Implementation and Evaluation of a New Shallow Convection Scheme in WRF

To help advance the understanding of the cloud life cycle and improve model skill in representing the effects of clouds, especially shallow clouds, Deng et al. (2003a and b) developed a mass-flux-based shallow convection parameterization (SCP) in MM5. In the prototype SCP, convection is triggered by the vertical velocities determined by various factors including the turbulent kinetic energy (TKE) within the planetary boundary layer (PBL). The convective parameterization closure is determined using a hybrid approach combining the boundary layer TKE and CAPE removal, depending on the depth of the convective updraft. In addition, there are two predictive equations for cloud water and cloud fraction of neutrally-buoyant clouds (NBC), with comprehensive cloud production and dissipation processes that take into account the cloud microphysical processes. The Penn State prototype SCP can transition to a deep convection regime when the cloud layer becomes sufficiently deep. On the other hand, the subgrid clouds produced by the shallow convection scheme can smoothly transition to explicit clouds when the grid cell becomes saturated. The SCP is currently being implemented into WRF, with the goal of improved forecasts of cloud fraction and radiation budgets. Here we will present some preliminary results, using both 1-D and 3-D WRF configurations, from several convective cases selected from both marine and continental environments. We will show that WRF with the Penn State SCP is able to reproduce the majority of features of the MM5 results, including cloud depth, cloud fraction and cloud water. We will discuss the interaction between the model-predicted partial cloudiness and atmospheric radiation, as well as future development plans that include extension to other TKE-based PBL schemes as well as non-TKE based schemes, and anticipated upgrades to the microphysics representation.