NTSB hearing is for the birds
While only one in five birdstrikes is ever reported, one unusual strike this past January 15, with an even more remarkable outcome, gained global attention and might bring advisories and eventually new certification and training standards to operators of all turbine-powered aircraft.
On June 9 the NTSB opened a three-day hearing to focus on the emergency ditching by an Airbus A320 (operating as US Airways flight 1549) on New York’s Hudson River just minutes after take- off from La Guardia Airport, following collision with a large flock of birds. All 150 passengers and the five crewmembers escaped alive, though one flight attendant and four passengers were seriously injured.
The captain’s and crew’s skill in ditching and evacuating the transport aircraft seized the public’s attention, and the captain’s testimony drew crowds and coverage. But the heart of the NTSB hearing, largely obscured by the narrow media focus, was to help lower the hazard from wildlife strikes, and to consider emergency landing procedures–particularly on water–for the crew of any turbine-engine aircraft.
The hearing opened with the now-celebrity Capt. Chesley Sullenberger, who listed the steps he took to find an emergency field and restart engines using the dual engine failure checklist, before deciding to make a managed descent into water. Sullenberger glided the jetliner to a river landing near a line of commercial piers to enable a swift response.
Not only the black box and cockpit voice recorder but the entire airframe were recovered, including the left engine, which separated from the wing and sank when it struck the water. The completeness and speed of the recovery has enabled an exhaustive review and even a computer animation of the entire flight path, which lasted five minutes and eight seconds from rotation at La Guardia to splashdown in the Hudson.
Experts from the Smithsonian Institution’s National Museum of Natural History concluded that it was a large flock of eight-pound Canada geese that began striking the Airbus a minute and 37 seconds into its climb. Such a bird is nearly three pounds larger than the bird that current engines are certified to ingest and still fly.
Witnesses at the hearing discussed the growing and largely unreported hazard of wildlife to turbine aircraft of all sizes, as well as research into avoiding such conflict, and called for more effective design of emergency checklists and training for crew resource management. The NTSB will use the input to form any recommendations it might issue when the investigation is complete. Though a final report might take months, the focus of NTSB questioning suggests that its recommendations will lead to new advisories and possibly to a call for a regulatory response.
Investigators questioned the adequacy of current certification requirements for turbine engines to withstand birdstrikes, as well as the voluntary nature of reporting such strikes under Parts 135 and 91, which has produced reports in only 21 percent of cases. Under Part 121 rules, the FAA and European agencies require operators to supply such data, though the frequency and comprehensiveness of such data collection was also brought into question.
Testifying that wildlife hazards near airports are growing worldwide, witnesses discussed the latest research for detection and to avoid collision.
The NTSB also considered the insight of witnesses regarding the most effective design of emergency checklists and suggested that training simulations might need to drill pilots in circumstances where adherence to the checklist produces no result, and when to abandon that list so as to focus on other priorities.
“Were you surprised when you saw birds in the windshield?” asked an opening panelist. “Yes,” replied Sullenberger, who then waited in disciplined silence for the next question. After Sullenberger’s testimony, half the audience, which included more than 40 media organizations, left. By the third day of hearings, fewer than 30 people remained in the audience as the hearings turned to the less sensational but wider-reaching topics of training, certification and reducing bird hazards.
Sullenberger testified that in training for a ditching, he was advised to set down near a vessel when possible, and the first thought that occurred to him, based on a sightseeing visit, was the USS Intrepid museum, a decommissioned aircraft carrier docked on the west side of midtown Manhattan. Captain Sullenberger said that 20,000 hours of flight experience had “internalized” what needed to be done, and this is why he had not exhibited “a startled response.” One panelist said he “wanted to bottle this mindset” for other pilots, but later in the hearing, experts debated whether memorizing versus reading checklists produces safer results, a question likely to receive a call for more study in the NTSB final report.
“Memory items should be minimized,” said Dr. Barbara Burian, a research psychologist at NASA.
Sullenberger alerted his passengers via the public-address system, “This is the captain, brace for impact.” Experts later questioned whether a more effective command in a future ditching might be to “brace for a water impact.”
