|
|
|
|
|
|
|
 |
Surface Ocean and Lower Atmosphere Study¡ªAir-Sea interactions and their climatic and environmental impacts
|
|
|
Identification of the chemically distinct groups of Atlantic aerosol particles based on K-means cluster analysis Wednesday 9th @ 1030-1050, Conference Room 7 Shan Huang* , Jinan University Laurent Poulain, Leibniz Institute for Tropospheric Research Zhijun Wu, Peking University Wolfram Birmili, Federal Environment Agency Hartmut Herrmann, Leibniz Institute for Tropospheric Research Alfred Wiedensohler, Leibniz Institute for Tropospheric Research Presenter Email: shanhuang_pku@163.com |
The marine aerosol plays an important role in global climate regulation and marine biogenic system. Aerosol particles in the marine boundary layer (MBL) are seldom from exclusively marine origins. Complex sources and atmospheric processes determine the chemical and physical properties of aerosol particles. In this study, chemically distinct groups are identified by k-means clustering, in an exploratory way, on chemical composition (provided by a High Resolution Time of Flight Aerosol Mass Spectrometer, HR-ToF-AMS) and particle number concentration (PNC) of MBL aerosol particles, and linked to the air mass origins. The analysis is based on physical and chemical measurements performed on board the German research vessel Polarstern during 4 research cruises over the Atlantic Ocean in 2011 and 2012, covering spatial range from 53˚S to 53˚N in two seasons (spring and autumn). In total 12 variables were used for clustering, including PNC of aerosol particles and 11 chemical components (ammonium, nitrate, chloride and BC directly provided by instruments, plus 5 organic components and 2 sulfate components with identified sources given by source apportionment models). With this algorithm, 9-cluster solution was selected as the best result and further combined to 4 groups with distinct chemical composition including 2 marine groups and 2 continental ones. The spatial distribution of these 4 chemically distinct groups can well correspond to the air mass origins. The resulting clusters are repeatable in the same season along the similar ship tracks, which makes the method useful for modelization on predicting physicochemical properties of MBL aerosol. |
|
|
|
|
|
|
|
|