Good afternoon, everyone. Wes, I appreciate your kind introduction and would like to thank you for inviting me to kick off the 2009 International Runway Safety Summit. On behalf of my fellow Board Members, Vice Chairman Chris Hart and Member Robert Sumwalt, and the 391 men and women of the NTSB, I would like to recognize the hard work of the FAA and AAAE in making this important event possible.
As the Chairman of the NTSB, I find many of my quotes in the news describe action or inaction that ultimately led to a transportation accident. Well, for the press in the room today, I will not disappoint you, there will be some truth-telling about what isn’t working and what we’d like to see changed. But since I am opening this conference, I’d like to begin by acknowledging that things do work well most of the time. If we look at the numbers, the runway incursion rate in the United States over the past 4 years stands at about 6 runway incursions per 100,000 tower operations and this year, the numbers are looking even better. While these incursions represent close calls and are measured in feet rather than miles, it is not due to luck that we avert disaster on a daily basis, it is because of robust procedures, safe designs, and well-trained and alert controllers and pilots that these accidents are prevented. Many of the people in this room can take credit for this safety record, my hat is off to you for what you’ve accomplished to date, but the next question we have to ask is how can we get from 6 per 100,000 to zero?
As you know, the Safety Board exists to learn from incidents and accidents, so we take a very close look at those handful per 100,000 of close calls that result in incursions, overruns and wrong runway or even taxiway, takeoffs and landings. Of course we investigate the accidents that result in fatalities, but in the five and a half years that I have been at the Safety Board, we have investigated scores of incidents alongside the FAA and many of the stakeholders in the aviation system. We have learned a great deal from the people that survived to tell the tale and unfortunately from those that tragically lost their lives. These investigations have identified areas for action when it comes to runway safety and frankly we don’t want any more accidents to learn from, we need to move forward with what we’ve learned already. When we investigate accidents and reiterate recommendations we made previously, everyone in the aviation community is accountable, because that subsequent accident was preventable.
Runway safety has been one of the NTSB's top priorities for many years, and four specific issues currently occupy prominent places on the Board's list of "Most Wanted" safety improvements. We have asked the FAA and the industry to:
- Give immediate warnings of probable collisions/incursions directly to cockpit flight crews.
- Require specific air traffic control clearance for each runway crossing.
- Install cockpit moving map displays or automatic systems to alert pilots of attempted takeoffs from taxiways or wrong runways.
- Require landing distance assessment with an adequate safety margin.
A couple of years ago, the Safety Board expanded our Most Wanted List issue area from “runway incursions” to a more encompassing “runway safety” issue area. We did this because we recognized that even though events that are classified as “runway incursions” still represent a significant threat to aviation safety, there are other scenarios in the airport environment that are equally hazardous. We’ve looked at technology that can aid operations on the airport surface and help provide additional situational awareness to the crew. In that area, I noticed that you’ll be dedicating several sessions this week to looking at human factors, emerging technologies as well as systemic changes on the horizon. We’ve reviewed analysis performed by the Flight Safety Foundation that shows runway excursions accounted for approximately 30 percent of accidents worldwide between 1995 and 2006. A runway excursion occurs when an aircraft either overruns or undershoots a runway while landing or taking off and during my time at the Board, we have investigated accidents involving large air carrier overruns during winter conditions at Midway, Cleveland and Traverse City to name a few. Although a 15-percent factor is included in the European Aviation Safety Agency and Joint Airworthiness Authorities operational requirements for contaminated runway landing performance, the FAA has looked for voluntary operator compliance with performing landing distance assessments and applying a 15-percent safety margin. However, runway overruns continue to occur in the United States when the flight crews have not performed a landing distance assessment before landing on a contaminated runway. We firmly believe that implementation of our recommendations, some which are over 10 years old, will reduce the chances of runway collisions, the risk of a pilot mistakenly selecting an incorrect runway or taxiway, as occurred in the August 27, 2006, Comair flight 5191 accident in Lexington, Kentucky, or the likelihood of an excursion, especially as we move into the winter season with contaminated runways.
