By Capt. Tom Duke (Ret.)
Air Line Pilot, September 2003, p. 20

In 1997, the White House, followed up by the National Civil Aviation Review Commission, challenged the FAA, U.S. airlines, and the aviation industry to reduce in 10 years the fatal accident rate by 80 percent over the 1994-96 rates per 100,000 departures. The challenge also extended to reducing the rate of overall accidents (fatal and nonfatal) and reducing the rate of fatalities in each major type of accident. Further, the airline industry, government, associations, and unions were to collaborate and provide the research to support that challenge. The FAA responded to the challenge with the project Safer Skies.

At the North American Aviation Safety Conference, held February 4-6 in Atlanta and sponsored by SAE Aerospace and the Flight Safety Foundation, both FAA Administrator Marion Blakey and NTSB member Carol Carmody congratulated the U.S. airline industry (including scheduled Part 135 carriers) for flying fatality-free in calendar year 2002. Blakey stated, "The goal of reaching an 80 percent reduction [in the U.S. airline accident rate] by 2007 is well on track, with a 49 percent reduction at the halfway point.

"Last month’s [January 2003] accident in Charlotte shows that we have a long way to go," Blakey said. "We must change our reactive nature and adopt a proactive approach involving data gathering, analysis, and action before accidents happen."

Here is how the Safer Skies challenges look for the Part 121 air carriers now in 2003, with less than 5 years to go. The Part 121 air carriers fly about 95 percent of all airline flying being tracked for the 80 percent reduction numbers. For simplicity and clarity, the data deal with the accidents behind the official FAA rates per 100,000 departures.

According to the FAA, fiscal years 1994–96 had 16 fatal accidents, including some involving Part 135 airplanes with 10 or more seats that now fly under Part 121. At the time, Boeing was warning that, at the then-current accident rate and forecast increases in departures, large jet hull losses would soon average one per week worldwide. Worldwide, pilots were being blamed as a primary cause in 67 to 80 percent of large jet accidents, according to Boeing data, and this was thought to approximate the U.S. experience. The challenge seemed formidable and necessary. To reach the goal of 80 percent reduction, the U.S. average of 5.3 fatal accidents per year would need to fall to fewer than 2 per year by 2007, considering the forecast increase in departures.

From fiscal year (FY) 1998 (Oct. 1, 1997, through Sept. 30, 1998) through half of FY 2003, 10 fatal Part 121 accidents occurred as follows:

• FY 1998

1. 12/28/97—B-747, Pacific Ocean. Turbulence passenger fatality—1 fatality

2. 11/03/98—Saab 340, Memphis, TN. Propeller strike—1 fatality

3. 06/01/99—MD-82, Little Rock, AR. Runway overrun—11 fatalities

4. 07/28/99—DHC-8, Little Rock, AR. Propeller strike—1 fatality

5. 01/31/00—MD-83, Pt. Mugu, CA. Loss of control, in flight—88 fatalities

6. 02/06/00—DC-8, Sacramento, CA. Loss of control, approach—3 fatalities

7. 11/20/00—A300, Miami, FL. Evacuation—1 fatality

8. 08/05/01—DHC-8, Washington, DC. Propeller strike—1 fatality

9. 11/02/01—A300, Belle Harbor, NY. Loss of control, in flight—283 fatalities

10. 01/08/03—Be-1900, Charlotte, NC. Loss of control, in flight—21 fatalities

The total number of accidents in each of these years was fewer than the number envisioned as the goal in 1997. The baseline years, FY 1994-96, had 5.3 accountable fatal accidents per year. Without any reduction in the rate of accidents, 53 total accidents would have occurred between 1997 and October 2007 (see Table 1 for the fatal accident experience from October 1997 to January 2003 and the original target as roughly forecast to be through 2007).

Fewer than two Part 121 fatal accidents have occurred each year since 1997. The 10-year goal of two accidents per year was met in 5 years. The U.S. airline industry could have suffered more than 20 accidents, or 4 per year on average, and still have met the original target for fatal Part 121 accidents. The 3-year period 1994-96 had 800 fatalities; 1997-2002 had 390. So far, 2003 has had 21 fatalities.

