A series of simulations were performed in order to test the accuracy of FEFLO for problems of flow and dispersion around buildings. The major case was extracted from the Evaluation of Modelling Uncertainty (project EMU). The simulations were run in "black box" mode, i.e. no tuning of parameters was allowed. The results obtained were in very close agreement with the available wind-tunnel data for the case selected. The accuracy is comparable with other codes. The results show that an Euler run with proper profile yields a fairly accurate answer.