This paper shows results of validation of the Atmospheric Dispersion Modell
ing System (ADMS) against the Indianapolis dataset, comparing arcwise maxim
a and analysing the results using the BOOT statistical package. There is go
od agreement between the quality 3 daytime observed and predicted concentra
tions. At night-time calculated concentrations depend critically on the val
ue of the surface roughness and minimum Monin-Obukhov length.