Integrated Source/Receptor-Based Methods for Source Apportionment and Area of Influence AnalysisEPA Grant Number: R832159
Title: Integrated Source/Receptor-Based Methods for Source Apportionment and Area of Influence Analysis
Investigators: Russell, Armistead G. , Odman, Mehmet Talat
Institution: Georgia Institute of Technology
EPA Project Officer: Chung, Serena
Project Period: December 27, 2004 through December 26, 2007
Project Amount: $444,899
RFA: Source Apportionment of Particulate Matter (2004) RFA Text | Recipients Lists
Research Category: Air Quality and Air Toxics , Particulate Matter , Air
Proposed is a study to capitalize upon recent advances in atmospheric modeling to develop and refine three innovative approaches to source attribution of particulate matter (PM): (1) high-order emission-based source apportionment modeling, (2) an Area-of-Influence method for inverting concentration-emission relationships in chemical transport models (CTMs) to determine the sources impacting each receptor, and (3) a four-dimensional data assimilation (FDDA) method for using CTMs to improve emissions characterization.
Primary objectives of this research are to: 1) extend a recently developed, non-linear source apportionment method for ozone to PM2.5/coarse and apply it to various locations to provide long-term, daily source apportionments (SAs) across the U.S.; 2) further develop the Area-of-Influence (AOI) analysis technique to quantitatively show the impacts of local vs. regionally generated PM by source type and provide information on how future emissions will impact PM in non-attainment areas; 3) refine and apply inverse modeling to improve emissions and source apportionment determinations and identify how to use SA methods when profiles change spatially, seasonally and with controls; and 4) inter-compare results from source-apportionment methods (receptor and source-oriented approaches). Specific hypotheses to be tested in achieving these objectives are that: 1) application of the sensitivity analysis-based techniques proposed here to both shorter term intensive measurements and long-term monitoring will lead to an identification of likely emissions inventory biases that can feed-back to provide more accurate source impacts; 2) emissions-based, second order, non-linear source apportionment using a chemical transport model will capture the spatially- and chemically-resolved impacts and quantify the interaction between sources, particularly after application of inverse modeling; and 3) detailed inter-comparison of emissions-based with receptor source apportionment results over a long-term period will identify the conditions under which the two approaches provide similar results and under what conditions results are likely to differ and why, allowing us to develop better founded assessments of the uncertainties in the various methods.
All three approaches would utilize sensitivity analysis in the CMAQ model using the Decoupled Direct Method in Three-Dimensions (DDM-3D), whose capability we would extend to include the non-linear response of PM to emissions. Modeling will be applied to the Eastern Supersites (ESP01/02) intensives to facilitate rigorous comparison with traditional receptor-oriented methods and over an extended three year period. While a primary focus of the study will be to best exploit the more abundant and detailed data from the Supersites, it is recognized that air quality managers and other stakeholders will seldom have such data at their disposal for conducting source apportionment. We therefore will integrate the more generally available observations from routine Speciation Trends Network (STN) and IMPROVE monitoring to identify issues specific to those monitors and their use for model evaluation and source apportionment.
Confidence in our ability to use both emissions and receptor-oriented air quality models to conduct source apportionment relies on their ability to provide results that are both consistent with each other and with the observations. This is rarely, if ever, achieved in practice, and few studies have ventured to do a comprehensive comparison. In this project, we would integrate both approaches to identify emissions biases, source impacts and estimates of uncertainty. A key result of general utility to other scientists and air quality managers, both, is a method to provide source impacts based upon chemical transport model results, while integrating the observations to get an improved estimate of how emissions actually do impact air quality. This is done, in part, by providing estimates of emissions biases on a source-by-source basis. A further benefit is a method to provide an “Area of Influence” of a source on its surroundings, with quantitative estimates of the source’s impact footprint. Another key benefit is that we will use the results, performance evaluations and intercomparisons to identify when specific models tend to work best. This information can be used by scientists and air quality managers for choosing modeling approaches and assessing how to use air quality model results for guiding policies.