 |
| Two pilots were killed when their planes collided during an aerobatic demonstration in Poland in 2007. |
We were returning after a lunch on Montauk, Long Island, about a 30-minute flight across the Sound and Rhode Island to my home airport south of Boston. As I climbed to 5,500 feet, I called Providence approach control to pick up advisories en route. The controller was busy at the time and told me to try back later and stay clear of his airspace. In the meantime, I could monitor the frequency to get a mental picture of the traffic in the area.
We were just crossing the Rhode Island shoreline when I heard another plane, a Cessna at 7,500 feet, call approach from the west en route to Newport, RI. The Cessna, like me, was told by the controller to try back later. The Cessna’s altitude seemed a bit odd to me because at 5,500 feet we were bumping up against the underside of an overcast. Unless there was a hole in the overcast that I couldn’t see, the Cessna would need to descend through it to land at Newport. The cloud layer made the Cessna invisible to me, and I was irritated at the fact.
Did the Cessna then announce he was beginning his descent into Newport? I don’t remember. Neither did I hear the controller warn of any conflict. Perhaps I was chatting with my copilot or preoccupied with something else. But in the next moment, my windscreen seemed filled with the profile of an airplane. I sucked in a breath, but before I could even react the moment passed. The plane was gone. My copilot and I just looked at each other. I doubt the pilot of the Cessna even saw us.
My copilot and I briefly debated whether to report the incident to the controller, but we decided to let it pass. The controller was busy. We were alive. Let’s get home.
Among all the things that can go wrong in the air, perhaps nothing strikes more terror in the heart of a pilot than the thought of a midair collision. The idea of metal slashing into metal, of helplessly falling to earth, and the near-certain death of multiple persons makes the blood run cold.
 |
| Artist's impression of the collision of two airliners over the Grand Canyon in 1956. |
While midair collisions have been a part of flying since the dawn of aviation, their occurrence didn’t become much of a public concern until the growth of airline service in the U.S., and even then it took tragedy to prompt aviation authorities to begin thinking about the problem. On June 30, 1956, a TWA Lockheed Super Constellation flying from Los Angeles to Kansas City and a United Airlines DC-7 flying from Los Angeles to Chicago collided over the Grand Canyon at 21,000 feet, killing all 128 persons on board.In the wake of the disaster, Congress opened hearings to probe the general problems of airspace and air traffic control management.
Though U.S. air traffic had more than doubled since the end of World War II, little had been done to expand the capacity of the air traffic control system or to increase safeguards against midair collisions. Sixty-five such collisions had occurred in the United States between 1950 and 1955. The lack of widespread radar coverage meant the air traffic control system did not have the ability to segregate traffic flying under visual flight rules from those flying under instrument flight rules, or slow-moving flights from faster ones. Aircraft flying under a flight plan simply radioed in their positions when passing over navigation aids or designated reporting points.
Following the Grand Canyon accident Congress reorganized the Civil Aeronautics Board to create the FAA, which subsequently expanded the network of radio navigation beacons that make up the national airway system, broadened radar coverage across the country, and put in place new rules requiring airliners to be in contact with air traffic controllers during most phases of flight.
While such measures undoubtedly improved navigation and control of airliners, the responsibility of collision avoidance still fell upon pilots’ ability to “see and avoid” other aircraft.
During primary flight training, pilots are trained how to visually “scan” the sky in search of other aircraft. Generally, this means taking vertical sections of the view out the windscreen and focusing on them for a second or two in a continuous sweep, trying to spot, as one of my flight instructors told me, “parts of the scenery that are moving.” There are obvious limitations to this technique. First is the design of airplanes. Cockpits are full of blind spots—window posts, door frames, engine canopies. And because of the general construction of planes—that is, long tubes—any aircraft behind, below or above most planes is impossible to see.
Second is the ability of the human eye to see very small objects. In the late 1980s, a study at the Massachusetts Institute of Technology looked at pilots' abilities to spot other aircraft in flight. The study tested the number of times pilots spotted or failed to spot aircraft and at what distance. It found that pilots could spot traffic only 56 percent of the time at a median range of just under one mile.
