Performance of a Gaussian model for centerline concentrations in the wake of buildings

This paper examines the performance and inherent limitations of a Gaussian plume model for predicting ground-level plume centerline concentrations in the wake of buildings. The Gaussian plume equation has been modified to incorporate building wake enhanced dispersion parameters. Model-predicted concentrations were compared to three sets of field observations. Predicted and observed concentrations were partitioned into groups by source-receptor distance, atmospheric stability class, and source-release height. The group analyses provided a way to identify sources of model error. The variability of errors was found to have values between 50 and 100 % of the mean observed concentration for groups where the mean error was small.