Abstract
Convective parameterization is one of the major factors responsible for biases in global climate model (GCM) simulations. At a scale of ~100 km or larger, there exists a quasi-equilibrium between convection and the large-scale forcing. At grey zone scales (e.g., ~10 km), convection becomes more stochastic within a GCM grid box, and many important assumptions in conventional parameterization break down. Therefore, the representation of stochasticity of convection is a critical scientific issue in high-resolution E3SM development, and a scale-aware convection scheme is also urgently needed. The parameterization of microphysical processes within convective clouds is another issue of fundamental importance in climate simulations. Both hydrometeor detrainment into anvil clouds and precipitation efficiency depend on microphysical processes in convective updrafts and downdrafts.
In this study, we will show some preliminary results from our earlier work, including testing a stochastic convective scheme in the NCAR CAM5 that largely eliminated the “too-much-drizzle and too little heavy rain” bias, and the effect of convective microphysics processes on water cycle and cloud radiative forcing. Then we will discuss our plans to 1) incorporate the stochastic convective parameterization and couple it with a deterministic convection scheme; 2) improve the scale-awareness of existing convection schemes; 3) enhance and incorporate a two-moment convective microphysics parameterization for use in the next generation E3SM. Tests of these enhancements of convective parameterization will be carried out at 100 km and 25 km resolutions as well as using the Regionally Refined Meshes (RRM) tool to test at 12 km resolution at Southern Great Plains (SGP) and Tropical Western Pacific (TWP). Since all these enhancements of convective parameterization are important to the simulation of clouds and precipitation, we expect them to have significant impacts on climate simulations in E3SM.