Bad as it was Monday in Oklahoma — and the danger persists — tornado researchers and operational meteorologists say it could have been worse. New levels of computer modeling accuracy are paying dividends in predicting these small, violent storms — perhaps the most difficult of weather forecasting problems.
"The hallmark was the National Weather Service forecasting warning process," Michael Foster, meteorologist-in-charge of the NWS forecast office in Norman, Oklahoma told a news conference Tuesday. The agency's sophisticated computer simulations at the National Environmental Modeling Center in Camp Springs, MD, provided "very specific long-lead forecast information starting about 7 days ago."
By Monday morning, several hours before any tornado had formed, regional models at the NWS Storm Prediction Center in Norman, and in Foster's local forecast office, were simulating a particular pattern of "radar reflectivity" that kept showing up.
"It was the fact that we had been doing that for several runs yesterday, and we kept getting the same general solution, that led us to put statements out to the public, and to the media, that we were anticipating very strong storms with tornadoes in the metropolitan area at rush hour."
Foster showed these two panels of simulated and observed tornado outbreaks that illustrate the accuracy of the forecast models. The panel on the left shows the simulated "radar reflectivities" from a model run at 2 p.m. The panel on the right is the actual radar image taken at the time of the storms, 5 p.m. The model was run over three hours before that time. "The similarities are striking," said Foster.
"The kind of thing we did yesterday is a confidence builder," he said, encouraging an experimental project at NOAA called "Warn on Forecast," which is designed to develop the ability to provide specific tornado forecasts an hour in advance. "It gives us confidence that we can see (in the models) the developments that may take place a few hours later."
Aside from the early warning capacity made possible by the models, Foster said the outbreak could have been worse:
If everything could have developed 15 miles west of where they did, especially between 4 and 6 o'clock in the evening, the big, well-developed tornadoes would have been on the ground and coming through the metropolitan area. As it turns out, for the most part, what I think we eventually will rate as the strongest tornadoes occurred in more rural areas… Even with the impacts it already had, it was a very close call for something much worse than that.
Images: National Weather Service