the u.s. air traffic control system wrestles with the

FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993 1

The U.S. Air Traffic Control SystemWrestles with the Influence of TCAS

Although TCAS II has been credited with averting inflight aircraft collisions, its implementation continues

to cause disagreements between pilots and controllers.

byV.J. Mellone and S.M. Frank

A significant increase in incident reporting relatedto Traffic Alert and Collision-avoidance System(TCAS II) implementation prompted a closer ex-amination of how TCAS II is affecting pilots andair traffic controllers.

In July 1992, the U.S. Federal Aviation Administra-tion (FAA) and the U.S. National TransportationSafety Board (NTSB) asked the U.S. National Aero-nautics and Space Administration’s (NASA) Avia-tion Safety Reporting System (ASRS) to complete adata base analysis of TCAS II incident reports.

The ASRS research team identified evidence ofincreasing air traffic controller consternation withand resistance to the implementation of TCAS IItechnology into the U.S. airspace system. Therewere also strong indications from the data set thatthe aviation community, government agencies andindustry may have unwittingly underestimated theinfluence of TCAS II avoidance maneuvers on airtraffic controllers and flight crews.

The study’s objectives were to analyze the ef-fects of TCAS II avoidance applications on

pilot-controller interactions and to suggest strat-egies that could promote a more harmonioushuman-technology interaction.

The study was limited to a randomly selected dataset of 174 ASRS reports from the January 1992-March 1993 period. The data set specifically in-cluded pilot and controller reports that involvedtraffic advisory (TA) and resolution advisory (RA)incidents.

A TA identifies nearby traffic meeting “certainminimum separation criteria,” according to an FAAadvisory circular (AC) issued in August 1993. AnRA provides flight crews with aural and displayinformation advising whether or not a particularmaneuver should be performed to maintain safeseparation from a threat aircraft.

The research addressed four issue areas: Were thereany significant changes in TCAS II reporting dur-ing the past three years? Do TCAS II avoidanceactions influence the air traffic control (ATC) sys-tem? Is there evidence of contention between pi-lots and controllers because of TCAS II applications?

FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 19932

Are there effective strategies that could enhancepilot-controller cooperation?

The ASRS reports were randomly drawn from apool of 780 TCAS II incidents from the same pe-riod involving TAs and RAs. The 174 reports ac-counted for approximately 11 percent of the 1,522TCAS II full-form reports in the ASRS data base.For reasons of maximum currency, the data setincluded 55 unprocessed reports from the January-March 1993 period. There were a total of 109 pilotreports, 51 controller reports and 14 reports whereboth a pilot and controller submitted reports on thesame TCAS II incident.

A coding instrument was developed to extract per-tinent information from the records. The codinginstrument addressed several areas. The key pointsare described in Table 1.

Development of the coding form required a numberof iterations of trial codings and comparisons tovalidate the instrument. A team of experienced ASRSpilot and air traffic controller analysts was assignedto analyze and code the 174 reports, and the com-pleted coding forms were entered into a data baseprogram for statistical analysis. The data was re-viewed, tabulated and summarized.

Since 1956, the aviation community has attempted,in the face of widely spaced, but successive com-mercial aircraft midair collisions, to conceive andimplement an airborne coll ision avoidance

system. In 1987, the U.S. Congress mandated theinstallation of TCAS II on all air carriers with 30or more passenger seats by the end of 1993. Todate, more than 4,000 air carrier and business air-craft are TCAS II equipped.

TCAS II is an aircraft-based airborne collisionavoidance system that provides information, inde-pendent of the ground-based ATC system, of theproximity of nearby aircraft. It alerts pilots,visually and aurally, to potentially threatening situ-ations (intruders) by monitoring the position, clo-sure rate, and altitude of nearby transponder-equippedaircraft. TCAS II offers the pilot both TAs andRAs. RAs are limited to vertical avoidance ma-neuvers if an intruder comes within approximately25 seconds to 30 seconds of closure.

Initial testing of TCAS II in 1984-85 by two majorair carriers, in a limited operational environment,determined that “TCAS II is safe and operationallyeffective, [and] operates harmoniously in the air trafficcontrol system.” The testing data indicated that “ap-proximately 50 percent of the alarms will be ‘pre-ventive’ and the pilot will not deviate from his flightpath. The average deviation will be approximately300 feet (91 meters) based on a 1,500 foot-per-minute(457 meters-per-minute) climb or descent.”1

The reporting of TCAS II incidents increased sig-nificantly from 1990-1992 (Figure 1). The increase

Table 1Key Traffic Alert and

Collision-avoidance System (TCAS II)Coding Points

P1 Pilot Use of TCAS II AdvisoryP2 Pilot-initiated Avoidance ActionsP3 Air Traffic Control (ATC)-initiated

Avoidance ActionsP4 Traffic Alert (TA)/Resolution Advisory (RA)

Causal FactorsP5 Pilot Communications to ATC about TA/RAP6 ATC Reactions to TA/RAP7 Pilot Comments About TCAS II IncidentP8 TCAS II Application: Prevented-caused

P9 Evidence of Pilot/Controller Contention

Source: Batelle/U.S. National Aeronautics and SpaceAdministration, Aviation Safety Reporting System


Figure 1

Traffic Alert and Collision-avoidanceSystem (TCAS II) Incident

Reporter Breakdown

1990 1991 1992

0

100

200

300

400

500

600

700

800

900

Pilot

Controller

Nu

mb

er o

f In

cid

ent

Rep

ort

s

Year


included a total of 99 TCAS II incident re-ports from controllers. This is the largestnumber of controller reports to ASRS on asingle topic since the August 1981 air trafficcontrollers’ strike. The total breakdown ofcontroller reporters on TCAS II incidentsincluded 37 reports from en route control-lers, 58 reports from terminal radar control-lers and four reports from tower controllers.The overall increase in TCAS II incidentreporting can also be attributed to the num-ber of air carrier and business aircraft thatare installing TCAS II equipment in accor-dance with the requirements of U.S. FederalAviation Regulation (FAR) 121.356 and, pro-spectively, FAR 91.221 and 135.180. Thereporting trend has continued upwards asflight crews and controllers experience vari-ous TCAS II avoidance advisory situations.

Figure 2 depicts the actions taken by pilots in con-junction with a TCAS II RA based on the data set.The majority of pilot actions were precipitated byRAs (92 percent).

Altitude changes ranged from 100 feet (30 meters) to1,600 feet (488 meters), as indicated in Figure 3,with an average altitude change of 628 feet (191meters). In contrast, the designers of TCAS II ex-pected altitude deviations to average 200 feet to 300feet (61 meters to 91 meters). Initial TCAS II simu-lation testing did not reveal that some aircraft woulddeviate by 1,000 feet (305 meters) or more and in-trude upon other occupied altitudes.

As shown in Figure 4 (page 4), the pilot informedATC after reacting to the RA in the majority ofincidents (93). In 15 incidents, the pilot did notnotify ATC about the RA maneuver. In 26 incidents,the pilot either forewarned ATC (9) or informed ATCwhile complying with the RA (17). A number ofcontroller reports indicate that this response patternis a major source of their concern because of theimpact of the unanticipated avoidance deviations onthe controllers’ air traffic situation.

It is here that the dual impacts of TCAS II applica-tions can be seen. Analysis of the reporter narra-tives verify that TCAS II has been instrumental inpreventing many near midair collisions (NMACs)and other conflicts. (Figure 5, page 4).

Reporter narratives also confirm that ATC work-load has increased. Pilot actions to avoid oneconflict sometimes precipitated a loss of separa-tion or airborne conflict with another non-intruderaircraft in the same vicinity (Figure 6, page 5).

There were indications in the reporter narrativesthat the process of assimilating TCAS II into theATC system was affecting pilot-controller interac-tions. Intensive coding analysis of the 174 reports

Pilot Use of Traffic Alert andCollision-avoidance System (TCAS II) Advisory

Source: Batelle/U.S. National Aeronautics and SpaceAdministration, Aviation Safety ReportingSystem

Figure 3

Change Altitude

Establish Visual Contact

Evasive Turn

None

0 20 40 60 80 100 120 140

137

29

11

11 188 citations froma total of 174 reports*

Number of CitationsThe number of citations may be greater than the number of reports because categories are not mutually exclusive.

