Testing Physical Parameterization Development in a Single-column Model Framework for Operational NWP

In preparation for operational NWP implementation, the process for testing advances in physical parameterizations requires a battery of tests within different modeling paradigms from ``unit test"-like configurations all the way through fully-coupled and cycled global runs at operational resolution. The sequence of increasingly complex tests can be thought of as a hierarchical testing framework, where subsequent, more complex testing paradigms build upon results from simpler ones. One implementation of this idea is the Physics Test Harness in use by the Global Model Test Bed (GMTB) project of the Developmental Testbed Center (DTC). A key component of this test harness is a single-column model (SCM) developed by GMTB. An SCM is a valuable tool for testing how a collection of physics schemes combined into a suite interacts and performs for a variety of meteorological conditions at a low computational cost. Further, while interpretation of the results is certainly dependent on the quality of the initial conditions and forcing, it is also uncomplicated by errors in three-dimensional flow or by numerical artifacts from a full dynamical core. The SCM developed by GMTB has the additional advantage of using the Common Community Physics Package (CCPP), a collection of vetted and supported physical parameterization schemes from which the operational community will assemble operational suites and pre-operational suites consisting of advanced options in the testing phase.

The GMTB SCM is initialized and forced from observational field campaign data assembled into case studies. GMTB provides a limited number of cases for use with the SCM from past Global Atmospheric System Studies (GASS) experiments to ongoing field campaigns, such as ARM's LES ARM Symbiotic Simulation and Observation (LASSO). External users are encouraged to generate new cases from other datasets and to share them with the broader community of users. Examples of using the SCM to analyze differences in physics suite performance in the context of several case studies will be presented.