Re­searchers were able to pre­dict vi­o­lent storm 90 min­utes be­fore it hit

Fore­cast sys­tem be­ing im­proved

Northwest Arkansas Democrat-Gazette - - NORTHWEST ARKANSAS - KELLY P. KISSEL

Re­searchers run­ning highly de­tailed sim­u­la­tions us­ing satel­lite images, radar and ground-based weather sta­tions were able to pre­dict a spe­cific vi­o­lent storm 90 min­utes be­fore it hit a western Ok­la­homa town and killed a man two months ago.

The Na­tional Se­vere Storms Lab­o­ra­tory in Nor­man, Okla., said Fri­day the still-ex­per­i­men­tal fore­cast sys­tem could give emer­gency plan­ners up to three hours’ no­tice of up­com­ing bad weather, in­clud­ing tor­na­does. Such no­tice is key for hos­pi­tals, schools and other places where large crowds gather.

“The the­o­ret­i­cal ground­work was laid in the 1980s and 1990s,” said Pa­trick Skin­ner, a re­search me­te­o­rol­o­gist with the Univer­sity of Ok­la­homa’s Co­op­er­a­tive In­sti­tute for Me­soscale Me­te­o­ro­log­i­cal Stud­ies. “When this project started in the late 2000s, there was some worry this wasn’t fea­si­ble. We are be­com­ing more con­fi­dent that we can do it.”

Af­ter gen­eral fore­casts showed that se­vere weather could de­velop in the east­ern Texas Pan­han­dle or western Ok­la­homa on May 16, Na­tional Se­vere Storms Lab­o­ra­tory me­te­o­rol­o­gists in cen­tral Ok­la­homa ran data through 36 sim­u­la­tions, tweak­ing data for each sce­nario.

“On May 16, we had a large num­ber that all pre­dicted the same evo­lu­tion of a su­per­cell. That gave us a lot of con­fi­dence,” Skin­ner said.

For the first time, the Na­tional Weather Ser­vice used a “Warn on Fore­cast” to no­tify emer­gency plan­ners around Elk City, Okla., that tor­nado warn­ings would likely be is­sued later in the day. Todd Lind­ley, the sci­ence op­er­a­tions of­fi­cer at the Nor­man weather of­fice, said the spe­cial weather ad­vi­sory was es­sen­tially “fore­cast­ing that we would be is­su­ing warn­ings.”

Beck­ham County, Okla., emer­gency man­agers started sirens 30 min­utes be­fore the tor­nado hit. Cur­rently, most com­mu­ni­ties re­ceive about 13 min­utes’ no­tice that a tor­nado is on the way.

“This could open up a new door for a very de­tailed in­creased lead time,” said Steve Koch, direc­tor of the Na­tional Se­vere Storms Lab­o­ra­tory. “That’s the di­rec­tion we’re head­ing.”

Us­ing com­puter mod­els to pre­dict weather isn’t new. Gen­eral fore­casts sent out ev­ery day are cre­ated with the help of com­put­ers, but de­tailed fore­casts such as the one at Elk City re­quire the re­searchers to run mil­lions of cal­cu­la­tions in a short pe­riod of time.

In ad­di­tion to images from satel­lites, radar units and weather sta­tions, “the one thing we need is com­pu­ta­tional power,” Skin­ner said. In their work, the lab uses about 1.2 ter­abytes of band­width a day to en­sure that mod­el­ing is done in a timely man­ner.

Af­ter all, if it takes 12 hours to run the var­i­ous sim­u­la­tions, the storm will be gone be­fore any pre­dic­tion comes out.

“We aim to do this in a half-hour,” Skin­ner said.

The May 16 tor­nado formed from a clas­sic su­per­cell formed in the rel­a­tively des­o­late Plains. Fore­cast­ing in other ar­eas of the coun­try might not be so easy be­cause storms can form un­der sev­eral con­di­tions such as cold fronts and sharp con­trasts between dry and hu­mid air.

Af­ter this year’s ex­per­i­ment pe­riod was over, a tor­nado hit Jones­boro in north­east­ern Arkansas on July 3. It was never de­tected ahead of time.

“We had a lit­tle spin-up tor­nado that never showed up on the radar,” said Jeff Pres­ley, direc­tor of the E911 sys­tem in Jones­boro. “It got re­ally dark. We an­tic­i­pated roads get­ting flooded, then peo­ple started call­ing say­ing there were power out­ages, and trees down, then a roof off an apart­ment com­plex.”

Such pop-up storms will of­ten elude fore­cast­ers.

“Smaller storms like the 30-sec­ond tor­nado, that is an ex­tremely big chal­lenge,” said Koch, the Se­vere Storms Lab­o­ra­tory direc­tor.

At a min­i­mum, the re­searchers said they ex­pect to do a bet­ter job read­ing the at­mos­phere ahead of a storm.

“It’s not a game changer at this point be­cause the sys­tem is still quite ex­per­i­men­tal, but it is a big step,” Lind­ley said.


A home sits de­stroyed in Elk City, Okla., on May 17 af­ter a tor­nado swept through the area, killing one. Re­searchers at the Na­tional Weather Cen­ter in Nor­man, Okla., said Fri­day it was able to tell 90 min­utes be­fore a spe­cific storm cell would cause sig­nif­i­cant weather in the area.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.