I don't mean to diminish the controllers, or to belittle the experience and dedication they bring to the job. The sight of a radar scope swarming with little ovals, each representing a flight, is indeed daunting. But what does it mean that control rooms can sound like trading floors? Perhaps only that air-traffic control has become less regimented, more human, and more complex than originally anticipated—which may even be for the best. Argue the risks as one might, there is no doubt that air-traffic control consists increasingly of informal solutions pieced together at the last moment to cope with an overwhelming flow. In that sense it is a typically American institution—the problem coming first, the attempt to manage it coming afterward.
On the most mechanical level, the most pressing issue that air-traffic controllers face is a surge in air traffic without a commensurate expansion of runway availability. Since 1978, when President Jimmy Carter deregulated the airlines, unleashing competition among them, the number of scheduled flights in the United States has grown by nearly 70 percent. And the growth has been lopsided: of the several thousand airplanes aloft during a typical daytime rush, most are headed for the same few cities. The busiest fifty airports, out of thousands of airports altogether, now account for more than 80 percent of the nation's traffic. This is not merely because those places are where people want to go but also because to stay competitive, airlines need efficient route structures centered on hubs—the now-familiar passenger-exchange airports that by their very nature increase the number of takeoffs and landings.
Newark, for instance, does double duty as an arrival point for New York City and as a northeastern hub for Continental and Federal Express. Faced with all the inbound airplanes, its approach controllers have no choice but to deal with them. They grapple with the core problems of overcrowded airspace: that flight is fast, fluid, and determinedly forward-moving; that every airport, airplane, and pilot is different; that thunderstorms, fog, wind, ice, snow, or merely low clouds can block a route or slow a runway; that even under a clear blue sky airline schedules push airports past their limits. The slightest bump then ripples backward, forcing the controllers to scramble. A flight may miss an early turnoff from a runway, or come in too fast or slow, or ignore a call on the radio, or jump out of line with an engine shut down. A new pilot may be unsure of the local procedures. An old pilot may get huffy and insist on having his way. These things happen constantly. The resulting complications are measured in wasted fuel, wasted money, and wasted time—but not in lives, or even in levels of danger.
Across controllers' radar screens I watched the targets move in short jumps, dragging identifying tags behind them: Lufthansa, United, Continental—dozens of airplanes at a time. By assigning headings and descent paths, the controllers for the Newark sector angled the flights down from the mid-altitude collection points known as arrival gates, joined them up to the south, and swept them into an arc that took them north past the airport and skirted LaGuardia's airspace before bending back around, straight in for the runways. The purpose of the arc was only secondarily to keep the airplanes apart. Its primary purpose was nearly the opposite: to give controllers the flexibility necessary to tighten the spacing, and to exploit the occasional gaps by shooting airplanes in from the side, pushing them toward the airport ahead of sequence.
The most basic geometry of air traffic dictates that departing airplanes fan out and so usually diverge, that cruising airplanes only sometimes cross, but that arriving airplanes must inevitably converge. The traffic compresses accordion-style as the airplanes slow toward their touchdown speeds. The compression does not mean that the airplanes are in danger of rear-ending each other: closing speeds are low between airplanes flying in the same direction. The formal separation requirements, which are measured in miles, are dictated ultimately by the civilian orthodoxy that requires one airplane to taxi clear of the end of a runway before the airplane behind it lands at the beginning. Military pilots routinely fly and land in formation, and safely. I don't mean that airline pilots should too, but the margins built into standard civilian procedures are large. That is why New York Approach controllers take their pride not in the collisions they avoid—an issue that almost never comes up in the manner civilians imagine—but in the pressure they keep on runways.
ONE might argue that I have described the wrong kind of controllers, that in order to understand the urgency of the work one must visit not the radar people funneling the traffic in but rather those at the receiving end of all that accumulating traffic pressure—for instance, those who stand guard at Newark International Airport itself, high atop its control tower.
When I climbed to the glass-walled cab where they work, the center of the operation was a slight young man with blond hair and birdlike reactions, whom I will call Dobkin. He wore a lightweight headset and held a transmitter switch down low in his right hand. It was Dobkin's turn on the frequency known as "local," which gave him responsibility for the airport's two parallel runways, the narrowly separated "22 Left" and "22 Right," running southwest beside the turnpike. The third runway, a short east-west reliever called "29," crossed the paths of the parallels and conflicted with their traffic. It was a cramped and awkward layout.
Dobkin said, "We work with what we've got. The parallels were built way too close for simultaneous approaches. We use the outer runway for arrivals. We use the inner runway for departures. We try to run the props over there on 29, keep them out of the way of the jets, but we can't cross them into the main approach. When the wind's light, we flip that runway back and forth, pump a load of departures to the west, then bring the inbounds around for landings to the east."
The idea that such an air-traffic controller is somehow in the business of "talking airplanes down" is part of a myth fostered specifically by the movies. Dobkin had ridden in cockpits a few times, but he knew little about the actual flying of airplanes. Between radio conversations he told me how he had come to the job after a stint as a controller in the Navy. It was a typically short story: he escaped home, learned a skill, grew tired of saluting, and hired on with the FAA because the FAA was hiring; he chose a control tower over a radar room because he liked to look at airplanes; he picked Newark for the money and the action, and sometimes wished he hadn't. He concluded his story with the false regret of a man proud of his skills: "So here I am ten years later, just another keeper of the concrete."
He had a high-strung personality, encouraged by the work. If the purpose of his game was simple—to squeeze the maximum possible use from these three runways—in execution it was fast-paced, complex, and competitive. He said, "You've got to use every chance, every gap, to move the traffic. Slam and jam. The job keeps you on your toes."
And safety? It intruded not as an active, minute-to-minute concern but as a set of rules within which one had to perform, the most basic of which was the restriction against simultaneous operations on a single runway. Perspective is needed here. The deadliest airline accident in history was a runway accident that occurred in 1977 when two 747s collided on Tenerife (in a fog, one taking off, the other taxiing across); other runway collisions have occurred. They have all, however, been freakish accidents resulting from multiple errors by both the controllers and the pilots. Except in the worst weather, or sometimes at night, pilots can easily see anyone lingering on the runway, and on their own initiative can delay their takeoffs, or if they are landing can add power and climb safely away from the ground. It is primarily because such go-arounds waste valuable landing slots, and further burden the final approach, that controllers work to avoid them. In other words, Dobkin took the timing seriously, but as an efficient practitioner of traffic flow rather than out of a sense of averting danger.