Billy Mitchell, a passenger who had been seated in 25A of the Airbus, testified that he and several passengers did not adopt the full brace position so as to keep an eye out the window to determine precisely when it would come.
“I’ve seen bracing in the movies but I’m not sure anyone is really clear on what it means to brace,” said Mitchell, who added that based on what he’d read of surviving automotive accidents, he also wanted to stay limp, rather than clamp hands to the seat as instructed by the safety card. Mitchell struggled for 10 seconds to find and remove his life vest beneath the seat, but could not. Once in the raft, Mitchell and others were worried that none had a knife with which to cut their tether to the wing as the aircraft sank. Panelists hinted via their questioning that such briefing points might benefit all passengers, including those on business jets.
Most survivable water impact accidents have occurred within 50 miles of shore or in a body of water such as a lake or river, and on approach or departure rather than from altitude. More than 179 fully certified U.S. airports are within five miles of a significant body of water. The NTSB concluded that the single most important factor in surviving a ditching is the ability of flight and cabin crew to assist passengers effectively.
The NTSB has previously recommended that in aircraft with low-density seating or seats spaced farther apart than on the airlines (many business jet cabins, for example) passengers should brace for impact by grasping their ankles or legs to reduce flailing or, if not able, to wrap their arms under their legs. Heads should face down in their laps and not turn to either side. Feet should be flat on the floor and slightly ahead of the seat edge.
Passengers facing to the aft, a typical configuration for some seats in business jets, should rest their head on the seat back or bulkhead behind them. The NTSB stressed that aft-facing passengers should not place their hands behind their head, as was advised at one time, but rather place hands in their laps or clutch the sides of their seats.
Questions arose in the NTSB hearing as to whether flight attendants would have prepared passengers better had they known they faced a water landing, rather than the generic “brace for impact.”
“We as an industry need to think about how we can better design these checklists, based on some hard empirical data,” said research psychologist Burian. “Often in real life the checklist does not work out the way you expect.” Burian suggested adding a training scenario where the pilot follows a checklist but the situation does not improve.
“You don’t want to create a checklist that gets people sucked into trying the same things over and over if they’re not working,” she said, suggesting that “opt-out” points be considered for certain checklists. Asked why she thought Sullenberger reacted well despite having no training in how to handle a dual engine failure low over water, she said that practice in making decisions generalizes across other areas, and recommended studies on how to reinforce such practice.
A computer animation of the flight path with time coding and the ATC transcript (not the cockpit voice recorder) is available online. A PowerPoint summary and photographs of the accident sequence, evacuation, rescue and recovery prepared by the NTSB investigator in charge (IIC) can be downloaded.
After ingesting the birds, both engines showed a decrease in fuel flow and rotor speed while the exhaust gas temperature (EGT) increased toward the redline, which most often suggests insufficient airflow through the engine. The pilots tried to restart both engines using the dual engine failure checklist, even though neither engine had flamed out. Passenger Mitchell testified that the engine was “a bonfire” but a witness from engine manufacturer CFM later said that the engine was not on fire, but that unburned fuel was passing through and then igniting.
The number-two engine could not be restarted and it shut down. Actions taken to restart the number-one engine actually degraded the rotor speeds and the fuel flow enough to prevent it from sustaining flight.
“The rate of descent of US Airways 1549 was much higher than that assumed for the ditching certification, at 13 feet per second instead of 3.5 feet per second, leading to external pressures estimated to be greater than twice the certification values,” concluded Airbus. Despite this high vertical velocity, all occupants were protected from major injury because the fuselage did not break apart either on impact or from the motion in the water. Though fast, a gentle flare helped the belly take the beating and reduce acceleration levels felt by the passengers, but parts of the sub-cargo floor were crushed, as were lower panels. The impact, and soon the pressure and suction of the water, cracked the skin and allowed water to enter.
Airbus provided the NTSB with a copy of its Getting to Grips with Cabin Safety, which emphasizes that cabin training covers ditching in large bodies of water and mainly with plenty of notice of the impending water landing. In a famous example in 1962 over Sitka Sound, Alaska, the crew had 45 minutes to prepare the passengers and actually rehearse each phase. All survived.
