Hailstones and Milestones

April 29, 1995. Around 10 p.m. a vicious thunderstorm rolls into Dallas-Ft. Worth airport. Marble to softball-sized hail pelts planes on the ground. More than 60 commercial jets must be removed from service for repair. The direct insured loss exceeds $20 million, and further loss from canceled flights over the next few weeks runs to $300 million.

What if? The question drives Kelvin Droegemeier's waking days. What if there had been four hours warning? Incoming flights could have been diverted, and with normal flight turnover, about two hours, virtually all the planes would have been removed from harm's way, cutting the airlines' loss and avoiding much of the subsequent inconvenience to travelers. With countless other severe storms as well, including the fierce tornadoes that rampage the Great Plains each spring, better forecasts could greatly reduce property damage and potentially save lives. Since 1986, severe weather has racked the insurance industry with unprecedented losses, and industry studies indicate that more than $14 billion a year could be saved with better forecasts.

Current forecasting gives about 30 minutes warning for storms, with fairly imprecise information about extent and severity. Droegemeier, who directs the Center for Analysis and Prediction of Storms (CAPS) at Oklahoma University, knows it's possible to do better. He and his colleagues are developing new storm-prediction capability that is setting milestones in the field. In 1995, using the CRAY T3D at Pittsburgh Supercomputing Center, their state-of-the-art computer model, the Advanced Regional Prediction System (ARPS), successfully predicted the location and structure of individual storms six hours in advance, the first time anywhere this has been accomplished. In 1996, ARPS did better yet, successfully forecasting the position and timing of storms seven hours in advance, even though the storms hadn't yet begun to form when the model was run.

Zeroing in on Stormy Weather

The weather reports we watch on TV derive from computer models at the National Centers for Environmental Prediction that predict atmospheric structure over the continental United States as often as every three hours. The local forecaster extracts from these models to show a regional map, covering perhaps several states, with predictions that weather the next day will be rainy, cloudy, sunny, etc.

In contrast, Droegemeier and his colleagues at CAPS work on storm-scale forecasting, a much tighter focus -- a few miles square in space and about 15 minutes in time -- that corresponds to the scale on which individual storms evolve. "What we're getting down to," says Droegemeier, "is to say that over Pittsburgh this afternoon from 3:30 to 3:50 there will be a thunderstorm with winds of 30 miles per hour, golfball-sized hail, two-and-a-half inches of rain, and by 3:50 it will be gone. And to give you that forecast six hours in advance."

Miles to Go Before They Sleep

CAPS' goal is to develop storm-scale forecasting to the point where it can be turned over to the
National Weather Service early in the next century. Notable among the challenges they have overcome since starting in 1988 is gathering the input data necessary to run a storm-scale model. CAPS innovations now make it possible to deduce all the initializing data -- pressure, temperature, wind speed and more -- from Doppler radar.

Since 1993, CAPS has carried out experiments during Oklahoma's spring storm season to see how ARPS works in an operational setting. Initializing data each morning feeds the computer model running in Pittsburgh. The output in turn feeds back to forecasters in Norman, Okla. Results have been encouraging. In 1993, running on the CRAY C90 with limited data and a limited version of the model, forecasters used ARPS information in an official National Weather Service forecast.

ARPS vs. Reality: June 8, 1995
The ARPS forecast (left) created at 1 p.m. for conditions at 7 p.m. compares well with an actual radar image (right) at 7 p.m. Color indicates rainfall intensity, increasing from light blue to pink. "The model did a remarkably good job," says Kelvin Droegemeier, "getting these storms in just about the right location, especially with most intensely rotating storms in the northeast Texas panhandle. And it predicted the north-south extent of the storm line up to Kansas. This represents a tremendous success."

In spring 1995, CAPS used more ARPS capability, and for the first time exploited parallelism on the CRAY T3D. "The T3D's distributed, globally addressable memory," says Droegemeier, "gave us the ability to run at storm scale over a big area at high resolution." The model did remarkably well, especially considering that initializing data was relatively crude -- lacking the resolution and accuracy radar can provide. As a result of this success, American Airlines is contracting with CAPS to test the new prediction technology as a "smoke alarm" for airports.

ARPS vs. Reality: May 24, 1996
The ARPS forecast (left) created at noon Oklahoma time for conditions at 5 p.m. compares well with an actual radar image at 5 p.m. Color indicates rainfall intensity, increasing from light blue to pink. The green square shows a three-kilometer fine-grid nested within the larger nine-kilometer grid of the forecast area. ARPS centers the small grid in the area where storms are likely to develop based on the large-grid forecast. Even though no storms were present in the forecast area at noon on May 24, ARPS successfully predicted the timing and location of the storm line that developed later that day.

In 1996, the spring operational test incorporated several model improvements, including more detail at ground level, that resulted in better temperature forecasts and more precision in predicting the time when storms would develop. With these improvements, ARPS successfully forecast storms on eight of the 10 days when they occurred, an unprecedented success rate. On several of these stormy days, data that fed the model showed no convection, the earmark of storm conditions, and the model still correctly forecast that storms would develop seven hours later. Just as importantly, the model also correctly predicted no storm on a day when conditions indicated a strong likelihood.

With this growing success record, CAPS researchers see their goals as within reach, although challenges remain, perhaps none bigger than the extreme computational challenge built into their work. "In meteorology," says Droegemeier, "getting results quickly is essential. If you can't predict the weather significantly faster than it evolves, the prediction has no value. If you're going to create a four-to-six hour forecast, you better do it in half an hour. Parallel computing and high-performance networking are crucial."

In 1996, it took about 100 minutes to run a seven-hour forecast using 256 processors of the T3D. Fine for this stage of development, says Droegemeier, but too long for real-world forecasting. The significantly improved performance of the CRAY T3E will further boost ARPS ability to forecast storm-scale phenomena. "We can do better," says Droegemeier, "and we will."

During spring 1995 and again in 1996, the Center for Analysis and Prediction of Storms at the University of Oklahoma used the Cray T3D to produce a daily weather forecast for parts of Oklahoma. PSC's visualization group processed data from selected forecasts into 3-D animations. The display shows a 152 x 152 kilometer area 35 kilometers in altitude viewed from the southeast. The storm structure shows regions of high water-density (white) with columns of high vorticity (red) that potentially spawn tornados. The streamlines, colored according to temperature, show ground-level wind velocity.


This animation, created with ARPS storm forecast data, shows development of a thunderstorm over Oklahoma. It covers a surface area 67 kilometers square extending 17 kilometers in altitude. The white isosurface shows high concentrations of combined cloud and rain water. The red isosurfaces represent columns of high vorticity, which have the potential to spawn tornadoes. The blue ribbons show wind velocity at two planes, about 600 and 4,800 meters above ground level.

Researchers: Kelvin K. Droegemeier, University of Oklahoma at Norman.
Hardware: Cray C90, Cray T3D
Software: Advanced Regional Prediction System (ARPS)
Keywords: weather, forecasting, storm, storms, storm prediction, storm warning, tornado, tornadoes, supercells, weather models, Advanced Regional Prediction System (ARPS), CAPS, Center for Analysis and Prediction of Storms.

Related Material on the Web:
More information about the Center for Analysis and Prediction of Storms
More information about the Advanced Regional Prediction System (ARPS)
Projects in Scientific Computing, PSC's annual research report.
On the Horizon: Accurate Storm Forecasts
PSC Wins Computerworld Smithsonian Award
Storm Forecasting Wins Discover Award

References, Acknowledgements & Credits