Code and configuration

The OpenIFS project at ECMWF provides versions of the OpenIFS model based on ECMWF's operational forecast model IFS (Integrated Forecast System).

OpenIFS does not include the observation processing system and the NEMO ocean model that are used in the operational IFS. With this exception, all of the forecast capability of IFS is also available in OpenIFS up to and including the operational forecast resolution. The OpenIFS models are designed to be easy to install on external computer systems, though the higher resolution forecasts (near operational resolution) will require the use of a high performance parallel computer.

Although the IFS can be run in a number of configurations, OpenIFS only supports forecast mode (NCONF=1) and interpolation mode (NCONF=927).

In addition to the 3D global forecasting model, a 1D single-column model version for process studies is part of OpenIFS.

Scientific capability

OpenIFS includes all the same forecasting capability of IFS but without any data assimilation and observation handling/processing. OpenIFS & IFS are the same with respect to:

  • Hydrostatic dynamical core.
  • ECMWF physics parametrization scheme.
  • ECMWF land surface model.
  • ECMWF wave model.

Unlike IFS, OpenIFS does not include:

  • 4D-Var and any observation handling.
  • Non-hydrostatic dynamics.
  • Adjoint / tangent-linear versions.
  • Ocean model.
  • Wave model (38r1 version only)
  • Ability to output Arpege file formats, only GRIB output is supported.

Support

Only operational configurations of IFS at ECMWF will be supported and for which documentation is available. Although we may endeavour to provide help on non-standard or untested configurations we cannot provide direct support for them.

Acknowledgements

The dynamical core IPR is shared with Meteo-France and collaborators. The IPR of the 'fullpos' and 'ddh' packages in OpenIFS are acknowledged as held by Meteo-France and we thank our colleagues for permission to include it in OpenIFS. Older versions of OpenIFS (before 48r1) used the FCM software for compilation and we gratefully acknowledge colleagues at the UK Met Office for their assistance.