Testimony of the Honorable Mark V. Rosenker
National Transportation Safety Board
Subcommittee on Aviation Operations, Safety and Security
Committee on Commerce, Science, and Transportation
United States Senate
Aviation Safety: FAA’s Role in the Oversight of Commercial Air Carriers
June 10, 2009
Good afternoon. With your concurrence, Mr. Chairman, I would like to begin my testimony with a short summary of the National Transportation Safety Board’s (NTSB) actions to date regarding the investigation of the accident involving Colgan Air flight 3407. I want to emphasize that this is still an ongoing investigation and that there is significant work left for our investigative staff. My testimony today will therefore out of necessity be limited to those facts that we have identified to date, and I will steer clear of any analysis of what we have found so far and avoid any ultimate conclusions that might be drawn from that information.
On February 12, 2009, about 10:17 p.m. eastern standard time, Colgan Air flight 3407, a Bombardier Dash 8-Q400, crashed during an instrument approach to runway 23 at Buffalo-Niagara International Airport, Buffalo, New York. The crash site was in Clarence Center, New York, about 5 nautical miles northeast of the airport, and was mostly confined to a single residential house. The flight was operating as a Part 121 scheduled passenger flight from Liberty International Airport, Newark, New Jersey.
The four crew members and 45 passengers were killed, and the aircraft was destroyed by impact forces and post crash fire. One person in the house was also killed and two individuals escaped with minor injuries.
The flight crew reported for duty on the day of the accident at 1:30 p.m. However, the crew’s first two flights of the day were cancelled because of high winds at the departure airport. The accident flight, which had been delayed due to weather, departed Newark at 9:18 p.m. with a planned arrival time of 10:21 p.m.
The captain was the pilot flying the aircraft, and the cruise altitude was 16,000 feet. During the ascent to 16,000 feet, all de-ice systems were selected on and stayed on throughout the flight. About 40 minutes into the flight, the crew began the descent portion of the flight.
At 9:54 p.m., the captain briefed the airspeed for landing, which was to be 118 knots with the flaps set to 15 degrees. At 10:10 p.m., the flight crew discussed the build-up of ice on the windshield. At 10:12 p.m., the flight was cleared to 2300 feet and at 10:14 p.m., the airplane reached the assigned altitude. Over the next two minutes, with the autopilot engaged, power was reduced to near flight idle and the airspeed slowed from about 180 to about 135 knots. At 10:16 p.m., the crew lowered the landing gear. About 20 seconds later, the first officer moved the flaps from 5 to 10 degrees. Shortly afterward, the stick shaker activated, and the autopilot disengaged. The stick shaker is a stall warning mechanism that warns of slow airspeed and an approaching stall should the pilot take no action to remedy the situation. In this case, the stick shaker activated more than 25 knots before the stall airspeed.
The flight data recorder data from the airplane indicate that the crew added about 75% of available engine power and the captain moved the control column aft. This action was accompanied by the airplane pitching up, and a roll to the left, followed by a roll to the right, during which time the stick pusher activated and the flaps were retracted.
At the time of the accident, the weather at Buffalo was: winds from 250 degrees at 14 knots, visibility 3 miles in light snow and mist, a few clouds at 1100 feet, ceiling overcast at 2100 feet, and temperature of 1 degree Celsius.
Examination of the flight data recorder data and performance models shows that some ice accumulation was likely present on the airplane prior to the initial upset event, but that the airplane continued to respond as expected to flight control inputs throughout the accident sequence.
The engines exhibited evidence of power at impact. Flight control continuity could not be established due to the extensive impact and fire damage to the airplane.
On May 12, 2009, the NTSB began a 3-day en banc public hearing on the accident. The NTSB swore in 20 witnesses to discuss the following topics:
I would like to note that these issues are not relevant to regional airlines alone. They are pertinent to every airline operation, major air carriers as well as regional air carriers.
The investigation is continuing with aircraft performance and simulation work, additional interviews, reviews of policies and procedures, and further examination of selected wreckage. We’ve identified numerous safety issues that we will explore in significant detail.
