Evaluation of Turbulence Depending Drag Coefficient in Plume Rise Model for Fire Smoke Dispersion

It is well-known that the correct modeling of the plume rise is fundamental for a proper description of pollutants dispersion, especially for highly buoyant plumes like those emitted by wildfires. In this work, we want to investigate the role of the drag coefficient (CD) in our plume rise scheme. We evaluate different formulations taken from the literature and suggest the best option for the plume rise scheme embedded in the Lagrangian Stochastic model SPRAY -WEB. As a matter of fact, previous works on plume rise models are based on the drag coefficient which is a quantity that should be determined. There is not a general consensus on the best values for CD. Some values suggested in the literature are simply constants that do not directly depend on the meteorological and turbulent variables. In order to evaluate the model performances obtained using different CD formulations we simulate a field experiment, carried out in August 2013 in Idaho (USA), in which prescribed fires were observed and the physical and chemical parameters were measured. As a matter of fact, plumes emitted from fires are affected by strong buoyancy and very low or even negligible initial vertical momentum. These conditions are very different from traditional plumes from stacks for which the classical plume rise models are built. For this reason, the case studies chosen for the tests are very challenging and they allow us to assess the scheme here proposed in the best way. Comparison among the results obtained with different CD parameterizations are presented and discussed. Three of the four CD models tested derive from an extension of the Stokes' law; the fourth one is a more refined model derived from the Shanks transformation of the Goldstein series. This last model seems to give better results as regards the maximum height of the plume, but is the one that underestimates most CO concentrations.