The series of tornadoes that killed at least six in Oklahoma and raised havoc in other parts of the Central Plains April 14-15 was preceded by a rare early, high-risk warning.
For the second time in U.S. history, the National Weather Service’s Storm Prediction Center issued a high-risk warning more than 24 hours in advance of the more than 75 tornadoes that later hit Oklahoma as well as Kansas, Iowa, Nebraska and north Texas.
High-risk warnings are issued on average two to five times a year but not with a 24-hour lead time. So what was different about these events? The storm was “well forecast” by the computer models as potentially dangerous and forecasters have been gaining confidence in forecasting every year, according to Russ Schneider, director of the National Oceanic and Atmospheric Administration’s Storm Prediction Center in Norman, Okla.
“The conditions were favorable, based on computer model guidance, for a potentially major tornado outbreak,” Schneider said. “When there is sufficient confidence, what we want is to state what we’re seeing in the data and what our expert forecast is so that people can prepare.”
The increased confidence comes from: a better understanding of science; improved observations, including satellite, radar and other observing systems; and greatly improved numerical forecast models and numerical modeling or ensemble forecasting. The latter is numerical modeling on a smaller scale (down to individual thunderstorms) that involves taking numerous computer forecasts and determining whether the storm will be isolated and a potential super cell or a solid line squall. Solid line storms tend to have less potential to become tornadoes than the isolated super cells.
“This will be an increasing trend within severe weather forecasts where we’ll be able to give the emergency management community a greater range of information on what the worst possible outcomes are and what some of the less severe outcomes are and the chances of each,” Schneider said.
He said most communication to the public will be the day of the event, but FEMA will continue to receive forecasts up to three days and sometimes four days in advance of storms. “We certainly want to stay responsible in our communication both to the emergency management community and to the public. We certainly don’t want to ‘over warn’ so to speak.”
The early spring storms like the one that prompted the early warning on April 13, are more likely candidates for early warnings than later storms when the jet stream weakens, Schneider said.
“We certainly won’t always be able to makes these kinds of decisions, particularly with a high-risk storm so many days in advance, although we are able to identify areas that may prove to be troublesome many days in advance, particularly very early in spring.”
Although Schneider sees a trend of more early warnings, he said predicting the future is a difficult business. “We don’t actually know the complete state of the atmosphere at any one moment with certainty, so we can’t calculate the future with certainty.”