Overall accident reduction

From calendar year (CY) 1998 through 2002, the United States had 240 total Part 121 accidents involving serious injury, substantial damage, fatalities, or destroyed aircraft. This works out to an average of 48 per year, or 4 per month. In the first 3 years, these accidents occurred once every 7 days. In CY 2001, the number of these accidents dropped to 38 (not counting terrorist acts) for the whole year, or 1 every 10 days. CY 2002 had 41 accidents, or 1 every 9 days. Even with the drastic reduction in departures since 9/11, this, too, shows progress, but airports and airlines could do more.

Turbulence (60) and ground collisions (49) made up more than half the total number of accidents in the first 5 years, when a Part 121 ground-collision and a turbulence-injury accident occurred about once a month. The approach and landing phases of flight had 60 accidents, or about 1 a month; and 17 hard landings and/or tailstrikes and one wing failure occurred. In 12 events, the airplane went off the side of the runway, and 11 events involved undershoots or overruns. Ten landing gear failures occurred, and six collisions with objects such as snow banks, deer, and trees. One airplane damaged propellers after failing to lower the gear for landing and successfully going around. Only one approach and landing accident, in Little Rock, resulted in fatalities. In the last 5 years, no fatal CFIT accidents, fatal runway incursions, or midair collisions occurred. Of every 27 accident investigations, 1 involved a fatal accident. Add three nonfatal destroyed airplane accidents requiring major investigations, at Subic Bay, Burbank, and Tallahassee, and one fatal or major accident occurred out of every 20 total accidents.

The greatest fatality producer, with 374 fatalities, was loss-of-control (LOC) in flight in three accidents. All involved airplane systems (flight controls) and flight crews’ inability to recover to save the airplane. LOC persists as a problem with 21 more fatalities at Charlotte on Jan. 8, 2003.

Did the accident rate improve because of CRM and AQP? LOSA? One Level of Safety? the mandate for airlines to have directors of safety? improved availability of aviation safety data through programs like GAIN, FOQA, and ASAP? conferences on corporate culture and fatigue? Or was it because of the collaborative efforts of government, union, and airline industry representatives to reduce the number of accidents attributed to controlled flight into terrain, approach and landing, midair collisions, and runway incursions or collisions? Over the last 5 years, many changes have taken place in the way we do business as a result of these and other projects, such as the Joint Safety Analysis Teams (JSATs) and Joint Safety Implementation Teams (JSITs) chartered by the government/industry Commercial Aviation Safety Team (CAST).

FAA Administrator Marion Blakey believes the "achievement is a result of safety pros who are acting as a team and know what they are doing."

Much of CAST’s success has come from the Safer Skies call for research projects to support the 80 percent reduction goal. The focused agenda required an organized strategic plan and process under CAST, which has taken on the extraordinary task of reviewing the applicable Part 121 (and some Part 135 and Part 129) airline accidents from 1987 to the most recent events. CAST merged—into one huge data-driven problem-identification and safety-enhancement/ implementation plan—existing Flight Safety Foundation approach and landing (ALAR) and controlled flight into terrain (CFIT) data, along with NTSB accident recommendations and FAA ongoing projects to reduce runway incursions, uncontained engine failures, in-flight loss of control, and turbulence/weather accidents.

CAST categorized the U.S. fatal accident history by type of accident and by the causes of accidents. The causes that resulted in the most fatalities were subject to special CAST study committees, or Joint Safety Analysis Teams, which analyzed root causes and selected actions that could correct each cause. The JSATs were formed using airline industry and government expertise, with many active ALPA pilots and staff members participating.

Accidents with multiple causes were subjected to multiple committee reviews. For example, an LOC accident that was caused by icing was reviewed by two committees, one considering LOC, and the other, weather/ice. Thus, all causes or issues were considered for problem identification and fixes.

Corrective actions that the JSATs cited and prioritized were then turned over to Joint Safety Implementation Teams (JSITs), which tested each proposed intervention for technical and economic feasibility and developed projects and plans for implementation. The most effective, most-bang-for-the-buck items made the final list of 84 projects. Thus, the JSATs and JSITs are data-driven groups that have provided the best priorities based on objective quantitative analysis of available safety information called for in the Safer Skies goals.

For 5 years, CAST has been defining and optimizing more than 700 identified problems and solutions in 90 or more accidents since 1987. By April 2003, CAST had completed 23 (see Table 2, above) of more than 100 possible safety enhancements, and 25 more are currently being implemented (see Table 3, opposite).

