Air Safety Foundation Executive Director Bruce Landsberg is a 5,000-plus-hour ATP.
The accident involving Comair 5191, the Bombardier Regional Jet that attempted to launch in late August from a short runway in Lexington, Kentucky, is still under investigation. The general media have moved on to other things, but for pilots, even before the NTSB makes a formal pronouncement, it's pretty clear what happened. There are lots of theories as to the cause and what could be done to prevent this in the future. Here are a few observations to consider.
Comair 5191 was cleared to depart from Runway 22 but lined up, incorrectly, on Runway 26. Runway 22 is 7,003 feet long, is 150 feet wide, and was fully lit except for centerline lights, which were notamed out of service. Runway 26 is 3,500 feet long, is also 150 feet wide however it was marked as only 75 feet wide and was unlit. The RJ went off the end of the shorter runway, may have become briefly airborne, and struck a berm and some trees. There were 49 fatalities and one critical injury.
The fact that the runway used by the RJ crew was unlit begs the question of how dark it really was at the takeoff time. The NTSB lists the accident time as 6:07 a.m., a half-hour before official civil morning twilight at 6:37 a.m., and nearly an hour before that day's official sunrise time of 7:05 a.m. On a truly dark night, an unlit runway is a not-so-subtle tip that maybe you're not where you're supposed to be, but as morning twilight starts filtering in, the hint becomes increasingly vague.
There will be pages of recommendations on all manner of things that could or should be changed to prevent accidents like this one. For those of us who fly single pilot in light aircraft and form the largest part of our own safety system, here are some thoughts to consider on your next flight. There is nothing new about getting onto the wrong runway; pilots have been getting misaligned occasionally since runways were numbered. Fortunately, accidents are extremely rare, but there are some techniques that would likely have prevented this disaster.
Think of taxiing as the first leg to a critical waypoint. It's just as important as any flight segment, and the consequences of a miscue or wrong turn can be inconvenient, expensive, or tragic. Multitasking on the ground is a balancing act, even more than in the air because of the extremely close time and distance tolerances. The distraction potential is significant and distractions are always present in accident chains.
Particularly at airports with more than a single parallel taxiway, always use an airport taxi diagram, if one is available. Taxi diagrams are published for just about all airports with control towers. VFR-only pilots can easily download taxi diagrams for free from the AOPA or AOPA Air Safety Foundation Web sites. For the few pilots who don't have a computer, chances are excellent that the FBO, hotel, or other business on or close to the airport will have access to the Web. Don't leave home without the taxi diagram!
In regard to the Comair accident, at this writing, there is some confusion about the accuracy of the taxi diagram and whether the crew had a current version — that will be resolved in the investigation. There was a notam on Lexington taxiway closures and construction, but it was probably buried in 20 pages of irrelevant notices about laser light shows and low-level unlit towers miles from airports not even remotely connected to the flight. I'll just say that the notam system continues as an unmitigated mess of which the FAA should be ashamed. It needs to be fixed soon. Truly important notams are buried among several hundred noncritical notices, and then pilots are hung out to dry when they don't see the need-to-know information.
Anytime there is more than one runway going more or less in the same direction, either parallel or within 90 degrees, the odds of lining up on the wrong one naturally increase. Equally troublesome are bad-karma runway numbers such as 13 and 31, 2 and 20, 23 and 32 that just beg for pilot confusion. Granted, they are not even close to being physically aligned, but just a short bout of dyslexia can be problematic, if not tragic. Any doubt about the path or limited visibility of intersections warrants a request for "progressive" taxi instructions. Verify and verify again.
In jet and turboprop crew operations, departure time is extremely busy as the crews configure the aircraft to go. If ever there was truth to the saying "time is money," this is it, imagine burning 10 to 20 gallons just while taxiing out, and then multiply that by the 850 flights that Comair alone launches every day. This is a noticeable expense. It applies to our airplanes as well — it's just that the numbers are smaller.
Pretakeoff prep includes programming the flight management system or GPS and avionics, negotiating or renegotiating with air traffic control, and just setting up the machine to fly. There is usually a division of duties where one pilot does the driving and the other does pretty much everything else.
