The crash last year of Korean Air Lines Flight 801 as it approached the runway at Guam underscores what many travelers have long sensed: The last few minutes of flight, in approach and landing, are the most dangerous.
Korean Air’s Boeing 747 jumbo jet might have arrived safely in Guam were it not for one factor. An important component of the instrument landing system that guides planes to the runway was out of service for maintenance.
That meant the pilots had to make a “nonprecision approach” to the airport using more basic instruments, a type of approach that pilots say is much more difficult and hazardous. The Korean Air jet crashed 3.3 miles short of the runway and killed 225 people. Only 28 survived.
Other factors, however, can– and have–caused crashes on approach and landing. Pilots can incorrectly set their altimeters, which must be adjusted for local barometric pressure, resulting in false altitude readings.
Pilots may misunderstand air controllers’ instructions and descend prematurely. Or, especially overseas, an approach may be poorly designed, bringing jets too close to hills near airports.
A Boeing Co. study of airline accidents around the world found that nearly 50 percent of fatal crashes occur on approach or landing. (The rest are spread fairly evenly among takeoff, climb, cruise and early descent.)
Another Boeing study focused on where 40 accidents or near-accidents occurred in relation to destination airfields. When Boeing superimposed the accidents on one graphic illustration, it showed the route to the threshold of the runway littered with red crosses, each one marking where a plane went down short of a safe landing.
In one 1995 case, an American Airlines jet coming into Bradley International Airport near Hartford, Conn., flew too low, through trees on a ridge top. Branches sucked into the engines knocked out power. But the jet made it to the runway, and the 78 people on board were unharmed.
The approach and landing problem is so acute that the aviation industry is in a major effort to seek ways to reduce risk.
The effort, dubbed ALAR for Approach and Landing Accident Reduction, is led by Flight Safety Foundation, an independent Alexandria, Va., group.
The effort is “critical to air safety,” said Stuart Matthews, chairman of Flight Safety Foundation.
Airline safety has improved steadily in recent years. Accidents that plagued carriers in the 1970s and 1980s, such as crashes because of wind shear and midair collisions, have almost become a thing of the past in the U.S. Sophisticated detection systems at airports and on jets warn pilots of wind shear. Radars coupled to jet computers alert pilots to any other planes on a potential collision course.
National Transportation Safety Board studies show that from 1993 through 1996, scheduled U.S. carriers averaged 0.02 fatal accidents per 100,000 flight hours, less than half the fatal-accident rate a decade earlier.
But approach and landing accidents remain the Achilles’ heel of airline safety. Such accidents are difficult to eradicate because the causes are so varied.
In a case apparently involving pilot error, a Boeing 737 flown by a Peruvian airline crashed four miles short of the Arequipa, Peru, airport last year, killing all 123 on board. Safety experts suspect the pilot set his altimeter improperly or made another error.
Such accidents have been less common in the U.S., mainly because of sophisticated instrument landing systems. Wally Roberts, a retired Trans World Airlines pilot and a consultant on navigational aids, said approaches made with the instrument landing systems are safer than nonprecision approaches.
In an instrument landing system, equipment at the threshold of the runway transmits an electronic beam. An airliner homes in on the beam miles away. In a cockpit “glide-slope” display, the beam generates a horizontal bar representing the runway. If the bar moves up in the instrument, it means the plane is coming in too low; if the bar moves down, the plane is coming in too high. This guides pilots in adjusting their glide-slope, or altitude, as they descend, allowing a descent at an even, constant rate.
In the older nonprecision approach, a pilot must watch two instruments. Distance-measuring equipment is checked to determine the distance from the runway. And the pilot must check the altimeter. At a certain distance from the runway, the plane should be at one altitude. At a closer distance, it should be at a lower altitude, and so on. Typically, this means the pilot has to “stair-step” the plane down.
At big airports, several alternate runways are available should the FAA take one runway’s ILS out of service. But Guam has only one runway long and strong enough to handle a 747, which Korean Air was flying. The FAA said it sent “Notices to Airmen” that the glide-slope equipment at Guam’s major runway would be out of service for several months.
Flying through heavy rain in August, the Korean Air jet hit the ground about 3.3 miles short of the runway. At that point, the jet should have been at 1,440 feet.
The American jet’s flight through the trees, in November 1995, is eerily similar to the Korean Air crash and may give investigators some clues. The main runway of Bradley International Airport has an instrument landing system. But apparently because of wind directions, the American pilots planned to land on a secondary runway, which didn’t have the system.
Because the McDonnell Douglas Corp. MD-83 twin-jet had left Chicago late, it wouldn’t arrive until nearly 1 a.m.
The co-pilot got barometric pressure at the airport from a radio report, but the report was outdated because pressure was falling so rapidly. So, the crew set altimeters that gave false readings. While the altimeters showed the jet at one altitude, the plane was lower.
As the jet wobbled toward the runway, the captain ordered the flaps back down, hoping a “balloon effect” would carry the plane a bit farther. It worked.