These 48 safety enhancements, when completed, predict a 73 percent fatal risk reduction for Part 121. Significant progress is being made.

ALPA continues to provide leadership in this process. As Capt. John Cox (US Airways), the Association’s Executive Air Safety Chairman, says, "ALPA’s pilots and professional staff are pledged to lead the way to meeting the Safer Skies goals."

A close look at the completed CAST enhancements reveals that 17 of the 23 call on pilot, air traffic controller, inspector, maintenance, or safety management participation and training.

Minimum safe-altitude warning systems (MSAW), terrain-avoidance warning systems (TAWS), visual glideslope indicators (VGSIs), new DMEs, and flightdeck upgrades in future-design airliners are the only completed enhancements calling for technical improvements designed to help the flightdeck crew and improve situational awareness.

Most enhancements are improved, more-foolproof upgrades of old technologies. The safety enhancement system remains highly dependent on human performance especially in the cockpit, at the air traffic controller station, and with those charged with enforcing and monitoring standard operating procedures and best practices and procedures.

Four of the nine fatal accidents and 374 of the 390 fatalities from October 1997 to October 2002 resulted from accidents involving loss of control (LOC) in flight. And the fatal Beech 1900 accident with 21 lives lost in Charlotte in January 2003 can be added to that total. Before the Safer Skies program, in FY 1994–96, 10 LOC accidents with 611 fatalities occurred; thus, the Safer Skies tracking data for Part 121 aircraft now contain 1,006 LOC fatalities.

To date, precious few enhancements have been implemented to prevent LOC accidents, except for simulator training and some training for flight inside the envelope. LOC accidents are highly problematical, involving both the aircraft systems and design and the pilots, who have very little chance to suddenly solve the problem and save the aircraft.

Of the 11 LOC projects that are awaiting the go-ahead and funding, 6 involve training, policy development, and human-factors/automation-interface studies.

Enhancements involving standard operation procedures and training include checklist usage; pilot-not-flying (PNF) duties; transfer of control; rushed or unstable approaches, rejected landings, and missed approaches; icing reports; and flight crew coordination. Incident analysis needs to be improved to supply critical safety information and reveal recurring intermittent failures. Manuals need to be updated better and faster, and pilots need better guaranteed proficiency on aircraft control with automation. The master training project involves inside-the-envelope advanced-maneuver ground and flight training dealing with the many aerodynamic and automation causal factors involved in LOC accidents.

The enhancements that score the highest for risk reduction for LOC accidents for existing airplanes involve prioritized risk-assessment methodology itself and the advanced-maneuver ground and flight training programs—training and procedures to overcome aircraft design flaws. By far, the highest-rated enhancements, however, were requiring envelope protection for new-design airplanes and installation of vertical situation displays. Unfortunately, these have been deemed too costly to install in the existing fleet of older aircraft, the current victims of LOC fatal accidents.

Five LOC enhancements involve technical and engineering projects—to minimize thrust asymmetry, improve manual flight control force disconnects, annunciate false aircraft autopilot responses and inappropriate disconnects, and provide low-speed protections. These are addressed in harmony with guidance that other existing working groups provide.

One project for new airplane designs would improve display and alerting systems to immediately show the flight crew what is happening that may lead to loss of control. Another project supports the development of new anti-icing certification criteria, mainly for non-hot-wing designs, for new airplanes. Another technical improvement will require vertical situation displays in new design aircraft (and possibly existing) cockpits. Finally, new airplane designs should require flight envelope protection for angle of attack/low speed, bank angle, and thrust asymmetry.

Flight crews can expect results from these projects in from 1 to 10 years after the project is "assigned resources." The most effective technical fixes will affect pilots only when they bid to fly a newly designed aircraft from the factory. More training and procedures will have to suffice for junior pilots. That will be the bulk of the help received from "outside the cockpit" for the near term.

Of the 14 approved CAST enhancements not involving LOC accidents, 6 result from pre-CAST recommendations on CFIT and ALAR projects that were absorbed and adopted by the 5-year process. The eight runway incursion enhancements were described in "CASTing About for a Solution," May.

