AAAR 37th Annual Conference October 14 - October 18, 2019 Oregon Convention Center Portland, Oregon, USA
Abstract View
Source Attribution Using Fourier Transform Infrared Spectroscopy in the Interagency Monitoring of Protected Visual Environments (IMPROVE) Network
ANDREW WEAKLEY, Alexandra Boris, Bruno Debus, Satoshi Takahama, Ann Dillner, University of California, Davis
Abstract Number: 752 Working Group: Source Apportionment
Abstract Ambient fine aerosols describe sub-micron condensed phase compounds (<2.5µm particle diameter) derived from primary emission sources (e.g., motor vehicle exhaust) and secondary gas-to-particle formation mechanisms in the atmosphere. Presently, only a fraction of the species comprising ambient fine aerosols are identified and quantified given limited theoretical knowledge of aerosol composition and consequent unavailability of standards. Fourier transform infrared (FT-IR) spectra have successfully probed functional group (FG) composition when applied to filter-bound fine aerosol samples collected from several US monitoring networks. As such, cluster analysis algorithms applied to FT-IR spectra may prove useful in attributing FG absorption to particular sources and/or understanding regio-temporal fluctuations in fine aerosol composition at the functional group level. To this end, a Gaussian Mixture Model (GMM) was fitted to approximately 20,000 Interagency Monitoring of PROtected Visual Environments (IMPROVE; 2015 sampling year) spectra following derivative transformation (to suppress baseline in the spectra), spectra normalization (to emphasize relative FG absorption), and a principal component analysis (to reduce dimensions of input data). Source-cluster assignments were accorded to changes in species-specific FT-IR absorption bands and supported by deviations from concentration normalized median fine carbon (i.e., OC and EC), nitrate and sulfate, and element (e.g., Si, Na) mass concentrations. Most clusters exhibited moderate-to-strong association between region (e.g., eastern US) and season including Midwest samples with elevated inorganic nitrate absorption (winter dominant), southwest samples distinguished by silica IR absorption attributable to African dust storms (spring dominant), eastern US samples showing elevated inorganic sulfate mass (summer), and wildfire samples collected from the northwest (summer). Other clusters, exhibiting less regional or temporal association, included samples impacts by marine aerosol as well as those connected to volcanic activity (AK and HI). Overall, FT-IR spectroscopy appears as a valuable tool for non-destructive aerosol source attribution.