2005 Progress Report: Emissions Inventory and Process Reconciliation Using Molecular Markers and Hybrid/Inverse Photochemical Modeling with Direct Sensitivity Analysis

EPA Grant Number: R831076
Title: Emissions Inventory and Process Reconciliation Using Molecular Markers and Hybrid/Inverse Photochemical Modeling with Direct Sensitivity Analysis
Investigators: Russell, Armistead G. , Odman, Mehmet Talat , Zheng, M.
Institution: Georgia Institute of Technology
EPA Project Officer: Chung, Serena
Project Period: November 1, 2003 through October 30, 2006 (Extended to September 30, 2008)
Project Period Covered by this Report: November 1, 2004 through October 31,2005
Project Amount: $449,899
RFA: Measurement, Modeling, and Analysis Methods for Airborne Carbonaceous Fine Particulate Matter (PM2.5) (2003) RFA Text |  Recipients Lists
Research Category: Air , Air Quality and Air Toxics , Particulate Matter

Objective:

The goal of this research project is to develop and improve integrated source/receptor-based methods for source apportionment of fine aerosol. The specific objectives are to: (1) conduct the optimal methodology for organic tracer analysis and source apportionment; (2) evaluate Community Multi-scale Air Quality (CMAQ) results using monitoring data and molecular marker based chemical mass balance-molecular marker (CMB-MM) results; (3) improve the emission inventories using CMAQ evaluation results; and (4) extend CMAQ functions for secondary organic aerosol (SOA) analysis and the Direct Decoupled Method.

Progress Summary:

Quantitative Uncertainty Study of CMB-MM Modeling Using the Monte Carlo Approach

Uncertainties of CMB-MM source apportionment results are caused, in part, by uncertainties in input data. To quantify uncertainties, two types of distributions, normal and log-normal distributions, were considered separately in source profiles and ambient data. The results of Monte Carlo simulations indicated higher uncertainties in the calculated source impacts when using either distribution as compared to those routinely calculated by CMB, especially for meat cooking. The uncertainties of the meat cooking contribution in organic carbon (OC) were 136 percent, 116 percent, and 46 percent for normal distribution, lognormal distribution, and CMB-MM output (default), respectively.

Conversion of OC to EC Ratios From IMPROVE Ambient Data to NIOSH Data in CMB-MM Input Files

There are two different, commonly used thermal carbon analysis methods to quantify OC and elemental carbon (EC) in ambient samples: the Interagency Monitoring of Protected Visual Environments (IMPROVE) and the National Institute for Occupational Safety and Health (NIOSH) method. Ambient measurements from the Southeast Aerosol Research Characterization (SEARCH) network typically used IMPROVE, but CMB-MM source profile measurements used the NIOSH method. To eliminate impacts on model results from these unmatched analysis methods, conversion factors between IMPROVE and NIOSH data were calculated by statistical analysis of about 130 SEARCH samples subjected to both approaches. Different factors were obtained at different sampling sites and in different seasons, and used to convert IMPROVE OC/EC of ambient data to NIOSH OC/EC. After conversion, CMB results showed that the diesel exhaust contribution in PM2.5 dropped, on average, by 65 percent and 76 percent for Jefferson Street summer and winter daily samples, respectively. Gasoline vehicle exhaust contribution, however, increased on average by 82 percent and 15 percent for the same daily samples at summer and winter, respectively. No significant changes were found regarding other major source contributions.

Intercomparison Between CMB-MM and CMB-Regular Modeling

To evaluate and improve CMB-MM applications and understand its limitations better, two receptor model approaches, inorganic species based CMB (CMB-Regular) and CMB-MM, were used simultaneously to conduct source apportionment of PM2.5 for SEARCH samples. Temporal (winter and summer) and spatial impacts (urban and rural) on source contributions were analyzed. Results indicated a few similarities in source categories and source contributions between both approaches. Motor vehicle exhaust and wood burning were the major primary sources of PM2.5 in the Southeastern United States. Obvious discrepancies, however, were found in the primary source contributions, including motor vehicle exhaust, paved road dust, and wood burning. Application of different fitting species in the CMB model was the main reason for these differences.