The six CFIT/ALAR projects mostly involve FAA administrative and procurement resources for perfecting technical hardware and software reliability and accuracy. The highest-rated risk reducer involved guidance to ensure continuing airworthiness (reliability) of critical system components—good news for pilots who know the effects that anomalies and emergencies have during critical phases of flight. When enhancements are in the cockpit sometime soon, they will enhance the ability of flightcrews to know where they are both vertically and laterally at all times with a one-step, faster, dependable perception-decision process. High-risk approaches and low-altitude operations will become much easier to fly. Some remaining CFIT/ALAR enhancements in R&D are costly, such as synthetic vision, and cockpit moving-map displays that would allow much-improved situational awareness and early warnings of traffic and navigational problems on both the ground and air.

Though things are looking good and on glidepath at the 5-year point, the last half of the Safer Skies countdown faces many obstacles. Fatal accidents are fewer than forecast, fatalities are reduced, overall accident trends are good, flight crew causal factors are less than perceived, and flight crews are ready to do more than their share of carrying the load of fixing the problems. The CAST members have performed a monumental task of uncovering the most vexing problems and identifying the best risk-reduction interventions for implementation to overcome recurring accidents. However, the pressing needs for aviation security and the downturn in airline economics may lead scarce funds to being placed elsewhere. Some safety experts are concerned that without blood on the runway motivation will be lost.

To counter the temptation to lose interest, the CAST process has given us disciplined research for reaching the Safer Skies goals. It has also provided a tracking and scoring system to measure results with yet another committee, called Joint Implementation Measurement-Data Analysis Team (JIMDAT). This team will be monitoring all the deadlines and keeping score on them, plus tracking the events that make them come true. If something new comes along, like another accident or new viable enhancement, JIMDAT will review it and make recommendations for research or prioritization. JIMDAT will work future accidents into the CAST process and keep up the pressure to stop accidents.

JIMDAT is also working on methodology for analyzing reports of incidents such as runway incursions, NTSB reports of investigated incidents (near accidents), and NASA ASRS-protected reports of unsafe conditions, and is working with airlines to develop a satisfactory process for using compiled ASAP and FOQA data. This is designed to convert CAST into a proactive group that will be able to work on accidents before they occur.

The JIMDAT is also keeping an eye on system efficiency measures that have safety connotations and enhancements attached. This includes the huge National Airspace System (NAS) improvements designed to avoid worse gridlock in the future that may very well have huge safety enhancements as a benefit.

This should keep the CAST process alive, if their activities receive the light of day in the "assignment of resources" scheme in the future. An ongoing, well-publicized report to airline CEOs, directors of safety, associations, and interested parties would help, whether the news is good or bad. Progress reports should become page one headlines to replace the tragedies now very well covered in the press. Progress or regression should not be kept quiet. Sadly, very few directors of safety or members of the press attended the North American Aviation Safety Conference in February.

At that Conference, the Air Transport Association’s Senior Vice-President for Aviation Safety and Operations illustrated that the money already spent on committed projects has reduced risk by 64 to 72 percent and has a chance to meet the 80 percent accident reduction goal by 2007. As much as $500 million per year is already being saved. Historically, airline accidents have cost about $90 per departure, every departure, all day, every day. Implementing all the enhancements in the CAST plan could result in a further savings of $540 million by 2007. However, in the next 5 years, with the current economy, maintaining a proactive momentum and funding for the CAST process and uncompleted enhancements may be extremely difficult. The bottom line is that safety more than pays for itself. A matured CAST process based on incident data (FOQA/ASAP) will continue to drive the consensus-based, data-driven decision process for the most highly leveraged safety improvements.

The CAST process may go international and become an ICAO methodology for reducing risk and accidents in other regions of the world, following the U.S. example. This should encourage continuing interest in CAST’s accomplishments. The new safety culture of open and no-jeopardy communications for safety is still in its infancy and is having growing pains. It bears watching in this safety evolutionary era. Will all the bad habits go away?

As FAA Administrator Blakey said, "We need to change one of the biggest historical characteristics of aviation safety improvements—our reactive nature. We must get in front of accidents, anticipate them, and use hard data to detect problems and disturbing trends." The most difficult part of the process is ahead—to get the job done before running out of gas (and money) and falling back into the old habit of waiting for the next surprise (even harder data) for motivation.