*Note:

Source: Batelle/U.S. National Aeronautics and SpaceAdministration, Aviation Safety Reporting System,

Figure 2

Pilot-initiated Altitude Change

1

100

150

200

250

275

300

400

500

600

700

800

1,00

01,

100

1,20

01,

300

1,40

01,

500

1,60

0

0

2

4

6

8

10

12

14

16

18

20

1 1

4

21

13

8

19

13

67

2 2 2

13

Average Altitude Change 628 Feet

Feet

Nu

mb

er o

f In

cid

ents

11 1


verified that a degree of contentiousness existed inapproximately 24 percent of the reports. Figure 7(page 5) shows that in a majority of the incidents(95), disagreement was not evident in the reporter’snarrative. However, in the 43 episodes related toTCAS II applications, and 27 episodes unrelatedto TCAS II applications, there was inferred orstated evidence of contention based on the narra-tive comments of the reporter.

TCAS II has definitely prevented or re-duced the threat of airborne conflicts asreflected by the majority of favorableASRS reports and evidenced by the fol-lowing statements from ASRS reports:

• “It is the captain’s opinion that TCASII saved about 100 lives today. I ama firm believer in TCAS II forever”(ASRS Pilot Report No. 212377);and,

• “Gap in updating of radar-displayedMode C misled me to think air carrier[name deleted] was still level at flightlevel 310 (31,000 feet [9,455 meters]).Situation was resolved by TCAS”(ASRS Controller Report No. 225920).

However, there have been side effects re-lated to the controllers’ sense of a loss ofcontrol of the aircraft nominally under

their jurisdiction. Apprehensive controllers increas-ingly wonder “Who is in control?” when TCAS II-equipped aircraft execute avoidance maneuvers orapply non-standard uses of TCAS II without priornotification or coordination with ATC.

In a recent U.S. congressional hearing on TCASII, Air Line Pilots Association (ALPA) President J.Randolph Babbitt testified that “line pilots have

strongly endorsed TCAS II and would em-phatically resist any efforts to reduce its op-erational effectiveness.”2 But, at the samehearing, National Air Traffic Control Asso-ciation (NATCA) President Barry Krasner omi-nously warned that TCAS II “is an airbornesystem that works improperly and erodes analready precarious margin of safety in theskies.”3 The pilot community, particularly ALPA,sees TCAS II as a “last ditch, they-may-have-hit-if-something-is-not-done piece of equip-ment that gives [the pilot] a way out if therest of our ATC system has somehow unac-countably failed.”4

Looking at the steps taken in implementingTCAS II, it appears that TCAS II designersmay not have fully anticipated all the waysthat pilots would act, and controllers wouldreact, once TCAS II was broadly implemented.More than 2.5 million revenue miles later,

Pilot Communications with Air Traffic Control(ATC) on Traffic Alert and Collision-avoidance

System (TCAS II) Resolution Advisory (RA)

Traffic Alert and Collision-avoidanceSystem (TCAS II) Advisory Prevented


Figure 5

Figure 4


Complied with RA — no ATCcounter-instructions

Informed ATC after RA response

Unknown

Warned ATC while complying

Pilot did not notify ATC

Complied with RA and ignoredATC instructions

Forewarned ATC before complying

Other

Unable to notify ATC becauseof frequency congestion

Ignored RA per ATC instructions

0 20 40 60 80 100 120

101

93

30

17

15

12

9

8

4

2

Number of Citations

291 citations froma total of 174 reports*

The number of citations may be greater than the numberof reports because categories are not mutually exclusive.

*Note:

Near midair collisions

Airborne Conflict/LessSevere

Loss of StandardSeparation

0 10 20 30 40 50 60

52

12

5

Number of Citations

69 citations froma total of 66 reports*

The number of citations may be greater than the number of reports because categories are not mutually exclusive.

*Note:


intruders, RA commands being contradicted bycounter-instructions from controllers, etc.

A statement in the NATCA News noted that “theGENOT (FAA General Notice RWA 3/141) is riddledwith ambiguities and contradictions that decreasethe safety margin. TCAS is not reliable enough forthe FAA to order controllers to take such a hands-off position.”8

The FAA recently announced its intention of man-dating TCAS II software logic changes (change 6.04)

with TCAS II installed in more than 70percent of the air carrier fleet and withseveral hundred RAs reported, the follow-ing issues have become evident:

Training and preparation have not beenadequate. Both pilot and controller initialTCAS II training has been inadequate forthe unexpected avoidance actions being en-countered. A recent NTSB recommendationto the FAA stated that “both controllers andpilots need more training in the traffic alertand collision-avoidance system.”5 There ap-pears to have been a lack of operationalhuman factors impact analysis preparatoryto implementing TCAS II system-wide.

One major air carrier’s TCAS II flightcrew indoctrination video indicated thatresolution command altitude changeswould typically require a 200-foot to 300-foot(61-meter to 91-meter) maneuver. The initialFAA controller TCAS II training video tape con-veyed that the implementation of TCAS II wouldbe “transparent to the controllers” and the RAswould be limited to “200 feet to 300 feet” inaltitude changes. However, both FAA and ASRSRA altitude-change data now confirm that theaverage altitude change is more than double whatwas originally anticipated.

An FAA report said “air traffic controllers haveexpressed major concern about the magni-tude of the altitude displacements in re-sponse to some corrective RAs. The dataprovide evidence that a problem exists inthe way the pilots use the system or in theway pilot training is implemented.”6

An ALPA statement noted: “Senior FAAofficials admitted [during the InternationalTCAS Conference, Jan. 7-9, 1992, Wash-ington, D.C.] that TCAS training for con-trollers and controller involvement ... havebeen too little, too late.”7

Ad hoc fixes have been necessary. Subse-quent fixes to unanticipated TCAS II logicdeficiencies convey an after-the-fact ap-proach to the overall problems with thetechnology, i.e., nuisance RAs, phantom

Traffic Alert and Collision-avoidance System(TCAS II) Advisory Caused

Evidence of Pilot/Controller Contention


Figure 7

Increased ATC*Workload

Loss of StandardSeparation

Airborne Conflict/LessSevere

Controlled Flight Towards Terrain

Near Midair Collisions

0 5 10 15 20 25 30 35 40

37

22

17

2

Number of Citations

82 citations froma total of 52 reports**


**Note:

4

*ATC: Air Traffic Control

Source: Batelle/U.S. National Aeronautics and Space Administration,Aviation Safety Reporting System

Figure 6

None Evident

Related to TCAS**Application

Based on ControllerComments

Based on Flight Crew Comments

Unrelated to TCAS**Application

Inferred by ASRS*Coding Analyst

0 10 20 30 40 50 60 70 80 90 100Number of Citations

95

43

36

31

27

19

*ASRS: Aviation Safety Reporting System**TCAS: Traffic Alert and Collision-avoidance System

251 citations froma total of 174 reports***


***Note:


by Dec. 31, 1993, through the release of a notice ofproposed rule making (NPRM) and the subsequentissuance of an airworthiness directive (AD). Thisfine tuning of TCAS II software logic shouldeliminate a majority of the nuisance RAs. However,the changes will, in some instances, reduce RA warningtime, and they have provoked strong reservationsfrom ALPA and others, as illustrated below:

• “There is some industry concern that pro-posed modif ica t ions . . . wi l l fur thercomplicate TCAS implementation andsafety risks;”9 and,

• Pilots are in a quandary abouthow to respond to TCAS IIalerts. Pilots are being placedin a dilemma about makingsplit-second decisions whetherto ignore controller adviso-ries on separated traffic or tofollow RA commands.

The following statements were takenfrom ASRS pilot reports:

• “Center later admonished usfor descending from altitudeand stated traffic was level … .” (ASRSPilot Report No. 205812);

• “We did not inform the controller of theevasive action in a timely manner” (ASRSPilot Report No. 212715); and,

• “ATC took offense to TCAS and its use inthe air traffic system. I have been findingthat ATC is not too fond of TCAS becauseit takes away their authority” (ASRS PilotReport No. 206966).

Controllers feel conflict about the appropriate re-sponse to RA-commanded deviations. Controllershave recently been instructed by the FAA to notcountermand RAs which could contradict avoidancecommands. In some instances, the controller becomesa passive observer to potential loss of separationsituations, contrary to all prior training and basicinstincts as an air traffic controller:

• “TCAS should only be used as an advisorytool for pilots and should not override air

traffic control instructions” (ASRS Con-troller Report No. 232487); and,

• “Pilot of air taxi [name deleted] elected torespond to RA from aircraft [name deleted]and descended [toward] aircraft [name de-leted]. This is insane at best and a blatantmisuse of technology” (ASRS ControllerReport No. 209177).

