This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| IFS (Integrated Forecasting System) | |
|---|---|
| Name | IFS (Integrated Forecasting System) |
| Developer | European Centre for Medium-Range Weather Forecasts |
| Initial release | 1990s |
| Latest release | ongoing |
| Programming language | Fortran, Python (programming language), C (programming language) |
| Platform | High-performance computing, ECMWF Cray Supercomputer |
| License | proprietary / research-use |
IFS (Integrated Forecasting System) The Integrated Forecasting System is a comprehensive numerical weather prediction and data assimilation suite developed and maintained by the European Centre for Medium-Range Weather Forecasts. It provides global and regional analyses and forecasts used by national meteorological services, research institutions, and operational centers such as Met Office, Deutscher Wetterdienst, Météo-France, and National Oceanic and Atmospheric Administration. IFS underpins seasonal to medium-range forecasting and supports climate reanalysis, emergency response, and commercial meteorological services.
IFS combines dynamical core models, physical parameterizations, data assimilation, and post-processing to produce deterministic and ensemble forecasts. It interoperates with systems like Copernicus Programme services, World Meteorological Organization frameworks, and national supercomputing facilities operated by European Commission, NASA, and Japan Meteorological Agency. Users include International Civil Aviation Organization, United Nations Office for Disaster Risk Reduction, and private vendors such as The Weather Company.
IFS evolved from post‑war numerical weather prediction efforts at institutions like Met Office and National Meteorological Center (United States), integrating advances from projects led by researchers associated with ECMWF and collaborations with European Space Agency and national services. Major milestones track adoption of satellite data from NOAA-AVHRR, MetOp, and GOES, the introduction of four-dimensional variational assimilation introduced in coordination with European Space Research Organisation initiatives, and transitions driven by computational milestones exemplified by machines from Cray Research and IBM installations. Cooperative projects with UK Research and Innovation and Horizon 2020 funded research accelerated coupling with ocean models used by Mercator Ocean.
IFS architecture is modular, comprising the dynamical core, physical parameterization suites, the 4D-Var and ensemble Kalman filter systems, and post-processing utilities. Core components interface with data management frameworks used by ECMWF and scheduling systems on HPC centers run by PRACE and national supercomputing centers such as NERSC. The stack integrates codebases in Fortran with workflow tools that mirror practices at European Organisation for Nuclear Research and Max Planck Institute computational groups. Auxiliary components link to land surface schemes developed in collaboration with European Space Agency projects and ice models from Norwegian Meteorological Institute.
IFS ingests observations from satellite missions like Sentinel (satellite constellation), MetOp, Himawari, ERS (satellite), and GOES-R Series, as well as in situ networks managed by Global Telecommunication System, Argo (oceanography), Radiosonde, and surface synoptic stations associated with World Meteorological Organization. Data assimilation employs 4D-Var and ensemble techniques developed alongside groups at University of Reading, ETH Zurich, and Princeton University. Collaboration with agencies such as NOAA, Japan Meteorological Agency, China Meteorological Administration, and Indian Meteorological Department ensures global observational coverage.
The dynamical core solves the primitive equations on spectral and gridpoint representations, integrating physics schemes for convection, radiation, cloud microphysics, and boundary-layer processes co-developed with research groups at Imperial College London, National Center for Atmospheric Research, University of Washington, and Institut Pierre-Simon Laplace. Coupling with ocean and sea-ice components uses model frameworks similar to those in NEMO (ocean model), CICE (sea ice model), and climate modules developed in coordination with Hadley Centre and Max Planck Institute for Meteorology. Parameterizations reflect research by teams at EUMETSAT and Lamont–Doherty Earth Observatory.
IFS outputs drive aviation forecasts for agencies like Eurocontrol and Federal Aviation Administration, support marine services for International Maritime Organization, and underpin hydrological forecasts for flood warning systems run by European Flood Awareness System and national civil protection agencies such as German Federal Office of Civil Protection and Disaster Assistance. Seasonal forecasts produced by IFS inform agricultural planning used by Food and Agriculture Organization, while climate reanalyses feed research at Intergovernmental Panel on Climate Change and institutions like Scripps Institution of Oceanography.
Verification employs standardized metrics coordinated via World Meteorological Organization task teams and intercomparison projects with Global Ensemble Forecast System and North American Mesoscale Forecast System. Studies by University of Oxford, University of Reading, and ETH Zurich evaluate skill scores, bias, and spread using datasets from ERA5, ECMWF Reanalysis (ERA-Interim), and independent observations from COSMIC and Buoy networks. Operational validation benchmarks against forecasts from Met Office Unified Model and NOAA GFS guide incremental improvements.
Ongoing R&D focuses on increasing resolution, coupling with biogeochemical and atmospheric chemistry models in partnership with Copernicus Atmosphere Monitoring Service, improving ensemble methods with contributions from Lionel Räisänen-style research groups, and exploiting machine learning techniques developed in collaboration with DeepMind, Google Research, and university labs at MIT and Stanford University. Future work addresses exascale adaptation for architectures from HPE, Cray and accelerators from NVIDIA, and integration with global observing systems promoted by Group on Earth Observations and Committee on Earth Observation Satellites.