CRM simulations of organized convection during TWP-ICE and their implications for cumulus parameterization

Poster PDF

Description

Most GCMs neglect the mesoscale organization of moist convection, which affects the diabatic heating profile, precipitation, the radiation budget, and the tropical general circulation. Cloud-resolving models (CRMs) can potentially provide insights into relationships between convection, stratiform rain and anvil regions, and the environment to guide parameterization development. CRMs are an imperfect standard, however, in part because of inadequacies in ice-phase microphysics. This is important because the primary physics distinction between the convective and stratiform regions is whether the updraft speed generally exceeds the ice crystal fall speed or vice versa. We have conducted simulations of the TWP-ICE active and break monsoon periods with WRF version 3.2 at 600 m horizontal resolution. We use updraft speed, cloud vertical extent, and surface rain rate criteria to partition the cloudy part of the WRF domain into convective, transition rain, stratiform rain, and anvil regions. The transition region adjacent to the convective region includes the air directly detrained from the parent convection, the air that is entrained into convective updrafts, and convective downdrafts and resulting boundary layer cold pools, all relevant to parameterization assumptions about how convection interacts with the “environment.” We note the following relevant features of the simulations for parameterization: (1) the strongest updrafts in the transition region are 1 m/s or greater, versus tens of cm/s in the stratiform region, suggesting that a mesoscale updraft parameterization should be initiated with detrainment of cumulus kinetic energy; (2) the relative humidity of lower troposphere transition region air is 10–20% higher than the domain mean when the domain is substantially sub-saturated, thus reducing the efficacy of entrainment; (3) convective area is a good predictor of stratiform rain area several hours later, but the lag is sensitive to the microphysics scheme used; (4) cumulus mass flux is an excellent predictor of stratiform region hydrometeor content 1–2 hours later, the lag also depending on the microphysics; (5) C-POL radar inferences of cluster component areas and hydrometeor type occurrence reveal inadequacies of different microphysics schemes; (6) temperature anomalies in the stratiform region are less than 0.5 K, implying that mesoscale updraft speed might be diagnosed in a GCM from the balance between diabatic heating and adiabatic cooling.