UI - Concentration of PM10 (LUISA Platform REF2014)

Land Use Regression mode (LUR), were used to simulate PM10 concentrations at very high resolution (100m) built using pollutant concentrations for 2010 from monitoring sites as dependent variable, and several parameters (independent variables) defined within a Geographic Information System (GIS). The LUR model was developed using Random Forest regression techniques (Breiman, 2001) and was used to predict evolution of concentrations of pollutants from 2010 to 2050, according to predicted changes in land use and population density data taken from the LUISA outputs (Baranzelli et al, 2015). The modelling did not consider for the prediction of future concentrations specific measures or policies implemented in order to reduce emissions.