Comparison of Optical Properties Calculated from a Dual-Spot Aethalometer and a Photoacoustic Spectrometer for Biomass Burning Aerosols during the GWISE Campaign

RYAN POLAND, Zachary McQueen, Omar El Hajj, Chase Glenn, Anita Anosike, Kruthika Kumar, Joseph O'Brien, Rawad Saleh, Geoffrey Smith, University of Georgia

     Abstract Number: 204
     Working Group: Instrumentation and Methods

Abstract
The Georgia Wildfire Simulation Experiment (GWISE) was designed to study the chemical and optical properties of particulates resulting from wild and prescribed fire smoke, which often contains large portions of highly absorbing “black carbon” (BC) and “brown carbon” (BrC). The controlled nature of these simulated wildfires allows for comparison of the optical properties of biomass burning aerosols measured by the commercially available seven-wavelength dual-spot AE33 aethalometer (Aerosol Magee Scientific) and a custom, four-wavelength photoacoustic spectrometer (MultiPAS-IV) (Fischer and Smith 2018). Under both wild and prescribed fire conditions, we show that the AE33 over-estimates absorption by a factor of ~ 2 at all wavelengths. We also find that over-estimation of absorbance at 370 nm and 470 nm increases with increasing filter attenuation (ATN) above ATN values of 60. This observation suggests that the dual-spot loading compensation algorithm (Drinovec et al., 2015) is not able to account for loading effects quantitatively when aerosol optical properties change significantly at high filter attenuation values. Consequently, we recommend using the dual-spot aethalometer with an attenuation threshold for filter spot advancement of ATN = 60, reduced from the default value of 120. We also find that the absorption Ångström exponent (ÅAE), a measure of the shape of the absorption spectrum, calculated from the AE33 data agrees well (within 7%) with values calculated from the MultiPAS-IV data. However, ÅAE’s calculated in the ultraviolet/blue (dominated by BrC) and red/near-infrared (dominated by BC) regions from the AE33 data are 23 and 32% high respectively, when compared to those calculated from the MultiPAS-IV data. The systematic errors observed for these ÅAEs could lead to over-prediction of biomass burning percentage commonly used for source apportionment (Sandradewi et al. 2008).