Yet most water-contact accidents occur after little warning during a takeoff or landing, and sometimes in darkness. Airbus recommended a “silent review,” with crewmembers asking themselves what commands should be used and what extra information be given to passengers before or during what it calls the impact phase, egress phase and survival phase.
As for ditching technique, the NTSB reemphasized the FAA/ AIM procedures, paragraph E(e), which underscore the critical importance of attention to power management (assuming there is still power to manage).
“The use of power in ditching is so important that when [the crew] is certain that the coast cannot be reached, the pilot should, if possible, ditch before fuel is exhausted,” advises the AIM. “The use of power in a night or instrument ditching is far more essential than under daylight contact conditions.”
The Costs of Birdstrikes
Ongoing research centers not on ditching but on preventing the type of bird and wildlife strikes that might force an emergency landing. Panelists at last month’s hearing heard testimony that aircraft collisions with wildlife cost the civil aviation industry in the U.S. about $625 million per year. Some 98 percent of strikes involve birds. Between 1988 and April this year, at least 229 people have died and 194 aircraft have been destroyed as a result.
Since 1990, the FAA has collected voluntary wildlife strike reports for a database that now holds 89,734 records. Yet in 2005 the USDA found that only 21 percent of known strikes are actually captured. In 1999 the NTSB issued Safety Recommendation A-99-091, proposing that all operators be required to report birdstrikes. The recommendation was not accepted, and at the time of January’s accident, birdstrike reporting was voluntary. The number of strikes reported annually in the U.S., though, increased from 1,759 in 1990 to 7,666 in 2007.
Researchers have concluded that modern turbofan-powered aircraft have quieter engines and are less obvious to birds, and fewer fowl are scared away. At the same time, habitat has been squeezed by development, making airports a more inviting patch of open grass and pavement for feeding and resting. Since 1980, commercial aircraft movements have grown about two percent yearly.
And there are more birds of nearly every type. Robert Ganley, manager of the FAA’s engine and propeller directorate standards staff, added that the goose population alone in the U.S. has increased four-fold.
“It’s fair to expect that the likelihood of bird encounters will increase,” said Ganley.
After certain types of wildlife encounter as described in 14 CFR Part 139.337, airports are required to conduct a hazard assessment using the guidance in AC 150/ 520033B. La Guardia Airport has studied different kinds of grass that might be less attractive to geese and has seen some reduction, but substantial hazards remain at airports with similar features.
Though La Guardia tries to drain any standing water quickly, it still attracts wildlife. The dense development nearby is attractive for perching, and the airport is trying to modify nesting points with bird deterrent devices. Signage is posted to cut down on feeding of wildlife and to contain food and waste. At La Guardia, two full-time bird supervisors monitor the tidal flat areas, with special attention to the spots near Runway 31.
Nationwide, 93 percent of wildlife strikes occur below 3,500 feet agl. Of the 26 large aircraft destroyed by birdstrikes since 1960, 24 were from collisions below 500 feet agl.
The FAA and the U.S. Navy Space and Naval Systems Center launched a research project in 2000 to assess low-cost, mobile radars. The FAA reported last month that “it is not yet known if this system would be capable of providing alerts that would be operationally suitable for making specific decisions on landing or takeoff,” though studies continue. Some scientists have suggested that landing lights can be manipulated to repulse birds.
Most current research aims to advance the capability of turbine engines to withstand bird ingestion.
“Our safety objective is to limit accidents due to multi-engine power loss, and significant accidents due to flocking bird ingestion,” said Ganley. The FAA target is for such an event to occur no more than once in a billion flight hours. “This rate is sometimes referred to as extremely improbable, and reflects the long-term FAA view of an acceptable risk,” he explained.
Along with the EASA, Ganley’s team has begun updating the bird ingestion rulemaking database, which contains records for 325 million civil turbine flights worldwide and about 8,100 reports over the last 30 years. Though Ganley did not present a timeframe, when the update is complete the status of birdstrike-related certification will be reevaluated and such data could drive policy changes.
