Deep convection is a multiscale process that influences the budgets of heat, moisture and momentum significantly. In global climate models, the thermodynamic effects of convection are normally treated by parametrization schemes, with a separate formulation for convective momentum transport (CMT). The schemes for current thermodynamic and momentum parametrizations are based on upright entraining plume models that do not account for vertically tilted mesoscale circulations, which characterize organized convection in sheared environments. The associated countergradient vertical transport of horizontal momentum fundamentally affects dynamical interactions between the convection and the mean flow. This study examines the CMT properties of simulated idealized mesoscale convective systems, including their sensitivity to horizontal resolution, domain size and lateral boundary conditions. It is found that, even for large domains, the horizontal gradient terms are important, especially the mesoscale pressure gradients, which are neglected in CMT parametrizations. A nonlinear analytic model provides a dynamical foundation for the effects of convective organization, including the role of the horizontal pressure gradient. It is found that a small computational domain affects the convective organization adversely by generating artificially large compensating subsidence and an unrealistic evolution of CMT. Finally, analyses of the cross‐updraught/downdraught pressure gradients expose significant uncertainties in their representation in contemporary CMT parametrization schemes.
Read full abstract