A unified representation of convective boundary-layer processes coupled to a probabilistic cloud closure: tests in an idealized GCM

We employ an eddy diffusivity/mass flux (EDMF) approach to represent all subgrid-scale convective and boundary-layer processes in a unified manner in a GCM. In this approach, convective and boundary-layer turbulence is decomposed into two components: smaller-scale eddies that are spatially quasi-homogenous, and updrafts that are spatially concentrated, with the former represented by a diffusive closure and the latter by a mass flux closure. We use the EDMF approach not only to represent boundary layer turbulence, as is previously been done, but also for shallow and deep convection, thus unifying the representation of these processes in a GCM such that they have well defined limits, for example, for a dry atmosphere. To this representation of subgrid-scale dynamics, we couple a probabilistic cloud closure based on Gaussian closures for budgets of adiabatically conserved quantities (liquid water potential temperature and total specific humidity).

This approach to the representation of subgrid-scale processes reduces the number of free parameters to be fitted compared with conventional approaches and gives them precise physical interpretations. We test this approach in an idealized moist GCM framework, We investigate the sensitivity of the results to different formulations of diffusivity coefficients, mass-flux coefficients, and lateral entrainment rates  and compare the results with large-eddy simulations.