There are indications of nonstandard use of TCASII by pilots. The ASRS has received a number ofTCAS II incident reports that clearly identify non-standard applications of TCAS II:

• “I cannot have pilots using TCASII for visual separation. There isno TCAS II separation” (ASRSController Report No. 202301);

• “Approach appeared to not havethe ‘big picture.’ TCAS was usedto ensure safe traffic separation”(ASRS Pilot Report No. 183286);

• “Cannot have pilot using TCASII to maintain spacing” (ASRSController Report No. 202301);and,

• “ALPA’s ATC committee cautions pilots... not to use TCAS to provide their own airtraffic control and self-separation.”10

The conclusions of the TCAS II incident analysisverified the following points:

• TCAS II has definitely enhanced flight safetyand is widely supported by the commercialpilot community;

• Air traffic controllers have been confusedand occasionally alarmed by the variety ofRA applications seen in terminal and enroute airspace;

• The requirements associated with the imple-mentation and subsequent nurturing of TCASII may not have fully anticipated the “growingpains” that have influenced both pilots andcontrollers;

There are

indications of

nonstandarduse of TCAS II

by pilots.


• Initial TCAS II training did not adequatelyprepare pilots and controllers for the sur-prises generated during RA situations;

• TCAS II applications have had an influ-ence on the respective roles of the pilotand the controller; and,

• Behavioral ramifications need to be fully evalu-ated in concert with ergonomic issues.

TCAS II Enhancements Suggested

Information provided to the ASRS by pilots andcontrollers suggests that the following enhance-ments should be given serious consideration:

• Mandate a requirement for high-performanceturbojet aircraft to reduce their vertical speedbelow 1,500 feet per minute (457 metersper minute) when 1,000 feet (305 meters)from their assigned level-off altitude;

• Require pilots to select TAmode only when establishedon the final approach forparallel approaches, in vi-sual meteorological condi-tions (VMC) at airportswhere the parallel runwaysare less than 4,300 feet(1,311 meters) apart;

• Require pilots to notifyATC whenever a TA be-comes a “yellow circle in-truder” to warn ATC of aprobable RA situation. A “yellow circleintruder” refers to a display on TCASinstrumentation that illuminates a solidamber circle when a TA is issued. Asflight crew workload and frequency con-gestion permits, the pilot should also citethe clock position, direction of flight,and altitude of the intruder;

• Require ATC to issue the pilot all Mode Ctraffic in his vicinity upon being informedof yellow circle intruder;

• Direct pilots not to exceed a 500-foot (152-meter) altitude change during an RA responseunless the intruder is displayed to be climb-ing or descending within 500 feet of the TCASII aircraft’s altitude;

• Expand and vigorously promote TCAS IIsimulation training for flight crews and con-trollers. The FAA training academy shouldconsider developing and distributing ATCTCAS II situational training scenarios forimmediate use at all terminal radar and enroute center dynamic simulation training labs;

• The FAA and major aviation organizationsincluding ALPA, Aircraft Owners and PilotsAssociation (AOPA), American Pilots’ As-sociation (APA), National Business AircraftAssociation (NBAA) and NATCA shouldsponsor joint pilot/controller area and regionalTCAS II workshops to promote cooperativedialogue between the constituencies; and,

• ATC facility managers, throughthe support of the air carriersand the Air Transport Associa-tion of America (ATA), shouldactively promote TCAS II fa-miliarization flights for control-lers. The opportunity to ridejumpseat, observe TCAS II situ-ations and engage in construc-tive dialogue with flight crewsshould promote a better under-standing of their respective pro-fessional concerns.

According to AC 120-55A, issuedin August 1993, TCAS II is “in-

tended to serve as a backup to visual collisionavoidance … and air traffic separation service.”

The AC adds: “For TCAS to work as designed,immediate and correct crew response to TCASadvisories is essential.” Flight crews, the AC said,are expected to respond to TCAS according to thefollowing guidelines:

• Respond to TAs by attempting to establishvisual contact with the intruder aircraft whilemaintaining ATC assigned clearance; and,

“For TCAS to

work as designed,

immediate andcorrect crew

response to TCAS

advisories isessential.”


• When an RA occurs, the pilot flying shouldrespond immediately “… unless the flightcrew has definitive visual” contact withthe aircraft causing the RA. The AC saidthe initial vertical speed response is ex-pected within five seconds and that alti-tude excursions “typically should be no morethan 300 [feet] to 500 feet [91 meters to152 meters] to satisfy the conflict.”

But the AC cautioned: “ATC may not know whenTCAS issues RAs. It is possible for ATC to un-knowingly issue instructions that are contrary [op-posite] to the TCAS RA indications. Safe verticalseparation may be lost during TCAS coordinationwhen one aircraft maneuvers opposite the verticaldirection indicated by TCAS and the other aircraftmaneuvers as indicated by ATC. As a result, bothaircraft may experience excessive altitude excur-sions in vertical-chase scenarios because of theaircraft maneuvering in the same vertical direc-tion. Accordingly, during an RA, do not maneuvercontrary to the RA based on ATC instruction.” ♦

[Editor’s Note: This article was adapted from Be-havioral Impact of TCAS II on the National AirTraffic Control System, a Battelle program reportfor NASA’s Aviation Safety Reporting System, April1993.]

References

1. Haines. The Journal of Air Traffic Control(October-December 1990).

2. Air Safety Week (August 10, 1992).

3. NATCA News (July/August 1992).

4. Hanson. Delta MEC Safety Newsletter (Sec-ond Quarter 1992).

5. Aviation Daily (March 26, 1992).

6. AIRINC Research Corp. Results of the TCASTransition Program (TTP). An interim reportprepared at the request of the U.S. FederalAviation Administration. December 1991.

7. Steenblik. “How Fares TCAS? Part I.” Air LinePilot (April 1992).

8. NATCA News (March 1993).

9. “Interim Reports Give TCAS Mixed Reviews.”Airport Operations (September/October 1992).

10. Steenblik. Air Line Pilot (April 1992).


[Editor’s Note: In the January 1991 Flight SafetyDigest, Gerard Bruggink first examined the safetyaspects of airline deregulation. He concluded thenthat a series of accidents following deregulationcould be a harbinger of poor performance to comeunless the airline industry quickly came to termswith the changes and upheaval that deregulationcaused. Following is Bruggink’s updated assess-ment based on recent accident statistics.]

During the three-year period 1990-1992, U.S. air-lines operating under Part 121 of the U.S. FederalAviation Regulations (FAR) set several records.

In number of departures and hours flown perfatal accident, this was the industry’s best three-year performance. And for the first time, the

number of aircraft occupant deaths was less thanthat in scheduled Part 135 (commuter) accidents.Financial losses, however, reached an all-timehigh despite an unprecedented level of passen-ger traffic.

The airlines’ financial plight is mentioned only be-cause it does not seem to fit the prevailing notionthat poor economics tend to provoke shortcuts thatlead to accidents. The staggering financial losses didnot have the corner-cutting effect typically associ-ated with economic survival because these losseswere an inherent part of competitive strategies.

The prediction that the Airline Deregulation Act of1978 would adversely affect air carrier safety isalso being questioned. Although deregulation’s

U.S. Airline Industry Safety Record ImprovesDespite Deregulation and Huge Financial Losses

The U.S. Airline Deregulation Act of 1978 raised concerns aboutits impact on safety. Although there was an increase in fatal accidents

after its enactment, trends have improved each year since.

byGerard M. Bruggink

Aviation Safety Consultant


destabilizing influence on the work force may haveplayed a role in several accidents, the overall trendsince 1978 has been one of improvement, with theexception of 1987-1989. That makes the 1990-1992 performance all the more remarkable andjustifies an emphasis on the more intangible ele-ments of success in accident avoidance and on theindividuals who deserve credit for this.

This article deals only with the safety performanceof U.S. carriers because a lack of consistency inthe worldwide reporting of exposureand accident data precludes meaning-ful comparison.

There is no agreed-upon method tomeasure safety. The common prac-tice of using past accident experi-ence as a yardstick is flawed becauseit equates the absence of accidentswith safety — without accountingfor the close calls and the hundredsof daily incidents and malfunctionsthat are absorbed by the system. How-ever, using those occurrences in asafety formula would not only dis-courage their reporting and investi-gation, but also create a bias againstthe most forthright carriers. This leavesaccident rates as an imperfect but credible mea-sure of safety performance.

A variety of accident rates have been used: permillion aircraft miles; per 100 million revenuepassenger kilometers; per 100,000 departures; per100,000 hours, etc. Most of them are expressed indecimal fractions that do not convey a clear pic-ture of year-to-year changes. A more informativeyardstick would be the use of hours flown peraccident, which is nothing but a variation of therate per 100,000 hours. To convert such a rate intothe number of hours flown per accident, divide100,000 by the rate. For example, an accident rateof .05 per 100,000 is the equivalent of 100,000 = 2million hours per accident. The same method isused to calculate the number of departures peraccident.

