News Release

Learning how to make better "nowcasts" of weather

Peer-Reviewed Publication

NASA/Marshall Space Flight Center--Space Sciences Laboratory

Three days after heavy summer rains caused a 100-year flood in the southeast part of Huntsville, Ala., a group of weather scientists met on the other side of town to discuss what they are doing to improve "nowcasts." The timing was entirely coincidental - the meeting had been scheduled for several months - but it helped point up the need for nowcasts, or forecasts of what the weather will do in the next few hours.

On the afternoon and night of June 27, the skies over Huntsville opened up and dropped up to 6 inches of rain in two hours on sections of southeast Huntsville. Dozens of homes suffered extensive flooding, and a woman drowned when her car stalled on a flooded road and she tried to walk out. A warning of even a few hours could have let the city brace for a storm that statistically happens about once a century.

Meteorologists can measure in detail what is happening now, and predict in general terms what will happen over the next week. They now want to see if they can combine the wide range of operational tools - including Doppler radars that even TV stations have - and powerful numerical models of the atmosphere running on high-speed desktop computers to predict in detail what will happen over the next 30 minutes to six hours.

The Collaborative Workshop on Mesoscale Modeling and Short-Term Convective Forecasting was held June 30-July 1 by the southeast regional office of the National Weather Service at the Global Hydrology and Climate Center (GHCC). Also attending were scientists from the National Center for Atmospheric Research and other research institutions.

"The purpose of this workshop is to get the research and operational communities together to discuss how what we - the researchers - do can be translated to the operational community and thus to the public in some way," explained Dr. Bill Lapenta, a researcher at the GHCC.

"A lot of times research can be very focused and does not develop a true application," he noted. While the GHCC is not involved in weather forecasting, "Our research tries to help the forecasters do their job" by providing improved tools.

Among the projects they discussed is the Summer Convective Rainfall in Alabama Prediction Experiment (SCRAPE). Tom Bradshaw of the National Weather Service Office in Birmingham described it as an experiment to improve short-term forecasts across North Alabama.

"We finally have all the tools at our disposal to look at the weather on a smaller scale," Bradshaw said. The tools include the WSR-88D Doppler radar, advanced weather satellites, and a range of other instruments, plus MM5 - Mesoscale Model 5 - the latest computer model of weather events covering several counties or an entire state and lasting for a few hours. From these, Bradshaw hopes for more accurate rainfall predictions - called quantitative precipitation forecasts - plus longer warnings and better nowcasts.

Test runs started in mid-June and began in earnest on July 1. They run through Sept. 15. Five NWS meteorologists in Birmingham have been detailed to work on SCRAPE nowcasts without being distracted by routine duties. They will integrate data from radar, weather satellites, and other tools in an MM5 model to produce a nowcast covering the next 6 to 12 hours.

Bradshaw hopes to get at least 40 nowcasts that his team will compare with what actually happens. The lessons will be fed back into operations, and then into SCRAPE 2 in 2000 when Bradshaw hopes to produce accurate 3-hour nowcasts.

Nowcasts today are limited, tending to be too broad and not taking in enough detail, Bradshaw said, or perishable in that they require a lot of work to update as the weather evolves. Because of that, "We tend to be a little dry in our forecasts," he continued, and in the summer, atmospheric convection tends to be highly irregular.

Developing accurate, durable nowcasts won't be easy, Bradshaw acknowledges: "We've got a challenge facing us."

###



Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.