The worst tornado outbreak to strike the Deep South in nearly two years came to an end yesterday after killing almost three dozen people and damaging hundreds of buildings. Meteorologists for the National Weather Service did an incredible job predicting the outbreak days before it happened, and they're directly responsible for saving countless lives.

Six Days' Notice

Residents in the areas affected by the worst tornadoes on Sunday and Monday received a heads-up six days before each of the respective outbreaks. Meteorologists' abilities to forecast (and detect) tornadoes has increased dramatically over the past few decades, and the advances were put on full display over the last week.

April 27 — Arkansas

The Storm Prediction Center outlined central and western Arkansas as at risk for severe weather on Tuesday, April 22, saying that "potentially significant supercell development with very large hail and tornadoes can be expected" if weather model guidance was correct.

Models consistently showed a major weather outbreak occurring over the central Plains — including most of Arkansas — on Sunday and refinements to the scope and duration of the outbreak were made with each subsequent forecast. Forecasters had enough confidence by early Friday morning, a full three days before the outbreak, to issue a "moderate risk" (4/5 risk) for severe weather. Issuing a moderate risk three days before a predicted severe weather event is exceedingly rare.

Forecasters finally pulled the trigger and issued a rare "high risk" (5/5 risk) the afternoon of the outbreak when it became apparent that the atmosphere was conducive to extremely strong, long-lived tornadoes.

April 28 — Alabama/Mississippi

By Wednesday, April 23, forecasters had seen enough consistency in the models to predict that Sunday's severe weather outbreak would continue into Monday further to the east. Their "bullseye" hardly budged over the six days they had the outbreak in the forecast. Again, they issued a "moderate risk" for severe weather early Saturday morning, three full days before the outbreak started.

Their Forecasts Verified Really Well

Overall, the forecasts issued by the Storm Prediction Center were pretty well aligned with the worst of the tornadoes in Arkansas and "Dixie Alley," a nickname for parts of northern/central Alabama and Mississippi that are susceptible to violent tornadoes.

Survey teams are still trying to determine the exact paths and strength of each tornado, and many of the tornado reports on these maps are for the same tornado since they traveled such long distances.

April 27 — Arkansas

The tornado outbreak in Arkansas was thankfully limited to one violent tornado that stayed on the ground for about two hours and tore a few dozen mile path through the center of the state. The supercell that spawned the tornado that tore through Mayflower and Vilonia formed right on the edge of the "high risk" area outlined by the SPC for violent tornado formation. Surveyors gave the tornado, which killed 15 people, a preliminary EF-3 rating on the Enhanced Fujita Scale, reserving the right to raise the final rating once they complete their tour of the damage.

The only way the forecast "busted" is that the outbreak wasn't really much of an outbreak, as there were only two tornadoes in the state that day.

All of the other tornadoes that formed across the United States on Sunday occurred in areas where the SPC said there would be a slight risk (2/5 risk) for severe weather.

April 28 — Alabama/Mississippi

The violent tornado outbreak in Alabama and Mississippi was more prolific and damaging than the one that occurred in Arkansas the day before it, spawning over two dozen tornadoes across the region and killing nearly a dozen people.

All of the largest, strongest tornadoes on April 28 occurred in the area the SPC placed under a "high risk" (5/5 risk) for severe weather on Monday. Survey crews are working to determine the strength and paths of the tornadoes that tore through the region, but many of the storms were rated strong EF-3 and EF-4s on the Enhanced Fujita Scale.

April 29 — A Bust in Alabama and Mississippi

On Tuesday, the atmosphere was conducive to the development of more violent, long-track tornadoes over much of the same areas in Alabama and Mississippi that saw the intense weather on Monday. The SPC issued a moderate risk for severe weather in anticipation of the outbreak, but thankfully, it was a bust: nothing happened.

A complex of thunderstorms along the Gulf Coast — the same one that produced copious amounts of flooding after dropping over a foot of rain in many spots — pumped enough cool, stable air into interior sections of the Deep South that it prevented the instability necessary to form dangerous thunderstorms that could produce large tornadoes.

Tornado Emergencies and Debris Balls

During the strongest tornado ever recorded, forecasters decided that they needed to use language stronger than "tornado warning" to get the public to take the threat seriously and seek shelter from the mile-wide wedge bearing down on Moore, Oklahoma. They used the term "tornado emergency" in the warning, and that set a precedent for future tornado safety in the United States.

A tornado emergency is issued by the National Weather Service when a large tornado is spotted moving into a populated area. It differs from a tornado warning as it assures people that a tornado really is coming and that they need to seek shelter to survive.

The National Weather Service issued at least five tornado emergencies on April 27 and 28, all after spotters sighted large tornadoes moving towards towns and cities, or Doppler radar left no doubt that a tornado was present.

Debris Balls on Doppler Radar

Anyone who's watched severe weather coverage on television has heard the word "debris ball" at one point or another. The term refers to an area that shows up inside a thunderstorm on Doppler radar created by the radar beam reflecting off of the immense amount of debris swirling around inside a tornado.

Meteorologists have a new technology to help differentiate between heavy rain, hail, and tornado debris, called "dual-polarization." I wrote an extended post about how dual-pol data helps meteorologists confirm tornadoes by spotting debris, and this technology was used several times to confirm the existence of a tornado without anyone actually seeing it on the ground.

Forecasters issue a tornado warning either when a tornado has been spotted on the ground, or when Doppler radar indicates strong rotation that could produce a tornado. While Doppler technology allows us to see tornadoes that we never would have known about until they formed a few decades ago, its major downfall is the false alarm rate. Almost 75% of all tornado warnings issued are false alarms. National Weather Service offices are actively working to decrease the false alarm rate, but the nature of weather radars means that there will always be some degree of tornado warnings that don't verify with a tornado.

All in all, meteorologists did a great job with this tornado outbreak, and as technology advances, so too will their ability to more accurately predict future disasters.

[Images via AP / author's compilation of SPC forecasts / SPC / SPC / Gibson Ridge via NWS Huntsville]