A GIS based approach to back trajectory analysis for the source apportionment of aerosol constituents and its first application

A new method has been developed to combine back trajectory statistics with a detailed land cover analysis. It provides numeric proxies for the residence times of sampled air masses above certain land cover classes (marine, natural vegetation, agricultural lands, urban areas, and bare areas), as well as further meteorological parameters (mean trajectory length, solar radiation along trajectory, and local height of the boundary mixing layer). The method has been implemented into a GIS-enabled database system to allow for an efficient processing of large datasets with low computational demands. A principal component analysis was performed on a dataset including the modelled residence times, the modelled meteorological parameters, some measured meteorological parameters (wind speed and temperature), and the concentrations of 10 particle constituents (inorganic ions and organic and elemental carbon) in 5 particle size ranges for 29 winter-and summertime samples at an urban background site in Leipzig, Germany. Six principal components could be extracted which together explained about 80% of the total variance in the dataset. The factors could be attributed to the influence of meteorology to continental background pollution, secondary formation processes in polluted air masses, wood burning, aged sea-salt, local traffic, and long-range transported crustal material. The modelled residence times and the meteorological parameters were generally consistent with the existing knowledge of specific particle sources and thereby facilitated and strengthened the interpretation of the factors. Moreover, they allowed for a clear distinction between continental background pollution and secondary formation processes, which has not been possible in previous source apportionment studies. The results demonstrate that the combined usage of back trajectory, land cover, and meteorological data by the presented method yields valuable additional information on the history of sampled air masses, which can improve the quality of source apportionment of atmospheric aerosol constituents.