During the hearing, the flight crew’s experience and training were examined. The captain received his type rating in the Dash 8 in November 2008, only a few months before the crash. He had a total flight time of 3,379 hours, with 1,030 hours as pilot-in-command and 110.7 hours in the Dash 8. The first officer received second-in-command privileges on the Dash 8 in March 2008. She reported 2,244 hours total pilot time with 774 hours in the Dash 8.
The captain had a history of multiple FAA certificate disapprovals involving flight checks conducted before his employment with Colgan. The captain did not initially pass flight tests for the Instrument flight rating (October, 1991), the Commercial Pilot certificate (May, 2002), and the multiengine certificate (April, 2004). In each case, with additional training, the captain subsequently passed the flight tests and was issued the rating or certificate.
In 1995, the NTSB issued 4 recommendations to the FAA to require an airline to evaluate an applicant pilot’s experience, skills, and abilities before hiring the individual. The FAA’s effort in response to these recommendations resulted in the Pilot Records Improvement Act (PRIA) of 1996 (Public Law 104-264, section 502, which is codified in 49 United States Code section 44703 (h), (i), and (j)). The PRIA required any company hiring a pilot for air transportation request and receive records from any organization that had previously employed the pilot during the previous 5 years. However, the PRIA does not require an airline to obtain FAA records of failed flight checks. Although validation of FAA ratings and certifications held by a pilot applicant is necessary in evaluating a pilot’s background, additional data contained in FAA records, including records of flight check failures and rechecks, would be beneficial for a potential employer to review and evaluate.
In 2005, the NTSB issued another recommendation to the FAA to require airlines, when considering an applicant for a pilot position, to perform a complete review of FAA airman records, including any notices of disapproval for flightchecks. In response to the NTSB’s recommendation, the FAA stated that Notices of Disapproval for flight checks for certificates and ratings are not among the records explicitly required by the Pilot Records Improvement Act (PRIA) of 1996, and therefore, to mandate that air carriers obtain such notices would require rulemaking or a change in the PRIA itself. The FAA indicated that such changes are likely to be time consuming and controversial. The FAA noted that some air carriers currently require applicants for pilot positions to sign a consent form permitting the FAA to release these records to the air carrier requesting them as part of the applicants' pre-employment screening. When this is done, the FAA furnishes these records to the air carrier without violating privacy laws. To date, the FAA has not issued any rulemaking to require airlines to obtain a release from all flight crew applicants to release their records to permit the airline to consider past performance in hiring decisions. These changes could also be made by modifying the statute, but to our knowledge, the FAA has not asked the Congress to do so. On November 7, 2007, the FAA issued Advisory Circular AC120-68D, which informs carriers that they can ask pilots to sign a consent form giving the carrier access to any Notices of Disapproval. The recommendation is currently classified “Open-Acceptable Alternate Response.”
The investigators also are pursuing why Colgan did not have a remedial training program in place as recommended in the FAA’s 2006 Safety Alert for Operators (SAFO) 06015, the purpose of which was to promote voluntary implementation of remedial training programs for pilots with persistent performance deficiencies.
Specifically, the SAFO provides guidance to safety directors on the development of programs to identify pilots with persistent performance deficiencies, those who have experienced multiple failures in training and proficiency checks. It was suggested that three objectives be accomplished: 1) review the entire performance history of any pilot in question; 2) provide additional remedial training as necessary; and 3) provide additional oversight by the certificate holder to ensure that performance deficiencies are effectively addressed and corrected.
The investigation is also exploring how commuting may have affected the pilots’ performance. Both pilots were based in Newark, New Jersey, but lived outside of the Newark area. The captain commuted to Newark from Tampa, Florida, three days before the accident, and spent the night in Colgan’s operations room the night before the accident. The first officer commuted from Seattle, Washington, on a “red eye” flight the night before the accident. She did not arrive into Newark until 6:30 a.m. the day of the accident flight, and there is evidence that she spent the day in the crew room.
Of the 137 Colgan pilots based at Newark in April 2009, 93 identified themselves as commuters. Forty-nine pilots have a commute greater than 400 miles, with 29 of these pilots living more than 1000 miles away.