Field Sampling and Experimental Study During Summer 2005

Two separate, more intensive, 1.5-week field studies were conducted in the Atlanta area during the summer of 2005. During the first study, ambient monitoring was conducted not only at the SEARCH and ASACA sites, but also two additional sites: directly next to the I-75/85 freeway corridor in Atlanta and at the Georgia Tech roof laboratory. The second period of study again added sampling at the Georgia Tech roof laboratory as well as additional sampling at the SEARCH Yorkville site. A pair of pulse code modulation (PCM) and Hi-Volume samplers was run at each site. Experiment processes on PM2.5 mass, OC, EC, ions, and metal concentrations have been completed. A few potential molecular markers of SOA will be identified and quantified in the samples using updated gas chromatography/mass spectrometry (GC/MS) technique. In particular, we can detect two compounds in our samples (2-methylthreitol and 2-methylerythritol). These two compounds were reported by Claeys, et al. (2004) as secondary tracers from photoxidation of isoprene. We identified these two tracers in our samples, but the quantification is not perfect because of the lack of the standards (semiquantitative). They are abundant in summer samples, but trace or nondetectable in winter samples. Their correlations, based on samples taken in June 2004, are shown in Figure 1.

Figure 1. Correlation Between Secondary Sulfate, Other or Unexplained Organic Matter From CMB Analysis, and Secondary Organic Tracers From Photooxidation of Isoprene (2-Methylthreitol and 2-Methylerythritol)

Improving Emission Inventories Using CMAQ Evaluation Results

The NEI 2001 (National Emission Inventory 2001) by EPA was used as basic emission inventories for CMAQ modeling. Based on CMAQ results, three major changes in emission inventories were made: (1) new point sources in Georgia; (2) new forest fire emissions using 2001 EPA temporal profiles and Visibility Improvement—State and Tribal Association of the Southeast (VISTAS) forest fire emissions for 2002; and (3) correcting monthly distribution of area source emissions. Errors in point sources in Georgia were recognized because there were unreasonably high contributions from natural gas combustions and industrial processes in source apportionment results in CMAQ. Forest fire emissions were replaced by emission inventories created by the VISTAS project. VISTAS developed detailed daily emissions for the southeastern United States, which are more realistic than the data included in NEI 2001. Monthly distribution of area sources were corrected because the contribution of fireplaces was too high in July. SMOKE version 1.5 had a problem with applying a monthly temporal profile to area sources, but SMOKE version 2.1 worked correctly. The final emission inventories resulted in more convincing results for source apportionment of organic aerosol.

Source-Receptor Analysis

Comprehensive comparison between source apportionment results of CMAQ and results of receptor models were made. CMB-MM, CMB-Regular, and Positive Matrix Factorization (PMF) were used as receptor models. Comparison was made using not only source apportionment results but also concentrations of each of the fitting species that were used in receptor models. Differences between two models varied with monitoring sites and source categories. Many factors were pointed out as reasons for differences such as accuracy of CMAQ simulation, conversion factors between organic matter and organic carbon, unaccounted sources in receptor models, accuracy of source profiles, and temporal variations. To improve those reasons, the following studies are needed: (1) inverse modeling using reconciled concentrations of species; (2) site specific OM to OC conversion factors; (3) modification of source profiles in receptor models and speciation profiles in CMAQ (SMOKE) model; and (4) identifying unknown sources in receptor models (see Lin, et al.; Yan, et al.; Park, et al.; and Marmur, et al.).

