Close Call -- O'Hare Airport Runway Scares

March 28, 2006 — -- The news that two sets of airliners came within a few hundred feet of colliding last week at Chicago O'Hare was more than unsettling. In fact, each incident was a very loud message from the underlying system, and one we need to heed.

Directing takeoffs and landings at a large airport equipped with straight, nonintersecting runways is challenging enough for the staff of any air traffic control tower, but when a heavy volume of traffic is combined with intersecting, crossing runways (as with O'Hare), the opportunity for a single human error to balloon into a major threat of collision is a real and present danger.

It's largely the volume of traffic at O'Hare that prompts the use of intersecting runways, although wind conditions can play a role. While the prime directive of running a two-intersecting-runway operation is rather straightforward (don't clear two aircraft in such a way that they could meet catastrophically in the intersection), in actual practice there are many complexities making that task less than simple.

Consider, for instance, what happens when a tower controller clears a small commuter jet for takeoff on a due west runway while a second tower controller (handling the runway running due south) is waiting to clear a heavy Boeing 747 for takeoff. The 747 will be cleared when the controller is sure that the jumbo's crew can't possibly reach the intersection before the commuter jet is safely airborne.

But, suppose the crew of the commuter jet is inordinately slow today. For whatever reason, suppose it ends up many seconds behind in starting the takeoff roll and, due to weather or darkness or workload, the controllers don't notice. Meanwhile, let's say the 747 has been very quick to power up and is now accelerating southbound, well ahead of the timeline its controller had expected. Suddenly, what had been a carefully interlaced takeoff with no potential hazard could become a near miss at the intersection, or worse.

While it will be up to the National Transportation Safety Board to study what went wrong at O'Hare last week, the fact that human error was involved does not solve the problem. Neither, of course, does the act of issuing blame. Controllers are human -- incredibly talented humans able to balance terribly complex mixes of aircraft in their minds and issue the right verbal commands to keep them moving safely -- but humans nonetheless.

That simply means that even though they perform at a level of near-perfection, mistakes will happen, and that fact is merely a starting point in constructing a series of barriers to catch human errors.

That's where the procedures followed by the controllers, future computer monitoring, and perhaps even physical changes to the airfield (such as building a new one) come in.

When you know a system has weaknesses (which is the message O'Hare's system was sending last week), safety engineering requires you to build new traps to catch the rare mistake. In O'Hare's case, that may mean some substantial changes in the way intersecting takeoffs and landings are handled. The bottom line is simply this: We've received too many warnings, and it's time to act.