Evaluation of Algorithms of Dispersion Coefficients with a Field Tracer Experiment over Complex Terrain

Abstract

The CALMET/CALPUFF modeling system is employed to simulate the dispersion and transport of tracer gas at a nuclear power plant site located in a complex valley and upland terrain where weak wind prevails. Using the surface and upper air meteorological observations obtained from a field experiment, the three-dimensional diagnostic wind fields are generated by CALMET module. Three different algorithms of dispersion coefficients are used to model the ground concentration distributions of tracer under different wind conditions, which are using measured turbulence velocity variances, similarity theory, and PG stability-dependent dispersion curves with coefficients modified according to the in situ turbulence measurement to calculate dispersion coefficients, respectively. The results show that turbulence and modified PG methods can better predict high observed concentrations than the similarity method, while all the three methods overpredict the low observed concentrations mainly due to underestimations on wind speed and overestimations on mixing layer heights in modeling. The turbulence and modified PG methods overestimate the observed peak concentrations by less than 30%, while the similarity method underestimates by about 20%. Overall, the turbulence and modified PG methods perform better than the similarity method, with less dependence of simulated concentration residues on wind speed and mixing height. From the viewpoint of engineering application, CALPUFF model with modified PG method to calculate dispersion coefficients is recommended at the site with hilly-valley complex terrain to simulate the transport and dispersion of gaseous effluent.