It is well recognized that the difficulty of properly simulating the effects of clouds and associated processes is one of the most significant limiting factors for global climate models. A large part of the problem is that the current climate models can not directly resolve or even permit clouds because of their low resolution. On the other hand, extensive climate simulations with global cloud-resolving models (GCRMs) will not be practical in the next decade. 

As an intermediate solution, the Multi-scale Modeling Framework (MMF) was introduced to global modeling. In an MMF, the cloud-scale processes are explicitly simulated by embedding a 2D CRM as a “super-parameterization (SP)” in each GCM grid column. This approach has shown its good ability to simulate various atmospheric phenomena that are closely linked to the cloud-scale processes, such as the diurnal cycle of precipitation, the Asian monsoon, and the Madden-Julian Oscillation. In spite of its encouraging performance in simulating these phenomena, however, the current MMF has inherent deficiencies. To mitigate the deficiencies of the current SP, the Q3D MMF has been introduced by Jung and Arakawa. The Q3D MMF is a major advance over the current SP: 1) The CRMs are not confined into the dynamical-core grid boxes. 2) The CRMs sense the three-dimensional large- and cloud-scale environment. 3) Two perpendicular sets of CRMs are used. 4) The Q3D MMF can simulate the vertical momentum transport due to both convection and gravity waves. 5) The CRMs also resolve the steep surface topography along the channel direction. Due to these features, the Q3D MMF is scale-aware, which means that it converges to a global CRM as the resolution of the GCM is increased.

In this project we will implement the Quasi-3D Multiscale Modeling Framework (Q3D MMF) as a physics option in ACME. The proposed model will use ACME’s spectral element dynamical core on a cubed sphere, and will be fully compatible with the ACME modeling and data infrastructure. In terms of CPU time, the computational cost of the proposed model will be slightly higher than the first-generation SP models, but much less than the cost of a GCRM.

We anticipate that the model will have the following unique capabilities, among others:

• The ability to simulate both orographic and non-orographic gravity-wave drag by solving the equation of motion on the CRM’s grid.
• The ability to simulate orographically enhanced precipitation, including extreme precipitation events in mountainous terrain.
• The ability to simulate the effects on ENSO of deep tropospheric momentum transport by convection and non-orographic gravity waves.
• The ability to directly simulate convective transports of chemical tracers and aerosols. Deep tropical convection plays a crucial role in vertical transport of chemical tracers and aerosols. Although we do not plan to pursue this area ourselves, we will welcome collaborations with other ACME investigators who are working on chemical transports. 

The Q3D version of ACME can bridge the gap between the much more expensive global cloud-resolving models (and conventional SP models). Extensive climate simulations with GCRMs will not be practical in the next decade or more. During that time, the Q3D MMF can be used as the best affordable alternative to a GCRM. The Q3D version of ACME will be a cutting-edge tool for simulations of current and future climates, including especially the hydrologic cycle on local and global scales, and extreme precipitation events in mountainous regions.