ON THE STATISTICAL-ANALYSIS OF MICROSCALE DISPERSION CALCULATIONS

Microscale models are valuable tools for estimating air pollutant conc entrations in urban areas. If however statistics of pollutant concentr ations (including annual mean, percentile and frequency of odour nuisa nce) are to be derived from model calculations a large number of simul ations for different meteorological situations are necessary. In this article recommendations are given explaining how to the number of simu lations can be reduced without a significant loss of accuracy in the r esulting air quality data. It is also shown that most applications req uire simulations for only four to twelve wind directions and two to th ree wind velocities. Following this recommendations the costs of air q uality studies in urban areas can be significantly reduced.