Improving runway safety has been a worldwide effort – no one is immune to the hazards encountered during surface operations. It is easy for those of us that work primarily in the United States, with its large number of aircraft and flight operations of all types, to become internally focused and perhaps not take full advantage of the experiences of airport operators and ATC service providers in other parts of the world. While doing a little homework for my remarks today I reviewed the "European Action Plan for the Prevention of Runway Incursions." The issues of hazard and risk addressed in the Eurocontrol plan were, as you might expect, very similar to those we encounter here in the United States. One of the biggest problem areas involves communications, the seemingly simple task of transferring information that humans perform constantly in all of our daily encounters. Deficient information transfer is an extremely common factor in runway safety incidents, both here and in other countries.
Those of us who reside in English-speaking countries are quite used to conducting all our business in English, so use of the worldwide standard language for ATC communications is fairly familiar and natural. Even so, pilots, vehicle operators, and controllers still manage to make mistakes and have misunderstandings. How much more difficult must it be to operate in an aviation environment where a controller is handling aircraft from several different countries and English is no one's native language? Tomorrow, you will hear from Captain Robert Bragg, a survivor of the worst aircraft accident in history. He will describe to you the chain of events that led to the collision of two Boeing 747s and the deaths of 583 passengers and crew – an accident that was partially attributable to communications issues. Perhaps we can learn something from those who encounter communications difficulties to a much greater degree than we do, yet still manage to operate safely day after day. I encourage you to take advantage of the opportunity provided by this conference to seek out professionals from other nations and gain their perspective on our common challenges.
In July 2000 the NTSB issued six recommendations to the FAA to amend various United States ATC procedures that, in the NTSB's judgment, unnecessarily added to the risks associated with airport surface operations. All but one of those six recommendations are still open with FAA responses in varying states of completion, and the remaining recommendation, regarding limitations on the use of position-and-hold procedures, has been closed – unacceptable action after the FAA declined to make the recommended changes. One of the recommendations asked the FAA to require the use of standard ICAO phraseology (excluding conditional clearances) for airport surface operations. In response, we were recently advised that the FAA soon plans to adopt a single change: the use of “line up and wait” instead of “position and hold” to instruct pilots to enter a runway and wait for takeoff clearance.
We needed to wait nine years for that?
It has taken over nine years to achieve partial acceptance of recommendations made in an area, runway safety, that is viewed as critical by both the FAA and the NTSB – and some of the FAA's recent responses to the NTSB on those July 2000 recommendations have asked for more time for further analysis. We really need to do better than that. Hundreds of thousands of passengers pass through the world's airports every single day. We owe it to them to address safety recommendations and suggestions, regardless of the source, in a timely and effective manner, and put safety improvements into practical use without delay. To do otherwise is to do less than our best work, and in the safety world, doing less than our best work is not acceptable performance.
Recent FAA statistics have shown a drop in reported category A and B runway incursions, those incidents where the circumstances indicate a significant risk of collision. Between fiscal years 2003 and 2008, there were 24 to 32 category A and B incursions reported each year. In FY 2009, only 12 such incidents were reported. While this is good news, it is too early to tell whether it is an aberration or the first sign of a long-term reduction in the number of these hazardous events. The NTSB has previously noted that the number of runway incursions typically tracks well with the overall number of flight operations in the system, going up when the number of operations increases and down when operations decrease. Given the recent economy-driven reductions in flight operations, we would expect the trend to continue and result in an observable decrease in the number of incursions. Looking at European statistics for reported category A and B incursions, there is precedent for rapid decreases in the numbers that proved to be short-lived. For example, in 1999 there were 7 reported A/B incursions, but in 2000 there were 41. In 2002, there were 19, but in 2003 there were 63. Since 2003, Europe has reported 48 to 69 A/B incursions each year. It will likely take a few years for statistics to confirm that our recent decrease in serious incursions is a trend, rather than an anomoly. In the meantime, we must continue our efforts to identify and mitigate the risks found in the airport environment. They’re still out there.