During post-accident interviews, the Newark regional chief pilot said no restrictions were placed on pilots regarding commuting, but pilots had to meet schedule requirements. Colgan has a commuting policy that is outlined in its Flight Crewmember Policy Handbook. The handbook states “a commuting pilot is expected to report for duty in a timely manner.” A previous edition of the handbook stated that flight crewmembers should not attempt to commute to their base on the same day they are scheduled to work. This statement is not in the current handbook edition. Additionally, Colgan’s procedures do not allow pilots to sleep in the operations room.
The investigation is examining whether conversations inconsistent with the sterile cockpit rule (which prohibits crew members from engaging in non-essential conversation below 10,000 feet) impacted the pilots’ situational awareness of the decreasing airspeed. For example, there was a 3-minute discussion on the crew’s experience in icing conditions and training; this conversation occurred just a few minutes before the stick shaker activated and while the crew was executing the approach checklist.
Another issue that the investigation is pursuing is whether fatigue may have affected the flight crew’s performance. We know that on the day of the accident, the captain logged into Colgan’s crew scheduling computer system at 3:00 a.m. and 7:30 a.m. And we know that the first officer commuted to Newark on an overnight flight and was sending and receiving text messages periodically the day of the accident.
At the time of the accident, Colgan had a fatigue policy in place. The fatigue policy was covered in the basic indoctrination ground school. Colgan did not provide specific guidance to its pilots on fatigue management.
On April 29, 2009, Colgan issued an operations bulletin on crewmember fatigue. The bulletin reiterated the company’s fatigue policy and provided information to crewmembers on what causes fatigue, how to recognize the signs of fatigue, how fatigue affects performance, and how to combat fatigue by properly utilizing periods of rest.
Once again, the issues we are exploring in the Colgan investigation are not new issues and are not unique to the regional airlines. The NTSB has previously issued recommendations on stall training, stick pusher training, pilot certification and recurrent training records, remedial training for pilots, sterile cockpit, situational awareness, pilot monitoring skills, low airspeed alerting systems, pilot professionalism, and fatigue. (See attachments.)
As you may know, the NTSB maintains a list of Most Wanted Transportation Safety Improvements. Issues on this list are selected for follow-up and heightened awareness because the Board believes they will significantly enhance the safety of the nation’s transportation system, have a high level of public visibility and interest, and will otherwise benefit from being highlighted on the Most Wanted List. Of the six aviation issues currently on the Most Wanted List, two issue areas are in some manner related to the Colgan investigation. I would like to briefly explain the two issue areas, and recent FAA activities in response.
Both of these issue areas currently have a red timeliness classification indicating that the FAA’s response has not been acceptable from the NTSB’s perspective. In many cases, the FAA’s response has been slow in coming, allowing important safety issues that the NTSB has identified to remain unresolved for a lengthy period of time. The FAA has recently indicated that actions are being taken in response to some of these recommendations, and the NTSB is currently reviewing this information. Some of the details, and recent FAA actions for each area are:
Finally, I would like to address pilot training issues. As you are aware, on January 12, 2009, the FAA published an NPRM titled, “Qualification, Service, and Use of Crewmembers and Aircraft Dispatchers.” The notice proposes to amend the regulations for flight and cabin crewmembers and dispatcher training programs in domestic, flag, and supplemental operations. Proposed changes include requiring the use of flight simulation training devices (FSTD) in traditional flight crewmember training programs and adding training requirements in safety-critical areas. In addition, the proposal reorganizes qualifications and training requirements in the existing rule by moving several sections of advisory information to the regulatory section. The NPRM also addresses issues raised in numerous safety recommendations issued to the FAA by the NTSB; 13 of these recommendations remain open.
On May 7, 2009, the NTSB provided comments to the NPRM. While the NTSB generally supports the proposed rule changes, we suggested additional requirements, including substantive changes that would improve or enhance crew and dispatcher procedures, qualifications, and training and the replacement of advisory circulars and other recommended guidance with regulatory changes mandating compliance.
At an April 7, 2009, presentation on the NPRM, the NTSB was briefed that the FAA principle regarding training is “Train like you fly, and fly like you train.” The NTSB agrees with this principle and with several proposed initiatives that are especially appropriate for flight operations in today’s environment. For example, the NTSB supports the NPRM’s proposals for adding a continuous analysis process and FSTDs to training programs, requiring special hazards and environment training, and establishing qualifications for training centers and other 14 Code of Federal Regulations (CFR) Part 119 facilities. The NTSB also concurred with the FAA that it is important for flight crewmembers to be trained and evaluated in a complete flight crew environment, which means that, during training for pilot flying and pilot monitoring roles, crewmembers should occupy the seats for—and perform the duties of—the position for which they are being trained.