A pivotal asset of the air transport industry is thepublic’s trust in the safety of the system. Thismeans that the industry is best served by a perfor-mance yardstick that is readily understood by the

public. Expressing safety in hours flown (or thenumber of departures) per accident would satisfythat requirement. This leaves one question open:Should all accidents be considered or just the fatalones?

The conspicuity of fatal accidents creates headlinesthat shape the public’s perception of aviation safety,and public concern about being in a fatal accidentshould be approached head-on. The most visual methodto accomplish this is to publish the number of hours

flown or departures per fatal accident.

According to a broad interpretation ofInternational Civil Aviation Organiza-tion (ICAO) definitions, a fatal acci-dent involves the death of any personas a result of the operation of an air-craft while there was intent for flight.The inconsistency in the application ofthis definition is easily demonstrated.A refueling mishap that results in afire and the death of the refueler whilethe aircraft is at the gate with passen-gers or crew members aboard is treatedas a fatal aircraft accident. When thesame mishap occurs in association withperiodic maintenance it is not consid-ered a fatal aircraft accident, although

the catastrophic potential is comparable and ofconcern to the traveling public.

The same definition may skew the accident statis-tics by not excluding work-related accidents thatdo not threaten the safety of the aircraft. Someexamples are propeller-to-person accidents involvingground personnel and fatal injuries associated withactivities at the gate, such as the towing or push-back of aircraft. These industrial mishaps shouldnot become part of the fatal aircraft accident sta-tistics just because they happened while there wasintent for flight. Once they go into the statisticalblender, these accidents assume the same weightas the collision of two Boeing 747s in Tenerife,Canary Islands, and distort the true risks of flying.

The U.S. air carrier accident experience presentedhere is derived from annual statistics compiled bythe U.S. National Transportation Safety Board(NTSB). Exposure data (hours flown and the num-ber of departures) were provided to the NTSB bythe U.S. Federal Aviation Administration (FAA).

A pivotalasset of the air

transport

industry is thepublic’s trust

in the safety of

the system.


All fatal accidents in Part 121 operations are usedin the calculations, including suicide and sabotageaccidents. No distinction is made between fixed-wing and rotary-wing aircraft, passenger or cargoservice, scheduled and non-scheduled operations,revenue and non-revenue flights.

The beginning of the first full year (1960) of jetoperations was the starting point for this analysis.The 33-year period 1960-1992 produced a total of240 fatal accidents. Of these, 16 were not used inthe calculations because they were more properlycharacterized as industrial accidents. These single-fatality mishaps include:

• Eight ground crew deaths during the tow-ing or push-back of aircraft;

• Three propeller-to-person strikes;

• Two ground crew deaths during mainte-nance work at the gate;

• One stowaway;

• One infant, accompanied by parents, strangledby a seatbelt; and,

• One passenger who fell from a truck-mounted platform during a maintenancedelay. His presence on that platform wasnot authorized.

Figure 1 is based on the hours flown per fatalaccident, using the averages of the 11 three-yearperiods between 1960 and 1992. The graph showsa constantly improving trend with the exceptionof the 1987-1989 period. The hours flown perfatal accident increased by a factor of 10 betweenthe first and the last three-year periods.

Since the introduction of deregulation (1978-1980), the hours flown per fatal accident in-creased from 1,649,000 to 3,627,000 during the1990-1992 period.

The reason for the 1987-1989 dip in the graph isthe unusually high number (eight) of fatalaccidents in 1989. This was the highest annual

All Operations Under U.S. Federal Aviation Regulations Part 121 (Airlines)And Scheduled Services Under Part 135 (Commuters)

1960-1992 (Three-year Averages)Hours Flown per Fatal Accident

Ho

urs

Flo

wn

pe

r F

ata

l A

cc

ide

nt

Time PeriodSource: Gerard M. Bruggink/U.S. National Transportation Safety Board

Figure 1

4.0 x 106

3.0 x 106

2.0 x 106

1.0 x 106

0

394,000


number of fatal accidents since 1974. Althoughdifferent combinations of three-year averages wouldmodify the shape of the graph to some extent, theperiod with 1989 in it would still have a dip.

With the introduction of the airline hub-and-spokesystem in the 1980s, many travelers became depen-dent on commuter air carriers operating under theprovisions of Part 135. Because the commuters arenow a vital element of the U.S. air transport system,it is appropriate to compare their safety performancewith that of Part 121 operators. Figure 1 (page 11)shows that there is a striking difference between thetwo operations during the past 15 years.

When the comparison is based on the number ofdepartures per fatal accident (Figure 2), the differ-ence becomes less extreme. Additional details aboutthe accident experience in the two forms of publictransportation during the 1990-1992 period are givenin Table 1. The conclusions are that the U.S. airtransport system operates at two distinct levels ofsafety; and that the hub-and spoke system offersthe public little or no choice in the matter. Thisdisparity has much to do with the lower regulatorystandards of Part 135.

The first accident linked to deregulation occurredin January 1982, when a Boeing 737 struck a bridge

in Washington, D.C., following a takeoff withice-contaminated wings and a thrust deficiency. Inits analysis of that accident, the NTSB did not citederegulation directly, but implicated its role. TheNTSB said the captain “missed the seasoning ex-perience, normally gained as a first officer, as aresult of the rapid expansion of (the carrier).” The

Departures per Fatal Accident1978-1990

De

pa

rtu

res

pe

r F

ata

l A

cc

ide

nt

Time Period

Figure 2

Table 1Exposure and Fatal Accident Data

Part 121 (Airlines) and Part 135(Commuters) Operations

1990-1992Part 121 Part 135

Hours Flown 36,273,000 6,689,000Departures 23,976,000 8,759,000Fatal Accidents 10 17Hours Flown per Fatal Accident 3,627,000 394,000*Departures per Fatal Accident 2,398,000 515,000Fatalities

Crew 19 22Passengers 73 80

* During the same time period, on-demand air taxisoperating under Part 135 flew 86,000 hours perfatal accident.

Source: Gerard M. Bruggink/U.S. NationalTransportation Safety Board

0

3.0 x 106

2.0 x 106

1.0 x 106

Source: Gerard M. Bruggink/U.S. National Transportation Safety Board


causal statement listed the crew’s “limited experi-ence ... in jet transport winter operations” as acontributing factor.

Before deregulation, it would have been unthinkablethat crew experience would be a factor in Part 121accidents. Carriers voluntarily went beyond the pre-scribed minimum qualification standards. However,these practices may have been compromised whenthe explosive increase in airline operations duringthe mid- and late-1980s required the massive hiringof ground and flying personnel. Two accidents in the1987-1989 period also support this view.

In November 1987, a McDonnell Douglas DC-9crashed at Denver Stapleton International Airportwhile taking off in a snowstorm. The NTSB attrib-uted the accident to contaminated wings and thefirst officer’s rapid takeoff rotation. In its find-ings, the NTSB stated: “Due to the relatively lowexperience levels of both crew members in theDC-9, the pairing of these pilots wasinappropriate.” The NTSB later madean unmistakable reference to the de-regulation factor: “The rapid growthof the aviation industry at a time whenfewer experienced pilots are in thework force has reduced the opportu-nity for a pilot to accumulate experi-ence before progressing to a positionof greater responsibility. This loss of‘seasoning’ has led to the assignmentof pilots who may not be operation-ally mature for positions previouslyoccupied by highly experienced pi-lots.”

The same theme runs through the B-737 accident at La Guardia Airport inNew York in September 1989, whichinvolved a rejected takeoff and a runway overrun.Although the official causal statement cites the captain’sfailures as the sole explanation for the accident, thereport and the safety recommendations leave no doubtabout the role played by the pairing of inexperiencedcrew members. The captain had about 140 hours as a737 captain; the first officer, who was making thenight takeoff, had 8.2 hours in the 737. This was alsohis first takeoff after 39 days of not flying.

Considering the scrupulous attention given to flightcrew experience in accident investigation, it should

come as no surprise that this is the only area wherethe influence of deregulation has been documented.In the equally important area of aircraft mainte-nance, the available data do not permit such anassessment despite the fact that six of the 16 fatalaccidents in 1987-1989 had maintenance implica-tions. Those implications include:

• The takeoff warning system was not op-erating in the two cases where the crewb e g a n t h e t a k e o f f i n t h e w r o n gconfiguration;

• There were two explosive decompressions,one caused by loss of a cargo door and theother caused by failure of the fuselage skin;

• A DC-9 crashed on takeoff after the cargodoor opened; and,

• There was one catastrophic engine failure.

