The dynamic coregionalization model in air quality risk assessment Francesco Finazzi University of Bergamo, Dept. of Information Technology and Mathematical Methods viale Marconi, 5 Dalmine 24044, Italy E-mail: francesco.nazzi@unibg.it Alessandro Fass University of Bergamo, Dept. of Information Technology and Mathematical Methods viale Marconi, 5 Dalmine 24044, Italy E-mail: alessandro.fasso@unibg.it Marian E. Scott University of Glasgow, School of Mathematics and Statistics 15 University Gardens Glasgow G12 8QW, Scotland E-mail: Marian.Scott@glasgow.ac.uk 1. Introduction Air pollution monitoring and mapping at country level are challenging tasks due to the large spatial scale and the amount of data involved. Nevertheless, they can be carried out successfully by considering the proper space-time statistical models and mapping techniques, which provide estimates of both the pollutant concentration and the respective uncertainty (Fass and Cameletti, 2010). Decision makers are called to take actions on the basis of the air quality assessment results in order to reduce the impact of air pollution on population health (Scott, 2007). Although uncertainty is crucial to analyze the results, it is not easily interpretable or transmutable into actions on its own, especially when the impact on population health must be evaluated. The aim of this paper is to provide a statistical framework for the country level population risk assessment connected with air pollution. In particular, the risk is evaluated by estimating the exceedance probability of pollutant concentration thresholds related to a set of airborne pollutants. The exceedance probability is mapped over space and time and, when convolved with the spatial population count distribution, it allows to derive aggregate risk indicator. The rest of the paper is organized as follows. Section 2 introduces the Dynamic Coregionalization Model which is used to map pollutant concentrations within a multivariate setting and to evaluate uncertainty. Section 3 describes the population risk assessment procedure applied in Section 4 to the Scottish air quality data for the year 2009. Conclusions and future works are reported in Section 5. 2. The dynamic coregionalization model The Dynamic Coregionalization Model (DCM) is a hierarchical multivariate space-time model introduced by Fass and Finazzi (2011). Let y i (s;t) be the concentration of the i  th pollutant at the spatial location s 2D R 2 and at time t 2 N + , 1  i  q. The model equation is the following: Int. Statistical Inst.: Proc. 58th World Statistical Congress, 2011, Dublin (Session CPS029) p.4537