Anelastic Convective Entities: Formulation and Properties for Revising Parameterization

A long-standing issue in the global climate models is the simulation of nighttime convection over land. Because of surface longwave radiative cooling, a layer of substantial convective inhibition (CIN) tends to occur over land after sunset. This CIN layer—in the conventional parameterization based on the local balance between buoyancy and updraft kinetic energy (KE)—prevents nighttime convection without an excessive trigger mechanism. Recently, a 1-D time-dependent model of an anelastic convective entity (ACE) is proposed to include the 1) dynamic entrainment, and 2) nonlocal acceleration. The latter arises from the pressure perturbation driven by the non-pressure-gradient-force (non-PGF) forcing including the buoyancy and mass flux convergence. Notably, the local balance between buoyancy and updraft KE is significantly modified by the proper inclusion of the pressure perturbation. In this work, preliminary results from the ACE prototype will be demonstrated for the characteristics of nonlocal dynamics, with supplemental analyses of cloud-resolving model (CRM) simulations to quantify the dependence of nonlocality on convective feature scales. The nonlocal dynamics tends to average over details of the non-PGF forcing (e.g., a CIN layer in a conditionally unstable environment) with the vertical range of influence being proportional to the horizontal size of the forcing, yielding robust solutions for the acceleration both in the ACE and in scale-dependent diagnostics of the CRM. Because of the vertically nonlocal effect—and the interaction of this with the surface boundary condition—a CIN layer is less inhibitive than would be inferred from a traditional plume model. Convection in the prototype ACE model can thus be initiated by the nonlocal buoyancy-driven acceleration even with a substantial CIN, tending to yield nighttime convection under conditions that a plume model would not. After initiation, in addition, the dynamic-induced nonlocal acceleration tends to yield bubble- or thermal-chain-like solutions resembling the CRM results in aspects that are quantified as a function of feature scale.