Keywords: atmospheric dispersion models, field tracer datasets, model evaluation
Evaluation of the ADMS, AERMOD, and ISC3 dispersion models with the OPTEX, Duke Forest, Kincaid, Indianapolis and Lovett field datasets
The model evaluation exercise addresses the question of whether the new models, ADMS and AERMOD, produce improvements over ISC3 when compared with a range of field observations. ADMS and AERMOD have similar state-of-the-art scientific components, whereas ISC3 contains 1960s technology. The five sets of field observations used in the statistical evaluation represent a cross-section of typical scenarios encountered by modellers. The OPTEX database deals with non-buoyant tracer releases within an oil refinery complex, and the Duke Forest database involves non-buoyant tracer releases from area and volume sources in an open field. The Kincaid, Indianapolis and Lovett databases all deal with buoyant plumes from tall stacks at power plants. However, the settings for each are quite different, since the Kincaid plant is surrounded by flat farmland, the Indianapolis plant is located in an urban environment, and the Lovett plant is situated in a valley surrounded by complex terrain with monitors at higher elevations than the stack. Analysis of the model performance measures suggests that ISC3 typically overpredicts and has a scatter of about a factor of three. Approximately 33% of its predictions are within a factor of two of observations. The ADMS performance is slightly better than the AERMOD performance and both perform better than ISC3. On average, ADMS underpredicts by about 20% and AERMOD underpredicts by about 40%, and both have a scatter of about a factor of two. Approximately 53% and 46% of the ADMS and AERMOD predictions, respectively, are within a factor of two of observations. Considering only the highest predicted and observed concentrations, ISC3 overpredicts by a factor of seven, on average, while ADMS and AERMOD underpredict by, on average, 20%.