Comparison of results from field measurements, wind tunnel experiments and numerical urban dispersion models

The emergence of many more, and increasingly sophisticated, mathematical models for environmental applications at the urban scale is paralleled by a growing public and regulatory awareness that many of these models have never undergone a procedure of careful and rigorous evaluation. Nevertheless they are used in making decisions with profound political and economic consequences.

It has been concluded in recent workshops concerning methodologies for model uncertainty analysis ( e.g. Workshop of the European Topic Centre on Air Quality in Rouen/France, Oct. 12, 1999; Workshop of the German Clean Air Commission in Heidelberg, Apr. 27, 1999), that a simple comparison of measurements (taken under continuously varying weather conditions at a specific point) with modelled concentrations (representative of a grid cell and generated with steady-state model inputs) comes close to the proverbial comparison of apples with oranges. This problem is particularly serious for urban-scale dispersion problems which are characterised by large local concentration gradients due to the receptor points being close to the sources.

Results from field measurements taken at an urban monitoring station, from corresponding wind tunnel experiments and from CFD calculations will be compared with each other and the differences discussed. Conclusions with respect to possible strategies for validation work will be drawn. Finally, an overwiev on recent quality assurance initiatives for urban air pollution models taken at European level will be given.