In addition, research shows that the average person has a reaction time of 12.5 seconds when confronted with a potential collision (see chart). In my case, flying in cruise at around 200 mph, if another plane of similar performance comes at me head-on, with a closing rate of 400 mph or nearly 7 miles per minute, my average 12.5 second window to react begins when the plane just comes into view at about a mile. Not a big margin.
 |
| Chart courtesy FAA |
Given the limitations of “see and avoid,” a lot of pilots resign themselves to the “Big Sky Theory” of traffic avoidance. That is, the sky is a big place, and statistically the odds of two small objects hitting each other are very remote. Call it the “I’m feeling lucky” concept of traffic avoidance.
Still, the odds of a midair collision are very long. A study written in 1993 calculated that roughly one flight in a million ends in a midair collision. FAA statistics collected since 1959 report an average of 30 midair collisions per year (an annual number that has come down dramatically the last three decades due to improved air space restructuring, air traffic control procedures, and technological advances, including the advent of GPS).
More telling perhaps is the number of near-collisions (aircraft passing within one mile of each other). On average, pilot-initiated reports to the FAA or NASA’s Aviation Safety Reporting System show there’s an average of around 200 near-miss incidents per year (“near miss” being the semantics chosen by the FAA to avoid using the more descriptive word “collision”). That number, however, is only as accurate as the percentage of pilots reporting every incident. The FAA Office of Safety has calculated that fewer than 20 percent of pilots report a near-miss incident (you can include me in that 80 percent). This puts the actual number of near-misses at a number of least 1,000 per year, and likely much more. All of which is cold comfort to those of us flying in areas crowded with airports and airplanes.
 |
| Image of radar controller's screen. |
The arrival of the jet age made the concept of see and avoid even more problematic. As passenger traffic boomed, not only did the number of planes filling the sky skyrocket, so did the closing speeds of those aircraft in potential conflict. General aviation underwent a similar boom, making the challenge of integrating aircraft with such large disparities in performance vastly more complex.
A number of horrific collisions involving airliners and small aircraft prompted a series of reforms in air traffic control and airspace design to improve separation of aircraft.
On September 9, 1969, Allegheny Airlines Flight 853, a McDonald Douglas DC-9 carrying a total of 82 passengers and crew, was descending into Indianapolis while being vectored by approach control when it collided with a single-engine Piper Cherokee at 3,500 feet. The student pilot of the Cherokee was flying a VFR solo cross-country flight and was not talking to approach control, but neither was he required to. All 83 aboard both planes were killed.
The controller handling the flight testified that he never saw the Cherokee on his radar screen. For non-airliners, radar controllers relied on primary returns—“skin paints” of reflected energy off the aircraft’s metal skin that illuminated their radar screen with an image of the aircraft but gave no other information such as altitude, airspeed and direction of flight. These radar systems—still in use today—are prone to blind spots and interference, and have limited range and reliability. Indianapolis, like most commercial airports, had a control area that consisted of a cylinder of air 10 miles in diameter up to 3,000 feet above ground level centered around the airport.
In its report, the National Transportation Safety Board cited the inadequacy of the air traffic control system in separating the mix of airliner and general aviation traffic in high-density areas, as well as the fundamental flaws of “see and avoid” in that context.
In response, the FAA moved to upgrade radar equipment so that controllers would be alerted if an imminent conflict between aircraft was detected. In 1970, it imposed redesigned airspace around the busiest airports, creating multi-tiered towers of airspace centered around the airport that extended out 20 or so miles. These towers, or Terminal Control Areas (relabeled Bravos in 1993), have a core column of airspace extending from the ground to as high as 10,000 feet. As you travel progressively further from the center, the tiers or shelves become thinner, leaving more open space beneath, like an inverted wedding cake—essentially funneling traffic to the airport. Aircraft need permission from air traffic control to enter the airspace and have to maintain two-way radio contact. Aircraft also need to be equipped with an onboard transponder, a radio device that transmits to radar screens the individual aircraft’s location, groundspeed and altitude.
 |
| A photographer caught this haunting image of the PSA jet in the moments before it crashed into a San Diego neighborhood. |
The technology and airspace restructuring improved traffic management and reduced the possibility of midair collisions. But the primary responsibility for collision avoidance still fell to pilots, especially in areas outside those covered by a TCA. This became tragically clear on September 25, 1978, when a Pacific Southwest Airlines Boeing 727 carrying 135 passengers and crew collided with a Cessna 172 at 2,600 feet while on approach into San Diego’s Lindbergh Field, an airport not covered by a TCA but a less restrictive Terminal Radar Service Area, in which the same basic radar services are offered, but not required.
