Search

Three-dimensional Lagrangian particle model, listed with tunnel experiment diffusion data, is used to compute the emitted tracer concentration in the base region of a two-dimensional hill. The results are used in a mass-consistent model to compute mean concentration fields and to interpolate the turbulence structure. The concentration fields depict topographical data on the flow structures in turbulence system.

Byline: E. Ferrero, D. Anfossi, G. Brusasca, G. Tinarelli The capabilities of the Lagrangian particle model LAMBDA in simulating tracer dispersion are shown. Two different datasets of experimental campaigns proposed in the Manno Workshop (Copenhagen and Lillestrom) are used to compare measured and calculated concentrations. LAMBDA, which is based on the Langevin equation, can use as input higher moments (like skewness and cross-correlations) of the atmospheric probability density function (p.d.f.). The model can take into account the time and space variation of all the input parameters and can use numerical schemes of different complexity depending on the number of available moments of the atmospheric p.d.f. However, it is able to correctly simulate pollutant dispersion using simpler input datasets and proper parametrisations, as was shown in previous papers. The comparisons of observed and predicted concentrations and the model evaluation indicate that LAMBDA can yield reasonably good results.

Byline: S Finardi, M.G. Morselli, G. Brusasca, G. Tinarelli Horizontal non-uniformity has important effects on the structure of the atmospheric boundary layer and, consequently, on pollutant transport and dispersion. These effects can be dominant in coastal sites or in the presence of complex terrain. Advanced 3-D atmospheric dispersion models can be used to calculate short-term pollutant concentrations in such complex situations. The use of new generation models (puff, Eulerian or particle models) in the frame of pollution control networks requires a real-time, fast and reliable evaluation of surface layer similarity scale variables, able to take account of the horizontal non-homogeneity of the computed parameters. A meteorological pre-processor has been built to calculate atmospheric turbulence scale parameters (u*, L, H0, zi, w*) over a 2-D domain, covering land-use variations, local meteorological data (surface data, SODAR, RASS), and results from diagnostic wind field models. The code implements updated methods to calculate similarity scale parameters, both on shore and over sea (lake) surfaces. The pre-processor is installed in the atmospheric pollution control network of a thermal power plant in Northern Italy, where a system consisting of a mass-consistent wind field model and a Lagrangian particle model gives real-time estimates of pollutant concentrations on an hourly base, using the local air quality control network data.

Byline: E. Ferrero, D. Anfossi, G. Tinarelli The results of numerical simulations of tracer dispersion in an urban stable boundary layer are presented. Both Gaussian approximation and fourth order closure are considered and a Gram-Charlier probability density function is used in our Lagrangian stochastic model. Different parameterisations for the boundary-layer height and different values of kurtosis are tested. The model is based on the generalised Langevin equation, whose coefficients are solution of the Fokker-Planck equation and plume rise is taken into account. The model accuracy is assessed by means of a model evaluation based on the predicted and observed crosswind integrated concentrations, maximum on the arc and standard deviation of the crosswind concentration distribution on the arc.

Byline: E. Ferrero, D. Anfossi, G. Tinarelli, M. Tamiazzo In this work, two different Lagrangian stochastic models for convective turbulence are compared and tested against the Indianapolis dataset. The two models are based on the generalised Langevin equation, whose coefficients are solutions of the Fokker-Planck equation. The first model uses the bi-Gaussian probability density function (PDF) following Luhar and Britter (1989) and the second one, proposed by us (Ferrero and Anfossi, 1998), is based on the Gram-Charlier PDF. Both models are able to take into account up to the third order moment of the wind velocity fluctuation vertical component and make use of the Gaussian assumption in the horizontal directions. Basically, the two models considered here are different versions of our stochastic dispersion model for flat terrain, LAMBDA. This means that plume rise is taken into account and that the correct boundary conditions are used. The results obtained are presented and their differences discussed.

Byline: S. Finardi, G. Tinarelli, A. Nanni, G. Brusasca, G. Carboni A three-dimensional modelling system, made by a mass-consistent wind field model and by a Lagrangian particle model, has been applied to a Mediterranean complex coastal site to describe the atmospheric dispersion of pollutants emitted by a thermal power plant. Different short-term episodes have been simulated to cover the seasonal climate variability. Model results have been compared with local SO2