Monday morning’s tornado outbreak across the Carolinas was one to remember. At least 16 tornadoes have been confirmed across South Carolina alone, a number that could grow as National Weather Service survey teams continue to inspect damage.

It was part of the same tornado outbreak that left 34 dead and spawned at least 105 tornadoes from the South to the Mid-Atlantic on Sunday and Monday.

Four of the tornadoes in South Carolina reached EF3 status on the 0 to 5 scale for twister intensity, with winds topping 136 mph. Nine deaths were reported as a result of storms across the Palmetto State.

But two of the EF3 tornadoes did something nobody was expecting and few have observed before: They combined, with one strong twister essentially getting absorbed by the other’s circulation. It’s an unusual event that has left many experts scratching their heads, but radar data could offer clues as to exactly what transpired.

An extreme event in the dead of night

Many of the tornadoes, especially across the Midlands and Piedmont, struck in the dead of night — their destruction cloaked in darkness. But Doppler radar served its role as a sentry, peering into storms at all hours. Some of what the radars illuminated was staggering.

“It looks like [this outbreak] is going to have the most significant tornadoes for us since 1984,” explained Frank Alsheimer, science and operations officer at the National Weather Service office in Columbia, S.C. “Looking at the ... setup, the [parameters] were the top of tornado situations around here. We were ramped up and expecting this.”

At one point, four tornado debris signatures, indicating tornado damage lifted into the air, lined up along a 20-mile stretch of storms, illustrating four simultaneous twisters. Perhaps the most remarkable were the northern two bright areas of debris, known as debris ball signatures, which dangled around each other before combining — hinting at a potential merger of two strong tornadoes. But there’s more to it than that.

“We’re not used to seeing things along those lines,” said Alsheimer, who could not remember an episode like this in his area before. He said that, while it’s impossible to confirm whether the tornadoes’ circulations merged without a detailed computer modeling study, there is evidence to support the idea that the twin funnels got extremely close to one another before some sort of interaction occurred.

Detailed timeline

The first signs of trouble came at 5:14 a.m., when the National Weather Service in Columbia, S.C., issued a tornado warning for parts of Burke County in Georgia and sections of Barnwell and Aiken counties in South Carolina. A hook, which can signal tornado formation, began appearing in the back of a severe thunderstorm, while the radar’s velocity mode revealed a dramatic change in wind speed and direction over a very short distance. This pointed to a storm that had very strong rotation within it.

That rotation tightened as it crossed Atomic Road (Route 125) to the northeast. At 5:24 a.m., the circulation was midway between Jackson and Barnwell.

Then at 5:28 a.m., Doppler radar detected debris about five miles east of Jackson. A Doppler radar product that looks for the types and sizes of objects that the radar beam bounces off detected an area with jagged, irregular objects flying around in the atmosphere. That was a telltale sign that a tornado was on the ground and causing serious damage.

During the next four minutes, that tornado and debris signature moved northeast, approaching the intersection of Williston Road and Tinker Creek road by 5:33 a.m. By now, however, radar revealed a new, rapidly developing circulation just a couple miles to the southeast.

At 5:35 a.m., a second tornadic debris signature appeared on radar beneath the second circulation. There were now two tornadoes on the ground. Debris from the northern twister, only about four or five miles away, began spreading out and falling to the ground north of its parent tornado.

At 5:39 a.m., both tornadoes crossed Route 278, on a collision course with Williston. It appears they narrowly avoided the intersection of Routes 78 and 39.

Debris from tornado No. 2, the southern one, began raining out of the sky to the north, including in and just east of Williston. The second tornado hit near the town of Elko at 5:45 a.m.

There may even have been a third vortex that briefly touched down near the “W” in Williston on the map below, but the National Weather Service has yet to confirm that.

“Essentially we were looking at three potential vortices on the ground at that time,” said Alsheimer.

The two confirmed tornadoes continued east. By 5:52 a.m., they were less than a mile apart as they struck Springfield, S.C. At that time, something unusual was happening.