“In conclusion, it’s not reasonable to expect that engines can accommodate all threats under all conditions. We’re shooting for a safe shutdown, with no fires and no release of materials through the engine compartment,” he summarized. “There’s an inherent conservatism built into the certification standards,” noting that if a 4.1-pound bird hits an engine certified for four-pound birds it will probably withstand the ingestion.
Mark Boutillier, an engineer in Ganley’s group at the FAA, said that the collection of birdstrike data for Part 121 operations is more rigorous and accurate than the wildlife-strike database, since it was mandatory as part of rulemaking. The Part 121 data is also enhanced by regular contributions from field representatives of the major airframe manufacturers who collect statistics yearly from the airlines.
Bird ingestion tests and certification apply to various sorting criteria, and the weight limit for each class of bird has grown since the certification standards were first introduced. Recommendations might emerge to raise them anew, following January’s strike of eight-pound geese.
There is a medium flocking-bird test, a large flocking-bird test and a large single-bird test, all performed under a series of critical flight conditions and at various locations on the airframe and engine. Birds that strike aircraft as a flock are rarer than single birdstrikes.
“Bigger engines do get bigger birds and more of them,” said Boutillier, citing a statistical correlation, explaining why the larger the engine, the higher the respective weight of bird it must sustain. “The regulatory/certification threshold has been rising in stages. Generally, the standard is to allow 20 minutes of continued operation with throttle movement and sufficient thrust to return to the airport,” he summarized. Original standards were issued in 1974, with major revisions in 2000 and 2007.
Amendment 33-24, effective on Nov. 16, 2007, introduced a new class of bird for testing: the large flocking bird whose weight requirements were four pounds, 4.5 pounds and 5.5 pounds, depending on engine size, and the run-on test (as a decreasing sliding scale) achieving 90 percent of maximum takeoff power for 20 minutes after ingestion. If the FAA finds that the bird ingestion rate is higher than the theoretical models have assumed–and this might happen if reporting becomes more complete–the certification standard will be reviewed.
Les McVey, a flight safety investigation engineer with CFM International, testified about the compliance standard for CFM56-5B4/P engines. “The standards are so that no uncontained failures occur, and that the aircraft remains controllable. Historically, small flocking birds have not been considered much of a hazard,” said McVey, noting that there are requirements to withstand flocks of small birds, but no testing.
Witnesses were questioned about after-market equipment that might contain or deflect birdstrikes, such as screens to install in front of the engine intake. McVey said that no such devices have been requested by manufacturers as part of their certification package, though the FAA has gotten requests from operators to evaluate add-on devices.
“We have concerns about such a device,” said Boutillier. “There may be turbulent flow upstream of the engines with a screen and that would make them more prone to a surge stall. They can also drive the airfoils closer to resonance. Icing conditions are another concern; a screen would probably accrue ice readily, and again there are distortion issues, and if the ice is shed it would pass through the engine and likely cause damage,” he continued.
McVey was asked if future engines could have internal sensors to gauge the effects of a large ingestion, possibly based on digital engine controls similar to those that detect an impending compressor stall. Better alerts to the crew, based on internal sensors, might add reaction time or give options, including the option to eliminate certain checklists as futile. McVey promised to study the idea carefully, but cautioned that he “does not want to take any options or authority away from the crew.”
As for improvements to design, McVey ticked off the progress in engines that’s been made in terms of added thrust, lower weight and better fuel efficiency, along with lower noise emissions, saying that these exist in a delicate balance.
“To try to improve single components [to handle] a very large bird could upset this balance. I don’t see any reasonable engineering solution. The heaviest bird [in a flock] we’re required to ingest in the core is 5.5 pounds, and these geese [that struck the Airbus] were eight pounds. This would not be an incremental change; it would be a major new technology and design.”
It’s impossible to say whether any engine under a new amendment would sustain the type of strike that befell the US Airways A320, with geese ingested by both engines concurrently, and of such a large size, added Boutillier. “The weight standard could increase,” he said. “We just hope that the threat does not increase
at the same time.”