Common modeling infrastructure

Common modeling infrastructure refers to software libraries that can be shared across multiple institutions in order to increase software reuse and interoperability in complex modeling systems. Early initiatives were in the climate and weather domain, where software components representing distinct physical domains (for example, ocean or atmosphere) tended to be developed by domain specialists, often at different organizations. In order to create complete applications, these needed to be combined together, using special coupling software that transferred and transformed data between the components. An additional challenge was that these models required supercomputers to run, and specialized software to address routine functions such as I/O, parallel data communications, and error handling.

In this context, the incentives for common modeling infrastructure included:^[1]

• the need for physical consistency across components, such as using the same physical constants and calendars
• the need for clear, standard software interfaces so components could be exchanged and introduced easily
• the desire to create systems in which the components of complex models were separable, so that controlled experiments could be performed in which only specific components were changed
• the desire to share as many of the libraries handling non-scientific modeling functions as possible

In the late 1990s and early 2000s, a series of journal papers and government reports described common modeling infrastructure as necessary to the competitiveness and evolution of the U.S. Earth science modeling community. These reports resulted in a number of new community projects. The Earth System Modeling Framework (ESMF) in the U.S. and the PRogramme for Integrated Earth System Modeling (PRISM) in Europe were two of the largest. Similar projects were initiated in related domains, including the Space Weather Modeling Framework and the Center for Integrated Space Weather Modeling in the space weather community.

In the late 2000s, the growing global interest in climate change refocused attention on the role of common modeling infrastructure in solving complex problems. Some of the specific factors that new common modeling infrastructure projects seek to address include:

• increased demand for access to climate model output and related information following the widespread dissemination of the conclusions of the IPCC Fourth Assessment Report in 2007
• the desire to model anticipated climate change impacts, often by linking together a variety of local, regional, and global models and datasets
• the exploration of weather modeling techniques and observational datasets to improve and validate climate model predictions^[2]

Active common modeling infrastructure projects include the Network Common Data Form (NetCDF) library, the Spherical Coordinate Remapping and Interpolation Package (SCRIP), the GFDL Flexible Modeling System (FMS), the OASIS coupler developed at CERFACS, and the multi-agency Earth System Modeling Framework (ESMF). The NOAA Global Interoperability Program is an integrative effort whose aim is to encourage coordination across such projects.

References

1. ^ Dickinson, R. E.; S. E. Zebiak, J. L. Anderson, M. L. Blackmon, C. DeLuca, T. F. Hogan, M. Iredell, M. Ji, R. B. Rood, M. J. Suarez, and K. E. Taylor (2002). "How Can We Advance Our Climate and Weather Models as a Community?". Bulletin of the American Meteorological Society 83: 431–434. doi:10.1175/1520-0477. http://ams.allenpress.com/perlserv/?doi=10.1175%2F1520-0477%282002%29083%3C0431%3AHCWAOW%3E2.3.CO%3B2&request=get-abstract. 
2. ^ Palmer, T. N.; F. J. Doblas-Reyes, A. Weisheimer, and M. J. Rodwell (2008). "Toward Seamless Prediction: Calibration of Climate Change Projections Using Seasonal Forecasts". Bulletin of the American Meteorological Society 89: 459–470. doi:10.1175/BAMS-89-4-459. http://ams.allenpress.com/perlserv/?request=get-abstract&doi=10.1175%2FBAMS-89-4-459&ct=1.