AAAR 36th Annual Conference October 16 - October 20, 2017 Raleigh Convention Center Raleigh, North Carolina, USA
Abstract View
Classification, Variation and Spatial Patterns of Mass Spectra Extracted from Plume Events Observed from Mobile Measurement by Aerodyne Aerosol Mass Spectrometer and Comparison with PMF Results
PEISHI GU, Zhongju Li, Qing Ye, Ellis Shipley Robinson, Allen Robinson, Albert A. Presto, Carnegie Mellon University
Abstract Number: 307 Working Group: Source Apportionment
Abstract Highly time-resolved field measurement of organic aerosol (OA) can be strongly influenced by short-duration plume events. Such plume events may contribute significantly to OA exposure in particular places where abundant human activities of different kinds coexists. However the origin of such OA plume events is not always clear and improved source attribution of these plumes could be used to develop policies to reduce human exposures. Aerodyne Aerosol Mass Spectrometer (AMS) data coupled with Positive Matrix Factorization (PMF) has been widely used for source apportionment of field measurement data. Typically, primary emission factors like hydrocarbon-like OA (HOA), cooking OA (COA) and biomass burning OA (BBOA) have been identified. PMF results represent an average factor composition, but may be insensitive to plume-to-plume variation within each source category. Some minor sources might also get ignored by the mathematical model due to their limited appearance in the typically huge dataset. In this study we explore the contribution of plume events to mobile sampling data collected from Aug 2016 to June 2017 in Pittsburgh, PA. The plumes are identified as peaks in the time series of OA concentration, while the average mass spectrum of a plume is extracted by calculating the difference of the mass spectra of the peak period and that of the immediate background before and after the plume events. Hundreds of well-defined plume events were identified and extracted, and then clustered into interpretable categories. Similar to results from PMF, HOA and COA are two major factors, but variations within each plume cluster are clearly observable. Such variation may partly result from the oxidation of fresh OA, or simply the inherent variability of the emissions. There are also plumes that resemble a mixture of multiple factors. Geospatial analysis reveals the spatial pattern of different types of OA plumes (including the minor sources), and how their oxidation level varies across the sampling domain.