Tornado No. 2 was swinging forward, as if making a left-hand turn to surge ahead of its northern counterpart. Tornado No. 1 was lagging behind. The “Fujiwhara Effect” was underway, a process by which two vortexes in close proximity orbit a common center until the weaker becomes absorbed into the dominant vortex’s circulation.

At 5:55 a.m., tornado No. 2 had surged ahead of tornado No. 1. Meanwhile, two additional tornadoes had touched down just a few miles to the south. At 5:56 a.m., the National Weather Service issued a new tornado warning, stating “several tornadoes possible.”

The 6:05 a.m. radar scan showed something remarkable; tornado No. 2 had entirely danced around tornado No. 1′s circulation, becoming the more northern vortex as the pair passed through Livingston, S.C.

Over the next 10 minutes, the two tornadoes combined into one large “debris ball” — indicating the possible merger of the two large, destructive tornadoes’ circulations.

What the ground surveys reveal

According to the official survey — which was still under review Wednesday — tornado No. 1 tracked 30.19 miles from the Savannah River Site to northeast of Springfield and was rated an EF3.

Tornado No. 2’s track was initially listed as beginning near Elko, S.C., at 5:46 a.m. — but that’s unlikely to be the true beginning of the twister’s rampage, since radar revealed it had continuously lofted debris in the air beginning five miles or so to the southwest, beginning at 5:35 a.m. In either case, that tornado was also rated an EF3.

However, the damage tracks show that the first tornado dissipated just before reaching the second EF3 twister, with the two circulations separated by just 1.14 miles. According to the NWS, tornado No. 1 “finally dissipated northeast of Springfield, just before reaching the path of [tornado No. 2] just to the east.”

You’ll see that after the interaction, the second tornado quickly strengthened, going from producing EF1 damage (yellow) to EF3 damage (orange). It also took a clear right turn as a direct result of the two funnels’ interaction.

There are two distinct explanations for what may have happened.

One explanation: a merger of sorts

While the damage track from the first tornado doesn’t lead continuously straight into tornado No. 2, there may still have been a merger of their areas of circulation. That could have strengthened the second tornado and may explain the sharp curve in its path as well.

The first tornado apparently dwindled some in intensity as its vorticity — or spin energy — was affected by the competition nearby. This may have allowed the other twister to usurp its vorticity, strengthening the parent circulation of tornado No. 2 and intensifying the surface circulation as well.

“It certainly appears there was ingesting of vorticity into that larger one,” said Alsheimer.

The other possibility: a “downdraft” surge

What may also have happened could be attributed partly to cool, descending air — a downdraft — within the storm, specifically north of the first tornado. This could have choked out tornado No. 1 and given the other nearby twister a bit more of an eastward shove. This is possible, but less likely.

Not your average tornado family

It’s not unusual for a thunderstorm to produce multiple tornadoes. It’s actually quite common with long-lasting, rotating supercell storms on the Plains. The storm’s lone area of rotation will produce a tornado, and that twister will eventually weaken and curve to the left before a new tornado touches down to the right.

But that’s not what happened here.

There were two distinct areas of rotation, called “mesocyclones,” within Monday morning’s storm. The first tornado never curved to the left; in fact, it was drawn to the right, as a result of its interaction with the other tornado.

In cyclic supercells, which exhibit the tornado trade-off discussed, it’s the dying funnel that makes the left turn. But this time around, tornado No. 2 — which curved left — dramatically strengthened.

Why is this so special?

What happened Sunday night in South Carolina may be an event that stands on its own.

Only a few instances of merging tornadic circulations have occurred before — mostly with rotating supercell thunderstorms as a result of a tornado family process. That was the case on March 13, 1990, when the remnants of a former F5 tornado were drawn into a new, strengthening tornado near Hesston, Kan. The new tornado produced F5 damage to the northeast.

A handful of other merging tornadoes have occurred before, all with tornado families in supercells.

A stunning example of two powerful tornadoes merging occurred in Pilger, Neb., on June 16, 2014 — and was broadcast live on the Weather Channel.

But Monday morning’s episode in South Carolina wasn’t a tornado family, and wasn’t from a supercell but rather two different areas of spin embedded within a squall line. Therefore, it may be an event scientists study for years.