to check the ability to model single particle chemical characteristics in any case.
Within the last two years both of these barriers to further progress in modeling particle properties at more nearly the single particle level have been relaxed. First, from the successful conclusion of the 1996 and 1997 Southern California field experiments described above, we now have for the first time data on single particle size and composition collected within the format of trajectory-oriented field studies that are suitable for use in studying particle evolution over time as air masses are transported across successive air monitoring sites. Such data can be used as an air quality model verification data set. Second, a new aerosol processes trajectory model has been developed that represents the atmospheric particles as a "source-oriented external mixture" in which particles from the different major source types in an urban area evolve separately from each other, thereby allowing the model to track particle-to-particle differences in chemical composition for particles of the same size in the atmosphere. That model also retains a description of individual particle composition, and can identify those particles that evolve by incorporation of water and heterogeneous liquid-phase chemistry separately from hydrophobic particles that do not easily grow upon humidification, for example.
The experiments conducted during the 1996 and 1997 field studies are ideally suited to providing the atmospheric data on single particle chemical composition needed to test this Lagrangian particle air quality model. The data also are well suited to the development and testing of the individual modules that form the building blocks within comprehensive aerosol processes air quality models in order to describe gas/particle partitioning and light scattering, for example. The objective of this project is to pursue such model evaluation studies, and in the process to learn more about the processes that govern particle evolution in the atmosphere.
The experimental design followed during the 1996 and 1997 field experiments was oriented toward following air parcel trajectories. Data analyses and model testing thus will be cast into a trajectory model format. One method of data analysis involves examination of the thermodynamic state of the atmospheric gas/particle mixture during the experiments. An investigation will be conducted to assess the extent to which alternative aerosol thermodynamic modules (e.g. the equilibrium model SCAPE vs. the kinetic model AIM) can represent the results of these experiments for the HNO3/NH3/HCI/H2SO4/aerosol Cl-/SO4=/NO3-/NH4+/Na+ system. A second phase of the research involves the application of complete air quality models to explain the relationship between emissions and air quality for both organic and inorganic particulate matter at nearly the single particle level. The air quality models used will be either the aerosol processes trajectory model of Kleeman et al. or novel chemical mass balance receptor models that relate source contributions to ambient particulate matter concentrations.
An air quality model that follows the evolution of single particles in the atmosphere has been combined with new emissions measurements and then used to predict the size distribution and chemical composition of the airborne fine particle mixture observed at Long Beach, Fullerton, and Riverside, CA, during September, 1996. Comparison of air quality model predictions to both filter-based and cascade impactor-based measurements of particle size and chemical composition show good agreement at all three air monitoring sites. Because the air quality model used tracks individual particles as they evolve over time, it is possible to also track the sources from which the primary seed particles originally were released, to study the source contributions to particulate matter size and composition at individual air monitoring sites, and to examine the secondary aerosol coatings of ammonium, sulfate, nitrate and organic compounds that have accumulated over time on particles released from different types of sources. Four major classes of particles are observed: (1) large mineral dust and road dust particles that accumulate only small amounts of secondary aerosol products; (2) small primary particles released initially in the 0.1-0.3 µm particle diameter size range from combustion sources (especially diesel vehicles and non-catalyst gasoline-powered vehicles) and food processing that grow by accumulation of secondary reaction products; (3) sea salt particles that are almost completely transformed by conversion from NaCl to NaNO3 during transport across the air basin, and (4) sulfate-containing non-sea salt background particles advected into the air basin from upwind over the ocean at a PM2.5 concentration of only 8 µg m-3 which by virtue of their size, solubility and initial presence in the air manage to grow by gas-to-particle conversion to produce a largely nitrate-containing aerosol having a PM2.5 concentration of 40 µg m-3 by the time that the air masses studied here reach Riverside, CA. These transformed non-sea salt background particles have a size distribution centered around 0.4-0.6 µm particle diameter, which is the optimal size for scattering visible solar radiation. As a result, these transformed background particles probably play a very significant role in the creation of the Los Angeles visibility problem. The behavior of the particle formation processes observed through modeling of the 1996 air pollution episodes studied here is qualitatively similar to results obtained by recently completed analyses of 1987 data from the SCAQS experiments, suggesting coherence of the nature of the Los Angeles fine particle problem over time.
As the progress reported above indicates, we are in many respects running ahead of our initially proposed schedule of accomplishments. However, even in the presence of this success, modifications to the scope of work are necessary, for the following reason. In our original proposal, we noted that additional support for air quality model development and testing was then being sought and that if additional funds were awarded from other sources then the scope of work under the present research program would be modified to avoid any duplication of effort. In mid-1998, we received an award from the USEPA as part of a special 12-investigator program designed to speed the construction and testing of combined air quality models for ozone and fine particles. We are presently in the process of separating the work under the ozone and fine particles grant from the work being done under the CAO-sponsored program. The trajectory modeling study described above was completed quickly by employing CAO funds for set-up of the model evaluation data, and ozone and fine particle grant funds for model execution, and will be reported to both sponsors as a result. In order to capitalize on the success of this approach, we propose to extend the modeling studies to include trajectory modeling of the 1997 Southern California experiments described in the (revised) statement of project objectives (in the original CAO proposal only the 1996 data were to be studied). Again we propose to use CAO funds for model evaluation data base creation, and ozone and fine particles grant funds for model execution. Evaluation of aerosol thermodynamic models and receptor models against the Southern California data will remain the exclusive province of the CAO project, while extension of aerosol processes models into a grid-based format will become the exclusive province of the ozone and fine particles project. In this way maximum efficiency will be maintained while the work performed under the two projects are distinguished from each other.
Lara Hughes, Michael Kleeman
Jonathan Allen