Evaluation of two dispersion models (ADMS-Roads and LASAT) applied to street canyons in Stockholm, London and Berlin

Data sets from street canyons in Stockholm, Berlin and London are used to investigate the performance of the Gaussian model ADMS-Roads1 and the Lagrangian model LASAT2 and the impact of different meteorological input, flow schemes and NO–NO2NO2 conversion schemes. ADMS-Roads tends to under-predict NOxNOx concentrations. Nevertheless, the more complex Lagrangian model is not always in better agreement with the observations. LASAT with a simpler flow scheme leads to higher concentrations of NOxNOx than the more complex DMK scheme in Stockholm and Berlin. Model runs based on time-series of stability classes render in several cases better results than those with constant neutral conditions. Two NO–NO2NO2 conversion schemes are used. The scheme after Romberg performs better for LASAT in most of the cases while the Derwent approach is more successful for ADMS-Roads. Although point to point correlations between model results and measurements are generally weak, both dispersion models are quite successful in predicting average values, maximum values as well as the occurrence of exceedances of threshold limits within the street canyons.