A Shallow Convection Scheme for 3-D Regional Scale Air-Quality Applications

Sub-grid scale shallow convection generally has been treated in a rather crude way in most meteorological models used for air-quality and other mesoscale applications. Often, shallow convection is represented as a simple mixing process (perhaps proceeding toward a target mixing-line profile), with cloud fraction diagnosed directly from the environmental relative humidity. While this approach is fairly reliable for global models, it is inadequate for regional air-quality applications which require detailed knowledge of cloud properties, such as the cloud base mass flux, number of clouds, depth, entrainment and detrainment profiles, vertical velocity of the updrafts, and liquid water profiles. These quantities are necessary to define how long parcels originating in the boundary-layer experience saturated conditions as they pass through the cloud, what water contents they are exposed to, and where those parcels are detrained from the cloud, which are required to drive aqueous reactions in air-quality models.

A detailed shallow-cloud parameterization suitable for air-quality applications has been under development at Penn State for the past four years. The cloud sub-model is a 1-D vertical scheme that uses a hybrid mass-flux closure that is based on the boundary-layer depth when clouds are very shallow and on the convective available potential energy (CAPE) when clouds become deeper. The parameterization determines updraft characteristics using vertical velocities calculated from the parcel buoyancy equation for an entraining/detraining convective cloud. When updraft mass is detrained, it is not allowed to mix directly with the environment, as is done in some other cloud sub-models. Rather, it detrains into a second class of clouds with nearly neutral buoyancy. The fractional area and liquid water content of these neutrally buoyant clouds (NBCs) are represented with two prognostic equations. The source terms are supplied by the updraft detrainment and the dissipation is modeled using horizontal and vertical mixing, ice settling, cloud-top entrainment instability and precipitation. The 1-D shallow cloud scheme has undergone extensive testing and evaluation in a variety of 1-D environments.

The shallow-cloud scheme now has been installed in the 3-D non-hydrostatic Penn State/ National Center for Atmospheric Research mesoscale model, known as MM5. Initial applications have been made for a marine environment near the Azores during the ASTEX experiment. This allowed evaluation of model performance in stratus, strato-cumulus and trade cumulus conditions, and verification with special observations from ASTEX. In a series of experiments it is shown that the MM5 with the cloud scheme can reproduce reasonably the character of the shallow convection regimes in the mid-ocean environment. The response of cloud fields in the vicinity of a front is examined and the importance of a realistic initial vertical thermodynamic structure is demonstrated.