Let me mention an event that really underscores the point that the numbers are only one indicator of risk. Only a few weeks ago, on October 19th at about 6 a.m., a Delta Airlines Boeing 767 completing a flight from Rio de Janeiro to Atlanta was cleared to land on runway 27R but instead landed on the parallel taxiway just north of the runway. It was dark, and visibility was reported at 10 miles. The NTSB’s investigation is not yet complete, but preliminary information indicates that neither the flight crew nor the tower controllers realized that anything was wrong until the aircraft was rolling out on the taxiway. Luckily, traffic was light at the time and there were no aircraft or ground vehicles ahead. Obviously, a heavy jet unexpectedly traveling across the wrong part of the world’s busiest airport at 150 knots represents a serious hazard, and with only a slight change in circumstances this incident could have become a catastrophic accident. It is also noteworthy that because they landed on a parallel taxiway rather than a parallel runway, it is not classified as an incursion. Of course, regardless of how the incident is classified, we are very interested in what cues led a professional crew to believe that the taxiway was their assigned runway, why the controllers were unable to detect the improper operation and intervene, and whether better technology can be of use in such scenarios. I have no doubt that if this incident had resulted in a fatal ground collision, there would be immediate and understandable calls for upgrades to both ATC and aircraft warning systems. There is no need to wait for an actual accident; this event is entirely sufficient to show we have room for improvement, and I expect that once the circumstances are fully understood, appropriate corrective actions will follow to prevent similar future events at Atlanta and other airports.
The NTSB has long been on record in support of the need to make warning information directly available to flight crews as well as controllers. Where systems have included such capabilities, the benefits have been clear. TCAS has practically eliminated midair collisions involving equipped aircraft. Testing of ground safety systems such as runway status lights has shown them to be both effective and well-accepted as means of providing crews with additional reaction time to deal with potential hazards affecting their aircraft. I’d like to show you an example of the type of situation where we believe that direct warning to involved flight crews would have increased safety. On May 26, 2007, a Republic Airlines Embraer regional jet and a Skywest Airlines Embraer turboprop, together carrying 92 people, nearly collided at the intersection of runway 1L and 28R at San Francisco International Airport.
The Skywest aircraft was landing on runway 28R. The tower controller forgot about Skywest, and cleared the Republic flight for takeoff from runway 1L. The conflict was detected by the tower AMASS system about 15 seconds before the aircraft were projected to cross paths. AMASS then provided both aural and visual alerts to the tower controllers. After a few seconds, the tower controller instructed Skywest to stop, but because of the limited warning time, the crew was unable to avoid entering the intersection and stopped on runway 1L directly in front of the departing Republic flight. A collision was averted through some very heads-up flying by the captain of the Republic flight, who assumed control of the aircraft and lifted off early. According to crew statements, the Republic jet overflew the Skywest turboprop by about 30 feet.
[ANIMATION - Available at http://ntsb.gov/events/2007/Most-Wanted-Progress/AnimationDescription.htm, second down the page]
Several seconds elapsed between the alert and the controller’s “Hold –hold –hold” instruction to Skywest 5741. In this scenario, those few seconds likely made the difference between the crew stopping short of runway 1L or stopping ON runway 1L. NTSB review of AMASS performance showed that the system detected the impending arrival of the Skywest flight and considered runway 28R occupied before the Republic aircraft was cleared for takeoff. That would have caused the system to note a potential conflict between runway 28R traffic and aircraft landing or departing on 1L because the two runways intersect. Consequently, if SFO’s runways had been equipped with runway status lights driven by the AMASS conflict detection rules, when the Republic crew was cleared for takeoff they would have been looking at 3000 feet of red centerline lights ahead, and would likely have questioned the controller about the clearance rather than beginning their takeoff roll. Even if the Republic aircraft had tried to take off, a direct warning to the Skywest crew would have given them several more seconds to stop the aircraft before entering runway 1L, thereby averting the potential collision.
Serious events are not limited to the major airports equipped with AMASS or ASDE-X. Some of the closest near-collisions recently investigated by the NTSB occurred at smaller airports, like Allentown-Bethlehem, Pennsylvania, and Fresno, California.
[Show slide 1 - 214 KB]
In the Allentown incident, a regional jet nearly collided with a landing Cessna 172 that did not turn off the runway where the tower controller expected it to. The controller cleared the regional jet for takeoff without ensuring that the runway was vacant. During takeoff roll, the alert first officer of the RJ spotted a light from the Cessna ahead and called for the captain to abort the takeoff. In this slide, the abort began at the marker on the lower left, and ended at the marker at the upper right.
