Almost everyone now knows that an extremely rare pair of tornadoes formed in Nebraska yesterday, destroying one small town and heavily damaging another. The side-by-side tornadoes are so rare that there are hardly any records of them happening before. How did these rare twin twisters form?

Well, what happened yesterday is essentially a twister birth defect. We saw two large, independent tornadoes form from the same thunderstorm and track side-by-side for an exceptionally long time across northeastern Nebraska. Meteorologists are going to study this storm for a long time, and nobody has a definite answer as to exactly why it happened, but I think I've developed a pretty reasonable explanation for it.

To get a rough idea of what happened, you have to know the basic structure of a supercell. Below I've paraphrased from my longer post back in April titled "What is a supercell?"

Under normal circumstances during a severe weather outbreak, a thunderstorm can develop into a supercell with the right amount of wind shear and instability. A supercell is a rotating thunderstorm that's acts like an efficient engine that's able to last for hours and travel hundreds of miles.

Wind shear is the changing of both the speed and direction of wind with height. Strong enough wind shear between the lower levels and the upper levels can create a horizontal tube of rotation in the air above the surface (Figure A). If the atmosphere can produce enough instability make air rise rapidly (called an updraft) and form a thunderstorm, the updraft can bend this tube of rotation into an arch, with one side rotating clockwise and the other counterclockwise.

The counterclockwise tube of now-vertical rotation almost always wins out, and it becomes what's known as the storm's mesocyclone, or the broad area of rotation that keeps the storm alive. The mesocyclone is often visible in the classic supercells that form Plains states; they are the cause behind the immense, picturesque structures that most people are familiar with. It is from these mesocyclones that tornadoes form in a supercell.

In some cases where an intense tornado forms from a supercell, the twister can be given a "nudge" by another shower or smaller storm merging with the supercell. That is what I believe happened here. A heavy shower moving much faster than the supercell itself got sucked into the inflow where the tornado was located, "nudging" the storm into producing a second tornado a mile or two away from the one already on the ground.

Let's take a look at the storm yesterday, following it on step by step on radar. The order of the images will always first show the base reflectivity (precipitation) and then the base velocity (winds). On the base velocity image, red shows wind moving northwest while green shows wind moving southeast. Bright red and bright green colors close together indicate strong rotation and the location of the tornadoes.

3:01PM

The supercell formed all alone out in the middle of northeastern Nebraska, allowing it to fully utilize all of the unstable air around it. In its beginning stages, the system looked more like a disorganized cluster of storms, but it appears that it was really two supercells forming side-by-side.

3:08PM

Seven minutes later, it's clear that the system consists of two distinct supercells each producing rotation. At this point, the tornado cluster is still 30 miles away from Pilger.

3:37PM

Almost thirty minutes later, the two supercells have merged into one and the dominant supercell (the one to the southeast) prevailed. This sweep of the radar shows the storm getting its act together as it's about to drop the first tornado about ten miles southwest of Stanton.

3:50PM

This is where the storm structure starts to get tricky. At this point we had a fairly strong tornado on the ground just west of Stanton, producing enough damage for the debris to show up on radar. While this tornado moved north through the western edge of the storm, another hook started to develop on the southern end of the supercell. This would ultimately produce one of the two tornadoes that hit Pilger.

4:04PM

At this point, there are two tornadoes on the ground. This is not the "twin tornadoes" yet. This storm produced three different tornadoes by the time it hits Pilger. Tornado #1, the original tornado, is northwest of Stanton, rapidly moving north and dying. The hook rapidly developed Tornado #2 just southeast of Stanton, and this is the main circulation that hits Pilger.

Take note of the showers just south of the supercell. They were moving much faster than the supercell itself, and merged into the hook minutes before the twin tornado formed.

4:12PM

This is the point where the large, damaging tornado began to move into Pilger. The tornado is clearly visible on base reflectivity as the deep red debris ball, and this is corroborated by the rotation on base velocity overlapping the debris ball, as well as the correlation coefficient (third image) showing a dark blue dot. The dark blue dot shows that the objects the radar is detecting are all different shapes and sizes — debris.

4:15PM

This is the point where the twin tornadoes were on the ground in and around Pilger, producing the images that we're all familiar with by now.

4:21PM

This radar sweep was taken just after the tornadoes crossed the highway, and just after the soon-to-be-infamous video of those idiotic storm chasers very nearly driving into one of the tornadoes in order to get pictures of it.

4:37PM

Shortly after the tornadoes crossed the highway north of Pilger, they began to dissipate as the supercell attempted to reorganize itself and head off towards Sioux City, Iowa.

I am almost certain that the second, twin tornado would not have formed if the group of heavy showers south of the supercell hadn't got caught in the inflow and sucked into the hook where the main tornado was located.

Here's an animated radar image (it may take a moment to load) of the tornadoes as they hit Pilger. Note that the second tornado does not form until the showers get sucked into the supercell.

A similar event occurred during the Moore, Oklahoma tornado back in May 2013, except that the merging convection didn't produce a second tornado. The tornado didn't explode into a monster until a heavy shower got sucked into the inflow and helped to spin the tornado up even stronger than it had been beforehand. Here's an animated radar image (again, it may take a moment to load) of that tornado.

Many meteorologists are noting on social media that this event bears a striking resemblance to one of the only other instances anyone can think of something like this happening — the Palm Sunday Outbreak back in 1965. During that event, Dunlap, Indiana was struck by a "twin-funneled" F4 tornado and a person in the area snapped this iconic photo of the two funnels as they collectively killed dozens of people.

Dr. Ted Fujita, the world's foremost tornado scientist, wrote a paper (caution: PDF file) in which he noted that the funnels you see above were two circulations of the same tornado. We're looking at one tornado split into two different funnels. In Pilger, Nebraska yesterday, we saw two distinct, independent tornadoes form within a few miles of each other within the same thunderstorm.

It was an incredibly rare event that bore tragic consequences for a small town in Nebraska. Meteorologists are going to write papers on it for years to come, and the event will hopefully shine a new light on how storms are able to produce tornadoes.

If you were wondering, tornado #1 near Stanton was at least an EF-3, and tornado #2 (the one that hit Pilger) was at least an EF-4, according to the NWS in Omaha. Meteorologists are finalizing their reports and should have preliminary data released by the end of the week.

[Top image by Roger Hill, others via Google Maps and NOAA, all radar images via Gibson Ridge]

You can follow the author on Twitter and send him an email here.