Even so, both planes were in contact with controllers, both planes were transponder equipped, and both planes had been alerted to each other’s location. In addition, the controller had an updated radar screen with a built-in conflict-alert system.
About three minutes before the collision, the descending PSA jet was advised of “traffic [at] twelve o’clock, one mile, northbound.” Then six seconds later, the controller reported the Cessna as “additional traffic” ahead at an altitude of 1,400 feet and climbing. The PSA jet pilots scoured the landscape ahead and below of them for the traffic. Twenty seconds later, the PSA jet reported “traffic in sight,” although no one will ever know exactly which of the two aircraft the pilots saw. At this point, the controller said “maintain visual separation,” which meant that the PSA jet pilots were responsible for maintaining separation from other aircraft.
The Cessna, piloted by an instructor and student practicing IFR procedures, was told “traffic at six o’clock [directly behind them], two miles, eastbound; a PSA jet inbound to Lindbergh Field, out of 3,200 [feet], has you in sight.” The Cessna acknowledged the advisory, though the jet’s position made it invisible to them.
Handed over to Lindbergh tower for it’s final approach, the PSA jet checked in and gave its position. The controller acknowledged and advised the jet that there was “traffic, twelve o’clock, one mile, a Cessna.” This was a minute before the collision.
The cockpit voice recorder revealed the pilots of the jet discussing where the Cessna might be. “Is that the one [we’re] looking at?” asked the captain. “Yeah, but I don’t see him now,” answered the copilot. The captain told the controllers, “Okay, we had it there a minute ago,” followed by, “I think he’s passed [or passing] off to our right.”
The PSA pilots continued to discuss the whereabouts of the Cessna. Thirty seconds before the collision, the plane’s landing gear was lowered. About the same time, a collision-alert alarm sounded at the controller’s station. The controller told the Cessna, “traffic in your vicinity, a PSA jet has you in sight, he’s descending for Lingbergh,” but relayed nothing to the PSA pilots, who clearly did not have the Cessna in sight.
During the investigation, the controller testified that he heard and saw the conflict-alert, and discussed it with his supervisor. Though the data blocks depicting the two planes overlapped and the altitude readouts were indiscernible, the controller did nothing to resolve the conflict. In its report, the NTSB says, “He said that he had pointed out the traffic to Flight 182; the flight crew had stated that they had the traffic in sight and that they would maintain visual separation from the Cessna. As far as he was concerned, there was no conflict, and therefore, no further action was required.”
The controller further testified that since the upgraded radars were installed the month before, the facility experienced an average of 13 conflict alerts per day. Controllers had grown accustomed to the alerts, many of which were nuisance alerts in which the aircraft were not close enough to require action or had passed by each other.
Just as the PSA jet began its turn toward the runway, its nosewheel struck the Cessna, flipping it upside down and into the airliner’s right wing. The impact ruptured one of the airliner’s internal fuel tanks, which exploded. The Cessna was ripped in half. The sound and impact reverberated through the cockpit. As the crippled plane pitched downward and began a roll to the right, the captain muttered “easy baby, easy baby.” He asked the copilot, who was flying the plane at the time, to assess the damage. “It’s bad,” the copilot said. “We’re hit, man; we are hit.”
As the ground rushed towards them, the captain notified Lindbergh tower, “Tower, we’re going down, this is PSA,” to which the controller responded, “Okay, we’ll call the equipment for you.”
In the moments before impact the captain told those in the cockpit to “brace yourself.” An unidentified voice muttered, “I love you, ma.”
