A numerical model for predicting the diffusion of sulfur dioxide in the atmosphere

A numerical model is proposed for predicting the dispersion of SOâ‚‚ over Nashville, Tennessee, during the winter. The model includes the effects of three-dimensional advection, of diffusion, and of variations in the topography. Also included in the numerical solution are the effects of chemical decay, of source strength, and of a correction term for the truncation error. A variable, vertical exchange coefficient is incorporated in the diffusive portion of the model. These terms represent the complete transport equation which is numerically integrated for a period of 2 hr in three-dimensional space. Initial conditions are based on aerometric and meteorological data acquired in Nashville, Tennessee, during the period 1958-1959. These data are used to establish the three-dimensional distribution of SOâ‚‚ and to derive a steady-state condition for the air flow and for the internal emissions of SOâ‚‚. A neutrally stable atmospheric condition is specified for a depth of 60 m. Seven different variations in the model are tested. In all cases, the predicted fields of SOâ‚‚ are aligned parallel to the air flow. One case illustrates the important role of vertical diffusion in dispersing SOâ‚‚. Another case shows that vertical advection can change local concentrations of SOâ‚‚ by as much as 20 percent of the predicted value. In 2 hr, chemical decay accounts for losses of the predicted value of SOâ‚‚ of 10 to 20 percent. It is shown that the correction term dampens out the small scale oscillations of concentration which result from truncation. Predicted concentrations of SOâ‚‚ are realistic and compare favorably with observed values. This study clearly demonstrates that numerical modeling of the diffusion of atmospheric pollutants is feasible, is a potentially fertile domain, and should be pursued in depth. Furthermore, the model offers a high degree of flexibility by permitting frequent changes to be made in the wind direction and speed, in the source strengths, and in the atmospheric stability. The need for carefully planned and coordinated aerometric and meteorological measurements is stressed. Such observations are mandatory if physically realistic models of atmospheric diffusion are to be developed and tested. Simultaneous sampling of atmospheric pollutants and meteorological conditions is important; this is especially true for the vertical dimension