An integrated TKE based eddy-diffusivity/mass-flux boundary layer scheme for the dry convective boundary layer

This study presents a new approach to the eddy-diffusivity/mass-flux (EDMF) framework for the modeling of convective boundary layers. At the roots of EDMF lays a decomposition of turbulent transport mechanisms into strong ascending updrafts and the remaining smaller-scale and more isotropic turbulent motions. The turbulent fluxes can be therefore described using two conventional approaches: mass-flux (MF) for the organized thermals and eddy-diffusivity (ED) for the remaining turbulent field. Since both MF and ED correspond to various regimes of the turbulent kinetic energy (TKE), it seems reasonable to couple the EDMF framework with a TKE closure. Such approach could allow for more physical and less arbitrary formulation of various parameters in the model. In this study the coupling is achieved through the use of a new parameterization for the lateral entrainment coefficient ε and the MF contribution to the buoyancy source of TKE. Some other important features of the EDMF parameterization presented here include a revised mixing length formulation and Monin–Obukhov stability scaling for the surface layer. The scheme is implemented in a 1D model. Several cases of dry convective boundary layers (CBL) with different surface sensible heat fluxes in the free-convection limit are investigated. Results are compared to large-eddy simulations (LES). Very good agreement between LES and 1D model is achieved with respect to mean profiles, boundary layer evolution, and updraft characteristic. Some disagreements between models are found to most likely relate to deficiencies in TKE simulation in the 1D model. Comparison with other previously established ε parameterizations shows that the new TKE based formulation leads to equally accurate, and in many respects better simulation of CBL. The very encouraging results obtained with the proposed EDMF framework indicate that full integration of EDMF with higher order closures is possible and can further improve boundary layer simulations.