Development of a Global Quasi-3-D Multiscale Modeling Framework for E3SM

 

Authors

Joon-Hee Jung — Colorado State University - Atmospheric Science Department
Celal Konor — Colorado State University
Donald Dazlich — Colorado State University
David A. Randall — Colorado State University

Category

Convective clouds, including aerosol interactions

Description

The Quasi-3-D Multiscale Modeling Framework (Q3D MMF) is a second-generation MMF, which has following advances over the first-generation MMF: 1) The cloud-resolving models (CRMs) that replace conventional parameterizations are not confined to the large-scale dynamical-core grid cells; instead, they are seamlessly connected to each other in the form of “channels” that wrap around the planet. 2) The CRMs sense the three-dimensional large- and cloud-scale environment. 3) Two perpendicular sets of CRM channels are used. 4) The CRMs can resolve the steep surface topography along the channel direction. The basic design of the Q3D MMF has been developed and successfully tested in a limited-area modeling framework. Currently, we are developing a global version of the Q3D MMF by coupling the dynamical core of E3SM and the 3-D nonhydrostatic anelastic Vector-Vorticity Model (VVM) that has been tested with the limited-area version to demonstrate its suitability for this application. As a part of the development, the VVM has been adapted to the cubed-sphere grid so that it can be applied to global channel domains consistent with the E3SM grid structure. We have successfully tested the VVM on such a global channel domain. The interface for coupling and the computational infrastructure are currently under development. On a parallel track of the development, we are working on the improvement of the dynamical formulations and physical parameterizations used in the VVM. In this poster, we will present the current status of the development of the global Q3D MMF. Also, we will present the results from the simulations of Radiative-Convective Equilibrium with a limited-area version of the VVM, and the comparison of these results with other models.