The Repository Hook Echo

Tuscaloosa 2011: 13 Minutes Missing From the Radar

October 22, 2026 · Runtime 15:13 · Watch on YouTube

On 27 April 2011 an EF4 tornado crossed Tuscaloosa and the edge of Birmingham, Alabama, on the ground for 80.68 miles; the National Weather Service counts 65 dead. The one radar that watched it from 70 km away has a single hole in its archive that day: 13 minutes, from 5:42 to 5:55 PM CDT, over the stretch through Concord and Pleasant Grove. We listed every scan it saved and measured what fell inside the gap. National Guard aerial video shows the path.

Chapters

  1. 0:00 Tuscaloosa from the air
  2. 0:19 The 13-minute hole
  3. 1:34 One storm, 3:38 PM
  4. 2:04 4:43: the tornado begins
  5. 3:13 A radar 70 km away
  6. 6:30 Warning 114, 5:31 PM
  7. 7:37 Two scans missing
  8. 8:16 What the assessment says
  9. 9:14 What fell inside the gap
  10. 11:27 The warnings and the gap
  11. 12:04 The words that changed
  12. 13:45 Sixty-five, sixty-six, sixty-four
  13. 14:06 What the record supports
  14. 14:54 Next: Hurricane Andrew

Transcript

Tuscaloosa from the air 0:00

Tuscaloosa, Alabama, April 27th, 2011. Thirteen minutes of the radar record are missing. The tornado that crossed this city killed sixty-five people, by the National Weather Service's count, and stayed on the ground for eighty miles. One radar watched it nearly all the way.

The 13-minute hole 0:19

We listed every scan that radar saved that day, from seven in the morning to one the next morning, and found one hole. This is the scan that began at 5:42. The next one in the archive begins at 5:55. The warnings put the storm's speed at fifty-five to sixty miles an hour. Thirteen minutes at that speed is about twelve miles.

So this episode lays that evening against the clock. What the radar could see, and from how far. What the warnings said, and when. And whether the missing scans changed any of it. Part of that, the record answers. Part of it, it cannot.

The federal assessment of the April 2011 tornadoes counts three hundred and sixteen deaths in five states. Two hundred and thirty-four were in Alabama. The forecast office in Birmingham alone issued forty-nine tornado warnings on April 27th.

Storms that morning had already cut power in some areas, and it was still off when the afternoon storms arrived. By the end of the day, more than a hundred high-tension towers were down, mostly across north Alabama. Cell phone and radio towers had been damaged or destroyed.

One storm, 3:38 PM 1:34

One storm produced the tornado in this story. The Birmingham office issued its first tornado warning for it at 3:38 that afternoon. Warnings for the same supercell stayed in effect until eight o'clock that night.

At 3:55, the office used the words Tornado Emergency for this storm for the first time, for the towns of Geiger and Panola, in the west of the state. That was another tornado. It was an hour and a quarter before the Tuscaloosa tornado reached the city.

4:43: the tornado begins 2:04

The Tuscaloosa tornado began at 4:43, in Greene County, near the town of Mantua, according to the Birmingham survey. Four minutes later, at 4:47, Tornado Warning number one hundred and nine went out for Tuscaloosa County.

The warning named Tuscaloosa, Northport, Holt, and Bryant-Denny Stadium. It put the storm near Mantua, moving northeast at fifty-five miles an hour. And it described the threat as a severe thunderstorm capable of producing a tornado.

The federal assessment later flagged that kind of wording. It found that several offices used the generic phrase, radar indicated, despite evidence of a tornado in progress. Birmingham was one of them.

At 4:58, an update used the words Tornado Emergency, and named Tuscaloosa. Radar and storm spotters were tracking a large and extremely dangerous tornado, it said, and it gave times. Northport around 5:05. Tuscaloosa around 5:10. At 5:08 another update reported visual confirmation of the tornado on the ground.

A radar 70 km away 3:13

The radar behind those words stands south of Birmingham. Tuscaloosa is about seventy kilometres from it, forty-four miles. It scans the sky in volumes, one about every four and a half minutes.

We computed where its lowest beam crossed the tornado over Tuscaloosa. The centre of the beam passed about one kilometre above the ground, and the beam itself was about one point two kilometres wide. The tornado, at its widest in Tuscaloosa County, was one point six.

The tornado was coming toward the radar. At Mantua, where it began, it was a hundred and ten kilometres away, and the centre of the lowest beam was nearly two kilometres up. Over Tuscaloosa, seventy-one kilometres. Near Pleasant Grove, about forty, with the centre of the beam a little over half a kilometre above the ground. The closest it came to the radar was around the time of the gap.

It was an older radar, without dual polarisation. It could say how much rain, hail or debris was out there, and how fast it was moving toward or away from the antenna. It could not say what the objects were. And its velocity, as recorded, tops out at thirty metres per second, then folds back.

