The rains are tapering off in South Carolina after a disastrous weekend that brought over two feet of rain and catastrophic flooding. Dams have been breached, rivers are at record flood stage, homes and cars are filled with water and multiple people have been reported dead in the disaster.

Authorities in South Carolina on Monday urged people to stay home if it was safe to do so, saying that flooding was expected to continue in more than half the state for several days. On Sunday, authorities responded to hundreds of reports of trees in roadways and hundreds of reports of flooded roads. Tens of thousands of sandbags were used by state and local agencies, while a stretch of Interstate 95 was shut down and traffic rerouted. Overnight, several cities and counties declared curfews, while others have declared states of emergency.

In the wake of a historic and devastating flood, South Carolinians are documenting their experiences with rising floodwaters on social media. (Jenny Starrs/The Washington Post)

The cities of Charleston and Columbia set new records for 24-hour, two-day and three-day rainfall totals. On Saturday alone, 11.5 inches of rain fell in Charleston, where, just five days into the month, it is now the rainiest October on record. Just a few miles northeast up Route 17, an astonishing 24.23 inches of rain fell near Mount Pleasant, S.C. In Huger, S.C., 21.04 inches fell. And rainfall totals over 16 inches are widespread across 10 counties from Columbia to Charleston.

According to statistics compiled by the National Oceanic and Atmospheric Administration, South Carolina’s torrential weekend rain has well surpassed a 1,000-year rainfall event — one that, on average, we would expect to see about every 1,000 years. A three-day, 1,000-year rainfall event for Charleston County would have been 17.1 inches. A four-day, 1,000-year event would have been 17.5 inches. Boone Hall Plantation, just north of Mount Pleasant, in Charleston County, reported more than 24 inches of rain through Sunday morning, which essentially blows NOAA’s 1,000-year events scale out of the water.

But last week’s much-anticipated Hurricane Joaquin, which at one point was threatening a Mid-Atlantic landfall with all of its storm surge, strong winds and torrential rainfall, tracked well offshore over the weekend. So how could something this devastating have happened without an actual hurricane landfall?

Hurricane Joaquin did play an indirect role in South Carolina’s weekend deluge, but there’s much more to this meteorological story.

As Hurricane Joaquin tracked north, well east of the coast, a separate, non-tropical low pressure system was setting up shop over the Southeast late last week. This system drew in a deep, tropical plume of water vapor off the tropical Atlantic Ocean. At the same time, this upper-level low pressure system tapped into the moist outflow of Hurricane Joaquin.

The moisture pipeline fed directly into a pocket of intense uplift on the northern side of the non-tropical vortex. Within this dynamic “sweet spot,” thunderstorms established a training pattern, passing repeatedly over the same location and creating a narrow corridor of torrential rain stretching from Charleston to the southern Appalachians.

The remarkable thing about this process is that it was sustained for three days.

The two players that gave rise to the great flood are shown in the dramatic water vapor image below. This satellite image was captured Sunday evening, while the flash flood-producing rainstorm over South Carolina was finally winding down. The enormous sweep of the non-tropical vortex is shown by alternating swirls of moist air (gray) and dry air (red). Hurricane Joaquin, much smaller in comparison, is completely embedded in a deep plume of tropical moisture. South Carolina’s heavy rain region, by comparison, is located in the tiny zone of interaction between these two mammoth, spinning vortexes.

At the surface, a low-pressure center developed east of the upper-level low along the coastline. A coastal front stretched out to its north, separating warm and very humid tropical air from cooler, drier air over land.

A jet-like plume of high-moisture air raced inland in the low parts of the atmosphere. The sodden air was lifted upward by the front and upper-level low, dumping repeated bouts of heavy rain. The entire pattern became stagnant beginning on Friday, typical of closed upper-level low centers as they occlude and become cut off from the surrounding weather patterns. Unstable air drawn off the warm Gulf Stream fueled towering complexes of thunderstorms, unleashing rain at rates approaching 3 inches per hour.

Some meteorologists have been calling this plume of rain a predecessor rain event, or “PRE,” which sometimes occurs ahead of tropical storms that interact with separate areas of low pressure and lingering surface fronts — exactly what Hurricane Joaquin did. Meteorologists do not agree on all the details, specifically whether the event can be more definitively linked to Hurricane Joaquin, and if this event can be defined as a PRE. The final determination must await more detailed analysis.

But here’s what we do know — at least eight key elements conspired to create a highly efficient, small-scale rain machine centered on South Carolina. We take each of these key elements one at a time, starting near the surface and working our way up in the atmosphere.

Very high precipitable water feeding into the upper-level low pressure system

Precipitable water is a measurement that meteorologists use to determine how much air moisture is available for rainfall. Precipitable water in the westward-surging plume, illustrated below, averaged 1.8 to 2.3 inches throughout the event, and at times peaked around 2.5 inches — very high for the month of October.

