NOAA storm researchers working to improve severe weather forecasts

Thunderstorms with lightning, hail, strong winds and tornadoes can be devastating, resulting in hundreds of deaths and millions of dollars in damage each year. National Oceanic and Atmospheric Administration (NOAA) researchers and forecasters in Norman, Okla., are working toward improving the tools used to predict such storms. Their aim is to provide the public more time to prepare for severe thunderstorm events.

NOAA researchers and forecasters are collaborating in an experiment to improve forecasts of when and where severe storms will occur. This spring, participants are evaluating several operational and experimental computer models and algorithms used by forecasters to determine which ones provide the best guidance.

Forecasters currently use a combination of computer model forecasts and observational tools, such as satellites and radars, to determine when to issue a severe thunderstorm watch. The experiment is designed to explore whether the computer models have improved enough to provide the guidance forecasters need to confidently issue watches several hours in advance, while the skies are still clear.

“We need to identify more clearly under what circumstances and with which models we can predict severe thunderstorm development with significant lead time and confidence,” said John S. Kain, a research meteorologist at NOAA’s National Severe Storms Laboratory and co-leader of the project with Paul R. Janish, science infusion meteorologist with NOAA’s Storm Prediction Center. The computer models include operational versions developed at the NOAA National Centers for Environmental Prediction Environmental Modeling Center and the NOAA Forecast Systems Laboratory, as well as research versions developed at the NOAA National Severe Storms Laboratory.

“This is a great opportunity for us to stretch our forecasts toward the limits of current science,” explained Russell Schneider, chief of the SPC’s science support branch. “During the experiment, forecasters are forced to decide whether to issue watches sooner, sometimes before they see evidence in the satellite and radar data that storms have begun to form. It requires detailed analysis of multiple computer model forecasts made possible by the National Weather Service’s super computer.”

The next day, participants evaluate the output from the different models, rating them on their usefulness, and compare their forecasts with what actually happened.

“This process helps the forecasters develop a better understanding of the numerical models that provide their primary source of forecast guidance,” Kain said. “And it helps researchers design more useful model guidance products for forecasters’ specific needs.”

NSSL and SPC have worked together on several projects in the past few years, since the SPC moved its operations to Norman from Kansas City in 1997. A key goal of these programs is to improve forecasts of meteorological phenomena by speeding up the transfer of new technology and research ideas into forecast operations. At the NSSL and SPC, this is accomplished by combining the skills and mutual research interests of research scientists and forecasters.

Visiting scientists also participating in the experiment are from NOAA’s Forecast Systems Laboratory, National Centers for Environmental Prediction Environmental Modeling Center, National Weather Service Norman Forecast Office, Iowa State University and the University of Oklahoma.

More information about the program is available online: