AAAR 33rd Annual Conference
October 20 - October 24, 2014
Rosen Shingle Creek
Orlando, Florida, USA
Abstract View
Positive Matrix Factorization Analysis of 47-years of Finnish Arctic Aerosol Composition
JAMES R. LAING, Philip K. Hopke, Eleanor F. Hopke, Liaquat Husain, Vincent A. Dutkiewicz, Jussi Paatero, Yro Viisinen, Clarkson University
Abstract Number: 260 Working Group: Source Apportionment
Abstract Week-long total suspended particle samples collected between 1964 and 2010 from Kevo, Finland were analyzed for trace metals, soluble trace metals, major ions and MSA, and black carbon (BC). Long-term datasets of aerosol chemical composition are very useful key for understanding and explaining past events. Significant changes have occurred in the Arctic in the past half century. In order to better understand these changes the 47-year complete data set was analyzed by Positive Matrix Factorization (EPA PMF5). The dataset was split into three time periods, 1964-1978, 1979-1990, and 1991-2010. Ten factors were determined for the 1964-1978 period, and 9 factors each resolved for the latter periods. All three time periods have two factors representing Cu-Ni-Co smelters, one dominated by Cu, the other by Ni and Co; a factor primarily consisting of Mo and W most likely due to mining, processing, and transportation of alkaline rocks on the Kola Peninsula; a factor of V, BC, and nss-SO4 characterizing residual oil combustion; a biogenic sulfate factor represented by methane sulfonic acid (MSA), a factor dominated by water soluble Fe and some Mn from the nearby iron mines and mills, and a factor consisting primarily of Sn. Similar factors containing the majority of As and some Re and Se were determined for the 1979-1990 and 1991-2010 time periods, while the 1964-1978 run contains additional inputs of Sb, Mo, and Tl. A Mn-Cd factor was found during the 1979-1990 and 1991-2010 period but not for the 1964-1978 time period. Unique factors for Ag, Au, and sea-salt (Na and Mg) were determined for the 1964-1978 time period. The combination of seasonal trend analysis and Potential Source Contribution Function (PSCF) analysis applied to the PMF results will help identify source profiles and their sources locations.