The Use of Positive Matrix Factorization with Conditional Probability Functions in Air Quality Studies: An Application to Hydrocarbon Emissions in Houston, Texas

Journal Article · Thu Jun 01 00:00:00 EDT 2006 · Atmospheric Environment, 40(17):3070-3091

DOI:https://doi.org/10.1016/j.atmosenv.2005.12.065· OSTI ID:883674

Xie, YuLong; Berkowitz, Carl M

As part of a study to identify groups of compounds (‘source categories’) associated with different processing facilities, a multivariate receptor model called Positive Matrix Factorization (PMF) was applied to hourly average concentrations of volatile organic compounds (VOCs) measured at five Photochemical Assessment Monitoring Stations (PAMS) located near the Ship Channel in Houston, Texas. The observations were made between June and October, 2003, and limited to nighttime measurements (21:00 pm – 6:00 am) in order to remove the complexity of photochemical processing and associated changes in the concentrations of primary and secondary VOCs. Six to eight volatile organic compounds source categories were identified for the five Ship Channel sites. The dominant source categories were found to be those associated with petrochemical, chemical industries and fuel evaporation. In contrast, source categories associated with on-road vehicles were found to be relatively insignificant. Although evidence of biogenic emissions was found at almost all the sites, this broad category was significant only at the Wallisville site, which was also the site furthest away from the Ship Channels area and closest to the northeast forest of Texas. Natural gas, accumulation and fuel evaporation sources were found to contribute most to the ambient VOCs, followed by the petrochemical emission of highly reactive ethene and propylene. Solvent / paint industry and fuel evaporation and emission from refineries were next in importance while the on-road vehicle exhaust generally contributed less than 10% of the total ambient VOCs. Specific geographic areas associated with each source category were identified through the use of a Conditional Probability Function (CPF) analysis that related elevated concentrations of key VOCs in each category to a network of grids superimposed on the source inventories of the VOCs.

• Research Organization: Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
• Sponsoring Organization: USDOE
• DOE Contract Number: AC05-76RL01830
• OSTI ID: 883674
• Report Number(s): PNNL-SA-46700; ATENBP; 600303000; TRN: US200614%%888
• Journal Information: Atmospheric Environment, 40(17):3070-3091, Vol. 40, Issue 17; ISSN 0004-6981
• Country of Publication: United States
• Language: English

Related Subjects

02 PETROLEUM
03 NATURAL GAS
AIR QUALITY
EVAPORATION
FACTORIZATION
FORESTS
HYDROCARBONS
INVENTORIES
MONITORING
NATURAL GAS
ORGANIC COMPOUNDS
PETROCHEMICALS
PROBABILITY
PROCESSING
PROPYLENE
receptor modeling
VOC
positive matrix factorization (PMF)
conditional probability function (CPF)
Houston
air quality