Aerosol and gaseous pollution measurements were carried out at an urban background site in the south of Italy located near an industrial complex. Collection of 24 h samples of PM10 and PM2.5 and successive chemical quantification of metals were performed. Data were compared with measurements taken at a suburban background site, located at 25 km distance. The comparison showed the presence of an industrial contribution with a well defined chemical emission profile, similar, in terms of metals content, to urban emissions. As this made difficult the quantitative characterisation of the contribution of the two sources to atmospheric PM, a statistical method based on the treatment of data arising from high temporal resolution measurements was developed. Data were taken with a micrometeorological station based on an integrating nephelometer (Mie pDR-1200) for optical detection of PM2.5 concentration, with successive evaluation of vertical turbulent fluxes using the eddy-correlation method. Results show that the contribution from the two sources (urban emissions and industrial releases) have a very different behaviour, with the industrial contribution being present at high wind velocity with short concentration peaks (average duration 4 min) associated to strong positive and negative vertical fluxes. The estimated contribution to PM2.5 is 2.3% over long-term averages. The urban emissions are mainly present at low wind velocity, with longer concentration peaks in the morning and late evening hours, generally associated to small positive vertical fluxes. The characterisation of the contribution was performed using deposition velocity V(d) that is on average -3.5 mm s(-1) and has a diurnal pattern, with negligible values during the night and a minimum value of around -9 mm s(-1) late in the afternoon. Results show a correlation between V(d), friction velocity and wind velocity that could be the basis for a parameterisation of V(d) to be used in dispersion codes.