All 137 people aboard the two planes were killed, and seven people on the ground in the neighborhood adjacent to Balboa Park. Another nine were injured, and 22 dwellings destroyed or damaged.
In addition to faulting the pilots for failing to spot the traffic, the NTSB cited deficiencies in air traffic control procedures and the reliance of “see and avoid” in high traffic areas. In response, the FAA quickly slapped a TCA over the San Diego area; it proposed expanding TCAs to 43 other airports; banned all practice approaches at TCA airfields; and put the primary responsibility of aircraft separation on TCA controllers unless requested by the pilot. It also pushed research into the development of an onboard device that could alert pilots to midair collision threats.
The aviation industry was still waiting for such a device by 1986. On August 31, Aeromexico Flight 498, a DC-9 on approach into Los Angeles International Airport collided with a Piper Cherokee that had strayed into LAX airspace, killing all 65 aboard the two planes and another 15 people on the ground. The Piper was not talking to controllers, but had a functioning transponder. The controller working the sector at the time of the collision became distracted dealing with another small aircraft that had violated the airspace when he noticed the Aeromexico jet had disappeared. He told investigators he never saw the Cherokee on his scope, which did not have a conflict-alert capability. Investigators determined that an “atmospheric inversion” made the Piper’s primary radar return invisible and his transponder return wasn’t displayed due to a system configuration.
Once again, after blaming the pilots for failing to see and avoid the other aircraft, the NTSB cited the limitations of the air traffic control system—which is to say, equipment glitches and controller error. In the wake of the accident, Congress mandated that airliners be equipped onboard collision avoidance devices.
Within two years, the first Traffic Collision Avoidance Systems (TCAS) were certified for use, and in 1989 the FAA required the installations of such devices on all airliners with more than 30 seats. The devices work in conjunction with the onboard transponders installed on aircraft. TCAS calculates the relative altitude and speed of nearby aircraft. If an aircraft gets close enough to pose a possible threat, a pilot will receive an audible “Traffic Alert” that includes the range and bearing of the threat. If the system detects that a collision may be imminent, it issues an alarm and a “Resolution Advisory” to either climb or descend (while instructing the opposing aircraft to do the opposite, if similarly equipped).
Since the advent of TCAS, there has not been a fatal collision involving an airliner in the United States. See and avoid still prevails, but it is now getting a heavy assist from technology.
 |
| Traffic overlay on GPS moving map. |
GPS is about to put similar capability into all planes. ADS-B, or Automatic Dependent Surveillance—Broadcast, is a satellite-based technology in which a plane’s GPS position is broadcast, allowing it to be received by either ground or airborne-based equipment. Such equipment has been required aboard aircraft in Europe since 2017, and as of January 2020 will be required aboard U.S. aircraft flying in controlled airspace.
The devices have been available in the U.S. for several years. I’ve been flying with one for a couple of years. Other traffic is overlaid on the moving map of my GPS. It works quite well. I can see traffic onscreen long before I can see it by naked eye—most of it before air traffic control calls it out. The icons display which direction the traffic is moving and whether it is climbing or descending, allowing me to judge the potential hazard it may pose.
The device is limited in that it only displays aircraft also equipped with ADS-B—an issue that should disappear beginning next year. At times, it will display phantom traffic, which is annoying but quickly disregarded, or alert pilots to traffic that poses no real threat. Pilots now have the ability to see the planes around them in a way never before possible, an incredibly reassuring thing as you are approaching, say, an airport.
There are drawbacks. As with a lot of technology, a pilot can become too dependent or complacent, letting themselves get mesmerized by the TV screen, neglecting things that require more immediate attention, such as navigation, managing the engine, evaluating the weather, preparing for the next phase of flight or—Good Lord—looking out the window.
In this context, see and avoid has an entirely different meaning: It’s about recognizing risks and doing something to minimize them. Flying a plane is a perishable skill, and as pilots it’s important we know how to fly the plane with or without the assistance of these gadgets, because when an investigator is sifting through the ruins, in all likelihood the gadget is not going to be blamed for the accident.
Comments
Post a Comment