Large-eddy simulation of the turbulent dispersion of a reactive plume froma point source into a neutral atmospheric boundary layer

We consider the dispersion from a point source in the neutral atmospheric b oundary layer. The emission from the point source consists of a species A w hich reacts with an ambient species B present in the boundary layer. The st udy is carried out by means of large-eddy simulation (LES). The main emphas is lies on the effect of turbulent mixing on chemistry in determining the c oncentration levels in the plume. First, we consider a relatively slow reac tion which is characterized by a reactive time scale that is much larger th an the turbulence time scale. Wind-tunnel experiments exist for this case ( Builtjes, 1983. In: Wispelaere, C. (Ed.), Air Pollution Modeling and its Ap plication II. Plenum Press, New York, pp. 59-84) and with these experimenta l data we validate our LES model. A comparison between computational and ex perimental data shows excellent agreement for the mean concentrations on th e plume centreline. Second, we consider the case of a relatively fast react ion which is characterized by a reaction time scale smaller than the turbul ent time scale. In this case the reaction is limited by the turbulent mixin g. First we consider the effect of mixing by the resolved scales obtained f rom the LES. The simulation results demonstrate that the total reaction rat e is reduced considerably compared to the ideal case in which species are c ompletely mixed. Next we consider the influence of subgrid scale mixing. To account for this influence a subgrid model for the chemistry is adopted. T he simulation results show that the subgrid scale mixing plays a prominent role close to the source. (C) 2000 Elsevier Science Ltd. All rights reserve d.