Keywords: atmospheric dispersion models, Bull Run experiment, model validation, model performance, model evaluation, Indianapolis experiment, model catalogue, air pollution, modelling, environmental pollution
Pilot study: extension of the Model Validation Kit
The so-called Model Validation Kit has been used as a basis for modelling exercises at workshops in Manno (1993), Mol (1994) and at the present workshop. It is a practical tool, meant to serve as a common frame of reference for modellers. It consists of three experimental datasets, accompanied by software for model evaluation. Before the present workshop, a pilot study was conducted, aimed at extending the original toolkit with datasets from two additional experimental campaigns. These 'new' datasets contain data from experimental campaigns at Indianapolis (urban) and Bull Run (moderately complex terrain). Both campaigns deal with emissions from a single stack, and include detailed tracer measurements from several hundred hours. The present paper gives an introduction to the two 'new' datasets. A large part of the paper, however, is used to convey general experiences on the preparation and use of performance-evaluation datasets. The context of this work is discussed; particular emphasis is on its relation to a planned catalogue of models, which will become available through the World Wide Web.