There is reason to believe that theinflux of new and less experiencedpersonnel, combined with competi-tive pressures, may have compromisedestablished maintenance standards dur-ing those years.

The safety performance of the air-lines during the 1990-1992 periodcan best be appreciated by a re-view of the 10 fatal accidents thatoccurred (Table 2, page 14). Threeof them involved cargo aircraft. TheDC-6 crash in Guatemala is includedin the performance calculations be-cause it was listed as a Part 121operation.

The trespasser who was struck by a Boeing 727 ona runway in Phoenix, Arizona (Table 2), was apatient from a nearby mental institution. The cap-tain rejected the night takeoff at a speed of 105knots and returned to the ramp.

The 84-year-old person injured in a turbulenceaccident (Table 2) died 20 days later. The air-craft experienced clear air turbulence while fly-ing under the overhang of a thunderstorm. Theseatbelt sign was on but the instruction was notenforced.

Beforederegulation, it

would have beenunthinkable that

crew experience

would turn up asa factor in Part

121 accidents.


The most unusual aspect of the seven accidentsinvolving passenger operations is the number ofrunway collisions. There were three: one duringtakeoff and two while landing. They are a dis-turbing reminder that airports remain high-riskareas.

A DC-9 and a Fokker F-28 crashed on takeoffbecause of ice or snow contamination on thewings. The F-28 had been deiced twice, but 35minutes elapsed between the last deicing andtakeoff. The DC-9 was not deiced.

The DC-8 accident at Swanton, Ohio (Table 2),occurred when control was lost during a missedapproach and was attributed to the captain’s ap-parent spatial disorientation, which resulted from

physiological factors and/or a failed attitudeindicator.

The NTSB was unable to determine the cause(s) ofa Boeing 737 accident in Colorado Springs, Colo-rado (Table 2). Several possibilities were explored,including a meteorological phenomenon and a mal-function in the aircraft’s control system.

What distinguishes the 1990-1992 fatal accidentrecord from the preceding period is not just thelower number of accidents, but the absence ofdocumented deregulation and proven maintenancefactors. The contrast between the two periods istoo great to ascribe to chance, and suggests thequestion: Does deregulation no longer affect theairlines’ safety performance? A definitive answer

Table 2Fatal Accidents and Fatalities

U.S. Federal Aviation Regulations Part 121 Air Carriers — All Operations1990-1992

Date Location Operator Service Aircraft Passenger Crew Other Nature of Accident1/18/90 Atlanta, Ga. Eastern Passenger B-727 0 0 1 Runway collision with

general aviationaircraft duringlanding.

3/13/90 Phoenix, Ariz. Alaska Passenger B-727 0 0 1 Aircraft strucktrespasser onrunway duringtakeoff.

5/5/90 Guatemala Translados Cargo DC-6 0 3 24 Crashed intoresidential area aftertakeoff.

10/3/90 Atlantic Ocean Eastern Passenger DC-9 1 0 0 Passenger died asresult of injuriesduring inflightturbulence.

12/3/90 Detroit, Mich. Northwest Passenger DC-9 7 1 0 Runway collision infog. One aircrafttaking off.

2/1/91 Los Angeles, Calif. USAir Passenger B-737 20 2 12 Runway collision withMetroliner afterlanding.

2/17/91 Cleveland, Ohio Ryan Cargo DC-9 0 2 0 Crashed on takeoff(contaminatedwings).

3/3/91 Colorado Springs, United Passenger B-737 20 5 0 Crashed onapproach (causeundetermined).

2/15/92 Swanton, Ohio Air Transport Cargo DC-8 0 4 0 Crashed duringmissed approach.

3/23/92 Flushing, NY USAir Passenger F-28 25 2 0 Crashed on takeoff(contaminatedwings).

Totals 73 19 38

Source: Gerard M. Bruggink/U.S. National Transportation Safety Board

Northwest Passenger B-727 0 0 0

International

Colo.


Although the

1990-1992 safety

record givesreasons for

satisfaction, there

can be noguarantee of a

similar or better

performance inthe future.

cannot be given because it will never be knownwhat the safety level might have been withoutderegulation. Nevertheless, there is a tentative ex-planation for the superior 1990-1992 performance.Using the number of departures in successivethree-year periods as a measure of growth sincethe start of deregulation, the diminishing rate ofexpansion in 1990-1992 becomes apparent:

Total Number of Departures1978-1980 16,327,0001981-1983 16,370,0001984-1986 19,408,0001987-1989 22,963,0001990-1992 23,976,000

In the 1990-1992 period, the number of departuresgrew by one million; in the two preceding periods,departures grew by three and 3.5 million respec-tively, which explains the massive hirings duringthose years. A reasonable conclusionis that by 1990-1992, the work forcehad become more experienced whilea more stable productivity tempo im-proved the harmony between job de-mands and skill levels.

With the widespread introduction ofthe jet transport in the early 1960s,the industry launched a vehicle thatwill still be dominating the skies overthe continents long after its super-sonic offspring have taken over theocean routes. No amount of technol-ogy can circumvent the laws of physicsthat set a practical limit on the speedsthat can be achieved economicallywithout sonic impact.

The accident record of the first-generation jets attests to the adjust-ment problems faced by manufacturers, airlinesand the FAA. It was a costly learning process thateased the introduction of subsequent generationsof jet aircraft, starting in the early 1970s. Theirincreasing reliability and maintainability produceda safety milestone in the mid-1970s, when thenumber of flying hours per fatal accident exceededone million for the first time.

At about the same time it became apparent that air-plane performance was maturing at a faster rate than

the performance of the humans in the system. TheNTSB’s sharpened focus on the human factors rolein accidents created new insights in the preventivepotential of training concepts, procedures and engi-neering solutions to problems such as controlled-flight-into-terrain (CFIT) accidents. The beneficialeffects of subsequent improvements in these catego-ries are evident in Figure 1 (page 3). By the mid-1980s, the number of hours flown per fatal accidentrose to 2.6 million.

The current maturity level of the subsonic transportis unquestionably the most constant contributor tothe industry’s safety performance, provided there iscompatibility between the competence of the workforce and the demands of the machine. True har-mony in that regard may not be achievable until aparticular make and model has become a stable fix-ture in the inventory of an airline. It would be un-wise to unsettle a developing competence level with

the rash introduction of technologiesthat increase the error potential in main-tenance and operations for the sake ofa fractional gain in airspeed or fuelconsumption. Nor should technologi-cal hubris be allowed to override theproven concepts of redundancy that haveserved the industry so well in the past.

Although the 1990-1992 safety recordgives reasons for satisfaction, there canbe no guarantee of a similar or betterperformance in the future. The elementof chance in the causation and preven-tion of accidents provides ample rea-son to temper confidence with wariness.Moreover, there is no predictable rela-tionship between the nature of an erroror compromise and the severity of itsconsequences. For example:

• In 1977, the seemingly innocuous term“O.K.,” followed by a short pause, wasinterpreted by a go-minded 747 captain inTenerife, Canary Islands, as confirmationof his (false) assumption that he was clearedfor takeoff; 574 people died in the Tenerifecollision; and,

• Recently, a deadheading captain prevented apotential disaster when he told a cabin atten-dant to warn the flight crew that the flight


spoilers were up. At the time, the flight wasnext in line for takeoff and the speed brakehandle was down and in the detent. The crewtaxied back to the gate where it was discov-ered that the speed brake control cable hadfailed.

Another factor that cannot be taken for granted isthe public’s trust in the safety of air travel. As longas the fear of flying is dormant, the public seemsto be more concerned with fares than the caliber ofa carrier and its equipment. Nevertheless, the public’strust can easily change into avoidance behaviorwhen a series of accidents draws widespread nega-tive attention to a particular operator, aircraft typeor type of operation. The FAA’s stamp of approvalon an aircraft or an operation has no commercialvalue unless the public also approves.

What brightens the outlook of the future is anapparent commitment to excellence at the workinglevel that goes beyond pay scale and contract.During the past three years, an average of 21,900flights a day were launched with unprecedentedresults in departures and hours flown per fatalaccident. Perfunctory task performance alone couldnot have accomplished this.

This overview of the safety performance of Part121 operators has highlighted some of the lessobvious elements of success in accident avoid-ance. Two of these are worth repeating:

• To capitalize on the safety potential of themature jet transport, the industry must pro-vide stability in the work environment; and,

• To meet its safety obligations, the industrymust establish experience criteria that

exceed the certification and qualification stan-dards in the FAR.

There are no reasons to suspect that the validity ofthese findings suffers from the acknowledged short-comings of basing safety performance on fatal ac-cident data. ♦

About the Author

Gerard M. Bruggink has been an aviation consultantsince his retirement in 1979 from the U.S. NationalTransportation Safety Board (NTSB). He haspublished more than 25 papers on aviation safety.