On larger aircraft, ground steering is done only by the captain, with a wheel or tiller located on the far left side of the cockpit (although some new aircraft are getting redundancy here, as well). This can make for an out-of-the-loop situation for the first officer (FO). In Lexington, according to the NTSB, the captain taxied out, lined the aircraft up on the wrong runway, and handed control over to the FO. In this scenario, the FO moves from "pilot-not-flying" to "pilot-flying" in a few seconds. Depending on company procedure, he may have been head down and dealing with the paperwork, checklists, and other myriad details that accompany a jet launch. Automation is wonderful, but sometimes the care and feeding of it adds to a high-workload environment. We'll learn more as the Flight 5191 cockpit voice recorder is analyzed.
Here are some recommendations that have worked for me in single-pilot operations over the years at both busy and backwoods airports. Get your clearance and program the avionics before engine start, if possible. If not, do it before starting to taxi. If that doesn't work because of the inability to talk to ATC on the ramp, do it at the run-up pad. The more complex the airport, or the darker or foggier the conditions, the more attention must be devoted to the all-important first leg of the trip: getting to the correct runway. Moving taxi maps can help a lot, but this may be a little above the budget for many of us. Taxiing and programming simultaneously invites mistakes in both arenas. Garmin's SafeTaxi is a big help — on the GPSMap 496.
A technique suggestion: Set the heading bug and the horizontal situation indicator (HSI) to the departure runway heading. Once launched, you can reset the bug and omni-bearing selector to the departure leg with little difficulty. This method encourages one final check upon runway lineup that the heading indicator, HSI, flight director, and bug are all in agreement. Don't have all that hardware? At least check the magnetic compass and directional gyro. Everything should point in the same direction as the runway numbers.
It was widely reported that in 1993 another airliner lined up on Lexington's Runway 26 instead of Runway 22, the identical situation as that of Flight 5191. The crew was looking at the weather radar and then checked the heading indicator just as the tower cancelled the takeoff clearance. A NASA Aviation Safety Reporting System report was filed and that was the end of it. The checks and balances in the system worked. The crew claims to have caught the error but the tower clinched the save.
Now fast-forward tens of thousands of takeoffs and more than a decade later. Is it the tower's responsibility to ensure that pilots take off on the proper runway? My take on it is "no and yes." The tower's primary job is to sequence traffic for arrivals and departures — the controller did that and fulfilled his responsibility. "Yes" comes from the obligation that on a time-available, workload-permitting basis, controllers should watch what's going on and help out as needed, as happened in the earlier Lexington event. It's a backup — not a guarantee.
Does safety require a minimum of two controllers in every tower all the time — one to manage ground ops and the other to deal with radar and other chores? FAA rules say two controllers are essential, but does it make sense during the midwatch when only a few aircraft are stirring? That might be exactly the time when the crew concept for the tower is essential.
On the other hand, if two controllers are required, might some communities with lower-activity airports decide to shutter their tower or limit hours of operation because of the higher staffing cost? What effect does that have on safety? Pulling two shifts back to back, as the Lexington tower controller reportedly did (finishing at 2 p.m. and returning to duty about midnight, with about two hours of sleep), should not be allowed. It doesn't work from any human-factors perspective.
It's unlikely that the airlines will, or should, change their departure procedures for every takeoff. Some airports are easy and some are not, but we should be thinking about the complexity of the taxi routing, the potential for a runway incursion, or the possibility of a misalignment. Some ground operations may require both pilots to devote full attention to taxiing. Timing is another aspect. When I've made mistakes on departures with either ATC or the aircraft, it is invariably because of rushing.
There are tens of millions of takeoffs every year and the system works very well. A cursory review of FAA statistics shows that there have only been about half a dozen attempted takeoffs on the wrong runway or on a taxiway at towered fields in the past several years. By any measure that's a pretty impressive record. In the final analysis, much of what we do in aviation is routine and repetitive. As pilots in command we also must remember that certain activities have to be done correctly every time or there will be a nonroutine outcome. Runway identification falls into that category.
For more information on runway operations, visit the Web site.
Related Articles