Researchers later unfolded those velocities. At 5:19, across the circulation, one side was moving away at fifty-seven metres per second and the other toward the radar at seventy-two. A difference of a hundred and twenty-nine, which the authors call very high for a beam that far away.

On the reflectivity scan, the strongest echo near the tornado's path was sixty-three dBZ at 5:10, and sixty-eight and a half in the next scan. That was the moment it crossed Tuscaloosa. Without dual polarisation, the radar could not say whether it was seeing hail, rain, or debris.

The segment through Tuscaloosa County was forty-one point seven nine miles long. The national archive counts forty-four direct deaths in it, and eight indirect. Eight hundred people were injured.

Of the forty-four direct deaths in Tuscaloosa County, the national archive's fatality records list twenty-two in permanent homes, thirteen in other permanent structures, eight in trailers, and one in a vehicle. All twenty in Jefferson County were in permanent homes. Across the outbreak, the assessment says, twice as many people died in mobile homes as in conventional homes, and officials could not determine the circumstances of more than half of the deaths.

The survey rates the tornado EF4, with peak winds estimated at a hundred and ninety miles an hour. Nobody measured that. It is an estimate from what the wind did to buildings and trees.

In all, the track was eighty point six eight miles. Six in Greene County, rated EF2, up to four hundred and fifty yards wide. Forty-one in Tuscaloosa County, EF4, up to seventeen hundred and sixty. Thirty-two in Jefferson County, EF4, and at the Interstate 65 crossing, twenty-six hundred yards. A mile and a half. A research paper puts the damage track at a hundred and thirty kilometres, two point three at its widest.

Warning 114, 5:31 PM 6:30

At 5:31, a new warning, number one hundred and fourteen, went out for Jefferson County. Storm spotters and the Weather Service were tracking a large and extremely dangerous tornado on the ground, it said, near Brookwood, twelve miles east of Holt.

At 5:36 the updates named Weller, North Johns, and McCalla. At 5:42 they named Hueytown, Bessemer, Fairfield, and the city of Birmingham.

In the scan that began at 5:42, the strongest echo near the track sits at sixty-nine dBZ, about thirteen kilometres southwest of Hueytown. That is the last scan before the gap.

This is how that scan looks. A hook of rain wraps around the southwest side of the storm, and at its tip sits a small knot, brighter than anything around it. Past sixty-five dBZ, the scale turns magenta. The assessment says forecasters read debris ball signatures on radar that day. On this radar, a knot like this could be hail, or it could be debris. The radar cannot say.

Two scans missing 7:37

This radar finished a volume every two hundred and seventy-five seconds, a median over eighteen hours. After 5:42 the next one should have come at about 5:47, and then at 5:52. Instead, the next begins at 5:55. Thirteen minutes. Two scans missing.

A scan is a slice in time. Between two slices, a hook can tighten, a core can grow, a tornado can turn. At one scan every four and a half minutes, a forecaster sees the storm in slices that thin. In this gap, the slices were nearly three times as far apart.

What the assessment says 8:16

The federal assessment mentions it in one sentence. The radar in Birmingham went down, it says, and the office's engineering technicians restored it after approximately ten minutes. It gives no time, and no cause.

The assessment says two of the region's radars failed that day. The other, at Hytop in northeast Alabama, lost communications at 5:15 PM and did not come back for the rest of the event.

Another passage adds that when the forecasters in Birmingham had to take shelter, the backup office in Georgia had trouble displaying this radar's low-level data, and could not take over. It issued a warning on Birmingham's behalf.

We cannot tell you the cause. We read every warning and update the office issued that evening, and none of them says more. We cannot prove that our hole is that outage. It is the only hole in the archive that day, and the assessment describes only one.

What fell inside the gap 9:14

Here is what fell inside the gap. In the scan before, the core of the echo is near North Johns. In the scan after, it is beside Pleasant Grove. Between them, nineteen point six kilometres. Twelve miles, in thirteen minutes.

That is fifty-six miles an hour. Between the other scans, from 4:43 to 6:09, the core moved at forty-five to seventy-eight miles an hour, a median of fifty-seven. The straight line between the two scans is what the warnings said it would do.

A word about method. For every scan we took the lowest tilt, looked within eight kilometres of the line the survey draws between its start and end points, and kept the strongest echo. That is the core of the echo, not the tornado itself, and the survey line is straight where the real track wiggles. That is why the speed is a range, not one number.

The survey narrative for Jefferson County says at least five people died in Concord, and at least eleven in Pleasant Grove. One more died at McDonald Chapel, and one in Smithfield Estates. Concord and Pleasant Grove both lie on that line.

At 5:51, in the middle of the gap, the office put out another update. It places the tornado where the 5:42 scan had it: near North Johns, seven miles southwest of Hueytown. At 6:00 the next one says near Minor and Fairfield, moving northeast at sixty miles an hour.

The scan that begins at 5:55 shows the core of the echo beside Pleasant Grove, at seventy-two dBZ. Here is the whole series. Fifty-seven at 4:43, when the tornado began. Sixty-three over Tuscaloosa. Sixty-nine in the last scan before the gap. Seventy-two in the first scan after it. Fifty-eight at 6:14, when it lifted. The highest value of the series is on the far side of the hole.

