Florence has been one weird storm.
The massive system threatened to grow to a Category 5, which would have had winds of 156 mph or more. No storm in recorded history had ever done that so far north, so close to South Carolina. Meanwhile, its uncertain path had computer models pointing fingers in different directions trying to predict it.
After landfall, it moved inland to the southwest, a direction that no hurricane expert had ever seen before.
Florence — a frightening storm larger than the Carolinas combined — was nothing if not unpredictable.
But National Hurricane Center forecasters nailed it. They indicated as early as six days out that the storm was most likely to strike where it did, in North Carolina near the South Carolina border, then turn south.
The performance lit up social media with accolades from meteorologists. Shea Gibson, the Charleston-based forecaster for the private company WeatherFlow, called it one of the more impressive forecast verifications he had ever seen.
Here's why the storm was so unpredictable, and why the Hurricane Center staff got it right:
Hurricanes get "steered" by shifting air pressure currents. They tend to move into currents of low pressure, pushed by circulations of high pressure. Florence tracked toward the United States while two high pressure currents moved across the nation, with only a narrow current of low pressure between them.
Timing was everything. What happened would be governed by the shifting strengths of the circulations as well as the storm.
"The Hurricane Center was able to gauge the ridge guidance (the shifts) just right," Gibson said.
Looking at the bigger picture, hurricanes in recent years have gotten tougher for the computer models to figure out. The storms are strengthening faster and in unexpected places. They are lasting longer and occasionally still not going where computer runs say they should.
In June, the National Oceanic and Atmospheric Administration released a report that said while it is too soon to scientifically determine whether climate change was having an impact on hurricanes in the Atlantic, the warming caused by fossil fuel burning might already have caused changes that "aren't confidently modeled" by the computers.
Forecast models are computer simulations of how the hurricane reacts to changes in the atmosphere and ocean. Each model is designed to weigh piles of data on various changes. Each model tends to read what's going on a little differently.
The notorious European, or Euro, model has generally been the most precise tracking recent storms, and its runs persistently forecast Florence to spin south down the South Carolina coast as a powerful storm instead of making landfall where it did.
That would have been a first-of-its-kind track — the reverse of what Hurricane Matthew did in 2016 — and with winds capable of damage that would dwarf the millions of dollars of destruction done by Matthew.
The model runs have gotten to be popular social media feed, and the Euro track quickly went viral, stoking anxiety, if not panic, in South Carolina.
But the forecasters knew that as good as the Euro model is, it also has a tendency to be very wrong when it's wrong. They weighed its runs with the other models, as well as their judgment of what they saw happening in the environment.
That's what they are paid to do. And the Hurricane Center forecasts have been consistently more reliable than any one computer model.
You can see for yourself in the verification critiques the Hurricane Center does of its work after each storm season, said meteorologist Bob Bright of the National Weather Service in Charleston. From year to year, the forecasts keep getting more accurate.
"It's just incredible," he said. "The Hurricane Center works from all that data, all those models runs and the staff's own experience. That's why it's important not to focus on the model runs."
Meteorologist Bob Henson, with the private company Weather Underground, echoed that.
"They're using their experience to ferret out which models are doing the best at the time. Their experience puts (the forecast) over the top," he said
On Saturday, with the remnant of Florence still spinning over South Carolina and four other storm systems to be monitored in the Atlantic and eastern Pacific oceans, "no one is available today" to comment on the forecaster's work with Florence, said Hurricane Center spokesman Dennis Feltgen. "Perhaps next week for this topic."
If there was one lapse in the Hurricane Center's read on Florence, the forecasters did not gauge very far in advance how much power the storm would lose as it made landfall and how quickly it would weaken afterward. The once nearly Category 5 storm fell to a Category 1 with winds of 90 mph.
But the reasons for that were pretty straightforward.
Dry, shearing winds worked their way into the storm unexpectedly, and at the last minute. Meanwhile, Florence got so big and strong the swells it pushed ahead of itself rolled into the Continental Shelf offshore "upwelling," or pushing up cooler deep waters. Hurricanes fuel themselves on the warmer surface waters.
"Sometimes these storms get so large so quickly it's hard for the core to strength. It might be it just wasn't able to take advantage" of the warm waters and other conditions that would have strengthened it, Henson said.
"We got really lucky with the intensity falling like it did, but we are seeing the predicted rainfall amounts verifying to create catastrophic flooding," Gibson said.
J. Emory Parker contributed to this report.