Non-local eddy diffusion and vertical mixing schemes for use in chemical transport models

Abstract

The atmospheric boundary layer (ABL) parameterization is one of the most important components of air quality and chemical transport models. The easiest and most commonly used method for the parameterization of vertical turbulent mixing in ABL is one that is analogous to molecular diffusion. These types of schemes are known as K-schemes, in which mixing occurs only between the closest model layers. This approach is suitable only when the scale of turbulent motion is much smaller than the scale of mean motion, i.e., during stable and neutral atmospheric conditions. K-schemes cannot describe the effects of large scale eddies, which are dominant in the convective boundary layer (CBL) and also cannot simulate counter gradient flows up to the gradient. Therefore, K-schemes are not recommended for the CBL. To avoid the disadvantages of K-schemes, a "non-local" turbulent kinetic energy (TKE) scheme can be used. However, the TKE scheme is able to represent the scale of the turbulence under any stability condition but the mixing is still local. It can also only represent small scales of turbulence, which reduces its reliability. During the convective condition, much of the mixing is caused by bounty plumes that originate in the surface layer and penetrate the capping inversion. In those circumstances, the ABL is characterized by narrow, fast-moving updrafts and wide, slow-moving downdrafts. This type of motion is usually parameterized using an asymmetrical convective mixing scheme (ACM). None of the above-mentioned schemes is able to represent mixing at all possible scales. However, a combination of the K-scheme and the ACM scheme can represent all of the spectrums of turbulence scales and is the most suitable to represent CBL mixing. In this chapter, a non-local TKE scheme, an ACM scheme and a linear combination of those schemes are comprehensively considered. To determine the features of the considered schemes, we performed numerical tests using Unified EMEP (European Monitoring and Evaluation Programme) chemical model and 1-D vertical diffusion models. The obtained outputs were compared with (i) the measured concentrations of the NO2 pollutant and (ii) the Large Eddy Simulation (LES). © 2012 by Nova Science Publishers, Inc. All rights reserved.