Application of Forward and Backward Atmospheric Dispersion Models using Measurements

A Lagrangian atmospheric dispersion model has been developed to predict the behavior of pollutants released into the air from the unexpected accident of the industrial or nuclear power plants. The random walk method in Lagrangian model is adopted in the dispersion model for the estimation of the atmospheric concentration distribution of the released pollutants. The basic advantages of that method are the simplicity, flexibility and the ability to produce relatively accurate results. In three-dimensional space, a particle transport due to the advection and the turbulent diffusion. And the movement of the particle is represented by the sum of the movements due to the advection and the turbulence. In the model, the atmospheric dispersion is evaluated by the motion of fictitious particles consisting of a deterministic part due to the mean wind and a stochastic part related to the turbulent flow. Forward and backward atmospheric models based on Lagrangian approach were applied to estimate unknown source regions and release rates of pollutants released into the air from unexpected accidents. Simulated results were compared with the measurements of a field tracer experiment performed at the Yeonggwang nuclear power plant in Korea in May 1996. The release point was first determined by using the backward dispersion model, and a unit release approach was used to estimate the release rates of the tracer at the release point. Calculated forward dispersion patterns are well presented the transport patterns by westerly wind. The time-varying concentrations were also simulated at the sampling points. Although, measurements and simulations for time-varying concentrations generally agreed, some discrepancy appeared due to the insufficient measurements of wind data during the experiment.