HRRR

HRRR
Name	HRRR
Developer	National Oceanic and Atmospheric Administration / National Centers for Environmental Prediction
Initial release	2014
Latest release	2020s
Programming language	Fortran, C
Operating system	Linux
License	Public domain (U.S. Government)
Website	National Weather Service product pages

HRRR

The High-Resolution Rapid Refresh (HRRR) is a near‑term, high‑resolution numerical weather prediction system designed for short‑range forecasting over the contiguous United States. It provides fine‑scale forecasts of atmospheric variables focused on convective events, aviation operations, and severe weather, delivering rapid updates to support agencies such as the Federal Aviation Administration, National Weather Service, and emergency management organizations. Built within the operational framework of National Oceanic and Atmospheric Administration modeling efforts, HRRR interfaces with satellite, radar, and surface networks to produce frequent assimilative forecasts for stakeholders including NOAA National Centers for Environmental Information, Air Traffic Control, and regional forecast offices.

Overview

HRRR is a convection‑allowing numerical model developed to provide hourly updated short‑range guidance with grid spacing typically around 3 km over the contiguous United States. It complements coarser systems such as the Global Forecast System, North American Mesoscale Model, and European Centre for Medium-Range Weather Forecasts outputs by resolving mesoscale and storm‑scale features critical to forecasts for events like derechos, squall lines, and mesoscale convective systems. End users include National Weather Service, Federal Aviation Administration, Department of Defense, regional National Oceanic and Atmospheric Administration partners, and academic research groups at institutions like University of Oklahoma and Colorado State University.

Model Development and Methodology

HRRR evolved from the Rapid Refresh (RAP) and leverages the nonhydrostatic dynamics and physics suites used in regional models maintained by National Centers for Environmental Prediction. Development involved collaborations with the Federal Aviation Administration, National Severe Storms Laboratory, and university partners to refine microphysics, boundary layer, and radiation parameterizations. HRRR uses explicit convection permitting schemes instead of parameterized deep convection, enabling depiction of hail, tornadoes, and updraft cores that are features of events studied at National Center for Atmospheric Research and in field campaigns such as VORTEX and PECAN. Methodological advances draw on research from Massachusetts Institute of Technology, Princeton University, and University of Washington on turbulence and cloud microphysics.

Data Assimilation and Input Sources

HRRR relies on hourly four‑dimensional data assimilation cycles ingesting a diverse suite of observations: Doppler radar radial velocities from the NEXRAD network, satellite radiances from GOES-R series, surface observations from the Meteorological Assimilation Data Ingest System, aircraft reports including Aircraft Meteorological Data Relay, and upper‑air soundings from National Weather Service radiosonde launches. Assimilation methods incorporate rapid update cycling and hybrid ensemble‑variational techniques influenced by research at University of Michigan and Scripps Institution of Oceanography. Ancillary inputs include land surface datasets derived from US Geological Survey products, sea surface temperatures from National Oceanic and Atmospheric Administration satellites, and boundary conditions from the Global Forecast System.

Forecast Products and Applications

Operational HRRR products include hourly gridded fields of precipitation, reflectivity, wind, temperature, humidity, convective available potential energy, and probabilistic severe‑weather indices used by Storm Prediction Center forecasters and Aviation Weather Center. Applications extend to airport planning at hubs like Hartsfield–Jackson Atlanta International Airport and Chicago O'Hare International Airport, wildfire smoke dispersion studies coordinated with U.S. Forest Service, wind energy forecasting for companies such as NextEra Energy, and hydrologic flash‑flood guidance used by the U.S. Geological Survey. Research communities use HRRR output for ensemble downscaling, coupling with urban canopy models studied at Massachusetts Institute of Technology and University of Illinois Urbana‑Champaign.

Performance and Verification

Verification studies compare HRRR against observations and models such as the North American Ensemble Forecast System and European Centre for Medium-Range Weather Forecasts ensembles, assessing metrics like equitable threat score, root‑mean‑square error, and probabilistic reliability. Independent evaluations by National Weather Service testbeds and academic groups at University of Oklahoma and Pennsylvania State University have documented HRRR strengths in short‑term precipitation placement and storm structure, while identifying biases in convective timing and warm‑season convective intensity. Ongoing verification efforts involve collaborations with NOAA Cooperative Institute for Research in the Atmosphere and the National Center for Atmospheric Research to refine microphysics and ensemble perturbation strategies.

Operational Implementation and Availability

HRRR is run operationally by National Centers for Environmental Prediction with frequent cycles providing hourly updates and forecasts extending to 18 hours, distributed through the National Weather Service dissemination systems, the Weather and Climate Operational Supercomputing System, and public data portals used by private vendors and research institutions. Model output is archived for post‑event analysis by NOAA National Centers for Environmental Information and supports decision support services for entities including FEMA, regional Department of Transportation offices, and commercial weather providers. Community contributions and code developments are coordinated with academic partners and federal laboratories such as National Severe Storms Laboratory and NOAA Geophysical Fluid Dynamics Laboratory.

Category:Numerical weather prediction models