WRF

WRF
Name	WRF
Developer	National Center for Atmospheric Research; National Oceanic and Atmospheric Administration; University Corporation for Atmospheric Research
Released	2000s
Programming language	Fortran; C; Python (interfaces)
Operating system	Linux; Unix; Windows (via ports)
Genre	Numerical weather prediction; atmospheric simulation
License	Open-source (various institutional licenses)

WRF is a numerical weather-prediction and atmospheric research model used for operational forecasting and scientific simulation. It acts as a bridge between observational programs such as Global Forecast System deployments and experimental projects at institutions like European Centre for Medium-Range Weather Forecasts and Jet Propulsion Laboratory. Researchers and forecasters employ it across domains including mesoscale meteorology, air quality studies, and climate downscaling in collaboration with organizations such as National Aeronautics and Space Administration and United States Geological Survey.

Overview

WRF is a mesoscale numerical model that provides capabilities for real-time forecasting and retrospective analysis used by agencies including National Weather Service, Met Office, and China Meteorological Administration. The system supports multiple dynamical cores, physics parameterizations, and data-assimilation interfaces comparable to frameworks like Weather Research and Forecasting Model (note: do not link model variants). It integrates with observational networks such as Global Telecommunications System feeds, Doppler radar networks, and satellite programs like GOES and METEOSAT for initialization and verification. WRF is designed to interoperate with community tools such as NetCDF, WRF-Chem extensions, and visualization packages developed by NCAR and university groups.

History and development

Development began in cooperative efforts among National Center for Atmospheric Research, NOAA laboratories, and university partners including Colorado State University and University of Oklahoma. Early milestones were tied to initiatives like the Atmospheric Model Intercomparison Project and software collaborations surrounding the Supercomputing community. Major releases followed iterative validation campaigns linked to field programs such as VORTEX and CALIPSO-related studies, and the codebase evolved alongside high-performance computing milestones exemplified by systems at Oak Ridge National Laboratory and NERSC. Contributions have come from international centers including Korean Meteorological Administration, Météo-France, and the Japan Meteorological Agency.

Model architecture and components

WRF's architecture includes a prognostic dynamical core, multiple physical-parameterization suites, and preprocessing/postprocessing utilities. The dynamical core supports advection schemes used in models like GFS and incorporates grid strategies including nested grids and nonhydrostatic formulations similar to designs used at ECMWF. Physics libraries contain microphysics schemes inspired by studies at Colorado State University and radiation modules informed by research from NCAR and NASA laboratories. The preprocessor ingests gridded datasets from sources including GFS, ECMWF Reanalysis, and regional analyses by European Reanalysis projects, while the postprocessor integrates with visualization tools developed at National Center for Atmospheric Research and universities such as Penn State.

Applications and use cases

Operational forecasting centers such as National Weather Service, Met Éireann, and Bureau of Meteorology deploy WRF configurations for short-range forecasting, severe-weather guidance, and hazardous-weather warnings. Research institutions use WRF for convection-permitting simulations in studies connected to projects like HyMeX and COARE, as well as renewable-energy assessments with partners like National Renewable Energy Laboratory. Environmental agencies deploy WRF-Chem integrations in air-quality modeling efforts coordinated with groups such as Environmental Protection Agency and European Environment Agency. Hydrologic coupling efforts tie WRF output to river models run by USGS and flood forecasting systems used by FEMA.

Performance and evaluation

WRF performance depends on configuration choices—dynamical core, microphysics, boundary-layer schemes—and on computational platforms from workstation clusters to petascale systems at Argonne National Laboratory. Evaluation campaigns reference verification metrics employed by World Meteorological Organization guidelines and intercomparison studies like Model Intercomparison Project for Climate subsets. Benchmarking typically compares WRF runs against analyses from ECMWF, GFS, and reanalysis products, and verification uses observational datasets from networks such as ASOS, AWOS, and research radars deployed in campaigns by NCAR and university consortia.

Licensing and community

WRF development is coordinated through community governance involving NCAR, NOAA, and academic contributors from institutions such as University of Washington and Penn State. Source distribution adheres to open-source policies negotiated with participating agencies and universities; contributors follow licensing practices similar to those used by scientific software projects hosted by UCAR and institutional repositories. A vibrant user community assembles at annual workshops sponsored by UCAR and AMS conferences, and collaborative development uses version-control workflows akin to models adopted by large scientific collaborations like CERN experiments.

Notable implementations and projects

Notable implementations include operational ensembles run by European Centre for Medium-Range Weather Forecasts partner centers, urban-scale deployments for projects led by MIT and Harvard in smart-city research, and regional climate downscaling efforts coordinated with IPCC assessment activities. Field campaigns that relied on WRF output include Tropical Warm Pool International Cloud Experiment and VORTEX2. Integrations with air-quality research were central to studies published in collaboration with EPA and international projects like EMEP. Large-scale computational adaptations have been demonstrated on supercomputers at Oak Ridge National Laboratory and NERSC for high-resolution ensembles used in hazard forecasting.

Category:Numerical weather prediction models