Emissions Inventory and Process Reconciliation Using Molecular Markers and Hybrid/Inverse Photochemical Modeling with Direct Sensitivity AnalysisEPA 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 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
The primary objectives of this research are to:
• Further develop and apply of a method for assessing emissions estimates of pollutant precursors and their impact on air quality by reconciling bottom-up and top-down emissions estimates using inverse modeling.
• Assess and improve the emissions inventory for primary organic particulate matter in the eastern United States with particular focus on the Southeast.
• Quantify the fraction of primary vs. secondary organic aerosol (SOA) and the fractions of SOA that are biogenic and anthropogenic. These results will be compared with results using other methods.
• Estimate the response of ambient PM2.5 to emissions changes by source category.
• Quantify uncertainties in emissions and source-receptor relationships for the major sources of primary organic matter and precursors to SOA.
• Assess the information added by using molecular markers in the inverse modeling and using longer periods with more routine measurements.
• Provide information on the impact of SOA parameters on simulated OC levels.
• Improve the current methodology by detecting more polar compounds and lower the detection limit for organic tracer analysis by silylation.
• Optimize the number of species applied in the chemical mass balance model so that only a subset of the important and necessary tracers will be included in the model.
• Investigate the sensitivity of the organic tracer-based receptor model technique by comparing the chemical mass modeling results with using different number of tracers, different source profiles, and including or excluding more inorganic tracers.
Specific hypotheses to be tested in achieving these objectives are that (1) inverse modeling can help identify and quantify emissions inventory errors, (2) a majority of the organic PM2.5 is primary in origin in both urban and rural areas throughout the Southeast (though SOA may dominate during periods of low visibility in the summer), (3) most of the SOA in the SE is from biogenic emissions, (4) specific molecular markers, and ratios of specific molecules, can be used to characterize the sources of OC, and (5) molecular markers increase the certainty of the inverse modeling approach.
Proposed is a combined simulation and field experimental project to identify errors in the estimate of primary particulate organic and inorganic carbon emissions and provide improved information on the formation of secondary organic fine particulate matter. CMAQ, with SAPRC99, an advanced treatment of SOA formation and gas-particle partitioning, and direct sensitivity analysis will be used. Inverse modeling will be accomplished using ridge regression along with the sensitivities to individual source emissions. The inverse air quality modeling will be used to estimate biases in emissions estimates using three periods–July/August 2001, January 2002 and August 1999. Our primary focus will be on the eastern United States, utilizing the eastern Supersites ESP01 and ESP02 data, and speciated organic analysis of filters from the EPA speciation network. Spatially detailed analysis will be conducted in the Atlanta. An updated, detailed PM emissions inventory is also being developed for Georgia, providing an improved starting point for the inverse modeling.
After the initial modeling study, where key tracer species and tracer species relations are hypothesized that characterize SOA origins, further organic molecular marker analysis will be conducted to evaluate the findings. Further bottom–up estimates will address biases found from inverse modeling, potentially reconciling differences. Part of the process in testing the inverse modeling, and source-oriented modeling in general, is comparison of the results from the CMAQ modeling with those from receptor modeling using molecular makers. As part of this, we plan to optimize the receptor modeling species being used and conduct an uncertainty analysis of the technique using Monte Carlo analysis. Finally, one of the current limitations of the molecular marker technique is the small amount of any one organic being present, such that longer sampling times are often required. We will address this by assessing silylation for improved quantification.
The research will identify errors in the estimate of primary particulate organic and inorganic carbon emissions and provide improved information on the formation of secondary organic fine particulate matter. In addition, the project will identify specific compounds that prove effective in source apportionment of organic PM, and develop a new, potentially more powerful approach for conducting molecular marker, organic tracer analysis using silylation. Further, we will develop a technique to lower the detection limits for some species measured as part of organic tracer analysis. Finally, we plan to conduct an uncertainty analysis of the molecular marker technique.