The effects of chemistry on the transport and the mixing of reacting scalars in the convective atmospheric boundary layer (CBL) are investigated. To do this, we use large-eddy simulation (LES) to calculate explicitly the different terms of the flux and (co)variance budget equations and to analyse in particular the role of the chemical term with respect to the thermodynamical terms. We examine a set of chemical cases that are representative of various turbulent reacting flows. The chemical scheme involves two reacting scalars undergoing a second-order reaction. In addition, we study a chemical cycle, based on a first- and a second-order reaction, to study the behaviour of chemical systems in equilibrium in turbulent flows. From the budget analysis, we found that the chemical terms become more relevant when the chemical timescale is similar to the turbulent timescale. In order to detemine the importance of the chemical terms, we compared these terms to the dynamical terms of the budget equations. For the flux of reactants, the chemical term becomes the dominant sink in the bulk of the CBL. As a result, flux profiles of reacting scalars have non-linear shapes. For the covariance, which accounts for the segregation of species in the CBL, the chemical term can act as a sink or source term. Consequently, reacting scalar covariance profiles deviate considerably from the inert scalar profile. When the chemistry is in equilibrium, the chemical term becomes negligible and therefore the flux and (co-)variance profiles are similar to those of inert scalrs. On the basis of the previous budget results, we develop a parameterisation that represents the segregation of reacting species in large-scale models under convective conditions. The parameterisation is applied to an atmospheric chemical mechanism that accounts for ozone formation and depletion in the CBL. We found a good agreement between the parameterisation and the LES results.
|Journal||Tellus Series B: Chemical and Physical Meteorology|
|Publication status||Published - 2003|
- large-eddy simulation
- modified k-model