Global Simulations from CAM with a Unified Convection Parameterization using CLUBB and Subcolumns

Abstract:

The newest version of the Community Atmosphere Model (CAM) will support subcolumns as a method to better couple sub-grid-scale convective and microphysical processes. We utilize this feature and samples from a PDF-based moist turbulence parameterization to produce a version of CAM where all convection (shallow, stratiform, and deep) is simulated with a single set of dynamic and microphysical equations. We call this version of the model CAM-CLUBB-SILHS, where CLUBB (Cloud Layers Unified By Binormals) is our higher-order closure convection and turbulence parameterization and SILHS (Subgrid Importance Latin Hypercube Sampler) is our sampler and the basis for our subcolumn generation.

Each physics timestep in this model, the CLUBB parameterization runs to calculate convective tendencies. In order to close the higher order moments, CLUBB calculates a new multi-variate PDF describing the subgrid distribution of moisture and temperature at each level. SILHS samples from that PDF and creates profiles of vapor, temperature, vertical velocity, cloud water and ice, and cloud water and ice number concentration. The microphysics scheme runs on each subcolumn seperately. The resulting tendencies are averaged together and returned to the model as a grid mean tendency. This use of subcolumns allows us to explicitly represent subgrid scale clouds and moisture distributions for microphysical calculations. Using this framework and no other convective parameterizations, we are able to produce stable, realistic, global atmospheric simulations in CAM. This study will present results from long-term atmospheric simulations, discuss the impact of subcolumns on the model, and show improvements in the model's tropical wave simulation.