Extend CMAQ Functions for Secondary Organic Aerosol Analysis and DDM

For secondary organic aerosol analysis, a literature review was conducted in Year 2. Three species were selected as secondary organic aerosol tracers. For SOA formed from monoterpenes, pinic acid and pinonic acid will be measured (Pankow, et al., 2001) and added to CMAQ SOA functions. For SOA formation from isoprene, 2-methyltetrol will be measured (Claeys, 2004). To add SOA tracers, detailed chemical species from biogenic sources are needed; monoterpene will be divided into a-Pinene, b-Pinene, and D3-Carene. To calculate detailed species emissions, we need to modify the Biogenic Emissions Inventory System (BEIS) in SMOKE. This will be done based either on a module in CMAQ-MADRID or emission factors in Helmig (1998). Update of CMAQ and Decoupled Direct Method (DDM) will be done in the third year.

Future Activities:

We will:

  • Conduct additional field experimental analysis in the winter (January).
  • Analyze 20 high-volume samples for organic tracers and perform data analysis.
  • Further conduct the optimal methodology for organic tracer analysis and source apportionment.
  • Update SOA treatment in CMAQ.
  • Modify DDM to treat the advanced SOA mechanism.
  • Check the air quality and inverse modeling of the13-month period.
  • Perform further updates of emissions inventory through reconciliation between an emission-based model and receptor models and through uncertainty analysis.

References:

Claeys M, Graham B, Vas G, Wang G, Vermeylen R, Pashynska V, Cafmeyer J, Guyon P, Andreae MO, Artaxo P, Maenhaut W. Formation of secondary organic aerosols through photooxidation of isoprene. Science 2004;303(5661):1173-1176.


Journal Articles on this Report : 1 Displayed | Download in RIS Format

Other project views: All 75 publications 40 publications in selected types All 40 journal articles
Type Citation Project Document Sources
Journal Article Park S-K, Marmur A, Kim SB, Tian D, Hu Y, McMurry PH, Russell AG. Evaluation of fine particle number concentrations in CMAQ. Aerosol Science and Technology 2006;40(11):985-996. R831076 (2005)
R831076 (2006)
R831076 (Final)
R830960 (Final)
R832159 (2005)
R832159 (2006)
R832159 (2007)
R832159 (Final)
  • Full-text: Taylor & Francis-Full Text HTML
    Exit
  • Abstract: Taylor & Francis-Abstract
    Exit
  • Other: Taylor & Francis-Full Text PDF
    Exit
  • Supplemental Keywords:

    hybrid/inverse PM modeling, emissions inventory analysis, molecular marker/organic tracer analysis, air toxics, particulate matter, aerosol analyzers, aerosol particles, air modeling, air quality modeling, air sampling, ambient air monitoring, ambient particle health effects, atmospheric particles, atmospheric particulate matter, carbon aerosols, carbon particles, direct sensitivity analysis, emissions, gas chromatography, health effects, human exposure, human health effects, mass spectrometry, measurement methods, molecular markers, particle dispersion, particle phase molecular markers, particle size, particulate matter mass, photochemical modeling,, RFA, Health, Scientific Discipline, Air, Ecosystem Protection/Environmental Exposure & Risk, particulate matter, Air Quality, air toxics, Environmental Chemistry, climate change, Air Pollution Effects, Risk Assessments, Monitoring/Modeling, Environmental Monitoring, Engineering, Chemistry, & Physics, Environmental Engineering, Atmosphere, carbon aerosols, air quality modeling, particle size, atmospheric particulate matter, health effects, aerosol particles, atmospheric particles, mass spectrometry, human health effects, ambient air monitoring, air modeling, air quality models, air sampling, gas chromatography, thermal desorption, carbon particles, air quality model, emissions, molecular markers, direct sensitivity analysis, particulate matter mass, human exposure, ambient particle health effects, particle phase molecular markers, photchemical modeling, aersol particles, particle dispersion, aerosol analyzers

    Relevant Websites:

    http://www.ce.gatech.edu/~trussell/lamda/ exit EPA

    Progress and Final Reports:

    Original Abstract
  • 2004 Progress Report
  • 2006 Progress Report
  • 2007 Progress Report
  • Final Report