We could call this tropical moisture channel an atmospheric river, and it formed between the closed upper-level low over northern Florida and a broad ridge across the northwestern Atlantic Ocean. This 175 mile-wide channel incorporated moisture from Hurricane Joaquin, which was located further east.

Strong, low-level onshore flow

Strong, low-level onshore flow — called a low-level jet — was embedded in the atmospheric river. This feature flowed consistently at around 50 mph from the east. The jet and its moisture are part of the low- and mid-level warm conveyor belt feeding into the closed low. On Sunday, this jet began lifting north, away from Charleston, where it had set up shop for over 48 hours.

An unstable air mass offshore

A narrow rain band with training rain cells became locked across central South Carolina this weekend. Feeding directly into this band was a moderately unstable air mass, with CAPE (convective available potential energy) values approaching 2000-2500 J/kg.

Having formed over the Gulf Stream, the unstable air promoted deep, strong updrafts in thunderstorms, enabling them to reach great heights. Cloud top temperatures dropped as low as minus-94 degrees. These updrafts, in turn, manufactured rain rates of 3 to 4 inches per hour, at times.

Coastal front

This front focused convergence and rising of air in a narrow zone along the coast, just off Charleston in the spot where the low-level jet intersected the front. This low-level “sweet spot” remained nearly stationary for 48 hours. It triggered a succession of thunderstorms that coursed inland before drifting north in the deeper southerly flow.

Stationary mid- and upper-level low pressure center

What began as an amplifying trough in the upper-level flow on Friday became a closed-off vortex during the weekend. This non-tropical vortex created a warm conveyor of moist, unstable air feeding from the east, shown below. Note also the compact, intense spin center of Joaquin on the right side of the figure.

Potent upper-level jet streak north of the closed low

This includes a jet streak, or pocket of fast-flowing air, and region of enhanced difluence (spreading of air) aloft. The jet streak’s entrance region (shaded red in the diagram below) is where air gets drawn upward very rapidly, like stoking the flames of a fire. This zone serves as the upper-level “sweet spot” for heavy rain production.

Perfect coupling of low and upper-level sweet spots

The combination of low-level convergence along the coast and the fanning-apart of the upper-level flow set up a narrow pocket of rapidly rising air. This zone is highlighted in the figure below, strongest in the middle atmosphere, where condensation of water vapor is highly efficient. Note the nearly perfect alignment of this upward motion center with the rain band illustrated in the radar image above.

A stationary, small-scale convective cloud

This is the heart of the rain-making machine. New thunderstorms erupted continuously along the coast, which were then whisked inland and north by the deeper flow of the upper vortex. They weakened as they moved out of the deep “sweet spot.” Then new thunderstorms bubbled up to the south and east, replacing the dissipated ones.

This cycling of rain cells within the same geographical region, called training, is a prime flash-flood generating mechanism, responsible for numerous other flood disasters throughout U.S. history.

Putting it all together

It’s useful to illustrate all these processes in a composite schematic. There are several synoptic-scale factors, including the non-tropical upper low and Hurricane Joaquin, both creating a rich plume of unstable, tropical air. The processes cascade down to the regional scale, culminating in training of convective elements across individual counties.

In this diagram, the gray cloudy region represents the mesoscale band of heavy rain.  Deep convective cells repeatedly form at “X” along the coastline (not shown, but implied by the location of the coastal front), move north, and decay at location “O.”   Note that the trigger point “X” remained anchored for 48+ hours!

Was this a PRE?

A predecessor rain event, or PRE, is defined as an area of heavy rainfall typically observed north of a hurricane. It is distinct from the hurricane but still indirectly tied to it.

In the past, PREs have been responsible for some big rains, including Hurricane Frances in 2004, in which a PRE caused costly flooding in New York City, and Tropical Storm Erin  in 2007, which inundated a large part of the Upper Midwest with several PREs. It’s thought that one out of three of Atlantic tropical cyclones making landfall in the U.S. generate some sort of PRE.

A simplified schematic of a generic PRE is shown below. It shows how mid- and upper-level moisture is drawn off the northern side of a tropical cyclone, toward the northeast. The moisture gets pulled into the dynamical “sweet spot” of an upper-level jet streak. Then one or more zones of heavy rain – distinctly separate from the main tropical storm’s main rain shield – develop. These are the PRE regions.

There’s been discussion among the meteorological community whether the South Carolina rainstorm classifies as a PRE, and we at the Capital Weather Gang suggest that this event includes at least some elements of a PRE.

The upper-level trough and jet streak are not located north of the tropical system, rather alongside it. But the non-tropical and tropical low pressure centers are widely separated, and Joaquin feeds moisture into the dynamic region of ascent in the jet stream disturbance. On the large-scale, they are connected, but the South Carolina rain maximum developed as a distinct, small-scale entity not at all connected to the main rain shield of Joaquin.

Thus, there are some nuances of the South Carolina event that must be carefully analyzed before it can be called a PRE for certain.

Wes Junker, Angela Fritz and Mark Berman contributed to this report.