The Complexity of the Air Traffic Control System

Part three of my series on air traffic control. I’m not an ATC professional; this is what I learned from a lot of reading. Sources are at the bottom, and corrections are welcome.

Part two walked through the pieces: towers, TRACONs, en route centers, oceanic control, and the equipment they use. On a clear Tuesday afternoon, those pieces mostly just work. This article is about what happens when they don’t get a clear Tuesday afternoon.

I’ve spent my career around software systems, and one lesson carries over well: the individual components are rarely the hard part. The hard part is the seams, where one piece hands work to another. Air traffic control is a system made almost entirely of seams.

The Handoff Problem

#

Every flight crosses multiple facilities, and every crossing is a handoff. The sending controller has to coordinate with the receiving controller, the pilot has to switch frequencies and check in, and everyone has to agree on altitude, route, and speed. Multiply that by thousands of flights an hour and you get a sense of the coordination involved.

Most of the time, it’s routine. When it isn’t (a missed frequency change, a misunderstood altitude, a handoff that doesn’t happen), the gap can open up fast. International flights add another layer: different procedures, different equipment, and controllers and pilots working in English as a second language.

Crowded Skies

#

Some airspace is simply packed. New York is the classic example: JFK, LaGuardia, and Newark sit within a few miles of each other, their arrival and departure paths are tangled together, and a single TRACON handles all of them along with dozens of smaller airports. It’s also chronically short-staffed. A 2023 report from the Department of Transportation’s Inspector General found the New York TRACON staffed at just 54 percent of the FAA’s target level.¹

When demand exceeds what the airspace can handle, controllers slow things down. They space arrivals further apart, put aircraft into holding patterns, and work with the national Command Center to hold flights on the ground before they ever take off. Ground delays are frustrating when you’re the one sitting at the gate, but they’re a lot better than circling for an hour.

Weather

#

Weather is the biggest single cause of delays, and it’s where the system’s flexibility gets tested. A line of summer thunderstorms across the Midwest can close off the routes that dozens of flights per hour were planning to use. Those flights have to go somewhere, which pushes more traffic into the airspace that’s still open, which makes that airspace harder to work. Winter storms shut down airports for deicing and snow removal. Volcanic ash, military exercises, and security restrictions can close airspace outright.

The Command Center and the airlines work through reroutes together, often using pre-planned “playbook” routes for common weather patterns. It’s a daily negotiation, and it’s one reason a storm in Chicago can delay your flight in Atlanta.

Emergencies

#

Then there are the moments nobody plans for: a medical emergency, an engine failure, a pressurization problem. The controller’s job shifts instantly from routine flow to getting one aircraft whatever it needs (priority handling, a direct route, the nearest suitable runway) while keeping everyone else safely out of the way. The pilot flies the airplane; the controller clears the path.

Three Cases Worth Knowing

#

Atlanta: Volume Every Day

#

Hartsfield-Jackson Atlanta International is the world’s busiest airport by passenger count, with about 108 million passengers in 2024. It also recorded 796,224 takeoffs and landings that year, which works out to roughly 2,200 a day.² Atlanta handles that volume with five parallel runways, which lets controllers run arrivals and departures on separate runways at the same time. The layout is a big part of why the airport can move as much traffic as it does.

September 11, 2001: Clearing the Sky

#

On the morning of September 11, the FAA did something it had never done before: it ordered every aircraft in US airspace to land at the nearest suitable airport. Controllers got more than 4,500 aircraft on the ground in roughly three hours, with no accidents.³ International flights headed for the US were turned back or diverted, many of them to Canada. NATCA, the controllers’ union, has called it the single greatest feat in air traffic control history, and it’s hard to argue with them.

US Airways Flight 1549: Four Minutes

#

On January 15, 2009, US Airways Flight 1549 hit a flock of Canada geese shortly after departing LaGuardia and lost thrust in both engines. Controller Patrick Harten immediately offered Captain Chesley “Sully” Sullenberger a return to LaGuardia’s Runway 13, then Teterboro’s Runway 1 in New Jersey. Sullenberger judged he couldn’t reach either one and told Harten they’d be in the Hudson. Less than four minutes passed between the bird strike and the landing on the river, and everyone on board survived.⁴

The pilots rightly got most of the attention. The part I find worth studying is the controller’s side: in a matter of seconds, Harten had to understand what had happened, offer realistic options, and coordinate with other facilities to clear the way for whichever one the pilots chose.

Where Technology Helps

#

Technology already takes some weight off controllers, and more is coming. The en route computer system continuously checks flight plans and projects them forward to warn controllers of conflicts before they happen. NASA’s Airspace Technology Demonstration 2 project tested a departure scheduling system at Charlotte Douglas International Airport; it let airlines wait at the gate with engines off instead of idling in a taxiway line, and it saved more than a million gallons of fuel over four years.⁵

Remote towers are another example. In April 2015, Örnsköldsvik Airport in Sweden became the first airport in the world to be controlled from a remote tower, with controllers working from a center in Sundsvall using cameras and sensors displayed on large screens.⁶ London City Airport followed with a fully remote tower in 2021, operated from the NATS center in Swanwick. These are camera-and-display systems, not artificial intelligence; there’s still a human controller watching.

The Common Thread

#

After reading about all of this, what struck me most is how much the system depends on judgment. The equipment tells controllers where aircraft are and where they’re going. It doesn’t tell them what to do when a storm line moves faster than forecast, or when an airliner loses both engines over Manhattan. That’s still a person, making a call, in seconds.

Next up: Modernization Efforts: Why Upgrading Air Traffic Control Is So Hard.

References

#