The NTSB is aware that, in the past, some considered upset recovery training to be inappropriate due to limitations in aerodynamic model fidelity of simulators; however, unusual attitudes do not equate to being outside the angle of attack and sideslip range of the aerodynamic model. Many, if not most, upsets occur well within this envelope. Therefore, the NTSB supports the “Airplane Upset Recovery Training Aid,” which is an FAA-industry effort referenced in the NPRM, and believes that training could be further improved by feedback to the pilot from the simulator. The training aid suggests that, in a scenario in which the pilot has maneuvered the simulator to an extremely high angle of attack or sideslip, there should be a change in the visual display when the aerodynamic envelope is exceeded; specifically, a color change would alert pilots that they are at an angle of attack or sideslip that should be avoided during recovery efforts.
The NTSB notes that some aircraft, such as the Saab 340 and the Bombardier CRJ, have experienced upsets due to premature stall caused by icing that disrupted the airflow over the wing or otherwise altered the aerodynamic stall characteristics of the wing or control surface. Because icing contamination can cause the critical angle of attack to be reduced considerably, these upsets can occur without warning. A stall roll-off departure from normal flight is often the flight crew’s first indication of an upset due to icing contamination; however, the NTSB has found that flight crews often do not apply decisive and timely recovery controls when this occurs, which results in prolonged upsets that increase the probability of ground impact. For aircraft that have experienced upsets due to icing contamination, the NTSB suggests that upset recovery training should include recognition of these excursions from normal flight attitudes and prompt application of proper recovery procedures.
Although the NPRM continues to encourage the traditional training approach to stall recovery (recovery from stick shaker), the NTSB is concerned that flight crews are not recognizing stalls and are not applying aggressive recovery procedures, as indicated by several aviation events. Among these events is the October 14, 2004, accident in which a Bombardier CL-600-2B19 crashed in Jefferson City, Missouri, when the flight crew was unable to recover after both engines flamed out as the result of a pilot-induced aerodynamic stall. Another example occurred during a December 22, 1996, accident in which a Douglas DC-8-63 experienced an uncontrolled flight into terrain in Narrows, Virginia, after the flying pilot applied inappropriate control inputs during a stall recovery attempt and the nonflying pilot failed to recognize, address, and correct these inappropriate control inputs. Because of examples like these, the NTSB advises that training in stall recovery should go beyond approach to stall to include training in recovery from a full stall condition. In addition, in cases when flight data are available (whether from flight tests or accidents/incidents), these data should be used to model stall behavior to facilitate training beyond the initial stall warning.
If the proposed rule becomes final, it would likely meet the intent of 5 of the 13 open safety recommendations related to crewmember training. The following is a list of the 13 recommendations and an explanation of whether or not the NPRM addresses each of them.
Amend 14 CFR Parts 121, 125, and 129 to require Traffic Alert and Collision Avoidance System [TCAS] flight simulator training for flight crews during initial and recurrent training. This training should familiarize the flight crews with TCAS presentations and require maneuvering in response to TCAS visual and aural alerts.
The NPRM contains requirements for TCAS training, as recommended. Therefore, the NPRM is responsive to the recommendation. If the NPRM (as currently presented) becomes a final rule, the NTSB would likely consider it an acceptable action, and the recommendation could be closed. The NTSB notes that this is currently the oldest open aviation recommendation.
Revise 14 CFR Section 121.445 to eliminate subparagraph (c), and require that all flight crewmembers meet the requirements for operation to or from a special airport, either by operating experience or pictorial means.
The NPRM proposes the following language for 14 CFR 121.1235(c): “The Administrator may determine that certain airports (due to items such as surrounding terrain, obstructions, or complex approach or departure procedures) are special airports requiring special airport qualifications and that certain areas or routes require a special type of navigation qualification.” In addition, special routes, areas, and airports for special operations are among the subjects in the NPRM’s list of required training. Therefore, the NPRM is responsive to the recommendation. If the NPRM (as currently presented) becomes a final rule, the NTSB would likely consider it an acceptable action, and this recommendation could be closed.