During his 10 years at the NTSB, Bruggink servedas field investigator, chief of the Human FactorsBranch, accident inquiry manager and deputy directorof the NTSB’s Bureau of Accident Investigation.He was an air safety investigator for the U.S.Army Safety Center and chief of the Human FactorsBranch at the Aviation Crash Injury Research Divisionof the Flight Safety Foundation. He was also aflight instructor, safety officer and accidentinvestigator for the U.S. Air Force and U.S. Armycontract flight schools.

Prior to immigrating to the United States, Brugginkwas a flight instructor, safety officer and accidentinvestigator for the Netherlands Air Force. A nativeof the Netherlands, he was a fighter pilot duringWorld War II. He saw action in the Pacific theaterand was interned as a prisoner of war.

Bruggink has been honored with the James H.McClellan Award from the Army Aviation Associationof America and the Jerome Lederer Award fromthe International Society of Air Safety Investigators.


Commercial Jet Fleet AccidentStatistics Reviewed

byEditorial Staff

Aviation Statistics

The Statistical Summary of Commercial Jet Air-craft Accidents, Worldwide Operations 1959-1992,published by the Boeing Commercial Airplane Group,Airplane Safety Engineering, and Commercial JetTransport Safety Statistics 1958-1992, publishedby the Douglas Aircraft Safety Data Office, FlightStandards and Safety Group, provide a wide spec-trum of statistical data relating to worldwide com-mercial jet fleet operations.

The Boeing report tracks nine manufacturers ofcommercial jet aircraft — 26 significant types (eightBoeing) and 10,916 aircraft in service as of Dec.31, 1992 (6,204 Boeing).

The Douglas statistics relate to the following air-craft: Airbus models A300/310/320, BAe (BAC) One-Eleven, BAe-146, Boeing models 707, 720, 727,737, 747, 757, 767, McDonnell Douglas models DC-8/9/10, MD-11, MD-80, Lockheed L-1011, and Fok-ker models F-28 and F-100.

Accidents serious enough to warrant attention inthe Douglas report amount to only one percent of

the 184,000 safety-related events tracked in theDouglas Safety Information System data base, whichincludes information from the U.S. National Trans-portation Safety Board (NTSB), the InternationalCivil Aviation Organization (ICAO),Airclaims MajorLoss Record, the U.K. Civil Aviation Authority,Flight Safety Foundation and local and nationalnews sources.

The Boeing summary of all accidents from 1959 to1992 (Figure 1, page 18) categorizes the worldwidecommercial jet fleet as follows:

Second generation: Boeing 727, Trident, BAeVC-10, BAe (BAC) One-Eleven, DC-9, 737-100/200, Fokker F-28;

Wide body (early): Boeing 747-100/200/300, DC-10, L-1011, A300; and,

New types: MD-11, MD-80, Boeing 737-300/400/500, 747-400, 757, 767, A310/320, BAe-146, andFokker F-100.


Source: Douglas Aircraft Company

All Accidents*Worldwide Commercial Jet Fleet — By Generic Group

Source: Boeing Commercial Airplane Group

Total Fleet and Accident Counts

Figure 2

B-737-100/200 (30)

-

B-737-300/400/500 (24)

B-737-100/200 (109)

B-747-400 (3)

B-747-100/200/300 (112)

B-757 (8)

B-767 (11)

DC-8 (133)

DC-9 (109) MD-80 (15)

DC-10 (53)

A300 (17), A310 (6), A320 (6)

BAe-146 (9),BAe (BAC) One-Eleven (52)

L-1011 (24)

F-100 (2), F-28 (55)

B-707/720 (309)

B-727 (182)

Figure 1

B-727 (45)

MD-80 (3)DC-9 (42)

DC-8 (43)

B-767 (1)

B-737-300/400/500 (6)

B-747 (15)

B-707/720 (72)

F-28 (18)

L-1011 (4)BAE-146 (1),

BAe (BAC) One-Eleven (11)

DC-10 (13)

B-727 (53)

B-737-100/200 (43)

B-737-300/400/500 (6)

B-747 (13)

B-767 (1)

DC-8 (60)

DC-9 (59)MD-80 (4)

DC-10 (16)A300 (5)

A320 (3)A310 (1)


BAE-146 (1),

L-1011 (4)

F-28 (27)

B-707/720 (100)

B-737-100/200 (1,021)

B-737-300/400/500 (1,235)

B-747-100/200/300 (678)

B-747-400 (205)

B-757 (485)

B-767 (457)

DC-8 (286)

DC-9 (847)

MD-80 (1,034)DC-10 (366)

MD-11 (76)A300 (363)

A310 (214)

A320 (349)


BAE-146 (188)

L-1011 (225)

F-28 (193)F-100 (114)

B-707/720 (230)

B-727 (1,575)

MANUFACTURER COUNT PCTBOEING 5,886 57.1%DOUGLAS 2,609 25.3%AIRBUS 926 8.9%LOCKHEED 225 2.2%BAe 360 3.5%FOKKER 307 3.0%

A300 (3)

A310 (1)A320 (3)

MD-11, F-100, B-757, B-747-400 AND MD-11 HAVE NO HULL LOSSES. MD-11, F-100 AND B-747-400 HAVE NO FATAL ACCIDENTS.

BAe (BAC) One-Eleven

F-28

1,239

10,313


B-707/720DC-8

B-727B-737-100/200

BAe (BAC) One-ElevenDC-9F-28

B-747-100/200/300DC-10A300

L-1011

BAe-146B-737-300/400/500

MD-80B-757F-100A320

MD-11A310

B-767B-747-400

The annual accident rate for second-generationaircraft, which peaked in 1965 at about 14.6 permillion departures, fell to less than four in 1969and has never risen above this figure. Wide bod-ies reached a peak of 19 in 1971, fell and thenincreased to about 9.5 in 1975, and have neverrisen above this rate again. Rates for new-typeaircraft climbed to about 12 from 1980 to 1981,then fell sharply to zero in 1982, after whichthey hovered about one or two.

Of a total of 1,239 accidents, 309 involved Boeing707/720s, 182 involved Boeing 727s, 133 involvedDC-8s, 112 involved Boeing 747-100s, 200s and300s, 109 involved DC-9s, and 109 involved Boeing737 100s and 200s, according to the Douglas reportfindings from 1958-1992 (Figure 2, page 18).

Of the worldwide commercial jet transport air-craft in service at year-end 1992, Boeing and

Accident RatesCommercial Jet Transport Aircraft

1960-1992


Figure 3

B-707/720 34 524,478DC-8 21 937,865

B-727 33 11,426,825B-737-100/200 45 10,725,255BAe (BAC) One-Eleven 6 904,818DC-9 9 8,548,491F-28 8 2,007,420B-747-100/200/300 3 2,650,057DC-10 4 1,967,675A300 1 2,082,913L-1011 1 1,194,372

BAe-146 1 1,286,971B-737-300/400/500 5 7,090,273MD-80 2 6,614,237B-757 0 1,986,332F-100 0 337,174A320 3 843,228MD-11 0 37,123A310 1 2,287,503B-767 1 1,967,675B-747-400 0 241,851

Douglas accounted for 57.1 percent and 25.3percent, respectively.

As for accidents per million departures from 1988 to1992 (Figure 3), 64.83 are attributed to the Boeing707/720 (narrow body), 22.39 to the DC-8 (narrowbody), 12.4 to the 747-400 (wide body), 9.95 to theBAe (BAC) One-Eleven (narrow body), and 9.81 toBoeing 747 models 100, 200 and 300 (wide body).The peak in accidents per million departures from1960-1992 occurred about 1961 (56), but has re-mained near 5.0 since 1969.

In terms of hull-loss accidents (Figure 4, page 20),Boeing’s statistics showed second-generation ac-cident rates from 1959-1992 peaking at about sixper million departures in 1965, then dropping andremaining at about one or two. Wide-body acci-dent rates reached a high of about 4.5 in 1972,then fell and rose, but never more than four.

LAST 10 YEARS


Hull-loss Accidents*Worldwide Commercial Jet Fleet — 1959-1992


Figure 5

Figure 4


Hull-loss Accidents*Worldwide Commercial Jet — By Generic Group


--

MD-80MD-11

Convair 880/990


F-28

MD-80

MD-11

BAe-146

DC-9737-100/200F-28

DC-10L-1011A300

BAe-146F-100


B-707/720DC-8

B-727B-737-100/200


B-747-100/200/300DC-10A300

L-1011

BAe-146B-737-300/400/500

MD-80B-757F-100A320

MD-11A310

B-767B-747-400

B-707/720 18 524,478DC-8 10 937,865


BAe-146 1 1,286,971B-737-300/400/500 5 7,090,273MD-80 2 6,614,237B-757 0 1,986,332F-100 0 337,174A320 3 843,228MD-11 0 37,123A310 1 2,287,503B-767 1 1,967,675B-747-400 0 241,851

New types jumped from zero to 12 in the 1980-1981 period, then dropped to zero in 1982, andonly started to climb slightly after 1986.

