Apportionment of Air Pollution by Source at a French Urban Site

Abstract

The development of air quality control strategies is a wide preoccupation for human health. In order to achieve thispurpose, air pollution sources have to be accurately identified and quantified. This case study is part of a scientificproject initiated by the French ministry of Ecology and Sustainable Development. Measurements of chemicalcomposition data for particles have been performed at a French urban site. The work presented in this paper splits intotwo main steps. In the first one, the identification of the source profiles is achieved by a Principal Component basedFactor Analysis (FA), followed by a rotation technique. Then, in the second step, a receptor modeling approach (usingPositive Matrix Factorization as an estimation method) allows us to evaluate the apportionment of particles by source.The results from these two statistical methods have enabled us to characterize and apportion fine particulate matteremissions by five sources. The exposition is accessible to readers with an intermediate knowledge of statistics; anexposure to factor and principal components analyses is useful but not strictly necessary.