Wolff, Jamie K., Louisa B. Nance, Barbara G. Brown and Ying-Hwa Kuo,
National Center for Atmospheric Research, Brad S. Ferrier, National Centers for
Environmental Prediction, and Clifford F. Mass, University of Washington
To facilitate the
transition of new science innovations from research to operations (R2O), the
Developmental Testbed Center (DTC) encourages the user community to contribute
new innovations back to the code base.
Through this process, a crucial aspect of the DTC mission can be
exercised: extensive testing and evaluation of promising codes emerging from
the research community can be performed in a common framework to demonstrate
the potential of new science and technologies for use in operations. For codes to progress from the research
to operational realm, several stages of testing must occur. In an effort to streamline the R2O
process, the DTC has laid out a testing protocol to engage the research
community in the steps necessary for preparing and nominating mesoscale
modeling code deemed ready for extensive DTC testing and evaluation in an
operationally similar environment.
Briefly, the first stage of testing will be conducted by the researchers
on high-impact or field program case studies. In an effort to provide a common framework for researchers
to demonstrate the merits of new developments, the DTC is establishing the
Mesoscale Model Evaluation Testbed (MMET), which will provide initialization
and observation data sets that can be used by the entire user community for
testing and evaluation at this initial stage. The DTC will utilize the same data sets from the MMET to
provide baseline results to the research community for specific operational
configurations. If improved
forecast accuracy is shown during the first stage of testing using reasonable
compute resources, the innovation may be recommended to continue on to the
second stage of testing. The
second stage of testing will be conducted by the DTC and would be more
extensive in nature and potentially include data assimilation cycling depending
on the target application. Along
with sharing the extensive test results with the user community, information
will be shared with interested operational entities. The ultimate decision to proceed to the pre-implementation
testing phase would be based on a variety of factors, including forecast performance
and computational resource requirements.
Through this process, the goal is to accelerate the rate at which new
technology is infused into operational weather forecasting.