An examination of airplane type showed the CometIV responsible for 9.63 accidents per million de-partures from 1959-1992 and the Convair 880/990, 9.06 (Figure 5, page 20). Rates for all othertypes were no greater than 6.03. Note that theBoeing report tracks airplane types not mentionedin the Douglas report, e.g., Comet IV and Caravelle.

A breakdown of the U.S. commercial jet fleet duringthe past decade produced the following numbers:16.30 accidents per million departures for theBoeing 707/720 and 3.02 for the DC-8 (less than twofor all other types). For the same period, non-U.S.commercial jet fleet, the Boeing 707/720 again hadthe most accidents per million departures, at 31.72;the DC-8 had 16.76, the Caravelle 14.93, the Trident18.99 (all others were less than five).

Hull-loss accidents from 1958-1992 totaled 419according to the Douglas report (Figure 2, page18) — 100 attributed to the Boeing 707/720; 60 tothe DC-8; 59 to the DC-9; and 53 to the Boeing727. The Boeing graph on hull-loss accidents permillion departures from 1988-1992 (Figure 6) putthe count at 34.32 for the Boeing 707/720, 10.66for DC-8, and 6.63 for BAe (BAC) One-Eleven(narrow body). For all other aircraft, the numberof accidents was less than four. Again, 1961 wasthe “worst” year in the 1960-1992 span, with about19 accidents per million departures, but after 1963the number never rose above five and during thelast 10 years it never passed two.

Fatal accident numbers were low for the 1959-1992 period studied in the Boeing report (Figure7, page 22). Second generation jumped from zeroto six from 1964-1965, then dropped and remainedat about one from 1970 to 1992. Wide-body num-bers showed more peaks and valleys, but never

Figure 6

Hull-loss Accident Rates — Commercial Jet Transport Aircraft1960-1992



B-707/720DC-8

B-727B-737-100/200


B-747-100/200/300DC-10A300

L-1011

BAe-146B-737-300/400/500

MD-80B-757F-100A320

MD-11A310

B-767

Fatal Accident Rates — Commercial Jet Transport Aircraft1960-1992


Figure 8

Fatal Accidents *Worldwide Commercial Jet Fleet — By Generic Group


Figure 7

--

MD-11

F-28


MD-80

BAe-146F-100

B-707/720 34 524,478DC-8 21 937,865


BAe-146 1 1,286,971B-737-300/400/500 5 7,090,273MD-80 2 6,614,237B-757 0 1,986,332F-100 0 337,174A320 3 843,228MD-11 0 37,123A310 1 2,287,503B-767 1 1,967,675B-747-400 0 241,851


extended past 5.5; and new types, with a high of abit over 12 in 1981, dropped to zero for 1982-1986and never climbed past one afterward.

According to the Douglas report, fatal accidents in theperiod 1958 to 1992 numbered 311 (Figure 2). Of thistotal, 72 were attributed to Boeing 707/720s, 45 toBoeing 727s; 43 to DC-8s; and 42 to DC-9s. Fatalaccidents per million departures from 1988-1992 were17.16 for Boeing 707/720s; for every other aircraft typethe number was less than five (Figure 8, page 22).

Fatal accident rates on commercial jet transport air-craft taken as a whole from 1960-1992 totaled about14 per million departures in 1961, then fell sharplyand rose to about five in the 1965-1966 period. Theydropped to two or less for 1967 to 1992 (Figure 8).

Accident and damage in the two reports were de-fined as follows:

An aircraft accident is an event associated withaircraft operation occurring between the time peopleboard an aircraft with the intention of flight untilall people have disembarked. It is considered anaircraft accident if, during this time, any persondies or is seriously injured as the result of beingin, on, or in direct contact with the aircraft or if theaircraft receives substantial damage. The Douglasreport excludes death from natural causes, fatal orserious injury to a passenger that is self-inflicted

or inflicted by another, injury to ground supportpersonnel before or after flight, and serious injuryinvolving stowaways or that is not a direct resultof aircraft operation.

Substantial damage affects structural strength, per-formance or flight characteristics of the aircraft thatwould require major repair or replacement of theaffected component. Engine failure limited to oneengine, bent fairings or cowlings, dents in the skin,damage to landing gear, wheels, tires, flaps, etc., arenot considered substantial damage.

Serious injuries are defined as requiring hospi-talization for more than 48 hours, beginning withinseven days of the injury; resulting in a fracture(except simple fractures of fingers, toes or nose);producing lacerations causing severe bleedingor nerve, muscle or tendon damage; involvinginjury to an internal organ or involving second-or third-degree burns over more than 5 percentof the body.

A fatal injury as defined results in death within 30days as a result of the accident. A hostile actionincludes sabotage, military action, suicide and ter-rorist acts that cause the accidents/injuries. U.S.and non-U.S. events are identified by the operator’snational registry, not the event’s location. A non-operational event occurs when there is no intentfor flight.♦


Publications Received at FSFJerry Lederer Aviation Safety Library

U.S. Guidelines Issued for Use of PortableElectronic Devices Aboard Aircraft

byEditorial Staff

New Reference Materials

U.S. Federal Aviation Administration. AdvisoryCircular 91.21-1, Use of Portable Electronic De-vices Aboard Aircraft. August 1993. 5 p.

This advisory circular (AC) provides informationand guidance to assist aircraft operators with com-pliance of Federal Aviation Regulations (FAR) Section91.21, prohibiting the operation of portable elec-tronic devices (PED) aboard U.S.-registered civilaircraft operated by a holder of an air carrier oper-ating certificate or any other aircraft while operat-ing under instrument flight rules.

Section 91.21 allows the operation of PEDs thatthe operator of the aircraft has determined willnot cause interference with the navigation orcommunication systems of the aircraft. The ACrecommends minimum procedures to provide meth-ods to inform passengers of permissible PEDoperation times, procedures to terminate the op-eration of the PEDs, procedures for reportinginstances of suspected or confirmed interferenceby PEDs, cockpit-to-cabin coordination, and pro-cedures for determining acceptability of thosePEDs that can safely be operated aboard theaircraft.

In addition, the AC recommends prohibiting theoperation of any portable electronic device duringthe takeoff and landing phases of flight. This ACcancels AC 91-47, Use of Portable ElectronicDevices—Radio Receivers, dated March 23, 1977.[Modified Purpose]

U.S. Federal Aviation Administration. AdvisoryCircular 21-36, Quality Assurance Controls forProduct Acceptance Software. August 1993. 6 p.

This advisory circular (AC) provides informa-tion and guidance concerning control of soft-ware, documentation and related digital input/output data designed for use in the acceptanceof airborne products. This AC addresses onlythose sections of Federal Aviation Regulations(FAR) Part 21, subparts F, G, K, and O whereinformation on computer-aided manufacturing(CAM), computer-aided inspection (CAI) andcomputer-aided test (CAT) software would behelpful to production approval holders (PAH)and parts suppliers during manufacture, inspec-tion and/or testing of materials and parts. [ModifiedPurpose]

Reports

U.S. General Accounting Office (GAO). AircraftCertification: New FAA Approach Needed to MeetChallenges of Advanced Technology. Report to theChairman, Subcommittee on Aviation, Committeeon Public Works and Transportation, U.S. Houseof Representatives. September 1993. 75p. Avail-able through the GAO**.

Keywords1. Airlines — Certification — United States.2. Airplanes — United States — Inspection.3. Aeronautics — United States — Technological

Innovations.


Summary: This report examines the U.S. FederalAviation Administration’s (FAA) process for certi-fying that transport aircraft designs meet safetystandards. Because the FAA is responsible for cer-tifying that new aircraft designs and systems meetsafety standards, it is crucial that FAA staff under-stand these new technologies.

The GAO was asked to examine whether FAAstaff are effectively involved in the aircraft cer-tification process and if they are provided theassistance and training needed to be competentin the increasingly complex technologies usedaboard the latest aircraft designs. According tothe report, the FAA has increased delegationduring the last 13 years, and its ability to effec-tively oversee or add value to the certificationprocess as well as understand new technologieshas been questioned by internal reviews and FAAand industry officials.

