Multiresolution decomposition and wavelet analysis of urban aerosol fluxes in Italy and Austria

Abstract

Observations of turbulent aerosol fluxes are fundamental to understand basic transport processes that govern changes of particle concentrations in the atmospheric boundary layer. The turbulent surface-atmosphere exchange of atmospheric particles can be quantified using several methods, including the eddy-covariance (EC) method and spectral flux estimation methods such as wavelet analysis and multiresolution decomposition. In this work, turbulent time series obtained by EC measurements in two different cities, Lecce (Italy) and Innsbruck (Austria), are spectrally analysed applying wavelet analysis and multiresolution decomposition, and the respective turbulent spectra are compared in these two European cities to quantify the contributions to turbulent fluxes in both the time and the frequency domains. As expected, particle emission is dominant in both cities following a similar diurnal cycle. Multiresolution decomposition reveals a similar cospectral peak of particle fluxes in both cities, with a median normalized frequency of n = 0.087 in Lecce and n = 0.086 in Innsbruck. Wavelet analysis shows that the 2–20 s time scales contribute very strongly to the particle flux in Lecce, while in Innsbruck the 20–200 s time scales are clearly dominant. In both cities, larger-sized eddies contribute only sporadically to turbulent aerosol fluxes. These results suggest that spectral similarity of urban particle number fluxes holds, to a large extent, even when comparing two very different urban environments and different meteorological conditions.