AAAR 32nd Annual Conference
September 30 - October 4, 2013
Oregon Convention Center
Portland, Oregon, USA
Abstract View
Retrospective Source Attribution for Source-Oriented Sampling and Toxicity
KEITH BEIN, Yongjing Zhao, Anthony Wexler, UC Davis
Abstract Number: 586 Working Group: Source Apportionment
Abstract In previous work, a novel system that uses single particle mass spectrometry to conditionally sample source-oriented particulate matter (PM) from the ambient atmosphere in real-time was designed and successfully implemented in two separate field studies (summer 2008 and winter 2009) conducted in Fresno, CA. The fundamental concept behind this technique is that single particle composition is a metric of particle source and thus sampling particles based on composition should be synonymous with sampling based on source. System operation relies on real-time pattern recognition in mixtures of single particle source signatures to control the actuation of different ChemVol samplers in a bank of samplers, where each ChemVol is associated with a unique composition signature, or combination of signatures. In the current work, a synthesis of data collected during these studies – including single particle composition, particle number distribution and wind direction – is used in retrospect to reconcile the actual source combinations contributing to the particles collected by each ChemVol. Source attribution is based on correlations between ChemVol sampling and wind direction and temporal variations in the actuation of different ChemVols, coupled to knowledge of single particle composition and the geospatial distribution and activity patterns of sources surrounding the site. Residential and commercial cooking, vehicular emissions, residential heating and highly processed regional background PM were identified as the major sources impacting the site. Different sources were observed to dominate in different seasons and the composition signatures of similar sources observed in different seasons were different. Results show that real-time patterns in single particle mixing state correctly identified specific sources and that these sources were successfully separated into different ChemVols for both seasons.