Refinement of atmospheric meshes to cloud-permitting scales requires robust physics parameterizations that converge numerically and avoid double-counting with explicitly-resolved processes.
Mesh refinement is desirable to better resolve key atmospheric processes such as tropical cyclones, mountains effects on precipitation, and convective cloud systems. However, refinement of the atmospheric mesh down to sizes small enough to resolve cloud systems present several key challenges, including
- How can cloud parameterizations designed to represent processes not resolved by the coarse meshes of the past be made more robust so that they do not continue to represent those processes when the mesh is sufficiently fine to explicitly resolve the same processes?
- Given the necessary reduction in the time step as the mesh is refined, does the integration of the model physical and dynamical processes converge with decreasing time step at a rate sufficiently fast to justify the increasing computational cost?
To address these challenges, this project is developing new and more robust cloud and aerosol parameterizations, methods to quantify the time integration numerical errors and the processes producing those errors, and new integration methods to reduce the errors. Powerful calibration methods are being developed to determine optimal values of the key parameters controlling the behavior of the schemes in global simulations of clouds, aerosols, and cloud-aerosol interactions.