image: A tornado touches down near Dodge City on May 24, 2016. The current lead time on issuing tornado warnings is about 15 minutes. More lead time from the Warn-on-Forecast system could save lives and property across the Great Plains and beyond, University of Kansas researchers found.
Credit: Lane Pearman, Wikimedia Commons
LAWRENCE — Researchers at the University of Kansas have shown the National Severe Storms Laboratory’s Warn-on-Forecast System(WoFS) has potential to help weather forecasters issue warnings to emergency managers and the general public well before tornado formation. Their study appears in the peer-reviewed journal Weather and Forecasting.
The current lead time on issuing tornado warnings is about 15 minutes. The difference could save lives and property across the Great Plains (the classic “Tornado Alley”) and other areas like the tornado-prone Southeast.
“Right now, when we do any kind of warnings or forecasts of tornadoes and severe weather, we use a warn-on-detection paradigm,” said co-author David Rahn, professor of geography & atmospheric studies at KU. “If we see rotation in real time on radar, or if a storm spotter reports that a tornado is happening, that is when warnings are issued. Typically, we know severe weather is coming days in advance. Three or four days out, we may know Kansas is going to have some sort of severe weather outbreak. As we get closer, we refine the forecast as models update. We narrow things down to watches, but we do not warn until danger is imminent or occurring.”
The idea behind the WoFS system is to increase lead time by not just warning on detection, but warning on predictions and forecasts of the event, according to Rahn. Today’s tornado warnings rely on radar data and regional spotters. By contrast, WoFS synthesizes observational data in real time and runs high-resolution simulations to gauge uncertainties and produce guidance for tornado and severe weather threats well before they form.
“The goal is to increase lead times well beyond the typical 15 minutes,” he said. “In some cases, this system can push lead times up to an hour before the event actually occurs.”
As of today, the WoFS system is in the experimental phase. Rahn and his colleagues analyzed 41 supercells resulting in reports of tornadoes of at least EF1 (on the Fujita scalefor tornado strength) or non-tornadic “significant hail” on the Great Plains. They aimed to determine the specific characteristics that could help determine if a storm resulted in a tornado or was just a dangerous supercell with large hail.
“The environment has to be just right to produce a tornado,” Rahn said. “One major finding is that the size and extent of pockets of environmental factors are better indicators than just maximum values. If you only look at a single maximum value, you may not see differences between tornadic and non-tornadic storms. But if you look at larger, well-defined areas of enhanced storm-relative helicity, those are markers of better-developed storms, which have a higher chance of tornado genesis.”
The KU researcher said storm-relative helicity is a low-level corkscrew motion fed into the storm.
“It’s like a spiraling football that, when it encounters the storm updraft, gets tilted vertically,” Rahn said. “That vertical rotation is what tornadoes tap into. The stronger and more extensive that motion is, the more likely a tornado becomes.”
Other factors that increased the likelihood of tornado genesis include lower cloud bases, stronger low-level wind shear and small pockets of enhanced rotation near the storm’s inflow region.
Because WoFS models these exact parameters at very fine spatial scales, they become signals of tornado formation up to an hour before a tornado forms.
The end users of WoFS will be forecasters with the National Weather Service, but the system also will be broadly available for free to meteorologists and the public.
“It’s primarily developed for National Weather Service forecasters, but it is cloud-based and accessible in real time,” Rahn said. “The National Weather Service is a public service, so its products are typically public and free. That is one of the benefits of having a National Weather Service that is not monetized and is for the public good.”
Rahn’s co-authors were lead author Jerod Kaufman, a KU graduate student now working for the National Weather Service; David Mechem, KU professor of geography & atmospheric science; Patrick Burke of the National Severe Storms Laboratory; Montgomery Flora of the National Severe Storms Laboratory; and Corey Potvin of the University of Oklahoma and the National Severe Storms Laboratory.
“The lead author was my master’s student, and this was his thesis project,” Rahn said. “He now works at the Wichita National Weather Service, which is a nice example of research transitioning into operations. This is cutting-edge research that integrates operational meteorology into the classroom. It shows how graduate research can transition directly into real-world forecasting and public service.”
As the WoFS system continues to undergo evaluation in live-forecasting scenarios, the hope is it will eventually save lives and property. Warnings can be issued to the public with more time for people to seek safety, while emergency services will have more opportunity for preparation.
“One major benefit of longer lead times is informing emergency managers at the county level,” Rahn said. “With more time, they can position resources, coordinate hospitals and prepare search-and-rescue operations. The mission of the National Weather Service is to protect lives and property. An hour of lead time versus 15 minutes is extremely useful for everybody.”
Journal
Weather and Forecasting