The report states that the FAA now delegates up to95 percent of certification activities to manufac-turers without defining critical activities in whichthe FAA should be involved. Moreover, the FAAhas not provided guidance on the necessary leveland quality of the oversight of designees. As aresult, FAA staff no longer conduct and overseesuch critical activities as the approval of test plansand analyses of hypothetical failures of systems.The report said that the FAA has not provided itsstaff the assistance and training needed to ensure

competence in new technologies. Although the FAAhas hired experts to assist staff in the certificationprocess, it has not identified critical points in thecertification process that require experts’ involve-ment, according to the report.

In addition, the report said the FAA’s training hasnot kept pace with technological advancements.The GAO found that only one of the 12 FAA engi-neers responsible for approving aircraft softwareattended a software-related training course betweenfiscal years 1990 and 1992. And although the FAAis developing a new training program, it may nothave the structure necessary to improve the staff’scompetence, the report states.

To ensure that FAA staff receive the technical trainingneeded, the GAO recommended that the secretaryof transportation direct the FAA to establish spe-cific training requirements for each certificationdiscipline. And to ensure that each staff membermeets those requirements, training should be keptas current as possible by identifying the training innew technologies that is available at universities,other government agencies and private industry,the report said. ♦

**U.S. General Accounting Office (GAO)Post Office Box 6012Gaithersburg, MD 20877 U.S.Telephone: (202) 275-6241

Updated Reference Materials (Advisory Circulars, U.S. FAA)

AC Number Month/Year Subject

90-66A 08/26/93 Recommended Standard Traffic Patterns and Practices for Aero-nautical Operations at Airports without Operating Control Towers(cancels AC 90-66, dated Feb. 27, 1975).

Federal Aviation Administration Orders

7400.2D 09/16/93 Procedures for Handling Airspace Matters (incorporating all re-quired changes due to the Airspace Reclassification Rule and can-celling Order 7400.2C, dated May 1, 1984).


Accident/Incident Briefs

Speed Brake Cable MisroutingProblem Identified

byEditorial Staff

The U.S. National Transportation Safety Board(NTSB) said that after the incident, it researchedthe service history of speed brake down controlcable failures on Boeing 737 aircraft and deter-mined that it was the third failure attributed tomisrouting.

“Instead of being routed on the pulley adjacent tothe spoiler ratio changer, the cable was riding atoptwo spacers that attach cable guards to the pulley,”the NTSB said. “Wear from riding on these spac-ers caused the cable to fail 14 inches [35 centime-ters] from the spoiler ration changer.”

The NTSB noted that if the cable fails on theground after the aircraft lands, all 10 spoilerpanels will remain extended when the lever is placedin the down position. “Since the takeoff configura-tion warning is based on control lever position andnot actual spoiler position, the crew may beunaware that spoiler panels are extended whentaking off,” the NTSB said.

The NTSB recommended inspections of all Boe-ing 737 aircraft to make sure the down controlcable is correctly installed.

Antenna Strike Damages TupolevOn Missed Approach

Tupolev Tu-154. Substantial damage. No injuries.

The Tupolev with 136 passengers and a crew ofnine on board was on a daylight instrument land-ing system (ILS) approach when it encountered

The following information provides an awarenessof problems through which such occurrences maybe prevented in the future. Accident/incident briefsare based on preliminary information from gov-ernment agencies, aviation organizations, pressinformation and other sources. This informationmay not be entirely accurate.

Air Carrier

Speed Brake Cable Misrouted

Boeing 737-300. No damage. No injuries.

The aircraft was taxiing for takeoff when a passen-ger, an off-duty captain, notified the crew that thewing spoiler panels were extended.

The flight immediately returned to the gate for amaintenance inspection. The inspection deter-mined that the speed brake down control cablefailed because it had been misrouted. The downcable had failed at a point where it rides over apulley in the right main landing gear well. If thespeed brake down cable fails, the spoiler panelsare prevented from retracting when the speedbrake control level is moved into the downposition.


heavy showers and turbulence about two nauticalmiles (four kilometers) from touchdown.

With engine power momentarily set at flight idle,the aircraft descended below the glidepath. A missedapproach was initiated at an altitude of 98 feet (30meters) above ground level (AGL). However, theright inboard flaps struck an ILS antenna at 20 feet(six meters) AGL, located about 2,133 feet (650meters) before the runway threshold. The flapswere retracted and the aircraft landed without in-cident after a go-around.

Air TaxiCommuter

Powerline Snags CommuterOn Night Approach

L-410 Turbolet. Substantial damage. No injuries.

The Turbolet with 13 passengers and a crew offour on board was on final approach at night to aEuropean airport when the top of the vertical sta-bilizer fin struck a high-tension powerline.

The pilot was able to execute a missed approachand the aircraft landed without further incident.An investigation cited a lack of crew coordinationin the cockpit, inadequate monitoring and poorairmanship as causes of the incident.

Engine Fire Linked to Air Taxi Crash

Piper PA-23 Aztec. Aircraft destroyed. Fourfatalities.

The twin-engine Aztec with three passengers onboard was departing on a daylight flight in theCaribbean when, shortly after takeoff, the pilotadvised air traffic controllers that one enginewas on fire.

Controllers cleared the aircraft for an emergencylanding, but there was no further contact with thepilot. The aircraft crashed in the water just west ofthe airport, killing the pilot and passengers. Aninquiry determined that the pilot likely lost controlof the aircraft during the engine-out emergency attoo low an altitude to recover.

CorporateExecutive

Course Drift Dooms TwinOn Climb-out

Cessna 421. Four fatalities and two seriousinjuries.

The twin-engine Cessna with five passengers onboard was climbing out after a night takeoffwhen it struck trees and snow-covered terrain at10,000 feet (3,050 meters) mean sea level (MSL).

The intended course heading was 110 degrees.The accident site was located eight miles fromthe airport on a heading of 130 degrees. Theweather was reported 500 feet (152 meters) over-cast, visibility eight miles (13 kilometers). Thepilot, who survived the crash and post-impactfire with serious injuries, told investigators hewas attempting to open his flight plan when theaircraft began impacting tall trees. Four passen-gers were killed and a fifth was seriouslyinjured.

Fuel Exhaustion ShortensFinal Approach

Cessna 310. Aircraft destroyed. One serious andthree minor injuries.

The Cessna was on final approach when an enginestopped because of fuel exhaustion. A few mo-ments later, the second engine stopped.


The pilot executed an emergency landing about10,499 feet (3,202) meters before the runway thresh-old. The pilot was seriously injured. The threepassengers received minor injuries.

OtherGeneralAviation

Airspeed Lost on Final

Beech 55 Baron. Substantial damage. No injuries.

The pilot of the twin-engine Baron was attemptingto land on a short field with a line of trees about 131feet (40 meters) from the runway threshold. Therunway was about 2,625 feet (800 meters) long.

Because of the runway length, the pilot flew atreduced speed for a short-field landing. At the treeline, the aircraft stalled and struck the ground. Thepilot and four passengers escaped without injuries.An inquiry determined that the pilot had exercisedpoor judgment and airmanship during the approachto an uncertified landing field.

Training Flight Ends Abruptly

Cessna 152. Aircraft destroyed. Two serious injuries.

The student pilot was being instructed in patternflying and touch-and-go landings. At about 30 feet(9 meters) above the runway during the fourthapproach, the aircraft yawed to the left as it flaredfor landing.

The instructor immediately called “I have con-trol,” and applied right rudder and power whileattempting to push the control column forward.She was unable to move the control column andrealized that the student was still gripping it. Be-fore the instructor could correct the situation, the

aircraft stalled and impacted the runway in a nosedown attitude. There was no fire, and the injuredinstructor was able to help the more seriouslyinjured student evacuate the aircraft.

Rotorcraft

Mountain Crash Blamed onPoor Planning

Aerospatiale SA 365 Dauphin. Aircraft destroyed.Eight fatalities.

The helicopter with six passengers on board de-parted in visual meteorological conditions, flewinto clouds and collided with a mountain aboutfive miles (eight kilometers) from the airport afterfour minutes of flight.

The aircraft was destroyed during the post-impact fire. An investigation determined the pilothad selected the wrong departure sector and hadplanned the flight poorly.

Cable Failure ThreatensTail Rotor Control

Sikorsky S-76. Minor damage. No injuries.

The Sikorsky was in cruise flight when the pilotobserved a loss of left anti-torque control.

Despite the partial loss of tail rotor control, thepilot was able to continue to an airport where arunning landing was executed. A subsequent in-spection found that a section of the tail rotor con-trol cable had failed near where it contacts a pulley.There were no injuries. ♦

the u.s. air traffic control system wrestles with the

Documents