Revise the certification standards for Part 25 and for Part 23 (commuter category) aircraft to require that a flight simulator, suitable for flight crew training under Appendix H of Part 121, be available concurrent with the certification of any new aircraft type.
The NPRM proposes a requirement that a flight simulator be available for training. The NTSB has previously indicated that such a requirement would be an acceptable alternative response to a design requirement for an aircraft. Therefore, if the proposed rule becomes final, the NTSB would likely consider it an acceptable action, and this recommendation could be closed.
Require, by December 31, 1997, operators that conduct scheduled and nonscheduled services under 14 CFR Part 135 in Alaska to provide flight crews, during initial and recurrent training programs, aeronautical decision-making and judgment training that is tailored to the company’s flight operations and Alaska's aviation environment, and provide similar training for Federal Aviation Administration principal operations inspectors [POI] who are assigned to commuter airlines and air taxis in Alaska, so as to facilitate the inspectors’ approval and surveillance of the operators’ training programs.
The FAA has previously indicated to the NTSB that the NPRM would include aeronautical decision-making and judgment in the crew resource management portion of the proposed training rule. However, this Safety Recommendation is specific to Part 135 operations in Alaska, while the NPRM addresses Part 121 operations. Therefore, the FAA has not supplied a satisfactory response. Thus, the NPRM, as drafted, would not meet the intent of this recommendation, and the status would remain “Open—Unacceptable Response.”
Develop a controlled flight into terrain training [CFIT] program that includes realistic simulator exercises comparable to the successful windshear and rejected takeoff training programs and make training in such a program mandatory for all pilots operating under 14 CFR Part 121.
The NPRM proposes to require special hazards training, including methods for preventing CFIT and approach and landing accidents. Therefore, if this requirement is included in the final rule, the NTSB would likely consider it an acceptable action, and the recommendation could be closed.
Require 14 CFR Part 121 and 135 operators to provide training to flight crews in the recognition of and recovery from unusual attitudes and upset maneuvers, including upsets that occur while the aircraft is being controlled by automatic flight control systems, and unusual attitudes that result from flight control malfunctions and uncommanded flight control surface movements.
The NTSB is pleased that, in response to Safety Recommendation A-96-120, the NPRM includes training on recognizing and recovering from “special hazards,” which are sudden or unexpected aircraft upsets. The NTSB interprets that this proposal would also include a requirement that gives FAA POIs the authority to review and require changes to training programs that do not adequately address a special hazard. Lack of such authority was a concern identified during the NTSB’s investigation of a November 12, 2001, accident involving American Airlines flight 587, an Airbus Industrie A300‑605R.(1) During this investigation, the NTSB learned that the POI knew that aspects of American Airlines’ training program had undesirable effects; however, he lacked the authority to force American Airlines to change its program.
In addition, a topic covered in the special hazards training section of the NPRM is recovery from loss of control due to airplane design, airplane malfunction, human performance, and atmospheric conditions. The “Upset Recognition and Recovery” section of the NPRM lists a number of items that should be covered, including catastrophic damage due to rapidly reversing controls and the use of light pedal forces and small pedal movements to obtain the maximum rudder deflection as speed increases.
This recommendation is currently classified “Open—Unacceptable Response” because of the FAA’s delayed response. Although the NPRM proposes requirements for Part 121 operators, similar action for Part 135 operators will be needed before Safety Recommendation A-96-120 can be closed.
Require air carriers to adopt the operating procedure contained in the manufacturer’s airplane flight manual and subsequent approved revisions or provide written justification that an equivalent safety level results from an alternative procedure.
The FAA has previously indicated to the NTSB that the NPRM would address the issues in this recommendation. However, the NTSB did not see any language in the NPRM that specifically addresses Safety Recommendation A-98-102, which currently is classified “Open—Acceptable Response” pending a requirement for the recommended action.
Amend 14 [CFR] 121.417 to require participation in firefighting drills that involve actual or simulated fires during crewmember recurrent training and to require that those drills include realistic scenarios on recognizing potential signs of, locating, and fighting hidden fires.
