A New Stochastic Eddy-Diffusivity/Mass-Flux Model: A Step Towards Unified Parameterization

We propose a physically consistent parameterization for weather and climate models, which represents different types of convection including dry, non-precipitating, and precipitating moist convection and boundary layer turbulence. The parameterization builds on a stochastic Eddy-Diffusivity/Mass-Flux (EDMF) approach and distinguishes between the local and non-local turbulence. The non-local turbulence is modeled with an ensemble of laterally entraining convective plumes. Condensation and precipitation within plumes are represented with a simple but realistic microphysical scheme. Lateral entrainment is modeled with a Monte-Carlo approach. The new parameterization is implemented in a single-column model (SCM). We show that it captures essential features of moist boundary layers ranging from stratocumulus to shallow and precipitating cumulus regimes. A detailed comparison of benchmark cases with large-eddy simulation (LES) results proves the value of the presented approach. The key advantages of the new parameterization are as follows. It can represent smooth transitions between different convective regimes including dry, shallow and deep convection without any arbitrarily defined trigger functions. It also allows us to simulate different types of convective plumes (that can be associated with different regimes) co-existing within the subgrid-scale domain.