Revisiting Kalnay's “Rules for Physics Interoperability” 25 Years Later

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Wednesday, 7 January 2015: 10:45 AM
229A (Phoenix Convention Center - West and North Buildings)
James D. Doyle, NRL, Monterey, CA; and M. Iredell, P. Tripp, J. Dudhia, T. Henderson, J. Michalakes, J. A. Ridout, J. Rosinski, S. Rugg, R. Adams Selin, T. R. Whitcomb, K. Lutz, and D. McCarren

One of the motivations for the Kalnay et al. (1989, BAMS, 620-623) rules for interchange of physical parameterizations was that “progress in operational numerical weather prediction and atmospheric modeling has been slowed by the difficulties of testing new schemes developed at different centers, and even within a center….”. These guidelines or rules to help facilitate the interchange of Fortran subroutines for physical parameterizations have been widely adopted over the past 25 years, and the operational and research numerical weather prediction (NWP) communities have benefited from this insightful contribution from Dr. Kalnay and collaborators. However, during this 25 year time span, computational hardware and software needed for operational NWP have changed considerably. For example, parallel computing technology, Message Passing Interface (MPI), accelerators, new Fortran standards (e.g. Fortran 90, Fortran 95, etc.), more complex physical parameterizations, have become more widely used in the past two decades. To address some of these new technical challenges, a multi-agency Physical Interoperability (PI) group has formed to modernize the Kalnay rules, as part of the National Unified Operational Prediction Capability (NUOPC). In this presentation, we will highlight some of the recommendations the PI group has made to the original Kalnay rules in order to address some of the new computational and technical challenges. Additionally, we will discuss the design and development of a software driver interface that will allow for the incorporation of physical parameterizations into operational and research NWP models in an efficient and standardized manner, much in line with the original vision of Kalnay et al. (1989).