An analysis and simulation of plume dispersion for time periods of 1-48 hours

All man-made emissions go through the plume stage on the mesoscale. The concentrations of primary pollutants depend directly on plume dispersion. Plume dispersion also has a critical influence on plume chemistry and the formation of secondary pollutants, which figure prominently in terms of harmful effects of air pollution. Thus, plume dispersion is of primary importance in the assessment of distribution of air pollution.;Spectral analysis of the radar boundary-layer profiler (915 MHz) wind data collected during June-July 1995 as part of the Southern Oxidants Study field program at Nashville, TN, has been performed to understand the missing part of the atmospheric energy spectrum that is not resolved or parameterized in the current grid models. The boundary-layer model simulations were performed with and without a four dimensional data assimilation. The profiler winds were used for four dimensional data assimilation. The wind flows obtained from the boundary-layer model simulations were used to compare the model-computed spectral density with that of observations. Based on the spectral analysis and numerical experiments, a parameterization has been developed for the unmodeled energy in the current models. The parameterization has been verified by simple numerical experiments and evaluated by comparison between model estimated plume dispersion and that derived from aircraft observations. The implications of this study are substantial for both regulatory and research agencies.;The modeling of mesoscale plume dispersion has been based largely on the assumption of a Gaussian distribution and parameterized dispersion coefficients. On diurnal scales such assumptions and parameterizations are known to be erroneous as vertical shear in the horizontal wind distorts the plume. A more realistic approach is based on Lagrangian particle dispersion modeling. Lagrangian particle dispersion models attempt to mimic atmospheric transport and dispersion by using an ensemble of particles (hypothetical fluid elements) to trace pollutant movements. Thousands of particles are released and particle movements are made using Lagrangian techniques. The resolvable wind components that move the particles come from observations or prognostic dynamic meteorological models (at typical horizontal grid spacing of 5-50 km). In grid models, energy is resolved on a minimum scales of two to four grid spaces. For a model using a horizontal grid spacing of 25 km, the minimum wavelength of the energy spectrum that is resolved is about 100 km. In the past, below the resolvable scale, the sub-grid movements were parameterized using only boundary-layer turbulence information that has a scale length comparable to the height of the planetary boundary-layer (approximately 1 km). Thus, if significant atmospheric energy exists at wavelengths between about 1 km and about 100 km, this energy is not represented in dispersion parameterizations, and can lead to underestimation of plume dispersion on the mesoscale.