The scans resume every four and a half minutes. At 6:14, the survey says, the tornado lifted just west of Alabama Highway 79, about two miles north of Tarrant.

The warnings and the gap 11:27

Now put the warnings next to the gap. The warning that covered Pleasant Grove went out at 5:31. At 5:42, the update named Birmingham. The first scan after the gap has the core beside Pleasant Grove at 5:55. So the warning was out roughly twenty minutes before the tornado arrived. The gap fell inside that time, not before it.

By the federal numbers, the warnings were early. Across the area the assessment studied, the mean lead time was twenty-two minutes. In Tuscaloosa, the watch had been in effect for more than three hours.

The words that changed 12:04

But the words changed from one product to the next. The assessment found that Tornado Emergency was used inconsistently from office to office, and from one product to another. Often it appeared in follow-up statements, when it was not in the original warning. That is what happened here. The warnings for Tuscaloosa and Birmingham did not contain the words. The updates after them did.

And people had heard warnings before. In the five years before 2011, between seventy-seven and eighty-six percent of the Birmingham office's tornado warnings, year by year, turned out to have no tornado.

Across three hundred and three warnings in the outbreak area, the assessment puts the probability of detection at eighty-nine percent, and the false alarm ratio at forty-nine percent. That is far below the Birmingham office's five-year record. Whatever people remembered, that day the warnings were right more often than not.

In Tuscaloosa, the assessment says, police cars drove up and down the streets with bullhorns. The Birmingham office also phoned the University of Alabama's emergency manager to say the campus would be struck. The university opened all its shelters, put warnings on building signs and buses, and used police loudspeakers.

Many people the assessment interviewed said they did not know where a community shelter was. A Tuscaloosa resident put it this way: we didn't know where to go. It was too wide and too strong.

Elsewhere in the state, in Hackleburg, about thirty-four people squeezed into a county storm shelter designed for eighteen. All thirty-four survived, the assessment says.

Sixty-five, sixty-six, sixty-four 13:45

Then the count. The Birmingham office says sixty-five people died in this tornado. The federal assessment's table says sixty-six. The national archive lists sixty-four direct deaths, and eight more, indirect, in Tuscaloosa County. None of these documents explains how the sixty-five was built.

What the record supports 14:06

Here is what the record supports. The warnings for Pleasant Grove and Birmingham were out before the radar lost its scans. What the gap took was the instrument's own record of the stretch where most of Jefferson County's deaths happened. And why it went missing is not in any public document we could find.

And here is what we do not know. Why the radar stopped. What those two scans would have shown. How the sixty-five was counted. The record has no answer to any of the four, and we would rather say so than guess.

So the question stays open. If you know the maintenance record of the Birmingham radar for that afternoon, we would like to read it. The tornado stayed on the ground for ninety-one minutes, and the record of it has thirteen minutes of silence in the middle.

Next: Hurricane Andrew 14:54

Next, a hurricane that was rated one category for ten years, and then changed. Hurricane Andrew, in August 1992, was reanalysed from the instruments' own records and raised to a Category Five. That is the next episode. Andrew.

Description and sources

The Birmingham radar (KBMX) saved 234 volumes between 7 AM and 1 AM, one about every 4.6 minutes, with exactly one exception: 22:42:40 to 22:55:39 UTC. The federal assessment says the radar 'went down' and was restored after about 10 minutes, with no time and no cause; we could not find the cause in any public document and say so. In the scan before the gap the echo core sits near North Johns; after it, beside Pleasant Grove, 19.6 km (12.2 mi) away: 56 mph, in line with the 45 to 78 mph between the other scans. The warning for Jefferson County went out at 5:31 PM and the update naming Birmingham at 5:42, so the gap fell inside the warning, not before it. The radar was about 71 km from Tuscaloosa, with its lowest beam about 1.0 km above the ground and 1.2 km wide; its velocity folds back at 30 m/s, and it had no dual polarisation. The record also disagrees on the toll: 65 (NWS Birmingham), 66 (federal assessment table), 64 direct plus 8 indirect (NCEI).

PRIMARY SOURCES

Hook Echo reads a disaster through the instruments that recorded it: radar, sensor, survey and report, with the primary sources above.

About the material. Video: Alabama National Guard (131st Mobile Public Affairs Detachment, Tuscaloosa aerial footage; Alabama National Guard search and recovery footage; 81st Readiness Division package) released through DVIDS, public domain. Every radar image is rendered by Hook Echo from NOAA's public archive of KBMX Level II data (NEXRAD, unidata-nexrad-level2 on AWS), in the National Weather Service colour scale. The gap, the echo-core positions and the beam geometry are our own measurements from those files. Every diagram is drawn in code from the sources listed. Nothing in this video is AI-generated imagery.

#HookEcho #Tuscaloosa #Tornado