干旱区生态环境知识资源中心

The hourly trails (trajectory points) of incoming air masses have been used in this study in order to compose independent variables for the quantification of regional PM10 and PM2.5 contributions in Marseille (Southern France), and also for the estimation of the atmospheric dispersion effect. These prediction variables were used as input for a Multiple Linear Regression Model (MLRM) in order to estimate daily PM10 and PM2.5 concentrations in Marseille. For a more exact localization of fine and coarse particle sources, during cold and warm period, our analysis was supplemented by the findings of Concentration Weighted Trajectory (CWT) algorithm on a 0.5 degrees.0.5 degrees resolution grid. A strong coherence was revealed among measured and estimated daily levels of PM10 and PM2.5; thus, the proposed MLRM can be a useful tool for assessing air quality in terms of atmospheric circulation. Increased PM contributions in Marseille from local and all-around emission sources were indicated by MLRM primarily within cold seasons. In addition, Northeast (NE) atmospheric circulation was associated by MLRM and CWT with extreme intrusions of exogenous particulate air pollution from Central Europe, during winter and early spring. Throughout warm period, the scarceness of NE airflows prevented the transportation of aerosols from continental Europe. Episodes of desert dust transportation from Northwest Africa (Algeria and Tunisia) had a clear footprint in the PMCOARSE (=PM10-PM2.5) fraction.