Multivariate receptor modelling to estimate fingerprints of the sources of pollution

We analyse environmental data collected by a number of stations located in Madrid area. The data are given in the form of concentrations of pollutants, which correspond to micro-particles and various chemical species. We use a multivariate receptor model to estimate the fingerprints of the sources and the amounts of pollution. The data collected over time at each station exhibit correlations in a regular way, due to meteorology. Therefore, we model the temporal dependence of the data by using a time series approach. To estimate parameters and uncertainties, we use the Markov Chain Monte Carlo method.