A Comparison of Six Highway Air Pollution Dispersion Models Using Synthetic Data.

摘要

This is the second of two studies conducted by the Transportation Systems Center (TSC) to test the performance of highway air pollution dispersion models,using synthetic data (i.e.,either measured or artificially constructed input data for models,consisting of traffic and meteorological parameters). In the first study,thirteen models were tested with a portion of the Airedale air quality data base. In the present study,six models (including five of the original thirteen) were tested with a new synthetic data base. The air pollution predictions of the six models were compared in pairs,and various measures of the difference between the predictions of each pair were calculated. A group of three models generating very similar predictions was discovered;these are called Consensus Models. In the first study,using the same analytical approach,these three models were also found to be Consensus Models,along with others which could not be tested in the present study. Synthetic data testing can only reveal the degree of agreement among model predictions. The capability of these models to predict real-world air pollution cannot be determined until an air-quality data base suitable for model validation become available.