You are here:
RECENT APPLICATIONS OF SOURCE APPORTIONMENT METHODS AND RELATED NEEDS
Citation:
Mukerjee, S. RECENT APPLICATIONS OF SOURCE APPORTIONMENT METHODS AND RELATED NEEDS. Presented at Environment Seminar Series, St. Louis, MO, October 4, 1992.
Impact/Purpose:
Overall Goal: To develop spatial analyses using limited network-based air quality and GIS and other ancillary spatial information to estimate exposures for epidemiologic studies.
Goal of NERL Contribution: To develop regression-based spatial models using said measures and ancillary information to predict such exposures at unmonitored locations.
Specific Objectives:
1. To determine whether ultrafine (<0.1 um), accumulation (0.1-0.7), and/or coarse (1-10 um) mode particle counts correlate with CO, NO2 and VOCs emitted from mobile and/or other urban sources using source apportionment modeling techniques.
2. To determine spatial associations among measured levels of NO2, VOCs, and (possibly) ultrafine/accumulation/coarse mode particle counts from mobile and other urban sources in El Paso. Spatial variability in ultrafine/accumulation/coarse mode particle concentrations will be determined using available PM, NO2, VOC and available surrogates of motor vehicle emissions. These measured or predicted spatial associations will then be used by NHEERL to ultimately assess impact of these particle counts and gaseous species on children's exposures in schools.
3. To evaluate accuracy of NO2 and VOC measurements using the passive badges to be deployed by EPA versus collocated FRM devices established by the State of Texas. In addition, to evaluate precision of collocated NO2 and VOC passive badge measurements.
4. To use spatial analysis concepts to evaluate their possible application in an EPA Region 6 study entitled "Air Toxics Data and Analysis and Development of a Predictive Model of Estimation of Ambient Vocs in Selected Census Tracts in Houston-Galveston, TX."
Description:
Traditional receptor modeling studies have utilized factor analysis (like principal component analysis, PCA) and/or Chemical Mass Balance (CMB) to assess source influences. The limitations with these approaches is that PCA is qualitative and CMB requires the input of source profiles; both the interpretation of factors and which profiles to use in CMB can be fairly subjective. Ambient-based, multivariate receptor models, such as the EPA Unmix receptor model and Positive Matrix Factorization (PMF), have an advantage over profile-based models in that they not only estimate source contributions like CMB but also generate source profiles for source identification.
In this presentation, a comparison of CMB and Unmix source contribution estimates will be presented using continuous measurements of volatile organic compounds (VOC) from an auto-GC. The VOC source estimates were then used to assess relationships to ultrafine particle number concentration data collected by a scanning mobility particle sizer and aerodynamic particle sizer at the same location.
Recent EPA PM Panel Studies have used ambient-based receptor models like Unmix and PMF to evaluate indoor and outdoor influences based on ambient, indoor, and personal exposure measurements. While this area shows promise in identifying potential exposure factors, research needs also exist, particularly with regards to measurement methods development.
The information in this document has been funded wholly or in part by the United States Environmental Protection Agency. It has been subjected to agency review and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.