Spectral analysis of air pollutants. Part 1: elemental carbon time series

An effective method to analyse different air pollution sources in an elemental carbon time series is presented. As a second feature, this technique allows a fast and efficient classification of monitoring sites. Time series of daily elemental carbon measurements at various urban locations have been evaluated with the corresponding power spectra. Typical and well-known periodicities caused by anthropogenic and meteorological influences have been identified using coherence and phase spectra. It will be shown that domestic heating by coal combustion appears as a 365 day periodicity, traffic contributes 3.5, 4.6 and 7 day peaks in the spectrum and elevated long range elemental carbon can be identified as characteristic peaks with periodicities in the range from 13 to 42 days. As the relative amplitudes of the various influences vary depending on the location of the measurement site in the urban area, the use of estimated power spectra helps to find the influence of traffic, domestic coal-heating and long range transport on the elemental carbon concentration.