Investigating Biogenic and Anthropogenic Aerosol Interaction at the Molecular Level – Insights from the Bankhead National Forest and Southern Great Plains

GREGORY W. VANDERGRIFT, Nathaniel Breaux, Darielle Dexheimer, Ashfiqur Rahman, Sijia Liu, Zezhen Cheng, Nurun Nahar Lata, Meghan C. Guagenti, Damao Zhang, Lindsay Yee, Allen Goldstein, Jason Surratt, Maria Zawadowicz, Chongai Kuang, Swarup China, Pacific Northwest National Laboratory

     Abstract Number: 296
     Working Group: Source Apportionment

Abstract
Ambient secondary organic aerosol (SOA) is the result of complex interactions between emissions/products of varied sources, namely anthropogenic and biogenic. SOA source is often inferred by examining a limited number of tracer species. While often effective, such approaches do not maximize the amount of compositional information that could be used to assess SOA source and mechanism, as hundreds to thousands of different individual molecular formulas may routinely be assigned per sample. Increased understanding of SOA may therefore be gained by leveraging untargeted analysis of the organic aerosol composition at the molecular level. Here, we first develop a machine learning model to assess sample source based upon untargeted molecular-level analyses of particle-phase samples collected from the Southern Great Plains (SGP). We leveraged the results from community chamber experiments for model training. Via correlations with known molecular tracers, the model then was used to identify novel molecular tracers in the untargeted datasets for various anthropogenic and biogenic sources. Related to the work conducted at SGP, untargeted analysis of the organic molecular composition of vertically resolved Bankhead National Forest (BNF) SOA is currently being conducted, with applied machine learning for more comprehensive evaluation of source interaction.