REAL-TIME PREDICTION OF POINT-SOURCE DISTRIBUTIONS USING AN ANEMOMETER-BIVANE AND A MICROPROCESSOR

An on-line air pollution prediction system has been developed. A single anemometer-bivane transmits wind data to a 64K RAM microprocessor. Turbulent components are filtered off and interpreted as horizontal and vertical diffusivities, while mean components are assumed to represent the entire local wind-field. A variable emission of pollutant is allowed in a defined area of up to 10 km multiplied by 10 km, and its distribution is simulated using a segmented Gaussian plume model. A continuous display shows ground-level concentration contours superimposed on a color map of the area. The picture is updated once per minute. A control box allows 'dumping' of the graphics picture to paper, 'zooming' to a smaller area for better definition, or examination of wind and model data on a text screen. Comparisons with short-range tracer experiments, and SO//2 measurements 5 km from a large source, showed that the system gave satisfactory predictions.