The NPRM addresses the substantive issues in this recommendation. Although the NPRM does not propose to revise 14 CFR 121.417, it contains training requirements on the actions to take in the event of fire or smoke in the aircraft, including realistic drills with emphasis on combating hidden fires. This training includes simulated locations of hidden fires, such as behind sidewall panels, in overhead areas and panels, or in air conditioning vents. The NPRM also contains firefighting training requirements for flight attendants, including operation of each type of installed hand fire extinguisher. This recommendation is currently classified “Open—Unacceptable Response” pending a requirement for the recommended action. If the requirements proposed in the NPRM are enacted in the final rule, the NTSB would likely consider it an acceptable action, and this recommendation could be closed.
Require all 14 [CFR] Part 121 and 135 air carriers to incorporate bounced landing recovery techniques in their flight manuals and to teach these techniques during initial and recurrent training.
Although the NPRM contains detailed requirements for training on landing, the NTSB did not see anything in the NPRM related to bounced landing recovery techniques. This recommendation is currently classified “Open—Acceptable Alternate Response” pending the results of a survey indicating that all operators’ training programs include the recommendations in a safety alert for operators.
Require that all 14 [CFR] Part 91K, 121, and 135 operators establish procedures requiring all crewmembers on the flight deck to positively confirm and cross-check the airplane’s location at the assigned departure runway before crossing the hold short line for takeoff. This required guidance should be consistent with the guidance in Advisory Circular 120-74A and Safety Alert for Operators 06013 and 07003.
The NPRM contains training requirements related to runway safety. Special hazards topics must include how to ensure that takeoff clearance is received and that the correct runway is being entered for takeoff before crossing the hold-short line. This recommendation is currently classified “Open—Unacceptable Response” because of continuing delays in the issuance of this NPRM. If the NPRM becomes final, the proposed requirement is partly responsive to this recommendation because it addresses only Part 121 operators. Action will still be needed for Part 135 and Part 91 subpart K operators before this recommendation can be closed.
Require air carriers to revise their cabin crew training manuals and programs to ensure that the manuals and programs state that a door must remain open while the air conditioning (A/C) cart is connected, advise that the A/C cart can pressurize the airplane on the ground if all doors are closed, and warn about the dangers of opening any door while the air conditioning cart is supplying conditioned (cooled or heated) air to the cabin.
The NPRM proposes a requirement for training that will familiarize cabin crewmembers with each aircraft on which they will work. Among these aircraft familiarization requirements are cabin pressurization indicators and systems. However, the NPRM does not fully address the recommended action because it only addresses specific actions to take when the door remains open while the A/C cart is connected. This recommendation is currently classified, and would remain, “Open—Acceptable Response” pending timely and acceptable revisions to Notice 8400.35 and Order 8900.1.
Require 14 [CFR] Part 121, 135, and Part 91 subpart K operators to include, in their initial, upgrade, transition, and recurrent simulator training for turbojet airplanes, (1) decision-making for rejected landings below 50 feet along with a rapid reduction in visual cues and (2) practice in executing this maneuver.
The NPRM proposes a requirement to use a simulator for training on rejected landing maneuvers, including the initiation of a rejected landing between 30 and 50 feet above the runway. Thus, the NPRM addresses the second part of this recommendation (“practice in executing this maneuver”). In addition, although the NPRM did not specifically address decision-making, this topic may be covered during training in the maneuver. Safety Recommendation A‑08‑16 is currently classified “Open—Response Received.” The NPRM partially responds to the recommendation because it addresses only Part 121, and not Part 135 or Part 91 subpart K, carriers. Action for Part 135 and Part 91 subpart K operators will still be needed before this recommendation can be closed.
Require 14 [CFR] Part 121, 135, and Part 91 subpart K operators to include, in their initial, upgrade, transition, and recurrent simulator training for turbojet airplanes, practice for pilots in accomplishing maximum performance landings on contaminated runways.
The NTSB did not find any language describing how to accomplish maximum performance landings on contaminated runways in the NPRM. In addition, any proposed requirements associated with this NPRM would only apply to Part 121 carriers and not Part 135 or Part 91 subpart K operators. This recommendation is currently classified “Open—Response Received.”
Mr. Chairman, this concludes my testimony, and I will be glad to answer questions you may have.
Recommendation history on: