The NOAA/EMC Vision for the Parameterization of Atmospheric Physical Processes in the Emerging National Unified Forecast System

The parameterization of physical processes is one of the largest sources of uncertainty in numerical prediction of climate and weather, and therefore one of the greatest challenges in improving such prediction. For most physical processes, e.g., moist convection, cloud microphysics, and turbulent mixing, multiple complex algorithms have been developed in attempts to find an optimal strategy for their representation, but these efforts have resulted in multiple different, but seemingly equally viable parameterizations rather than a single optimal solution.

The vision for a national Unified Forecast System (UFS) includes unification of physical parameterizations. This could include a merging of parameterizations that have traditionally been designed to represent distinct physical processes (e.g., dry boundary-layer convection and deep moist convection), but the initial emphasis will be on unifying individual approaches for parameterizing the same physical process(es).

The strategy for this effort will be to identify the strengths and weaknesses of different parameterizations for the same process(es), combine the strengths, and mitigate the weaknesses. A key tool in this effort will be a hierarchical testing framework that is being developed at the Global Model Test Bed (GMTB). This framework will provide an unprecedented opportunity to examine precisely how individual parameterizations represent key physical processes with different approaches and how these representations uniquely interact with progressively more complex and non-linear dynamical systems. It is anticipated that the hierarchy of tests will provide a systematic and reliable framework for identifying and combining best practices, allowing us to optimize performance of unified parameterizations across a spectrum of dynamical systems. The ultimate goal is to develop an optimally-performing suite of parameterizations that is insensitive to model resolution and appropriately sensitive to atmospheric aerosols and other constituents that have been largely ignored for many applications of numerical weather prediction.

In this presentation, the short term goals and longer term strategic plan for representing unresolved physical processes in the national UFS will be discussed