A Shallow-Convection Parameterization for Mesoscale Models. Part II: Verification and Sensitivity Studies

Abstract

Formulations describing a new shallow-convection parameterization intended for mesoscale models have been described in a companion paper, Part I. In the present paper the convection scheme is tested and evaluated against observed datasets in both marine and continental environments. Additional experiments explore the sensitivity of the scheme to changes in model vertical resolution, key parameters of the closure, and the cloud-dissipation mechanisms. The new shallow-convection scheme uses a hybrid mass flux closure that adjusts linearly between two closure options based on (i) boundary layer turbulent kinetic energy (TKE) for very shallow clouds and (ii) convective available potential energy (CAPE) for deep clouds. Meanwhile, cloud mass detrained from convective updrafts acts as the source for a second class of subgrid clouds having nearly neutral buoyancy, which can persist for hours. Performance of the convection submodel is found to be quite reasonable in four cases covering a range of conditions. In a marine application taken from the Atlantic Stratocumulus Transition Experiment (ASTEX), a 1D column of air undergoes Lagrangian advection as it gradually transitions from a stratus environment at 41°N to a trade-cumulus environment at 30°N. Characteristics of the simulated cloud fields agree rather well with ASTEX observations in this weakly forced environment, including distributions of cloud fraction, cloud water, precipitation, and cloud liquid water path, and with simulations from other cloud-predicting models. The three other cases simulate various continental convective regimes (stratocumulus, cumulus humilis, and cumulonimbus) observed at the Atmospheric Radiation Measurement (ARM) Program Cloud and Radiation Testbed (CART) Southern Great Plains (SGP) ARM CART Central Facility in Lamont, Oklahoma. Verifications of the evolving cloud area, base height, cloud depth, and liquid water pathlength in these cases show that the shallow-convection scheme can adapt to different synoptic environments and the rapidly changing conditions associated with strong surface fluxes.