Identification and characterisation of sources of PM10 in Madrid (Spain) by statistical methods

A source apportionment study was performed to characterise PM10 sources in the urban area of Madrid. Receptor modelling techniques were applied to obtain the different source contributions, identifying four main sources that contribute to most of the analysed PM10 mass, namely: road traffic, crustal/mineral, secondary and marine aerosol. PM10 mean concentration and chemical profile were obtained for each source category by multilinear regression analysis. Road traffic was found to be the major contributor to the PM10 mass (48%), together with crustal (26%) and secondary particles (18%). A minor contribution has been associated to marine aerosol (3%), while 5% of the PM10 mass has been attributed to non-identified sources. An analysis of the geographical origin was performed for remote sources. Residence time probability values were calculated based on air mass back trajectories information. Different elements have been selected as tracers of the sources. Desertic regions from North Africa have been identified as the most significant source area for crustal particles, with a high content of Al and Ti. Atlantic ocean regions have been identified as weak to intermediate source areas of Na, supporting the marine source identified by the factor analysis. Then the influence of remote natural sources on the PM10 levels in central Spain has been detected by the use of single components as tracers, even at urban sites highly affected by traffic. (C) 2003 Elsevier Ltd. All rights reserved.