Subfilter-Scale Processes and the Simulation of Convective Clouds in the Terra Incognita

The terra incognita, also known as the gray zone, of numerical simulations refers to a range of model grid spacing that is of the order of the scale of the energetic turbulence. Challenges for large eddy simulations (LES) arise in the terra incognita because the traditional models of subfilter-scale (SFS) turbulence in LES were designed for the grid spacing much smaller than the energy-containing eddies. As documented in previous simulations of atmospheric convection, LES in the terra incognita can over-predict updraft velocities and the amount of cloud water/ice in deep convection, and lead to grid-dependent organization of vertical motions in some cases. Here we use an LES model (CM1) to simulate moist convection with both high and terra-incognita resolutions, and with different SFS turbulence models. The turbulence models include both simple schemes, such as the Smagorinsky-based eddy-viscosity model, and more sophisticated models, such as ones using reconstruction and explicit filtering. Their performance of simulating clouds in the terra incognita is compared. The high-resolution data is then further used to evaluate the roles played by SFS processes in causing the dynamical and physical errors of simulating clouds in the terra incognita. The effects of SFS mixing of momentum, heat and cloud water/ice are analyzed, and the relevant feedbacks that may reinforce those effects, such as radiative process and surface fluxes, are also explored. Our analysis here on the effects of SFS mixing in the convective atmosphere will aid the development of more physically realistic turbulence models in the future.