An effective simple site-optimised urban air pollution model for predicting CO, PM 10 , NO and NO 2 concentrations is developed for the topographically complex region of the Aosta Valley in Italy. The good results suggest that such a model could be used to 'downscale' mesoscale-forecasted surface conditions to give real-time site-specific pollution predictions.

This paper investigates the use of the Site-Optimized Semiempirical (SOSE) air pollution model to identify the surface wind measurement site characteristics that yield the best air pollution predictions for urban locations. It compares the modelling results from twelve meteorological sites with varying anemometer heights, located at different distances from the air pollution measurements and exhibiting different land use characteristics. The results show that the index of agreement (IA) between observed and predicted concentrations can be improved from 0.4 to 0.8 by using the most compared to the least representative wind data as input to the air pollution model. Although improvements can be achieved using wind data from a site closer to the air quality monitoring site, choosing the closest wind site does not necessarily yield the best results, especially if the meteorological station is located in a region of complex land use. In addition, both the height of the anemometer and the openness of the terrain surrounding the anemometer were found to be equally important in obtaining good model predictions. The simple SOSE model can therefore be used to complement regulatory meteorological guidelines by providing a quantitative assessment of wind site representativeness for air quality applications in complex urban environments.