NOAA’s weather forecasts go hyper-local with next-generation weather model

This is a comparison of two weather forecast models looking six hours ahead for the New Jersey area. Image on left shows the forecast which doesn't distinguish localized hazardous weather. Image on right shows the new HRRR (High-Resolution Rapid Refresh) model that clearly depicts where local thunderstorms (yellow and red coloring) are likely.
This is a comparison of two weather forecast models looking six hours ahead for the New Jersey area. Image on left shows the forecast which doesn't distinguish localized hazardous weather. Image on right shows the new HRRR (High-Resolution Rapid Refresh) model that clearly depicts where local thunderstorms (yellow and red coloring) are likely. (Credit: NOAA)
VIDEO: NOAA launches new tool to improve weather forecasts. (Credit: NOAA)

Today, meteorologists at NOAA’s National Weather Service are using a new model that will help improve forecasts and warnings for severe weather events. Thanks to the High-Resolution Rapid Refresh (HRRR) model, forecasters will be able to pinpoint neighborhoods under threat of tornadoes and hail, heavy precipitation that could lead to flash flooding, or heavy snowfall and warn residents hours before a storm hits. It will also help forecasters provide more information to air traffic managers and pilots about hazards such as air turbulence and thunderstorms.

Developed over the last five years by researchers at ESRL's Global Systems Division (GSD), the HRRR is a NOAA research to operations success story. It provides forecasters more detailed, short-term information about a quickly developing small-scale storm by combining higher detail, more frequent radar input and an advanced representation of clouds and winds. The HRRR model forecasts are run in high-resolution every hour using the most recent observations with forecasts extending out 15 hours, allowing forecasters to better monitor rapidly developing and evolving localized storms.

High-Resolution

Hyper-local forecasts are possible with the HRRR because of higher resolution. The HRRR’s spatial resolution is four times finer than what is currently used in hourly updated NOAA models offering a more precise prediction of a storm’s location, formation, and structure. Using the HRRR, forecasters have an aerial image in which each pixel represents a neighborhood instead of a city. “This increase in resolution from eight to two miles is a game-changer,” added Lapenta.

What Goes In…

The HRRR starts with a full three-dimensional picture of the atmosphere one hour before the forecast and then brings in observations from surface stations, commercial aircraft, satellites, and weather balloons to create a more detailed and balanced starting point for the forecast. Another key innovation for the HRRR is adding in radar data every 15 minutes during that hour to help the model “know” where precipitation is ongoing. Integrating atmospheric data gathered before a model run, including radar data at a two mile resolution, provides a more accurate picture of what is happening in the atmosphere at the start of the forecast. This helps predict changes to storms and development of new storms faster than current models.

…And What Comes Out

The HRRR’s hourly output includes more frequent snap-shots, in 15 minutes intervals, of the atmosphere. With this information forecasters can better anticipate and predict the onset of a storm and critical details of its evolution, allowing for earlier watches and warnings.

“The HRRR model will provide forecasters a powerful tool to help them inform communities about evolving severe weather,” said Stan Benjamin, Ph.D., a research meteorologist at GSD who led the research team that developed the model. "Being able to warn the public of weather hazards earlier and with greater detail is an outstanding return from NOAA's investment in research and observation systems."

Many NOAA scientists were involved with testing, optimizing, and implementing the model, including experts at NOAA’s National Weather Service and its National Centers for Environmental Prediction. NOAA’s partners at the Cooperative Institute for Research in Environmental Science at the University of Colorado at Boulder and the Cooperative Institute for Research in the Atmosphere at Colorado State University, Fort Collins helped with development. NOAA researchers partnered with users such as the Federal Aviation Administration, the National Center for Atmospheric Research, and the Department of Energy to significantly improve forecasts for aviation, energy among other industries through the HRRR model.

“Implementation of the HRRR is just one of many model improvements made possible with NOAA’s boost in its supercomputing power for weather prediction,” said Louis Uccellini, Ph.D., director, National Weather Service. “With advances in our forecast models, like the HRRR, we’re moving toward building a Weather-Ready Nation by improving our forecasts, providing better information to decision makers, and helping communities become more weather-ready and resilient against severe weather events.”

Contacts: