Development and evaluation of a simple dispersion model for an atmospheric point source

Abstract

Atmospheric dispersion modelling is an essential tool for studying the effects of air pollution and is very frequently used to set regulations for the introduction of new sources of pollution in the environment. In this project, a simple Gaussian dispersion model is developed to study the dispersion of a plume from a stack. Three versions of the model are created; in each one a different formulation is used to calculate the horizontal and vertical dispersion coefficients. The model is then evaluated using the Thompson wind tunnel data - a data set collected from experiments carried out in the U.S. Environmental Protection Agency wind tunnel. Validation is made through linear regression and statistical analysis. Results show that the model slightly underestimates pollutant concentration as the plume is modelled to disperse much quicker than it does in reality; however there is a fair linear correlation between the observed data and the predicted values. A relation between height of stack and model performance is also observed. Finally, some modifications to the horizontal dispersion parameters are implemented using curve fitting techniques. Such modifications indeed show an improvement in the results output by the model.