[Show slide 2 - 62 KB]
During the maximum effort stop, the RJ slid past the Cessna on the left side of the runway. Inspection of the runway surface showed that the RJ created about 1220 feet of skid marks. Here we see the marks created as the aircraft swerved off the centerline.
[Show slide 3 - PDF, 55 KB]
The crew estimated that they were still moving about 40 mph when they passed the Cessna, and missed colliding with it by about 10 feet. If not for the alert response of the first officer, it is likely that this would have been a fatal high-speed runway collision.
For several years the FAA has been evaluating various technologies that can provide a direct warning to pilots of a runway incursion risk. Whether it is Runway Status Lights (RWSLs), Final Approach Runway Occupancy Signal (FAROS), or efforts to develop a minimum operational performance specification for an ADS-B based surface alerting application that would provide a direct cockpit warning of a runway incursion risk, the FAA is taking commendable actions, but they are just too slow. Many of these systems are still being evaluated. In 2011, 11 years after the issuance of the recommendation for a direct warning to the cockpit, only 22 airports will have RWSLs installed. The FAA’s testing of and plans to develop standards for a system based on ADS-B to provide a direct cockpit warning are also commendable, although the FAA’s proposed requirements for ADS-B do not include a requirement for ADS-B In, which will be necessary for such a system.
The FAA announced that it will provide funding for users who agree to equip their aircraft with an Electronic Flight Bag, which includes Moving Map Displays, or an aural runway alerting system. Again, The FAA’s program to encourage users to equip their aircraft with moving map displays is commendable, but it is not a requirement, and the program is limited to no more than $5 million. As a result, the program is not likely to result in widespread adoption of moving map technology.
Most everyone in this room has some role in runway safety. This is what I would like you to think about while you’re here this week: what can you do either at the conference or right after you get back to the office that will help move some aspect of runway safety forward? Is there something “stuck” in your in-box or somewhere else in the process that you can shake loose? Frederick Brooks, a well-known project manager and engineer, asked, “How does a project get to be a year late? One day at a time.” I’ll follow his lead: How do safety improvements end up taking ten years to deliver? They get delayed one day at a time…and every one of those days may be the day when a preventable accident occurs as the result of something we were “just about ready to fix.”
Gatherings such as this one, which bring together professionals from many disciplines and many countries to discuss common problems, can help all of us to succeed in delivering real safety results. Runway safety is not the sole responsibility of the Administrator or even a tower manager or a particular controller; runway safety begins when an airport is designed, standard phraseology is used, and well before pilots brief an approach. Safety is accomplished through cooperation, collaboration and commitment. I have been briefed on the ATSAP program and believe that opening the lines of communication between all of the stakeholders to solve problems before they result in accidents is a step in the right direction.
I note there will be a session on detecting systemic risk and another one on safety management systems. I suspect that good information and feedback will be critical to these efforts. Perhaps it is time to reinvent our approach to runway safety. If we are going to go from six per 100,000 to zero, we will have to identify ways to improve not just individuals, but the system. We can no longer focus on who made the last mistake in the chain of events. In the end, it doesn’t matter to the victims of an accident if the last mistake was made by a controller or by a pilot, but that the error was not detected in time to prevent the accident from occurring. We must move away from the processes that were developed, years ago, with a view toward who should be punished for the error, to a system that is focused on how to identify and fix the problems as they arise, such as fatigue for both pilots and controllers, another issue on our Most Wanted List.
A more collaborative approach would treat the problems the same, irrespective of who made the last mistake. Earlier this year, the aviation community recognized the effectiveness of a system approach to safety, when the Commercial Aviation Safety Team was awarded the 2008 Robert J. Collier trophy for their work which resulted in a 65% reduction in the U.S. fatal commercial aviation accident rate from 1997-2007 (including two years with no commercial scheduled airline fatalities). This accomplishment is notable, because they achieved these significant gains despite the fact that the US safety record was already considered very good in 1997.
CAST developed system solutions for system problems. Perhaps it is time for runway safety programs to adopt a similar approach. That may be our best hope for the path to zero!
Along with my fellow NTSB member Robert Sumwalt, who is a panelist in the next plenary session, I look forward to discussing runway safety or any other aspect of the NTSB’s work with you. Thank you for inviting me to open the conference, and I appreciate your kind attention.